Obstet Gynecol Clin N Am

34 (2007) xvii–xviii

Foreword

William F. Rayburn, MD
Consulting Editor

This issue of the Obstetrics and Gynecology Clinics of North America,

guest edited by Henry Galan, MD, pertains to emergencies that can occur
in obstetrics and gynecology. An obstetrician-gynecologist may be con-
fronted with a sudden emergency at any time, either at the hospital or in
the outpatient setting. Prompt corrective action is necessary, whether it is
severe postpartum hemorrhage, acute chest or abdominal pain, or an ana-
phylactic reaction to an injection in the office. Preparing for an emergency
requires planning, provision of resources, awareness of early warning signs,
and specialized trainees who are aware of what to do in an emergency.
Certain emergencies, such as a massive pulmonary embolus or a complete
abruptio placentae, can be sudden and potentially catastrophic. Standard-
ized responses will increase the efficiency and quality of care. A protocol
should provide a full evaluation of the problem and clearly communicate
the patient care issue. Periodic drills may lead to a more standard response
with a favorable outcome.
Planning for potential emergency events such as anaphylactic shock or
cardiopulmonary resuscitation can be complex. At a minimum, it should
involve an assessment of suspected risks related to the underlying condition.
All physicians should be familiar with the ‘‘crash cart.’’ By placing necessary
items in one place, time is not lost in gathering supplies. A small kit can be
created for handling allergic reactions. As with a crash cart, this kit must be
maintained regularly to ensure that supplies are current.
It becomes clear with any emergency when to call for help. Activation of
a response team before a full arrest may lead to improved survival and less

0889-8545/07/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ogc.2007.08.005 obgyn.theclinics.com
xviii FOREWORD

need for an intensive care admission. Rapid correction of problems is

better met with a small emergency team whose members talk with each other
and share information. Although a leader must coordinate the response, all
members of the team should be empowered to practice together. By practicing
together, barriers hindering communication and teamwork can be overcome.
Adult learning theory, as described in this issue by its distinguished panel
of contributors, supports the value of experiential learning. Training can
entail a sophisticated simulated environment or a customary work space
with a mock event. Emergency drills allow physicians and others to practice
principles of effective communication in a crisis. Our desire is that this issue
will attract the attention of providers caring for those women at risk for
emergencies. Practical information provided herein will hopefully aid in
the development and implementation of more-specific and individualized
treatment plans.

William F. Rayburn, MD
Department of Obstetrics and Gynecology
University of New Mexico School of Medicine
MSC10 5580
1 University of New Mexico
Albuquerque, NM 87131-0001, USA
E-mail address: wrayburn@salud.unm.edu
 Obstet Gynecol Clin N Am
34 (2007) xix–xxi

Preface

Henry L. Galan, MD
Guest Editor

Every medical or surgical specialty has emergencies that are somewhat

specific to that specialty. This is also true in obstetrics and gynecology.
However, several characteristics set the specialty of Ob/Gyn apart from
all others. Not only can nearly all of the emergencies seen in other specialties
be seen in the field of Ob/Gyn, but pregnancy also brings a new and unique
dimension to emergency situations in our specialty. Three primary charac-
teristics of Ob/Gyn set it apart from other fields of medicine when it comes
to emergencies: (1) it is the only specialty committed completely to women;
(2) it is the only specialty in which a single emergent event can threaten the
lives of two individuals, the mother and her fetus; and (3) an otherwise com-
pletely healthy patient may succumb purely to a pregnancy-related compli-
cation. It is these three general themes that drive the topics in this issue of
the Obstetrics & Gynecology Clinics of North America.
The authors contributing to this issue were invited to cover topics that are
of particular interest to them and in which they are considered leaders. They
have utilized the best available evidence and their own experience to provide
the reader with knowledge of and guidance through these emergency condi-
tions. Considerable focus is given to the physiological changes in pregnan-
cies that impact emergency conditions.
Several of the articles in this issue are related to hemorrhage, which, be-
cause of the 600 cc/min uterine blood flow at term, can be massive. Gyamfi
and Berkowitz launch this issue by guiding us through the challenges of

0889-8545/07/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ogc.2007.08.004 obgyn.theclinics.com
xx PREFACE

caring for the Jehovah’s Witness patient who refuses the medically indicated
blood transfusion. Fuller and Bucklin provide the basics of blood product
transfusion and its application to the hemorrhaging patient. Teal and Mukul
review first-trimester bleeding, which itself can be massive and without the
benefit of having reached the full maternal expansion of blood volume seen
later in pregnancy. Monga and Kilpatrick address the physiologic and
physical changes of the abdomen and contents within related to pregnancy,
which are dramatic and impact the differential diagnosis, diagnostic proce-
dures, and thresholds for surgical exploration. Oyelese, Scorza, Mastrolia,
and Smulian provide guidelines for the management of postpartum hemor-
rhage, including the newer B-Lynch and Bakri balloon procedures, followed
by the expert descriptions by Banovac, Lin, Shah, White, Pelage, and Spies
of interventional radiologic approaches to hemorrhage.
Of all the obstetric-related emergencies, few match the profound mater-
nal cardiovascular collapse and disseminated intravascular coagulation of
amntiotic fluid embolism, which is discussed in depth by Sheffield and Staf-
ford. Gottlieb and I review risk factors and management of shoulder dysto-
cia, which most often rears itself in without warning and carries risk for
long-term fetal sequelae and medical-legal action. Muench and Canterino
thoroughly review catastrophic and noncatastrophic trauma in pregnancy
with emphasis on evaluation of the trauma patient and how physiologic
changes impact the evaluation. Gardner and Atta conclude the emergencies
articles with a review of cardiopulmonary resuscitation with a focus on the
effect of physiologic changes in pregnancy and which may be an end result
of any of the above-mentioned emergencies.
While not always presenting as acutely or urgently as some of the afore-
mentioned emergencies, several medical conditions and social circumstances
predispose pregnant patients to serious and life-threatening events. Guinn,
Abel, and Tomlinson provide information on sepsis, the leading cause of
death in the critically ill patient. Conway and Parker review the most serious
condition in the diabetic patient, diabetic ketoacidosis. Pregnancy is a known
thrombogenic state with great potential for adverse events; Lockwood and
Rosenberg guide the reader through thromboembolic disease. Gunter draws
our attention sharply to the prevalence, dangers, and the need for height-
ened awareness of domestic partner violence and provides us everyday tools
with which to address this problem in our office practice. This issue con-
cludes with an article by Shwayder reviewing the medical-legal implications
of obstetric emergencies and strategies for prevention of legal action in the
setting of an adverse event.
I would like to add a personal note of gratitude to all the gifted individ-
uals contributing to this issue of the Obstetrics & Gynecology Clinics of
North America and to Carla Holloway of Elsevier for her patience and pro-
fessionalism. Most of all, on behalf of my fellow authors, I would like to
thank our patients, students, nurses, and house staff, from whom we learn
so much about our beautiful specialty. This gift allows us to push the
 PREFACE xxi

frontiers of knowledge and provide the best care possible for the next mom
and unborn baby that we encounter.

Henry L. Galan, MD
Department of Obstetrics and Gynecology
University of Colorado at Denver Health Sciences Center
Academic Office 1, 12631 East 17th Avenue, Rm 4001
Aurora, CO 80045, USA
E-mail address: henry.galan@uchsc.edu
 Obstet Gynecol Clin N Am
34 (2007) 357–365

Management of Pregnancy
in a Jehovah’s Witness
Cynthia Gyamfi, MD*, Richard L. Berkowitz, MD
Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology,
Columbia University Medical Center, 622 West 168th Street,
PH-16, New York, NY 10032, USA

The refusal of blood products by Jehovah’s Witnesses makes this group

a unique obstetric population with the potential for disastrous perinatal out-
comes secondary to hemorrhage. Obstetric hemorrhage is the second leading
cause of maternal mortality in the United States after pulmonary embolism
[1]. Singla and colleagues [2] reported on maternal mortality amongst Jeho-
vah’s Witnesses who refuse all blood products. When this group develops an
obstetric hemorrhage, they have a 44-fold increased risk of death.
The care of these patients must be meticulously coordinated to achieve
good pregnancy outcomes. This involves coordination of care with the
patient’s primary care provider, maternal–fetal medicine specialist, anesthe-
siologist, and possibly other subspecialists to reduce perinatal morbidity and
mortality.
To provide comprehensive care to patients who are Jehovah’s Witnesses,
the care provider should understand the background of their belief system.
Charles Russell founded the group in 1872 in Pennsylvania [3]. Many of the
followers’ beliefs are based on literal translations of the Bible. Genesis 9 and
Leviticus 17 state that one cannot eat the blood of life; these passages are
interpreted to include the exchange of blood products [4]. For the Jehovah’s
Witness, receiving blood products may lead to excommunication and eternal
damnation [3], and an individual who offers to transfuse blood is considered
by many members of the sect to be acting through the devil’s influence. Un-
derstanding these facts is crucial when caring for patients who are Jehovah’s
Witnesses.

* Corresponding author.
E-mail address: cg2231@columbia.edu (C. Gyamfi).

0889-8545/07/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ogc.2007.06.005 obgyn.theclinics.com
358 GYAMFI & BERKOWITZ

Addressing the risk of hemorrhage

As the editors of Williams Obstetrics have reemphasized over many edi-
tions, ‘‘Obstetrics is ‘bloody business’!’’ [5]. The incidence of postpartum
hemorrhage is difficult to quantify because of varying definitions. However,
it has been estimated to occur in 4% of vaginal deliveries and 6% to 8% of
cesarean deliveries [5]. The need for blood transfusion is fairly common.
Klapholz [6] reported a 2% transfusion rate for women who delivered at
Beth Israel Hospital in 1986. Rouse and colleagues [7] reviewed over
23,000 primary cesarean deliveries and found that the rate of transfusion
in that population was 3.2%. Among patients with a previous cesarean
delivery, Landon and colleagues [8] found that transfusion was more likely
with a trial of labor than with an elective repeat cesarean, 1.7% versus 1.0%,
respectively (odds ratio: 1.71; 95% CI, 1.41–2.08, P!.001).
Because the risk of requiring blood transfusion is not negligible, the po-
tential for transfusion should be discussed with all obstetrical patients dur-
ing their prenatal care. The policy at Columbia University Medical Center is
to ask all new obstetrical patients whether they will accept a blood transfu-
sion in an emergency situation. Without specifically asking about religion,
this serves to open the dialog about transfusion and can identify patients
who hold fast to the beliefs of the Jehovah’s Witnesses.
The authors have previously shown that there are varying degrees of ad-
herence to the doctrine of blood refusal amongst Jehovah’s Witnesses [9]. In
a study of pregnant Jehovah’s Witnesses, almost 50% indicated, when a re-
view of health care proxies was undertaken, that they would accept some
form of blood or blood products [9]. This means that, rather than assuming
that a Jehovah’s Witness will not accept any blood products, the clinician
must inquire as to the specific beliefs of the individual patient. Strong famil-
ial and church pressures can influence a patient’s decision in the presence of
others. This is why it is important for the clinician to be alone with the pa-
tient when discussing her wishes. At the very minimum, the patient should
be asked about whether she will be willing to accept any or all of the follow-
ing: whole blood, fresh frozen plasma, cryoprecipitate, albumin, isolated
factor preparations, nonblood plasma expanders, hemodilution, and cell-
saver. At the authors’ institution, this inquiry is presented in the form of
a checklist, which is then signed by the patient and included in the patient’s
chart. Additionally, a statewide health care proxy is signed.

Prenatal care
For a variety of reasons, identification of a patient who will not accept
blood, and the discussion about which products, if any, she is willing to ac-
cept, should be undertaken at the first prenatal visit. First, most obstetric
patients are young and healthy and may not consider themselves to be at
risk to hemorrhage. It is important to explain to the patient what puts her
 MANAGEMENT OF PREGNANCY IN A JEHOVAH’S WITNESS 359

in this category. A discussion of the health care proxy and blood product
checklist requires extensive education because the average person is not
familiar with the terms ‘‘nonblood plasma expanders’’ or ‘‘cell-saver.’’ In
most cases the patient will want to discuss this with her family and/or
church leaders, so there will be a delay in signing the checklist. An early dis-
cussion allows the patient a chance to make an informed decision. Second,
identification and treatment of an existing anemia are very important in the
care of these patients. Because the treatment of anemia is a slow process,
aggressive early management may obviate the need for transfusion later.
Finally, a physician has to be both willing and able to allow a properly ed-
ucated patient to die once she has indicated that she prefers death over
transfusion. It is always difficult for a physician, who has been trained to
save lives, to accept a patient’s decision that can lead to her death. If a phy-
sician does not want to participate in the care of such a patient, she should
be transferred to the practice of a physician associated with a tertiary care
center, and consultation should be obtained with a maternal–fetal medicine
specialist. The transferring physician is obligated to ensure that another
physician has agreed to accept the patient. This may be difficult to arrange
in an emergency situation, so early transfer of the patient’s care is extremely
prudent.

Evaluation and treatment of anemia

When a Jehovah’s Witness presents for her first prenatal visit, a complete
blood count with platelets should be included in the routine prenatal labo-
ratory tests, and the patient should be started on iron and folic acid supple-
mentation. The goal should be to maintain her hematocrit above 40% [10].
Once that level has been achieved, a patient can sustain a 2-L peripartum
blood loss, and is unlikely to require transfusion. If the initial hematocrit
is below this level, a workup for potential causes of anemia should be initi-
ated. If iron deficiency is documented, the dose of iron supplementation can
be adjusted accordingly, and a stool softener should be prescribed. Iron is
best absorbed through the gastrointestinal tract in an acidic medium, so
vitamin C, or simply orange juice, should be taken along with the iron pills.
Foods high in heme content, such as meat, poultry, and fish, should
be encouraged [4]. Vegetarian diets are low in heme, and tannins found
in tea and phylates in bran can decrease the absorption of iron [11]; so it
is important to supplement this subgroup.
Many patients complain of constipation while taking iron supplementa-
tion. This can lead to noncompliance. An easy way to assess whether a pa-
tient is taking her iron supplements is to ask her about the color of her stool,
which should be markedly darker if iron is being consumed. One strategy to
encourage compliance is to prescribe a stool softener in addition to iron. In
women who cannot or will not take oral iron, parenteral iron is a reasonable
alternative. Intravenous iron has traditionally been discouraged because
360 GYAMFI & BERKOWITZ

iron dextran can lead to anaphylactoid reactions. Iron sucrose, however, is

considered a safer alternative, with hypersentivity reactions estimated at
0.005% compared with 0.2% to 3% for iron dextran [12]. A test dose is
not required before administration of iron sucrose, but it should not be con-
sidered the first-line agent for treatment of anemia because adverse drug
events other than hypersensitivity are common [12].
Erythropoietin may also be administered to an obstetrical patient with
a hematocrit of less than 40% who has not responded to iron supplementa-
tion [10]. Erythropoietin stimulates the bone marrow to maximize red blood
cell production. Recombinant erythropoietin is available either in the form
of epoetin alfa or darbepoetin alfa. Both of these drugs are erythropoesis-
stimulating agents (ESAs) that increase hemoglobin in a similar fashion.
Darbepoetin is more expensive, but can be dosed less frequently than epoe-
tin alfa [13]. ESAs should be stopped once the hemoglobin is greater than
12 g/dL because adverse cardiovascular events can occur above that level
[14]. Not all Jehovah’s Witnesses accept these medications because each is
packaged with 2.5 mL of albumin per dose. To help the patient make
an informed decision, a discussion should ensue about how the medication
works and how it is constituted.

Review blood products and their alternatives

Another key element in the initial prenatal visit is a comprehensive dis-
cussion about what blood products the patient may be willing to accept
and the available alternatives. As mentioned earlier, this conversation
should occur in the absence of outside influences that may alter the woman’s
responses. This is the appropriate time to review the checklist of blood and
blood products, described earlier, to see which of these, if any, is acceptable.
Next, a discussion of autologous blood donation should ensue [4]. Autol-
ogous blood donation involves optimizing the patient’s hematocrit with oral
iron supplementation (or erythropoietin, if this is acceptable) [4] and then
having her donate her own blood at least 72 hours (but ideally, 2 weeks) be-
fore elective cesarean delivery or the estimated date of confinement. After
appropriate testing, the blood is stored and held for the patient. It will be
discarded if not used at the time of delivery [15]. This process is somewhat
tedious, but if the patient is willing to accept her own blood, it could be life-
saving [15].
In addition to allogenic blood or blood products, other options should
also be discussed with the patient. Cell salvage systems can be employed
as a form of intraoperative autologous blood donation [4,16]. Cell-saver sys-
tems allow for free blood in the abdomen to be aspirated, filtered, and then
reinfused into the patient perioperatively [16]. Such systems use centrifugal
cell separators that segregate the red cells from the plasma, wash the red
cells with normal saline, and prepare them for reinfusion. Clotting is pre-
vented by using a double-lumen tube with one lumen providing suction
 MANAGEMENT OF PREGNANCY IN A JEHOVAH’S WITNESS 361

and the other providing a constant flow of anticoagulant [16]. Using a cell-
saver system during a cesarean delivery carries the potential risk that fetal
cells, amniotic fluid, and debris may enter the maternal circulation if they
are not properly filtered by the system, theoretically predisposing the patient
to amniotic fluid embolism (AFE) [17]. However, researchers have shown
that the filtration system used by these devices can limit the amount of par-
ticulate matter in the blood to be reinfused to a concentration equal to that
of maternal venous blood [18–20].
Although the use of cell salvage systems has been shown to be safe and
potentially life-saving, they are unfortunately still underused in obstetrics
because of the theoretical risk of AFE [18,21,22]. The obstetric literature
contains hundreds of cases where a cell-saver system was used safely [22],
and an American College of Obstetrics and Gynecology (ACOG) technical
bulletin advocates the use of these systems during cesarean delivery associ-
ated with major hemorrhage such as that which occurs with placenta accreta
[21]. An extensive MEDLINE search from 1966 to the present using the key
words ‘‘cell salvage,’’ ‘‘cell saver,’’ ‘‘obstetrics,’’ and ‘‘amniotic fluid embo-
lism’’ in various combinations revealed only one case report containing
a possible association with cell salvage and maternal death [23]. The patient
was a Jehovah’s Witness with hemolysis–elevated-liver-enzymes–low-
platelets (HELLP) syndrome. Preoperatively, she was anemic and thrombo-
cytopenic with a hemoglobin of 7.1 g/dL and a platelet count of 48,000/mL.
Intraoperatively, she developed clinical signs of disseminated intravascular
coagulopathy (DIC). The estimated blood loss was 600 mL, and she received
200 mL of salvaged blood. She died 10 minutes later from a cardiac arrest,
and an autopsy never confirmed AFE. It is likely that the combination of
severe anemia and DIC was the cause of that death, but this cannot be
verified.

Techniques employed by anesthesiologists

To complete the overview of alternatives to blood and blood products, an
anesthesia consult should be obtained to discuss some additional techniques
available to combat massive blood loss. Ideally, there should be a core
group of obstetric anesthesiologists involved in the patient’s care who are
familiar with the relevant therapeutic options and well versed in the imple-
mentation of intraoperative alternatives to blood administration in women
experiencing massive intraoperative bleeding. All the anesthesiologists
involved should be comfortable with the management plans because the
patient’s refusal to accept blood may result in her death on the operating
table. If a member of that group does not feel that he or she can withhold
a transfusion, a covering physician should be immediately available to
take over if needed. This arrangement prevents confusion and conflict in
the case of an emergency situation.
362 GYAMFI & BERKOWITZ

Intraoperative techniques to combat massive hemorrhage include normo-

volemic hemodilution, controlled hypotensive anesthesia, sedation, and
muscle paralysis. Normovolemic hemodilution involves removing whole
blood in the immediate preoperative period and replacing it with crystalloid
or colloid [4]. This causes a decrease in the viscosity of the patient’s circulat-
ing blood and increases tissue perfusion. Because the circulating blood con-
tains a reduced number of red cells, there is a shift of the oxygen dissociation
curve to the right, which optimizes the oxygen-carrying capacity of those
cells [16]. Once the perioperative blood loss has been curbed, the patient’s
whole blood can be replaced. This technique has been used safely in some
pregnant patients [18]. Controlled hypotensive anesthesia involves reducing
the mean arterial pressure to 50 mm Hg [4]. This is the minimum require-
ment for tissue perfusion, and reduces the amount of blood loss by lowering
the arterial pressure in the setting of substantial intraoperative hemorrhage.
Sedation and muscular paralysis have also been used both peri- and postop-
eratively to decrease oxygen consumption [4].
If the pregnant Jehovah’s Witness is scheduled for a cesarean delivery
with the potential for more than average blood loss (eg, in the case of a pre-
vious myomectomy or a known placenta accreta) consultation with inter-
ventional radiology for preoperative pelvic placement of balloon catheters
is an option to be considered.

Blood substitutes
An ideal substitute for blood would be a compound that could both act
as a volume-expander and have a high oxygen-carrying capacity. Such com-
pounds exist, but are in limited use in the United States because of several
shortcomings. Perfluorocarbons are under investigation for the delivery of
oxygen to tissues [24]. These compounds have a 10- to 20-fold increase in
oxygen-carrying capacity when compared with water, but they are very un-
stable at room temperature, and there is limited information on their use in
pregnancy [25]. Stroma-free hemoglobin is another potential blood substi-
tute. However, it has been shown to cause hypertension and renal damage,
and there are no reports of its use in pregnancy [26].
Recombinant activated factor VIIa has been used to treat obstetric
hemorrhage. This clotting factor is indicated for patients with demon-
strated factor VII deficiency, and its use in obstetrics remains controver-
sial. Factor VIIa promotes hemostasis by ultimately leading to the
formation of fibrin through an increase in thrombin formation [27].
Although there are case reports of successful use in the treatment of obstetric
hemorrhage [27,28], recombinant activated factor VIIa has been associated
with the development of thromboembolic events [29]. Considering the
hypercoagulable state of pregnancy, one should only use this drug as a last
resort.
 MANAGEMENT OF PREGNANCY IN A JEHOVAH’S WITNESS 363

Once the various therapeutic options have been discussed, the patient
should also be made aware that, in the case of a significant postpartum hem-
orrhage, a hysterectomy might be necessary. This should be performed much
earlier than would be the case in women who accept blood transfusions. The
potential need for hysterectomy is part of a routine consent once any patient
is admitted to a labor floor, but in the case of a Jehovah’s Witness, there
should be a much lower threshold for definitive surgical management if hem-
orrhage ensues [10]. At the authors’ institution, obstetric patients who refuse
blood transfusion are not candidates for elective procedures, such as tubal li-
gation, and they are informed of this during the antepartum period. Addi-
tionally, women who refuse to accept blood or blood products are not
considered to be candidates for attempted vaginal birth after cesarean be-
cause of the increased risk for blood transfusion in this group of patients [8].

End of life decisions

Once a Jehovah’s Witness has declared what forms of management are
acceptable to her, the next step involves making end-of-life decisions and
assigning next of kin to her children [10]. This serves not only to convey
to the patient the importance and potential consequences of blood refusal,
but also to prevent a court order reversal of such refusal. It is important
that the patient understands that the refusal to accept blood or blood prod-
ucts substantially increases her risk of both morbidity and mortality if major
hemorrhage occurs. She should feel comfortable that with appropriate early
prenatal care her condition can be optimized before the intrapartum period;
but she must also know that even with the best ‘‘alternatives’’ to blood
transfusion, she still could bleed to death.
The remainder of the patient’s prenatal care involves reassessment of her
hematocrit at least once a trimester with treatment of anemia as indicated.
As stated, the goal is to maintain a hematocrit above 40% so that even a rel-
atively large amount of peripartum blood loss will be better tolerated. Ap-
propriate consultation should be completed in the antepartum period, with
an initial maternal–fetal medicine consult obtained before 28 weeks. The
blood products checklist and health care proxy should be signed and placed
in the patient’s chart.

Summary
In the successful management of a pregnant Jehovah’s Witness, many is-
sues must be addressed beyond those normally required for routine prenatal
care. The clinician who undertakes such care should be well versed in the
potential complications related to blood refusal, the antepartum manage-
ment of anemia, and the intrapartum management of obstetric hemorrhage.
Furthermore, these patients should be delivered in a tertiary care center
because this increases their options for obtaining alternative management
364 GYAMFI & BERKOWITZ

of hemorrhage. A woman who is well informed about her options can then
decide exactly what she wants done in the event of a life-threatening obstet-
rical hemorrhage.

References
[1] Chang J, Elam-Evans LD, Berg CJ, et al. Pregnancy-related mortality surveillancedUnited
States, 1991–1999. MMWR Surveill Summ 2003;52:1–8.
[2] Singla AK, Lapinski RH, Berkowitz RL, et al. Are women who are Jehovah’s Witnesses at
risk of maternal death? Am J Obstet Gynecol 2001;185:893–5.
[3] Harrison BG. Visions of glory: a history and memory of Jehovah’s Witnesses. New York:
Simon and Shuster; 1978.
[4] Gyamfi C, Yasin SY. Preparation for an elective surgical procedure in a Jehovah’s Witness:
a review of the treatments and alternatives for anemia. Prim Care Update Ob Gyns 2000;7:
266–8.
[5] Cunningham FG, Hauth JC, Leveno KJ, et al, editors. Williams obstetrics. 22nd edition.
New York: The McGraw-Hill Companies, Inc.; 2005.
[6] Klapholz H. Blood transfusion in contemporary obstetric practice. Obstet Gynecol 1990;75:
940–3.
[7] Rouse DJ, MacPherson C, Landon M, et al. for the National Institues of Child Health and
Human Development Maternal-Fetal Medicine Units Network. Blood transfusion and
cesarean delivery. Obstet Gynecol 2006;108:891–7.
[8] Landon MB, Hauth JC, Leveno KJ, et al. for the National Institues of Child Health and Hu-
man Development Maternal-Fetal Medicine Units Network. Maternal and perinatal out-
comes associated with a trial of labor after prior cesarean delivery. N Engl J Med 2004;
351:2581–9.
[9] Gyamfi C, Berkowitz RL. Responses by pregnant Jehovah’s Witnesses on health care prox-
ies. Obstet Gynecol 2004;104:541–4.
[10] Gyamfi C, Gyamfi MM, Berkowitz RL. Ethical and medicolegal considerations in the
obstetric care of a Jehovah’s Witness. Obstet Gynecol 2003;102:173–80.
[11] Centers for Disease Control and Prevention. Recommendations to prevent and control iron
deficiency in the United States. MMWR Recomm Rep 1998;47(RR-3):1–29.
[12] Silverstein SB, Rodgers GM. Parenteral iron therapy options. Am J Hematol 2004;76:74–8.
[13] Morreale A, Plowman B, DeLattre M, et al. Clinical and economic comparison of epoetin
alfa and darbepoetin. Medscape Today. Available at: http://www.medscape.com/
viewarticle/472685_4. Accessed March 29, 2007.
[14] Aranesp prescribing information. Available at: http://www.aranesp.com/professional/
prescribing_information.jsp#dosage. Accessed March 28, 2007.
[15] Yamada AH, Lieskovsky G, Skinner DG, et al. Impact of autologous blood transfusion on
patients undergoing radical prostatectomy using hypotensive anesthesia. J Urol 1993;149:
73–6.
[16] Desmond MJ, Thomas MJG, Gillon J, et al. Perioperative red cell salvage. Transfusion 1996;
36:644–51.
[17] Fuhrer Y, Bayoumeu F, Boileau S, et al. Evaluation of the blood quality collected by cell
saver during cesarean section. Ann Fr Anesth Reanim 1996;15(8):1162–7.
[18] Bernstein HH, Rosenblatt MA, Gettes M, et al. The ability of the Haemonetics 4 Cell Saver
System to remove tissue factor from blood contaminated with amniotic fluid. Anesth Analg
1997;85(4):831–3.
[19] Catling SJ, Williams S, Fielding AM. Cell salvage in obstetrics: an evaluation of the ability of
cell salvage combined with leucocyte depletion filtration to remove amniotic fluid from
operative blood loss at caesarean section. Int J Obstet Anesth 1999;8:79–84.
 MANAGEMENT OF PREGNANCY IN A JEHOVAH’S WITNESS 365

[20] Waters JH, Biscotti C, Potter PS, et al. Amniotic fluid removal during cell salvage in the
cesarean section patient. Anesthesiology 2000;92:1531–6.
[21] ACOG Committee opinion. Number 266, January 2002: placenta accreta. Obstet Gynecol
2002;99(1):169–70.
[22] Catling SJ, Joels L. Cell salvage in obstetrics: the time has come. BJOG 2005;112:131–2.
[23] Oei SG, Wingen CB, Kerkamp HEM. Cell salvage: how safe in obstetrics? [letter]. Int J Ob-
stet Anesth 2000;9:143.
[24] Victorino G, Wisner DH. Jehovah’s Witnesses: unique trauma population. J Am Coll Surg
1997;184:458–68.
[25] Karn KE, Ogburn PL Jr, Julian T, et al. Use of a whole blood substitute, Fluosol-DA 20%,
after massive postpartum hemorrhage. Obstet Gynecol 1985;65:127–30.
[26] Bartz RR, Przybelski R. Blood substitutes. eMedicine. Available at: http://www.emedicine.
com/med/topic3198.htm. Accessed March 29, 2007.
[27] Prosper SC, Goudge CS, Lupo VR. Recombinant factor VIIa to successfully manage dis-
seminated intravascular coagulation from amniotic fluid embolism. Obstet Gynecol 2007;
109:524–5.
[28] Pepas LP, Arif-Adib M, Kadir RA. Factor VIIa in puerperal hemorrhage with disseminated
intravascular coagulation. Obstet Gynecol 2006;108:757–61.
[29] O’Connel K, Wood J, Wise R, et al. Thromboembolic adverse events after use of recombi-
nant human coagulation factor VIIa. JAMA 2006;295:293–8.
 Obstet Gynecol Clin N Am
34 (2007) 367–388

Intimate Partner Violence

Jennifer Gunter, MD
Department of Obstetrics/Gynecology, Kaiser Northern California, 2238 Geary Boulevard,
San Francisco, CA 94115, USA

Intimate partner violence (IPV) is a pattern of psychological, economic,

and sexual coercion of one partner in a relationship by the other that is
punctuated by physical assaults or credible threats of bodily harm [1,2]. It
is a universal health crisis affecting women of every economic, social, cul-
tural, and racial background. The World Health Organization (WHO)
Multi-Country Study of Women’s Health and Domestic Violence Against
Women indicates that the lifetime prevalence of IPV varies significantly
by country and region, ranging from 13% to 71% [3]. Estimates of the prev-
alence in the United States vary significantly because of underreporting and
differences in methods of collection with the lifetime prevalence ranging
from 23% to 60%, with an annual prevalence of up to 17% and an esti-
mated 5.3 million IPV incidents per year [4–10]. IPV is the most common
cause of nonfatal injury for women with 21% of the female population re-
porting ever receiving some type of injury and 9% reporting a severe injury
[6,11]. IPV is truly an obstetrics gynecology emergency as 50% of murdered
women are killed by a current or previous partner. Murder is among the five
most common causes of death for women ages 15 to 34 and is the leading
cause of maternal mortality [12,13].

The scope of the problem

The definition of IPV, also known as domestic violence, encompasses
both physical and sexual violence in addition to psychological abuse, eco-
nomic coercion, stalking, and threats of violence both sexual and nonsexual.
There are many misperceptions concerning personality or socioeconomic
status of women who are victimized; every woman who has ever been part-
nered in a heterosexual or same-sex relationship is at risk [7].
IPV is characterized by what has become known as the cycle of violence
that starts with tension-building or arguing that escalates into battering,

E-mail address: jennifer.gunter@kp.org

0889-8545/07/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ogc.2007.06.010 obgyn.theclinics.com
368 GUNTER

followed by a ‘‘honeymoon phase,’’ which is characterized by excuses, gifts,

and/or denial (Fig. 1). Many ask: ‘‘Why don’t women just leave?’’ The
reasons are complex and involve both intangibles and barriers to leaving,
such as shame, guilt, love, self-esteem, hopelessness, depression, economic
dependency, lack of support systems, social isolation, fear, and negative
experiences with medical professionals and the legal system. In addition,
changing behavior is a dynamic process. A continuum of predictable stages
has been identified as individuals attempt to change behavior (Fig. 2)
[14–17]. These stages account for such responses as denial, acknowledgment
of the problem, planning for action, enacting the plan, and maintenance.
Returning to a previous stage is a frequent occurrence and many women
leave a harmful relationship as many as eight times before securing a perma-
nent break [14,17].
The lifetime prevalence of IPV in the United States ranges from 25% to
60% with an annual prevalence of 4% to 17% [4–8,18–21]. IPV is the most
common cause of nonfatal injury for women. In a given year, approximately
1.5 million women in the United States are victimized. On a global scale,
millions of women are assaulted every day [3–5,11,19,21]. The two most
common forms of abuse are emotional (84%) and psychological (68%).
However, 43% to 60% of women report physical violence. The most com-
mon violent act is a slap [3,18,19].
IPV has subclassifications based on risk of injury and potential lethality
[1,3,11,19]. Severe IPV involves being hit with a fist, kicked, dragged,
choked, threatened, burned, or injured with a weapon with a lifetime prev-
alence among ever-partnered women ranging from 4% to 49% [1,3,11]. In
the United States, at least 21% of women report an injury as the result of
IPV and up to 46% of women seen in the emergency room for violence-
related injuries are injured by a current or former partner [19,22,23]. Under-
reporting of these injuries is common because many women do not seek care
and screening for IPV is suboptimal, even in emergency room settings, so

Tension Building Battering

(blaming, arguing, jealousy) (verbal threats, sexual abuse,
physical battering, use of weapons)

Honeymoon
(excuses, gifts, denial)

Fig. 1. IPV cycle of violence.

 INTIMATE PARTNER VIOLENCE 369

Precontemplation: not aware, denial, minimizing problem

Contemplation: aware and considering changes

Preparation: making plans

Action: enacting plans

Maintenance: keeping the new actions as part of daily activities

Fig. 2. Stages of change. Returning to a previous stage is expected, is not a failure, and may
happen several times as people learn more about their problems and how best to approach
them. (Data from Refs. [14–17].)

many assaults are unrecognized [21–23]. The lifetime prevalence of sexual

abuse by a current or previous partner ranges from 10% to 50%. Among
ever-abused women, 40% to 80% report a sexual assault, which may have
been the result of direct physical force or of fear of implied violence
[3,24]. Femicide, the murder of a woman, is a leading cause of death for
women and 40% to 50% of these murders are perpetrated by a current or
previous intimate partner [12,18,25,26].

At-risk populations
While any woman who has ever been partnered is at risk for IPV, some
populations are at increased risk, including pregnant women, adolescents,
and the disadvantaged. Women who are at increased risk often have addi-
tional barriers to leaving, such as a greater degree of financial and emotional
dependency and greater social isolation [14,27].
Up to 45% of pregnant women report a history of IPV and the preva-
lence of IPV during pregnancy ranges from 6% to 22% [3,28–35]. It is im-
portant for clinicians to include women seeking pregnancy termination in
this high-risk population because 22% of women seeking pregnancy termi-
nation report a history of abuse in the preceding 12 months and 24% to
35% report a history of substantial conflict and fights with the man involved
with the current pregnancy [32–34]. Of all the assault-related injuries re-
ported for women of reproductive age, 10% occurred during pregnancy
and women who are assaulted during pregnancy are three times more likely
to be hospitalized as compared with women who are assaulted and not preg-
nant [36]. Women who are pregnant are three times more likely to be a victim
370 GUNTER

of an attempted or successful femicide as compared with abused nonpreg-

nant controls [37]. Trauma is the leading cause of maternal death and femi-
cide is the most common cause of injury-related death, most often
perpetrated by an intimate partner [37–43].
The increased incidence of IPV-related abuse, assaults, and femicide dur-
ing pregnancy is most likely multifactorial. Pregnancy is associated with in-
creased personal, medical, and financial stress. Pregnancy is also a period
when attention is focused on the pregnant woman, which means the partner,
and potential batterer, gets less attention. Furthermore, pregnancy may also
mark a change in the relationship. Unplanned pregnancy may be a marker
for sexual assault as a significant percentage of women who are victimized
by IPV are raped by their partners. Meanwhile, many other women become
pregnant out of fear of implied violence, they fail to ask their male partner
to use a prophylactic, or are afraid or unable to see a health care provider
for a prescription contraceptive [3].

Adolescents
The incidence of IPV is highest among younger women, particularly be-
tween the ages of 15 and 19 [3,44–47]. Dating violence is a significant prob-
lem in this population with more than 90% of teens reporting verbal abuse,
25% reporting physical abuse, and 14% victimized by sexual abuse [14,
44–47]. Femicide, most often perpetuated by an intimate partner, is the
number-one cause of death for African American women ages 15 to 24 and
the second most common cause of death for white women ages 15 to 24
[12,18,47]. In addition to injuries, the consequences of IPV for adolescent
women include anxiety, anger control issues, suicide ideation, substance
abuse, unsafe sex, and unhealthy weight control behaviors [48–51]. Young
maternal age is an independent risk factor for IPV during pregnancy and,
among adolescents who are pregnant, IPV is associated with a more-
than-threefold increased risk of repeat pregnancy within 12 months [52].

Disadvantaged populations
IPV affects women of every race and ethnicity, regardless of socioeconomic
status. However, some women have additional vulnerabilities and greater bar-
riers to leaving based on social, economic, or physical factors. In the United
States, victimization rates are highest for African American women, women
who live in urban areas, and those with lower household incomes [53]. In ur-
ban areas, the exposure to violence in general is greater and it has been hypoth-
esized that this may cause desensitization, leading to acceptance or
rationalization of IPV by both victim and perpetrator [14,49,54,55]. Poverty,
higher in inner-city regions and among minority women, increases financial
dependency on an abusive partner and creates additional barriers to leaving,
such as difficulties in finding new housing and obtaining resources for civil lit-
igation. Minority women report a higher prevalence of negative experiences,
 INTIMATE PARTNER VIOLENCE 371

including racism, with institutional resources and law enforcement. These

negative experiences further inhibit IPV reporting because these women as-
sume they will not get the type of assistance they need or they fear that their
partner may be victimized by racism [14,55–58].
The prevalence of IPV varies among cultures. However it is more pre-
valent in some societies and in some cultures many women report that the
violence is justified [3]. Acceptance of battering is higher among women
from provincial and rural settings and among those who have previously
experienced abuse, suggesting that some women may learn to adapt to their
violent situations and, either because of societal pressure or because of
acceptance of their situation, do not recognize themselves as victims [3].
This is an important consideration for immigrant women who may have dif-
ferent understandings of what constitutes IPV as it is ‘‘normalized’’ in some
cultures. Communication barriers, social isolation, lower awareness of
IPV-related services, and lack of direct questioning by clinicians add further
barriers for immigrant women [14,58–61]. Women with no family in the
United States are three times more likely to be physically injured by their
partner as compared with women with family in the country. Immigration
laws further increase the risk of victimization; IPV is higher among women
who report that their partners refuse to change their immigrant status,
among those who are threatened by their spouse with deportation, and
for women on spousal visas who are unable to work [60,61].
Aboriginal women–that is, women descended from indigenous peoples of
North America, report a higher prevalence of IPV and in some communities
it is estimated that 60% to 90% of women are battered and up to 57% sex-
ually abused [14,62–65]. Aboriginal women are more likely to be victims of
severe IPV with more than 40% reporting injuries and are eight times more
likely to be a victim of femicide as compared with non-aboriginal women
[63–66]. Like women in other minority populations, aboriginal women
experience double discriminationdas a woman and as a minority [14]. In
addition, for many minority women, regardless of race, ethnicity, country
of origin, culture, or aboriginal status, culturally appropriate services for
victims of IPV often do not exist.
Women with disabilities are more vulnerable to abuse and face more bar-
riers in attempting to escape abuse. Challenges encountered by women with
disabilities include an inability to physically defend themselves, a high
dependency on partners for physical needs, difficulties in reporting abuse
because of communication barriers, an inability without assistance to phys-
ically leave a dwelling and go to a shelter, and a high economic dependency
on their partner. The prevalence of IPV is likely significantly underestimated
in this population. However, it is believed to be at least 40% higher than in
the general population with the risk of severe IPV and sexual assault also
significantly higher [14,67–69].
Women who are economically disadvantaged are at increased risk of vi-
olence independent of other risk factors, such as race, aboriginal status,
372 GUNTER

pregnancy, age, and immigrant status [7,49,54,63,70–72]. The associations

between income and IPV are complex, and are most likely different for
each woman. However, economically disadvantaged women, compared to
women with average financial means, have more difficulty hurdling financial
barriers to health care, are less likely to have access to health care, and there-
fore are less likely to be screened for IPV.

Consequences of IPV
The consequences of IPV are far-reaching and range from injuries to the
perpetuation of gender inequality [3,14,73]. The immediate medical sequelae
of IPV include trauma, sexually transmitted diseases, unplanned pregnancy,
and death. Abused women, compared to other women, have a higher inci-
dence of headaches, back pain, vaginal bleeding, vaginal infections, pelvic
pain, dyspareunia, urinary tract infections, eating disorders, abdominal
pain, gastrointestinal disorders, depression, suicide, substance abuse, anxiety,
and chronic somatiform disorder [39,73–78]. Medical consequences that may
not be immediately appreciated include the psychological harm of shame or
guilt, stress-related illness, and post-traumatic stress disorder. Other issues
of concern include noncompliance with medical recommendations and lack
of treatment or exacerbation of medical conditions because of insufficient ac-
cess to health care either due to shame, fear of discovery, or restriction of ac-
cess to health care by an abuser to maintain control [14,72].
It is estimated that IPV costs $5.8 billion annually in the United States,
with $4.1 billion for direct medical care and mental health services; a study
conducted in a closed-model health maintenance organization indicates that
IPV increases the cost per member per year by $1700 [9,79]. Costs increased
most among women who reported physical abuse. However, elevated costs
are also associated with sexual and emotional abuse, and cost of care in-
creased both for women currently experiencing abuse and for those who re-
ported a history of IPV [79].
The maternal sequelae of IPV during pregnancy include maternal mor-
bidity from injuries, exacerbation of medical conditions due to restricted ac-
cess, depression, and mortality because pregnant women are more likely to
die as victims of femicide than from any obstetric cause [13,14,39–43,80].
Women who are victimized by IPV during pregnancy have an increased
risk of spontaneous abortion and an increase in perinatal complications,
such as low birth weight, preterm labor and delivery, preterm rupture
of membranes, insufficient weight gain, and urinary tract infections
[14,29,31,80–84]. One quarter to one half of women who are physically
abused during pregnancy report that they were kicked or punched in the ab-
domen. These women had increased rates of placental abruption and ante-
partum hemorrhage [3,14,29,37,80–84]. In addition, violence during
pregnancy results in delayed entry into prenatal care [14,29,80–84].
 INTIMATE PARTNER VIOLENCE 373

The medical sequelae of IPV also extend to children; in homes with IPV,
child abuse occurs in up to 70% of families. Thirty-nine percent of victim-
ized women report that their children witnessed the attack and during 61%
of these attacks the mother was injured [85–87]. Children who witness vio-
lence not only are at risk of injury, but are also more likely to have behav-
ioral problems, problems in school, and such problems as substance abuse,
anxiety, aggression, enuresis, depression, and suicidality [74,85–89]. In addi-
tion, batterers often use child custody as a forum to continue the abuse with
harassing and retaliatory legal actions [86,90].
Women victimized by IPV experience significant economic hardship.
They may miss work because of injuries, fear, stalking, court appearances,
custody hearings, and litigation and they may incur more expenses with
new housing and legal bills from divorce and child custody petitions.
Women who leave violent situations are four times more likely to report
housing instability, such as late rent or mortgage payments and frequent
moves, because of the inability to obtain affordable housing or lack of
own housing [91]. Housing ramifications can be severe as 50% to 60% of
homeless women report a history of IPV [92,93].

Diagnosing IPV
Whom to screen?
With a lifetime prevalence of 25% to 60% and a 21% lifetime risk of
injury, women who are currently victims of IPV and those who have previ-
ously been abused are likely to be encountered regularly in both acute-care
and office-based settings [4–8,18–23]. Accordingly, the American College of
Obstetrics and Gynecology (ACOG) recommends routine screening at
annual exams, family planning visits, and preconception visits [29,94,95].
Routine screening for IPV is also endorsed by the Society of Obstetricians
and Gynecologists of Canada, the American Medical Association, the
American Academy of Family Physicians, and numerous other national
medical associations and government agencies [10,14,96,97]. The Joint Com-
mission, formerly the Joint Commission on Accreditation of Healthcare
Organizations (JCAHO), initiated standards for IPV screening in 2004
(JCHAO standard PC.3.10 on victims of abuse).
Factors that increase a woman’s risk for IPV include young age, previous
episodes of IPV, and disability. This means that some patients may require
more frequent screening. Enhanced surveillance is specifically recommended
during pregnancy because of the increased risk of IPV and its association
with both maternal and fetal morbidity and mortality [14,29,40,94,98].
Screening in pregnancy should occur at the first prenatal visit, at least
once a trimester, and at the postpartum visit [14,29,94,99,100]. In addition,
there are ‘‘red flags’’ that should raise suspicion of IPV and prompt screen-
ing. These ‘‘red flags’’ include injuries that are inconsistent with the history,
374 GUNTER

frequent missed appointments, repeated visits with vague complaints, and

chronic pain [14,29,95,98].
The US Preventative Services Task Force and the Canadian Task Force
on Preventative Health Care do not recommend routine screening for IPV
because of ‘‘limited evidence as to whether interventions reduce harm to
women,’’ because few studies have addressed the negative sequelae of
screening, and because few interventions have proven successful [101–103].
However, support for screening, both routine and when symptoms suggest
possible abuse, is high among women who have been victimized by IPV
[104,105]. In addition, many variables affect how a patient responds to
screening for IPV. Such variables include the stages of change, fear of repri-
sal, self-esteem, previous experiences with the medical and legal systems,
skill of the provider, and format used to screen [14,98,104–107]. The evi-
dence for the efficacy of specific interventions for IPV are unclear and the
most appropriate outcome measures have not been identified. Such mea-
surements could track access to advocacy services, frequency of abusive ep-
isodes, or injury rates. Such measurements would vary depending on stages
of change and many other unique factors for each woman [14,104–108].
Many women identify the act of screening itself as helpful and possibly use-
ful in helping a woman move forward in the stages of change [3,14,104–107].
Barriers to leaving are multifactorial and unique for each woman. Health
care professionals do not necessarily have the ability to provide the desired
health outcome because freedom from violence for many women involves
complex financial, social, and legal issues. Furthermore, leaving a violent
partner does in guarantee freedom from further violence as many women
are stalked, abused, assaulted, and even murdered by former partners
[4,5,12,109]. Many significant health problems have ineffective interven-
tions. One such problem is smoking, which is the most common preventable
cause of death in the United States with only a 14% to 20% long-term quit
rate. Yet the US Preventative Services Task Force recommends that clini-
cians screen all adults for tobacco use and provide tobacco cessation inter-
ventions [102,110].

How to screen?
Screening involves not only asking the right questions, but also docu-
menting findings and providing information to victims about safety,
options, and interventions. A useful pneumonic developed by the Massachu-
setts Medical Society is RADAR with each letter representing one of its five
directives: RdRoutinely inquire about violence; AdAsk direct questions;
DdDocument findings; AdAssess safety; and RdReview options and re-
ferrals. To ensure both safety and accuracy A woman must not be in the vi-
cinity of a partner or family member when screened, and questions should
be posed in a nonjudgmental manner. A sound universal policy is to
make sure every patient has time alone with his or her health care
 INTIMATE PARTNER VIOLENCE 375

professional. Also, it is best to routinely use a medical interpreter and not

a family member if there are language barriers. As staff and patients alike
become familiar with these routines, patients will be less likely to be anxious
about being singled out for questioning and a perpetrator who presents with
his or her partner will be less likely to become suspicious.
A variety of questionnaires, both oral and written, have been designed.
How a patient is screened significantly affects response rates, with a 12-
month prevalence of IPV ranging from 1% to 19%, depending on the
method used [4–8,10,19,21]. The most common ones cited include the Part-
ner Violence Screen (PVS), the Women Abuse Screening Tool (WAST), the
SAFE tool, a two-question emergency department tool, and the Conflicts
Tactics Scale (CTS), which is considered the gold standard (Box 1)
[10,14,111–114]. All of these questions are closed-ended with yes–no or short
responses; only the WAST asks about violence in an indirect manner and
then progresses to direct abuse-related questions. In addition, there is a ver-
bal, less structured patient-centered approach that involves picking up on
verbal and nonverbal cues, such as a patient comment about stress, a chronic
pain complaint, or another issue. Then questions can be framed using the
patient’s own description: ‘‘You have described a lot of stress. How is
that handled at home?’’ The response may lead to further questions and re-
sponses that uncover serious problems [114]. Single questions about being
afraid produce lower results, with only 8% of victims correctly identified;
only 50% of women who survive an attempted homicide by partner per-
ceived their risk and women who are precontemplative may not perceive
risk at all [113,115].
When compared with the gold standard CTS the three-question PVS has
a 71% sensitivity and an 85% specificity. The PVS and WAST have similar
sensitivities. However, the written WAST may yield a lower prevalence
[10,111]. Studies are conflicting as to the optimal method of screening
with some suggesting that patients prefer a written questionnaire and others
supporting the less structured, individually tailored, patient-centered ap-
proach, which appears to be preferred, although non-direct screening may
have a lower sensitivity [3,10,104–107,114,116,117]. Women report that
they want their physician to be sympathetic and caring, so it is possible
that health care professionals who do not have the same training as IPV re-
searchers may ask direct questions with a different tone and manner or they
respond differently to positive screens, thus reducing satisfaction with this
approach [104,105,116].

Barriers to screening
Voluntary screening by verbal questions and subsequent documentation
in the medical record are often considered ‘‘usual care.’’ However, this
method results in the lowest screening rates with only 8% to 45% of women
in the emergency room and 10% to 42% in office-based settings screened
376 GUNTER

Box 1. IPV screening tools

Partner Violence Screen
1. Have you been hit, kicked, punched, or otherwise hurt by
someone within the past year? If so, by whom?
2. Do you feel safe in your current relationship?
3. Is there a partner from a previous relationship who is making
you feel unsafe?
Antenatal Psychological Assessment
1. Within the past year, or since you have become pregnant,
have you been hit, slapped, kicked, or otherwise physically
hurt by someone?
2. Are you in a relationship with a person who threatens or
physically hurts you?
3. Has anyone forced you to have sexual activities that made you
feel uncomfortable?
SAFE tool
S for spouse: How would you describe your spousal relationship?
A for arguments: What happens when you and your partner
argue?
F for fights: Do fights result in you getting hit, shoved, or hurt?
E for emergency: Do you have an emergency plan?
Emergency department screening tool
1. Have you ever been hit, slapped, kicked, or otherwise
physically hurt by your partner?
2. Have you ever been forced to have sexual activities?
The Woman Abuse Screening Tool
1. In general, how would you describe your relationship? A lot of
tension? Some tension? No tension?
2. Do you and your partner work out arguments with great
difficulty? With some difficulty? With no difficulty?
3. Do arguments ever result in you feeling down or bad about
yourself? Often? Sometimes? Never?
4. Do arguments ever result in hitting, kicking, or pushing?
Often? Sometimes? Never?
5. Do you ever feel frightened by what your partner says or
does? Often? Sometimes? Never?
6. Has your partner ever abused you physically? Often?
Sometimes? Never?
7. Has your partner ever abused you emotionally? Often?
Sometimes? Never?
 INTIMATE PARTNER VIOLENCE 377

according to ACOG guidelines, with more than 50% of providers not

screening at all for IPV and one third screening only if a patient presents
with a bruise or laceration; younger women, who are at greatest risk for
IPV and subsequent injury, are screened the least [10,21,112,118–122]. The
low screening rates for IPV imply that significant barriers for health care
professionals impede routine screening. Attitudes toward IPV, training in
residency, and comfort with screening vary significantly and screening for
IPV by health care professionals is related to their preparedness, both edu-
cational and experiential [119–121,123,124]. Studies using oral screening for
IPV use a trained professional (physician, nurse, or research assistant) who
has background in the area, is trained to screen, and is surveying many
patients and thus is well prepared both by education and experience. Addi-
tional barriers for health care professionals include brevity of visits, lack of
access to services, misconceptions about typical victims, frustration because
the victim may not leave dangerous situation, and lack of understanding of
mandatory reporting laws [119–124].
Women who have been or who currently are victims of IPV have signifi-
cantly more frequent contacts with the health care system as compared with
woman who have never been abused. This means that barriers to screening
for IPV translate into many missed opportunities for detection and interven-
tion [125,126]. In a study of identified female IPV victims, 64% presented at
least once to an emergency department in the year of the index assault (with
the median number of visits four), but only 23% were correctly identified as
victims of IPV [125]. Among women murdered by a current or previous inti-
mate partner, 40% sought medical care in the emergency room within the pre-
ceding 12 to 24 months. Thus, there are many potential opportunities to offer
intervention for many of the women who are at the greatest risk [127,128].
There are also patient barriers to identification, including denial, past
failures with medical and legal systems, shame, cultural and language bar-
riers, fear of reprisal, low self-esteem, and desire to protect the perpetrator
[3,14,15,17,98,104–107]. Many women do not recognize that they are victims
of abuse or they underestimate their risk; violence becomes normalized
through exposure and psychological abuse leaves many women with shame
and self-doubt [14–17,27,98,113–116]. If a woman does not recognize her
situation as abusive, it is important to raise the issue but to not push too
far to prevent alienation [14,104–107,114]. Posters in bathrooms and printed
material in waiting rooms can also help raise awareness among women who
are precontemplative.

How to respond
If a patient responds yes to screening for IPV the following four steps
should occur: (1) show support, (2) perform a risk assessment, (3) document
injuries, and (4) discuss solutions [10,14,29,74,95–99]. Statements of support
378 GUNTER

might vary if the patient is screening positive for current abuse versus past
or lifetime abuse. For patients currently in violent relationships, statements
may include:
‘‘I believe what you are saying.’’
‘‘No one deserves to be treated that way.’’
‘‘I am so sorry. I would like to help.’’
‘‘It must be hard to be treated that way.’’
‘‘It’s not your fault.’’ [14,74,98]
For patients who are no longer in a violent situation, examples of useful
statements include:
‘‘That must have been a difficult time.’’
‘‘Some women have health consequences from such stress.’’
‘‘Do you have any ongoing concerns regarding a previous relationship?’’
[14,74,98]
The next step is to perform a risk assessment. A variety of factors have
been identified that are associated with increased risk of injury and lethality
(Box 2) [4,11,14,25,40,74,109,115,128–131]. Factors associated with an in-
creased risk of femicide include the perpetrator’s access to a gun, previous

Box 2. Risk factors for injury and lethality

Demographic factors
 Age 15–24
 Minority population
 Pregnant
 Women with disabilities
Assault factors
 Perpetrators threats of suicide or homicide
 Gun in house
 Choking
 Stalking
 Previous or current injury with a weapon
 Sexual assault
 Abuse of family pets
 Increase in severity or frequency of violence
 Fear for personal safety or life
 Violence outside the house
Relationship factors
 Recent separation
 Separation for a new partner
 Perpetrator stepchild living in the house
 INTIMATE PARTNER VIOLENCE 379

threat with a weapon, abuse during pregnancy, stalking, choking, forced

sex, perpetrator’s stepchild living in the home, estrangement from a control-
ling partner, victim having left the relationship for another partner, and per-
petrator threats of suicide [4,11,14,25,40,74,109,115,128–131]. Abuse does
not always escalate in a predictable pattern and many women are severely
injured or murdered without any known risk factors. However, if a patient
screens positive, there is a significant increased risk of injury and death
[14,25,74,98,115,128,130,132]. Women often underestimate the risk of their
situation, a factor probably compounded by many factors, including stages
of change, fear, and lack of alternatives to leaving; almost 50% of women
who are severely injured or murdered by their partner did not appreciate
that they were at risk [25,37,115,129,130]. If any risk factors are present
for injury or lethality, be clear that the patient may be in imminent danger,
be clear about the need to leave, and document the conversation and recom-
mendations in the medical record.
Patients should be asked if they have any injuries as a result of IPV. If so,
such injuries should be documented because this is important for legal
follow-up. Injuries should be photographed. If a camera is not available,
draw the injuries. Common injury patterns include defensive wounds;
central injuries; multiple injuries; bruises in various stages of healing; and
injuries to head, neck, and mouth. During pregnancy, the abdomen is
more likely to be involved [14,29,31,74,99]. Document size of lesions, color,
bruising, and who the patient identified as the perpetrator; it is important to
be as specific as possible and to use quotations, such as ‘‘John Smith hit me
on March 3rd in the afternoon.’’ [2,14,74,98]. This is an important legal
point as the medical record is not hearsay, and a well-documented chart
can be very helpful with orders of protection, prosecution, and child
custody. Unfortunately, documentation of risk and safety assessment is
often neglected; in one study, only 4% of identified victims had any IPV
documentation in the medical record and less than 2% had documentation
of risk assessment [132]. Documentation is also essential for those working in
a hospital setting because screening for IPV is a Joint Commission measure
and failure to screen or to document risk assessment and recommendations
has resulted in exposure to medical malpractice claims [132,133].
Many health care providers are uncomfortable dealing with IPV; they
may be unfamiliar with best screening practices, uncomfortable responding
to those who screen positive, and unaware of available and appropriate
interventions. To raise the comfort level of medical staff and improve
screening rates, educational programs are available that incorporate specific
training and tools for response [14,98,118–121,123,124]. When a woman
screens positive for IPV, the provider, after acknowledging the positive
response, should asked directly how he or she can help. It is important to
frame provider responses and interventions in consideration of the stages
of change and not to alienate patients who are precontemplative. However,
discussing IPV can provide important validation for many victims and may
380 GUNTER

help them move forward to a contemplative phase. Posters and flyers in

bathrooms with the number for the National Domestic Violence Hotline
(1-800-799-SAFE) can also raise awareness and provide information for
women about interventions [3,14,16,17,98,104–107,116,134–136]. Women
who screen positive for IPV benefit from safety strategies and information
on local resources, legal steps, and advocacy [16,98,104–107,135,136]. A per-
sonal safety plan, both for work and home, should be developed; some con-
siderations are money, medications, extra keys, and copies of important
documents with a trusted friend or in a safe deposit box, and a code word
for friends and coworkers to trigger help [14,98]. Contact numbers for local
advocacy services and other resources should be provided and, if available,
the services of a social worker should be offered. Many victims of IPV indi-
cate that legal information would also be helpful and so reporting of IPV to
the police and orders of protection should also be discussed [98,135,136].

Mandatory reporting
Many states have injury reporting requirements for assault-related in-
juries and for injuries resulting from firearms, knives, or other weapons.
California, Colorado, Kentucky, New Hampshire, and Rhode Island each
have specific mandatory reporting laws for IPV [137,138]. In Rhode Island,
reporting is for data collection purposes only with no identifying informa-
tion passed along. In New Hampshire, a patient can object to the release
of the information to the police unless there was a gunshot wound or serious
bodily injury [137]. In California, Colorado, and Kentucky, IPV must be
reported regardless of patient objections. However, in all states health
care providers should encourage women to report the violence to law
enforcement.
While support for universal IPV screening is very high among women
with and without a history of abuse, concerns have been raised that manda-
tory reporting affects patient autonomy and confidentiality, may deter
victims from disclosing IPV or seeking medical care, and may possibly
increase the risk of retaliation [104,135,136,139,140]. In one state with man-
datory reporting, 12% of women attending an inner-city emergency depart-
ment indicated that, with mandatory reporting, they would be less likely to
seek care for an IPV-related injury while 27% said they would be more likely
to seek care [141]. Studies show that survivors of IPV have very high support
for universal screening and physician reporting with patient approval, but
have mixed support for mandatory reporting with 44% to 68% of women
with a history of abuse opposing mandatory reporting that does not allow
for consideration of patients wishes [139,140]. In states with mandatory re-
porting, if a patient objects, it is important to ask why, to try to address any
concerns, and to relay the patient’s objections and reasons to the authorities.
In many states, the witnessing of IPV by a child is considered child abuse
and as such requires mandatory reporting. Because definitions of witnessing
 INTIMATE PARTNER VIOLENCE 381

domestic violence vary significantly by state and may change, it is important

to know the local statute information [142]. State-specific reporting require-
ments are available from the US Department of Health and Human Services
Administration for Children and Families (www.childwelfare.gov/
systemwide/laws_policies/search/index.cfm).

Do interventions work?
A woman increases her likelihood of accessing an intervention and im-
proving her health by talking with a health care provider about abuse
[143]. Interventions that have proven to be effective in reducing subsequent
abuse include a stay by the woman for at least one night in a shelter with
advocacy, the issuance of permanent restraining orders, and the arrest of
the perpetrator [108]. Therapies targeting the batterer, such as cognitive be-
havioral therapy, mandatory counseling, and rigorous monitoring, have not
proven effective. Therefore, the main focus of the intervention should be
helping the patient recognize the abuse and providing assistance to leaving
[144,145]. Executing interventions is out of the hands of the medical pro-
vider and access to advocacy, shelters, and response from the legal system
varies by community. In addition, the current legal system relies more on
batter intervention than on victim support to prevent future violence. While
women can obtain orders of protection, such orders do not prevent batterers
from purchasing guns. There are also many complicating factors, such as
denial, social isolation, language barriers, finances, children, pets, hous-
ing, employment, self-esteem, and fear. So, in many studies, screening
does not translate into change. However, most victims of IPV report
a high degree of satisfaction with screening because it acknowledges the
problem [3,104,105,107,140]. Interventions frequently fail because the prob-
lem of IPV is complex and the solution involves much more than just walk-
ing out the door.

Summary
IPV has a lifetime prevalence of approximately 60% and is a leading
cause of morbidity and mortality for women of all reproductive ages, espe-
cially among younger women and during pregnancy. Providers should rec-
ognize that every woman who has ever been partnered is at risk for IPV and
should screen appropriately, with increased surveillance during pregnancy
and the postpartum period. Despite these recommendations, most providers
do not screen according to ACOG guidelines. However, educational efforts
improve provider confidence in screening. When a woman screens positive
for IPV, it’s important to consider the stages of change; to frame the re-
sponse appropriately; to perform a risk assessment; to discuss interventions,
including a safety plan; and to document in the medical record accordingly.
382 GUNTER

Those providers in states with mandatory screening must also report posi-
tive screens as indicated. Screening has yet to translate into reduced rates
of abuse, indicating that IPV is not simply a medical problem, but involves
complex psychological, financial, familial, cultural, and legal issues. Regard-
less, victims of IPV appreciate screening by medical professionals and indi-
cate that simply asking the questions is helpful and supportive. Society’s
approach to IPV can be also be framed by the stages-of-change model;
only recently has society moved past the precontemplative phase as IPV is
now recognized as a major health problem for women. However, society
is still trying to understand how best to approach the problem and offer
the most effective interventions.

References
[1] Violence against women. WHO Consultation, Geneva 5-7 February 1996. Geneva World
Health Organization. 1996 document FRH/WHD/96.27. Available at: http://whqlibdoc.
who.int/hq/1996/FRH_WHD_96.27.pdf. Accessed April 8, 2007.
[2] Saltzman LE, Fanslow JL, McMahon PM, et al. Intimate partner violence surveillance:
uniform definitions and recommended data elements version 1.0. Atlanta (GA): National
Center for Injury Prevention and Control, Centers for Disease Control and Prevention;
1999.
[3] The WHO Multi-Country Study of Women’s Health and Domestic Violence Against
Women. Summary of initial reports on prevalence, health outcomes, and women’s re-
sponses. Geneva World Health Organization; 2005.
[4] Tjaden P, Thoennes N. Prevalence, incidence, and consequences of violence against women:
findings from the National Violence Against Women Survey. Research in brief. Washing-
ton, DC: US Department of Justice, Office of Justice Programs; 1988. NCJ 172837.
[5] Bachman R. Incidence rates of violence against women: a comparison of the redesigned Na-
tional Crime Victimization Survey and the 1985 National Family Violence Survey. Harris-
burg (PA): VAWnet, a project of the National Resource Center on Domestic Violence/
Pennsylvania Coalition Against Domestic Violence. Available at: http://www.vawnet.org.
Accessed August 12, 2007.
[6] Bensley L, MacDonald S, Van Eenwyk J, et al. Prevalence of intimate partner violence and
injuries: Washington, 1998. MMWR Morb Mortal Wkly Rep 2000;49:589–92.
[7] Moracco KE, Runyan CW, Bowling JM, et al. Women’s experiences with violence: a na-
tional study. Women’s Health Issues 2007;17:3–12.
[8] Bonomi AE, Thompson RS, Anderson M, et al. Ascertainment of intimate partner violence
using two abuse measurement frameworks. Inj Prev 2006;12:121–4.
[9] National Center for Injury Prevention and Control. Costs of intimate partner violence
against women in the United States. Atlanta (GA): Centers for Disease Control and Preven-
tion; 2003.
[10] McMillan HL, Wathern CN, Jamieson E, et al. Approaches to screening for intimate
partner violence in health care settings: a randomized trial. JAMA 2006;296:530–6.
[11] Kyriacou BM, Anglin D, Taliaferro E, et al. Risk factors for injury to women from domes-
tic violence. N Engl J Med 1999;341:1892–8.
[12] Heron MP, Smith BL. Deaths: leading causes for 2003. National Vital Statistics Reports,
vol. 55, No.10. Hyattsville (MD): National Center for Health Statistics; 2007.
[13] Horon IL, Cheng D. Enhanced surveillance for pregnancy-associated mortalitydMary-
land, 1993–1998. JAMA 2001;285:1455–9.
[14] Cherniak D, Grant L, Mason R, et al. Intimate Partner Violence Consensus Statement,
Society of Obstetricians and Gynecologists of Canada No. 157, April 2005.
 INTIMATE PARTNER VIOLENCE 383

[15] Prochaska J, DiClemente C, Nordos J. In search of how people change: application to

addictive behaviors. Am Psychol 1992;47:1102–14.
[16] Burke JG, Denison JA, Gielen AC, et al. Ending intimate partner violence, an application
of the transtheorectical model. Am J Health Behav 2004;28:122–32.
[17] Brown J. Working toward freedom from violence: the process of change in battered women.
Violence Against Women 1997;3:5–26.
[18] Bhandari M, Dosanjh S, Tornetta P, et al. Musculoskeletal manifestations of physical
abuse after intimate partner violence. J Trauma 2006;61:1473–9.
[19] Wilt S, Olson S. Prevalence of domestic violence in the United States. J Am Med Womens
Assoc 1996;51:77–82.
[20] Coker AL, Derrick C, Lumpkin JL, et al. Help-seeking for intimate partner violence and
forced sex in South Carolina. Am J Prev Med 2000;19:316–20.
[21] Trautman DE, McCarrthy ML, Miller N, et al. Intimate partner violence and emergency
department screening: computerized screening versus usual care. Ann Emerg Med 2007;
49:526–34.
[22] Grisso JA, Schwarz DF, Hirschinger N, et al. Violent injuries among women in an urban
area. N Engl J Med 1999;341:1899–905.
[23] Rennison CM, Welchans S. Bureau of Justice Statistics special report: intimate partner
violence. Washington, DC: The Bureau of Justice Statistics, US Department of Justice,
May 2000.
[24] McFarlane J, Malecha A, Watson K, et al. Intimate partner sexual assault against women:
frequency, health consequences, and treatment outcome. Obstet Gynecol 2005;105:99–108.
[25] Campbell JC, Webster D, Kozol-McLain J, et al. Risk factors for femicide in abusive
relationships: results from a multisite case control study. Am J Public Health 2003;93:
1089–97.
[26] Greenfield LA, Rand MR, Craven D, et al. Violence by intimates: analysis of data on crimes
by current or previous boyfriends, girlfriends, or spouses. Washington, DC: US Depart-
ment of Justice; 1998.
[27] Bornstein RF. The complex relationship between dependency and domestic violence. Am
Psychol 2006;61:595–606.
[28] Martin SL, Mackie L, Kupper LL, et al. Physical abuse of women before, during, and after
pregnancy. JAMA 2001;285:1581–4.
[29] ACOG technical bulletin.
[30] Gazmarian J, Lazorick S, Spitz A, et al. Prevalence of violence against pregnant women.
JAMA 1996;275:1915–20.
[31] Stewart DS, Cecutti A. Physical abuse during pregnancy. CMAJ 1993;149:1257–63.
[32] Evins G, Chescheir N. Prevalence of domestic violence among women seeking abortion
services. Women’s Health Issues 1996;6:204–10.
[33] Glander SS, Moore ML, Michielutte, et al. The prevalence of domestic violence among
women seeking abortion. Obstet Gynecol 1998;91:1002–6.
[34] Fisher WA, Singh SS, Shuper PA. Characteristics of women undergoing repeat induced
abortion. CMAJ 2005;172:637–41.
[35] Norton LB, Peipert JF, Lima B, et al. Battering in pregnancy: an assessment of two screen-
ing methods. Obstet Gynecol 1995;85:321–5.
[36] Weiss HB, Laawrence BA, Miller TR. Pregnancy associated assault hospitalizations.
Obstet Gynecol 2002;100:773–80.
[37] McFarlane J, Campbell JC, Sharps P, et al. Abuse during pregnancy and femicide: urgent
implications for women’s health. Obstet Gynecol 2002;100:27–36.
[38] Gunter J. Trauma in pregnancy. Contemporary therapy in obstetrics and gynecology. In:
Ransom SB, Dombrowski MP, Evans MI, et al, editors. Philadelphia (PA): W.B. Saunders
Company; 2002. p. 128–31.
[39] Granja AC, Zacarias E, Bergstrom S. Violent deaths: the hidden face of maternal mortality.
Br J Obstet Gynaecol 2002;109:5–8.
384 GUNTER

[40] Shadigian EM, Bauer ST. Pregnancy-associated deaths: a qualitative systemic review of
homicide and suicide. Obstet Gynecol Surv 2005;60:183–90.
[41] Krulewitch CJ, Pierre-Louis ML, de Leno-Gomez R, et al. Hidden from view: violent
deaths among pregnant women in the district of Columbia, 1988–1996. J Midwifery
Womens Health 2001;46:4–10.
[42] Parsons LH, Harper MA. Violent maternal deaths in North Carolina. Obstet Gynecol
1999;94:990–3.
[43] Harper M, Parsons L. Maternal deaths due to homicide and other injuries in North Caro-
lina: 1992–1994. Obstet Gynecol 1997;90:920–3.
[44] Matud MP. Dating violence and domestic violence. J Adolesc Health 2007;40:295–7.
[45] Centers for Disease Control and Prevention. Youth risk behavior surveillance United
States, 2005. Surveillance Summaries 2005. MMWR 2006;55(No. SS-5):1–108.
[46] Munoz-Rivas MJ, Grana JL, O’Leary KD, et al. Aggression in adolescent dating relation-
ships: prevalence, justification, and health consequences. J Adolesc Health 2007;40:
298–304.
[47] Rennison CM. Bureau of Justice statistics special report: intimate partner violence and age
of victim, 1993–1999. Washington, DC: US Department of Justice, 2001.
[48] Siverman JG, Raj A, Mucci LA, et al. Dating violence against adolescent girls and associ-
ated substance abuse, unhealthy weight control, sexual risk, behavior, pregnancy, and
suicidality. JAMA 2001;286:571–9.
[49] Kennedy AC. Urban adolescent mothers exposure to community, family, and partner vio-
lence: prevalence, outcomes, and welfare policy implications. Am J Orthopsychiatry 2006;
76:44–54.
[50] Halpern CT, Oslak SG, Young ML, et al. Partner violence among adolescents in opposite-
sex romantic relationships: findings from the national longitudinal study of adolescent
health. Am J Public Health 2001;91:1679–85.
[51] Jezl DR, Molidor CE, Wright TL. Physical, sexual, and psychological abuse in high school
dating relationships: prevalence rates and self-esteem issues. Child Adolesc Soc Work J
1996;13:69–87.
[52] Jacoby M, Gorenflo D, Black E, et al. Rapid repeat pregnancy and experiences of interper-
sonal violence among low-income adolescents. Am J Prev Med 1999;16:318–21.
[53] Rennison CM. Criminal victimization, 1999. Bureau of Justice Statistics. Washington, DC:
Department of Justice; 1999.
[54] Raghavan C, Mennerich A, Sexton E, et al. Community violence and its direct, indi-
rect, and mediating effects on intimate partner violence. Violence Against Women
2006;12:1132–49.
[55] Harvey W. Homicide among young black adults: life in the subculture of exasperation. In:
Hawkins DF, editor. Homicide among black Americans. Lanham (MD): University Press;
1986. p. 153–71.
[56] Campbell D, Sharps PW, Gary F, et al. Intimate partner violence in African American
women. Online J Issues Nurs 2002;7(1):5.
[57] Wyatt G. Socio-cultural and epidemiological issues in the assessment of domestic violence.
Journal of Social Distress and the Homeless 1994;3:7–21.
[58] Bauer HM, Rosriguez MA, Quiroga SS, et al. Barriers to health care for abused Latina and
immigrant women. J Health Care Poor Underserved 2000;11:33–44.
[59] Rodriguez MA, Sheldon WR, Bauer HM, et al. The factors associated with disclosure of
intimate partner violence to clinicians. J Fam Pract 2001;50:338–44.
[60] Raj A, Silverman JG. Immigrant South Asian women at greater risk for injury from inti-
mate partner violence. Am J Public Health 2003;93:435–7.
[61] Raj A, Silverman JG, McCleary-Sills J, et al. Immigration policies increase South Asian im-
migrant women’s vulnerability to intimate partner violence. J Am Med Womens Assoc
2005;60:26–32.
 INTIMATE PARTNER VIOLENCE 385

[62] Cohen M, Maclean H. Violence against Canadian women in National Women’s Health
Surveillance Report. Journal of Society of Obstetricians and Gynaecologists of Canada
2003;25:499–504.
[63] Malcoe LH, Duran BM, Montgomery JM. Socioeconomic disparities in intimate partner
violence against Native American women: a cross-sectional study. BMC Med 2004;2:20.
[64] The National Clearinghouse on Domestic Violence. Family violence in aboriginal commu-
nities: an aboriginal perspective. [CatH7221/150–1997E]; Ottawa, Canada: Health Canada;
1997.
[65] Green K. Family violence in aboriginal communities: an aboriginal perspective. Ottawa
(Canada): National clearinghouse on family violence; 1997.
[66] Trainer C, Mihorean K, editors. Family violence in Canada: a statistical profile 2001.
Ottawa (Canada): Ministry of Industry; 2001.
[67] Brownridge DA. Partner violence against women with disabilities: prevalence, risk, and
explanations. Violence Against Women 2006;12:805–22.
[68] Cohen MM, Forte T, Du Mont J, et al. Intimate partner violence among Canadian women
with activity limitations. J Epidemiol Community Health 2005;59:8340839.
[69] Forte T, Cohen MM, Du Mont J, et al. Psychological and physical sequelae of intimate
partner violence among women with limitations in their activities of daily living. Arch
Womens Ment Health 2005;8:248–56.
[70] Wauchope BA, Strauss MA. Physical punishment and physical abuse of American chil-
dren: incidence rates by age, gender, and occupational class. Physical violence in American
families: risk factors and adaptations to violence in 8,145 families. New Brunswick (NJ):
Transaction Publishers; 1990. p. 133–48.
[71] Vest JR, Catlin TK, Chen JJ, et al. Multistate analysis of factors associated with intimate
partner violence. Am J Prev Med 2002;22:156–64.
[72] Bullock L, Bloom T, Davis J, et al. Abuse disclosure in privately and medicaid funded preg-
nant women. J Midwifery Womens Health 2006;51:361–9.
[73] Campbell J, Snow Jones A, Dienemann J, et al. Intimate partner violence and physical
health consequences. Arch Intern Med 2002;162:1157–63.
[74] Eisenstat SA, Bancroft L. Domestic violence. N Engl J Med 1999;341:886–92.
[75] McCauley J, Kern DE, Kolodner K, et al. The ‘‘battering syndrome’’: prevalence and clin-
ical characteristics of domestic violence in primary care internal medicine practices. Ann
Intern Med 1995;123:737–46.
[76] Drossman DA, Leserman J, Nachman G, et al. Sexual and physical abuse in women with
functional or organic gastrointestinal disorders. Ann Intern Med 1990;113:828–33.
[77] Paranjape A, Heron S, Thompson M, et al. Are alcohol problems linked with an increase in
depressive symptoms in abused, inner-city African American women? Womens Health
Issues 2007;17:37–43.
[78] Weisesheimer RL, Schermer CR, Malcoe LH, et al. Severe intimate partner violence and
alcohol use among female trauma patients. J Trauma 2005;58:22–9.
[79] Jones AS, Dienemann J, Schollenberger J, et al. Long-term costs of intimate partner
violence in a sample of female HMNO enrollees. Womens Health Issues 2006;16:
252–62.
[80] Silverman JG, Decker MR, Reed E, et al. Intimate partner violence victimization prior to
and during pregnancy among women residing in 26 U.S. states: associations with maternal
and neonatal health. Am J Obstet Gynecol 2006;195:140–8.
[81] Murphy CC, Schei B, Myhr TL, et al. Abuse: a risk factor for low birth weight? A systematic
review and meta-analysis. CMAJ 2001;164:1567–72.
[82] Curry MA, Perin N, Wall E. Effects of abuse on maternal complications and birth weight in
adult and adolescent women. Obstet Gynecol 1998;92:530–4.
[83] Newberger EH, Barkan SE, Lieberman ES, et al. Abuse of pregnant women and ad-
verse birth outcomes: current knowledge and implications for practice. JAMA 1992;
267:2370–2.
386 GUNTER

[84] Parker B, McFarlane J, Socken K. Abuse during pregnancy: effects of maternal complica-
tions and birth weight in adult and teenage women. Obstet Gynecol 1994;84:323–8.
[85] Bowker LH, Arbitell M, McFerron JR, On the relationship between wife beating and child
abuse. In: Yillo and Gofrad. Feminist perspectives on wife abuse 1998;158:162.
[86] National Coalition Against Domestic Violence. Children and domestic violence. Washing-
ton, DC: NCADV, July 2007. Available at: www.ncadv.org/files/childrenandchildcustody.
pdf. Accessed August 10, 2007.
[87] Family violence in Canada: A statistical profile. Canadian Center for Justice Statistics.
Minister of Industry Statistics, Canada. Catelogue no. 85-224-XIE.
[88] Kitzmann KM, Gaylord NK, Holt AR, et al. Child witnesses to domestic violence; a meta-
analytic review. J Consult Clin Psychol 2003;71:339–52.
[89] Wolfe DA, Crooks CV, McIntyre-Smith A, et al. The effects of children’s exposure to
domestic violence: a meta-analysis and critique. Clin Child Fam Psychol Rev 2003;6:
171–87.
[90] Straus RB. Supervised visitation and family violence. Fam Law Q 1995;229:232–3.
[91] Pavao J, Alvarez J, Baumrind N, et al. Intimate partner violence and housing instability.
Am J Prev Med 2007;32:143–6.
[92] Browne A, Bassuk SS. Intimate partner violence in the lives of homeless and poor housed
women: prevalence and patterns in an ethnically diverse sample. Am J Orthopsychiatry
1997;67:261–78.
[93] Sheehan MA. An interstate compact on domestic violence: what are the advantages? Juve-
nile and Family Justice Today 1993;1:12–3.
[94] American College of Obstetrics and Gynecology. Screening tools for Intimate partner vio-
lence. Available at: www.acog.org/departments/dept_notice.cfm?recno¼17&bulletin¼585.
Accessed April 1, 2007.
[95] Primary and preventative care: periodic assessments. ACOG Committee Opinion No.
357. American College of Obstetricians and Gynecologists. Obstet Gynecol 2006;108:
1615–22.
[96] American Medical Association Policy Statement on Family and Intimate Partner Violence
H-515.965. Available at: http://www.ama-assn.org/apps/pf_online/pf_online. Accessed
May 1, 2007.
[97] American Academy of Family Physicians. Family violence and abuse. Available at: http://
www.aafp.org/x16506.xml. Accessed April 3, 2007.
[98] The Family Violence Prevention Fund. National consensus guidelines on identifying and
responding to domestic violence victimization in health care settings. San Francisco
(CA): Family Violence Prevention Fund; 2004.
[99] Intimate partner violence during pregnancy, a guide for clinicians. Centers for Disease
Control and Prevention. Available at: http://cdc.gov/reproductivehealth/violence/Intimate
PartnerViolence/index.htm. Accessed May 1, 2007.
[100] Psychosocial risk factors: perinatal screening and intervention. ACOG Committee opinion
No. 343 American College of Obstetricians and Gynecologists. Obstet Gynecol
2006;469–77.
[101] Berg SO. U.S. Preventative Services Task Force. Screening for family and intimate partner
violence: recommendation statement. Ann Fam Med 2004;2:156–60.
[102] The guide to clinical preventive services 2006. Recommendations of the U.S. preventative
services task force. Available at: www.ahrq.gov/clinic/pocketgd.pdf. Accessed August 10,
2007.
[103] Wathen CN, MacMillan HL. Canadian Task Force on Preventative Health Care. Preven-
tion of violence against women. Recommendation statement from the Canadian Task
Force on Preventative Health Care. CMAJ 2003;169:582–4.
[104] Boyle SK, Schneider FD, Ivy L, et al. Patients’ advice to physicians about intervening in
family conflict. Ann Fam Med 2005;3:248–54.
 INTIMATE PARTNER VIOLENCE 387

[105] Zink T, Eler N, Jacobson J, et al. Medical management of intimate partner violence consid-
ering the stages of change: precontemplation and contemplation. Ann Fam Med 2004;2:
231–9.
[106] Chang JC, Decker M, Martin SL, et al. What happens when health care providers ask about
intimate partner violence? A description of consequences from the perspectives of female
survivors. J Am Med Womens Assoc 2003;58:76–81.
[107] Rodriguez MA, Quiroga SS, Bauer HM. Breaking the silence: battered women’s perspec-
tives on medical care. Arch Fam Med 1996;5:153–8.
[108] Wathen CN, MacMillan HL. Interventions for violence against women: a scientific review.
JAMA 2003;289:589–600.
[109] Tjaden P, Thoennes N. Stalking in America: findings from the National Violence Against
Women Survey. Washington, D.C.: National Institute of Justice; 1998. Available at: http://
www.ncjrs.gov/pdffiles/169592.pdf. Accessed May 1, 2007.
[110] Schroeder SA. What to do with a patient who smokes. JAMA 2005;294:482–7.
[111] Feldhaus KM, Koziol-McLain J, Amsbury HL, et al. Accuracy of 3 brief screening ques-
tions for detecting partner violence in the emergency department. JAMA 1997;277:
1357–61.
[112] Rhodes KV, Drum M, Anliker E, et al. Lowering the threshold for discussions of domestic
violence: a randomized controlled trial of computer screening. Arch Intern Med 2006;166:
1107–14.
[113] Peralta R, Flaming MF. Screening for intimate partner violence in a primary care setting:
the validity of ‘‘feeling safe at home’’ and prevalence results. J Am Board Fam Pract 2003;
16:525–32.
[114] McCord-Duncan EC, Floyd M, Kemp EC, et al. Detecting potential intimate partner
violence: what approach do women want? Fam Med 2006;38:416–22.
[115] Nicolaidis C, Curry AM, Ulrich Y, et al. Could we have known? A qualitative analysis of
data from women who survived and attempted homicide by an intimate partner. J Gen
Intern Med 2003;16:788–94.
[116] Feder GS, Hutson M, Ramsay J, et al. Women exposed to intimate partner violence: expec-
tations and experiences when they encounter health care professionals: a review of qualita-
tive studies. Arch Intern Med 2006;166:22–37.
[117] Fulfer JL, Tyler JJ, Choi NJ, et al. Using indirect questions to detect intimate partner vio-
lence: the SAFE-T questionnaire. J Interpers Violence 2007;22:238–49.
[118] Horan DL, Chapin J, Klein L, et al. Domestic violence screening practices of obstetrician-
gynecologists. Obstet Gynecol 1998;92:785–9.
[119] Intimate partner violence provider survey: Virginia 2006. Division of injury and violence
prevention, office of Family Health Services, Virginia Department of Health.
[120] Lal S, Walker M, MacDonald S, et al. Spouse abuse in pregnancy: a survey of physicians’
attitudes and interventions. J SOGC 1999;21:565–72.
[121] Gutmanis I, Beynon C, Tutty L, et al. Factors influencing identification and response
to intimate partner violence a survey of physicians and nurses. BMC Public Health
2007;7:12.
[122] McCloskey LA, Lichter E, Ganz ML, et al. Intimate partner violence and patient screening
across medical specialties. Acad Emerg Med 2005;12:712–22.
[123] Waalen J, Goodwin MM, Spitz AM, et al. Screening for intimate partner violence by health
care providers. Barriers and interventions. Am J Prev Med 2000;19:230–7.
[124] Short LM, Surprenant Z, Harris JM. A community-based trial of online intimate partner
violence CME. Am J Prev Med 2006;30:181–5.
[125] Kothari CL, Rhodes KV. Missed opportunities: emergency department visits by police-
identified victims of intimate partner violence. Ann Emerg Med 2006;47:190–9.
[126] Bergman MD, Brismar B. A 5-year follow-up study of 117 battered women. Am J Public
Health 1991;81:1486–9.
388 GUNTER

[127] Wadman MC, Muellman RL. Domestic violence homicides: ED use before victimization.
Am J Emerg Med 1999;17:689–91.
[128] Sharps PW, Kozoil-McLain J, Campbell J, et al. Health care providers missed opportuni-
ties for preventing femicide. Prev Med 2001;33:373–80.
[129] Dobash RE, Dobash RP, Cavanagh K, et al. Lethal and nonlethal violence against an
intimate female parter: comparing male murderers to nonlethal abusers. Violence against
women 2007;13:329–53.
[130] Aldridge ML, Browne KD. Perpetrators of spousal homicide: a review. Trauma violence
abuse 2003;4:265–76.
[131] Campbell JC. Helping women understand their risk in situations of intimate partner
violence. J Interpers Violence 2004;19:1464–77.
[132] Datner EM, Baren JM, Sites FD, et al. Universal screening for domestic violence: inability
to prove JCAHO-mandated screening makes an immediate impact. Acad Emerg Med 2002;
9:512–3.
[133] Brown-Cranstoun J. Kringen v. Boslough and Saint Vincent Hospital: a new trend for
professionals who treat victims of domestic violence? J Health Law 2000;33:629–55.
[134] Petersen R, Moracco KE, Goldstein KM, et al. Moving beyond disclosure: women’s per-
spectives on barriers and motivators to seeking assistance for intimate partner violence.
Women Health 2004;40:63–76.
[135] Chang JC, Cluss PA, Ranier L, et al. Health care interventions for intimate partner
violence: what women want. Womens Health Issues 2005;15:21–30.
[136] Chang J, Cluss P, Ranieri A, et al. What women want from health care interventions for
intimate partner violence. Abstr Acad Health Serv Res Health Policy Meet 2002;19:5.
[137] Mandatory reporting of domestic violence by health care workers. The family violence pre-
vention fund. Available at: www.endabuse.org/health/mandatoryreporting/. Accessed
April 15, 2007.
[138] Houry D, Sachs CJ, Feldhaus KM, et al. Violence-inflicted injuries: reporting laws in the
fifty states. Ann Emerg Med 2002;39:56–60.
[139] Rodriguez MA, Sheldon WR, Rao N. Abused patient’s attitudes about mandatory report-
ing of intimate partner abuse injuries to police. Women Health 2002;35:135–47.
[140] Gielen AC, O’Campo PJ, Campbell JC, et al. Women’s opinions about domestic violence
screening and mandatory reporting. Am J Prev Med 2000;19:279–85.
[141] Houry D, Feldhaus K, Thorson AC, et al. Mandatory reporting laws do not deter patients
from seeking medial care. Ann Emerg Med 1999;34:336–41.
[142] U.S. Department of Health and Human Services, Child Information Gateway. Children
and domestic violence: summary of state laws 2004. Available at: www.childwelfare.gov/
systemwide/laws_policies/statutes/domviol.cfm. Accessed April 3, 2007.
[143] McCloskey LA, Lichter E, Williams C, et al. Assessing intimate partner violence in health
care settings leads to women’s receipt of interventions and improved health. Public Health
Rep 2006;121:435–44.
[144] Babbock JC, Green CE, Robie C. Does batterers’ treatment work? A meta-analytic review
of domestic violence treatment. Clin Psychol Rev 2004;23:1023–53.
[145] Dunford FW. The San Diego Navy Experiment: an assessment of interventions for men
who assault their wives. J Consult Clin Psychol 2000;68:468–76.
 Obstet Gynecol Clin N Am
34 (2007) 389–402

Approach to the Acute Abdomen

in Pregnancy
Charlie C. Kilpatrick, MDa,b,*, Manju Monga, MDa
a
Department of Obstetrics, Gynecology and Reproductive Sciences,
University of Texas Houston Medical School Houston, TX, USA
b
Lyndon Baines Johnson Hospital, 5656 Kelley Street, Houston, TX 77002, USA

Assessment of the pregnant woman with abdominal pain should be un-

dertaken in an expedient and thorough manner. An acute abdomen may
be the result of one of many gastrointestinal, gynecologic, urologic, or ob-
stetric causes. These situations often require surgical intervention, and delay
in diagnosis and intervention only worsens the outcome for the mother and
her fetus.

Physiologic changes in pregnancy

Certain anatomic and physiologic changes specific to pregnancy may
make the cause of the pain difficult to ascertain. As the gravid uterus en-
larges, it becomes an abdominal organ at around 12 weeks’ gestation and
compresses the underlying abdominal viscera. This enlargement may
make it difficult to localize the pain and may also mask or delay peritoneal
signs [1]. The laxity of the anterior abdominal wall may also delay peritoneal
signs. Alterations in gastrointestinal function are thought to be mediated by
elevated levels of sex steroids. Progesterone decreases lower esophageal
sphincter pressure and small bowel motility [2]. A decrease in progesterone
has also been linked to a subjective increase in appetite [3]. Colonic empty-
ing slows in pregnancy but the cause is not quite as clear. A decrease in
lower esophageal sphincter pressure leads to heartburn, gastroesophageal
reflux, and even stricture formation. Delayed gastric emptying can lead to
increased gastric residual volume, and possibly to nausea and vomiting,

* Corresponding author. Lyndon Baines Johnson Hospital, 5656 Kelley Street, Houston,
TX 77002.
E-mail address: charles.c.kilpatrick@uth.tmc.edu (C.C. Kilpatrick).

0889-8545/07/$ - see front matter. Published by Elsevier Inc.

doi:10.1016/j.ogc.2007.06.002 obgyn.theclinics.com
390 KILPATRICK & MONGA

reflux, and pulmonary aspiration with general anesthesia. The slow colonic
transit time may lead to constipation and, subsequently, pain [4].
Pregnancy also affects the urologic system. The ureters become dilated as
early as the first trimester and remain dilated into the postpartum period [5].
There are two plausible explanations for this. According to the first expla-
nation, an increase in progesterone relaxes the smooth muscle of the ureter,
slowing peristalsis, and thus leading to dilatation. According to the second
explanation, the pregnant uterus may also compress the ureters, leading to
dilatation; this effect is more pronounced on the right because the overlying
colon protects the left ureter. This distension may lead to urinary stasis,
increasing not only the risk of urolithiasis but also infection.
Other physiologic changes may affect clinical presentation of abdominal
pain in pregnancy. Increased progesterone increases respiratory drive; total
minute ventilation increases because of a larger tidal volume while respira-
tory rate is unchanged [6]. Chest films frequently show an increased cardio-
thoracic ratio largely due to the gravid uterus displacement of the
diaphragm. This results in an overall decrease in functional residual capac-
ity. These changes result in an increase in Po2 and a decrease in Pco2, result-
ing in a mild respiratory alkalosis. In the third trimester of pregnancy,
normal Pco2 is 27 to 32 mm Hg, and normal pH is greater than 7.4 [7].
Cardiac output in the pregnant state increases by 17% at high altitudes
to as much as 40% at sea level [8]. The increase, which begins early in
pregnancy and peaks in the second trimester, is mostly directed to the
uterus [9]. This is accompanied by a decrease in systemic vascular resis-
tance, which leads to an increase in the resting pulse of about 10 to 15
beats per minute above baseline. Pregnancy is also associated with
a 25% increase in red cell volume and 40% increase in plasma volume
[10], which peaks around 28 to 32 weeks. These changes lead to the so-
called ‘‘physiologic anemia of pregnancy.’’ It is not uncommon to see a he-
moglobin less than 11.0 with a normal mean corpuscular volume (MCV)
and mean corpuscular hemoglobin concentration (MCHC), although the
increased demand for iron during pregnancy may manifest as an iron-de-
ficiency anemia, with a low MCV and MCHC. Given the increase in total
blood volume, if intraperitoneal hemorrhage is suspected, clinical signs of
hypotension and tachycardia indicate massive intravascular losses of at
least 25% of total blood volume.
Beyond 20 weeks’ gestation, the compressive effects of the uterus on the
inferior vena cava can lead to a decrease in venous return, subsequent
decrease in preload, and ultimately to a decrease in cardiac output. The
decrease in cardiac output can be as much as 25% to 30% [9]. This decrease
is more often seen when the patient is in a supine position and may manifest
as complaints of dizziness and syncope. Fortunately, this is easily corrected
by lateral displacement of the gravid uterus.
Hemostatic changes also add to the challenge of evaluating and caring for
pregnant women. Pregnancy produces a thrombogenic state, with two- to
 APPROACH TO THE ACUTE ABDOMEN IN PREGNANCY 391

threefold increases in fibrinogen levels. Other clotting factors, VII, VIII, IX,
X, and XII, can increase by as much as 20% to 1000%, peaking at term [11].
Levels of von Willibrand factor increase by as much as 400% at term [12].
Prothrombin and factor V levels remain unchanged while levels of factors X
and XIII decline, along with a decrease in protein S activity and subsequent
increase in resistance to activated protein C [11]. Pregnancy is therefore as-
sociated with an increased tendency for thrombosis. Use of thrombo-embo-
lism deterrent (TED) hose and sequential compression devices should be
considered in all pregnant women undergoing nonobstetric surgery during
pregnancy.
Infection may be more difficult to assess during pregnancy, as white
blood cell counts increase to a normal range of 10,000 to 14,000 cells/
mm3 [13]. In labor, white blood cell counts may be as high as 20,000 to
30,000 cells/mm3 [14]. By 1 week postpartum, the white blood cell count
should return to normal.

Diagnostic procedures
‘‘Don’t penalize her for being pregnant!’’ Never is this phrase truer than
when evaluating a pregnant woman who may require surgical intervention.
Radiologists often approach the pregnant patient with trepidation, but ra-
diologists are not alone. Among obstetricians, the use of radiologic proce-
dures is viewed with undo fear. In a study by Ratnapalan [15],
obstetricians’ perception of potential fetal harm by CT scan and conven-
tional radiograph was unrealistically high. Usually it is unnecessary delay
in diagnosis that leads to untoward outcomes. Ultrasound and MRI are
not associated with ionizing radiation, have not been shown to have any
deleterious effects on pregnancy, and should be used when feasible. While
ionizing radiation exposure can lead to cell death, carcinogenesis, and ge-
netic effects or mutations in germ cells [16], no single diagnostic radio-
graph procedure results in radiation exposure to a degree that would
threaten the well-being of the developing preembryo, embryo, or fetus, ac-
cording to the American College of Radiology [17]. Exposure to less than
5 rad has not been associated with an increase in fetal anomalies or preg-
nancy loss [18,19].
Information gleaned from atomic bomb survivors shows the greatest risk
to the fetus is exposure at 8 to 15 weeks’ gestation [16], with radiation-
induced mental retardation the highest specific potential danger. Risk in-
creases linearly as exposure rises above 20 rad. Most of the procedures or-
dered in evaluation of the pregnant woman have much lower doses than
5 rad. When possible, always shield the abdomen during diagnostic proce-
dures and counsel patients on the baseline risks of known adverse events,
such as miscarriage, genetic disease, congenital anomalies, and growth re-
striction. Listed in Table 1 are common diagnostic radiologic procedures
and the dose of ionizing radiation to the fetus [16].
392 KILPATRICK & MONGA

Table 1
Estimated fetal exposure from some common radiologic procedures
Procedure Fetal exposure
Chest radiograph (two views) 0.02–0.07 mrad
Abdominal film (single view) 100 mrad
Intravenous pyelography O1 rada
Hip film (single view) 200 mrad
Mammography 7–20 mrad
Barium enema or small bowel series 2–4 rad
CT scan of head or chest !1 rad
CT scan of abdomen and lumbar spine 3.5 rad
CT pelvimetry 250 mrad
a
Exposure depends on the number of films.
Data from American College of Obstetricians and Gynecologists. Guidelines for diagnostic
imaging during pregnancy. ACOG Committee opinion No. 299. Obset Gynecol 2004;104:649.

Anesthesia during pregnancy

Elective, nonobstetric surgery should be avoided if possible during preg-
nancy. Surgery safely delayed from the first to the second trimester avoids
the period of organogenesis and highest pregnancy loss [20]. When possible,
regional analgesia is favored over general anesthesia as the maternal mortal-
ity is 16 times higher with general [21]. The effect of anesthesia on the fetus
remains unclear without good evidence to suggest a clear relationship be-
tween outcome and type of anesthesia [22]. There is little evidence that
any drug used during general anesthesia is a proven teratogen in humans,
and this should be relayed to the patient to alleviate any anxiety she may
have before surgery [23]. There is an increased chance of pulmonary aspira-
tion and all pregnant women should be treated as though they have a full
stomach. Premedication with citrate and histamine-2 receptor blockers is
warranted.
The rate of preterm labor after nonobstetric surgery during pregnancy
tends to increase with increasing gestational age and depends on the type
and duration of the procedure. In a review of over 720,000 births during
a 9-year period in Sweden, nonobstetric surgery complicated 0.75% of
pregnancies [24]. The incidence of preterm delivery increased by 46% in
those complicated by surgery compared with those not complicated by sur-
gery. While some advocate prophylactic tocolytic therapy, others argue
that there is no benefit [25,26] and no consensus exists; each case should
be individualized. Tocolysis is not recommended in the presence of mater-
nal infection.
Intraoperative fetal monitoring has also been suggested by some, but
there are no comparative studies to suggest that this improves fetal outcome.
According to the American College of Obstetricians and Gynecologists, this
decision should be individualized and made by the surgeon and obstetrician
 APPROACH TO THE ACUTE ABDOMEN IN PREGNANCY 393

who is consulted [27]. The authors do not routinely recommend intraoper-

ative fetal monitoring. Logically, fetal heart rate monitoring is an indirect
reflection of uteroplacental blood flow, so careful attention to avoid hypo-
tension during the surgery, with the goal of maintaining systolic blood pres-
sure within 20% of baseline and a left or right uterine displacement of the
uterus off the vena cava are recommended [28]. At the authors’ institution,
pre- and postoperative fetal heart rate is documented with careful attention
to end tidal carbon dioxide and maternal blood pressure, heart rate, and ox-
ygenation during the procedure.

Laparoscopic surgery
The safety and timing of laparoscopic surgery in pregnancy is another
area where better studies are needed. Based on retrospective evaluation
and survey data, laparoscopy is comparable to laparotomy in safety during
pregnancy [24,29]. Laparoscopy is associated with decreased hospital stay,
quicker return of bowel function, less postoperative pain, quicker time to
ambulation, and smaller chance of wound infection and hernia [30]. Access
to the peritoneal cavity must be based on the size of the uterus. Some inves-
tigators suggest the use of Hasson’s trochar [31], although others feel com-
fortable with Veres needle insufflation [32]. A surgeon experienced in
laparoscopic surgery is required. In general, when planning the procedure,
an open laparoscopic procedure using Hasson’s trochar and a more upward
placement of the laparoscopic camera to a supraumbilical location appears
logical, as there has been a report of Veres needle placement into the amni-
otic cavity with insufflation at 21 plus weeks with subsequent fetal loss [33].
Otherwise, insufflation and camera placement in the midclavicular line, 1 to
2 cm inferior to the costal margin may be considered. The goal is to avoid
the gravid uterus and to limit pneumoperitoneal pressure to no more than 12
to 15 mm Hg in an attempt to decrease the likelihood of fetal acidosis. In
studies of pregnant ewes, the carbon dioxide used for insufflation was ab-
sorbed across the peritoneum into the maternal blood stream and across
the placenta, leading to fetal respiratory acidosis and ultimately hypercapnia
[34]. This can be corrected with careful anesthetic attention to maternal ven-
tilation. Some have proposed arterial blood gas determination of the mother
over routine capnography [35]. Others suggest that reliance on maternal end
tidal carbon dioxide should be sufficient, but that more invasive monitoring
may be needed in those with a history of cardiovascular or pulmonary dis-
ease [36]. Keeping the intraperitoneal pressure at 12 to 15 mm Hg may pre-
clude adequate visualization, especially in the obese patient or those with
adhesive disease from prior surgery, and must be kept in mind when plan-
ning surgery. After insufflation is performed, the placement of other trocars
depends on the procedure and the size of the gravid uterus. Besides the con-
cern of carbon dioxide absorption, the pneumoperitoneum itself may de-
crease venous return, cardiac output, and ultimately uteroplacental blood
394 KILPATRICK & MONGA

flow. As gestation progresses, the likelihood increases that the pneumoper-

itoneum will decrease venous return, cardiac output, and uteroplacental
blood flow. The optimal gestational age at which to perform laparoscopic
surgery is unclear, but an upper limit of 26 to 28 weeks has been suggested
[37]. Recently, in a case series with 18 women undergoing laparoscopy in the
third trimester, there was no fetal loss [38].
The most commonly reported indications for nonobstetric surgery in the
pregnant patient are appendicitis, cholelithiasis, persistent ovarian cyst, and
ovarian torsion [30].

Appendicitis
Appendicitis affects 1 in 1500 pregnancies and is the most common rea-
son for nonobstetrical surgical intervention in pregnancy [39]. The inci-
dence, cause, diagnosis, and management are similar to those affecting the
nonpregnant patient, with some notable exceptions. The location of the ap-
pendix has traditionally been described as rising in the peritoneal cavity as
the uterus enlarges, beginning around 12 weeks, and reaching the iliac crest
by 24 weeks [40,41]. More recently this has been challenged in a prospective
study comparing the location of the appendix in women undergoing cesar-
ean at term, in pregnant women undergoing appendectomy, and in nonpreg-
nant women undergoing appendectomy, with no difference in appendix
location among the three groups [42]. The most common presenting com-
plaint of the patient suspected of having appendicitis is right lower quadrant
pain [39]. Anorexia, nausea, and vomiting with initial periumbilical pain are
similar in the pregnant and nonpregnant state. Fever may also be present.
As discussed earlier, the white blood cell count may increase during preg-
nancy and leukocytosis does not always indicate appendicitis, but an in-
creased number of bands is more indicative of a pathologic process [1].
Careful physical examination is key to making the diagnosis. Gross perito-
neal signs with rebound and guarding are not normal in pregnancy, al-
though laxity of the anterior abdominal wall and an enlarged uterus may
delay these signs. A high clinical suspicion is therefore needed when evalu-
ating a pregnant patient for appendicitis. Delay in diagnosis remains the
leading cause of morbidity in this disease process. An unruptured appendix
is associated with a fetal loss rate of around 3% to 5% with little effect on
maternal mortality, in contrast to a fetal loss rate of 20% to 25% and ma-
ternal mortality rate of 4% with ruptured appendicitis [43,44]. When history
and physical examination are not conclusive, prompt imaging may be help-
ful; undue delay only increases fetal and maternal morbidity. Some studies
support the use of ultrasound by an experienced sonographer in the diagno-
sis of appendicitis in pregnancy. In a blinded prospective study, Poortman
and colleagues [45] found a similar sensitivity and specificity in diagnosing
appendicitis in 199 patients with the use of graded compression sonography
and unenhanced focused single-detector helical CT. Helical CT has the
 APPROACH TO THE ACUTE ABDOMEN IN PREGNANCY 395

advantage over traditional CT of less ionizing radiation to the fetus (re-

ported as 300 mrad), however, only case series describing the use of helical
CT in pregnancy have been reported. Sonography is technically more diffi-
cult, so the radiologist must be experienced. Given the dynamic nature of
the test, the pictures cannot be reliably reevaluated with ultrasound, and
a ruptured appendix is not as clearly visualized [45]. When using compres-
sion ultrasound, the diagnosis of appendicitis is made when there is a non-
compressible, blind-ended tubular structure in the right lower quadrant
greater than 6 mm in diameter. The use of ultrasound is limited to less
than 35 weeks, as the graded compression technique is not able to visualize
the appendix clearly and is less useful later in pregnancy [46]. If ultrasound is
not available or not interpretable, CT of the abdomen with oral and intra-
venous contrast can be used. This is the best studied modality for diagnosing
appendicitis, with the radiologist looking for right lower quadrant inflamma-
tion, an enlarged nonfilling tubular structure, and/or fecalith. On MRI, the
radiologist looks for an enlarged fluid-filled appendix greater than 7 mm in
diameter. Recently, retrospective studies have suggested that MRI of the ap-
pendix is useful in delineating the presence of appendicitis in pregnant
women, but the small number of patients in these studies limits the inference
that can be drawn [41,47]. The diagnosis is best made based on clinical sus-
picion by history and examination and a negative appendix at the time of sur-
gery is justifiable, especially given the morbidity in pregnancy associated with
delay in treatment. As discussed earlier, laparoscopy appears to be as safe as
laparotomy for the treatment of this disease and has become the standard at
some institutions [48]. Another benefit of diagnostic laparoscopy is that it
can decrease the number of false-positive appendectomies performed [1].

Gallbladder disease
Biliary sludge and gallstone formation is common, occurring in up to 31%
and 2% of pregnancies, respectively. While most patients remain asymptom-
atic, 28% manifest with pain [49]. It has been suggested that the increase in
sex steroids during pregnancy delays gallbladder emptying, precipitating the
development of stones. Despite this, the incidence of acute cholecystitis does
not increase during pregnancy. Biliary colic presents with episodic postpran-
dial right upper quadrant pain and abdominal ultrasound documents chole-
lithiasis. Acute cholecystitis presents with right upper quadrant pain,
anorexia, nausea, vomiting, and fever. Physical examination usually reveals
a tender right upper quadrant, and/or a positive Murphy’s sign (pain in the
right midclavicular line upon deep inspiration). Differential diagnosis in-
cludes appendicitis, hepatitis, pancreatitis, right-sided pneumonia, intraab-
dominal abscess, and, rarely, acute fatty liver of pregnancy. On laboratory
analysis, an elevated white blood cell count with the presence of bandemia,
and sometimes elevation of liver enzymes (particularly direct bilirubin) point
toward the diagnosis. Abdominal ultrasound may reveal gallstones,
396 KILPATRICK & MONGA

inflammation of the gallbladder, and dilatation of the common bile duct.

Management begins with admission to the hospital, intravenous antibiotics,
adequate hydration, and instructions to withhold liquids or solids by mouth.
Conservative management of acute cholecystitis is championed in pregnancy
unless evidence of pancreatitis, ascending cholangitis, or common bile duct
obstruction is noted. Endoscopic retrograde cholangiopancreatography
(ERCP) can be safely performed in pregnancy with little ionizing radiation
exposure to the fetus if the patient has cholangitis or pancreatitis due to
a common bile duct stone [50]. If the patient fails to respond to conservative
management, has repeated bouts of biliary colic, or has gallstone pancreatitis
or cholangitis that is not amenable to ERCP, surgical intervention should be
considered. Laparoscopic cholecystectomy during pregnancy is the most
common laparoscopic procedure performed in pregnancy, and ideally is per-
formed in the second trimester. There are several reports in the literature of
cases performed in the first trimester with few instances of fetal loss [48,51].
In a review of the literature, fetal loss is low, except in cases associated with
acute pancreatitis, suggesting the underlying disease process and not the sur-
gery itself increases mortality [30].

Other gastroenterologic conditions

Small bowel obstruction in pregnancy complicates 1 in 3000 pregnancies.
The most common causes are adhesions, followed by volvuluses, intussus-
ceptions, and hernias [1]. Presenting symptoms include nausea, vomiting,
and abdominal distension. This clinical entity should not be confused with
hyperemesis gravidarum, as delay in diagnosis and timely surgery can lead
to increased fetal and maternal mortality. Nausea and vomiting, accompa-
nied by peritoneal signs should never be considered normal in pregnancy.
In a report of the literature of 66 cases of bowel obstruction in pregnancy,
there were 4 maternal deaths [52]. The fetal mortality rate in this review was
26%. Diagnosis of bowel obstruction is made with serial examinations and
abdominal series. Initial management is conservative but with worsening
clinical symptoms, surgery should not be delayed. Careful maintenance of
fluid, electrolyte, and nutritional balance is essential. A midline vertical in-
cision to expose the peritoneal cavity is suggested and there is no place for
laparoscopy.
Pancreatitis in pregnancy complicates 1 in 3000 pregnancies, most com-
monly secondary to cholelithiasis [53]. The classic presentation includes up-
per abdominal pain, sometimes with radiation to the back and often
relieved by leaning forward, accompanied by nausea, and vomiting. Most
cases occur in the third trimester and are mild and self-limiting, but may
progress to multisystem disease [54]. Diagnosis is made based on symptoms
and elevations of pancreatic amylase and lipase. Imaging in the form of ultra-
sound to look for evidence of gallstone formation is prudent, and CT scan is
rarely needed. Treatment is usually nonoperative and supportive, with bowel
 APPROACH TO THE ACUTE ABDOMEN IN PREGNANCY 397

rest, nasogastric suction, pain medicine, and repletion of electrolytes and

fluids. In a review of 43 cases, Perdue and colleagues [52] noted that most pa-
tients did well with supportive treatment, with symptoms resolving in about 5
days, and without any maternal deaths. Surgical intervention should be
strongly considered in all trimesters for gallstone pancreatitis. In a review
of 30 patients presenting with acute pancreatitis in pregnancy, 70% of those
with gallstone pancreatitis were noted to relapse without surgery [55]. The
differential diagnosis is similar to that for acute cholecystitis in pregnancy.

Nephrolithiasis
Symptomatic nephrolithiasis complicating pregnancy is an infrequent oc-
currence, reported as 1 in 3300 deliveries in one retrospective review [56].
Symptoms include lower abdominal and flank pain, sometimes accompa-
nied by nausea and vomiting. Fever is present if there is associated upper
urinary tract infection. There may be a history of dysuria, frequency, and
often gross hematuria. Twenty percent will have a history of renal colic.
On physical examination, costovertebral angle tenderness may be elicited
and urinalysis reveals hematuria in 75% to 95% of cases. It is postulated
that the increase in blood volume and subsequent glomerular filtration
rate increases excretion of calcium. This and the previously described uri-
nary stasis appear to promote urinary calculi in pregnancy, although
some reports indicate no increase in renal colic in pregnancy [57]. Perhaps
this is explained by a propensity for stone formation but decreased symp-
tomatology due to ureteric dilatation. Management is conservative with hy-
dration, adequate analgesia, and straining the urine for calculi. Spontaneous
passage occurs in 85% of cases [58]. In evaluating for the presence of a cal-
culus, abdominal ultrasound is safe, but may not result in adequate visual-
ization because the ureters are difficult to visualize in pregnancy. The use of
the resistance index in some series has helped to increase the sensitivity in
abdominal ultrasound, but is limited to the first 48 hours. Before the proce-
dure, anti-inflammatories should be withheld [59]. If the renal arterial resis-
tance index is not diagnostic, and symptoms do not abate, a one-shot
intravenous pyelogram can be helpful in confirming the diagnosis. Radia-
tion exposure to the fetus is a tenth that of CT of the renal system [59].
Rarely is further action needed, but if symptoms do not resolve, urologic
consultation for ureteral stent placement may be necessary. Rarely should
nephrostomy tubes be required. There is minimal effect on fetal or maternal
morbidity.

Adnexal masses in pregnancy

While causes of an acute abdomen in pregnancy are frequently gastroin-
testinal, the ovary can be a source of pathology as well. Unlike gastrointes-
tinal disorders, in which delay of treatment can lead not only to fetal
398 KILPATRICK & MONGA

morbidity and mortality, but also to maternal morbidity and mortality, it is

unclear whether the persistent adnexal mass poses such a risk. The incidence
of torsion in case reports and series varies from less than 1% to 22% [60]. As
the incidence of first-trimester ultrasound increases, so does the diagnosis of
adnexal masses, ranging from 0.2% to 2.9% [61]. When followed, adnexal
masses at or larger than 5 cm in diameter visualized in the first trimester
spontaneously resolve about 70% to 85% of the time, which suggests a func-
tional nature to the cyst [62]. Using pooled data of over 65,000 women
screened for adnexal masses and followed, there were only 6 cases of torsion
requiring surgical intervention (0.01%) [62]. This information seems to war-
rant conservative management of adnexal masses incidentally found on ul-
trasound. Also, surgery for benign adnexal masses in pregnancy is
associated with a higher rate of preterm labor than that following expectant
management [60].
Ovarian torsion in pregnancy can be confused with other intraperitoneal
processes. It most often presents with lower abdominal pain that may be
waxing and waning in nature. Symptoms may appear to be out of propor-
tion to physical examination. There may be associated nausea, vomiting and
fever. An adnexal mass may be difficult to palpate later in pregnancy be-
cause of the gravid uterus. A high index of suspicion is required to make
the diagnosis. Frequently a leukocytosis is seen. The differential diagnosis
includes ectopic pregnancy, ruptured hemorrhagic cyst, appendicitis, endo-
metriomas, and degenerating fibroid. Ultrasound evaluation with Doppler
examination may aide in providing more information by indicating an ad-
nexal mass and sometimes free peritoneal fluid. The presence of Doppler
flow does not exclude the diagnosis of torsion [63]. The diagnosis is usually
made at the time of surgery by encountering an ovary and fallopian tube
with a bluish, blackish appearance. Surgery may be performed laparoscopi-
cally or by open laparotomy [64]. Previously there was concern for the risk
of ovarian vein thrombus formation at the time of torsion, especially in the
pregnant patient, with the risk of subsequent pulmonary embolus. Manually
examining the infundibulopelvic ligament at the time of surgery was recom-
mended for the presence of cords. If present, salpingoophorectomy at the
time of torsion was recommended to guard against pulmonary embolus.
More recently, the use of salpingoophorectomy in such cases has been chal-
lenged. In a series of 102 patients with adnexal torsion, of whom 25% were
pregnant, detorsion by untwisting the infundibulopelvic ligament was un-
dertaken. Laparoscopy was used in two thirds of the cases, with only 5 re-
quiring reoperation for subsequent torsion, and no documented cases of
pulmonary emboli [65].

Uterine fibroids
Uterine fibroids are another cause of abdominal pain that may compli-
cate pregnancy. Fibroids are present in 2.7% to 4% of pregnancies when
 APPROACH TO THE ACUTE ABDOMEN IN PREGNANCY 399

discovered on second-trimester ultrasound examination [66,67]. This might

be an underestimation. As the number of first-trimester ultrasounds in-
crease, it is likely that the reported incidence of fibroids in pregnancy also
increase. It was once theorized that pregnancy, and the influence of in-
creased sex steroids, namely estrogen, cause hypertrophy of uterine fibroids.
Serial ultrasound shows that most fibroids remain the same size [68] or
shrink [69] during gestation. Pain is theorized to be due to degeneration
of the fibroid as it outgrows the blood supply [70] and may require hospital-
ization in 5% to 15% of women [71]. Patients typically present complaining
of lower abdominal pain, and may have nausea, vomiting, and fever, thus
mimicking other gastrointestinal disorders. Leukocytosis may be present,
and on physical examination there is usually tenderness over the area of
the fibroid, and sometimes frank peritoneal signs [72]. Treatment is usually
conservative, including short-term (48-hour) administration of indometha-
cin [73]. Rarely is surgery required or recommended.

Summary
 Numerous physiologic changes in pregnancy may affect the presentation
of abdominal pain in pregnancy. A high index of suspicion must be used
when evaluating a pregnant patient with abdominal pain.
 General anesthesia is considered safe in pregnancy with little evidence to
suggest teratogenic or harmful effects to the fetus. Intraoperative mon-
itoring and tocolytics should be individualized with little evidence to
support their usefulness.
 Laparoscopic surgery should be performed in the second trimester when
possible and appears as safe as laparotomy, but more studies are needed
to delineate the rates of fetal loss and preterm labor.
 If indicated, diagnostic imaging should not be withheld from the preg-
nant patient.
 Appendectomy and cholecystectomy, in the hands of experienced lapa-
roscopists, appear to be safe in pregnancy.
 The reported incidence of adnexal masses and fibroids in pregnancy may
increase with increasing use of first-trimester ultrasound. Conservative
management, with surgical management postpartum, appears reason-
able in most cases.

References
[1] Parangi S, Levine D, Henry A, et al. Surgical gastrointestinal disorders during pregnancy.
Am J Surg 2007;193:223–32.
[2] Ryan JP, Pellechia D. Effect of ovarian hormone pretreatment on gallbladder motility in
vitro. Life Sci 1982;31(14):1445–9.
[3] Loprinzi CL, Ellison NM, Schaid DJ, et al. Controlled trial of megestrol acetate for the treat-
ment of cancer anorexia and cachexia. J Natl Cancer Inst 1990;82(13):1127–32.
400 KILPATRICK & MONGA

[4] Baron TH, Ramirez B, Richter JE. Gastrointestinal motility disorders during pregnancy.
Ann Intern Med 1993;118(5):366–75.
[5] Cormier CM, Canzoneri BJ, Lewis DF, et al. Urolithiasis in pregnancy: current diagnosis,
treatment, and pregnancy complications. Obstet Gynecol Surv 2006;61(11):733–41.
[6] Yannone ME, McCurdy JR, Goldfien A. Plasma progesterone levels in normal pregnancy,
labor, and the puerperium. II. Clinical data. Am J Obstet Gynecol 1968;101(8):1058–61.
[7] Lim VS, Katz AL, Lindheimer MD. Acid–base regulation in pregnancy. Am J Physiol 1976;
231(6):1764–9.
[8] Kametas NA, McAuliffe F, Krampl E, et al. Maternal cardiac function during pregnancy at
high altitude. BJOG 2004;111(10):1051–8.
[9] Stone K. Acute abdominal emergencies associated with pregnancy. Clin Obstet Gynecol
2002;45(2):553–61.
[10] Chesley LC. Plasma and red cell volumes during pregnancy. Am J Obstet Gynecol 1972;
112(3):440–50.
[11] Lockwood CJ. Pregnancy-associated changes in the hemostatic system. Clin Obstet Gynecol
2006;49(4):836–43.
[12] Bremme KA. Haemostatic changes in pregnancy. Best Pract Res Clin Haematol 2003;16(2):
153–68.
[13] Kuvin SF, Brecher G. Differential neutrophil counts in pregnancy. N Engl J Med 1962;266:
877–8.
[14] Acker DB, Johnson MP, Sachs BP, et al. The leukocyte count in labor. Am J Obstet Gynecol
1985;153(7):737–9.
[15] Ratnapalan S, Bona N, Chandra K, et al. Physicians’ perceptions of teratogenic risk associ-
ated with radiography and CT during early pregnancy. AJR Am J Roentgenol 2004;182(5):
1107–9.
[16] Guidelines for Diagnostic Imaging During Pregnancy. ACOG committee opinion Number
299. American College of Obstetricians and Gynecologists 2004;104:647–51.
[17] Gray JE. Safety of diagnostic radiology exposures. In: Janower ML, Linton WS, editors.
American College of Radiology. Radiation risk: a primer. Reston (VA): ACR; 1996. p. 15–7.
[18] Brent RL. The effect of embryonic and fetal exposure to x-ray, microwaves, and ultrasound:
counseling the pregnant and nonpregnant patient about these risks. Semin Oncol 1989;16(5):
347–68.
[19] Osei EK, Faulkner K. Fetal doses from radiological examinations. Br J Radiol 1999;72(860):
773–80.
[20] Wyatt PR, Owolabi T, Meier C, et al. Age-specific risk of fetal loss observed in a second
trimester serum screening population. Am J Obstet Gynecol 2005;192(1):240–6.
[21] Hawkins JL, Koonin LM, Palmer SK, et al. Anesthesia-related deaths during obstetric
delivery in the United States, 1979–1990. Anesthesiology 1997;86(2):277–84.
[22] O’Rourke N, Kodali BS. Laparoscopic surgery during pregnancy. Curr Opin Anaesthesiol
2006;19(3):254–9.
[23] Kuczkowski KM. Nonobstetric surgery during pregnancy: what are the risks of anesthesia?
Obstet Gynecol Surv 2004;59(1):52–6.
[24] Mazze RI, Kallen B. Reproductive outcome after anesthesia and operation during preg-
nancy: a registry study of 5405 cases. Am J Obstet Gynecol 1989;161(5):1178–85.
[25] Allen JR, Helling TS, Langenfield M. Intraabdominal surgery during pregnancy. Am J Surg
1989;158(6):567–9.
[26] Kort B, Katz VL, Watson WJ. The effect of nonobstetric operation during pregnancy. Surg
Gynecol Obstet 1993;177(4):371–6.
[27] Nonobstetric Surgery in Pregnancy. ACOG Committee Opinion Number 284. American
College of Obstetricians and Gynecologists 2003;102:431.
[28] Steinbrook RA. Anaesthesia, minimally invasive surgery and pregnancy. Best Pract Res Clin
Anaesthesiol 2002;16(1):131–43.
 APPROACH TO THE ACUTE ABDOMEN IN PREGNANCY 401

[29] Reedy MB, Kallen B, Kuehl TJ. Laparoscopy during pregnancy: a study of five fetal out-
come parameters with use of the Swedish Health Registry. Am J Obstet Gynecol 1997;
177(3):673–9.
[30] Al-Fozan H, Tulandi T. Safety and risks of laparoscopy in pregnancy. Curr Opin Obstet
Gynecol 2002;14(4):375–9.
[31] Carter JF, Soper DE. Operative laparoscopy in pregnancy. JSLS 2004;8(1):57–60.
[32] Rollins MD, Chan KJ, Price RR. Laparoscopy for appendicitis and cholelithiasis during
pregnancy: a new standard of care. Surg Endosc 2004;18(2):237–41.
[33] Friedman JD, Ramsey PS, Ramin KD, et al. Pneumoamnion and pregnancy loss after sec-
ond-trimester laparoscopic surgery. Obstet Gynecol 2002;99(3):512–3.
[34] Hunter JG, Swanstrom L, Thornburg K. Carbon dioxide pneumoperitoneum induces fetal
acidosis in a pregnant ewe model. Surg Endosc 1995;9(3):272–7 [discussion: 277–9].
[35] Curet MJ, Allen D, Josloff RK, et al. Laparoscopy during pregnancy. Arch Surg 1996;
131(5):546–50 [discussion: 550–1].
[36] Bhavani-Shankar K, Steinbrook RA, Brooks DC, et al. Arterial to end-tidal carbon dioxide
pressure difference during laparoscopic surgery in pregnancy. Anesthesiology 2000;93(2):
370–3.
[37] Fatum M, Rojansky N. Laparoscopic surgery during pregnancy. Obstet Gynecol Surv 2001;
56(1):50–9.
[38] Upadhyay A, Stanten S, Kazantsev G, et al. Laparoscopic management of a nonobstetric
emergency in the third trimester of pregnancy. Surg Endosc 2007;21(8):1344–8.
[39] Mourad J, Elliot JP, Erikson L, et al. Appendicitis in pregnancy: new information that con-
tradicts long-held clinical beliefs. Am J Obstet Gynecol 2000;182(5):1027–9.
[40] Baer J, Reis R, Arens R. Appendicitis in pregnancy with changes in position and axis of nor-
mal appendix in pregnancy. JAMA 1932;98:1359–63.
[41] Oto A, Srinivisan PN, Ernst RD, et al. Revisiting MRI for appendix location during preg-
nancy. AJR Am J Roentgenol 2006;186(3):883–7.
[42] Hodjati H, Kazerooni T. Location of the appendix in the gravid patient: a re-evaluation of
the established concept. Int J Gynaecol Obstet 2003;81(3):245–7.
[43] Firstenberg MS, Malangoni MA. Gastrointestinal surgery during pregnancy. Gastroenterol
Clin North Am 1998;27(1):73–88.
[44] Doberneck RC. Appendectomy during pregnancy. Am Surg 1985;51(5):265–8.
[45] Poortman P, Lohle PN, Schoemaker CM, et al. Comparison of CT and sonography in the
diagnosis of acute appendicitis: a blinded prospective study. AJR Am J Roentgenol 2003;
181(5):1355–9.
[46] Lim HK, Bae SH, Seo GS. Diagnosis of acute appendicitis in pregnant women: value of
sonography. AJR Am J Roentgenol 1992;159(3):539–42.
[47] Pedrosa I, Levine D, Eyvazzadeh AD, et al. MR imaging evaluation of acute appendicitis in
pregnancy. Radiology 2006;238(3):891–9.
[48] Affleck DG, Handrahan DL, Egger MJ, et al. The laparoscopic management of appendicitis
and cholelithiasis during pregnancy. Am J Surg 1999;178(6):523–9.
[49] Maringhini A, Ciambra M, Baccelliere P, et al. Biliary sludge and gallstones in pregnancy:
incidence, risk factors, and natural history. Ann Intern Med 1993;119(2):116–20.
[50] Tham TC, Vandervoort J, Wong RC, et al. Safety of ERCP during pregnancy. Am J Gastro-
enterol 2003;98(2):308–11.
[51] Muench J, Albrink M, Serafini F, et al. Delay in treatment of biliary disease during preg-
nancy increases morbidity and can be avoided with safe laparoscopic cholecystectomy.
Am Surg 2001;67(6):539–42 [discussion: 542–3].
[52] Perdue PW, Johnson HW Jr, Stafford PW. Intestinal obstruction complicating pregnancy
[review]. Am J Surg 1992;164(4):384–8.
[53] Ramin KD, Ramin SM, Richey SD, et al. Acute pancreatitis in pregnancy. Am J Obstet
Gynecol 1995;173(1):187–91.
402 KILPATRICK & MONGA

[54] Boakye MK, Macfoy D, Rice C. Alcoholic pancreatitis in pregnancy. Obstet Gynaecol 2006;
26(8):814.
[55] Swisher SG, Hunt KK, Schmit PJ, et al. Management of pancreatitis complicating preg-
nancy. Am Surg 1994;60(10):759–62.
[56] Butler EL, Dashe JS, Ramus RM. Symptomatic nephrolithiasis complicating pregnancy.
Obstet Gynecol 2000;96(5 Pt 1):753–6.
[57] Drago JR, Rohner TJ Jr, Chez RA. Management of urinary calculi in pregnancy. Urology
1982;20(6):578–81.
[58] Stothers L, Lee LM. Renal colic in pregnancy. J Urol 1992;148(5):1383–7.
[59] McAleer SJ, Loughlin KR. Nephrolithiasis and pregnancy. Curr Opin Urol 2004;14(2):
123–7.
[60] Leiserowitz GS. Managing ovarian masses during pregnancy. Obstet Gynecol Surv 2006;
61(7):463–70.
[61] Condous G, Khalid A, Okaro E, et al. Should we be examining the ovaries in pregnancy?
Prevalence and natural history of adnexal pathology detected at first-trimester sonography.
Ultrasound Obstet Gynecol 2004;24(1):62–6.
[62] Yazbek J, Salim R, Woelfer B, et al. The value of ultrasound visualization of the ovaries dur-
ing the routine 11–14 weeks nuchal translucency scan. Eur J Obstet Gynecol Reprod Biol
2006;132(2):154–8.
[63] Albayram F, Hamper UM. Ovarian and adnexal torsion: spectrum of sonographic findings
with pathologic correlation. J Ultrasound Med 2001;20(10):1083–9.
[64] Yuen PM, Ng PS, Leung PL, et al. Outcome in laparoscopic management of persistent
adnexal mass during the second trimester of pregnancy. Surg Endosc 2004;18(9):1354–7.
[65] Cohen SB, Wattiez A, Seidman DS, et al. Laparoscopy versus laparotomy for detorsion and
sparing of twisted ischemic adnexa. JSLS 2003;7(4):295–9.
[66] Exacoustos C, Rosati P. Ultrasound diagnosis of uterine myomas and complications in preg-
nancy. Obstet Gynecol 1993;82(1):97–101.
[67] Qidwai GI, Caughey AB, Jacoby AF. Obstetric outcomes in women with sonographically
identified uterine leiomyomata. Obstet Gynecol 2006;107(2 Pt 1):376–82.
[68] Muram D, Gillieson M, Walters JH. Myomas of the uterus in pregnancy: ultrasonographic
follow-up. Am J Obstet Gynecol 1980;138(1):16–9.
[69] Hammoud AO, Asaad R, Berman J, et al. Volume change of uterine myomas during preg-
nancy: do myomas really grow? J Minim Invasive Gynecol 2006;13(5):386–90.
[70] Cunningham FG, Gant NF, Leveno KJ, et al. Abnormalities of the reproductive tract. In:
Williams obstetrics. 21st edition. New York: McGraw-Hill; 2001. p. 926–32.
[71] Ouyang DW, Economy KE, Norwitz ER. Obstetric complications of fibroids. Obstet Gyne-
col Clin North Am 2006;33(1):153–69.
[72] Fogata ML, Jain KA. Degenerating cystic uterine fibroid mimics an ovarian cyst in a preg-
nant patient. J Ultrasound Med 2006;25(5):671–4.
[73] Dildy GA 3rd, Moise KJ Jr, Smith LG Jr, et al. Indomethacin for the treatment of symptom-
atic leiomyoma uteri during pregnancy. Am J Perinatol 1992;9(3):185–9.
 Obstet Gynecol Clin N Am
34 (2007) 403–419

Current Management of Ectopic

Pregnancy
Liberato V. Mukul, MD*,
Stephanie B. Teal, MD, MPH
Department of Obstetrics and Gynecology, University of Colorado at Denver and Health
Sciences Center, Academic Office 1, B198-2, 12631 East 17th Avenue,
P.O. Box 6511, Aurora, CO 80045, USA

Ectopic pregnancy, which is any pregnancy implanted outside the uterine

cavity, remains the leading cause of pregnancy-related first-trimester death
among women in the United States. Fertilization of the ovum occurs in
the fallopian tube. As the zygote divides, it becomes first a morula and
then a blastocyst, normally arriving in the uterine cavity and beginning im-
plantation on day 6 after fertilization. Anything that delays or impedes
tubal transport may allow implantation to begin while the blastocyst is still
in the tube; approximately 97% of ectopic pregnancies are tubal in location.
Ectopic pregnancies represent approximately 2% of all pregnancies [1,2].
This estimate is conservative, as the analysis did not include patients whose
condition was diagnosed and managed exclusively as outpatients. While the
incidence of ectopic pregnancy has continued to increase, the case fatality
rate has dropped from 69% in 1876 [3], to 0.35% in 1970, and to 0.05%
in 1986. The death rate for African American and other minority women
remains over double that for white women, and the highest death rate occurs
in the 15- to 19-year-old age group [4].
With documented intrauterine pregnancy, the risk of coexisting ectopic
(heterotopic pregnancy) is approximated at 1 case in 10,000 patients to 1 case
in 30,000 [5,6]. This risk increases to approximately 1 case in 100 patients
if the woman is being treated for infertility [7].

Risk factors
Risk factors for ectopic pregnancy are strongly associated with condi-
tions that cause alterations to the normal mechanism of fallopian tubal

* Corresponding author.
E-mail address: liberato.mukul@uchsc.edu (L.V. Mukul).

0889-8545/07/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ogc.2007.07.001 obgyn.theclinics.com
404 MUKUL & TEAL

transport. It is postulated that the more damage that occurs to the fallopian
tube, the higher the risk for developing an ectopic pregnancy. This damage
may result from a number of factors, such as infection, surgery, congenital
anomalies, or tumors. Many potential risk factors have been reported in the
literature, some with good evidence and others with less convincing data.
There is good evidence to support the following as risk factors for develop-
ing an ectopic pregnancy: history of previous ectopic pregnancy, previous
tubal surgery, tubal ligation, tubal pathology, in utero diethylstilbestrol
exposure, and current use of an intrauterine device (IUD) [8].
In a 1996 meta-analysis, Ankum and colleagues [8] reported an odds ratio
of 6.6 (95% CI, 5.2–8.4) with a history of a previous ectopic pregnancy.
Barnhart and colleagues [9] in 2006 confirmed previous reports that a history
of previous ectopic pregnancy was the strongest risk factor associated with
ectopic pregnancy. A history of one previous ectopic pregnancy conferred
an odds ratio of 2.98 (95% CI, 1.88–4.73) and a history of two ectopic preg-
nancies increased the risk to 16% overall (odds ratio 16.04; 95% CI, 5.39–
47.72). Table 1 presents a comparison of the odds ratios evaluated in these
two studies.
Reconstructive tubal surgery has also been shown to be a high risk factor
for ectopic pregnancy with an odds ratio of 4.7 [8]. Reconstructive tubal sur-
gery is closely linked to the underlying tubal damage caused by a previous
ectopic pregnancy or pelvic inflammatory disease. The complexity of surgi-
cal restoration of the damaged tube correlates with subsequent risks of de-
veloping an ectopic pregnancy [10]. The underlying risk factors, and not the
surgery itself, are the likely major contributing factors in these cases. Pa-
tients who have undergone tubal reanastomosis are also at risk for ectopic
pregnancy. In one study, 6.6% of patients were diagnosed with an ectopic
pregnancy after undergoing tubal reanastomosis. The same study also found
that patients who had a history of tubal occlusion by cautery were at higher
risk than those who had reversals after noncautery methods [11].
Tubal ligation failures also confer a high risk for ectopic pregnancy. The
US Collaborative Review of Sterilization prospectively followed 10,863
women electing tubal sterilization. Thirty-three percent of post-sterilization
pregnancies occurring in this population (47 out of 143 pregnancies) were
ectopic; all but 1 were tubal. The risk was highest in patients who had a tubal
ligation using bipolar cautery, and in women sterilized under the age of 30.
The risk of ectopic pregnancy in these patients was 31.9 per 1000 procedures
compared with 1.2 per 1000 procedures in patients who had a postpartum
salpingectomy [12]. The increased risk with bipolar cautery is most likely as-
sociated with fistula formation of the fallopian tube leading to subsequent
failure. There are currently no data on the risk of ectopic pregnancy after
hysteroscopic sterilization.
The use of both hormonal and nonhormonal contraceptive methods con-
fers protection against ectopic pregnancy [13]. This includes the use of both
hormonal and nonhormonal IUDs. However, should a patient get pregnant
 CURRENT MANAGEMENT OF ECTOPIC PREGNANCY 405

Table 1
Risk factors for ectopic pregnancy
Ankum Barnhart
Risk factor (odds ratio; 95% CI) (odds ratio; 95% CI)
High risk factor
Previous ectopic 6.6; 5.2–8.4 2.9; 1.9–4.7 (if O2 ectopic
pregnancy pregnancies: 16.0;
5.4–47.7)
Previous tubal surgery 4.7; 2.4–9.5 Not reported
History of tubal 9.3; 4.9–18.0 Not reported
ligation
In utero DES exposure 5.6; 2.4–13.0 Not reported
Current use of IUD 4.2–45.0 Not reported
Moderate risk factor
History of PID 2.5; 2.1–3.0 1.5; 1.1–2.1
History of infertility 2.5–21.0 Not reported
Smoking 2.5; 1.8–3.4 Not reported
History of gonorrhea 2.9; 1.9–4.4 See below
History of chlamydia 2.8; 2.0–4.0 See below
Weak or no association
Outpatient treatment Not reported 1.22; 0.6–2.6
chlamydia/gonorrhea
Sexual partners O1 2.1; 1.4–4.8 Not reported
Coitarche !18y 1.6; 1.1–2.5 Not reported
Past use of IUD 1.6; 1.4–1.8 1.1; 0.6–1.9
History of TAB 1.6; 1.0–1.6 0.99; 0.6–1.6
Nontubal surgery 1.5; 1.1–2.6 0.95; 0.67–1.4
Prior cesarean section 0.56; 0.3–1.1 Not reported
Abbreviations: DES, diethylstilbestrol; PID, pelvic inflammatory disease; TAB, threatened
abortion;
Adapted from Ankum WM, Mol BW, Van der Veen F, et al. Risk factors for ectopic preg-
nancy: a meta-analysis. Fertil Steril 1996;65(6):1093–9; and Barnhart KT, Sammel MD, Gracia
CR, et al. Risk factors for ectopic pregnancy in women with symptomatic first-trimester preg-
nancies. Fertil Steril 2006;86(1):36–43.

while using an IUD, her risk of an ectopic pregnancy rises dramatically,

with reported odds ratios of 4.2 to 45 [13,14]. Some studies have reported
a potentially small increased risk of ectopic pregnancy in past users of an
IUD, but more current, well-controlled research indicates there is no
increased risk with previous IUD use [9,13].
Previous genital tract infection is the major cause of tubal damage and
infertility. A history of previous cervical infection with Neisseria gonorrhea
or Chlamydia trachomatis and pelvic inflammatory disease has been linked
to increased risk for ectopic pregnancy [8,15]. A recent study found that
a previous history of pelvic inflammatory disease had an odds ratio of 1.5
(95% CI, 1.11–2.05) for ectopic pregnancy [9]. This study specifically looked
at patients treated for N gonorrhea or C trachomatis in the outpatient setting
versus those requiring inpatient treatment for pelvic inflammatory disease.
The investigators found that patients who received outpatient treatment
406 MUKUL & TEAL

for N gonorrhea and/or C trachomatis did not have an increased risk for ec-
topic pregnancy (odds ratio 1.22; 95% CI, 0.6–2.6). These findings suggest
that the insult to the normal tubal transport mechanism may be greater
when patients present with symptoms or findings that require inpatient man-
agement. Hillis and colleagues [15] reported that repeated chlamydia infec-
tions increased the risk for ectopic pregnancy. The odds ratio after two
infections was 2.1 and rose to 4.5 after three infections.
A history of nontubal pelvic surgery has been inconsistently reported to
confer a potential increased risk for ectopic pregnancy [16–18]. Barnhart
and colleagues [9] in 2006 found no strong association for nontubal surgery
(including cesarean section) and ectopic pregnancy. In addition, there was
also no association between a history of voluntary interruption of preg-
nancy (therapeutic abortion), regardless of number, and ectopic pregnancy.
This study did not mention appendectomy as a risk factor, but in another
study, a history of an appendectomy was more commonly reported in cases
of ectopic pregnancy [19].
Diethylstilbestrol exposure in utero has been shown to confer a ninefold
increased risk of ectopic pregnancy [20]. Other potential risk factors include
smoking, young age at coitarche, multiple sexual partners, vaginal douch-
ing, and infertility [8,21]. Many of these risk factors likely act through a com-
mon pathway of tubal damage by infectious or environmental agents.

Location
The most common location for an ectopic pregnancy is in the fallopian
tube. Other less common sites include the abdomen, ovary, cervix, and
the interstitial portion of the fallopian tube. In one study, over 95%
occurred in the fallopian tube in the following locations: ampulla (70%),
isthmus (12%), fimbria (11.1%), and interstitium/cornua (2.4%). The re-
maining sites of ectopic pregnancies were ovarian (3.2%), abdominal
(1.3%), and cervical (!1%) [22]. Identifying the location of an ectopic is
important for therapy, but may be very challenging. Ultrasound remains
the best method to diagnose location. The location of an ectopic pregnancy
may alter the approach to treatment and subsequent follow-up. Depending
on location, a combination of medical and surgical treatment may need to
be employed. This review will focus on the management and treatment of
tubal ectopic pregnancy.

Presentation
The classic triad of abdominal pain, amenorrhea, and vaginal bleeding
should always alert the clinician to evaluate for an ectopic pregnancy. Un-
fortunately the diagnosis may be quite challenging because the presentation
of an ectopic pregnancy can vary significantly. In one study, the percentage
 CURRENT MANAGEMENT OF ECTOPIC PREGNANCY 407

of patients who presented with ectopic pregnancy with abdominal pain was
98.6%, amenorrhea 74.1%, and irregular vaginal bleeding 56.4%. Abdom-
inal tenderness (97.3%) and adnexal tenderness (98%) were the most com-
mon physical findings [23]. Barnhart and colleagues [9] reported an increased
odds ratio for ectopic pregnancy in patients presenting with first-trimester
symptoms if moderate to severe bleeding (odds ratio 1.42; 95% CI,
1.04–1.93) and pain (odds ratio 1.42; 95% CI, 1.06–1.92) were present.
Although these signs and symptoms are common, the clinical presenta-
tion of ectopic pregnancy can vary significantly from the classic presenta-
tion. Physical examination findings may also reveal a change in vital
signs, such as tachycardia or orthostatic changes; cervical motion tender-
ness; adnexal/uterine tenderness (from blood irritating the peritoneal sur-
faces); or a palpable mass. Physical examination findings may also be
unremarkable or subtle. Ectopic pregnancy can also mimic other conditions,
such as spontaneous abortion, early pregnancy failure, ruptured corpus
luteal cyst, and infection. Thus, in the setting of a positive pregnancy test,
ectopic pregnancy should always be high on the clinician’s differential diag-
nosis. In clinical scenarios of patients with known high risk factors for
ectopic pregnancy, some investigators have advocated early screening for
ectopic pregnancy once they have a positive pregnancy test [24].

Diagnosis
Early diagnosis can reduce the mortality and morbidity associated with
ectopic pregnancy. Following the history and physical examination, the
two most important diagnostic tests in evaluating for an ectopic pregnancy
are transvaginal ultrasound (TVUS) and a serum human chorionic gonoda-
trophin (hCG) level. The sensitivity and specificity of combining these tests
has been reported to range from 95% to 100% [25–27].
The first step in the diagnosis of an ectopic pregnancy is to evaluate for an
intrauterine pregnancy. Confirmation of an intrauterine pregnancy almost
definitively rules out an ectopic pregnancy; the risk of a heterotopic pregnancy
is one for every 10,000 to 30,000 spontaneous pregnancies [5,6]. However, in
the setting of assisted reproductive technologies the risk can rise to 1% [7].
TVUS can identify intrauterine pregnancy at a gestation of 5.5 menstrual
weeks at nearly 100% accuracy [28]. At 4.5 to 5 weeks, the first ultrasound
marker of intrauterine pregnancy is a gestational sac with a ‘‘double decid-
ual sign’’ (double echogenic rings around the sac) [29]. The yolk sac appears
next at 5 to 6 weeks and remains until about 10 weeks. The embryo (fetal
pole) and cardiac activity can be first detected at about 5.5 to 6 weeks. A
potentially confounding ultrasound finding is a pseudosac. This is described
as a collection of fluid within the endometrial cavity that is usually localized
centrally within the uterus. This can be potentially mistaken for an intra-
uterine gestational sac. A pseudosac is the result of endometrial bleeding
from decidualized endometrium in the setting of an extrauterine pregnancy
408 MUKUL & TEAL

[30]. Unfortunately, identification of a pseudosac is not diagnostic of an ec-

topic pregnancy, has a high false-positive rate, and thus cannot be relied on
to make the diagnosis of an ectopic pregnancy [31].
In the absence of a reliable last menstrual period, the hCG level is instru-
mental in the evaluation of ectopic pregnancy. The concept of a discriminatory
zone should be used to help facilitate ultrasound findings. The discriminatory
zone is defined as the level of hCG at which an intrauterine pregnancy should
be visualized. With abdominal ultrasound, most radiologists use 6500 mIU/
mL, but this has been further refined with the use of TVUS, reducing the
discriminatory zone to 1500 to 2500 mIU/mL [30,32]. The exact cutoff to
use depends on the success of the institution in diagnosing the discriminatory
zone, the quality of the equipment, and the expertise of the sonographer.
When the hCG level has reached the discriminatory zone and an intra-
uterine pregnancy cannot be diagnosed, an extrauterine pregnancy should
be highly suspected. An exception to this would be in cases of multiple ges-
tations. Patients at risk for multiples, such as those using assisted reproduc-
tive technologies, can be carefully followed to a higher discriminatory zone
[33]. The detection of an abnormally rising or declining hCG has also aided
in the diagnosis of an ectopic pregnancy. Kadar and colleagues [34] first re-
ported on the concept of a ‘‘doubling’’ hCG in normal pregnancies every 1.4
to 2.1 days, with a minimum 66% rise in 2 days. More recently the hCG
curves have been redefined. The lower limit of a normal rise for a normal
pregnancy has been reported to be 53% in 2 days. A rise lower than this
is highly suggestive of an abnormal pregnancy [35]. While abnormally rising
hCG levels are useful to distinguish an abnormal pregnancy, normally rising
hCG levels do not rule out ectopic pregnancy. The same researchers recently
reported hCG profiles for women diagnosed with an ectopic pregnancy.
They reported that the number of women with ectopic pregnancy who expe-
rienced a rise in hCG (60%) was similar to those with a decrease in hCG
(40%) and that there was no definitive way to characterize the pattern of
hCG for women with an ectopic pregnancy [36].
In situations where there is no definitive ultrasound diagnosis of an intra-
uterine pregnancy and the hCG level is above the discriminatory zone, uter-
ine evacuation is indicated to differentiate between an early pregnancy
failure (miscarriage) and an ectopic pregnancy. In these cases, women
have an equal chance of being diagnosed either with a miscarriage or ectopic
pregnancy [37]. The same study reported that the presumed diagnosis of ec-
topic pregnancy was incorrect nearly 40% of the time. The addition of uter-
ine evacuation to the treatment algorithm (Fig. 1) can help minimize the
inadvertent administration of methotrexate to patients with early pregnancy
failures without a significant difference in complication rates or cost [38].
Uterine evacuation is superior to Pipelle endometrial biopsy in the diagnosis
of ectopic pregnancy and should be the method employed [39]. In the
absence of chorionic villi, an ectopic pregnancy is likely and medical or
surgical treatment is indicated.
 CURRENT MANAGEMENT OF ECTOPIC PREGNANCY 409

Pregnant with cramping/bleeding, hemodynamically stable

Ultrasound

EP Nondiagnostic Viable IUP Abnormal IUP

Treat Quant hCG > DZ Quant hCG < DZ PNC D&C

D&C

- CV + CV Serial hCG

Treat EP
Normal rise Normal fall Abnormal rise

TVUS when Close monitoring

hCG>DZ

Resolution

IUP EP Abnormal IUP Nondiagnostic

D&C
PNC Treat D&C

- CV + CV

Treat EP

Fig. 1. Evaluation of the symptomatic first-trimester pregnancy. CV, chorionic villi; D&C, di-
lation and curettage; DZ, discriminatory zone; EP, ectopic pregnancy; IUP, intrauterine preg-
nancy; PNC, prenatal care.

The usefulness of a single progesterone level to diagnose ectopic preg-

nancy has been debatable. During the first 8 to 10 weeks, progesterone is
produced by the corpus luteum and remains relatively stable. A progester-
one level above 25 ng/mL is usually consistent with a normal pregnancy
(97% sensitivity), while a progesterone level less than 5 ng/mL has been
shown to be 99% specific in confirming an abnormal pregnancy. Unfortu-
nately, the lower limit cannot differentiate between an early pregnancy fail-
ure and an ectopic pregnancy [40]. In 1998, a meta-analysis of 26 studies
concluded that progesterone alone is not sufficient to diagnose ectopic preg-
nancy with good reliability [41].

Treatment
After the diagnosis is made, several factors influence the decision to treat
an ectopic pregnancy medically or surgically. If the patient is unstable, then
immediate surgical treatment via laparotomy or laparoscopy is necessary. In
410 MUKUL & TEAL

the past, laparotomy with salpingectomy was considered the gold standard,
but with the availability of minimally invasive technology and increasing
physician skill, laparoscopy is now the treatment of choice [42]. Laparos-
copy is associated with a faster recovery, shorter hospitalization, reduced
overall costs, and less pain, bleeding, and adhesion formation. In a hemody-
namically stable patient, surgery is still the preferred route for heterotopic
pregnancy, tubal rupture, or imminent risk of rupture. Other indications
for surgery include no desire for or an inability to comply with medical
treatment, contraindication to methotrexate, and failure of medical treat-
ment. Surgery should also be considered for patients with conditions that
seem to predispose to failure of medical therapy, such as a tubal pregnancy
greater than 5 cm or fetal cardiac activity seen on TVUS [43,44]. These fac-
tors are considered in more detail below.

Salpingectomy versus salpingostomy

Once the decision is made to proceed to the operating room, the surgeon
must decide on the appropriate surgical technique. Often this decision must
be made in the operative suite. Thus, appropriate preoperative counseling is
important. Taking into consideration risk factors, patient desire for future
fertility, and the condition of the patient also helps guide the intraoperative
decision. Salpingectomy is the segmental or entire removal of the fallopian
tube. The indications for removing the tube include recurrent ectopic preg-
nancy in the same tube, a severely damaged tube, uncontrolled bleeding (be-
fore or after salpingostomy), heterotopic pregnancy, and lack of desire to
bear more children.
Salpingostomy is the method of choice in women of reproductive age who
wish to preserve their fertility. Salpingostomy is typically performed by mak-
ing an incision on the antemesenteric border of the fallopian tube at the point
of maximal distension. The use of vasopressin before incision has been re-
ported to reduce bleeding and operative time in some studies, but has also
been found to not be significant in others [45,46]. Removing the product of
conception by hydrodissection is recommended, along with avoiding exces-
sive handling of the tube and excessive cautery to prevent potential further
damage to the fallopian tube. The rate of intrauterine pregnancy is improved
in patients having linear salpingostomy versus salpingectomy, although the
recurrent ectopic pregnancy rate is also higher [47–49].

Persistent ectopic pregnancy

One of the potential hazards of conservative surgical management of
ectopic pregnancy with salpingostomy is persistent ectopic pregnancy. The
risk of persistent ectopic pregnancy after salpingostomy is reported to be
 CURRENT MANAGEMENT OF ECTOPIC PREGNANCY 411

2% to 11% with laparotomy and 5% to 20% with laparoscopy [32,50]. The

increased rate in patients treated by laparoscopy is thought to be associated
with the learning curve of laparoscopy. Because of the potential risk of tubal
rupture and hemorrhage, some investigators recommend following weekly
hCG serum levels to ensure complete resolution [51]. If the hCG level pla-
teaus, methotrexate is usually indicated as the first option, followed by sal-
pingectomy if medical treatment fails. Some investigators have advocated
the use of prophylactic methotrexate after salpingostomy to reduce the
risk of persistent ectopic pregnancy [52,53]. Risk factors for salpingostomy
failure, such as an ectopic pregnancy less than 2 cm, or rapidly rising preop-
erative hCG levels, may help guide the decision to administer prophylactic
methotrexate after salpingostomy [54]. Small masses, by preventing com-
plete evacuation of the ectopic pregnancy, may potentially place patients
at higher risk for persistent ectopic pregnancy.

Medical management
Before the mid-1980s treatment for ectopic pregnancy was exclusively
surgical. The first case report of methotrexate for the treatment of ectopic
pregnancy appeared in 1982 [55]. Many other agents have been used with
varying rates of success. Prostaglandins, dactinomycin, etoposide, hyperos-
molar glucose, anti-hCG antibodies, potassium chloride, and mifepristone
have all been described in the literature [56].
Methotrexate has been the most successful method of medical manage-
ment for ectopic pregnancy and is currently the medical treatment of choice.
Methotrexate for ectopic pregnancy was proposed after the observation that
actively replicating trophoblasts in gestational trophoblastic disease were
successfully treated with methotrexate [57]. Methotrexate is a folinic acid
antagonist that binds to the catalytic site of dihydrofolate reductase inhibit-
ing the synthesis of purines and pyrimidines, thus interfering with the syn-
thesis of DNA and cell replication [58].
Hemodynamically stable patients are eligible for medical management
with methotrexate. The inclusion and exclusion criteria for administration
of methotrexate are listed in Boxes 1 and 2 [59]. The initial treatment regi-
mens for ectopic pregnancy consisted of multiple doses of methotrexate with
citrovorum rescue. Stovall and colleagues [60] in 1989 demonstrated a suc-
cess rate of 96% with their multiple-dose regimen. Their protocol consisted
of intramuscular methotrexate, 1 mg/kg of actual body weight alternating
with citrovorum rescue factor 0.1 mg/kg. Methotrexate was continued
only until there was a 15% decline in the level of hCG. These investigators
then observed that most of their patients treated with the multidose regimen
had declining levels of hCG before receiving the second and/or third dose of
methotrexate [61]. This led to the publication of the development of the
single-dose regimen without citrovorum rescue [62]. Table 2 describes the
412 MUKUL & TEAL

Box 1. Criteria for receiving methotrexate

Absolute indications
 Hemodynamically stable without active bleeding or signs of
hemoperitoneum
 Patient desires future fertility
 Nonlaparoscopic diagnosis
 Patient able to return for follow-up care
 General anesthesia poses risk
 Patient has no contraindications to methotrexate
Relative indications
 Unruptured mass •3.5 cm at greatest dimension
 No fetal cardiac activity
 b-hCG limit does not exceed a predetermined value (6–15 K)

Adapted from American College of Obstetricians and Gynecologists (ACOG).

Medical management of tubal pregnancy. Int J Gynaecol Obstet 1999;65:99;
with permission.

single-dose methotrexate regimen. The single-dose protocol uses 50 mg/m2

of patient body surface area, administered intramuscularly. Lipscomb [63]
later reported the University of Tennessee’s experience with their first 315
patients treated with single-dose methotrexate and reported an overall suc-
cess rate of 91.1%.

Single-dose versus multidose protocol

There is currently no consensus as to which methotrexate protocol should
be used [59]. The overall success rate reported in the literature for both pro-
tocols is approximately 90% [64]. In a recent randomized trial of 108 pa-
tients, the success rate with a single dose was 88.9% compared with
92.6% for multidose patients [65]. This was not considered statistically sig-
nificant (odds ratio 0.64; 95% CI, 0.17–2.1) and no differences in side effect
profiles were reported. In a systematic review, women treated with the sin-
gle-dose regimen were reported to have a higher failure rate (odds ratio 4.74;
95% CI, 1.77–12.62) [66]. The data obtained for this review were from case
series and not randomized controlled studies. In addition, it is difficult to as-
certain whether there may have been selection bias between patients receiv-
ing single- versus multidose regimens. The review did confirm that success
was inversely associated with hCG levels for both protocols. Given the cur-
rent available data, the single-dose methotrexate protocol appears to have
similar efficacy and side effect profile while making the least impact on
resources of patients and providers.
 CURRENT MANAGEMENT OF ECTOPIC PREGNANCY 413

Box 2. Contraindications to medical therapy

Absolute
 Breastfeeding
 Immunodeficiency
 Abnormal creatinine (>1.3 mg/dL), aspartate aminotransferase
(twice the normal value)
 Alcoholism or liver disease
 Preexisting blood dyscrasias
 Peptic ulcer disease
 Active pulmonary disease
 Known sensitivity to methotrexate
Relative
 Gestational sac >3.5 cm
 Cardiac activity

Adapted from American College of Obstetricians and Gynecologists (ACOG).

Medical management of tubal pregnancy. Int J Gynaecol Obstet 1999;65:99;
with permission.

Predictors of success
Various predictors of success with methotrexate have been reported in
the literature. Limited and anecdotal evidence has attributed success par-
tially or entirely to such factors as hCG levels, ectopic size, fetal cardiac
activity, progesterone levels, and free peritoneal blood in the cul-de-sac.
Lipscomb and colleagues [44] reviewed their experience and reported that
high hCG and progesterone levels and, the presence of fetal cardiac activity,
were associated with higher failure rates. They further concluded that the
single best predictor for success with methotrexate was the initial hCG level.
In counseling patients who receive a single-dose methotrexate regimen, it is
important to consider the available data on failure rates (Table 3). Patients
with an hCG below 5000 mIU/mL had the best success with methotrexate.

Table 2
Single-dose methotrexate protocol
Day Therapy
0 hCG  dilation and curettage
1 hCG, aspartate aminotransferase, serum urea nitrogen/creatinine,
complete blood cell count, Rh, methotrexate (50 mg/m2)
4 hCG
7 hCG
Data from Stovall TG, Ling FW, Gray LA. Single-dose methotrexate for treatment of ec-
topic pregnancy. Obstet Gynecol 1991;77(5):754–7.
414 MUKUL & TEAL

Table 3
Success rates by hCG
Serum b-hCG Success rate
!1000 98% (118/120)
1000–1999 93% (40/43)
2000–4999 92% (90/98)
5000–9999 87% (39/45)
10,000–14,999 82% (18/22)
O15,000 68% (15/22)
Data from Lipscomb GH, McCord ML, Stovall TG, et al. Predictors of success of metho-
trexate treatment in women with tubal ectopic pregnancies. N Engl J Med 1999;341(26):
1974–8.

Patients with hCG levels between 5000 mIU/mL and 9999 mIU/mL had
failure rates of 13%, increasing to 18% with an hCG between 10,000
mIU/mL and less than 14,999 mIU/mL. Above 15,000 mIU/mL, the failure
rates rose to 32%. This study also concluded that a large ectopic and the
presence of free peritoneal blood were not associated with higher failure
rates. There is currently no set defined limit above which methotrexate
should not be administered, but based on available data, the higher failure
rates with hCG levels above 5000 mIU/mL need to be taken into
consideration.

Surveillance
Once the decision is made to proceed with medical management, it is im-
portant to counsel patients about potential side effects (Box 3) and the need
for close follow-up. The day of methotrexate administration is considered
day 1 (see Table 2). Patients receiving the single-dose protocol then need
to follow up on day 4 and 7 for additional laboratory draws and reevalua-
tion. The day-4 hCG level can plateau or rise before a decrease begins. It is
not uncommon to see a rise in the day-4 hCG level because of the continued
production of hCG from syncytiotrophoblasts, despite cessation of hor-
mone in the cytotrophoblast [67]. A study looking at the predictability of
day-4 hCG on success of methotrexate found no association with success
of treatment or the need for potential surgical intervention [68].
Many patients (33%–60%) also experience abdominal pain (‘‘separation
pain’’) 3 to 7 days after administration of methotrexate [48,69,70]. Separation
pain is thought to be secondary to tubal abortion or an expanding hematoma
within the fallopian tube [71]. This is usually self-limited and most patients can
be managed conservatively with nonsteroidal anti-inflammatory agents. Pa-
tients who report no relief with supportive measures should be immediately
evaluated to rule out tubal rupture. The majority of methotrexate-treated ec-
topic pregnancies can be associated with an increase in size by TVUS, likely
representing hematoma formation within the tube. This finding does not reli-
ably predict treatment failure unless other signs of rupture are present [72,73].
 CURRENT MANAGEMENT OF ECTOPIC PREGNANCY 415

Box 3. Side effects associated with methotrexate

Drug related
 Nausea
 Vomiting stomatitis
 Gastric distress
 Dizziness
 Reversible alopecia (rare)
 Severe neutropenia (rare)
 Pneumonitis
 Vaginal bleeding
 Increase in abdominal pain
 Increase in hCG levels from day 1 to day 4

Data from American College of Obstetricians and Gynecologists (ACOG).

Medical management of tubal pregnancy. Int J Gynaecol Obstet 1999;65:97–103.

Signs of treatment failure include significantly worsening abdominal pain

(despite change in hCG levels), signs of hemodynamic instability, less than
a 15% decline between day-4 and day-7 hCG levels, and increasing or pla-
teauing hCG levels after the first week of treatment [59]. In a study of rup-
tured ectopic pregnancies, tubal rupture was encountered more frequently in
women with no previous history of ectopic pregnancies [74], suggesting that
surveillance of patients at presumed lower risk should be just as diligent as
for patients with known risk factors. The same study also reported a rupture
rate of greater than 11% in patients with hCG levels less than 100 mIU/mL.
If no signs of treatment failure are present by day 7 and there is a decline
of 15% between day 4 and day 7, weekly hCG levels are recommended until
complete resolution (hCG !15 mIU/mL) is seen [61,63]. If on day 7 the
drop in hCG is not greater than 15% from day 4, and if the patient is clin-
ically stable, a second dose of methotrexate with weekly follow-up is sug-
gested. In general, a second dose is needed in 15% to 20% of patients,
with less than 1% requiring more than two doses [63,66]. The average
time to resolution (hCG !15 mIU/mL) for patients successfully treated
with single-dose methotrexate was 33.6 days [63].

Expectant management
Expectant management of ectopic pregnancy has been employed with
rates of reported in the range of 48% to 100%. That large gap in rates is
in part due to the differences in inclusion criteria [48,75]. In one study, ex-
pectant management was most successful (32 of 33) in women with hCG
416 MUKUL & TEAL

levels less than 175 mIU/mL [76]. In subjects with hCG greater than 175
mIU/mL, only 41 out of 74 were managed successfully. In a situation of
a clinically stable patient with hCG less than 175 mIU/mL, indeterminate
TVUS, and declining hCG levels, it may be reasonable to employ expectant
management. On the other hand, given the low complication rate of meth-
otrexate, many clinicians opt for medical treatment over expectant
management.

Summary
While mortality from ectopic pregnancy has dropped precipitously be-
cause of improved diagnostic and management techniques, it remains a sig-
nificant gynecologic emergency, and delay in diagnosis or treatment can
be catastrophic. Diagnosis rests on maintaining a high index of suspicion
for women with symptomatic complaints in the first trimester, or women
without complaints but with risk factors, such as a prior ectopic preg-
nancy, an IUD in situ, or pregnancy following assisted reproductive tech-
nology. Algorithms, such as that shown in Fig. 1, identify how combined
use of hCG measurement, TVUS, and examination of uterine contents
after confirming nonviability may be used to efficiently prevent under-
or over-treatment. Choice of the best management technique, ranging
from expectant, to outpatient medication, to conservative versus radical
surgery, is based on the patient’s clinical condition; factors related to
the ectopic, such as size, evidence of rupture, or rate of hCG rise; and
the patient’s wishes.

References
[1] Ectopic pregnancydUnited States, 1990–1992. MMWR Morb Mortal Wkly Rep 1995;
44(3):46–8.
[2] Grimes DA. The morbidity and mortality of pregnancy: still risky business. Am J Obstet
Gynecol 1994;170(5 Pt 2):1489–94.
[3] Classic pages in obstetrics and gynecology. John Stubbs Parry. Extra-uterine pregnancy: its
causes, species, pathological anatomy, clinical history, diagnosis, prognosis, and treatment.
Am J Obstet Gynecol 1974;118(1):136.
[4] Lawson HW, Atrash HK, Saftlas AF, et al. Ectopic pregnancy in the United States,
1970–1986. MMWR CDC Surveill Summ 1989;38(2):1–10.
[5] Reece EA, Petrie RH, Sirmans MF, et al. Combined intrauterine and extrauterine gestations:
a review. Am J Obstet Gynecol 1983;146(3):323–30.
[6] Condous G. Ectopic pregnancydrisk factors and diagnosis. Aust Fam Physician 2006;
35(11):854–7.
[7] Ludwig M, Kaisi M, Bauer O, et al. Heterotopic pregnancy in a spontaneous cycle: do not
forget about it! Eur J Obstet Gynecol Reprod Biol 1999;87(1):91–3.
[8] Ankum WM, Mol BW, Van der Veen F, et al. Risk factors for ectopic pregnancy: a meta-
analysis. Fertil Steril 1996;65(6):1093–9.
 CURRENT MANAGEMENT OF ECTOPIC PREGNANCY 417

[9] Barnhart KT, Sammel MD, Gracia CR, et al. Risk factors for ectopic pregnancy in women
with symptomatic first-trimester pregnancies. Fertil Steril 2006;86(1):36–43.
[10] Lavy G, Diamond MP, DeCherney AH. Ectopic pregnancy: its relationship to tubal recon-
structive surgery. Fertil Steril 1987;47(4):543–56.
[11] Seiler JC. Factors influencing the outcome of microsurgical tubal ligation reversals. Am
J Obstet Gynecol 1983;146(3):292–8.
[12] Peterson HB, Xia Z, Hughes JM, et al. The risk of ectopic pregnancy after tubal sterilization.
U.S. Collaborative Review of Sterilization Working Group. N Engl J Med 1997;336(11):762–7.
[13] Mol BW, Ankum WM, Bossuyt PM, et al. Contraception and the risk of ectopic pregnancy:
a meta-analysis. Contraception 1995;52(6):337–41.
[14] Rossing MA, Daling JR, Voigt LF, et al. Current use of an intrauterine device and risk of
tubal pregnancy. Epidemiology 1993;4(3):252–8.
[15] Hillis SD, Owens LM, Marchbanks PA, et al. Recurrent chlamydial infections increase the
risks of hospitalization for ectopic pregnancy and pelvic inflammatory disease. Am J Obstet
Gynecol 1997;176(1 Pt 1):103–7.
[16] Marchbanks PA, Annegers JF, Coulam CB, et al. Risk factors for ectopic pregnancy. A pop-
ulation-based study. JAMA 1988;259(12):1823–7.
[17] Michalas S, Minaretzis D, Tsionou C, et al. Pelvic surgery, reproductive factors and risk of
ectopic pregnancy: a case controlled study. Int J Gynaecol Obstet 1992;38(2):101–5.
[18] Parazzini F, Tozzi L, Ferraroni M, et al. Risk factors for ectopic pregnancy: an Italian case-
control study. Obstet Gynecol 1992;80(5):821–6.
[19] Nordenskjold F, Ahlgren M. Risk factors in ectopic pregnancy. Results of a population-
based case-control study. Acta Obstet Gynecol Scand 1991;70(7–8):575–9.
[20] Goldberg JM, Falcone T. Effect of diethylstilbestrol on reproductive function. Fertil Steril
1999;72(1):1–7.
[21] Tulandi T, Sammour A. Evidence-based management of ectopic pregnancy. Curr Opin
Obstet Gynecol 2000;12(4):289–92.
[22] Bouyer J, Coste J, Fernandez H, et al. Sites of ectopic pregnancy: a 10 year population-based
study of 1800 cases. Hum Reprod 2002;17(12):3224–30.
[23] Alsuleiman SA, Grimes EM. Ectopic pregnancy: a review of 147 cases. J Reprod Med 1982;
27(2):101–6.
[24] Mol BW, Hajenius PJ, Ankum WM, et al. Screening for ectopic pregnancy in symptom-free
women at increased risk. Obstet Gynecol 1997;89(5 Pt 1):704–7.
[25] Aleem FA, DeFazio M, Gintautas J. Endovaginal sonography for the early diagnosis of
intrauterine and ectopic pregnancies. Hum Reprod 1990;5(6):755–8.
[26] Ankum WM, Van der Veen F, Hamerlynck JV, et al. Laparoscopy: a dispensable tool in the
diagnosis of ectopic pregnancy? Hum Reprod 1993;8(8):1301–6.
[27] Cacciatore B, Ylostalo P, Stenman UH, et al. Suspected ectopic pregnancy: ultrasound find-
ings and hCG levels assessed by an immunofluorometric assay. Br J Obstet Gynaecol 1988;
95(5):497–502.
[28] Gracia CR, Barnhart KT. Diagnosing ectopic pregnancy: decision analysis comparing six
strategies. Obstet Gynecol 2001;97(3):464–70.
[29] Bradley WG, Fiske CE, Filly RA. The double sac sign of early intrauterine pregnancy: use in
exclusion of ectopic pregnancy. Radiology 1982;143(1):223–6.
[30] Seeber BE, Barnhart KT. Suspected ectopic pregnancy. Obstet Gynecol 2006;107(2 Pt 1):
399–413.
[31] Ahmed AA, Tom BD, Calabrese P. Ectopic pregnancy diagnosis and the pseudo-sac. Fertil
Steril 2004;81(5):1225–8.
[32] Fylstra DL. Tubal pregnancy: a review of current diagnosis and treatment. Obstet Gynecol
Surv 1998;53(5):320–8.
[33] Kadar N, Bohrer M, Kemmann E, et al. The discriminatory human chorionic gonadotropin
zone for endovaginal sonography: a prospective, randomized study. Fertil Steril 1994;61(6):
1016–20.
418 MUKUL & TEAL

[34] Kadar N, Caldwell BV, Romero R. A method of screening for ectopic pregnancy and its
indications. Obstet Gynecol 1981;58(2):162–6.
[35] Barnhart KT, Sammel MD, Rinaudo PF, et al. Symptomatic patients with an early viable
intrauterine pregnancy: HCG curves redefined. Obstet Gynecol 2004;104(1):50–5.
[36] Silva C, Sammel MD, Zhou L, et al. Human chorionic gonadotropin profile for women with
ectopic pregnancy. Obstet Gynecol 2006;107(3):605–10.
[37] Barnhart KT, Katz I, Hummel A, et al. Presumed diagnosis of ectopic pregnancy. Obstet
Gynecol 2002;100(3):505–10.
[38] Ailawadi M, Lorch SA, Barnhart KT. Cost-effectiveness of presumptively medically treating
women at risk for ectopic pregnancy compared with first performing a dilatation and curet-
tage. Fertil Steril 2005;83(2):376–82.
[39] Barnhart KT, Gracia CR, Reindl B, et al. Usefulness of pipelle endometrial biopsy in the di-
agnosis of women at risk for ectopic pregnancy. Am J Obstet Gynecol 2003;188(4):906–9.
[40] McCord ML, Muram D, Buster JE, et al. Single serum progesterone as a screen for ectopic
pregnancy: exchanging specificity and sensitivity to obtain optimal test performance. Fertil
Steril 1996;66(4):513–6.
[41] Mol BW, Lijmer JG, Ankum WM, et al. The accuracy of single serum progesterone measure-
ment in the diagnosis of ectopic pregnancy: a meta-analysis. Hum Reprod 1998;13(11):
3220–7.
[42] Hajenius PJ, Mol BW, Bossuyt PM, et al. Interventions for tubal ectopic pregnancy.
Cochrane Database Syst Rev 2000;2:CD000324.
[43] Tulandi T. Ectopic pregnancy. Semin Reprod Med 2007;25(2):83–4.
[44] Lipscomb GH, McCord ML, Stovall TG, et al. Predictors of success of methotrexate treat-
ment in women with tubal ectopic pregnancies. N Engl J Med 1999;341(26):1974–8.
[45] Vermesh M, Silva PD, Rosen GF, et al. Management of unruptured ectopic gestation by lin-
ear salpingostomy: a prospective, randomized clinical trial of laparoscopy versus laparot-
omy. Obstet Gynecol 1989;73(3 Pt 1):400–4.
[46] Ugur M, Yesilyurt H, Soysal S, et al. Prophylactic vasopressin during laparoscopic salpin-
gotomy for ectopic pregnancy. J Am Assoc Gynecol Laparosc 1996;3(3):365–8.
[47] Mol BW, Matthijsse HC, Tinga DJ, et al. Fertility after conservative and radical surgery for
tubal pregnancy. Hum Reprod 1998;13(7):1804–9.
[48] Yao M, Tulandi T. Current status of surgical and nonsurgical management of ectopic preg-
nancy. Fertil Steril 1997;67(3):421–33.
[49] Job-Spira N, Bouyer J, Pouly JL, et al. Fertility after ectopic pregnancy: first results of a pop-
ulation-based cohort study in France. Hum Reprod 1996;11(1):99–104.
[50] Seifer DB, Diamond MP, DeCherney AH. Persistent ectopic pregnancy. Obstet Gynecol
Clin North Am 1991;18(1):153–9.
[51] Farquhar CM. Ectopic pregnancy. Lancet 2005;366(9485):583–91.
[52] Gracia CR, Brown HA, Barnhart KT. Prophylactic methotrexate after linear salpingos-
tomy: a decision analysis. Fertil Steril 2001;76(6):1191–5.
[53] Graczykowski JW, Mishell DR Jr. Methotrexate prophylaxis for persistent ectopic preg-
nancy after conservative treatment by salpingostomy. Obstet Gynecol 1997;89(1):118–22.
[54] Kemmann E, Trout S, Garcia A. Can we predict patients at risk for persistent ectopic preg-
nancy after laparoscopic salpingotomy? J Am Assoc Gynecol Laparosc 1994;1(2):122–6.
[55] Tanaka T, Hayashi H, Kutsuzawa T, et al. Treatment of interstitial ectopic pregnancy with
methotrexate: report of a successful case. Fertil Steril 1982;37(6):851–2.
[56] Carson SA, Buster JE. Ectopic pregnancy. N Engl J Med 1993;329(16):1174–81.
[57] Sand PK, Stubblefield PA, Ory SJ. Methotrexate inhibition of normal trophoblasts in vitro.
Am J Obstet Gynecol 1986;155(2):324–9.
[58] Barnhart K, Coutifaris C, Esposito M. The pharmacology of methotrexate. Expert Opin
Pharmacother 2001;2(3):409–17.
[59] ACOG. Medical management of tubal pregnancy. 2007 Compendium of Selected Publica-
tion, 1998.
 CURRENT MANAGEMENT OF ECTOPIC PREGNANCY 419

[60] Stovall TG, Ling FW, Buster JE. Outpatient chemotherapy of unruptured ectopic preg-
nancy. Fertil Steril 1989;51(3):435–8.
[61] Lipscomb GH. Medical therapy for ectopic pregnancy. Semin Reprod Med 2007;25(2):93–8.
[62] Stovall TG, Ling FW, Gray LA. Single-dose methotrexate for treatment of ectopic preg-
nancy. Obstet Gynecol 1991;77(5):754–7.
[63] Lipscomb GH, Bran D, McCord ML, et al. Analysis of three hundred fifteen ectopic preg-
nancies treated with single-dose methotrexate. Am J Obstet Gynecol 1998;178(6):1354–8.
[64] Lipscomb GH, Givens VM, Meyer NL, et al. Comparison of multidose and single-dose
methotrexate protocols for the treatment of ectopic pregnancy. Am J Obstet Gynecol
2005;192(6):1844–7 [discussion: 1847–8].
[65] Alleyassin A, Khademi A, Aghahosseini M, et al. Comparison of success rates in the medical
management of ectopic pregnancy with single-dose and multiple-dose administration of
methotrexate: a prospective, randomized clinical trial. Fertil Steril 2006;85(6):1661–6.
[66] Barnhart KT, Gosman G, Ashby R, et al. The medical management of ectopic pregnancy:
a meta-analysis comparing ‘‘single dose’’ and ‘‘multidose’’ regimens. Obstet Gynecol
2003;101(4):778–84.
[67] Thompson GR, O’Shea RT, Harding A. Beta HCG levels after conservative treatment of
ectopic pregnancy: is a plateau normal? Aust N Z J Obstet Gynaecol 1994;34(1):96–8.
[68] Gabbur N, Sherer DM, Hellmann M, et al. Do serum beta-human chorionic gonadotropin
levels on day 4 following methotrexate treatment of patients with ectopic pregnancy predict
successful single-dose therapy? Am J Perinatol 2006;23(3):193–6.
[69] Stovall TG, Ling FW. Single-dose methotrexate: an expanded clinical trial. Am J Obstet
Gynecol 1993;168(6 Pt 1):1759–62 [discussion: 1762–5].
[70] Lipscomb GH, Stovall TG, Ling FW. Nonsurgical treatment of ectopic pregnancy. N Engl
J Med 2000;343(18):1325–9.
[71] Lipscomb GH, Puckett KJ, Bran D, et al. Management of separation pain after single-dose
methotrexate therapy for ectopic pregnancy. Obstet Gynecol 1999;93(4):590–3.
[72] Brown DL, Felker RE, Stovall TG, et al. Serial endovaginal sonography of ectopic pregnan-
cies treated with methotrexate. Obstet Gynecol 1991;77(3):406–9.
[73] Atri M, Bret PM, Tulandi T, et al. Ectopic pregnancy: evolution after treatment with trans-
vaginal methotrexate. Radiology 1992;185(3):749–53.
[74] Saxon D, Falcone T, Mascha EJ, et al. A study of ruptured tubal ectopic pregnancy. Obstet
Gynecol 1997;90(1):46–9.
[75] Stovall TG, Ling FW. Expectant management of ectopic pregnancy. Obstet Gynecol Clin
North Am 1991;18(1):135–44.
[76] Elson J, Tailor A, Banerjee S, et al. Expectant management of tubal ectopic pregnancy: pre-
diction of successful outcome using decision tree analysis. Ultrasound Obstet Gynecol 2004;
23(6):552–6.
 Obstet Gynecol Clin N Am
34 (2007) 421–441

Postpartum Hemorrhage
Yinka Oyelese, MD*, William E. Scorza, MD,
Ricardo Mastrolia, MD, John C. Smulian, MD, MPH
Division of Maternal-Fetal Medicine, Department of Obstetrics, Gynecology,
and Reproductive Sciences, University of Medicine and Dentistry of New Jersey-Robert
Wood Johnson Medical School, Clinical Academic Building, 125 Paterson St,
New Brunswick, NJ 08901, USA

‘‘She died in childbirth.’’ These haunting words have echoed throughout

the ages. Hemorrhage probably has killed more women than any other
complication of pregnancy in the history of mankind. Annually, an esti-
mated 150,000 maternal deaths worldwide result from obstetric hemorrhage
[1,2]. The majority of these are from postpartum hemorrhage (PPH). In
countries with less developed medical facilities and limited access to blood
transfusion services, obstetric hemorrhage continues to take a tremendous
toll on women’s lives. In fact, in both Africa and Asia, PPH is the leading
cause of pregnancy-related mortality [1]. During the past century, in the
developed world, maternal deaths resulting from obstetric hemorrhage
have dropped precipitously, mainly because of the advent of blood transfu-
sions, fluid management, coagulation factor replacement, and improved
surgical techniques. A significant proportion of deaths from PPH are poten-
tially preventable. At least one study has indicated that 90% of deaths from
PPH were preventable. Thus, those caring for pregnant women must be
aware of the risk factors for PPH and be prepared to deal aggressively
with this complication when it does occur. This article focuses on the etiol-
ogy, prediction, prevention, and management of PPH.

Definition
PPH traditionally has been defined as blood loss in excess of 500 mL after
a vaginal delivery and 1000 mL after a cesarean delivery. Such traditional
definitions are not that helpful, however, because studies have demonstrated

* Corresponding author.
E-mail address: yinkamd@aol.com (Y. Oyelese).

0889-8545/07/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ogc.2007.06.007 obgyn.theclinics.com
422 OYELESE et al

that the average blood loss is about 500 mL at a vaginal delivery and 1000 mL
at cesarean delivery [3]. Furthermore, there is consistent evidence that
obstetricians frequently underestimate blood loss at delivery. Using the
traditional definitions, at least one half of deliveries would be categorized
as having PPH. Perhaps a more useful definition of PPH would include
blood loss sufficient to cause symptoms of hypovolemia, a 10% drop in
the hematocrit after delivery or to require transfusion of blood products
[4]. Such loss occurs in approximately 4% of vaginal deliveries and 6% of
cesarean deliveries [5]. The majority of PPH occurs within the first 24 hours
after delivery and is called ‘‘primary PPH.’’ Secondary PPH occurs between
24 hours and 6 weeks after delivery.

Clinical implications
PPH is associated with significant morbidity and mortality. In fact, it is
the leading cause of death in pregnancy worldwide and is second only to
thromboembolic events in Europe and North America. Hypovolemic shock,
blood transfusion and its attendant complications, surgical injury, fever,
renal and hepatic failure, acute respiratory distress syndrome, disseminated
intravascular coagulopathy, loss of fertility, and Sheehan’s syndrome are
among the consequences of PPH.

Relevant physiology
To understand the causes and management of PPH, it is important first
to understand the mechanisms by which excessive blood loss is prevented
during normal pregnancy. Blood flow to the gravid uterus at term is 800
to 1000 mL/min, and large amounts of blood can be lost rapidly. Without
mechanisms to minimize blood loss, maternal exsanguination could occur
rapidly. After delivery of the placenta, the uterus contracts. Because the
myometrial fibers run in different directions, contraction of these fibers
occludes blood vessels, preventing blood loss. This contraction, rather
than formation of clot or aggregation of platelets, is the major mechanism
for hemostasis after delivery. Thus, if the uterus is well contracted immedi-
ately after delivery, and hemorrhage develops, the bleeding is most likely the
consequence of a genital tract laceration or injury. Strategies to treat pri-
mary PPH first must ensure uterine contraction and then identify and repair
any genital tract injuries.

Maternal adaptation during pregnancy

Maternal blood volume expands by 40% to 50% during pregnancy, the
result of a rise in both plasma volume and red blood cell mass. This increased
blood volume to some extent protects the mother from the consequences of
 POSTPARTUM HEMORRHAGE 423

hemorrhage during and following delivery. Thus, following delivery,

a woman may lose up to 20% of her blood volume before clinical signs
become apparent. In volume-contracted conditions such as pre-eclampsia,
women may be more vulnerable to the effects of blood loss at delivery and
may decompensate more quickly.

Risk factors and etiology

Risk factors for PPH are listed in Box 1. A history of PPH in a prior
pregnancy, abnormal placentation, and operative delivery rank among the
most important risk factors [6]. More direct causes of PPH are listed in
Box 2. Essentially, they may be categorized into two groups: those in which
the uterus is not contracted, and those in which it is. By far the most
common cause of early PPH, contributing to approximately 80% of cases,
is uterine atony. If the uterus is contracted, the leading causes of primary
PPH are genital tract trauma and pathologic placentation. Secondary
PPH is caused most frequently by retained products, subinvolution of the
uterus, and uterine infection. Coagulopathy is a relatively uncommon cause
of primary PPH; it typically occurs when one of the other causes already has
produced significant blood loss. The retained dead fetus syndrome,
described in most obstetrics texts, clinically manifests about 6 weeks after fetal
death and is rarely seen in modern obstetrics. Congenital coagulation

Box 1. Risk factors for postpartum hemorrhage

Prior postpartum hemorrhage
Advanced maternal age
Multifetal gestations
Prolonged labor
Polyhydramnios
Instrumental delivery
Fetal demise
Placental abruption
Anticoagulation therapy
Multiparity
Fibroids
Prolonged use of oxytocin
Macrosomia
Cesarean delivery
Placenta previa and accreta
Chorioamnionitis
General anesthesia
424 OYELESE et al

Box 2. Causes of postpartum hemorrhage

Primary causes
Uterine atony
Genital tract lacerations
Retained products
Abnormal placentation
Coagulopathies and anticoagulation
Uterine inversion
Amniotic fluid embolism
Secondary causes
Retained products
Uterine infection
Subinvolution
Anticoagulation

disorders such as Von Willebrand’s disease or specific factor deficiencies (fac-

tors II, VII, VIII, IX, X, and XI) are uncommon individually but as a class of
disorders may be present more frequently than commonly thought.

Uterine atony
Uterine atony may result from overdistension of the uterus, as occurs
with polyhydramnios, multifetal gestations, and fetal macrosomia. Other
causes of uterine atony include the myometrial laxity that is associated
with multiparity, prolonged labor, use of large quantities of oxytocin, toco-
lytic therapy, and general anesthesia.

Genital tract trauma

Upper genital tract trauma most often is the result of uterine rupture,
which may result from separation of a prior cesarean or myomectomy scar.
There also may be bleeding from direct uterine injury at the time of cesarean
birth or through injury of associated vascular structures such as the uterine
artery or broad ligament varicosities. Lower genital tract trauma includes
perineal, cervical, or vaginal lacerations, which may occur spontaneously
or result from episiotomy, obstetric maneuvers, or operative instrumented
deliveries.

Prediction and prevention of postpartum hemorrhage

Perhaps the most important aspect of the management of PPH is its
prediction and prevention. In all pregnant women, early in pregnancy,
 POSTPARTUM HEMORRHAGE 425

a detailed history should be taken to determine whether or not the patient

has risk factors for PPH (see Box 1). In addition, the patient should be ques-
tioned regarding any religious beliefs that may lead to the patient’s declining
blood transfusions. Any history of heavy menses or bleeding abnormalities
should be noted carefully. In all women, especially those who have identified
risk factors, anemia should be corrected before delivery.
Women identified as being at risk should be delivered at a center with
facilities for blood transfusion and with properly trained obstetric and anes-
thesiology personnel. Prolonged labor should be avoided if at all possible.
Any anticoagulation agents used during pregnancy should be stopped
before the onset of labor. Large-bore (at least 18-gauge) intravenous cathe-
ters should be inserted when labor is established. Patients having protracted,
difficult labors and those who have intrapartum intraamniotic infection also
should be considered at risk of PPH. Immediately following delivery of the
placenta, uterotonic agents should be given and uterine massage performed
to minimize the chance of bleeding from uterine atony. The active manage-
ment of the third stage has been shown to reduce the risk of PPH. Fluid
replacement should be timely and adequate. The Joint Commission on
Accreditation of Health care Organizations recommends that regular clini-
cal drills be conducted to enhance the management of PPH [7]. In addition,
there is evidence that training of health care personnel on estimating blood
loss improves the accuracy of blood loss estimation [8].

Personnel in management of postpartum hemorrhage

The basic principles of PPH management involve relieving the causative
factors (especially surgically correctable injuries) and prompt replacement of
intravascular volume, blood, and coagulation factors as needed. Perhaps the
most important aspect in the management of PPH is the attitude of the
attendant in charge. It is critical to maintain equanimity in what can be
a chaotic and stressful environment. Confusion and paralysis of assistants
may result if too many orders are given at once and are not directed to spe-
cific individuals. Assistants should be designated with specific tasks; instruc-
tions should be clear, distinct, and brief. Only support staff with a crucial
role should be in the room. An excessive number of well-meaning individ-
uals increases the ambient noise, adds to confusion, and opens the door
to communication errors. The newborn infant should be removed from
the room by nursery personnel, and it usually is appropriate to have any
family members who are present accompany the infant. In massive PPH,
it is important to inform and mobilize all necessary staff. These personnel
include the most experienced obstetrician and anesthesiologist available,
the operating room staff, nursing staff, the hematologist/blood bank staff,
critical care/intensive care staff, and, where available, interventional radiol-
ogy personnel. Finally, having a readily available obstetric hemorrhage
procedure tray that contains all the instruments that could be needed for
426 OYELESE et al

the management of PPH along with personnel familiar with the instruments
may help improve outcomes [9].

Initial therapy
Prompt recognition of excessive bleeding after delivery is crucial. A
healthy woman may lose 10% to 15% of her blood volume without
a drop in blood pressure [4]. The initial finding is a very modest increase
in pulse rate. By the time her blood pressure drops appreciably, the woman
frequently has lost at least 30% of her blood volume. Thus, depending on
vital signs alone to make a diagnosis of PPH, or to determine its severity,
may be misleading. Initial therapy should be aimed at simultaneous aggres-
sive fluid and blood replacement to maintain adequate circulating volume
and direct treatment of the cause of the hemorrhage. Several wide-bore
intravenous catheters should be inserted, and aggressive volume replace-
ment should be commenced.
The first interventions should be directed toward ensuring that the uterus
is contracted. Often uterine contraction can be achieved initially by biman-
ual compression. Manual exploration of the uterus should be performed to
ensure that there are no retained secundines. The bladder should be emp-
tied, and uterotonic agents should be administered. If the uterus is well
contracted, the lower genital tract (cervix and vagina) should be examined
carefully to determine whether there are any lacerations. This examination
requires good exposure, adequate lighting, good pain relief, and a competent
assistant. This often is best done in an operating room. If genital tract
trauma is identified, and the uterus is well contracted, these lacerations
should be repaired promptly. It is important to keep up with volume
replacement.

Medical treatment of postpartum hemorrhage

Medical treatment of PPH comprises two main categories: (1) medica-
tions that cause uterine contraction, and (2) medications that promote coag-
ulation or correct abnormalities of coagulation. This discussion focuses, for
the most part, on uterotonic medications that promote uterine contraction.

Medical therapies that cause uterine contraction

Oxytocin
Oxytocin is the most common medications used to achieve uterine
contraction and thus is the first-line agent for prevention and treatment of
PPH [4]. It may be administered intramuscularly or intravenously. The par-
enteral dose is 10 mg. Oxytocin generally is well tolerated and has few side
effects, but rapid intravenous push may, rarely, contribute to hypotension.
Oxytocin also is commonly administered by an intravenous infusion of 10
 POSTPARTUM HEMORRHAGE 427

to 20 units in 1000 mL of lactated Ringer’s solution, with the infusion rate

titrated to achieve adequate uterine contraction. Oxytocin, a nonapeptide
produced in the neurohypophysis, has biologic similarity to antidiuretic hor-
mone; therefore large doses administered with large volumes of fluid may re-
sult in water toxicity.

Ergot alkaloids
Ergot alkaloids such as methylergonovine rapidly induce strong tetanic
uterine contractions. They also have been used widely as first-line agents
in the prevention and treatment of PPH [4]. They may be given orally or
parenterally. In cases of PPH, the intramuscular route is the route of choice
with dosages of up to 0.2 mg. These medications may cause significant rapid
elevation of the blood pressure and thus are contraindicated in patients who
have hypertension or pre-eclampsia. Except in very unusual circumstances,
intravenous use should be avoided.

Prostaglandins
The 15-methylated prostaglandin F2a analog carboprost is a potent
uterotonic agent that has a long duration of action. It may be administered
in a 250-mg dose intravenously, intramuscularly, or injected directly into the
myometrium. The dose may be repeated every 15 to 20 minutes up to a total
of 2 mg, although a single dose is effective in most patients. Increased doses
up to 500 mg can be used if the initial 250-mg doses are ineffective. This
prostaglandin agent may cause bronchoconstriction and elevation in blood
pressure and therefore is contraindicated in asthmatics and patients who
have hypertension. It also has significant gastrointestinal side effects and
may cause diarrhea, nausea, and vomiting as well as fever.
Misoprostol, an inexpensive, relatively new prostaglandin E1 analog, is
used in obstetrics primarily for cervical ripening and induction of labor. It
is a potent uterotonic and has been used for both the prevention and treat-
ment of PPH. Meta-analyses have found that misoprostol is less effective
than ergot alkaloids and oxytocin in the prevention of PPH and that miso-
prostol has more side effects [10–12]. Studies, however, have found that
misoprostol is highly effective in the treatment of PPH caused by uterine
atony [13–16]. Misoprostol may be administered by the oral, vaginal, or rec-
tal route [17]. The typical dosage for the treatment of PPH is 400 to 1000 mg
[14,17]. Side effects include diarrhea and fever.

Surgical therapy
Surgical therapies may be divided into four groups: (1) those that
decrease blood supply to the uterus, (2) those that remove the uterus, (3)
those aimed at causing uterine contraction or compression, and (4) those
that tamponade the uterine cavity.
428 OYELESE et al

Surgical techniques the reduce uterine blood flow

Uterine artery ligation
Uterine artery ligation is one of the easiest and most effective surgical
measures for controlling PPH refractory to initial attempts to control the
bleeding. This technique is particularly useful when excessive bleeding
occurs during cesarean section. A large curved needle with an absorbable
#1 suture is directed anterior to posterior through the myometrium, approx-
imately 1 to 2 cm medial to the broad ligament. The suture then is directed
posterior to anterior through a cleared avascular space in the broad liga-
ment close to the lateral border of the uterus and tied. The suture may be
passed from posterior to anterior if doing so facilitates an easier
approach. The suture usually is placed at the level of the internal cervical
os (which lies at the junction of the corpus and the lower uterine segment)
but, depending on ease and safety, may be placed higher or lower. The tech-
nique is a mass ligature, and the uterine artery does not have to be dissected
or mobilized. Personal experience supported by the literature has proven
efficacy in 75% of cases of severe PPH [18–20]. Successful pregnancy follow-
ing uterine artery ligation can be expected [21].
A few case reports have described a vaginal approach to uterine artery liga-
tion [22,23]. In this technique, an anterior colpotomy is created, the bladder is
reflected cephalad, and caudad traction is placed on the cervix with sponge
forceps. Traction on the cervix is maintained in a direction contralateral to
the side on which the uterine artery ligation is to be performed. The uterine
artery then is ligated at the insertion into the uterus [22,23]. Ureteral injury,
bleeding, and hematoma formation are potential complications that have
raised concerns about the safety of the operation [24]. The abdominal
approach is performed under direct visualization, has a documented low com-
plication rate, a high success rate, and a large amount of literature supporting
its validity, making it the approach of choice.

Ovarian artery ligation

The anastomosis of the ovarian vessels with the uterine vessels can be li-
gated near the insertion of the utero-ovarian ligament. Alternatively the
ovarian artery can be ligated directly between the medial margin of the
ovary and the lateral aspect of the fundus in the area of the utero-ovarian
ligament. A stepwise combination of unilateral and then bilateral ligatures
starting with the uterine artery and working to the ovarian vessels can be
an orderly and effective strategy [25].

Hypogastric artery ligation

Ligation of the internal iliac (hypogastric) artery should be performed
only by an experienced surgeon who is familiar with pelvic anatomy and,
most importantly, with the retroperitoneal course of the ureters. In the
United States this procedure is performed less often than in the past [4],
 POSTPARTUM HEMORRHAGE 429

perhaps because the procedure is more complicated and requires more time
than uterine artery ligation, has potential serious complications, and, if not
successful, may delay recourse to hysterectomy [26]. This procedure, how-
ever, is effective in perhaps two thirds of cases in which a woman wishes
to maintain her fertility [27,28]. If this procedure fails, it is important to pro-
ceed quickly to more definitive therapy (ie, hysterectomy) [29].
Several approaches can be taken to access the retroperitoneal space to
locate the anterior division of the internal iliac artery. The round ligament
can be divided, the area between the infundibulopelvic ligament and the round
ligament can be incised, direct incision into the posterior peritoneum can be
performed, with care taken to avoid the ureters, and a primary retroperitoneal
approach can be employed. The ureter is reflected medially, the areolar tissue
in the retroperitoneal space is dissected away carefully, and the branching of
the common iliac artery into its external and internal branches is identified.
The internal iliac artery should be grasped with a Babcock clamp and gently
elevated. Then a large silk suture is passed beneath the artery about 2 to
3 cm distal to the bifurcation where the anterior division of the hypogastric
artery is located. Only a blunt-tipped instrument such as a Mixter clamp
should be used to avoid a disastrous puncture of the vessels, especially the in-
ternal iliac vein. The tip of the clamp should be passed in a medial-to-lateral
direction to reduce further the likelihood of vessel injury. The suture is tied,
but the artery is not divided. It is preferable to ligate the anterior division
because ligation may decrease the amount of collateral flow that can ensue
to the area of distribution; however, this vessel is not always readily obvious.

Surgical techniques that remove the uterus

Hysterectomy
Hysterectomy is required in the management of PPH in approximately
1 in 1000 deliveries [30,31]. The procedure should be reserved for cases in
which other measures have failed, and the American College of Obstetri-
cians and Gynecologists recommends that if hysterectomy is performed
for uterine atony, there should be documentation of first attempting other
therapies [4]. In most cases of suspected placenta accreta, however, hyster-
ectomy should be the primary management, especially when the woman
does not desire future fertility [32]. Seventy percent of peripartum hysterec-
tomies follow cesarean delivery, with the remaining 30% performed after
vaginal delivery. In the past most hysterectomies were performed for uterine
atony. Now, however, the increasing frequency of placenta accreta associ-
ated with the dramatic rise in the rate of cesarean sections has made morbid
placental adherence the most common indication for peripartum hysterec-
tomy [31–35]. Even in the modern era, maternal mortality associated with
emergency peripartum hysterectomy can be as high as 5% [35,36].
The technique of peripartum hysterectomy is similar to that performed in
gynecology, but the vascular changes of pregnancy demand a significantly
430 OYELESE et al

modified technique. The blood flow to the uterus is tremendous, and minor
errors acceptable in gynecologic surgery may lead to a life-threatening situ-
ation in an obstetric hysterectomy. There is considerable potential for injury
to adjacent structures, particularly the ureters and bladder. The precise tech-
nique used depends on whether the surgery is performed with a stable
patient or in one who is rapidly losing massive quantities of blood. In the
first situation, it is good practice to keep pedicles small and ensure that
they are carefully and doubly ligated. The engorged and edematous tissues
that exist following delivery can cause vessels tied within large pedicles to
slip and retract, which may lead to massive bleeding. In the latter, more
emergent situation, rapid control of blood loss calls for quick clamping
and cutting until the bleeding is controlled or the uterus is removed. Only
when hemostasis is secured are the pedicles tied off. The risk of injury to
adjacent structures is greater when hysterectomy is performed rapidly in
a blood-filled field. Urinary tract injuries complicate 5% to 22% of peripar-
tum hysterectomies, with the bladder being the most frequently involved
structure [37,38]. Tissue malacia can develop, particularly in cases of
placenta accreta, rendering a wet-cardboard consistency to the uterus and
parametria. In cases of suspected placenta accreta, placenta previa with
prior cesarean sections, or other cases in which there is a high probability
of hemorrhage, preoperative placement of a three-way Foley catheter con-
nected to a bladder-irrigation infusion can be useful in identifying injuries
to the bladder. The drainage port can be clamped, and an infusion into
the catheter of sterile saline containing indigo carmine or sterile milk at
room temperature is commenced. A temperature difference will be noticed
between the bladder and adjacent structures. Injury may be detected by ob-
serving fluid or dye leaking into the operative field. Distending the bladder
also helps define tissue planes between the uterus, bladder, paravesical, and
parametrial areas. The authors have found large, noncrushing angulated
Glassman intestinal clamps invaluable because they prevent the tearing
into pedicles that often occurs with crushing clamps. These clamps can be
placed along almost the entire length of the lateral margin of the uterus, pro-
viding uterine traction, compressing the uterine vessels, and providing
a stopgap measure while bleeding is assessed and hemostasis is being
achieved. If a ureter is grasped in this clamp inadvertently, a crush injury
is much less likely to ensue than when crushing clamps are used [39].
Because the cervix frequently is involved with a complete placenta previa,
total hysterectomy generally is the operation of choice; however, supracer-
vical hysterectomy may be preferable, especially when the bleeding is caused
by uterine atony, when removal of the cervix is not essential for hemostasis,
or when there is difficulty maintaining the patient in stable condition. The
cervico-vaginal junction can be identified either by placing a finger through
the uterine incision and hooking the finger between the cervical rim and the
vaginal wall or by palpating the upper vagina, pinching to palpate the
cervix.
 POSTPARTUM HEMORRHAGE 431

Surgical techniques that cause uterine compression

The B-Lynch stitch and other uterine compression sutures
In 1997 Christopher B-Lynch and colleagues first reported an innovative
approach to the surgical management of PPH in a series of five patients [40].
This surgical technique is based on the principle that a contracted uterus
does not bleed. The suture is sometimes referred to as the ‘‘brace suture’’
because of its resemblance to trouser suspenders. The B-Lynch suture
aims at compressing the uterus in women in whom bimanual compression,
administration of uterotonic agents, and other early interventions have
failed. Following their initial report of the technique, B-Lynch and associ-
ates, and others, have published several reports documenting wide success
in stopping uterine hemorrhage and preventing hysterectomy [41–45]. This
technique is performed most easily at the time of cesarean section. It
requires that the uterine incision be reopened. Following vaginal delivery,
a laparotomy must be performed, and the lower uterine segment must be
opened through a transverse incision. The technique begins using a rapidly
absorbable suture on a large, curved needle, taking a bite approximately
3 cm medially from the lateral margin of the uterus and 3 cm below the
inferior edge of the uterine incision [46]. The needle then exits about 4 cm
from the lateral margin of the uterus and 3 cm above the superior edge of
the uterine incision. The suture is drawn over the serosal surface of the
fundus and then down the posterior aspect of the uterus to the level of
the uterine incision on the opposite anterior wall. A horizontal bite is taken,
entering and exiting 3 to 4 cm from the lateral margins of the uterus. Next
the suture is drawn back over the serosal surface of the fundus, down the
anterior wall, and a bite is taken 3 cm from the superior edge of the uterine
incision and 4 cm from the lateral margin. The needle exits 3 cm below the
inferior edge, approximately 3 cm from the lateral margin. The suture is tied
firmly, compressing the uterus directly. Initial reports suggested that the
procedure was safe and associated with no significant morbidity. Subse-
quently, however, there have been reports of severe uterine necrosis, infec-
tions, and other complications following this technique [47–49]. Erosion
of the suture through the uterine wall into the cervical canal also has been
described [50]. The B-Lynch suture is best used for PPH resulting from uter-
ine atony; the successful use of this technique in controlling PPH associated
with placenta previa accreta also has been described [51]. Successful term
pregnancies following the B-Lynch technique have been reported [52,53].
Similar compression techniques have been described by Ouahba and col-
leagues [54], Cho and colleagues [55], Ghezzi and colleagues [56], and Hay-
man and colleagues [57]. The hemostatic suturing technique of Cho and
colleagues [55] often is referred to as ‘‘box’’ suturing [55]. In this procedure
the anterior and posterior uterine walls are sutured together so that the
space in the uterine cavity is eliminated. At an arbitrary point in an area
of heavy bleeding, a straight needle with an absorbable suture is passed
432 OYELESE et al

through the anterior wall of the uterus, exiting on the serosal surface of the
posterior wall. The needle is reinserted several centimeters lateral to the exit
in the posterior wall and is drawn right through the uterus to the serosal sur-
face of the anterior uterine wall. The needle then is redirected 2 to 3 cm
above the second exit point, from anterior to posterior as described previ-
ously. The suturing is completed by passing the needle 2 to 3 cm to the
side of the previous exit point through the uterine walls and tied securely,
forming a box. Several of these sutures can be placed from the fundus to
the lower uterine segment, as needed. Cho and colleagues [55] reported
success with this technique, avoiding hysterectomy in 23 women who had
not responded to other conservative methods. These authors and others
also noted a return to normal fertility in women treated by this technique
[55,58]. A case report has described the formation of uterine synechiae
following this procedure [59].
Hayman [57] reported a technique that combined modifications of both
B-Lynch and Cho techniques and employed compression by suturing the
anterior and posterior walls of the uterus. In this method, which has the
advantage of not requiring that the uterus be opened after vaginal delivery,
the needle is passed from the anterior wall through the posterior wall about
2 cm medial to the lateral margin of the uterus. The suture then is tied over
the fundus. Four such sutures are placed, two on each lateral border of the
uterus. In addition, isthmic-cervical compression sutures can be placed
below the bladder reflection by driving a #2 absorbable suture on a straight
needle anterior to posterior and then reinserting the needle 2 cm medially
posterior to anterior and tying the suture. An instrument such as a clamp
can be placed between the areas to be sutured to ensure patency of the cer-
vical canal [57].

Techniques for uterine tamponade

A variety of techniques have been used to tamponade the uterine cavity.
These techniques include uterine packing [60], the umbrella pack, the Seng-
staken-Blakemore balloon, and a variety of other adapted balloons and
packs. Some obstetricians have used a large, inflated Foley catheter [61].
Condous and Arulkumaran [62] described the use of the tamponade test
to determine whether an intrauterine balloon would be effective in the man-
agement of PPH and to select patients who required further surgery. The
Sengstaken-Blakemore tube, with the stomach end cut off, was inserted
into the uterine cavity and then inflated with 75 to 150 mL of saline. If
bleeding stopped after inflation of the balloon, the woman was considered
not to require further surgery. Seror and colleagues [63] used an intrauterine
Sengstaken-Blakemore tube inflated with 250 mL of saline in 17 women who
had PPH that had not responded to conventional conservative therapy.
Hemorrhage was controlled in 71% of cases, and further surgery was
avoided in 88% of cases [63]. A similar technique using a Rusch urological
 POSTPARTUM HEMORRHAGE 433

hydrostatic catheter, which can be inflated with 500 mL of saline, has been
used successfully to manage PPH refractory to conventional therapies
[64,65]. In a variation of this technique, Bakri and colleagues [66] developed
a commercially available balloon for use in the management of PPH. These
authors claim that the balloon may be used successfully in the management
of hemorrhage caused by placenta previa.
Uterine packing also has been used successfully in controlling PPH in the
past but is used infrequently in more modern obstetrics [60,67–70]. Nonethe-
less, the technique may be very effective in stopping postpartum bleeding
and avoiding hysterectomy. The packs generally are removed 24 to 48 hours
after delivery.

Pelvic pressure packing

In 1926, Logothetopulos described a pelvic pressure pack also known as
a mushroom, umbrella, or parachute pack for the control of PPH [71]. This
pack is filled with gauze swabs and is inserted in the pelvis with the stalk
passing out into the vagina. Gravity traction is applied to the end of this
stalk, thereby pressing the pack against the pelvic vessels. This technique
is rarely used today but may have a role in massive hemorrhage that has
not to responded to other therapies [71,72]. Occasionally, when all else
has failed, packing the pelvic cavity with swabs in bags at laparotomy
with enough pressure to tamponade bleeding vessels and closing the abdom-
inal incision is effective in stopping hemorrhage. The patient may be reop-
erated on in 24 to 48 hours to remove the sponges. This procedure carries
significant risks including infection and bowel ischemia/infarction.

Uterine artery and internal iliac artery embolization

Embolization in obstetrics was described first for the control of intracta-
ble PPH [73]. The procedure, performed by an interventional radiologist,
now is used widely in obstetrics and gynecology. Numerous reports have
documented the efficacy of this technique in controlling life-threatening
PPH [74–81]. In the most common approach, the femoral artery is catheter-
ized, and the catheter is passed under fluoroscopic guidance into the anterior
branch of the internal iliac artery or into the uterine artery. These catheters
may contain balloons at their tips, which may be inflated to occlude blood
flow to the uterus. An occlusive material then is injected under fluoroscopy
until arterial flow to the uterus ceases. Typical embolic agents include
absorbable gelatin sponge and clear acrylic microspheres. Side effects and
adverse reactions include inadvertent embolization of collateral structures
leading to necrosis and gangrene, allergic reactions, and renal impairment.
Embolization requires a skilled interventional radiologist and some degree
of stability in the patient. The catheters may be placed prophylactically in
the radiology suite in patients at risk of severe hemorrhage such as those
who have placenta accreta. In general, it is considered best to wait to
434 OYELESE et al

embolize vessels until after the fetus is delivered. Embolization also may be
performed as an emergent procedure in the operating room, using a C-arm.
There have been numerous reports of successful subsequent pregnancies
after uterine or internal iliac artery embolization, although these patients
may be at risk of intrauterine growth restriction or recurrence of hemor-
rhage [79,82].

Special situations
Magnesium sulfate
Women who have received prolonged therapy with magnesium sulfate for
seizure prophylaxis in pre-eclampsia or for tocolysis may be at increased risk
for PPH caused by uterine atony. This type of PPH may not respond well to
usual pharmacologic therapies. Should hemorrhage occur in these situa-
tions, any remaining magnesium sulfate infusions should be stopped, and
calcium carbonate can be administered, which may help the myometrium
contract. Seizure prophylaxis can be resumed later if the mother has been
stabilized and there is no further bleeding.

Uterine inversion
Uterine inversion occurs in approximately 1 in 2000 deliveries and gener-
ally is the result of overenthusiastic attempts to deliver the placenta by cord
traction or fundal pressure before complete placental separation. Inversion
of the uterus may lead to massive postpartum hemorrhagic shock. The
condition is treated by aggressive fluid/blood replacement and uterine
replacement. A variety of techniques have been used to replace the uterine
fundus. These include manual replacement and the use of hydrostatic pres-
sure. Uterine replacement may require general anesthesia and uterine relax-
ant agents.

Morbid adherence of the placenta

Morbid adherence of the placenta (placenta accreta/increta/percreta) is
an increasingly common cause of severe PPH and has become the leading
cause for peripartum hysterectomy [33]. Typically the placenta does not
separate following the delivery, and attempts to separate it are accompanied
by torrential hemorrhage. A multidisciplinary team approach has the poten-
tial to reduce morbidity and mortality [32]. The key to a good outcome lies
in prenatal diagnosis and planned delivery in a center with good blood
transfusion services [32]. Placenta accreta should be suspected in any patient
who has had a prior cesarean and who has a low-lying placenta or placenta
previa. The diagnosis can be made sonographically based on the following
findings: (1) prominent echolucent vascular spaces in the placenta giving
it a ‘‘Swiss-cheese’’ appearance; (2) thinning of the placenta-myometrial
 POSTPARTUM HEMORRHAGE 435

border; (3) protrusion of the placenta into the bladder; and (4) abnormal
turbulent Doppler flow in the vascular spaces and on the surface of the
bladder [83]. It is recommended that no attempts be made to separate the
placenta [32]. The uterus should be opened through a fundal incision and
hysterectomy performed with the placenta in situ. Embolization of the
uterine or internal iliac vessels after delivery of the baby and before the hys-
terectomy may reduce blood loss greatly [32].

Transfusion therapy
The first documented successful transfusion of human blood was
performed by James Blundell in 1825 for a woman dying from PPH [84].
His interest in blood transfusion had been stimulated when he attended
a woman who died from PPH 7 years before [84]. Since that first experience,
transfusion of blood has been a critical component of life-saving resuscita-
tion in PPH.
Recommendations for transfusion based on laboratory values and
changes in vital signs alone are reasonable in a nonpregnant bleeding
patient, but the obstetric patient experiencing rapid heavy blood loss that
cannot be stemmed is subject to sudden decompensation and exsanguina-
tion. Hypovolemic shock, defined as poor tissue perfusion associated with
hypoxia, first must be treated with replacement of vascular volume. Crystal-
loid solutions such as Ringer’s lactate are readily available, inexpensive, and
easily administered. Crystalloids should be administered as a volume three
times the estimated blood loss, because they have a lower oncotic pressure
than plasma and rapidly leave the vascular tree to the extravascular space.
Although colloids have a higher oncotic pressure and can be administered
in less volume, there is little difference in clinical response, and postresusci-
tation diuresis is better with crystalloids. Life can be sustained, temporizing,
by keeping the circulating volume replete and the cardiac pump primed.
Whole blood is rarely used for transfusion, but it has several advantages.
It contains all the coagulation factors. In urgent situations, uncrossed
O-negative blood may be administered. Type-specific blood is preferable.
Packed red blood cells (PRBCs) are the primary transfusion product used
to increase the oxygen-carrying capacity. A typical volume of about
300 mL is mixed with normal saline before infusion. Diluting PRBCs with
Ringer’s lactate can cause calcium to precipitate with the citrate used as
a preservative in stored blood. A single unit of PRBCs can be expected to
raise the hemoglobin and hematocrit by 1 g and by 3%, respectively, in
a nonbleeding patient.
Fresh-frozen plasma is a secondary transfusion product indicated mainly
in states of coagulopathy or with massive transfusion. It comes in 250-mL
units and contains all the coagulation factors, especially fibrinogen. One
unit will raise the fibrinogen level by 10 mg/% in a nonbleeding patient.
It is reasonable to consider transfusing 1 unit of fresh-frozen plasma to
436 OYELESE et al

every 4 units of PRBCs in an actively bleeding patient, but the clinical cir-
cumstances guided by fibrinogen level, prothrombin time, and activated
partial thromboplastin time should dictate the amount transfused.
Cryoprecipitate is a tertiary transfusion product that contains as much
fibrinogen as a unit of fresh-frozen plasma but in a volume of only about
15 mL. It also contains factor VIII, factor XIII, and von Willebrand’s fac-
tor. It also will raise the fibrinogen level about 10 mg/% per unit. Its main
indication for transfusion is in a hemorrhaging patient who is volume
replete but has low fibrinogen levels. A large amount of fibrinogen can be
administered in a small volume using cryoprecipitate.
Platelets also are a tertiary transfusion product and are administered to
heavily bleeding patients who have thrombocytopenia. Platelets are stored
at room temperature on an oscillator in the blood bank and have a short
shelf life of 3 to 5 days. Blood banks preferably issue single-donor platelets
with a volume of about 300 mL. A unit of single-donor platelets raises the
platelet count by 30,000 to 60,000 in a nonbleeding patient. Platelet packs,
which usually consist of 6 units, are less preferred because of the increased
risk of developing platelet antibodies and blood-borne infection, but the vol-
ume and increase in platelet count are similar. The goal of platelet therapy is
to stimulate coagulation and maintain a platelet count of 50,000 to 100,000.
Developments in the field of transfusion medicine have led to new prod-
ucts that hold promise now and in the future. Human recombinant activated
factor VII (rfVII) has been approved by the Food and Drug Administration
(FDA) for treatment of bleeding associated with hemophilia A and B and
congenital factor VII deficiency. Case reports are accumulating describing
successful use of rfVII in the control of life-threatening hemorrhage after
other standard measures have failed [85–87]. It has been successful in stop-
ping hemorrhage in cases of amniotic fluid embolus, disseminated intravas-
cular coagulopathy, placenta previa, placenta accreta, uterine atony, and
hemolysis, elevated liver enzymes, and low platelets syndrome. The dose
of rfVII has varied from 16.7 to 120 mg/kg. A review of the literature
suggests that a dose of 70 to 90 mg/kg could be sufficient to stop 75% of
cases of refractory PPH [88]. Factor VII interacts with tissue factor at
a site of vascular injury; this interaction activates factors X and IX, leading
to a burst of thrombin that in turn leads to a functioning fibrin clot. Platelet-
dependant clotting mechanisms also are stimulated by rfVII [85,89]. It must
be remembered that although rfVII seems to be very promising for treat-
ment of PPH, it is an off-label use, complications have been reported, and
the actual incidence of complications in the setting of obstetric hemorrhage
is unknown. Documented complications include thrombosis, disseminated
intravascular coagulation, and myocardial infarction [90]. The pharmacy
costs of rfVII may be as high as several thousand dollars for a dose of 70
mg/kg.
Blood substitutes have been in development for more than a decade, and
some have been approved for use overseas and in veterinary medicine.
 POSTPARTUM HEMORRHAGE 437

Hemoglobin-based oxygen carriers have been prepared from various sour-

ces; at present the most promising are derivatives of either bovine hemoglo-
bin or outdated human packed red blood cells. Hemopure, produced by
Biopure Corporation of Cambridge, Massachusetts, polymerizes hemoglo-
bin obtained from a specially managed herd of cattle into long chains that
resist filtration in the kidney. It has been approved for use in South Africa
for use in general surgery. Its complementary veterinary product, Oxyglo-
bin, has been approved by the FDA and is in current use in the United
States for canine transfusion. Polyheme, developed by Northfield Laborato-
ries, Evanston, Illinois, another erythrocyte-free hemoglobin under trials, is
the product of cross-linked polymers of human hemoglobin. The potential
advantages of these products are a shelf life of about 1 year, lack of need
for cross matching, and decreased risk for the development of antibodies
to red blood cell surface membrane antigens. Potential risks include vascular
reactivity resulting in increased systemic and pulmonary artery pressure and
neurotoxicity. The newer generation of polymerized hemoglobins has not
caused the nephrotoxicity associated with earlier preparations.
An alternative approach to deliver oxygen to the tissues without red
blood cells involves perfluorocarbon emulsions. Oxygent is one such prod-
uct developed by Alliance Pharmaceutical subsidy San Diego, California.
The perfluorocarbon is mixed with lecithin and buffer salts and then is ho-
mogenized. When a high concentration of oxygen (70%–100%) is inspired,
the oxygen is dissolved in the infused perfluorocarbon emulsion. The oxygen
is not carried as with hemoglobin derivatives, but the dissolved oxygen is
able to diffuse into tissues. As with hemoglobin-based oxygen carriers,
this product does not require cross matching, does not carry risk of anti-
bodies developing against antigens in the red cell membrane, can be stored
at room temperature on the floor, and has a longer shelf life than banked
blood. There has been concern about cerebral vascular events with this
product. It is retained in the reticuloendothelial system and can result in re-
ticuloendothelial suppression. Lowering of the platelet count also has been
noted. Another potential disadvantage is its short half-life of about 12 to 24
hours [91].
These innovative products may have a role in the future, but at present
none are approved for clinical use in humans in the United States. It may
take several years before any of these products is perfected and gains
FDA approval.

Summary
The incidence of PPH can be reduced drastically by anticipation and
preventive measures. When PPH does occur, the resulting morbidity and
mortality can be prevented in most cases by early recognition and aggressive
and appropriate management.
438 OYELESE et al

References
[1] Khan KS, Wojdyla D, Say L, et al. WHO analysis of causes of maternal death: a systematic
review. Lancet 2006;367:1066–74.
[2] AbouZahr C. Global burden of maternal death and disability. Br Med Bull 2003;67:1–11.
[3] Pritchard JA, Baldwin RM, Dickey JC, et al. Blood volume changes in pregnancy and the
puerperium. II. Red blood cell loss and changes in apparent blood volume during and follow-
ing vaginal delivery, cesarean section, and cesarean section plus total hysterectomy. Am J
Obstet Gynecol 1962;84:1271–82.
[4] American College of Obstetrics and Gynecology practice bulletin: clinical management
guidelines for obstetrician-gynecologists number 76, October 2006: postpartum hemorrhage.
Obstet Gynecol 2006;108:1039–47.
[5] Combs CA, Murphy EL, Laros RK Jr. Factors associated with postpartum hemorrhage
with vaginal birth. Obstet Gynecol 1991;77:69–76.
[6] Kominiarek MA, Kilpatrick SJ. Postpartum hemorrhage: a recurring pregnancy complica-
tion. Semin Perinatol 2007;31:159–66.
[7] Preventing infant death and injury during delivery. Sentinel Event Joint Commission on
Accreditation of Healthcare Organizations ALERT No. 30.
[8] Dildy GA 3rd, Paine AR, George NC, et al. Estimating blood loss: can teaching significantly
improve visual estimation? Obstet Gynecol 2004;104:601–6.
[9] Baskett TF. Surgical management of severe obstetric hemorrhage: experience with an obstet-
ric hemorrhage equipment tray. J Obstet Gynaecol Can 2004;26:805–8.
[10] Amant F, Spitz B, Timmerman D, et al. Misoprostol compared with methylergometrine for
the prevention of postpartum haemorrhage: a double-blind randomised trial. Br J Obstet
Gynaecol 1999;106:1066–70.
[11] Gulmezoglu AM, Forna F, Villar J, et al. Prostaglandins for prevention of postpartum hae-
morrhage. Cochrane Database Syst Rev 2004;1:CD000494.
[12] Villar J, Gulmezoglu AM, Hofmeyr GJ, et al. Systematic review of randomized controlled
trials of misoprostol to prevent postpartum hemorrhage. Obstet Gynecol 2002;100:
1301–12.
[13] El-Refaey H, Nooh R, O’Brien P, et al. The misoprostol third stage of labour study: a rand-
omised controlled comparison between orally administered misoprostol and standard man-
agement. BJOG 2000;107:1104–10.
[14] Mousa HA, Alfirevic Z. Treatment for primary postpartum haemorrhage. Cochrane Data-
base Syst Rev 2003;1:CD003249.
[15] Surbek DV, Fehr PM, Hosli I, et al. Oral misoprostol for third stage of labor: a randomized
placebo-controlled trial. Obstet Gynecol 1999;94:255–8.
[16] Bamigboye AA, Hofmeyr GJ, Merrell DA. Rectal misoprostol in the prevention of postpar-
tum hemorrhage: a placebo-controlled trial. Am J Obstet Gynecol 1998;179:1043–6.
[17] O’Brien P, El-Refaey H, Gordon A, et al. Rectally administered misoprostol for the treat-
ment of postpartum hemorrhage unresponsive to oxytocin and ergometrine: a descriptive
study. Obstet Gynecol 1998;92:212–4.
[18] O’Leary JL, O’Leary JA. Uterine artery ligation in the control of intractable postpartum
hemorrhage. Am J Obstet Gynecol 1966;94:920–4.
[19] O’Leary JL, O’Leary JA. Uterine artery ligation for control of postcesarean section hemor-
rhage. Obstet Gynecol 1974;43:849–53.
[20] O’Leary JA. Uterine artery ligation in the control of postcesarean hemorrhage. J Reprod
Med 1995;40:189–93.
[21] O’Leary JA. Pregnancy following uterine artery ligation. Obstet Gynecol 1980;55:112–3.
[22] Hebisch G, Huch A. Vaginal uterine artery ligation avoids high blood loss and puerperal
hysterectomy in postpartum hemorrhage. Obstet Gynecol 2002;100:574–8.
[23] Philippe HJ, d’Oreye D, Lewin D. Vaginal ligature of uterine arteries during postpartum
hemorrhage. Int J Gynaecol Obstet 1997;56:267–70.
 POSTPARTUM HEMORRHAGE 439

[24] Baggish MS. Vaginal uterine artery ligation avoids high blood loss and puerperal hysterec-
tomy in postpartum hemorrhage. Obstet Gynecol 2003;101:416–7 [author reply: 417–8].
[25] AbdRabbo SA. Stepwise uterine devascularization: a novel technique for management of
uncontrolled postpartum hemorrhage with preservation of the uterus. Am J Obstet Gynecol
1994;171:694–700.
[26] Clark SL, Phelan JP, Yeh SY, et al. Hypogastric artery ligation for obstetric hemorrhage.
Obstet Gynecol 1985;66:353–6.
[27] Joshi VM, Otiv SR, Majumder R, et al. Internal iliac artery ligation for arresting postpartum
haemorrhage. BJOG 2007;114:356–61.
[28] Das BN, Biswas AK. Ligation of internal iliac arteries in pelvic haemorrhage. J Obstet
Gynaecol Res 1998;24:251–4.
[29] Evans S, McShane P. The efficacy of internal iliac artery ligation in obstetric hemorrhage.
Surg Gynecol Obstet 1985;160:250–3.
[30] Habek D, Becarevic R. Emergency peripartum hysterectomy in a tertiary obstetric center:
8-year evaluation. Fetal Diagn Ther 2007;22:139–42.
[31] Zelop CM, Harlow BL, Frigoletto FD Jr, et al. Emergency peripartum hysterectomy. Am J
Obstet Gynecol 1993;168:1443–8.
[32] Oyelese Y, Smulian JC. Placenta previa, placenta accreta, and vasa previa. Obstet Gynecol
2006;107:927–41.
[33] Kastner ES, Figueroa R, Garry D, et al. Emergency peripartum hysterectomy: experience at
a community teaching hospital. Obstet Gynecol 2002;99:971–5.
[34] Castaneda S, Karrison T, Cibils LA. Peripartum hysterectomy. J Perinat Med 2000;28:
472–81.
[35] Kwee A, Bots ML, Visser GH, et al. Emergency peripartum hysterectomy: a prospective
study in The Netherlands. Eur J Obstet Gynecol Reprod Biol 2006;124:187–92.
[36] Al-Sibai MH, Rahman J, Rahman MS, et al. Emergency hysterectomy in obstetricsda re-
view of 117 cases. Aust N Z J Obstet Gynaecol 1987;27:180–4.
[37] Smith J, Mousa HA. Peripartum hysterectomy for primary postpartum haemorrhage: inci-
dence and maternal morbidity. J Obstet Gynaecol 2007;27:44–7.
[38] Lau WC, Fung HY, Rogers MS. Ten years experience of caesarean and postpartum hyster-
ectomy in a teaching hospital in Hong Kong. Eur J Obstet Gynecol Reprod Biol 1997;74:
133–7.
[39] Kinzler WL, Scorza W, Schen-Schwarz S, et al. Second-trimester cervical pregnancy present-
ing as a failed labor induction. Obstet Gynecol 2000;96:839.
[40] CBL, Coker A, Lawal AH, et al. The B-Lynch surgical technique for the control of massive
postpartum haemorrhage: an alternative to hysterectomy? Five cases reported. Br J Obstet
Gynaecol 1997;104:372–5.
[41] Allahdin S, Aird C, Danielian P. B-Lynch sutures for major primary postpartum haemor-
rhage at caesarean section. J Obstet Gynaecol 2006;26:639–42.
[42] Danso D, Reginald P. Combined B-Lynch suture with intrauterine balloon catheter
triumphs over massive postpartum haemorrhage. BJOG 2002;109:963.
[43] Ferguson JE, Bourgeois FJ, Underwood PB. B-Lynch suture for postpartum hemorrhage.
Obstet Gynecol 2000;95:1020–2.
[44] Habek D, Kulas T, Bobic-Vukovic M, et al. Successful of the B-Lynch compression suture in
the management of massive postpartum hemorrhage: case reports and review. Arch Gynecol
Obstet 2006;273:307–9.
[45] Wohlmuth CT, Gumbs J, Quebral-Ivie J. B-Lynch suture: a case series. Int J Fertil Womens
Med 2005;50:164–73.
[46] Price N, B-Lynch C. Technical description of the B-Lynch brace suture for treatment of mas-
sive postpartum hemorrhage and review of published cases. Int J Fertil Womens Med 2005;
50:148–63.
[47] Price N, Lynch C. Uterine necrosis following B-Lynch suture for primary postpartum hae-
morrhage. BJOG 2006;113:1341 [author reply: 1342].
440 OYELESE et al

[48] Treloar EJ, Anderson RS, Andrews HS, et al. Uterine necrosis following B-Lynch suture for
primary postpartum haemorrhage. BJOG 2006;113:486–8.
[49] B-Lynch C. Partial ischemic necrosis of the uterus following a uterine brace compression su-
ture. BJOG 2005;112:126–7.
[50] Grotegut CA, Larsen FW, Jones MR, et al. Erosion of a B-Lynch suture through the uterine
wall: a case report. J Reprod Med 2004;49:849–52.
[51] Harma M, Gungen N, Ozturk A. B-Lynch uterine compression suture for postpartum hae-
morrhage due to placenta praevia accreta. Aust N Z J Obstet Gynaecol 2005;45:93–5.
[52] Habek D, Vranjes M, Bobic Vukovic M, et al. Successful term pregnancy after B-Lynch
compression suture in a previous pregnancy on account of massive primary postpartum
hemorrhage. Fetal Diagn Ther 2006;21:475–6.
[53] Api M, Api O, Yayla M. Fertility after B-Lynch suture and hypogastric artery ligation. Fertil
Steril 2005;84:1810–28.
[54] Ouahba J, Piketty M, Huel C, et al. Uterine compression sutures for postpartum bleeding
with uterine atony. BJOG 2007;114:619–22.
[55] Cho JH, Jun HS, Lee CN. Hemostatic suturing technique for uterine bleeding during cesar-
ean delivery. Obstet Gynecol 2000;96:129–31.
[56] Ghezzi F, Cromi A, Uccella S, et al. The Hayman technique: a simple method to treat post-
partum haemorrhage. BJOG 2007;114:362–5.
[57] Hayman RG, Arulkumaran S, Steer PJ. Uterine compression sutures: surgical management
of postpartum hemorrhage. Obstet Gynecol 2002;99:502–6.
[58] Chen CP. Use of the hemostatic multiple square suturing of the uterus for control of massive
postcesarean section hemorrhage and preservation of fertility. Acta Obstet Gynecol Scand
2001;80:976.
[59] Wu HH, Yeh GP. Uterine cavity synechiae after hemostatic square suturing technique.
Obstet Gynecol 2005;105:1176–8.
[60] Wittich AC, Salminen ER, Hardin EL, et al. Uterine packing in the combined management
of obstetrical hemorrhage. Mil Med 1996;161:180–2.
[61] Goldrath MH. Uterine tamponade for the control of acute uterine bleeding. Am J Obstet
Gynecol 1983;147:869–72.
[62] Condous GS, Arulkumaran S, Symonds I, et al. The ‘‘tamponade test’’ in the management of
massive postpartum hemorrhage. Obstet Gynecol 2003;101:767–72.
[63] Seror J, Allouche C, Elhaik S. Use of Sengstaken-Blakemore tube in massive postpartum
hemorrhage: a series of 17 cases. Acta Obstet Gynecol Scand 2005;84:660–4.
[64] Johanson R, Kumar M, Obhrai M, et al. Management of massive postpartum haemorrhage:
use of a hydrostatic balloon catheter to avoid laparotomy. BJOG 2001;108:420–2.
[65] Keriakos R, Mukhopadhyay A. The use of the Rusch balloon for management of severe
postpartum haemorrhage. J Obstet Gynaecol 2006;26:335–8.
[66] Bakri YN, Amri A, Abdul Jabbar F. Tamponade-balloon for obstetrical bleeding. Int J
Gynaecol Obstet 2001;74:139–42.
[67] Maier RC. Control of postpartum hemorrhage with uterine packing. Am J Obstet Gynecol
1993;169:317–21 [discussion: 321–3].
[68] Hsu S, Rodgers B, Lele A, et al. Use of packing in obstetric hemorrhage of uterine origin.
J Reprod Med 2003;48:69–71.
[69] Nwagha UI, Okaro JM, Nwagha TU. Intraoperative uterine packing with mops: an effec-
tive, but under utilized method of controlling post partum haemorrhage-experience from
South Eastern Nigeria. Niger J Med 2005;14:279–82.
[70] Druzin ML. Packing of lower uterine segment for control of postcesarean bleeding in
instances of placenta previa. Surg Gynecol Obstet 1989;169:543–5.
[71] Cassels JW Jr, Greenberg H, Otterson WN. Pelvic tamponade in puerperal hemorrhage.
A case report. J Reprod Med 1985;30:689–92.
[72] Robie GF, Morgan MA, Payne GG Jr, et al. Logothetopulos pack for the management of
uncontrollable postpartum hemorrhage. Am J Perinatol 1990;7:327–8.
 POSTPARTUM HEMORRHAGE 441

[73] Heaston DK, Mineau DE, Brown BJ, et al. Transcatheter arterial embolization for control
of persistent massive puerperal hemorrhage after bilateral surgical hypogastric artery liga-
tion. AJR Am J Roentgenol 1979;133:152–4.
[74] Soncini E, Pelicelli A, Larini P, et al. Uterine artery embolization in the treatment and
prevention of postpartum hemorrhage. Int J Gynaecol Obstet 2007;96:181–5.
[75] Merland JJ, Houdart E, Herbreteau D, et al. Place of emergency arterial embolisation in
obstetric haemorrhage about 16 personal cases. Eur J Obstet Gynecol Reprod Biol 1996;
65:141–3.
[76] Ojala K, Perala J, Kariniemi J, et al. Arterial embolization and prophylactic catheterization
for the treatment for severe obstetric hemorrhage. Acta Obstet Gynecol Scand 2005;84:
1075–80.
[77] Vegas G, Illescas T, Munoz M, et al. Selective pelvic arterial embolization in the manage-
ment of obstetric hemorrhage. Eur J Obstet Gynecol Reprod Biol 2006;127:68–72.
[78] Cheng YY, Hwang JI, Hung SW, et al. Angiographic embolization for emergent and pro-
phylactic management of obstetric hemorrhage: a four-year experience. J Chin Med Assoc
2003;66:727–34.
[79] Hong TM, Tseng HS, Lee RC, et al. Uterine artery embolization: an effective treatment for
intractable obstetric haemorrhage. Clin Radiol 2004;59:96–101.
[80] Ornan D, White R, Pollak J, et al. Pelvic embolization for intractable postpartum hemor-
rhage: long-term follow-up and implications for fertility. Obstet Gynecol 2003;102:904–10.
[81] Alvarez M, Lockwood CJ, Ghidini A, et al. Prophylactic and emergent arterial catheteriza-
tion for selective embolization in obstetric hemorrhage. Am J Perinatol 1992;9:441–4.
[82] Casele HL, Laifer SA. Successful pregnancy after bilateral hypogastric artery ligation.
A case report. J Reprod Med 1997;42:306–8.
[83] Comstock CH. Antenatal diagnosis of placenta accreta: a review. Ultrasound Obstet Gyne-
col 2005;26:89–96.
[84] Baskett TF. James Blundell: the first transfusion of human blood. Resuscitation 2002;52:
229–33.
[85] Prosper SC, Goudge CS, Lupo VR. Recombinant factor VIIa to successfully manage
disseminated intravascular coagulation from amniotic fluid embolism. Obstet Gynecol
2007;109:524–5.
[86] Franchini M, Lippi G, Franchi M. The use of recombinant activated factor VII in obstetric
and gynaecological haemorrhage. BJOG 2007;114:8–15.
[87] Karalapillai D, Popham P. Recombinant factor VIIa in massive postpartum haemorrhage.
Int J Obstet Anesth 2007;16:29–34.
[88] Pepas LP, Arif-Adib M, Kadir RA. Factor VIIa in puerperal hemorrhage with disseminated
intravascular coagulation. Obstet Gynecol 2006;108:757–61.
[89] Branch DW, Rodgers GM. Recombinant activated factor VII: a new weapon in the fight
against hemorrhage. Obstet Gynecol 2003;101:1155–6.
[90] Aledort LM. rFVIIadits thrombogenicity. Thromb Haemost 2000;84:522–3.
[91] Jahr JS, Nesargi SB, Lewis K, et al. Blood substitutes and oxygen therapeutics: an overview
and current status. Am J Ther 2002;9:437–43.
 Obstet Gynecol Clin N Am
34 (2007) 443–458

Blood Component Therapy in Obstetrics

Andrea J. Fuller, MD*, Brenda Bucklin, MD
Department of Anesthesiology, University of Colorado Denver Health Sciences Center,
4200 E. 9th Avenue, B-113, Denver, CO 80262, USA

Hemorrhage is the leading cause of intensive care unit admission and one
of the leading causes of death in the obstetric population [1]. This empha-
sizes the importance of a working knowledge of the indications for and com-
plications associated with blood product replacement in obstetric practice.
This article provides current information regarding preparation for and ad-
ministration of blood products, discusses alternatives to banked blood in the
obstetric population, and introduces pharmacological strategies for treat-
ment of hemorrhage.

Preparing for transfusion

Preparing for an obstetric hemorrhage requires the drawing of a blood
sample from the patient to obtain crossmatched blood. The first step in
the process of preparing blood is determining ABO type and the presence
or absence of Rh factor. To determine ABO type, the blood is mixed with
commercially available antibodies that react with A or B antigens on the pa-
tient’s erythrocytes, causing agglutination [2]. The Rh factor status is also
classified by this method. Then the blood type is confirmed by mixing the
patient’s blood with cells that contain A or B antigens. Because most people
have antibodies to antigens that they lack (ie, type-AB patients do not have
antibodies and type-O patients have anti-A and anti-B antibodies), aggluti-
nation will occur when antigen–antibody complexes are present.
Following typing, blood is screened for common antibodies. Screening
involves mixing the recipient’s blood with commercially available antigens.
If red blood cell agglutination or hemolysis occurs, antibodies are present
and must be characterized. This initial ‘‘type and screen’’ takes approximately
45 minutes and is best for patients at low risk for requiring blood transfusion

* Corresponding author.
E-mail address: andisamf@msn.com (A.J. Fuller).

0889-8545/07/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ogc.2007.06.003 obgyn.theclinics.com
444 FULLER & BUCKLIN

[2]. The most recent American Society of Anesthesiologists Practice

Guidelines for Obstetric Anesthesia [3] state that a routine blood crossmatch
is not necessary for healthy and uncomplicated parturients for vaginal or
operative delivery. The decision whether to order or require a blood type
and screen, or crossmatch, should be based on maternal history, anticipated
hemorrhagic complications (eg, placenta accreta in a patient with placenta
previa and previous uterine surgery), and local institutional policies.
Patients should undergo blood crossmatching when blood transfusion is
imminent or likely. To crossmatch blood, the recipient’s blood is mixed with
the donor’s to mimic the transfusion (serologic crossmatch) [2]. This process
detects antibodies in the Kell, Duffy, Kidd, and MN groups as well as an-
tibodies that are present in low titers and that do not agglutinate easily
[4]. Blood crossmatching typically takes an additional 15–45 minutes after
the blood has been typed and screened [2].
In an emergency where the patient requires transfusion before type-spe-
cific or crossmatched blood can be obtained, type-O blood can be adminis-
tered. In obstetric patients, it is especially important to administer type-O,
Rh-negative blood because of the risk of Rh sensitization. Crossmatched
blood should be administered as soon as it is available because the estimated
risk of a hemolytic transfusion reaction with this emergency blood has been
reported to be as high as 5%, although publications with trauma patients
report much lower complication rates [2,5].
The American Society of Anesthesiologists Task Force on Obstetric An-
esthesia and the American College of Obstetricians and Gynecologists [6]
recommend that all facilities providing obstetric care be prepared to manage
hemorrhagic emergencies [3]. Immediate availability of such equipment as
hand-inflated pressure bags, an automatic rapid infusion system, a fluid
warmer, and a forced-air warming device is recommended. Knowledge of
blood bank capability is paramount and resources vary depending on the
hospital. Therefore, it is essential to know the time required for obtaining
type-O, type-specific, and crossmatched blood as well as platelet and clot-
ting factor availability. Response to massive hemorrhage takes a coordinated
effort between clinicians and the blood bank; it is helpful to have a massive
hemorrhage protocol outlined before an emergency occurs [7]. Facilities
should also consider writing and posting such a protocol in addition to run-
ning clinical drills on obstetric hemorrhage scenarios [6,8].
For patients who are at risk for bleeding or who are actively hemorrhag-
ing, the importance of adequate intravenous access cannot be emphasized
enough. Flow through an intravenous cannula is directly proportional to
the fourth power of the radius and inversely proportional to the length.
For these reasons, one or more short, large-bore peripheral intravenous
catheters are often preferable to central venous access with a longer catheter
(such as a double- or triple-lumen catheter). An arterial line can be
extremely helpful during a hemorrhagic emergency, both for beat-to-beat
monitoring of blood pressure and for obtaining frequent laboratory tests.
 BLOOD COMPONENT THERAPY IN OBSTETRICS 445

Determining when to transfuse

Determining the point at which a patient requires blood transfusion can
be difficult. Many factors, including vital signs, ongoing blood loss, and co-
existing disease should be considered. Estimating blood loss during and af-
ter delivery can also be difficult and is often underestimated because the
blood is not always contained in one space and because amniotic fluid is
present. As a result, postpartum hemorrhage is not clearly defined. How-
ever, an estimated blood loss greater than 500 mL for a vaginal delivery
and 1000 mL for a cesarean delivery are typical definitions used to describe
postpartum hemorrhage [6].
The American College of Surgeons separates the severity of hemorrhagic
shock into classes based on vital signs and mental status [9]. Signs and symp-
toms of inadequate perfusion due to hypovolemia are presented in Table 1
and include tachycardia, decreased pulse pressure, tachypnea, decreased
urine output, and an altered mental status ranging from anxious to lethargic
[9,10]. While the physiologic changes of pregnancy (eg, increased blood

Table 1
Signs and symptoms in patients with obstetric hemorrhage
Severity ACS Blood loss % Blood
of shock class Signs and symptoms (mL) volume lost Notes
None Class I None Up to 750 10–15
Mild Class II Tachycardia 750–1500 15–25 Volume replacement
(!100 bpm); mild with crystalloid
hypotension; and/or colloid
normal or [ pulse
pressure
(peripheral
vasoconstriction)
Moderate Class III Tachycardia 1500–2000 25–40 Transfusion
(100–120 bpm); probable
hypotension
(systolic blood
pressure 80–100
mm Hg); Y pulse
pressure; anxiety,
confusion; oliguria
Severe Class IV Tachycardia O2000 O40 Transfusion
(O120–140 bpm; probable; massive
hypotension transfusion
(systolic blood possible
pressure
!80 mm Hg);
Y pulse pressure;
confusion,
lethargy; anuria
Abbreviations: ACS, American College of Surgeons; bpm, beats per minute.
Data from Refs. [7,9,10].
446 FULLER & BUCKLIN

volume) can limit the utility of this table, classes III and IV hemorrhage in-
dicate significant hypoperfusion and almost always require transfusion [9].
Historically, patients were transfused to keep the hemoglobin concentra-
tion greater than 10 mg/dL [11]. This practice has been challenged by a re-
cent study demonstrating decreased mortality in critically ill patients who
were transfused at lower hemoglobin thresholds (ie, transfusions adminis-
tered with hemoglobin concentrations less than 7 g/dL) [12]. On the other
end of the spectrum, Karpati and colleagues [13] found an approximately
50% incidence of myocardial ischemia in intensive care patients admitted
with postpartum hemorrhage and hypovolemic shock. Risk factors for
myocardial ischemia in this population were a hemoglobin of 6.0 g/dL or
lower, systolic blood pressure of 88 mm Hg or lower, diastolic blood pres-
sure of 50 mm Hg or lower, and a heart rate greater than 115 beats per
minute [13].
The purpose of packed red blood cell (PRBC) administration is to increase
the oxygen-carrying capacity of blood. According to the American Society
of Anesthesiologists Task Force on blood product replacement, PRBC trans-
fusion is rarely indicated with a hemoglobin level greater than 10 g/dL and is
almost always indicated with a hemoglobin level less than 6 g/dL [14]. Table 2
outlines the indications for PRBC and other blood products.
A recent survey of anesthesiologists and obstetrician/gynecologists found
that the transfusion threshold for most providers is 7 to 8 g/dL, with the
anesthesiologists transfusing at 7.5 g/dL and obstetricians at 8 g/dL [15].
While the clinical situation should dictate when to transfuse red blood cells,
a threshold in the range of 6.5 to 8.5 g/dL appears prudent given current data.

Disseminated intravascular coagulation

Disseminated intravascular coagulation (DIC) occurs when an inciting
event initiates the biodegradation of fibrinogen and clotting factors, result-
ing in hemorrhage and microvascular thrombosis. Obstetric disorders asso-
ciated with DIC include amniotic fluid embolism, placental abruption,
retained products of conception, eclampsia, and abortion [16]. Disseminated
intravascular coagulation is commonly associated with obstetric hemor-
rhage and causes profuse bleeding due to inadequate blood clot formation.
Therefore, obstetric care providers must consider the need for platelet and/
or clotting factor administration in a hemorrhaging patient, especially when
a condition associated with DIC is present.

Platelets
Platelets are usually available in six- to nine-unit equivalents from apheresis
or whole blood. One unit of platelets increases the platelet count by 5000 to
10,000 cells/mL in the absence of platelet destruction [7]. Platelet transfusion
 BLOOD COMPONENT THERAPY IN OBSTETRICS 447

Table 2
Blood product information
Indications for
Product Contents administration Notes
Packed red Red blood cells  Improve Type-specific and
blood cells oxygen-carrying crossmatched blood
capacity preferred
 Almost always for
hemoglobin
!6 g/dL
 Rarely for
hemoglobin
O10 g/dL
Platelets Platelets  Microvascular Blood product most
bleeding with often associated
platelet counts with bacterial
!50,000 cells/mL contamination
Fresh frozen All plasma proteins  Microvascular Must be thawed before
plasma and clotting factors bleeding due to administration
clotting factor (20–30 min)
deficiency
 International
normalized ratio
O2 normal
 Activated partial
thromboplastin time
O1.5 normal
Cryoprecipitate Factor VIII and  Microvascular Can also be used to
fibrinogen bleeding due to treat congenital
fibrinogen deficiency fibrinogen
 Fibrinogen deficiencies or von
!80–100 mg/dL Willebrand’s disease
when clotting
factors are
unavilable

is rarely indicated when the platelet count is greater than 100,000 cells/mL, but
should be considered when there is excessive bleeding with platelet counts
less than 50,000 cells/mL [14]. While it is possible to transfuse ABO-incom-
patible platelets, these cells may have a shorter life span [2]. Rh compati-
bility should be considered in the obstetric population and Rh immune
globulin should be administered if Rh-positive platelets are administered
to an Rh-negative individual [17].

Clotting factors
Fresh frozen plasma (FFP) is collected from whole blood or plasma
apheresis after platelets and cells are removed. It contains all plasma
proteins and clotting factors. FFP is stored at 18 C to 30 C and must
be thawed before administration. Thawing takes 20 to 30 minutes. In the
448 FULLER & BUCKLIN

obstetric setting, common indications for FFP are treatment of microvascu-

lar bleeding due to coagulopathy and/or factor deficiency following massive
transfusion. Additional indications include reversal of warfarin, correction
of isolated factor deficiencies when specific factor concentrates are unavail-
able, and antithrombin III deficiency in patients receiving heparin [14]. Rec-
ommendations for FFP administration include measurement of the
activated partial thromboplastin time (aPTT) and prothrombin time before
administration and when the prothrombin time and international normal-
ized ratio are greater than two times normal and/or the aPTT is greater
than 1.5 times normal [14]. Because anti-ABO antibodies are present in
plasma, ABO compatibility should be considered when transfusing FFP
[2]. For example, a patient with type-AB blood should not receive type-O
plasma because of the presence of anti-A and anti-B antibodies [2].
Cryoprecipitate is extracted from slowly thawing FFP. It is rich in factor
VIII and fibrinogen and is used to treat microvascular bleeding in the pres-
ence of fibrinogen deficiency, which most commonly occurs because of DIC
or massive transfusion. Ideally, a fibrinogen level should be obtained before
administration of cryoprecipitate. Fibrinogen concentrations greater than
150 mg/dL usually do not require cryoprecipitate, but fibrinogen concentra-
tions less than 80 to 100 mg/dL indicate need for transfusion [14]. Cryopre-
cipitate can also be administered for treatment of congenital fibrinogen
deficiencies or bleeding in patients with von Willebrand’s disease when fac-
tor concentrates are unavailable [14]. Because cryoprecipitate has only
a small amount of plasma, ABO compatibility is unnecessary [2].

Autologous blood donation

Because of concern about cost-effectiveness, routine autologous blood
donation is not recommended for routine obstetric deliveries [11]. However,
autologous blood transfusion is a viable option for patients at risk for peri-
partum hemorrhage, especially those with rare antibodies who will be diffi-
cult to transfuse with compatible homologous blood. Autologous blood
donation during pregnancy has been shown to have minimal maternal he-
modynamic effects with maintenance of fetal umbilical artery systolic/dia-
stolic ratio [18].
Yamada and colleagues [19] published an analysis of 82 patients with pla-
centa previa after implementation of an autologous blood donation proto-
col. They found that women who did not donate blood prepartum had
a four times greater rate (12% versus 3.1%) of peripartum homologous
blood transfusion. They recommended beginning the blood donation at
32 weeks’ gestation with removal of 400 mL per week to achieve a total
stored volume of 1200 to 1500 mL [19]. In the study, patients who donated
autologous blood had a higher overall rate of blood transfusion, with 71%
receiving blood peripartum compared with 12% of patients who received
homologous blood. While autologous blood has a slightly smaller incidence
 BLOOD COMPONENT THERAPY IN OBSTETRICS 449

of bacterial contamination, the risk of ABO mismatching is similar for both

autologous and homologous blood [20,21]. Thus, administration of autolo-
gous blood should not be viewed as innocuous and should be administered
for the same indications as banked blood.

Acute normovolemic hemodilution

Acute normovolemic hemodilution is a technique involving collection of
autologous blood immediately before surgery or delivery. Normovolemia is
maintained by intravenous fluid administration with colloid or crystalloid.
The volume of colloid administered should be equal to the volume of blood
withdrawn. When crystalloid is administered, the volume should be three
times the volume of blood removed [22,23]. When blood is subsequently
lost, it has less red blood cell mass and the blood removed can be returned
to the patient as needed.
Because the blood is collected and stored at the bedside for immediate re-
infusion, the risks of bacterial contamination and administrative error asso-
ciated with autologous blood storage are significantly reduced. This
technique has been successfully reported in patients at risk for blood loss
during cesarean delivery, with an average of 1000 mL of blood collected
just before the surgery [22]. In this study, no patients experienced symptoms
of nausea, vomiting, dizziness, or lightheadedness and there were no abnor-
malities in vital signs or fetal heart rate [22].

Intraoperative cell salvage

Another alternative to allogenic banked blood is the use of an intraoper-
ative cell salvage device, or cell saver. This technique involves suctioning of
blood from the operative field followed by cell washing, suspension in saline,
and reinfusion to the patient [24]. Concerns about its possible association
with amniotic fluid embolism (AFE) have made this technique controversial
[25–27].
The cause of the coagulopathy and cardiovascular collapse associated
with AFE is unclear [28]. Tissue factor is present in amniotic fluid and plays
a role in the initiation of coagulation, prompting speculation that tissue fac-
tor is responsible for the DIC associated with AFE. The effectiveness of
a commonly available cell saver system to remove functionally active tissue
factor from blood contaminated with amniotic fluid has been demonstrated
[29]. Fetal squamous cells, meconium, and other particulates have also been
implicated in the development of AFE [30]. Waters and colleagues [26] dem-
onstrated that when cells are washed and a leukocyte depletion filter is used,
the resulting blood has a concentration of fetal squamous cells similar to
a preoperative maternal blood sample [26].
In a multicenter historical cohort study, 139 patients received autologous
blood transfusion during cesarean delivery via intraoperative cell salvage
450 FULLER & BUCKLIN

technique with no patients experiencing AFE or adult respiratory distress

syndrome [31]. While the investigators concluded that their study had
enough power to detect a clinically significant increase in AFE, it is still pos-
sible that this rare event can be associated with this technology. In fact,
one case report exists of a patient who developed hypoxia, cardiovascular
collapse, and death minutes after infusion of cell saver blood follow-
ing cesarean delivery. The patient had coexisting diseases, including
hemolysis–elevated-liver-enzymes–low-platelets (HELLP) syndrome, so the
exact cause of death was unclear. However, a clinical diagnosis of AFE was
made [32].
Fetal hemoglobin is present in the processed cell saver blood, raising con-
cerns about maternal alloimmunization and the potential for problems with
subsequent pregnancies [26,27]. Rh mismatch is particularly important and
anti-D immune globulin should be administered to Rh-negative mothers
who receive salvaged blood [24]. Because the exact volume of fetal blood ad-
ministered to the mother via the cell saver is highly variable, a Kleihauer-
Betke test should be considered to allow for dose adjustment of Rh immune
globulin [33].
Critics caution that because the inciting factors in AFE are unknown and
the incidence is so low, the safety of salvaged blood cannot be proven
[25,27,32]. Furthermore, because obstetric hemorrhage can be unpredict-
able, availability of equipment and skilled personnel is a significant draw-
back to intraoperative cell salvage [25,27]. However, this technique has
been used safely in many patients and should be considered in patients at
high risk for hemorrhage who would be difficult to crossmatch or object
to blood transfusion (eg, a Jehovah’s Witness with a known placenta ac-
creta) [6,24,34].

Massive transfusion
Massive transfusion is defined as administration of greater than 10 units
of packed red blood cells [35]. Because large amounts of blood products will
be needed, it is important to notify the blood bank when massive hemor-
rhage occurs in an obstetric patient. A massive hemorrhage protocol can
be extremely helpful, especially one that outlines how blood products will
be transported to the obstetric suite and how clotting factors will be pre-
pared in a timely way [36]. Clear communication between personnel, espe-
cially the obstetrician, anesthesiologist, and nursing staff regarding
ongoing blood loss and the continued need for blood products is important.
The massively bleeding patient must be reassessed frequently to deter-
mine the efficacy of treatment as well as to identify correctable complica-
tions. Massive transfusion is associated with the ‘‘bloody vicious cycle,’’
which was originally used to describe coagulopathy following trauma [35].
Active hemorrhage is worsened by coagulopathy, which is caused by meta-
bolic acidosis and core hypothermia. The treatment of the hemorrhage with
 BLOOD COMPONENT THERAPY IN OBSTETRICS 451

red cell transfusion can worsen the coagulopathy by diluting platelets and
clotting factors as well as contributing to hypothermia and acidosis [35].
In a prospective analysis of trauma patients receiving greater than 10 units
of PRBC, approximately 50% developed coagulopathy [35]. Patients who
also had a core temperature of less than 34 C and persistent metabolic ac-
idosis had an even higher incidence of life-threatening coagulopathy [35]. In
obstetrics, the exact incidence of coagulopathy with massive transfusion is
unknown, but may be even higher given the high incidence of DIC in the
obstetric population. For these reasons, platelets and coagulation factors
must be administered to the massively bleeding patient. Core temperature
must be measured and every effort made to warm both the patient and
blood products being administered. Other complications associated with
massive transfusion are discussed later in this article and include hypocalce-
mia and hyperkalemia.

Errors and transfusion

While patients are often highly concerned about the infectious risks asso-
ciated with blood transfusion, patients are actually at more risk for compli-
cations resulting from ABO incompatibility errors and similar mixups
unrelated to infections [20,21,37]. A survey of transfusion errors in New
York state over a 10-year period found the incidence of erroneous adminis-
tration to be one for every 19,000 red blood cell units administered with
blood being administered to the wrong recipient representing 38% of the er-
rors [20]. The incidence of ABO incompatibility errors ranges from one for
every 38,000 units administered to one in every 138,000 units administered
[20,21,37]. Overall, the incidence of a fatal reaction due to erroneous ad-
ministration is approximately one in every 1,500,000 units of blood
administered [37]. In many cases, multiple errors are involved and can
include phlebotomy errors, patient misidentification, sample mislabeling,
and laboratory errors [20,37]. Clearly, vigilance is required of all personnel
involved in blood product administration and is paramount to keeping these
risks at an absolute minimum.

Hemolytic reactions
The most serious complication arising from erroneous blood product
administration is an acute hemolytic reaction. This occurs as a result of
the recipient’s circulating antibodies destroying the donor’s red blood cells.
An acute hemolytic reaction is characterized by fever, urticaria, nausea,
chest and flank pain, hyperkalemia, hypotension, DIC, hemoglobinemia,
and acute renal failure [4,7]. If an acute hemolytic reaction is suspected,
the transfusion should be stopped immediately with initiation of supportive
care, including blood pressure support, aggressive intravenous fluid
452 FULLER & BUCKLIN

replacement, diuresis, and alkalinization of the urine [4,7]. Laboratory stud-

ies, including urine and plasma hemoglobin, an antibody screen, coagula-
tion parameters, and blood counts should be obtained [4,7]. The blood
being infused should be sent to the blood bank with a sample of the patient’s
blood to confirm incompatibility.
A delayed hemolytic reaction is due to extravascular hemolysis of donor
erythrocytes. It results from the presence of antibodies from previous trans-
fusions or pregnancy in recipient serum that were at levels too low to be
detected during the crossmatch [4]. Clinical manifestations occur approxi-
mately 1 week after a seemingly compatible transfusion and are character-
ized by anemia, mild fever, increased unconjugated bilirubin, jaundice,
hemoglobinuria, decreased haptoglobin, and spherocytosis on the blood
smear [4,7]. Because the hemolysis is extravascular, the reaction is much
less severe than an immediate hemolytic reaction and the symptoms are
self-limited [4].

Transfusion-transmitted infectious disease

The incidence of transfusion-transmitted infectious diseases has de-
creased dramatically over the last 20 years, mainly because of improved do-
nor screening and technological advances in blood bank testing. Of
particular importance has been the development of nucleic acid testing for
viral pathogens, such as HIV, hepatitis C, and hepatitis B. Historically,
transfusion-transmitted viral infections posed a large risk to recipients.
The incidence of such infections is now so low that mathematical models
must be used to estimate the incidence of pathogen transmission. The cur-
rent incidence of various infectious diseases associated with blood transfu-
sion is summarized in Table 3. The estimated risk of HIV is one infection
for every 2,135,000 units of blood administered and that of hepatitis C is

Table 3
Estimated incidence of transfusion-associated disease
Incidence (incidence of disease/units
Transfusion-associated disease of blood administered)
HIV 1:2,135,000 [38]
Hepatitis C virus 1:1,935,000 [38]
Hepatitis B virus 1:200,000a [38]
West Nile virus Incidence varies seasonally and
geographically; approximately 1:1,000,000 [39]
Chagas’ disease Rare
Malaria Rare
Variant Creutzfeldt-Jakob disease Rare
Bacterial contamination 1:12,000 for platelets; 1:500,000 for red blood
cells [11]
a
Estimate made before introduction of nucleic acid testing.
 BLOOD COMPONENT THERAPY IN OBSTETRICS 453

one for every 1,935,000 units administered [38]. The genetic diversity of HIV
is increasing and constant surveillance of the blood supply is required to op-
timize detection of this virus and keep transfusion-associated transmission
at its current rate [39].
Other potentially infectious agents are continually surfacing. Transfu-
sion-associated transmission of West Nile virus was first reported in 2002
and prompted the development of nucleic acid testing, especially in locales
with high West Nile virus activity [39]. Variant Creutzfeldt-Jakob disease
is an emerging concern, with one probable case of transfusion-associated
transmission prompting exclusion of blood donors who have spent more
than 6 months in the United Kingdom from 1980 to 1996 [39]. In parts of
the world where variant Creutzfeldt-Jakob disease transmission is a signifi-
cant concern, plasma treated with a solvent-detergent can be imported from
the United States to minimize the risk [40].
Trypanosoma cruzi, the pathogen responsible for trypanosomiasis
(ie, Chagas’ disease), can also be transmitted via blood transfusion. This dis-
ease is a growing concern in the United States because the parasite can sur-
vive the cold storage and cryopreservation of blood products. While the
incidence of transmission remains low, screening tests are being improved
with potential universal screening of blood donations in the future [39].
Transfusion-associated transmission of malaria, another parasitic illness, re-
mains a potential threat, with approximately three cases per year in the
United States [39]. Currently, because laboratory screening tests lack accu-
racy, the risk is reduced by excluding donors who have recently traveled to
endemic areas [39].
Bacterial contamination of blood products is the most common cause of
acute transfusion-associated mortality from an infectious agent [39]. Bacte-
rial contamination occurs most often with platelets, with an estimated inci-
dence of one for every 12,000 units of blood administered [11]. This is due to
the fact that platelets must be stored at room temperature and therefore
have a higher potential for supporting bacterial growth than do other blood
products. The most frequent contaminating organism is Yersinia entercoli-
tica for red blood cells and Staphylococcus aureus for platelets [11]. The clin-
ical presentation ranges from mild fever to acute sepsis leading to death.
Bacterial contamination should be suspected and antibiotic therapy consid-
ered in patients who develop a fever within 6 hours after platelet transfusion
[11].

Transfusion-associated acute lung injury

Transfusion-associated acute lung injury (TRALI) is an acute respiratory
distress syndrome occurring within 2 to 6 hours after transfusion [41–43].
It is characterized by noncardiogenic pulmonary edema manifesting as hyp-
oxia with bilateral infiltrates on chest radiograph [41–44]. The true incidence
454 FULLER & BUCKLIN

of TRALI is unknown because it is difficult to distinguish from other forms

of acute lung injury and it often occurs in patients with multiple coexisting
illnesses [43]. However, it is not a rare entity, and is estimated to occur once
in every 2000 to 5000 transfusions of blood or blood products [43]. Accord-
ing to the Food and Drug Administration, TRALI is the leading cause of
death from transfusions in the United States [43].
The leading hypothesis for the pathogenesis of TRALI is antibody-medi-
ated [45]. A donor HLA or granulocyte-specific antibody is transfused into
a recipient who possesses the corresponding leukocyte antigens [45]. This
antibody–antigen interaction then initiates a cascade of cellular activity in
the lung, resulting in endothelial damage and capillary leakage into alveoli
[45]. Women who have had multiple pregnancies and patients with a history
of prior transfusions are the most likely donors to be implicated in cases of
TRALI [46]. Of particular interest in the obstetric setting is the association
of TRALI in children whose mothers act as directed blood donors. In such
cases, TRALI presumably stems from the development of antibodies toward
paternally derived antigens present in the offspring [47]. An alternate hy-
pothesis for the pathogenesis of TRALI involves two events, the first being
a preexisting clinical condition in the patient that causes activation of the
pulmonary endothelium [45]. The second event is the transfusion of biolog-
ically active substances that cause neutrophil activation and lead to pulmo-
nary endothelial damage and alveolar edema [45].
Treatment for TRALI is supportive, with mechanical ventilation required
for most patients. Small tidal volumes are recommended. Hypotension is
generally responsive to intravenous fluid but diuretic administration can
worsen the patient’s condition [42].

Miscellaneous complications
Other complications associated with blood transfusion are associated
with the citrate phosphate dextrose (CPD) used as an anticoagulant preser-
vative in PRBC. In massive transfusion, citrate can bind plasma calcium and
lead to hypocalcemia, causing hypotension, tetany, and cardiac arrhythmias
[4]. Plasma calcium levels should be measured during massive transfusion
and hypocalcemia treated with intravenous calcium chloride [4].
Another potential complication associated with the CPD preservative is
acidosis. The pH of stored blood is approximately 7.0 because of the preser-
vative and can decrease to 6.9 during storage because of the metabolism of
glucose to lactate [4]. It is unclear whether the acidity of banked blood con-
tributes to acidosis in the patient. When massive transfusion is required,
therapy should be guided by frequent blood gas analysis [4,35].
Hyperkalemia can occur with PRBC administration because of passive
diffusion of potassium out of the red blood cells during storage. In patients
with normal renal function, the excess potassium is usually transported back
into the cells or excreted in the urine. However, potassium levels should also
 BLOOD COMPONENT THERAPY IN OBSTETRICS 455

be measured in patients requiring transfusion. If EKG changes, such as

peaked T waves and wide PR and QRS intervals, are observed, the patient
must be treated for hyperkalemia [4].
Because blood is stored at 1 C to 6 C, hypothermia can result from blood
transfusions, especially during massive transfusion. Extreme hypothermia
can result in impaired coagulation, decreased tissue perfusion, arrhythmias,
and decreased drug activity [7,48–50]. Because of this, temperature monitor-
ing, active warming, and use of a blood warmer are imperative when mas-
sive transfusion is required [3,14].

Activated recombinant factor VII

A promising new alternative to blood component therapy is recombinant
activated factor VII (rFVIIa). This drug is identical in structure and func-
tion to human factor VIIa and was originally developed to prevent or con-
trol bleeding in patients with hemophilia A or B with inhibitors to factors
VIII or IX. However, the drug has been used in other situations with uncon-
trolled bleeding, including life-threatening obstetric hemorrhage [49,51,52].
Several case reports and reviews have described decreased blood product re-
quirements in surgical and trauma patients with uncontrolled bleeding with
the administration of rFVIIa [53,54]. The mechanism of action of rFVIIa is
to augment the intrinsic clotting pathway by binding with tissue factor and
directly activating factors IX and X [49,54]. The use of rFVIIa for post-
partum hemorrhage is off-label and therefore the dose is based on case
reports. The most commonly reported effective dose is 50 to 100 mg/kg
intravenously every 2 hours until hemostasis is achieved, with the vast
majority of patients requiring only one dose [6,49].
It is important to ensure adequate levels of platelets and clotting factors
(by administration of blood products if necessary) because rFVIIa increases
clotting by acting on these substrates [49,52,55]. While the optimal timing of
rFVIIa administration is not known, reports suggest improved outcome
when rFVIIa is administered relatively early in a hemorrhagic emergency
[49,52,55]. Furthermore, the activity of rFVIIa is reduced during hypother-
mia and acidosis, highlighting the importance of its use before the patient
develops some of the consequences of massive transfusion [49].
Because rFVIIa is derived from recombinant technology and not from
human proteins, there is no risk of viral transmission from the drug [49].
The most commonly reported adverse events associated with factor VIIa ad-
ministration are thrombosis, including cerebrovascular accidents, myocar-
dial infarction, pulmonary embolism, and clotting of indwelling devices
[56]. Most occur within 3 hours of administration of the last drug dose
[56]. More information regarding the off-label use of this new product for
obstetric hemorrhage is needed and will surely become available as its use
increases. For now, rFVIIa should be considered, if available, in a hemor-
rhagic emergency.
456 FULLER & BUCKLIN

Summary
Hemorrhagic emergencies are common in obstetrics. Blood component
therapy should be administered to treat specific conditions, such as inade-
quate oxygen delivery, microvascular bleeding, and coagulation factor defi-
ciency. Alternatives to banked blood include autologous blood donation,
normovolemic hemodilution, and intraoperative cell salvage. These should
be considered in patients who are difficult to crossmatch and/or who refuse
banked blood. Recombinant factor VIIa is a new adjunct for treatment of
massive hemorrhage and should be considered, if available.

References
[1] Heinonen S, Tyrvainen E, Saarikoski S, et al. Need for maternal critical care in obstetrics:
a population based analysis. Int J Obstet Anesth 2002;11(4):260–4.
[2] Yazer MH. The blood bank ‘‘black box’’ debunked: pretransfusion testing explained. CMAJ
2006;174(1):29–32.
[3] Practice guidelines for obstetric anesthesia: an updated report by the American Society of
Anesthesiologists task force on Obstetric Anesthesia. Anesthesiology 2007;106(4):1–21.
[4] Petrovich CT. Hemostasis and hemotherapy. In: Barash PG, editor. Clinical anesthesia.
3rd edition. Philadelphia: Lippincott-Raven Publishers; 1996. p. 189–217.
[5] Dutton RP, Shih D, Edelman BB, et al. Safety of uncrossmatched type-O red cells for resus-
citation from hemorrhagic shock. J Trauma 2005;59(6):1445–9.
[6] ACOG practice bulletin: clinical management guidelines for obstetrician-gynecologists
Number 76, October 2006: postpartum hemorrhage. Obstet Gynecol 2006;108(4):1039–47.
[7] Santoso JT, Saunders BA, Grosshart K. Massive blood loss and transfusion in obstetrics and
gynecology. Obstet Gynecol Surv 2005;60(12):827–37.
[8] Preventing infant death and injury during delivery. Sentinel event alert No. 30. Joint Com-
mission on Accreditation of Healthcare Organizations. Available at: http://www.
jointcommission.org/SentinelEvents/SentinelEventAlert/sea_30.htm. Accessed February
13, 2007.
[9] Shock. American College of Surgeons committee on trauma: advanced trauma life support
for doctors, vol. 7. Chicago: American College of Surgeons; 2004. p. 69–102.
[10] Mayer DC, Spielman FJ, Bell EA. Antepartum and postpartum hemorrhage. In:
Chestnut DH, editor. Obstetric anesthesia: principles and practice. 3rd edition. Phila-
delphia: Elsevier Mosby; 2004. p. 662–82.
[11] Goodnough LT, Brecher ME, Kanter MH, et al. Transfusion medicine: first of two parts.
N Engl J Med 1999;340(6):438–47.
[12] Hebert PC, Wells G, Blajchman MA, et al. A multicenter, randomized, controlled clinical
trial of transfusion requirements in critical care. N Engl J Med 1999;340(6):409–17.
[13] Karpati PCJ, Rossignol M, Pirot M, et al. High incidence of myocardial ischemia during
postpartum hemorrhage. Anesthesiology 2004;100(1):30–6.
[14] Practice guidelines for perioperative blood transfusion and adjuvant therapies. An updated
report by the American Society of Anesthesiologists task force on perioperative blood trans-
fusion and adjuvant therapies. Anesthesiology 2006;105:198–208.
[15] Matot I, Einav S, Goodman S, et al. A survey of physicians’ attitudes toward blood trans-
fusion in patients undergoing cesarean section. Am J Obstet Gynecol 2004;190(2):462–7.
[16] Bick RL. Disseminated intravascular coagulation: a review of etiology, pathophysiology,
diagnosis, and management: guidelines for care. Clin Appl Thromb Hemost 2002;8(1):1–31.
[17] Menitove JE. Immunoprophylaxis for D- patients receiving platelet transfusions from
Dþ donors? Transfusion 2002;42(12):136–8.
 BLOOD COMPONENT THERAPY IN OBSTETRICS 457

[18] Droste S, Sorensen T, Price T, et al. Maternal and fetal hemodynamic effects of autologous
blood donation during pregnancy. Am J Obstet Gynecol 1992;167(1):89–93.
[19] Yamada T, Mori H, Ueki M. Autologous blood transfusion in patients with placenta previa.
Acta Obstet Gynecol Scand 2005;84(3):255–9.
[20] Linden JV, Wagner K, Voytovich AE, et al. Transfusion errors in New York State: an
analysis of 10 years’ experience. Transfusion 2000;40(1):1207–13.
[21] Andreu G, Morel P, Forestier F, et al. Hemovigilance network in France: organization and
analysis of immediate transfusion incident reports from 1994 to 1998. Transfusion 2002;
42(10):1356–64.
[22] Grange CS, Douglas J, Adams TJ, et al. The use of acute hemodilution in parturients under-
going cesarean section. Am J Obstet Gynecol 1998;178(1):156–60.
[23] Monk TG. Acute normovolemic hemodilution. Anesthesiol Clin North America 2005;23(2):
271–81.
[24] Waters JH. Indications and contraindications of cell salvage. Transfusion 2004;44(Suppl 12):
40S–4S.
[25] Clark V. Facilities for blood salvage (cell saver technique) must be available in every obstetric
theatre. Int J Obstet Anesth 2005;14(1):50–2.
[26] Waters JH, Biscotti C, Potter PS, et al. Amniotic fluid removal during cell salvage in the
cesarean section patient. Anesthesiology 2000;92(6):1531–6.
[27] Weiskopf RB. Erythrocyte salvage during cesarean section. Anesthesiology 2000;92(6):
1519–22.
[28] Clark SL, Hankins GDV, Dudley DA, et al. Amniotic fluid embolism: analysis of the na-
tional registry. Am J Obstet Gynecol 1995;172(4):1158–69.
[29] Bernstein HH, Rosenblatt MA, Gettes M, et al. The ability of the Haemonetics 4 Cell Saver
System to remove tissue factor from blood contaminated with amniotic fluid. Anesth Analg
1997;85(4):831–3.
[30] Petroianu GA, Altmannsberger SH, Maleck WH, et al. Meconium and amniotic fluid
embolism: effects on coagulation in pregnant mini-pigs. Crit Care Med 1999;27(2):348–55.
[31] Rebarber A, Lonser R, Jackson S, et al. The safety of intraoperative autologous blood col-
lection and autotransfusion during cesarean section. Am J Obstet Gynecol 1998;179(3):
715–20.
[32] Oei SG, Wingen CBM, Kerkkamp HEM. Cell salvage: how safe in obstetrics? Int J Obstet
Anesth 2000;9:143.
[33] Harkness UF, Spinnato JA. Prevention and management of Rh D isoimmunization. Clin
Perinatol 2004;31(4):721–42.
[34] Thomas D. Facilities for blood salvage (cell saver technique) must be available in every
obstetric theatre. Int J Obstet Anesth 2005;14(1):48–50.
[35] Cosgriff N, Moore EE, Sauaia A, et al. Predicting life-threatening coagulopathy in the mas-
sively transfused trauma patient: hypothermia and acidoses revisited. J Trauma 1997;42(5):
857–62.
[36] Malone DL, Hess JR, Fingerhut A. Massive transfusion practices around the globe and
a suggestion for a common massive transfusion protocol. J Trauma 2006;60(Suppl 6):
S91–6.
[37] Stainsby D. Errors in transfusion medicine. Anesthesiol Clin North America 2005;23(2):
253–61.
[38] Dodd RY, Notari EP IV, Stramer SL. Current prevalence and incidence of infectious disease
markers and estimated window-period risk in the American Red Cross blood donor popu-
lation. Transfusion 2002;42:975–9.
[39] Fiebig EW, Busch MP. Emerging infections in transfusion medicine. Clin Lab Med 2004;
24(3):797–823.
[40] Chekrizova V, Murphy WG. Solvent-detergent plasma: use in neonatal patients, in adult and
paediatric patients with liver disease, and in obstetric and gynaecological emergencies.
Transfus Med 2006;16(2):85–91.
458 FULLER & BUCKLIN

[41] Goodnough LT. Risks of blood transfusion. Anesthesiol Clin North America 2005;23(2):
241–52.
[42] Moore SB. Transfusion-related acute lung injury (TRALI): clinical presentation, treatment,
and prognosis. Crit Care Med 2006;34(Suppl):S114–7.
[43] Toy P, Popovsky MA, Abraham E, et al. Transfusion-related acute lung injury: definition
and review. Crit Care Med 2005;33(4):721–6.
[44] Nathens AB. Massive transfusion as a risk factor for acute lung injury: association or
causation? Crit Care Med 2006;34(Suppl 5):S144–50.
[45] Mair DC, Hirschler N, Eastlund T. Blood donor and component management strategies to
prevent transfusion-related acute lung injury (TRALI). Crit Care Med 2006;34(Suppl 5):
S137–43.
[46] Palfi M, Berg S, Ernerudh J, et al. A randomized controlled trial of transfusion-related acute
lung injury: is plasma from multiparous blood donors dangerous? Transfusion 2001;41(3):
317–22.
[47] Yang X, Ahmed S, Chandrasekaran V. Transfusion-related acute lung injury resulting from
designated blood transfusion between mother and child: a report of two cases. Am J Clin
Pathol 2004;121(4):590–2.
[48] DeLoughery TG. Coagulation defects in trauma patients: etiology, recognition, and ther-
apy. Crit Care Clin 2004;20(1):13–24.
[49] Karalapillai D, Popham P. Recombinant factor VIIa in massive postpartum haemorrhage.
Int J Obstet Anesth 2007;16:29–34.
[50] Eddy VA, Morris JA, Cullinane DC. Hypothermia, coagulopathy, and acidosis. Surg Clin
North Am 2000;80(3):845–54.
[51] Bouwmeester FW, Jonkhoff AR, Verheijen RHM, et al. Successful treatment of life threat-
ening postpartum hemorrhage with recombinant activated factor VII. Obstet Gynecol 2003;
101(6):1174–6.
[52] Biss TT, Hanley JP. Recombinant activated factor VII (rFVIIa/NovoSeven) in intractable
haemorrhage: use of a clinical scoring system to predict outcome. Vox Sang 2006;90(1):
45–52.
[53] Martinowitz U, Kenet G, Segal E, et al. Recombinant activated factor VII for adjunctive
hemorrhage control in trauma. J Trauma 2001;51(3):431–9.
[54] Hedner U, Erhardtsen E. Potential role for rFVIIa in transfusion medicine. Transfusion
2002;42(1):114–24.
[55] Clark AD, Gordon WC, Walker ID, et al. ‘‘Last-ditch’’ use of recombinant factor VIIa in
patients with massive haemorrhage is ineffective. Vox Sang 2004;86(2):120–4.
[56] OConnell KA, Wood JJ, Wise RP, et al. Thromboembolic adverse events after use of
recombinant human coagulation factor VIIa. JAMA 2006;295(3):293–8.
 Obstet Gynecol Clin N Am
34 (2007) 459–479

Early Goal Directed Therapy

for Sepsis During Pregnancy
Debra A. Guinn*, David E. Abel, Mark W. Tomlinson
Northwest Perinatal Center, 9701 SW Barnes Road, Suite 299,
Portland, OR 97225, USA

Sepsis is the leading cause of death in critically ill patients in the United
States and is among the 10 leading causes of death overall [1–6]. The costs as-
sociated with sepsis are staggering, approaching $17 billion dollars annually
as sepsis accounts for 2% to 11% of all hospital admissions [2,3,7]. The an-
nual rate of sepsis is estimated at 240 to 300 cases per 100,000 population,
and this rate has increased over the past decade [3,5–7]. This increase is attrib-
uted in part to an aging population, greater antimicrobial resistance, and the
increased use of invasive procedures, immunosuppressive drugs, chemother-
apy, and transplantation. Annually, over 750,000 cases are thought to occur,
and estimates for the year 2010 are projected at 934,000 cases per year [3,8].
Historically, imprecise definitions of the terms bacteremia, septicemia,
sepsis, and septic shock hindered the ability to establish an early diagnosis
in the evolving process of sepsis [9]. These terms were often used inter-
changeably in both the general and obstetric literature. This imprecision
makes study comparisons difficult. Furthermore, the lack of clear definitions
has hampered the ability to understand the pathophysiology of sepsis and
the development of successful therapy [3]. In 1992, the American College
of Chest Physicians/Society of Critical Care Medicine (ACCP/SCCM) pub-
lished a consensus report based on a panel convened to standardize the
definitions for the classification of sepsis [10]. Despite the specific diagnostic
criteria, considerable overlap remained. Thus, in 2001, an international
group of critical care specialists met to provide some resolution to the
case definition dilemma [11]. The results were standardized definitions pub-
lished in 2003 and shown in Table 1 [12]. Widespread use of these definitions
has helped clarify the epidemiology and outcomes of persons with sepsis.
Studies looking at sepsis during pregnancy are particularly difficult to
analyze because of the retrospective nature of the data, small numbers,

* Corresponding author.
E-mail address: dguinn@whallc.com (D.A. Guinn).

0889-8545/07/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ogc.2007.06.009 obgyn.theclinics.com
460 GUINN et al

Table 1
Definitions of sepsis
Condition Definition
Infection A microbial phenomenon characterized by an
inflammatory response to the presence of
microorganisms or the invasion of normally sterile
host tissue by those organisms
Bacteremia Presence of viable bacteria in the blood; may be
transient and of no clinical significance; presence
alone not sufficient to diagnose sepsis
Sepsis Systemic inflammatory response to infection
Systemic inflammatory Widespread inflammatory response defined by two
response syndrome or more of the following:
Temperature O38 C or !36 C
Pulse O90 beats/min
Respiratory rate O20/min or PaCO2 !32 mm Hg
White blood cell count O12,000 mm3 or
!4000 mm3 or O10% immature (band) forms
Severe sepsis Sepsis with associated organ failure
Septic shock Sepsis with hypotension refractory to fluid
resuscitation
Data from Levy MM, Fink MP, Marshall JC, et al, for the International Sepsis Definitions
Conference. The 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Confer-
ence. Crit Care Med 2003;31:1250–6; and American College of Chest Physicians/Society of
Critical Care Medicine Consensus Conference. Definitions for sepsis and organ failure and
guidelines for the use of innovative therapies in sepsis. Crit Care Med 1992;20:864–74.

and different methodologies. The varying definitions of bacteremia, sepsis,

septic shock, and the systemic inflammatory response syndrome over time
make data comparison difficult. Furthermore, mortality estimates from
sepsis during pregnancy may be underestimated partly because, in many
studies, sepsis is neither defined nor classified as an infectious cause of mor-
tality [13].
A recent review of the global burden of maternal sepsis conducted by the
World Health Organization (WHO) highlighted the problem of imprecise
nonuniform terminology [14]. WHO collected published data and examined
regional office databases from around the world. Reporting was not
uniform. As expected, rates of obstetrical sepsis differed dramatically in devel-
oping and developed countries. The incidence of ‘‘sepsis’’ varied from a low of
0.96 to a high of 7.04 per 1000 women age 15 to 49 years. Similarly, estimated
mortality rates ranged from 0.01 to 28.46 per 100,000 women age 15 to 49
years. Despite the obvious limitations of combining data from varied sources,
common themes were noted. These are delineated in Box 1. Lack of access to
prenatal care is strongly associated with higher sepsis rates. In developing
countries, malaria, HIV, and community-acquired pneumonia are common
‘‘nonobstetric’’ causes of sepsis during pregnancy [15]. Obstetrical sepsis is
primarily the result of pelvic infections due to choriomanionitis, endometritis,
wound infections, septic abortion, or urinary tract infections [9,16–19].
 SEPSIS DURING PREGNANCY 461

Box 1. Bacterial infections associated with septic shock

in the obstetric patient
Obstetric
Chorioamnionitis
Postpartum endometritis (more common after cesarean section)
Septic abortion
Septic pelvic thrombophlebitis
Cesarean wound infection
Episiotomy infections
Nonobstetric
Appendicitis
Cholecystitis
Urinary tract infections
Pyelonephritis (perinephric abscess, renal calculi)
Pneumonia
HIV
Malaria
Invasive procedures
Necrotizing fasciitis
Infected cerclage
Postchorionic villus sampling/amniocentesis (septic abortion)
Miscellaneous
Toxic shock syndrome

Chorioamnionitis rates are strongly associated with preterm delivery and

number of vaginal examinations [15]. Endometritis and wound infections
are common complications of cesarean delivery and are probably underre-
ported in most series because the diagnosis is frequently made as an outpa-
tient following discharge [14,20]. The most common risk factor for
maternal sepsis is cesarean delivery [14]. The cesarean rate appears to be in-
creasing worldwide. It is estimated that the increasing rates of cesarean will
result in increased numbers of women diagnosed with infection and sepsis.
Septic abortion is also common throughout the world. Access to birth con-
trol and ‘‘legalized’’ abortion significantly influence the rates of septic abor-
tion and maternal sepsis [15]. Septic abortion is not only associated with
maternal morbidity and mortality short term, but is associated with second-
ary infertility and chronic pain in survivors [14]. Asymptomatic bacteriuria,
urinary tract infections, and pyelonephritis rates all increase during preg-
nancy [18,19,21,22]. When any of these are left untreated, sepsis can occur.
Pregnant patients who present with sepsis, regardless of the cause, are at
high risk for delivery during their admission [23]. Outcomes for the baby
462 GUINN et al

depend upon the gestational age at the time of delivery and the presence of
neonatal infection. The prognosis for the mother’s recovery from septic
shock is favorable, particularly when compared with prognoses for nonob-
stetric patients [1,3,5,24,25]. For the gravid patient, the factors contributing
to a decreased rate of septic shock as well as a favorable prognosis in the
face of septic shock include a younger patient profile with fewer comorbid-
ities and organisms that are usually responsive to common broad-spectrum
antimicrobials [16,26]. In addition, a common site of infection in the pregnant
patient is the pelvis, a location amenable to medical and surgical intervention.
These characteristics also lead to a lower mortality rate [16,26–28].
The following is a review of the microbiology, pathophysiology, and
management guidelines to reduce morbidity and mortality from obstetrical
sepsis.

Microbiology and risk factors

Most obstetric patients who develop bacteremia do not develop sepsis. In
multiple studies, the prevalence of bacteremia in the obstetric population is
estimated to be 7.5 per 1,000 admissions, of which 8% to 10% develop sep-
sis [29–35]. Although gram-negative bacteria are commonly identified in
patients with sepsis, gram-positive species have emerged as a predominant
pathogen over the last decade [5]. Still, in the obstetric patient with sepsis,
many studies have shown aerobic gram-negative rods to be the principal
etiologic agents followed by gram-positive bacteria and mixed or fungal
infections [16,17,25]. Ledger and colleagues [31] identified gram-negative
bacteremia in approximately 3 out of 1000 obstetric admissions. The most
frequently recovered organisms were Escherichia coli, Enterococci, and
Beta-hemolytic streptococci. The most commonly isolated anaerobes were
Peptostreptococci, Peptococci, and Bacteroides. Blanco and colleagues [29]
noted that in 176 bacteremic obstetric patients, E coli accounted for 57%
of cases with group B streptococci noted in 28%. Endotoxin, a complex
lipopolysaccharide present in the cell wall of gram-negative bacteria, is
released when the bacteria are lysed, thereby initiating the inflammatory
cascade. Gram-positive species cause a similar response by release of
exotoxin. In the nonpregnant population, 15% to 20% of severe sepsis cases
are polymicrobial in nature [2,7]. Less is known about the obstetric popula-
tion. However, multiple organisms are commonly implicated in pelvic
infection. Box 2 lists the organisms that have been identified in obstetric
septic shock.

Pathophysiology
In general, the pathophysiology of sepsis is complex and not completely
understood. The severity of the condition is determined not only by the
 SEPSIS DURING PREGNANCY 463

Box 2. Pathogens implicated in obstetric septic shock

Gram-positive cocci
Pneumococcus
Streptococcus, groups A, B, and D
Staphylococcus aureus
Gram-negative rods
Escherichia coli
Hemophilus influenzae
Klebsiella species
Enterobacter species
Proteus species
Pseudomonas species
Serratia species
Gram-positive rods
Listeria monocytogenes
Anaerobes
Bacteroides species
Clostridium perfringens
Fusobacterium species
Peptococcus
Peptostreptococcus
Fungal species

virulence of the offending organism, but also by a number of host factors,

including age, genetic factors, site of infection, and the presence of comorbid
conditions. Our understanding primarily comes from critically ill general
medical or surgical patients. Specific information relating to pregnant
patients is limited as these women are typically not included in most studies
because the condition in pregnancy is rare, because of concern for the
developing fetus, and because of confounding by the significant physiologic
changes associated with pregnancy.
The goal of the host inflammatory response is to localize and eliminate
any invading organisms. With microbial infection a complex cascade of
events occurs. Macrophages and neutrophils are activated, which in turn
directly release inflammatory mediators and activate CD4 T cells. These
cells then release proinflammatory cytokines, including tumor necrosis
factor-a (TNF-a) and interleukin-1 (IL-1), which have a variety of physio-
logic functions aimed at containing and eliminating the infection. Cytokines
further recruit other macrophages and neutrophils. The cytokines also lead
to the generation of oxygen free radicals, production of proteases, induction
464 GUINN et al

of nitric oxide synthase, release of vasoactive hormones, increase in

endothelial permeability, activation of the coagulation cascade, and inhibi-
tion of the fibrinolytic system. Although these various physiologic responses
are important in containing an infection, they can lead to host tissue
damage. To minimize this, CD4 T cells are also stimulated to release anti-
inflammatory cytokines, which attempt to keep the system in check. It is
the failure to control the inflammatory response as well as over-expression
of the anti-inflammatory response that leads to the pathologic events
associated with sepsis [1,6,36].
Although the inflammatory cascade exerts its effects throughout the body
simultaneously, the hemodynamic responses are prominent early in the
clinical course. Vasodilation and a decrease in systemic vascular resistance
(SVR) occur at least in part because of increased nitric oxide production.
The ‘‘relative’’ hypovolemia stimulates the baroreceptors, which in turn
activate the sympathetic nervous system, resulting in tachycardia. The
tachycardia combined with the decreased SVR results in increased cardiac
output. In response to this, vasoconstrictive hormones, such as vasopressin
and endothelin, are released. The renin-angiotensin system is also activated.
These responses attempt to maintain vascular tone along with increasing the
intravascular volume by increased sodium and water reabsorption in the
kidneys [36].
The proinflammatory cytokines TNF-a and IL-1 also have procoagulant
effects. The coagulation cascade is activated by TNF-a–induced release of
tissue factor from endothelial cells. This ultimately results in thrombin pro-
duction. Thrombin along with thrombomodulin activates protein C, which
acts to inhibit coagulation. By down-regulating thrombomodulin, TNF-a
inhibits this anticoagulation balance in the system. Fibrinolysis is also
decreased by a TNF-a– and IL-1–stimulated increase in PAI-1, an inhibitor
of fibrinolysis [1].
Complex physiologic adaptations happen during pregnancy and they
must be understood when managing the pregnant patient with sepsis. In
the cardiovascular system, the heart rate increases, peripheral vasodilation
occurs, leading to a decrease in blood pressure and an increase in cardiac
output [37]. These changes may not only mask the initial presentation of
sepsis, but can further aggravate decreased organ perfusion seen in the septic
patient [1,36]. Myocardial depression seen in advanced sepsis further com-
plicates the situation [6].
Red cell mass and plasma volume both increase in pregnancy with
a greater increase in the latter resulting in red blood cell dilution and thus
anemia. Albumin and protein concentrations also decrease, likely because
of the same mechanism. The decrease in albumin and protein concentrations
results in a lower colloid osmotic pressure, leading to increased interstitial
fluid [37]. These physiologic alterations are generally protective in preg-
nancy. However, in certain pathologic states, pregnant patients are more
susceptible to develop pulmonary edema [1,38].
 SEPSIS DURING PREGNANCY 465

Changes in the respiratory system include increased tidal volume associ-

ated with a decrease in residual volume and functional reserve capacity.
There is also a slight decrease in total lung capacity. The respiratory rate
may increase slightly. The vital capacity is unaffected. Minute ventilation
increases, leading to a decrease in PaCO2 and a compensatory decrease in
serum bicarbonate to maintain a normal pH. The net result is a compensated
respiratory alkalosis. These physiologic adaptations are beneficial in a nor-
mal pregnancy. However, in the setting of sepsis and/or respiratory failure,
these changes predispose the patient to rapid declines in oxygenation and
decreased ability to compensate or buffer a metabolic acidosis [38].
In the kidneys, the renal plasma flow and the glomerular filtration rate
(GFR) increase, resulting in decreased serum levels of blood urea nitrogen
and creatinine. Normal nonpregnant serum levels of these metabolites
may suggest mild renal compromise in the pregnant patient. The renal col-
lecting system dilates because of smooth muscle relaxation. The growing
uterus may add to the dilation by mechanical obstruction of the ureter.
This leads to urinary stasis and an increased incidence of asymptomatic bac-
turia, which in turn is associated with an increased risk of pyelonephritis if
left untreated [37,38].
Throughout the gastrointestinal tract, smooth muscle tone is decreased,
leading to increased esophageal reflux, decreased gastric emptying, and
delayed intestinal transit. Pregnant patients are thus more susceptible to
aspiration pneumonia [38]. Transaminases and bilirubin values are slightly
decreased in pregnancy, while alkaline phosphatase levels are increased
due to placental production. Lactate dehydrogenase levels are unchanged.
Understanding these biochemical changes is important when trying to inter-
pret laboratory values in the septic obstetrical patient [37].
White blood cell counts increase throughout pregnancy while platelet
counts decrease. A number of changes occur in the clotting cascade as
well. Factors VII, VIII, IX, X, and XII; von Willebrand’s factor; and fibrin-
ogen all increase, while protein S decreases. Protein C and antithrombin III
remain unchanged. There is also a decrease in the activity of the fibrinolytic
system mediated by an increase in the plasminogen activator inhibitors
1 and 2 (PAI-1 and -2). As a result of the increased clotting factors along
with the decreased anticoagulant activity and decreased fibrinolysis, preg-
nancy is associated with an increased risk of thrombosis [39,40]. The net
effect is a procoagulant state leading to an increase in the risk of thrombosis
and the potential exacerbation of disseminated intravascular coagulation
(DIC) [1].
Many of these changes are further accentuated during labor and delivery.
Heart rate and respiratory rates may increase with the exertion and stress of
labor. Intravenous fluids are typically given. White blood cell counts are
even more increased. Epidural use is common and prolonged use is
associated with a rise in maternal temperature [41]. Large fluid shifts and
significant blood loss can be seen with delivery. Cesarean delivery is
466 GUINN et al

common today and is a factor associated with a greater risk of hemorrhage,

infection, and thrombosis [20,42]. Diagnosis and management during labor
and delivery can thus be particularly challenging.
Acute respiratory distress syndrome (ARDS) is a severe and life-threaten-
ing complication seen in sepsis. Although ARDS has multiple causes, not all
of which are infection-related, the most common cause is reported to be
sepsis. Pneumonia has been reported to be the most common infection
associated with ARDS. The condition has also been associated with
pyelonephritis and chorioamnionitis/endomyometritis [27,38,43]. ARDS-
complicating sepsis is associated with maternal mortality rates ranging
from 10% to 50% [27,43,44]. Some of this wide variation is due to the small
number of patients in the reported case series. Generally, the mortality rate
associated with ARDS appears to be lower in pregnant patients than in the
nonpregnant population [38].
Although hemodynamic compromise is a hallmark of severe sepsis and
the physiologic changes occurring during pregnancy can theoretically aggra-
vate the condition, significant hemodynamic instability is not as common as
respiratory failure [27]. Reports describing sepsis complicated by renal fail-
ure in pregnancy are limited and difficult to separate from reports about
renal failure complicating other conditions. The mortality rate in these
patients tends to be lower than would be expected by the severity of the
patient’s condition [27]. This is in contrast to the nonpregnant population
where the combination of sepsis and renal failure is associated with a mortal-
ity rate as high as 70% [36].
Because of the procoagulant state associated with sepsis, DIC is not
uncommon in severe sepsis and frequently complicates patients with multi-
organ failure. It can lead to microthrombi in the glomerulii, resulting in
renal failure. These two complications together–DIC and renal failure–are
associated with a mortality rate of 75% among nonpregnant patients. In
the pregnant patient, the combination of DIC and respiratory failure in sep-
sis is associated with increased maternal mortality [44]. Microthrombi can
also occur in the liver, leading to failure. This failure can decrease protein
production, further decreasing colloid oncotic pressure and aggravating
DIC because of decreased production of clotting factors [15]. In the gastro-
intestinal tract, tissue hypoxia can lead to bleeding and pancreatitis [27].
Multiorgan failure in nonpregnant patients with sepsis is associated with
mortality rates in excess of 70%. The highest rates have been associated with
involvement of three or more organ systems [5,28]. Other factors influencing
survival in patients with multiorgan failure in this population are age, asso-
ciated comorbid medical conditions, and duration of the organ failure [28].
There appears to be an increased risk of mortality in the pregnant popula-
tion as well. However, it is not of the same magnitude [27,44]. Acute
Physiology and Chronic Health Evaluation (APACHE II) scores are used
to estimate risk of mortality in intensive care unit patients. The scores are
not reliable for septic pregnant patients and tend to overestimate their
 SEPSIS DURING PREGNANCY 467

risk of mortality [15]. Therefore, counseling regarding prognosis and

decisions regarding therapy should be based on the presence of multiorgan
failure, the presence of comorbid conditions, and the potential to identify
and treat the source of infection [15].

Management
Early recognition and prompt, aggressive therapy is crucial to reduce
maternal and fetal morbidity and mortality in women with suspected sepsis.
To help standardize effective resuscitation strategies for persons with
suspected sepsis, the Surviving Sepsis Campaign was initiated in October
2002. Subsequently, the working group has expanded and revised its recom-
mendations [45,46]. Therapeutic bundles have been developed for early
resuscitation (0–6 hours) and management (6–24 hours). The bundles
were developed using evidence-based medicine principals. Table 2 reviews
the strength of the evidence and the basis of the recommendations. Fig. 1
is an overview of early goal directed therapy (EGDT). Implementation
of EGDT improves survival and is cost-effective in a variety of settings
[47]. The following is a summary of the recommendations of the Surviving
Sepsis Campaign with some caveats as they apply to the obstetrical
population.

Diagnosis
A thorough history and physical examination is required to evaluate
potential sources of infection. Ideally, cultures should be obtained before in-
stituting antibiotic therapy to identify suspected pathogens, to monitor
effectiveness of therapy, and to guide appropriate use of antibiotics [45].

Table 2
Evidence-based medicine guidelines rating systems
Grade Basis of Grade
Recommendations
A Supported by at least two level I investigations
B Supported by at least one level I investigation
C Supported by level II investigations only
D Supported by at least one level III investigation
E Supported by level IV or V evidence
Evidence
I Large randomized control trial with clear-cut results
II Small, randomized trials with uncertain results
III Nonrandomized, contemporaneous controls
IV Nonrandomized, historical controls, and expert opinion
V Case series, uncontrolled studies, and expert opinion
468 GUINN et al

Suspected infection

The High Risk Patient: SBP<90 mmHg after

20-40 cc/kg volume challenge or lactic acid
> 4 mmole/liter

Antibiotics within 1 hour and Source Control

<8 mmHg
CVP Crystalloid

>8-12 mmHg

Decrease <65 or >90 mmHg Vasoactive

Oxygen MAP
Agent(s)
Consumption
>65-90 mmHg

>70%
ScvO2 <70% Packed red blood
cells to Hct >30% <70%

>70%
Ionotrope (s)

Goals
No Achieved

Fig. 1. Overview of early goal directed therapy. CVP, central venous pressure; MAP, mean ar-
terial pressure; ScvO2, central venous oxygen saturation; SBP, systolic blood pressure. (From
Otero RM, Nguyen HB, Huang DT, et al. Early goal-directed therapy in severe sepsis and sep-
tic shock revisited concepts, controversies, and contemporary findings. Chest 2006;130(5):
1579–95; with permission.)

A grade D recommendation supports the practice of obtaining two blood

cultures be obtained with at least one drawn peripherally and one drawn
through a central access device if available. If the clinical situation warrants
it, cultures of additional sites, including those from urine, wounds, respira-
tory secretions, and cerebrospinal fluid, should be performed. In the gravid
woman with preterm labor, ruptured membranes or suspected chorioamnio-
nitis, amniocentesis should be performed. In addition to culture, the
amniotic fluid can be sent for Gram stain, white count, and glucose levels
[48]. The results of these markers are available within hours, whereas culture
results usually take days. In selected hospitals, evaluation of the proinflam-
matory cytokines (TNF-a, IL-1 beat, and IL-6) may be available [49,50].
 SEPSIS DURING PREGNANCY 469

Cervical cultures and/or placental cultures may also be useful. In postpar-

tum women, transcervical endometrial sampling has been well described
[51]. In the authors’ experience, endometrial cultures obtained transcervi-
cally are frequently contaminated by normal vaginal flora, particularly
anaerobes, and are not particularly helpful.

Initial resuscitation
Once severe sepsis is suspected, EGDT has been shown to improve sur-
vival, according to grade B evidence. Grade B evidence also suggests that
during the first 6 hours of resuscitation (early therapy), the goals should in-
clude all of the following: central venous pressure (CVP) of 8 to 12 mm Hg,
mean arterial pressure of greater than or equal to 65 mm Hg, urine output
greater than 0.5 mL/kg/h, and central venous (superior vena cava) or mixed
venous oxygen saturation of greater than or equal to 70% (Table 3) [45,52].
These goals were established in nonpregnant patients. Either crystalloids or
colloids can be used for volume expansion, according to grade C evidence.
Crystalloids have a larger volume of distribution and may result in more
edema than colloids [45]. Large amounts of fluid (6–10 L) may be required
initially [47]. Blood pressure, pulse rate, urine output, oxygen saturation,
and fetal status can be used to judge clinical response to fluid challenges.
CVP and pulmonary artery wedge pressure measure cardiac filling
pressures. Their use in general is limited because of errors in routine mea-
surement and confounding from use of mechanical ventilation and increased
abdominal pressure [45,53]. Their use in pregnancy has been widely
described in obstetrical patients [54–57]. In gravid women, the CVP and
pulmonary artery wedge pressures are not reliably related [54,57,58]. CVP
levels may be normal in gravidas with left ventricular dysfunction or pulmo-
nary edema. In contrast, the CVP may be elevated in women with no
evidence of pulmonary edema [54,57,59]. No studies specifically evaluate

Table 3
EGDT goals and normal values in pregnancy
Normal third-trimester
Measures Resuscitation Goals physiologic valuesa
Central venous pressure 8–12 mm Hg 4–10 mm Hg
Mean arterial pressure R65 mm Hg 84–96 mm Hg
Urine output O0.5 mL/kg/h Minimum 0.5 mL/kg/h
Mixed venous oxygen O70% O80%b
saturation
Heart rate Decreasing in response to 83 (10) beats/min
treatment
a
Normal values in pregnancy. Data from Norwitz ER, Robinson JN, Malone FD. Preg-
nancy-induced physiologic alterations. In: Dildy GA III, Belfort MA, Saade G, et al, editors.
Critical care obstetrics. 4th edition. Malden (MA): Blackwell Science; 2004. p. 19–42.
b
Dependent upon cardiac output, fraction of inspired oxygen, and oxygen consumption.
Data from Refs. [45,52,47] for EGDT goals and [37] for normal values in pregnancy.
470 GUINN et al

the use of CVP in obstetrical patients with sepsis. Nonetheless, the authors
believe it is reasonable in the setting of suspected sepsis to use CVP measure-
ments to guide initial fluid resuscitation in women with low CVP and
evidence of hypoperfusion. Consideration may be given to using a pulmo-
nary artery catheter in gravid, septic women with preeclampsia and/or car-
diomyopathies where volume expansion may increase the risk of pulmonary
edema and/or ARDS [54,57,59,60].
Patients receiving mechanical ventilation may benefit from a higher tar-
geted CVP pressure of 12 to 15 mm Hg [45]. This may also be required in
patients with increased abdominal pressure, which has increasing relevance
with advancing gestation. Displacement of the uterus using lateral tilt or use
of a hip roll minimizes aorto-caval compression and improves venous return
to the heart.
Grade B evidence suggests that, if patients do not respond to volume
expansion and if a central venous oxygen saturation or mixed venous oxygen
saturation of greater than or equal to 70% is not achieved within 6 hours of
diagnosis, transfusion with packed red blood cells to achieve a hematocrit of
greater than or equal to 30% and/or administration of a dobutamine
infusion (maximum of 20 mg/kg/min) is indicated. If time allows, type-specific,
CMV-safe (leukoreduced) transfusions are preferred. There is no contrain-
dication to using inotropes and/or vasopressors in gravid patients. Inotropes
and vasopressors are administered by standardized protocols. The infusions
are titrated upward to achieve the desired increase in blood pressure and/or
cardiac output. Dobutamine, the first choice inotrope for patients with
sepsis who have evidence of low cardiac output despite adequate filling
pressures [45,53,61–64], is a potent inotrope with modest vasodilatory prop-
erties. Dobutamine increases cardiac contractility and improves cardiac
output without a significant increase in heart rate. In patients with severe
shock, vasopressors may be required to correct hypotension. The most
commonly used vasopressors are dopamine, norepinephrine, epinephrine,
and phenylephrine. The two agents specifically recommended in the
Surviving Sepsis Guidelines for sepsis with refractory hypotension are dopa-
mine and norepinephrine [45]. Dopamine increases mean arterial pressure
and cardiac output because of an increase in stroke volume and heart
rate. Norepinephrine improves mean arterial pressure by its vasoconstrictive
properties. Dopamine and norepinephrine can reduce blood flow to the
periphery, the gut, and the uterus. Thus, close monitoring is required.
Vasopressin may be considered in patients with refractory shock. In the set-
ting of shock, vasopressin is administered at a rate of 0.01 to 0.04 U/min
[45,63].

Antibiotic therapy
Grade D evidence supports the initiation of antibiotic therapy as soon as
possible. The initial selection of empiric antibiotics is based on the patient’s
 SEPSIS DURING PREGNANCY 471

history, known drug allergies, physical examination, underlying disease, and

clinical condition. It is essential that physicians are knowledgeable regarding
community and hospital-specific antibiotic resistance patterns when
initiating empiric therapy. Patients with severe sepsis or septic shock
warrant broad-spectrum therapy until the causative organism or organisms
are identified and their susceptibilities defined [45]. Once a suspected source
or pathogen is identified, restricting the number of antibiotics is appropri-
ate. However, most patients with sepsis or septic shock have negative blood
cultures [45]. This may be more common in pregnancy where antibiotic use
is high [21,23,65]. If a positive blood culture is identified in the setting of
suspected obstetrical sepsis, caution should be used in narrowing antibiotic
coverage before recovery. The majority of obstetrical infections are polymi-
crobial [7,24,29,34,66]. One investigator hypothesized that in an obstetrical
population, ‘‘the microbe found in blood culture represent only the tip of an
iceberg of pathogens at the original site of infection’’ [18,66]. Using clinical
findings and other culture results, the clinician ultimately needs to decide
whether to continue broad-spectrum therapy, narrow therapy, or stop
therapy. Consultation with microbiology and infectious disease specialists
is recommended.
The task of selecting the appropriate antibiotic regimen in pregnancy is
further complicated by several factors. Pharmacokinetic studies of antibiotic
dosing in pregnancy are limited. All of the physiologic adaptations to
normal pregnancy can have an impact drug availability, concentration,
and effectiveness [21,67–69]. In particular, the increased volume of distribu-
tion and changes in absorption and distribution can affect drug levels. In
general, antibiotics that are primarily eliminated by renal excretion result
in lower serum levels during pregnancy as compared with those for the
nonpregnant patient. The half-life of certain antibiotics is shorter. Transpla-
cental passage occurs to some degree for all antibiotics according to the
physicochemical properties of the drug. When choosing antibiotic regimens,
attempts should be made to maximize effectiveness and minimize fetal harm.
However, there may be specific cases where the ‘‘best’’ drug for the mother
may not be ‘‘safe’’ for the fetus. In these rare situations, consultation with
maternal–fetal medicine specialists is strongly recommended. Most antibi-
otics are safe to use in pregnancy. A comprehensive review of specific
antibiotics in pregnancy is beyond the scope of this article. In general, the
classes of antibiotics to avoid if a suitable alternative exists are tetracyclines,
fluoroquinolones, and erythromycin estolate.
In septic pregnant women with obstetrical infections, most infections
are the result of mixed flora, including gram positives, gram negatives,
and anaerobes. Therefore, the approach to antibiotics is generally to
administer broad-spectrum antibiotics using two to three agents. The
most commonly cited combination is ampicillin, gentamicin, and clindamy-
cin or metroniadazole [23,68]. There is limited data on efficacy and safety
of once-daily dosing of gentamicin during pregnancy. Once-daily or
472 GUINN et al

twice-daily dosing options are preferred postpartum [67]. Given the

increased volume of distribution and clearance, drug levels are indicated
in critically ill patients.

Source control
Once resuscitative measures are initiated, evaluation for a specific focus
of infection that may be amenable to source control measures is essential.
Transporting unstable patients to radiology may not be safe. Ultrasound
at the bedside can be an invaluable tool. It is recognized, based on grade
E evidence, that emergent intervention for necrotizing soft tissue infection
or intestinal ischemia is essential to reduce morbidity and mortality [45].
Most obstetrical infections are amenable to source control measures. In
women with chorioamnionitis, delivery should be accomplished as soon as
possible, regardless of the gestational age. Obstetricians must use clinical
judgment in determining route of delivery. Vaginal delivery is preferred in
the patient who has a favorable cervix and/or is laboring spontaneously.
If a long induction of labor is anticipated, cesarean may be a better choice
in the hemodynamically stable patient. General anesthesia is necessary in
cases where urgent delivery is required because of fetal distress that is not
responsive to maternal resuscitation or in cases where cesarean is indicated
and the mother is hemodynamically unstable [41]. Intubation of the gravid
patient can be particularly difficult because of airway edema and anatomical
challenges. Intubation of the gravid patient is also associated with an
increased risk of gastric aspiration [70]. Skilled anesthesiologists familiar
with the particular challenges of pregnancy should be present for intubation
and surgery. Preoxygenation and rapid-sequence induction with cricoid
pressure are essential to control the airway and to reduce the risk of
aspiration [70]. Patients who develop respiratory failure and sepsis following
a seizure or intubation should receive antibiotic coverage for potential
aspiration pneumonia.

Adjunctive measures
The use of corticosteroids in sepsis is controversial. The Surviving Sepsis
Campaign endorsed the use of intravenous corticosteroids (hydrocortisone
200–300 mg/d for 7 days in three to four divided doses or by continuous
infusion) in patients with septic shock who require vasopressors. This
recommendation is supported by grade A evidence. [45]. Hydrocortisone
is not contraindicated in pregnancy. Other investigators recommend a corti-
cotropin stimulation test to identify patients who would benefit from
corticosteroids [71]. Cortisol levels and response to corticotropin may be
influenced by pregnancy. Therefore, the authors recommend empiric
therapy with corticosteroids in the septic gravid patient. In cases where
preterm delivery of a viable fetus is likely, antenatal corticosteroid adminis-
tration with betamethasone (12 mg intramuscularly every 24 hours times
 SEPSIS DURING PREGNANCY 473

two) or dexamethasone (6 mg intramuscularly every 12 hours times four) is

also recommended [72].
The use of recombinant human activated protein C is also controversial.
Patients at high risk for death (multiorgan failure, septic shock, or sepsis-in-
duced ARDS) are potential candidates, according to grade B evidence [45].
Recombinant human activated protein C increases the risk of bleeding and
is contraindicated in patients with active bleeding or recent surgery (within
30 days) [45,73,74]. Its use in obstetrical patients is limited to case reports
[75,76]. Pregnancy should not be an absolute contraindication to its use.
However, most septic gravid women are not candidates because of threat-
ened labor or recent cesarean [23]. The implications for the fetus and new-
born are largely unknown.
Thresholds for transfusion following initial resuscitation have not been
established in pregnant patients. In general, a transfusion threshold of 7.0
to 9.0 g/dL is reasonable, according to grade B evidence. This may not be
sufficient for the antepartum patient with high oxygen consumption and
real potential for blood loss due to delivery. In an individual patient, the
decision to transfuse should be based on maternal and fetal status. The fetal
heart rate tracing and/or biophysical profile may be used as an indirect mea-
sure of maternal oxygen delivery and uterine blood flow. If there is evidence
of non-reassuring fetal heart rate and the patient has not responded to fluid
resuscitation, mothers should be transfused liberally. The benefit to correct-
ing clotting abnormalities in patients with no evidence of bleeding is also
unclear [45]. A patient with platelet levels of less than 5000/mm3 should
be transfused regardless of apparent bleeding [45]. In patients who are
candidates for surgery or invasive procedures, coagulation defects should
be corrected and platelets transfused to a level greater than 50,000/mm3
preoperatively [45].
Patients with sepsis are at high risk for sepsis-induced acute lung injury
and ARDS. Initiation of lung-protective ventilation is important. According
to recommendations based on grade B evidence, ‘‘low’’ tidal volumes should
be used (6 mL/kg of predicted body weight) while maintaining end-inspira-
tory plateau pressures less than 30 cm H2O (grade B) [45]. To accomplish
these goals, permissive hypercapnia (allowing PaCO2 ‘‘above normal’’)
may be necessary. These goals were established in nonpregnant patients.
If introduced slowly, increases in carbon dioxide are well tolerated in the
absence of raised intracranial pressure or a severe metabolic acidosis
[23,70,77–79]. In pregnancy, a normal PaCO2 is 26 to 32 mm Hg and serum
bicarbonate is 18 to 21 mEq/L [37]. The respiratory alkalosis associated with
pregnancy is essential for diffusion of carbon dioxide to occur from the fetus
to the mother. Over time, the fetus gains the capacity to increase bicarbon-
ate production in the kidneys and adapt to ‘‘respiratory acidosis.’’ Mechan-
ical ventilation with low tidal volumes is probably well tolerated in most
cases. However, permissive hypercapnia for prolonged periods may result
in fetal acidosis. If hypoxia is also present, the potential for fetal
474 GUINN et al

myocardial dysfunction and death exists. Fetal acidosis can be identified

by the presence of late decelerations on fetal heart monitoring, depressed
biophysical profile score, or abnormal Doppler studies. Ultimately, the
only way to improve fetal status short of delivery is to improve maternal
status. Attempts to minimize respiratory complications are essential to
improve fetal and neonatal outcome. A comprehensive review of respiratory
strategies to manage ARDS and weaning from mechanical ventilation is
beyond the scope of this article. Several important references are included
for review [23,70,77–79]. Unfortunately, to the authors’ knowledge, all of
the randomized trials evaluating respiratory therapies excluded pregnant
women. Therefore, definitive recommendations are limited.
Patients who require mechanical ventilation require anxiolytics, analge-
sics, sedatives, and/or neuromuscular blockade [23,45,77–79]. These classes
of drugs can all cross the placenta and may result in decreased fetal heart
rate variability and depressed fetal movements [80,81]. Fetal assessment
may be difficult. Attempts to minimize sedation and/or paralysis are recom-
mended to avoid complications and to assess neurologic status [45].
Hyperglycemia is a common complication of sepsis. Insulin infusions
should be used to maintain blood glucose less than 150 mg/dL , according
to grade D evidence [45]. In surgical patients, maintaining glucose values
between 80 and 110 mg/dL improved survival rates [74]. These target values
are appropriate for pregnant women. Frequent monitoring to avoid hypo-
glycemia and ketosis is important.
Septic patients are at risk for thrombosis. This risk is magnified in
pregnancy and in the postpartum period [40]. All pregnant women should
wear external compression stockings or intermittent compression devices.
Limited data exist regarding the optimal mechanical device in pregnancy.
Recommendations based on grade A evidence suggest that patients without
evidence of coagulopathy should receive prophylaxis with unfractionated
heparin, if delivery is likely to occur within 12 hours of administration, or
low molecular weight heparin, if delivery is not anticipated that soon
[40,41]. Stress ulcer prophylaxis with H2 receptor inhibitors should be
prescribed, according to grade A evidence [45]. These agents are not contra-
indicated in pregnancy. Nutritional therapy should be administered
enterally, if possible [82]. In patients with contraindication to tube feeding,
hyperalimentation can be administered peripherally or centrally. Increased
caloric and nutritional demands in pregnancy must be met. Consultation
with nutritionists helps develop an individualized plan of care.

Fetal monitoring and obstetrical interventions

In general, women with obstetrical sepsis are young and healthy. Aggres-
sive intensive care should be employed with the goal of optimizing maternal
and fetal health. Interventions that improve maternal hemodynamic stabil-
ity and oxygen delivery to the fetus should result in improved maternal and
 SEPSIS DURING PREGNANCY 475

fetal outcomes. Decisions regarding timing of delivery and interventions to

prolong pregnancy using tocolytics are complex. For pregnancies presenting
before viability, optimizing maternal health is paramount. Pregnancies near
term can be delivered once the maternal status is stable. For pregnancies
between 24 and 32 weeks’ gestation, decisions should be made based on
maternal prognosis and family desires. Operative intervention on behalf
of the fetus in an unstable mother increases maternal morbidity and mortal-
ity. Perimortem cesarean is an option. However, the neonatal outcomes
depend on several important factors, including timing from arrest to
delivery, fetal reserves, and the presence of fetal infection [80].

Summary
Sepsis is a leading cause of death in pregnancy, particularly in the devel-
oping world, and results in significant perinatal mortality. These deaths
occur despite the younger age of pregnant patients, the low rate of comorbid
conditions, and the potential for effective interventions that should result in
rapid resolution of illness. To date, no ‘‘evidence-based’’ recommendations
are specific to the pregnant patient who is critically ill or septic. Until
pregnant women are included in therapeutic trials in the intensive care
unit, particularly in the setting of sepsis, therapy will remain empiric and
anecdotal with the potential for excess morbidity and mortality. Optimal
care for the septic patient requires a multidisciplinary team with expertise
in obstetrics, maternal–fetal medicine, critical care, infectious disease,
anesthesia, and pharmacy. Coordination of care and good communication
amongst team members is essential. Incorporation of EGDT for suspected
sepsis into obstetric practice would seem to be essential to optimize maternal
and neonatal outcomes.

References
[1] Fernandez-Perez ER, Salman S, Pendem S, et al. Sepsis during pregnancy. Crit Care Med
2005;33(Suppl):S286–93.
[2] Sands KE, Bates DW, Lanken PN, et al. Epidemiology of sepsis syndrome in 8 academic
medical centers: Academic Medical Center Consortium Sepsis Project Working Group.
JAMA 1997;278:234–40.
[3] Angus DC, Linde-Zwirble WT, Lidicker J, et al. Epidemiology of severe sepsis in the United
States: analysis of incidence, outcome and associated costs of care. Crit Care Med 2001;29:
1303–10.
[4] Hoyert DL, Arias E, Smith BL, et al. Deaths: final data for 1999. National vital statistics re-
ports. vol. 49. No. 8. Hyattsville, MD: National Center for Health Statistics, 2001. (DHHS
publication no. (PHS) 2001-1120 PRS 01-0573).
[5] Martin GS, Mannino DM, Eaton S, et al. The epidemiology of sepsis in the United States
from 1979 through 2000. N Engl J Med 2003;348:1546–54.
[6] Hotchkiss RS, Karl IE. The pathophysiology and treatment of sepsis. N Engl J Med 2003;
348:138–50.
476 GUINN et al

[7] Sheffield JS. Sepsis and septic shock in pregnancy. Crit Care Clin 2004;20:651–60.
[8] Increase in National Hospital Discharge Survey rates for septicemiadUnited States,
1979–1987. MMWR Morb Mortal Wkly Rep 1990;39:31–4.
[9] Fein AM, Duvivier R. Sepsis in pregnancy. Clin Chest Med 1992;13:709–22.
[10] Bone RC, Balk RA, Cerra FB, et al. Members of the American College of Chest Physicians/
Society of Critical Care Medicine Consensus Conference Committee. American College of
Chest Physicians/Society of Critical Care Medicine Consensus Conference. Definitions for
sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit
Care Med 1992;20:864–74.
[11] Levy MM, Fink MP, Marshall JC, et al, for the International Sepsis Definitions Conference.
The 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference.
Crit Care Med 2003;31:1250–6.
[12] American College of Chest Physicians/Society of Critical Care Medicine Consensus Confer-
ence. Definitions for sepsis and organ failure and guidelines for the use of innovative thera-
pies in sepsis. Crit Care Med 1992;20:864–74.
[13] Chang J, Elam-Evans LD, Berg CJ, et al. Pregnancy-related mortality surveillance: United
States, 1991–1999. MMWR Surveill Summ 2003;52:1–8.
[14] Dolea C, Stein S. Global burden of maternal sepsis in the year 2000. Evidence and informa-
tion for policy (EIP). Geneva: World Health Organization; 2003. Available at: http://www.
who.int/entity/healthinfo/statistics/bod_maternalsepsis.
[15] Vasquez DM, Estenssoro E, Canales JS, et al. Clinical characteristics and outcomes of ob-
stetric patients requiring ICU admission. Chest 2007;131:718–24.
[16] Gonik B. Septic shock in obstetrics. Clin Perinatol 1986;13:741–54.
[17] Maupin RT. Obstetric infection disease emergencies. Clin Obstet Gynecol 2002;45:
393–404.
[18] Mabie WC, Barton JR, Sibai B. Septic shock in pregnancy. Obstet Gynecol 1997;90:553–61.
[19] Smaill F, Vazquez JC. Antibiotics for asymptomtic bacteriuria in pregnancy. Update of
Cochrane Database Syst Rev 2007;3.
[20] NIH State-of-the-Science Conference statement on cesarean delivery on maternal request.
NIH Consensus Scientific Statements. 2006. Mar 27–29;23(1):1–29.
[21] ACOG Practice Bulletin. Prophylactic antibiotics in labor and delivery. Number 47: October
2003.
[22] Heresi GA, Arroliga AC, Weidemann HP, et al. Pulmonary artery catheter and fluid man-
agement in acute lung injury and the acute respiratory distress syndrome [abstract ix]. Clin
Chest Med 2006;27(4):627–35.
[23] Jenkins TM, Troiano NH, Graves CR, et al. Mechanical ventilation in an obstetric popula-
tion: characteristics and delivery rates. Am J Obstet Gynecol 2003;188:549–52.
[24] Simpson KR. Sepsis during pregnancy. J Obstet Gynecol Neonatal Nurs 1995;24:550–6.
[25] Leonardi MR, Gonik B. Septic shock. In: Dildy GA III, Belfort MA, Saade G, et al, editors.
Critical care obstetrics. 4th edition. Malden (MA): Blackwell Science; 2004. p. 562–80.
[26] Freid MA, Vosti KL. The importance of underlying disease in patients with gram-negative
bacteremia. Arch Intern Med 1968;121:418–23.
[27] Afessa B, Green B, Delke I, et al. Systemic inflammatory response syndrome, organ failure,
and outcome in critically ill obstetric patients treated in an ICU. Chest 2001;120:1271–7.
[28] Knaus WA, Draper EA, Wagner DP, et al. Prognosis in acute organ failure syndrome. Ann
Surg 1985;202:685–93.
[29] Blanco JD, Gibbs RS, Castaneda YS. Bacteremia in obstetrics: clinical course. Obstet
Gynecol 1981;58:621–5.
[30] Cavanagh D, Knuppel RA, Sheperd JH, et al. Septic shock and the obstetrician/gynecolo-
gist. South Med J 1982;75:809–13.
[31] Ledger WJ, Norman M, Gee C, et al. Bacteremia on an obstetric-gynecologic service. Am J
Obstet Gynecol 1975;121:205–12.
 SEPSIS DURING PREGNANCY 477

[32] Wernstein MP, Murphy JR, Reller LB, et al. The clinical significance of positive blood
cultures: a comparative analysis of 500 episodes of bacteremia and fungemia in adults.
Rev Infect Dis 1983;5:54–60.
[33] Reimer LG, Reller LB. Gardnerella vaginalis bacteremia: a review of thirty cases. Obstet
Gynecol 1984;64:170–2.
[34] Monif GR, Baer H. Polymicrobial bacteremia in obstetric patients. Obstet Gynecol 1976;48:
167–9.
[35] Gibbs RS, Blanco JD. Streptococcal infections in pregnancy: a study of 48 bacteremias. Am
J Obstet Gynecol 1981;140:405–11.
[36] Schrier RW, Wang W. Acute renal failure and sepsis. N Engl J Med 2004;351:159–69.
[37] Norwitz ER, Robinson JN, Malone FD. Pregnancy-induced physiologic alterations. In:
Dildy GA III, Belfort MA, Saade G, et al, editors. Critical care obstetrics. 4th edition. Mal-
den (MA): Blackwell Science; 2004. p. 19–42.
[38] Cole DE, Taylor TL, McCollough DM, et al. Acute respiratory distress syndrome in preg-
nancy. Crit Care Med 2005;33(Suppl):S269–78.
[39] Mason BA. Systemic inflammatory response syndrome and acute respiratory distress
syndrome. In: Dildy GA III, Belfort MA, Saade G, et al, editors. Critical care obstetrics.
4th edition. Malden (MA): Blackwell Science; 2004. p. 329–45.
[40] Lockwood CJ. Pregnancy-associated changes in the hemostatic system. Clin Obstet Gynecol
2006;49:836–43.
[41] ACOG Practice Bulletin. Obstetric analgesia and anesthesia. Number 36-July 2002.
[42] Krivak TC, Zorn KK. Venous thromboembolism in obstetrics and gynecology. Obstet
Gynecol 2007;109:761–77.
[43] Catanzarite V, Willms D, Wong D, et al. Acute respiratory distress syndrome in
pregnancy and puerperium: causes, courses, and outcomes. Obstet Gynecol 2001;97:
760–4.
[44] Chen CY, Chen CP, Wang KG, et al. Factors implicated in outcome of pregnancies
complicated by acute respiratory failure. J Reprod Med 2003;48:641–8.
[45] Dellinger RP, Carlet JM, Masur H, et al. Surviving sepsis campaign guidelines for manage-
ment of severe sepsis and septic shock. Crit Care Med 2004;32(3):858–73.
[46] Osborn TM, Nguyen HB, Rivers EP. Emergency medicine and the surviving sepsis
campaign: an international approach to managing severe sepsis and septic shock. Ann
Emerg Med 2005;46:228–31.
[47] Otero RM, Nguyen HB, Huang DT, et al. Early goal-directed therapy in severe sepsis and
septic shock revisited concepts, controversies, and contemporary findings. Chest 2006;
130(5):1579–95.
[48] Romero R, Jimenez C, Lohda AK, et al. Amniotic fluid glucose concentration: a rapid and
simple method for the detection of intraamniotic infection in preterm labor. Am J Obstet
Gynecol 1990;163:821–30.
[49] Yoon BH, Romero R, Kim CJ, et al. Amniotic fluid interleukin-6: a sensitive test for
antenatal diagnosis of acute inflammatory lesions or preterm placenta and prediction of
perinatal morbidity. Am J Obstet Gynecol 1995;172:960–70.
[50] Gomez R, Romero R, Nien JK, et al. A short cervix in women with preterm labor and intact
membranes: a risk factor for microbial invasion of the amniotic cavity. Am J Obstet Gynecol
2005;192:678–89.
[51] Duff WP, Gibbs RS, Blanco JD, et al. Endometrial culture techniques in puerperal patients.
Obstet Gynecol 1983;61(2):217–22.
[52] Rivers E, Nguyen B, Havstad S, et al, for the Early Goal Directed Therapy Collaborative
Group. Early goal directed therapy in the treatment of severe sepsis and septic shock.
N Engl J Med 2001;345:1368–77.
[53] Beale RJ, Hollenberg SM, Vincent JL, et al. Vasopressor and inotropic support in septic
shock: an evidence-based review. Crit Care Med 2004;32(11 Suppl):S455–65.
478 GUINN et al

[54] Cotton DB, Gonik B, Dorman KF, et al. Cardiovascular alterations in severe pregnancy-
induced hypertension: relationship of central venous pressure to pulmonary capillary wedge
pressure. Am J Obstet Gynecol 1985;151:762–4.
[55] Gonik B, Cotton DB, Spillman T, et al. Peripartum colloid osmotic pressure changes: effects
of controlled fluid management. Am J Obstet Gynecol 1985;151:812–5.
[56] Jones MM, Longmire S, Cotton DB, et al. Influence of crystalloid versus colloid infusion on
peripartum colloid osmotic pressure changes. Obstet Gynecol 1986;68:659–61.
[57] Clark SL, Cotton DB. Clinical indications for pulmonary artery catheterization in the
patient with severe preeclampsia. Am J Obstet Gynecol 1988;158:453–8.
[58] Bolte AC, Dekker GA, van Eucl K, et al. Lack of agreement between central venous pressure
and pulmonary capillary wedge pressure in preeclampsia. Hypertens Pregnancy 2001;19:
261–71.
[59] ACOG technical bulletin. Invasive hemodynamic monitoring in obstetrics and gynecology,
1992: number 175.
[60] ACOG practice bulletin. Diagnosis and management of preeclampsia and eclampsia 2003:
number 33.
[61] Ngan Kee WD, Khaw KS. Vasopressors in obstetrics: what should we be using? Current
Opin Anesthesiol 2006;19(3):238–43.
[62] Coons JC, Seidl E. Cardiovascular pharmacotherapy update for the intensive care unit. Crit
Care Nurs Q 2007;30(1):44–57.
[63] Barrett LK, Singer M, Clapp LH. Vasopressin: mechanisms of action on the vasculature in
health and in septic shock. Crit Care Med 2007;35(1):33–40.
[64] Holmes CL. Vasoactive drugs in the intensive care unit. Curr Opin Crit Care 2005;11(5):
413–7.
[65] Centers for Disease Control and Prevention. Prevention of perinatal group B streptococcal
disease: a public health perspective. MMWR Recomm Rep 1996;45(RR-7).
[66] Kankuri E, Kurki T, Carlson P, et al. Incidence, treatment and outcome of peripartum
sepsis. Acta Obstet Gynecol Scand 2003;82:730–5.
[67] French LM, Smaill FM. Antibiotic regimens for endometritis after delivery. Cochrane
Database Syst Rev 2007;2.
[68] Hopkins L, Smaill F. Antibiotic regimens for management of intraamniotic infection.
Cochrane Database Syst Rev 2007;2.
[69] Philipson A. Pharmacokinetics of antibiotics in pregnancy and labour. Clin Pharmacokinet
1979;4(4):297–309.
[70] Muckart DJ, Bhagwanjee S. Ventilation and the critically ill parturient. Best Pract Res Clin
Obstet Gynaecol 2001;15:541–56.
[71] Annane D, Sebille V, Troche G, et al. A 3 level prognostic classification in septic shock based
on cortisol levels and cortisol response to corticotrophin. JAMA 2000;283:1038–45.
[72] National Institutes of Health Report on the consensus development conference on the effect
of corticosteroids for fetal maturation on perinatal outcome. Bethesda (MD): National In-
stitute of Child Heealth and Human Development; November 1994. Publication #NIH
95–3784.
[73] Abraham E, Laterre PF, Garg R, et al, for the Administration of Drotrecogin Alfa (Acti-
vated) in Early Stage Severe Sepsis Study Group. Drotrecogin alfa (activated) for adults
with severe sepsis and a low risk of death. N Engl J Med 2005;353:1332–41.
[74] Van Cromphaut S, Wilmer A, Van den Berghe G. Management of sepsis [reply to the editor].
N Engl J Med 2007;356:1179–80.
[75] Meidve L, Csitar IK, Molnar Z, et al. Recombinant human activated protein C treatment of
septic shock syndrome in a patient at 18th week of gestation: a case report. Am J Obstet
Gynecol 2005;193:864–5.
[76] Mikaszewska-Sokolewicz M, Mayzner-Zawadzka E. Use of recombinant human activated
protein C in treatment of severe sepsis in a pregnant patient with fully symptomatic ovarian
hyperstimulation syndrome. Med Sci Monit 2005;11:CS27–32.
 SEPSIS DURING PREGNANCY 479

[77] The Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as
compared with traditional tidal volumes for acute lung injury and the acute respiratory
distress syndrome. N Engl J Med 2000;342:1301–8.
[78] Wheeler AP, Bernard GR. Acute lung injury and the acute respiratory distress syndrome:
a clinical review. Lancet 2007;369(9572):1553–64.
[79] Girard TD, Bernard GR. Mechanical ventilation in ARDS: a state-of-the-art review. Chest
2007;131(3):921–9.
[80] Norwitz ER, Robinson JN, Malone FD. Fetal considerations in the critically ill gravida. In:
Dildy GA III, Belfort MA, Saade G, et al, editors. Critical care obstetrics. 4th edition. Mal-
den (MA): Blackwell Science; 2004. p. 673–95.
[81] Norwitz ER, Robinson JN, Malone FD. Fetal effects of drugs commonly used in critical
care. In: Dildy GA III, Belfort MA, Saade G, et al, editors. Critical care obstetrics. 4th
edition. Malden (MA): Blackwell Science; 2004. p. 696–713.
[82] Marik PE, Zaloga GP. Early enteral nutrition in acutely ill patients: a systematic review. Crit
Care Med 2001;29:2264–70.
 Obstet Gynecol Clin N Am
34 (2007) 481–500

Thromboembolism in Pregnancy
Victor A. Rosenberg, MD*, Charles J. Lockwood, MD
Department of Obstetrics, Gynecology and Reproductive Sciences,
Yale University School of Medicine, 333 Cedar Street,
PO Box 208063, New Haven, CT 06520, USA

Of the potential clinical emergencies an obstetrician/gynecologist will

confront, venous thromboembolism, which includes deep venous thrombo-
sis and pulmonary embolus, has been associated with the highest risk for
maternal and fetal morbidity and mortality. In the most recent Centers
for Disease Control and Prevention data available, thromboembolism was
shown to be responsible for 19.6% of pregnancy-related deaths in the
United States as compared with 17.2% for hemorrhage [1]. Venous throm-
boembolism is estimated to complicate between 0.5 and 1 in 1000 pregnan-
cies per year in the United States [2–10]. More recent evidence suggests that
the risk is evenly divided among each of the trimesters [3,11], with an even
higher risk in the postpartum period [12,13]. In addition, cesarean delivery
confers a five- to ninefold higher risk over vaginal delivery [13,14].
Essentially, every pregnant patient is at risk for a venous thromboem-
bolic event and the risk is estimated to be five- to 10-fold higher than for
the nonpregnant patient. From a teleological perspective, the adaptation
of the maternal hemostatic system to pregnancy (to prevent hemorrhage
at the time of delivery) predisposes women to an increased risk of thrombo-
embolism. Particular women seem to be at yet an even higher risk for venous
thromboembolism in pregnancy. These women include multiparous
patients, obese gravidas, women who have postpartum endometritis, and
those with a history of venous thromboembolism or underlying thrombo-
philia. It is estimated that the recurrence risk in pregnancy is between 5%
and 16% for women with a history of a venous thromboembolism [11,15]
and may be related to the presence or absence of underlying maternal
thrombophilia [16]. Others have demonstrated a 7.5% recurrence risk in
pregnancy if the first venous thromboembolism was unprovoked, related
to pregnancy, or related to use of oral contraceptives [17]. In contrast, there

* Corresponding author.
E-mail address: victor.rosenberg@yale.edu (V.A. Rosenberg).

0889-8545/07/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ogc.2007.06.006 obgyn.theclinics.com
482 ROSENBERG & LOCKWOOD

was no recurrence if the first venous thromboembolism was related to other

transient risk factors [17]. Therefore, one must consider routine thrombo-
prophylaxis in selected obstetrical patients [15,17–19].
However, despite prophylaxis, venous thromboembolism can occur in
pregnancy and clinicians must have a heightened surveillance for this poten-
tial emergency. Diagnostic tests must be readily available and there should
be no delay in initiating treatment when appropriate [20]. Treatment goals
should include preclusion of further thrombus propagation and pulmonary
embolism and prevention of recurrent venous thromboembolism and long-
term complications, including venous insufficiency, pulmonary hyperten-
sion, right-sided heart failure, and ‘‘post-thrombotic syndrome’’ [21].
This article focuses on the clinical emergency posed by deep venous
thrombosis and pulmonary embolism in pregnancy. The article begins
with a brief review of the physiologic changes that predispose pregnant
women to a thrombotic event. The article then reviews the signs and symp-
toms that should alert the clinician to the possibility of a thromboembolic
event, and then presents an algorithm outlining specific diagnostic tests
that guide the clinician to the correct diagnosis. The article then reviews rec-
ommended treatment regimens while attempting to resolve some of the con-
troversies regarding optimal anticoagulation therapy in pregnancy. The
article ends with a brief look at future directions, including innovative diag-
nostic tests that may be safer and easier to perform than current ones.

Physiology and pathophysiology of hemostasis in pregnancy

The adaptation of the maternal hemostatic system to pregnancy predis-
poses women to an increased risk of venous thromboembolism. Pregnancy
produces the components of Virchow’s triad, including an increase in vascu-
lar stasis, changes in the coagulation system, and vascular injury. Other risk
factors for thrombosis involve inherited thrombophilias, including mutations
in the factor V Leiden and prothrombin genes; deficiencies in protein C,
protein S, and antithrombin III; and acquired maternal thrombophilias,
such as the condition known as antiphospholipid antibody syndrome
[12,22,23]. It is estimated that an underlying thrombophilia is present in
at least 50% of those who develop a deep venous thrombosis or pulmonary
embolism in pregnancy [23]. Therefore, a thorough understanding of the
coagulation and fibrinolytic systems and their inhibitors with specific
relation to pregnancy is essential.

Physiology
Platelet aggregation and vasoconstriction are the initial responses to hem-
orrhage following vascular disruption and endothelial damage. By limiting
the size of the requisite plug required to obstruct blood flow through the
vascular defect, vasoconstriction limits blood flow to promote platelet
 THROMBOEMBOLISM 483

plug formation. Integrins bound to platelet membranes adhere to subendo-

thelial laminin, fibronectin, and vitronectin, and circulating von Wille-
brand’s factor mediates platelet attachment by binding to both platelet
GPIb/IX/V receptors and subendothelial collagen in damaged vessels [24].
Platelet adhesion then triggers calcium-dependent protein kinase C activa-
tion, which induces thromboxane A2 (TXA2) synthesis and platelet granule
release. The a-granules contain von Willebrand’s factor and various clotting
factors while dense-granules contain adenosine diphosphate and serotonin,
which together with thromboxane A2 (TXA2), exacerbate vasoconstriction
and platelet activation. The latter process activates platelet GPIIB/IIIa
receptors to promote aggregation by forming inter-platelet fibrinogen, fibro-
nectin, and vitronectin bridges [25]. Epinephrine, arachidonic acid, and
platelet activating factor can also activate platelets.
Tissue factor, a glycoprotein bound to cell membranes, is the primary ini-
tiator of hemostasis and the coagulation cascade. Tissue factor is expressed
constitutively by epithelial, stromal, and perivascular cells throughout the
body. Tissue factor is also expressed, particularly in pregnancy, in endome-
trial stromal cells and uterine decidua [26,27]. Clotting is initiated by the
binding of tissue factor to factor VII after vascular injury and can be exter-
nally activated by thrombin, factor IXa, factor Xa, or factor XIIa [26,27].
The coagulation cascade is initiated and, ultimately, thrombin cleaves fibrin-
ogen to fibrin monomers, which self-polymerize and are cross-linked via
thrombin-activated factor XIIIa.
The counterpoise of the hemostatic system is the anticoagulant system.
The tissue factor pathway inhibitor is the first agent in this system and
acts on the factor-Xa–tissue factor–factor-VIIa complex to inhibit tissue-
factor–mediated clotting [28]. However, factor XIa can bypass this block
and sustain clotting for some time. As a result, additional endogenous anti-
coagulant molecules are required to avoid thrombosis, including activated
protein C, protein S, and protein Z.
Fibrinolysis is initiated by tissue-type plasminogen activator (tPA), which
cleaves plasminogen to generate plasmin. Plasmin, in turn, cleaves fibrin
into fibrin degradation products (FDPs). These FDPs can also inhibit
thrombin action, a favorable result when limited, but when generated in
excess can contribute to disseminated intravascular coagulation. Inhibitors
of fibrinolysis include a-2-plasmin inhibitor and type-1 and -2 plasminogen
activator inhibitors (PAI-1 and -2), which inactivate tPA. The endothelium
and uterine decidua are primary sources of PAI-1, while the placenta pro-
duces PAI-2 [29]. The thrombin-activatable fibrinolysis inhibitor modifies
fibrin to render it resistant to inactivation by plasmin [30].

Pathophysiology
Changes in decidual and systemic hemostatic systems occur in pregnancy,
likely to meet the hemorrhagic challenges poised by implantation,
484 ROSENBERG & LOCKWOOD

placentation, and the third stage of labor. Decidual tissue factor and PAI-1
expression increase in response to progesterone, providing a potent local
system of hemostasis to prevent hemorrhage. In addition, levels of placental
PAI-2, circulating levels of fibrinogen, and levels of factors VII, VIII, IX, X,
and XII and of von Willebrand’s factor increase considerably in gestation
[29–32]. While these mechanisms serve to generally prevent puerperal hem-
orrhage following significant uterine vascular trauma at the time of delivery,
they predispose to thrombosis, a tendency aggravated by maternal
thrombophilias.
Inherited thrombophilias refer to a genetic tendency to venous thrombo-
embolism. Disorders include the factor V Leiden and prothrombin gene
G20210A mutations, antithrombin deficiency, and protein C and S defi-
ciencies. Acquired thrombophilias include the antiphospholipid antibody
syndrome, which is characterized by the presence of antibodies directed
against plasma proteins bound to anionic phospholipids.
The antiphospholipid antibody syndrome is responsible for 14% of ve-
nous thromboembolism in pregnancy [33,34]. The lifetime prevalence of
arterial or venous thrombosis is approximately 30%, with an event rate of
1% per year [35]. The risks of venous thromboembolism are highly depen-
dent upon the presence of other predisposing factors, including pregnancy,
estrogen exposure, surgery, and infection. There is a 5% risk of a thrombotic
event in pregnancy even with prophylaxis [36].
The inherited thrombophilias are a heterogeneous group of genetic disor-
ders often associated with a personal or family history of venous thrombo-
embolism. Such a history is an important modifier of projected risk.
Thrombophilias are divided into high-risk thrombophilias and low-risk
thrombophilias based on the overall risk of venous thromboembolism. Be-
cause of the association between thrombophilias and recurrent venous
thromboembolism in pregnancy, the authors routinely obtain a comprehen-
sive thrombophilia evaluation on patients diagnosed with venous thrombo-
embolism in pregnancy. However, because functional levels of protein C,
protein S, and antithrombin are altered in pregnancy, abnormally low levels
should be confirmed 6 weeks postpartum before a diagnosis of a deficiency is
made.

Diagnosis of deep vein thrombosis

Clinicians must have a high baseline index for suspicion of deep venous
thrombosis in pregnancy because many of the common clinical signs and
symptoms, such as lower extremity edema, are also common findings in nor-
mal pregnancy. A timely diagnosis of deep venous thrombosis is crucial be-
cause up to 24% of patients with untreated deep venous thrombosis will
develop a pulmonary embolism [37]. A life-threatening pulmonary embo-
lism usually originates from a clot in the deep veins of the pelvis and legs,
including the internal iliac, femoral, and popliteal veins [7].
 THROMBOEMBOLISM 485

Common clinical features of deep venous thrombosis include lower ex-

tremity edema, pain, difficulty with ambulation, warmth, and erythema.
However, the diagnostic sensitivity of these clinical signs and symptoms
is at best 50% and the diagnosis of deep venous thrombosis is confirmed
in less than a third of patients with these complaints [38,39]. Therefore, pa-
tients who present with any of these complaints warrant a full diagnostic
workup. Diagnostic tests for evaluation of suspected deep venous throm-
bosis include D-dimer assays, venous color Doppler ultrasound, magnetic
resonance venography, CT, and, less commonly, contrast venography
[40,41].

D-dimer assays
D-dimer assay testing may be used as a screening test and/or in combina-
tion with venous ultrasound to facilitate diagnosis and prediction of
a thromboembolic event. D-dimer is a product of the degradation of fibrin
by plasmin. Therefore, elevated levels indicate increased thrombin activity
and increased fibrinolysis following fibrin formation [42]. The assay employs
monoclonal antibodies to detect D-dimer fragments. Commercial assays
available include at least three accurate and reliable products: two rapid en-
zyme lined immunosorbent assays and a rapid whole-blood assay.
Though quite reliable in the exclusion of deep venous thrombosis in the
nonpregnant patient [43,44], the value of D-dimer testing in pregnancy is
somewhat controversial because D-dimer levels increase with gestational
age and, in the postpartum period, even in the absence of venous thrombo-
embolism [45–48]. This makes it difficult to assign a ‘‘normal’’ cutoff. Most
studies report a sensitivity ranging from 85% to 97% but a specificity of
only 35% to 45% [21,49]. In addition, there appears to be a wide variation
in D-dimer assay results depending on the specific test used. These factors
have led some investigators to conclude that the literature does not support
the general use of D-dimer assays as a stand-alone test for the diagnosis of
deep venous thrombosis in pregnancy [50]. However, others argue that
D-dimer testing is likely to have a higher negative predictive value in preg-
nancy and therefore it has a role in the initial triage of patients with sus-
pected deep venous thrombosis. In patients with a negative D-dimer assay
and a low clinical probability of deep venous thrombosis, further testing
may be unnecessary (Fig. 1). Several elaborate scoring systems (not vali-
dated in pregnancy) have been proposed to help classify patients as either
low or high risk for deep venous thrombosis [51,52]. Another approach is
to categorize patients as low risk if there is another reasonable clinical expla-
nation for their symptoms and there are no major risk factors, such as re-
cent major abdominal surgery, late pregnancy and postpartum, varicose
veins, malignancy, and reduced mobility [53]. In addition, there may be a
role for D-dimer testing to identify women at high risk for recurrent venous
thrombosis [42].
486 ROSENBERG & LOCKWOOD

Clinical Suspicion of DVT

D-dimer AND Venous Ultrasound

Negative VUS
Negative D-dimer Abnormal VUS
Negative VUS

low clinical high clinical

suspicion suspicion Therapy
End of
workup
Repeat VUS in 1 week MR Venogram

Abnormal Normal Abnormal

No Therapy

Fig. 1. Diagnostic algorithm for deep venous thrombosis. DVT, deep venous thrombosis;
MR, magnetic resonance; VUS, venous ultrasound.

Venous ultrasound
Compression ultrasound aided by color flow Doppler imaging involves
the use of firm pressure applied to the ultrasound transducer to detect an
intraluminal filling defect of the major venous systems of the legs, including
the common femoral, superficial femoral, greater saphenous, and popliteal
veins. Noncompressibility of the venous lumen is the most accurate ultra-
sound criteria for thrombosis [38]. Venous ultrasound to detect deep venous
thrombosis has been well studied in pregnancy [54]. It is noninvasive, easy to
perform, and can be repeated if necessary without any restrictions. Sensitiv-
ity and specificity of venous ultrasound in the detection of proximal deep
venous thrombosis is estimated at 95% and 96%, respectively [41,55]. There
is a slightly lower sensitivity (75%–90%) in detecting more distal thrombosis
in the leg [41,56].

Other modalities
It is estimated that in up to 3% of patients, venous ultrasound is not tech-
nically possible [57], and in some patients, despite negative ultrasound re-
sults, clinical suspicion remains high. Magnetic resonance venography and
CT of the pelvis and lower extremities may be a viable alternative in these
patients. Magnetic resonance direct thrombus imaging was shown in
a blinded study of nonpregnant patients to have a sensitivity of 94% to
96% and specificity of 90% to 92% for the detection of deep venous throm-
bosis with similar results for calf deep venous thrombosis. MRI was well tol-
erated and interpretation was highly reproducible [58–60]. The reported
 THROMBOEMBOLISM 487

experience with MRI as a diagnostic modality for pregnant patients with

deep venous thrombosis is extremely limited [61] and there is only limited
safety data [62]. Thus, while magnetic resonance venography is promising,
additional studies are needed before it can be routinely recommended. In
the nonpregnant patient, CT of the pelvis and lower extremities to diagnose
deep venous thrombosis is a useful modality with a reported sensitivity and
specificity similar to ultrasound [63–65]. However, there is no reported expe-
rience with this modality in pregnancy and the natural preference during
gestation is to test with ultrasound, which does not involve a risk of radia-
tion exposure to the fetus.
Contrast venography involves the injection of radio-opaque dye into the
vein below the site of the suspected thrombus. Imaging is then used to iden-
tify a filling defect [66]. However, the relative ease and noninvasive nature of
compression ultrasound has made this more invasive test somewhat obsolete
[21].

Workup of patients with suspected deep venous thrombosis

A diagnostic algorithm is presented in Fig. 1 to guide the clinician in the
workup of a pregnant patient with a suspected deep venous thrombosis.

Diagnosis of pulmonary embolus

Timely diagnosis of pulmonary embolus in pregnancy is critical because
of the potential for a catastrophic maternal and fetal outcome if overlooked.
If the clinical suspicion is high, consideration should be given to empiric an-
ticoagulation until the workup is completed [7]. Likewise, a precise diagnosis
is vital to prevent unnecessary treatment of pulmonary embolism because
treatment is associated with side effects for both the mother and fetus. Ac-
curate imaging is essential, but fetal radiation exposure during diagnostic
procedures often provokes unfounded anxiety for the clinicians involved
[67].
An array of clinical, biochemical, and radiological tests is available to aid
in the investigation of pulmonary embolism in pregnancy. Because, accord-
ing to estimates, 70% of patients with proven pulmonary embolism have
a proximal deep venous thrombosis, the basic workup begins with compres-
sion venous ultrasound if there are any signs or symptoms of thrombosis of
the lower extremities. If a deep venous thrombosis is confirmed, then pulmo-
nary embolism can be assumed, and treatment can be initiated without fur-
ther workup [7,68]. If venous ultrasound is nondiagnostic or not performed,
traditional teaching (based on older research) focused on the ventilation–
perfusion (VQ) scan as the primary modality to diagnose pulmonary embo-
lism in pregnancy. However, more recent studies support CT pulmonary
angiography (CTPA) as the favored diagnostic tool. In fact, the most recent
guidelines from the British Thoracic Society recommend CTPA as the initial
488 ROSENBERG & LOCKWOOD

lung imaging modality in pregnancy for nonmassive pulonary embolism

[53].

Clinical signs and symptoms

Traditional clinical hallmarks of pulmonary embolism, including dysp-
nea, tachycardia, tachypnea, pleuritic chest pain, and syncope or near-syn-
cope are present in up to 90% of patients found to have a pulmonary
embolus. However, these clinical signs and symptoms lack specificity and
generate a broad differential diagnosis [69,70]. Other more objective mea-
sures, such as low oxygen saturation on pulse oximetry, abnormal arterial
blood gas (ABG), abnormal chest radiograph, abnormal EKG, and abnor-
mal echocardiogram, have also been proposed.
Low oxygen saturation on pulse oximetry or ABG has a limited role in
the assessment of pregnant patients with suspected pulmonary embolism.
These tests are useful in elderly populations, but lack diagnostic accuracy
in younger patients, including pregnant patients [71]. Studies have shown
that up to 20% of patients with a documented pulmonary embolism, had
PO2 measurements on ABG greater than 85 mm Hg [70]. The alveolar-
arterial gradient may be a more sensitive indicator of pulmonary embolism
in nonpregnant patients with 86% of patients with documented pulmonary
embolism having an alveolar-arterial gradient greater than 20 [70]. How-
ever, 58% of pregnant women with documented pulmonary embolism had
a normal alveolar-arterial gradient [72].
Abnormalities on EKG, including the classic S1-Q3-T3 changes, may be
present in 70% to 90% of patients with pulmonary embolism but are con-
sidered nonspecific findings [73,74]. Other EKG findings, such as new-onset
atrial fibrillation and right bundle branch block or right axis deviation, are
typically later findings after pulmonary embolism and are more suggestive
of significant cardiopulmonary compromise. Absence of abnormal EKG
findings should not reassure a clinician who has a reasonable suspicion of
pulmonary embolism [75].

Initial imaging modalities

The chest radiograph may be abnormal in up to 85% of affected patients.
Common findings include effusions, infiltrates, and atelectasis. The ‘‘classic’’
wedge-shaped infiltrate (Hampton’s hump) or decreased vascularity (West-
ermark’s sign) are, in fact, rare findings [70,76]. Chest radiograph may be
helpful in excluding other competing diagnoses, including pneumonia, pul-
monary edema, pleural effusions, and pneumothorax.
Echocardiographic abnormalities of right ventricular size and function
are present in a significant number of patients with acute large pulmonary
embolism. Typical findings include a dilated and hypokinetic right ventricle
and tricuspid regurgitation. Transesophageal imaging may enhance diag-
nostic accuracy [77–79]. A recent observation is that the release of cardiac
 THROMBOEMBOLISM 489

troponins can detect acute right heart strain from right ventricular muscle
damage in major pulmonary embolism. However, the role of cardiac tropo-
nins in decision-making is limited and they are of no diagnostic value in
nonmassive pulmonary embolism [80–83].

D-dimer
As with the evaluation of patients with suspected deep venous thrombo-
sis, D-dimer is a sensitive, but not specific test for pulmonary embolism. In
nonpregnant patients, a negative D-dimer has a negative predictive value of
95%, but only a 25% specificity [76]. However, as mentioned previously in
the discussion regarding the diagnosis of deep venous thrombosis, abnormal
cutoffs are difficult to assign in pregnancy because D-dimer levels increase
with gestational age, and in the postpartum period, even in the absence of
venous thromboembolism [45–48]. A negative D-dimer probably has
a role in the exclusion of pulmonary embolism in patients with a low clinical
suspicion (see description of risk assessment above), but the assay should
not be performed in those with high clinical probability of pulmonary
embolism [53].

Pulmonary angiogram
For many years the ‘‘gold standard’’ in diagnosing an acute pulmonary
embolism was pulmonary arteriography. Sensitivity approaches 100%,
though the ability to detect segmental and subsegmental lesions is consid-
ered diminished. The procedure involves catheterization of the pulmonary
artery via a femoral or internal jugular approach and noting a filling defect
via radiograph or fluoroscopy. This procedure carries significant risk,
including 0.5% mortality risk and 3% complication rate, primarily due to
the risks of contrast injection and catheter placement. Complications
include groin hematoma, cardiac perforation, renal failure, and respiratory
failure [69,84–86]. This apparent potential for morbidity led to an intensive
effort over the past several years to identify a diagnostic modality that
would be safer and easier to perform without sacrificing sensitivity.

Ventilation–perfusion scan
VQ imaging is a well-established diagnostic modality in the workup of
a suspected pulmonary embolus in pregnancy and for many years it was
the most frequently employed test in this subgroup of patients [67]. The
test involves comparative imaging of the pulmonary vascular beds and air-
spaces using radiolabeled markers injected intravenously and as inhaled
gases. Patients are then categorized into different diagnostic probability cat-
egories, including low, intermediate, high, normal, and indeterminate [38].
Any outcome other than high probability or normal requires further testing.
Radiation dose can be minimized in pregnancy by using a half-dose perfu-
sion scan and only using ventilation imaging if the perfusion scan is
490 ROSENBERG & LOCKWOOD

abnormal [87]. Unfortunately, VQ scans are time-consuming and the sensi-

tivity varies widely depending on the degree of clinical suspicion [7].
The Prospective Investigation of Pulmonary Embolism Diagnosis
(PIOPED) study looked at the diagnostic accuracy of VQ scans in nearly
1000 nonpregnant patients with suspected pulmonary embolism. High-
probability VQ scans correlated with pulmonary embolism in 87.2% of
cases. However, only 41% of patients with pulmonary embolism had
high-probability scans, yielding a sensitivity of 41% and a specificity of
97% [74]. In addition, it is estimated that over 10% of patients with
a low-probability scan were found to have a pulmonary embolus on subse-
quent imaging. In the largest published study of VQ scans in pregnancy [68],
fewer than 5% of pregnant patients had high-probability scans, almost 25%
had indeterminate scans that required further evaluation, and more than
70% had normal scans. This is quite different than in the nonpregnant pop-
ulation where 40% to 70% of scans are nondiagnostic [67].

CT pulmonary angiography
CTPA employs intravenous contrast injection to highlight the pulmonary
vasculature while using the latest generation of fast multislice scanners
[53,76]. Much of the reluctance to use CTPA in pregnancy revolves around
potential radiation exposure to the fetus. In fact, the authors’ radiology col-
leagues often cite unfounded concerns regarding radiation exposure as a rea-
son to refuse to perform CT scan and to promote VQ as the primary
imaging modality.
In a recent study, Winer-Muram and colleagues [88] calculated the mean
fetal radiation dose from helical chest CT by using maternal–fetal geome-
tries obtained from healthy pregnant women and comparing the calculated
CT doses with the doses reported with VQ scan. They found that the aver-
age fetal radiation dose is higher with VQ scan than with CT scan in all tri-
mesters of pregnancy. As a corollary, in a survey of health professionals to
determine their knowledge of dositometry in the workup of pulmonary em-
bolism, only 58% appreciated correctly that a VQ scan delivers a higher fe-
tal dose of radiation than that delivered by CT pulmonary angiography [89].
Interestingly, the survey population included medical trainees, radiologists,
nuclear physicians, medical physicists, and pulmonologists. Lastly, in a sur-
vey of the PIOPED II investigators, only 31% recommended CT as the pri-
mary imaging test [90], but 75% of respondents in a conflicting study use CT
angiography in pregnant patients [91].
CTPA is a well-validated diagnostic modality with a sensitivity and spec-
ificity between 94% and 100%. In a systematic review of available studies,
the negative likelihood ratio of pulmonary embolism (pulmonary embolism
confirmed by additional imaging) after a negative or inconclusive CT was
0.07; and the negative predictive value was 99.1%. The investigators con-
clude that the clinical validity of CTPA to diagnose pulmonary embolism
 THROMBOEMBOLISM 491

is similar to the clinical validity of pulmonary angiography [92], and missed

diagnoses are rare [93]. Others have suggested that in patients with a low
clinical suspicion of pulmonary embolism, CTPA has a greater discrimina-
tory power than VQ scanning, but in patients with a high clinical suspicion,
CTPA and VQ scan perform similarly [94].
CT is not only safe during pregnancy but also accurate for the diagnosis
of pulmonary embolism in main, lobar, and segmental pulmonary arteries
[88]. The latest CT technology and techniques are more accurate than VQ
technology in identifying peripheral thrombus [53]. CT was also found to
be the most cost-effective modality in diagnosing pulmonary embolism in
pregnancy with a cost of $17,208 per life saved, compared with $35,906
per life saved for a VQ scan [95].
Given the safety data presented above and the relative ease in obtaining
a CT versus a VQ scan, the authors prefer CTPA as the initial diagnostic ap-
proach to suspected pulmonary embolism in pregnancy. CTPA is easier to per-
form, is readily available even in off hours, and rarely requires any follow-up
imaging. In fact, many radiology departments have sufficient confidence in the
sensitivity of their CT imaging to also forgo formal contrast pulmonary angi-
ography. Another advantage to CT over VQ scan is the ability to detect other
disorders that may be responsible for the patient’s symptoms, including pul-
monary edema, pneumonia (consolidation), and pleural effusions [53].

Magnetic resonance angiography

Magnetic resonance angiography (MRA) uses gadolinium injection dur-
ing magnetic resonance scanning to visualize the pulmonary vasculature.
Newer generation MRI with faster imaging acquisition times have enabled
the use of this technique. While initial studies were promising with report-
edly high sensitivity and specificity [96], in a prospective study of 141
patients with suspected pulmonary embolism, the overall sensitivity was
only 77% when compared with pulmonary angiography [97]. Still, others
have proposed a combination of chest MRI and lower extremity magnetic
resonance venogram as a way to detect 13% more cases of thromboembo-
lism [98]. Unfortunately, no reported studies have examined the use of mag-
netic resonance to diagnose pulmonary embolism in pregnancy.

Workup of patients with suspected pulmonary embolism

A diagnostic algorithm is presented in Fig. 2 to guide the clinician in the
workup of a pregnant patient with a suspected pulmonary embolism.

Treatment of venous thromboembolism in pregnancy

Whether manifested as a deep venous thrombosis or pulmonary embo-
lism, acute venous thromboembolism in pregnancy requires immediate med-
ical therapy. Initial steps in the management of pulmonary embolism
492 ROSENBERG & LOCKWOOD

Suspected Pulmonary Embolus

Clinical Suspicion

Low High

D-dimer D-dimer CT pulmonary angiography (CTPA)

negative positive (or VQ scan if CTPA not available)

Venous Venous
CTPA CTPA
ultrasound ultrasound
positive negative
negative positive

No Therapy No Therapy
Treatment for PE

Fig. 2. Diagnostic algorithm for pulmonary embolism. PE, pulmonary embolism.

include oxygen support, blood pressure stabilization, and an assessment of

the patient’s cardiovascular and respiratory status [7,53]. Consultation with
the intensive care unit service may be appropriate and transfer to the inten-
sive care unit should be considered, depending on nursing and physician
resources in the unit where the patient is located. Close monitoring for
evidence of right-sided cardiac failure in cases of massive pulmonary emboli
is warranted [53].
The mainstay of medical treatment is anticoagulation. While conven-
tional treatment recommendations called for unfractionated heparin as
the suggested therapy in pregnancy, low molecular weight heparin
(LMWH) has emerged as the superior alternative based on more recent
studies. Warfarin is seldom a treatment for acute venous thromboembolism
in pregnancy given the drug’s risk of teratogenicity [20,53], though this risk
is greatest between the sixth and 12th weeks of pregnancy. There is also
a risk for fetal hemorrhage with warfarin use.

Unfractionated heparin
Unfractionated heparin promotes anticoagulation by inhibiting platelet
aggregation and by enhancing and increasing antithrombin and factor Xa
inhibitor activity [99]. The initial bolus dose and maintenance dosing are cal-
culated and titrated to achieve an activated partial thromboplastin time
(aPTT) at 1.5- to two-times normal [18,99,100]. Standard nomograms are
readily available from hospital pharmacies. Once therapeutic dosing is
achieved, the aPTT must be periodically monitored to confirm adequate
dosing. The potential side effects from unfractionated heparin include hem-
orrhage, osteoporosis, and thrombocytopenia.
 THROMBOEMBOLISM 493

Osteoporosis, or clinically significant bone loss, has been traditionally

quoted as an adverse effect of long-term anticoagulation with heparin dur-
ing pregnancy. Dahlman [101] reported that the incidence of vertebral frac-
tures in 184 women treated with unfractionated heparin during pregnancy
was 2.2%. Additionally, the mean duration of heparin prophylaxis in the
women who had osteoporosis and spinal fracture was only 17 weeks (range:
7–27 weeks).
Heparin-induced thrombocytopenia (HIT) occurs in approximately 3%
of patients receiving unfractionated heparin. Type I, or the immediate
form, occurs within days of exposure and is typically self-limited. Type II,
the immunoglobulin type, is rare and usually occurs 5 to 14 days after the
initiation of therapy [102]. The authors therefore typically monitor platelet
counts as follows: complete blood cell count on day 3, then on each of days
7 through 10, and then monthly after starting anticoagulation. A 50% de-
cline in platelet count from the pretreatment level suggests a type II reaction
and is an indication to promptly discontinue the heparin. Consultation with
hematology would then be recommended for acceptable alternative
therapies.
The cumbersome dosing requirements, the need for frequent aPTT mon-
itoring, the need for long-term hospitalization, and concerns regarding side
effects, including osteoporosis, osteopenia, and HIT, have led many author-
ities to recommend LMWH as the primary anticoagulation in pregnancy
(see below) [20,53]. However, in certain rare circumstances, the authors pre-
fer unfractionated heparin over LMWH. These include circumstances in-
volving patients who are hemodynamically unstable due to massive
pulmonary embolism [53], patients at significant risk for bleeding (eg, imme-
diately postoperation patients or patients with antepartum placental abnor-
malities), and patients close to term who may require regional anesthesia
and/or cesarean delivery. These patients are potentially better served by un-
fractionated heparin because of its shorter half-life and ease of reversibility
with such agents as protamine sulfate. Protamine can be given as an intra-
venous infusion and dosing is based on residual circulating heparin.

Low molecular weight heparin

LMWHs, including enoxaparin and dalteparin, have established safety
profiles in pregnancy and are emerging as the anticoagulant of choice for
many indications, including acute venous thromboembolism [20,53,103–
107]. LMWHs have potential advantages over unfractionated heparin
because they have a lower incidence of HIT [102], a more predictable
dose response, and a lower incidence of bone loss related to use. Shefras
and colleagues [108] performed serial bone mineral density measurements
in women treated with LMWH during pregnancy. Mean bone loss was
5.6% and 5.1%, depending on the dose, but this was not statistically differ-
ent from the mean bone loss in the control group of pregnant patients who
494 ROSENBERG & LOCKWOOD

were not exposed to LMWH (3.1%). In a randomized open study of unfrac-

tionated heparin versus LMWH in pregnancy, Pettila [109] showed that
bone mineral density, as measured by serial dual energy x-ray absorptiom-
etry scan up to 3 years postpartum, was significantly lower in the unfractio-
nated heparin group versus the LMWH group. However, there was no
difference between the LMWH group and healthy controls that were not
exposed to heparin therapy.
Many studies have examined the efficacy of LMWH versus unfractio-
nated heparin. In a prospective observational study, Rodie and colleagues
[110] demonstrated the safety of enoxaparin for the treatment of acute
venous thromboembolism in pregnancy. Few patients needed modification
of the initial dose to maintain a therapeutic anti-Xa activity. Jacobson
and associates [104] found similar results, but suggested that approximately
10% to 20% higher doses of LMWH may be needed in pregnancy. In
a meta-analysis of 11 randomized trials comparing LMWH to unfractio-
nated heparin [103], LMWH reduced mortality rates over 3 to 6 months
of patient follow-up (odds ratio: 0.71), had favorable results with regard
to major bleeding complications, and had equivalent efficacy to unfractio-
nated heparin in preventing thromboembolic recurrences. Other reviews
[106,111] have had similar conclusions. In a decision model, Gould and col-
leagues reported (in nonpregnant patients) LMWH to be more cost-effective
than unfractionated heparin in the treatment of acute deep venous thrombo-
sis [112].
One controversial area with regard to the use of LMWH in pregnancy is
the necessity to monitor therapeutic levels (ie, anti-Xa levels). In nonpreg-
nant patients, monitoring is generally not required because anticoagulant ef-
fects are predictable [111]. However, in pregnancy, the increased glomerular
filtration rate in the kidney may explain the apparent need for increased dos-
ing to maintain therapeutic levels reported in the literature [113]. In addi-
tion, there is a greater variability with regard to binding, distribution, and
metabolism of LMWH in pregnancy.
The authors’ preferred treatment for acute venous thromboembolism in
pregnancy is LMWH. Though some have proposed outpatient therapy as
a viable option outside of pregnancy [114], initial hospitalization is recom-
mended in a gravid patient. The authors start with enoxaparin at a dose
of 1 mg/kg subcutaneously given twice a day. The authors typically follow
anti-Xa levels monthly and adjust the LMWH dosing to achieve a peak anti-
Xa level of 0.6 to 1.0 U/mL (3–4 hours after injection). The authors also
prefer twice-daily over once-daily dosing. It is recommended to continue
therapeutic anticoagulation for at least 20 weeks. If this period expires be-
fore the end of pregnancy or the postpartum period, prophylactic anticoa-
gulation should be initiated unless the patient has another indication for
the continuation of therapeutic anticoagulation, such as a high-risk throm-
bophilia. Prophylactic anticoagulation should be continued for up to 6 weeks
postpartum.
 THROMBOEMBOLISM 495

Though the risk of HIT is lower with LMWH, the authors still monitor
platelet counts by checking a complete blood cell count on day 3, once between
days 7 through 10, and then monthly after starting anticoagulation. Finally,
the authors typically convert patients to unfractionated heparin at 36 weeks
in anticipation of labor and possible regional anesthesia as regional anesthesia
is contraindicated within 18 to 24 hours of therapeutic LMWH administra-
tion. Patients should be advised to hold their anticoagulation at the onset of
labor. Heparin should be discontinued 24 hours before induction of labor
or planned cesarean section. If spontaneous labor occurs in women receiving
unfractionated heparin, careful monitoring of the aPTT is required [20].
In the postpartum period, prophylactic anticoagulation should be re-
started 3 to 6 hours after vaginal delivery and 6 to 8 hours after uncompli-
cated cesarean delivery. The authors either continue enoxaparin (40 mg
daily) or transition to oral anticoagulant therapy with warfarin. Warfarin
should be dosed to achieve an international normalized ratio of 2.0 to 3.0
and enoxaparin must be continued for 5 days and until the international
normalized ratio is therapeutic for 2 days. Because of the need with warfarin
therapy for frequent monitoring of the international normalized ratio, most
patients prefer to simply continue the enoxaparin.

Summary
Venous thromboembolism is one of the most critical clinical emergencies
an obstetrician/gynecologist will confront. An understanding of the physiol-
ogy and pathophysiology of hemostasis and thrombosis in pregnancy is es-
sential and allows the clinician to predict which patients are at highest risk.
Prompt recognition and diagnosis of venous thromboembolism with con-
temporary imaging modalities allow for the timely initiation of appropriate
therapy to prevent further maternal and fetal morbidity.

References
[1] Chang J, Elam-Evans L, Berg C, et al. Pregnancy-related mortality surveillancedUnited
States, 1991–1999. MMWR Surveill Summ 2003;52:1–8.
[2] Greer I. Thrombosis in pregnancy: maternal and fetal issues. Lancet 1999;353(9160):
1258–65.
[3] Gherman RB, Goodwin TM, Leung B, et al. Incidence, clinical characteristics, and timing
of objectively diagnosed venous thromboembolism during pregnancy. Obstet Gynecol
1999;94(5 Pt 1):730–4.
[4] Ginsberg J, Brill-Edwards P, Burrows R, et al. Venous thrombosis during pregnancy: leg
and trimester of presentation. Thromb Haemost 1992;67:519–20.
[5] James K, Lohr J, Deshmukh R, et al. Venous thrombotic complications of pregnancy.
Cardiovasc Surg 1996;4:777–82.
[6] Kierkegaard A. Incidence and diagnosis of deep vein thrombosis associated with preg-
nancy. Acta Obstet Gynecol Scand 1983;62:239–43.
[7] Martin S, Foley M. Intensive care in obstetrics: an evidence-based review. Am J Obstet
Gynecol 2006;195(3):673–89.
496 ROSENBERG & LOCKWOOD

[8] Rutherford S, Montoro M, McGehee W, et al. Thromboembolic disease associated with

pregnancy: an 11-year review (SPO Abstract). Obstet Gynecol 1991;164:286.
[9] Simpson E, Lawrenson R, Nightingale A, et al. Venous thromboembolism in pregnancy
and the puerperium: incidence and additional risk factors from a London perinatal data-
base. BJOG 2001;108:56–60.
[10] Stein P, Hull R, Patel K, et al. D-dimer for the exclusion of acute venous thrombosis and
pulmonary embolism: a systematic review. Ann Intern Med 2004;140:589–602.
[11] Blanco-Molina A, Trujillo-Santos J, Criado J, et al. Venous thromboembolism during
pregnancy or postpartum: findings from the RIETE registry. Thromb Haemost 2007;
97(2):186–90.
[12] McColl M, Ramsay J, Tait R, et al. Risk factors for pregnancy associated venous throm-
boembolism. Thromb Haemost 1997;78:1183–8.
[13] Macklon N, Greer I. Venous thromboembolic disease in obstetrics and gynecology: the
Scottish experience. Scott Med J 1996;41:83–6.
[14] Lindqvist P, Dahlback B, Marsal K. Thrombotic risk during pregnancy: a population
study. Obstet Gynecol 1999;94(4):595–9.
[15] Tengborn L, Bergqvist D, Matzsch T, et al. Recurrent thromboembolism in pregnancy and
puerperium. Is there a need for thromboprophylaxis? Am J Obstet Gynecol 1989;160(1):90–4.
[16] Brill-Edwards P, Ginsberg JS, Gent M, et al. Safety of withholding heparin in pregnant
women with a history of venous thromboembolism. Recurrence of Clot in This Pregnancy
Study Group. N Engl J Med 2000;343(20):1439–44.
[17] De Stefano V, Martinelli I, Rossi E, et al. The risk of recurrent venous thromboembolism in
pregnancy and puerperium without antithrombotic prophylaxis. Br J Haematol 2006;
135(3):386–91.
[18] Barbour LA, Smith JM, Marlar RA. Heparin levels to guide thromboembolism prophy-
laxis during pregnancy. Am J Obstet Gynecol 1995;173(6):1869–73.
[19] Quiñones J, James D, Stamilio D, et al. Thromboprophylaxis after cesarean delivery: a
decision analysis. Obstet Gynecol 2005;106(4):733–40.
[20] Bates S, Greer I, Hirsh J, et al. Use of antithrombotic agents during pregnancy: the seventh
ACCP conference on antithrombotic and thrombolytic therapy. Chest 2004;126:627S–44S.
[21] Krivak T, Zorn K. Venous thromboembolism in obstetrics and gynecology. Obstet Gyne-
col 2007;109(3):761–77.
[22] Gerhardt A, Scharf R, Beckmann M, et al. Prothrombin and factor V mutations in women
with a history of thrombosis during pregnancy and the puerperium. N Engl J Med 2000;342:
374–80.
[23] Grandone E, Margaglione M, Colaizzo D, et al. Genetic susceptibility to pregnancy-related
venous thromboembolism: roles of factor V Leiden, prothrombin G20210A, and methyle-
netetrahydrofolate reductase C677T mutations. Am J Obstet Gynecol 1998;179(5):1324–8.
[24] Ruggeri Z, Dent J, Saldivar E. Contribution of distinct adhesive interactions to platelet ag-
gregation in flowing blood. Blood 1999;94:172–8.
[25] Pytela R, Pierschbacher M, Ginsberg M, et al. Platelet membrane glycoprotein IIb/IIIa:
member of a family of Arg-Gly-Asp-specific adhesion receptors. Science 1986;231:1559–62.
[26] Nemerson Y. Tissue factor and hemostasis. Blood 1988;71:1–8.
[27] Preissner K, de Boer H, Pannekoek H, et al. Thrombin regulation by physiological inhib-
itors: the role of vitronectin. Semin Thromb Hemost 1996;165:1335–41.
[28] Broze G. The rediscovery and isolation of TFPI. J Thromb Haemost 2003;1:1671–5.
[29] Schatz F, Lockwood C. Progestin regulation of plasminogen activator inhibitor type-1 in
primary cultures of endometrial stromal and decidual cells. J Clin Endocrinol Metab
1993;77:621–5.
[30] Lockwood C, Krikun G, Schatz F. The decidua regulates hemostasis in the human endo-
metrium. Semin Reprod Endocrinol 1999;17:45–51.
[31] Bremme K. Haemostatic changes in pregnancy. Baillieres Best Pract Res Clin Haematol
2003;16:153–68.
 THROMBOEMBOLISM 497

[32] Hellgren M, Blomback M. Studies on blood coagulation and fibrinolysis in pregnancy, dur-
ing delivery and in the puerperium. Gynecol Obstet Invest 1981;12:141–54.
[33] Ginsberg J, Wells P, Brill-Edwards P, et al. Antiphospholipid antibodies and venous throm-
boembolism. Blood 1995;86(10):3685–91.
[34] Girling J, de Swiet M. Inherited thrombophilia and pregnancy. Curr Opin Obstet Gynecol
1998;10:135–44.
[35] Garcia-Fuster M, Fernandez C, Forner M, et al. Risk factors and clinical characteristics of
thromboembolic venous disease in young patients: a prospective study. Med Clin (Barc)
2004;123:217–9.
[36] Branch D, Silver R, Blackwell J, et al. Outcome of treated pregnancies in women with anti-
phospholipid syndrome: an update of the Utah experience. Obstet Gynecol 1992;80:
612–20.
[37] Wessler S. Medical management of venous thrombosis. Annu Rev Med 1976;27:313–9.
[38] Hirsh J, Hoak J. Management of deep vein thrombosis and pulmonary embolism: a state-
ment for healthcare professionals from the council on thrombosis (in consultation with the
council on cardiovascular radiology), American Heart Association. Circulation 1996;93:
2212–45.
[39] Sandler D, Martin J, Duncan J, et al. Diagnosis of deep-vein thrombosis: comparison of
clinical evaluation, ultrasound, plethysmography, and venoscan with X-ray venogram.
Lancet 1984;8405:716–9.
[40] Stein PD, Hull RD, Pineo G. Strategy that includes serial noninvasive leg tests for diagnosis
of thromboembolic disease in patients with suspected acute pulmonary embolism based on
data from PIOPED. Prospective investigation of pulmonary embolism diagnosis. Arch In-
tern Med 1995;155(19):2101–4.
[41] Douketis JD, Ginsberg JS. Diagnostic problems with venous thromboembolic disease in
pregnancy. Haemostasis 1995;25(1–2):58–71.
[42] Eichinger S. D-dimer testing in pregnancy. Pathophysiol Haemost Thromb 2003;33(5–6):
327–9.
[43] Kelly J, Hunt BJ. A clinical probability assessment and D-dimer measurement should be
the initial step in the investigation of suspected venous thromboembolism. Chest 2003;
124(3):1116–9.
[44] Wells PS, Anderson DR, Ginsberg J. Assessment of deep vein thrombosis or pulmonary
embolism by the combined use of clinical model and noninvasive diagnostic tests. Semin
Thromb Hemost 2000;26(6):643–56.
[45] Ghirardini G, Battioni M, Bertellini C, et al. D-dimer after delivery in uncomplicated preg-
nancies. Clin Exp Obstet Gynecol 1999;26(3–4):211–2.
[46] Francalanci I, Comeglio P, Liotta A, et al. D-dimer concentrations during normal preg-
nancy, as measured by ELISA. Thromb Res 1995;78(5):399–405.
[47] Ballegeer V, Mombaerts P, Declerck PJ, et al. Fibrinolytic response to venous occlusion
and fibrin fragment D-dimer levels in normal and complicated pregnancy. Thromb Hae-
most 1987;58(4):1030–2.
[48] Nolan TE, Smith RP, Devoe LD. Maternal plasma D-dimer levels in normal and compli-
cated pregnancies. Obstet Gynecol 1993;81(2):235–8.
[49] Bounameaux H, de Moerloose P, Perrrier A, et al. Plasma measurement of D-dimer as a di-
agnostic aid in suspected venous thromboembolism: an overview. Thromb Haemost 1994;
71:1–6.
[50] Heim S, Schectman J, Siadaty M, et al. D-dimer testing for deep venous thrombosis: a meta-
analysis. Clin Chem 2004;50(7):1136–47.
[51] Wells P, Anderson D, Rodger M, et al. Evaluation of D-dimer in the diagnosis of suspected
deep-vein thrombosis. N Engl J Med 2003;349:1227–35.
[52] Wells P, Hirsh J, Anderson D, et al. A simple clinical model for the diagnosis of deep-vein
thrombosis combined with impedance plethysmography: potential for an improvement in
the diagnostic process. J Intern Med 1998;243:15–23.
498 ROSENBERG & LOCKWOOD

[53] British Thoracic Society guidelines for the management of suspected acute pulmonary
embolism. Thorax 2003;58(6):470–83.
[54] Polak JF, Wilkinson DL. Ultrasonographic diagnosis of symptomatic deep venous throm-
bosis in pregnancy. Am J Obstet Gynecol 1991;165(3):625–9.
[55] Kassai B, Boissel J, Cucherat M, et al. A systematic review of the accuracy of ultrasound in
the diagnosis of deep venous thrombosis in asymptomatic patients. Thromb Haemost 2004;
91:655–66.
[56] Gottlieb R, Widjaja J, Tian L, et al. Calf sonography for detecting deep venous thrombosis in
symptomatic patients: experience and review of the literature. J Clin Ultrasound 1999;27:415–20.
[57] Palareti G, Cosmi B, Legnani C. Diagnosis of deep vein thrombosis. Semin Thromb
Hemost 2006;32(7):659–72.
[58] Moody AR. Magnetic resonance direct thrombus imaging. J Thromb Haemost 2003;1(7):
1403–9.
[59] Fraser DG, Moody AR, Morgan PS, et al. Diagnosis of lower-limb deep venous thrombo-
sis: a prospective blinded study of magnetic resonance direct thrombus imaging. Ann Intern
Med 2002;136(2):89–98.
[60] Carpenter J, Holland G, Baum R, et al. Magnetic resonance venography for the detection
of deep venous thrombosis: comparison with contrast venography and duplex Doppler
ultrasonography. J Vasc Surg 1993;18:734–41.
[61] Spritzer CE, Evans AC, Kay HH. Magnetic resonance imaging of deep venous thrombosis
in pregnant women with lower extremity edema. Obstet Gynecol 1995;85(4):603–7.
[62] Kanal E, Shellock FG. Policies, guidelines, and recommendations for MR imaging safety
and patient management. SMRI Safety Committee. J Magn Reson Imaging 1992;2(2):
247–8.
[63] Loud PA, Katz DS, Klippenstein DL, et al. Combined CT venography and pulmonary an-
giography in suspected thromboembolic disease: diagnostic accuracy for deep venous eval-
uation. AJR Am J Roentgenol 2000;174(1):61–5.
[64] Garg K, Kemp JL, Wojcik D, et al. Thromboembolic disease: comparison of combined CT
pulmonary angiography and venography with bilateral leg sonography in 70 patients. AJR
Am J Roentgenol 2000;175(4):997–1001.
[65] Duwe KM, Shiau M, Budorick NE, et al. Evaluation of the lower extremity veins in patients
with suspected pulmonary embolism: a retrospective comparison of helical CT venography
and sonography. 2000 ARRS Executive Council Award I. American Roentgen Ray Soci-
ety. AJR Am J Roentgenol 2000;175(6):1525–31.
[66] Heijboer H, Cogo A, Buller H, et al. Detection of deep vein thrombosis with impedance
plethysmography and real-time compression ultrasonography in hospitalized patients.
Arch Intern Med 1992;152:1901–3.
[67] Matthews S. Short communication: imaging pulmonary embolism in pregnancy: what is the
most appropriate imaging protocol? Br J Radiol 2006;79(941):441–4.
[68] Chan W, Ray J, Murray S, et al. Suspected pulmonary embolism in pregnancy: clinical pre-
sentation, results of lung scanning, and subsequent maternal and pediatric outcomes. Arch
Intern Med 2002;162(10):1170–5.
[69] Fedullo P, Tapson V. The evaluation of suspected pulmonary embolism. N Engl J Med
2003;349:1247–56.
[70] Stein P, Terrin M, Hales C, et al. Clinical, laboratory, roentgenographic, and electrocardio-
graphic findings in patients with acute pulmonary embolism and no pre-existing cardiac or
pulmonary disease. Chest 1991;100:598–603.
[71] Green R, Meyer T, Dunn M, et al. Pulmonary embolism in younger adults. Chest 1992;101:
1507–11.
[72] Powrie RO, Larson L, Rosene-Montella K, et al. Alveolar-arterial oxygen gradient in acute
pulmonary embolism in pregnancy. Am J Obstet Gynecol 1998;178(2):394–6.
[73] The urokinase pulmonary embolism trial: a national cooperative study. Circulation 1973;
47(Suppl II):1–108.
 THROMBOEMBOLISM 499

[74] Value of the ventilation/perfusion scan in acute pulmonary embolism. Results of the pro-
spective investigation of pulmonary embolism diagnosis (PIOPED). The PIOPED investi-
gators. JAMA 1990;263:2653–9.
[75] Rodger M, Makropoulos D, Turek M, et al. Diagnostic value of the electrocardiogram in
suspected pulmonary embolism. Am J Cardiol 2000;86:807–9.
[76] Tapson V, Carroll B, Davidson B, et al. The diagnostic approach to acute venous throm-
boembolism. Clinical practice guideline. American Thoracic Society. Am J Respir Crit
Care Med 1999;160:1043–66.
[77] Come P. Echocardiographic evaluation of pulmonary embolism and its response to thera-
peutic interventions. Chest 1992;101:151S–62S.
[78] Gibson N, Sohne M, Buller H. Prognostic value of echocardiography and spiral com-
puted tomography in patients with pulmonary embolism. Curr Opin Pulm Med 2005;
11:380–4.
[79] Pruszczyk P, Torbicki A, Pacho R, et al. Noninvasive diagnosis of suspected severe pulmo-
nary embolism: transesophageal echocardiography vs spiral CT. Chest 1997;112:722–8.
[80] Giannitsis E, Muller-Bardorff M, Kurowski V, et al. Independent prognostic value of cardiac
troponin T in patients with confirmed pulmonary embolism. Circulation 2000;102(2):211–7.
[81] Meyer T, Binder L, Hruska N, et al. Cardiac troponin I elevation in acute pulmonary em-
bolism is associated with right ventricular dysfunction. J Am Coll Cardiol 2000;36(5):
1632–6.
[82] Konstantinides S, Geibel A, Olschewski M, et al. Importance of cardiac troponins I and T
in risk stratification of patients with acute pulmonary embolism. Circulation 2002;106(10):
1263–8.
[83] Douketis JD, Leeuwenkamp O, Grobara P, et al. The incidence and prognostic significance
of elevated cardiac troponins in patients with submassive pulmonary embolism. J Thromb
Haemost 2005;3(3):508–13.
[84] Dalen J, Brooks H, Johnson L, et al. Pulmonary angiography in acute pulmonary embo-
lism: indications, techniques, and results in 367 patients. Am Heart J 1971;81:175–85.
[85] Mills S, Jackson D, Older R, et al. The incidence, etiologies, and avoidance of complications
of pulmonary angiography in a large series. Radiology 1980;136:295–9.
[86] Stein P, Athanasoulis C, Alavi A, et al. Complications and validity of pulmonary angiog-
raphy in acute pulmonary embolism. Circulation 1992;85:462–8.
[87] Balan KK, Critchley M, Vedavathy KK, et al. The value of ventilation-perfusion imaging
in pregnancy. Br J Radiol 1997;70(832):338–40.
[88] Winer-Muram H, Boone J, Brown H, et al. Pulmonary embolism in pregnant patients: fetal
radiation dose with helical CT. Radiology 2002;224(2):487–92.
[89] Groves A, Yates S, Win T, et al. CT pulmonary angiography versus ventilation-perfusion
scintigraphy in pregnancy: implications from a UK survey of doctors’ knowledge of radi-
ation exposure. Radiology 2006;240(3):765–70.
[90] Stein P, Woodard P, Weg J, et al. Diagnostic pathways in acute pulmonary embolism: rec-
ommendations of the PIOPED II Investigators. Radiology 2007;242(1):15–21.
[91] Schuster M, Fishman J, Copeland J, et al. Pulmonary embolism in pregnant patients: a sur-
vey of practices and policies for CT pulmonary angiography. AJR Am J Roentgenol 2003;
181(6):1495–8.
[92] Quiroz R, Kucher N, Zou K, et al. Clinical validity of a negative computed tomography
scan in patients with suspected pulmonary embolism: a systematic review. JAMA 2005;
293(16):2012–7.
[93] Moores LK, Jackson WL Jr, Shorr AF, et al. Meta-analysis: outcomes in patients with sus-
pected pulmonary embolism managed with computed tomographic pulmonary angiogra-
phy. Ann Intern Med 2004;141(11):866–74.
[94] Hayashino Y, Goto M, Noguchi Y, et al. Ventilation-perfusion scanning and helical CT in
suspected pulmonary embolism: meta-analysis of diagnostic performance. Radiology 2005;
234(3):740–8.
500 ROSENBERG & LOCKWOOD

[95] Doyle NM, Ramirez MM, Mastrobattista JM, et al. Diagnosis of pulmonary embolism:
a cost-effectiveness analysis. Am J Obstet Gynecol 2004;191(3):1019–23.
[96] Meaney J, Weg J, Chenevert T, et al. Diagnosis of pulmonary embolism with magnetic res-
onance angiography. N Engl J Med 1997;336:1422–7.
[97] Oudkerk M, van Beek EJ, Wielopolski P, et al. Comparison of contrast-enhanced magnetic
resonance angiography and conventional pulmonary angiography for the diagnosis of pul-
monary embolism: a prospective study. Lancet 2002;359(9318):1643–7.
[98] Kluge A, Mueller C, Strunk J, et al. Experience in 207 combined MRI examinations for
acute pulmonary embolism and deep vein thrombosis. AJR Am J Roentgenol 2006;
186(6):1686–96.
[99] Hirsh J. Heparin. N Engl J Med 1991;324:1565–74.
[100] Raschke R, Reilly B, Guidry J, et al. The weight-based heparin dosing nomogram com-
pared with a ‘‘standard care’’ nomogram. A randomized controlled trial. Ann Intern
Med 1993;119:874–81.
[101] Dahlman TC. Osteoporotic fractures and the recurrence of thromboembolism during preg-
nancy and the puerperium in 184 women undergoing thromboprophylaxis with heparin.
Am J Obstet Gynecol 1993;168(4):1265–70.
[102] Warkentin T, Greinacher A. Heparin-induced thrombocytopenia: recognition, treatment,
and prevention. The Seventh ACCP Conference on Antithrombotic and Thrombolytic
Therapy. Chest 2004;126:311S–37S.
[103] Gould M, Dembitzer A, Doyle R, et al. Low-molecular-weight heparins compared with un-
fractionated heparin for treatment of acute deep venous thrombosis. A meta-analysis of
randomized, controlled trials. Ann Intern Med 1999;130(10):800–9.
[104] Jacobsen A, Qvigstad E, Sandset P. Low molecular weight heparin (dalteparin) for the
treatment of venous thromboembolism in pregnancy. BJOG 2003;110(2):139–44.
[105] Lepercq J, Conard J, Borel-Derlon A, et al. Venous thromboembolism during pregnancy:
a retrospective study of enoxaparin safety in 624 pregnancies. BJOG 2001;108:1134–40.
[106] Sanson B, Lensing A, Prins M, et al. Safety of low-molecular-weight heparin in pregnancy:
a systematic review. Thromb Haemost 1999;81(5):668–72.
[107] Ginsberg J, Hirsh J, Turner D, et al. Risks to the fetus of anticoagulant therapy during preg-
nancy. Thromb Haemost 1989;61:197–203.
[108] Shefras J, Farquharson RG. Bone density studies in pregnant women receiving heparin.
Eur J Obstet Gynecol Reprod Biol 1996;65(2):171–4.
[109] Pettila V, Leinonen P, Markkola A, et al. Postpartum bone mineral density in women
treated for thromboprophylaxis with unfractionated heparin or LMW heparin. Thromb
Haemost 2002;87(2):182–6.
[110] Rodie V, Thomson A, Stewart F, et al. Low molecular weight heparin for the treatment of
venous thromboembolism in pregnancy: a case series. BJOG 2002;109(9):1020–4.
[111] McColl M, Greer I. Low-molecular-weight heparin for the prevention and treatment of
venous thromboembolism in pregnancy. Curr Opin Pulm Med 2004;10(5):371–5.
[112] Gould M, Dembitzer A, Sanders G, et al. Low-molecular-weight heparins compared with
unfractionated heparin for treatment of acute deep venous thrombosis. A cost-effectiveness
analysis. Ann Intern Med 1999;130(10):789–99.
[113] Barbour L, Oja J, Schultz L. A prospective trial that demonstrates that dalteparin require-
ments increase in pregnancy to maintain therapeutic levels of anticoagulation. Am J Obstet
Gynecol 2004;191:1024–9.
[114] Wells PS, Kovacs MJ, Bormanis J, et al. Expanding eligibility for outpatient treatment of
deep venous thrombosis and pulmonary embolism with low-molecular-weight heparin:
a comparison of patient self-injection with homecare injection. Arch Intern Med 1998;
158(16):1809–12.
 Obstet Gynecol Clin N Am
34 (2007) 501–531

Shoulder Dystocia: An Update

Amy G. Gottlieb, MD*, Henry L. Galan, MD
Division of Maternal Fetal Medicine, Department of Obstetrics and Gynecology,
University of Colorado Health Sciences Center, 4200 East 9th Avenue,
B-198, Denver, CO 80262, USA

Shoulder dystocia is something of an enigma: It is poorly defined, ulti-

mately unpredictable, and, once encountered, difficult to treat given the ab-
sence of a proven management algorithm. Because of these facts, and
because shoulder dystocia is frequently associated with permanent birth-
related injuries, it remains one of the most terrifying obstetric emergencies.
The injuries carry potentially daunting medical implications for the patient
and family and are among the most litigated issues in obstetrics [1,2]. The
rare occurrence of shoulder dystocia makes management difficult to teach
during training because severe shoulder dystocia is often handled by attend-
ing obstetricians. However, all practicing clinicians must be prepared to
manage this unpredictable event.

Definition
The American College of Obstetricians and Gynecologists (ACOG)
defines shoulder dystocia as a delivery that requires ‘‘additional obstetric
maneuvers following failure of gentle downward traction on the fetal head
to effect delivery of the shoulders’’ [3]. Many authors use a definition similar
to the ACOG definition [4–10]. Others simply defer to the clinician’s judg-
ment and/or require the clinician to record the term ‘‘shoulder dystocia’’
in the chart [11–14]. Still others include various combinations of the preced-
ing definitions [15–18]. Some divide shoulder dystocia into mild and severe
based upon the number of maneuvers employed [19].
In trying to objectively define shoulder dystocia, Spong and colleagues
[20] proposed defining shoulder dystocia as a ‘‘prolonged head-to-body
delivery time (eg, more than 60 seconds) and/or the necessitated use of
ancillary obstetric maneuvers.’’ The 60-second interval was selected because,

* Corresponding author.
E-mail address: amy.gottlieb@uchsc.edu (A.G. Gottlieb).

0889-8545/07/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ogc.2007.07.002 obgyn.theclinics.com
502 GOTTLIEB & GALAN

in their study, it was approximately two standard deviations above the mean
value for head-to-body time for uncomplicated deliveries. The group sug-
gested that an objective definition would facilitate future studies regarding
prevention and management of shoulder dystocia [21]. Despite this recom-
mendation, shoulder dystocia remains an entity without a clear definition.

Epidemiology
The lack of a uniformly accepted criteria for shoulder dystocia contrib-
utes to its varying incidences found in the literature, which range from
0.2% to 3% [22]. Ethnic differences have also been reported, with African
American women [14] and ‘‘non-Caucasian’’ [23] women reported to have
increased incidence, while a study examining the 1-year incidence of shoul-
der dystocia in California reports that Hispanic patients have a decreased
incidence of shoulder dystocia [24]. One study from Singapore reported
that a birth weight above 3600 g (almost the 90th percentile for this popu-
lation) conferred a relative risk of shoulder dystocia 16.1 times higher when
compared with pregnancies with birth weight below 3600 g [25]. Yet another
study from France concludes that after controlling for confounding factors,
ethnic origin was not an independent factor associated with shoulder dystocia
[26]. The above articles underscore the importance, when researching and re-
porting on shoulder dystocia, of establishing a uniform definition for shoulder
dystocia and a precise definition of the population being studied.

Risk factors
Macrosomia
The known risk factors include macrosomia and fetal anthropometric
variations, maternal diabetes and obesity, operative vaginal delivery, precip-
itous delivery and prolonged second stage of labor, history of shoulder
dystocia or macrosomic fetus, postterm pregnancy, and advanced maternal
age. Macrosomia, like shoulder dystocia, has no uniformly accepted defini-
tion. Proposed definitions for macrosomia include cases where the infant is
large for its gestational age (greater than the 90th percentile for a given ges-
tational age) or weighs more than a specific cut-off limitdmost commonly
4000 g [15,17,27–31] or 4500 g [32–34]. ACOG supports the use of the
4500-g cutoff to diagnose macrosomia because, at this weight, sharp in-
creases are seen in risks of morbidity for infants and mothers [35]. No matter
the definition used, the most serious complication for macrosomic infants is
shoulder dystocia [35], and this risk clearly increases with increasing birth
weight. Nesbitt and colleagues [24] reviewed the 1-year incidence of shoulder
dystocia in California, and reported the percentages of spontaneous births
of nondiabetics complicated by shoulder dystocia as 5.2% for infants weigh-
ing 4000 to 4250 g, 9.1% for those weighing 4250 to 4500 g, 14.3% for those
 SHOULDER DYSTOCIA: AN UPDATE 503

weighing 4500 to 4750 g, and 21.1% for those weighing 4750 to 5000 g. A
Swedish study of newborns from 1973 to 1984 weighing greater than or
equal to 5700 g reported a 40% incidence of shoulder dystocia [36]. Despite
the increasing risk of shoulder dystocia with macrosomia, nearly half of
shoulder dystocia cases occur with birth weight of less than 4000 g [37,38].
While correlating birth weight with shoulder dystocia is convenient for
the sake of retrospective research, no one has been able to consistently iden-
tify the macrosomic fetus antenatally. Methods used to predict the macro-
somic fetus include assessment of maternal risk factors (such as diabetes,
prior history of macrosomic infant, maternal prepregnancy weight, weight
gain during pregnancy, multiparity, male fetus, gestational age, gestational
age greater than 40 weeks, ethnicity, maternal birth weight, maternal height,
maternal age younger than 17 years, and positive 50-g glucose screen with
a negative result on the 3-hour glucose tolerance test [39]), clinical examina-
tion, and ultrasound measurement of the fetus [35]. While it may seem intu-
itive that ultrasound measurements are superior to clinical examination in
the prediction of macrosomia, this is not the case and an error of up to
20% must be taken into account when performing ultrasound near term
[23]. Chauhan and colleagues [40], after a prospective study of over 100 par-
ous women in active labor, concluded that maternal estimates of birth
weight were within 10% of the actual birth weight in 69.8% of cases, com-
pared with 66.1% for clinical estimates and 42.4% for ultrasonography
(femur length and abdominal circumference). These results are further
validated by a prospective study reporting the sensitivity of clinical and ultra-
sonographic prediction of macrosomia (defined as birth weight O4000 g) as
68% and 58%, respectively [41].
The usefulness of ultrasonography for prediction of macrosomia is fur-
ther limited by the fact that fetal weight prediction is less accurate at higher
birth weights. For example, Hadlock’s formula to predict fetal weight has
a mean absolute percent error of 13% for infants greater than 4500 g, com-
pared with 8% for non-macrosomic infants [42]. Using a definition of mac-
rosomia of 4500 g, existing formulas require that an estimated fetal weight
must exceed 4800 g for the fetus to have a greater than 50% chance of being
macrosomic [35,43,44]. Investigators from Iceland [45] and France [46]
attempted unsuccessfully to predict shoulder dystocia based upon ultraso-
nographic measurements of the humerospinous distance and newborn
shoulder length. Improved methods to estimate fetal weight are critical in
identification of the fetus at risk for shoulder dystocia.
Over 50 formulas exist to calculate estimated fetal weight by ultrasound.
The formula proposed by Cohen and colleagues [47] involves subtracting the
biparietal diameter from the abdominal diameter (abdominal circumference
divided by 3.14). They reported that a value greater or equal to 2.6 cm in
infants of diabetic mothers has ‘‘excellent sensitivity, specificity, and predic-
tive value in identifying those fetuses at high risk of birth injury.’’ Elliott and
colleagues [48] reported that, in their study involving infants of diabetic
504 GOTTLIEB & GALAN

mothers, performing cesarean section for all fetuses with a chest-diameter–

biparietal-diameter of 1.4 cm or more would reduce the incidence of trauma-
tive morbidity from 27% to 9%. Winn and colleagues [49] studied which
fetal ultrasonographic parameter best correlates with the neonatal bisacro-
mial diameter and concluded that the fetal chest circumference (at the level
of the four-chamber heart) was most accurate.
Also, while several investigators have reported that various measure-
ments by three-dimensional ultrasonography improves the accuracy of
birth weight prediction [50–52], these results are not universally accepted
and the limited availability of three-dimensional ultrasounds and clinicians
trained in three-dimensional ultrasonography limits its clinical usefulness.
Based upon level A evidence, ACOG states that ‘‘the diagnosis of fetal
macrosomia is imprecise.’’ The ACOG further states that ‘‘for suspected
fetal macrosomia, the accuracy of estimating fetal weight using fetal biom-
etry is no better than that obtained by clinic palpation (Leopold’s maneu-
vers)’’ [35].

Diabetes
Maternal diabetes is an independent risk factor for shoulder dystocia
[3,9,21,35,53–55]. One study demonstrated that, at any incremental birth
weight above 3500 g, the cumulative incidence of shoulder dystocia was sig-
nificantly greater among diabetic than nondiabetic patients [56]. A second
study, this one by Langer and colleagues [37], made a similar findingd
that, at any incremental birth weight above 3750 g, the cumulative incidence
of shoulder dystocia was significantly greater among diabetic than nondia-
betic patients. Langer and colleagues [37] go on to report that when com-
pared gram-for-gram, the perinatal mortality rate, incidence of birth
injuries, and incidence of shoulder dystocia are increased in diabetic
mothers. Diabetes mellitus confers a risk for shoulder dystocia six times
that of the normal population [55], and in births in which the shoulder di-
agnosis is made, the risk of adverse neonatal outcome is higher when mater-
nal diabetes is present [24].
Why is it that infants of diabetic mothers are at increased risk of shoulder
dystocia and resulting birth injury? Some investigators have proposed that
anthropometric differences in macrosomic infants of diabetic and nondia-
betic mothers are to blame [57,58]. McFarland and colleagues [58] report
that macrosomic infants of diabetic mothers are characterized by larger
shoulder and extremity circumferences, decreased head-to-shoulder ratio,
higher body fat, and thicker upper-extremity skin folds compared with non-
diabetic control infants of similar birth weight and birth length. As men-
tioned above, Cohen and colleagues [19] actually quantified sonographic
fetal asymmetry in diabetic patients. Whatever the cause of the increased
risk of shoulder dystocia in this population, intensive treatment of diabetes
reduces the risk of macrosomia and shoulder dystocia [59–62].
 SHOULDER DYSTOCIA: AN UPDATE 505

Operative vaginal delivery

While a few studies have not reported an association between shoulder dys-
tocia and operative vaginal delivery [21,25], the overwhelming conclusion is
that operative vaginal delivery (especially midpelvic extraction) significantly
increases the risk of shoulder dystocia [7,9,10,12,14,18,24,30,53,55,62–65]
with odds ratios ranging from 4.6 to 28.0 depending on the station at applica-
tion and other risk factors. Multiple studies state that vacuum confers an
increased risk when compared with forceps delivery [7,14,53,64,65] and that
the sequential use of forceps and vacuum further increases the risk of shoulder
dystocia and brachial plexus injury [66,67].
Many of the risk factors for shoulder dystocia are interrelated. For exam-
ple, diabetes, both gestational and insulin-dependent, occurs more fre-
quently in older mothers, in women with higher parity, and in those with
a previous large infant [68]. Belfort and colleagues [18] performed multiple
regression analysis and found that only three factors remained statistically
significant for shoulder dystocia: birth weight, diabetes, and operative vag-
inal delivery. They produced a formula incorporating birth weight, 1-hour
glucola, and operative vaginal delivery and found a sensitivity and specific-
ity of 84% and 80%, respectively. Moreover, significant associations per-
sisted when height of fundus, which can be measured antenatally, and
carbohydrate intolerance, which includes pregestational diabetics, were
substituted for birth weight and 1-hour glucola, respectively. They propose
that this model may be useful in the design of prospective studies for man-
aging suspected macrosomia.

Minor risk factors

As previously mentioned, there are multiple minor risk factors for shoul-
der dystocia. Many reports about such factors reveal conflicting results
regarding their significance. In 1990, O’Leary and Leonetti [69] proposed
the dictum: ‘‘once a shoulder dystocia, always a cesarean.’’ That is, any
woman who has a delivery involving shoulder dystocia should subsequently
have all her babies delivered by cesarean. The reported incidence of recur-
rent shoulder dystocia among women with a previous shoulder dystocia
ranges between 1.1% [63] and 16.7% [70]. These studies are all retrospective,
however, and ‘‘one might surmise patients with the worse shoulder dysto-
cias, greatest complications, and biggest babies may have been selected in
future pregnancies to have cesarean section and therefore not appear in
the retrospective analysis’’[55]. The Australian Carbohydrate Intolerance
Study in Pregnant Women, known as the ACHOIS trial [30], actually found
no association between a prior birth complicated by shoulder dystocia and
the risk of shoulder dystocia. ACOG states that ‘‘because most subsequent
deliveries will not be complicated by shoulder dystocia, the benefit of univer-
sal elective cesarean delivery is questionable in patients who have a history
506 GOTTLIEB & GALAN

of shoulder dystocia’’ [3]. The issue of recurrent shoulder dystocia, like

many issues related to shoulder dystocia, is unclear.
It may seem counterintuitive that both a precipitous delivery [53,71,72]
and prolonged labor pattern [4,12,21,63] have been associated with in-
creased incidence of shoulder dystocia. Some authors [71,72] propose
that precipitous delivery is associated with absence of truncal rotation
into an oblique diameter. This, in turn, leads to a persistent anteroposterior
location of the fetal shoulders at the pelvic brim. On the opposite end of the
spectrum, prolonged labor pattern has been associated with between a three-
fold [63] and sevenfold [12] increased risk of shoulder dystocia. Beall and col-
leagues [21] divided patients into primigravid and multigravid groups and
significance for prolonged second stage remained only in the multigravid
group. McFarland and colleagues [73] matched 276 shoulder dystocia cases
with 600 controls and found no association between labor abnormalities
and shoulder dystocia. Other investigators also report no relationship be-
tween prolonged labor pattern and shoulder dystocia [9,16,26,53,74]. This
discrepancy could be due to variations in study design, study population,
or labor management [14].
Several clinical entitiesdmaternal obesity, prolonged pregnancy, ad-
vanced maternal age, male fetal genderdare associated with macrosomia
and, therefore, shoulder dystocia. ACOG states, and many investigators con-
cur, that ‘‘maternal obesity is associated with macrosomia and, thus, obese
women are at risk for shoulder dystocia’’ [3,26,74–76]. However, after control-
ling for confounding effects, such as fetal macrosomia, previous macrosomic
infant, midpelvic instrumental delivery, and/or coexisting medical complica-
tions (such as diabetes and/or hypertension), multiple logistic regression per-
formed in various studies reports that maternal obesity actually is not
significant as an independent risk factor for shoulder dystocia [9,16,21,77].
Similarly to maternal obesity, prolonged pregnancy also increases the risk
of macrosomia [39] and therefore, according to some investigators, shoulder
dystocia [78]. Baskett and Allen [63] reported that prolonged pregnancy in-
creased the risk of shoulder dystocia threefold. Also, however, just as
evidence relating maternal obesity and shoulder dystocia is conflicting, there
are similarly conflicting reports about prolonged pregnancy and the risk of
shoulder dystocia with some reports finding no independent relationship
between postdatism and shoulder dystocia [9,16,21,24,54].
Advanced maternal age is associated with increasing incidences of coex-
isting medical disease, including diabetes (both gestational and pregesta-
tional [68]) and obesity. Therefore, it makes sense that advanced maternal
age confers an increased risk of shoulder dystocia [55]. However, using
logistic regression analysis in a large population (n ¼ 75,979), Langer and
colleagues [37] found no significant contribution of maternal age on the in-
cidence of shoulder dystocia.
The incidence of male gender in shoulder dystocia series (55%–68%) is
greater than incidence of male gender in the general obstetric population
 SHOULDER DYSTOCIA: AN UPDATE 507

[55,79–82]. The reason for this discrepancy is unclear. Dildy and Clark [55]
postulated that birth weight, which is established to be greater in male new-
borns, places them a greater risk of macrosomia. It is also plausible that
a difference in anthropomorphic dimensions between male and female in-
fants exists, as seen between infants of women with and without diabetes.
This issue deserves further study.
The use of oxytocin, which is rather prevalent in many labor and delivery
units for labor augmentation, has been associated with increased risk of
shoulder dystocia [83]. It is likely not oxytocin augmentation alone that
causes shoulder dystocia, but its use is probably associated with labor
dystocia and fetal macrosomia [55]. With regards to maternal obesity, pro-
longed pregnancy, advanced maternal age, male fetal gender, and oxytocin
augmentation, it is unclear whether their relationships with shoulder dysto-
cia is an independent entity or a result of confounding variables. Whatever
the relationship, it is clear that the predictive value of these risk factors is not
high enough to be useful in a clinical setting [3].

Method of delivery
Despite O’Leary and Leonetti’s insistence that ‘‘once a shoulder dystocia,
always a cesarean,’’ recent reports have suggested that cesarean may not al-
ways be the prudent choice for deliveries following a shoulder dystocia. As
mentioned above, the benefit of universal elective cesarean is questionable in
patients with a history of shoulder dystocia and the counseling of the patient
should play into the decision-making process [3]. What about management
in patients (with or without a history of shoulder dystocia) who appear to be
carrying a macrosomic fetus? Rouse and colleagues [84] constructed a deci-
sion analytic model to compare three policies in both diabetic and nondia-
betic patients: (1) management without ultrasound, (2) ultrasound and
elective cesarean delivery for estimated fetal weight of 4000 g or more,
and (3) ultrasound and elective cesarean delivery for estimated fetal weight
of 4500 g or more. The study compared rates of shoulder dystocia, rates of
permanent brachial plexus injury, the number and cost of additional cesar-
ean births, and the potential cost savings for averting permanent brachial
plexus injury. In the nondiabetic population, the study found that for
each permanent brachial plexus injury prevented by the 4500-g policy, an
increase of 8.5% in the cesarean birth rate with an additional cost of $8.7
million. The findings of the 4000-g policy in nondiabetics increased the ce-
sarean rate 11.5% with an additional cost of $4.9 million. Expressed in other
terms, one maternal death would result for every 3.2 brachial plexus injuries
prevented. They concluded that a policy of elective cesarean birth in nondi-
abetics at a cutoff of either 4000 or 4500 g would be medically and econom-
ically unsound. They were unable to adapt the model for an estimated fetal
weight of 5000 g because of the paucity of data available to analyze. Anal-
ysis in the diabetic population, with a macrosomia threshold of 4500 g
508 GOTTLIEB & GALAN

reported that 443 cesareans and $930,000 are required to prevent one
permanent brachial plexus injury, which they reported as ‘‘more tenable,
although the absolute merits of the approach are debatable.’’ In a later
publication, this group questions whether prophylactic cesarean delivery
for fetal macrosomia diagnosed by means of ultrasonography is a ‘‘Faustian
bargain’’ [85]. A recent publication analyzed mode of delivery and survival
of macrosomic infants and concluded that cesarean delivery may reduce the
risk of neonatal death in infants weighing over 5000 g [28]. All of the above
findings are based upon birth weight and not estimated fetal weight.
Gonen and colleagues [32] performed a retrospective assessment of a policy
at their institution that recommended cesarean delivery for macrosomia (esti-
mated fetal weight 4500 g or more) and found an insignificant effect on the in-
cidence of permanent brachial plexus palsy. Cesarean delivery is not 100%
reliable for averting permanent brachial plexus palsy, as will be discussed later
[86–88]. Improving the clinician’s ability to accurately predict fetal weight
could increase the benefit of elective cesarean delivery. Now, however, the
concept of prophylactic cesarean to prevent shoulder dystocia and its perma-
nent sequelae has not been supported by clinical or theoretic data [89]. Based
on expert opinion, ACOG states that planned cesarean delivery to prevent
shoulder dystocia may be considered for suspected fetal macrosomia when
there is an estimated fetal weight of 5000 g in women without diabetes [3].
Studies regarding induction of labor (IOL) are divided into three cate-
gories: IOL for macrosomia in nondiabetic patients, IOL for macrosomia
in diabetic patients, and IOL for prevention of macrosomia in diabetics.
Several, small retrospective studies report that labor induction in nondia-
betic patients appears to at least double the risk of cesarean delivery without
reducing shoulder dystocia or newborn morbidity [17,27,90–92]. Gonen and
colleagues [93] performed a prospective study where patients at term with
ultrasonic fetal weight estimation of 4000 to 4500 g were randomized to ei-
ther induction of labor or expectant management. There was no difference
in either the number of cesarean deliveries, shoulder dystocia, or neonatal
morbidity. Furthermore, Nassar and colleagues [16] report vaginal delivery
is achievable in 88.9% of pregnancies allowed to labor with infants weighing
4500 g or more, at the expense of a 15.5% risk of shoulder dystocia, a 3%
risk of brachial plexus injuries, and a 7.7% risk of perineal trauma. ACOG
recommends that, as induction does not improve maternal or fetal out-
comes, suspected fetal macrosomia in nondiabetic patients is not an indica-
tion for induction of labor.
Herbst [94] performed a cost-effective analysis including three strategies
of managing infants with an estimated fetal weight of 4500 g. Based solely
on cost, expectant treatment was the preferred strategy at a cost of
$4014.33 per injury-free child, versus an elective cesarean delivery cost of
$5212.06 and an induction cost of $5165.08. This suggests that expectant
treatment is the most cost-effective approach to treatment of the fetus
with suspected macrosomia in nondiabetic patients.
 SHOULDER DYSTOCIA: AN UPDATE 509

Management of patients with diabetes must account for glucose control.

Lurie and colleagues [95] compared outcomes after implementation of a pol-
icy in which labor was induced at 38 to 39 weeks’ gestation for insulin-
requiring diabetic patients in comparison to pregnancies managed
expectantly. Although the incidence of shoulder dystocia was higher in
the group managed expectantly, the difference did not reach statistical sig-
nificance. Kjos and colleagues [68] demonstrated that, in pregnancies com-
plicated by insulin-dependent diabetes, expectant management does not
reduce the incidence of cesarean birth and actually leads to an increased
prevalence of large-for-gestational-age infants and shoulder dystocia.
They conclude that if delivery is not pursued at 38 weeks, careful monitoring
of fetal growth is essential. Partially because other groups have not con-
firmed these results, ACOG does not specifically recommend induction of
labor for diabetics. ACOG states, however, that ‘‘expectant management
beyond the estimated due date is generally not recommended’’ [96]. They
further recommend, based on expert opinion, that diabetic patients with
estimated fetal weight greater than 4500 g may be offered prophylactic
cesarean delivery [3].
Nesbitt and colleagues [24] are among many investigators to report an in-
creased risk of shoulder dystocia with operative vaginal delivery, especially
midpelvic extraction. Obstetric patients should be counseled regarding the
risks, benefits, and alternatives of both forceps and vacuum extraction
before labor. While in labor, the entire clinical scenario should be taken
into account whenever making the decision to proceed with operative vag-
inal delivery as there are certainly times when the risks of cesarean delivery
outweigh the risks of operative vaginal delivery. It seems prudent to use
forceps or vacuum with caution in the setting of suspected macrosomia.

Intrapartum management
Upon arrival to labor and delivery, estimated fetal weight should be al-
ways be documented. Despite the notion that estimations have an inherent
margin of error, legal texts [97–99] and journals [100] have maintained that
a physician’s failure to assess fetal weight during pregnancy or labor consti-
tutes a deviation from standards of practice [89]. Along these same lines, an
assessment of the adequacy of the patient’s pelvis should be performed and
documented either at a prenatal visit or on labor and delivery. While care of
laboring patients may include recording the labor curve, the labor parto-
gram is not predictive of shoulder dystocia [73].
If the clinician is concerned about a possible shoulder dystocia, certain
‘‘shoulder precautions’’ can be employed. This generally includes position-
ing the patient in the dorsal lithotomy position [101] with the bed ‘‘broken
down’’ such that the patient’s buttocks are at the end of the bed [102], emp-
tying the patient’s bladder before delivery, ensuring the presence of an extra
nurse or other clinician, and having a stool immediately available in case
510 GOTTLIEB & GALAN

suprapubic pressure is indicated. One study evaluated whether prophylactic

use of McRoberts maneuver (exaggerated hyperflexion of the patient’s legs)
and suprapubic pressure was beneficial in reducing head-to-body delivery
time. The investigators randomized pregnancies with estimated fetal weight
over 3800 g to either undergo prophylactic maneuvers or deliver in dorsal
lithotomy with additional maneuvers being employed only if necessary,
and demonstrated that prophylactic maneuvers were not beneficial. They
did state that ‘‘the only apparent advantage of performance of prophylactic
maneuvers was that an overt diagnosis of shoulder dystocia was avoided in
a number of patients’’ [101]. Furthermore, the use of McRoberts maneuver
before the clinical diagnosis of shoulder dystocia does not significantly
change the traction forces applied to the fetal head during vaginal delivery
in multiparous patients [103]. While ‘‘shoulder precautions’’ seem reason-
able, many shoulder dystocias are encountered in the absence of risk factors;
therefore, all practitioners should be prepared to manage this obstetric
emergency at every delivery.

Maneuvers
How much time do I have?
Unfortunately, there is no one superior algorithm to manage shoulder
dystocia. Typically, shoulder dystocia is heralded by the classic ‘‘turtle
sign’’; after the fetal head is delivered, it retracts back tightly against the
maternal perineum [71]. Shoulder dystocia, as mentioned above, is typically
not diagnosed until downward traction fails to deliver the shoulders. At this
point, one of the major concerns is: How much time can elapse without risk-
ing fetal hypoxic injury? Insult to the fetus from hypoxia results from com-
pression of the neck and central venous congestion, as well as compression
of the umbilical cord, reduced placental intervillous flow from prolonged
increased intrauterine pressure, and secondary fetal bradycardia [102]. Stal-
lings and colleagues [6] report that shoulder dystocia resulted in statistically
significant but clinically insignificant reduction in mean umbilical artery gas
parameters (pH of 7.23 versus 7.27). Wood’s [104] work in 1973 reports a de-
crease of 0.14 pH U/min during trunk delivery. Such a drop would suggest
that a pH less than 7.00 might occur with a delay in delivery as short as 2 or
3 minutes. Stallings and colleagues [6] analyzed Wood’s results and reported
that they are of limited value in regard to fetal acidosis because the method-
ology involves inappropriate extrapolation. Stallings and colleagues further
report that their data suggest the change in fetal pH after the onset of shoul-
der dystocia is probably slower than previously thought.
Ouzounian and colleagues [105] analyzed 39 cases of shoulder dystocia, 15
with neonatal brain injury and 24 without. They reported that the mean inter-
val in the injured group was 10.6 minutes compared with 4.3 minutes in the un-
injured group. On the basis of a receiver-operating characteristic curve, the
 SHOULDER DYSTOCIA: AN UPDATE 511

investigators stated that a threshold interval of 7 or more minutes had a 67%

sensitivity and 74% specificity for predicting brain injury. Allen and
colleagues [106] reported that head-to-body interval of 6 or more minutes
was the only significant predictor of low Apgar scores at 5 minutes in vaginal
deliveries that resulted in permanent brachial plexus injury. These infants,
however, did not appear to be at imminent risk of permanent central neuro-
logic dysfunction. While it is reasonable to assume that permanent central
neurologic dysfunction is associated with prolongation to head-to-shoulder
interval thresholds, there is no clear consensus for the amount of time allowed
to safely resolve a shoulder dystocia.
When a shoulder dystocia is encountered, the clinician must first designate
a care-team member to mark the time. Tracking the time is necessary both for
documentation and to allow periodic reassessment of the situation in case of
severe shoulder dystocia. Clinicians’ first reaction to difficult delivery is to ex-
ert considerably larger forces than normal, thereby possibly increasing the risk
of fetal injury [107]. However, the clinician (and the other staff) must remain
calm and proceed through maneuvers to resolve the dystocia. Adequate ancil-
lary staff, including nursing staff, pediatricians, anesthesiology staff, and other
obstetricians, if available, should be called to the room. Having a prearranged
protocol in place involving a team approach to management of shoulder dys-
tocia can help all team members be aware of their role (Box 1).

McRoberts maneuver
According to ACOG [3], the performance of the McRoberts maneuver
(Figs. 1 and 2), with or without suprapubic pressure, is a reasonable initial
approach to shoulder dystocia. The McRoberts maneuver does not change
the actual dimension of the maternal pelvis (see Figs. 1 and 2); it straightens

Box 1. Maneuvers for shoulder dystocia

Initial maneuvers
McRoberts
Suprapubic pressure
Episiotomy?
Woods’ corkscrew
Rubin’s
Delivery of posterior arm
Gaskin position
Last-resort maneuvers
Intentional clavicular fracture
Zavanelli
Symphysiotomy
Hysterotomy
512 GOTTLIEB & GALAN

Fig. 1. The McRoberts maneuver. This maneuver involves hyperflexion of the maternal thighs
against the abdomen, usually involving two assistants, each of whom grasps a maternal leg.

the sacrum relative to the lumbar spine, allowing cephalic rotation of the
symphysis pubis sliding over the fetal shoulder [108]. Suprapubic pressure
(Fig. 3) assists in dislodging the anterior shoulder [71]. Gonik and colleagues
[109] demonstrated that McRoberts positioning reduced delivery force up to
37% for endogenous load (maternal force) and up to 47% for exogenous
loads (clinician applied), thereby decreasing brachial plexus stretching.
This group also noted greater stretching with endogenous versus exogenous
force. Along these same lines, Buhimschi and colleagues [110] reported that
use of McRoberts position almost doubled the intrauterine pressure devel-
oped by contractions alone.

Fig. 2. The McRoberts maneuver does not change the actual dimension of the maternal pelvis.
Rather, the maneuver straightens the sacrum relative to the lumbar spine, allowing cephalic
rotation of the symphysis pubis sliding over the fetal shoulder.
 SHOULDER DYSTOCIA: AN UPDATE 513

Fig. 3. Suprapubic pressure. Suprapubic pressure is applied directing the anterior shoulder
downward and laterally. If possible, pressure should be directed from the side of the fetal spine
toward the face. Pressure should be applied by an assistant with either the palm or fist.

The McRoberts maneuver is often done with the application of suprapu-

bic pressure (see Fig. 3), which involves an assistant other than the primary
delivering provider to apply pressure to the anterior shoulder of the fetus
just cephalad to the pubic symphisis so that the shoulder is pushed anteri-
orly relative to the fetus. The success of McRoberts in resolving shoulder
dystocia (used either alone or in combination with suprapubic pressure) is
reported between 42% and 58% [5,15,111]. McRoberts positioning has
risks, however. Continued attempts at McRoberts maneuver during severe
shoulder dystocia are often associated with increasing traction, which can
lead to increased risk of brachial plexus injury [63,112]. Gherman and
colleagues [113] published a case report involving symphyseal separation
and transient femoral neuropathy associated with the McRoberts maneuver.
In spite of this reported case, the investigators still recommend the McRo-
berts position as the initial technique in management of shoulder dystocia
but caution against ‘‘overly aggressive hyperflexion and abduction of the
maternal thighs onto the abdomen.’’

Episiotomy?
Shoulder dystocia is typically a ‘‘bony’’ obstruction and not a result of
obstructing soft tissue [3]. Management by episiotomy or proctoepisiotomy
has been associated with a nearly sevenfold increase in the rate of perineal
514 GOTTLIEB & GALAN

trauma without benefit of reducing the occurrence of neonatal depression or

brachial plexus palsy [114,115]. The decision to cut a generous episiotomy or
proctoepisiotomy must be based upon clinical circumstances, such as a nar-
row vaginal fourchette in a primigravid patient or the need to perform fetal
manipulation [70].

Woods’ corkscrew and Rubin’s maneuvers

The Woods’ corkscrew maneuver (Fig. 4) involves the practitioner
abducting the posterior shoulder by exerting pressure onto the anterior
surface of the posterior shoulder (see Fig. 4). The Rubin’s maneuver (reverse
Woods’) entails the practitioner applying pressure to the posterior surface of
the most accessible part of the fetal shoulder (ie, either the anterior or pos-
terior shoulder) to effect shoulder adduction (Fig. 5) [116]. Gurewitsch and
colleagues [117] developed a laboratory birthing simulator and determined
that the anterior Rubin’s maneuver required less traction and produces less
brachial plexus stretch than McRoberts positioning or posterior Rubin’s.
They encourage an emphasis on practicing the Rubin’s maneuver during
training so clinicians are familiar with its use.

Delivery of posterior arm

Delivery of the posterior arm was first described by Barnum [118] in 1945.
To perform the maneuver, pressure should be applied by the delivering
provider at the antecubital fossa to flex the fetal forearm. The arm is subse-
quently swept out over the infant’s chest and delivered over the perineum
(Figs. 6 and 7). Rotation of the trunk to bring the posterior arm anteriorly
is sometimes required. Grasping and pulling directly on the fetal arm and
applying pressure onto the midhumeral shaft should be avoided when
possible, as bone fracture may occur [119], although these injuries typically
heal without any long-term morbidity [102]. Kwek and Yeo [102]
recommend placing traction on the posterior axilla to help facilitate delivery
of the posterior arm. Posterior arm delivery effectively creates a 20% reduc-
tion in shoulder diameter and, according to Poggi and colleagues [120],
reduces the obstruction by more than a factor of two when compared
with McRoberts position. Poggi further recommends prioritizing posterior
arm delivery in management algorithms and states that, when the trunk fails
to deliver after posterior arm delivery, clinicians should proceed directly to
emergent techniques, such as intentional clavicular fracture, cephalic
replacement, or symphysiotomy.

Gaskin position
Several investigators propose placing the patient in the ‘‘all-fours’’ (or
Gaskin) position (Fig. 8) to help resolve shoulder dystocia [121,122]. Bruner
and colleagues [121] report a series of 82 consecutive cases of shoulder dys-
tocia managed by moving the laboring patient to her hands and knees.
 SHOULDER DYSTOCIA: AN UPDATE 515

Fig. 4. The Woods’ corkscrew maneuver. This maneuver involves applying pressure to the
clavicular surface of the posterior arm, allowing rotation (A) such that the anterior shoulder
dislodges (B) from behind the maternal symphysis. Curved arrow shows rotation. Straight
arrow shows manual rotation of infant’s body in coordination with rotation by hand below.
516 GOTTLIEB & GALAN

Fig. 5. The Rubin’s maneuver. This maneuver involves applying pressure to the most accessible
part of the fetal shoulder (ie, either the anterior or posterior shoulder) to effect shoulder adduc-
tion (A). (B) Curved arrows shows rotation of fetal shoulders.

Sixty-eight women (or 83%) delivered without need for any additional ma-
neuvers with no increase in maternal or fetal morbidity. The ‘‘all-fours’’ po-
sition exploits the effects of gravity and increased space in the hollow of the
sacrum to facilitate delivery of the posterior shoulder and arm [122].

Walcher’s position
Walcher’s position, a reverse form of McRoberts position, in which the
thighs are hyperextended, results in downward displacement of the
 SHOULDER DYSTOCIA: AN UPDATE 517

Fig. 6. Delivery of the posterior arm. To deliver the posterior arm, pressure should be applied
at the antecubital fossa to flex the fetal forearm. The forearm or hand is subsequently grasped
and the arm swept out over the infant’s chest and delivered over the perineum. Rotation of the
trunk to bring the posterior arm anteriorly is sometimes required. (A) First, turn fetal head to
allow entry of practitioner’s hand to facilitate manipulation. (B) Second, support fetal head
with one hand and sweep second hand posteriorly. (C) Next, flex infant’s arm at antecubital
fossa to allow practitioner to grasp posterior forearm or hand. (D) Deliver posterior arm.
This allows rotation of the fetus with the goal of disimpacting the anterior shoulder. (E) Further
rotate fetus to facilitate delivery.
518 GOTTLIEB & GALAN

Fig. 7. (A, B) This figure shows delivery of the posterior arm with facilitation of delivery by
hysterotomy. The intra-abdominal hand can be used to rotate the anterior shoulder to allow
vaginal delivery; or a Zavanelli maneuver can be performed subsequently, allowing cesarean
delivery.

symphysis pubis by 1.0 to 1.5 cm [123]. While it is mentioned in some of the

older literature as a maneuver to help relieve shoulder dystocia, there are no
recent case series or reports in the literature about its use and it is not men-
tioned in the most recent ACOG bulletin on shoulder dystocia [3].

Clavicular fracture
Intentional clavicular fracture has been described, mostly in older litera-
ture, by applying upward digital pressure on the fetal clavicle against the
maternal pubic ramus. Although this would decrease the bisacromial diam-
eter, there is significant risk of damage to the brachial plexus and pulmonary
vasculature. Additionally, cleidotomy, which involves separation of the
clavicle with a blade or pair of scissors, is probably best reserved following
intrauterine death [102] as it is technically difficult to perform and carries
significant fetal risks [71].

Zavanelli maneuver
For catastrophic shoulder dystocia, cephalic replacement, hysterotomy,
and symphysiotomy are last-resort options. Cephalic replacement (Zavanelli
maneuver) is essentially a reversal of the delivery process whereby the fetal
 SHOULDER DYSTOCIA: AN UPDATE 519

Fig. 8. The Gaskin position. The ‘‘all fours’’ position exploits the effects of gravity and in-
creased space in the hollow of the sacrum to facilitate delivery of the posterior shoulder and
arm.

neck is flexed, restitution is reversed, the head is rotated back to the occi-
pito-anterior position, and digital pressure is applied to replace the head
within the uterine cavity. The use of tocolytics (eg, terbutaline or nitroglyc-
erine) can be used along with halothane or other general anesthetic agents to
facilitate successful completion of the maneuver, which is followed by a ce-
sarean delivery [71,102]. Among the 59 reported cases of attempted cephalic
replacements described by O’Leary [124], only 6 (10.2%) were unsuccessful.
Sandberg [125] reviewed 12 years’ worth of literature on the Zavanelli
maneuver and reported an overall 92% success rate. While Sandberg men-
tioned numerous injuries in these infants, the conclusion was that most of
these injuries were due to pre-Zavanelli manipulations and protracted hyp-
oxia. Reported maternal complications include both uterine and vaginal
rupture but, again, Sandberg states that these injuries cannot be directly
attributed to the Zavanelli procedure. He concludes that ‘‘in most cases of
cephalic replacement, the Zavanelli maneuver appears to be simple and suc-
cessful, even without prior experience.’’ Despite this review, ACOG states
that the Zavanelli maneuver is associated with a significantly increased
risk of fetal morbidity and mortality and of maternal morbidity and that
520 GOTTLIEB & GALAN

it should only be performed in cases of severe shoulder dystocia unrespon-

sive to more commonly used maneuvers [3].

Symphysiotomy
Due to the significant maternal morbidity associated with symphysiot-
omy, including bladder neck injury and infection, it should only be used
as a last attempt to preserve fetal life [102,126]. To perform a symphysiot-
omy, the patient should be placed in an exaggerated lithotomy position
with proper support of the legs. Then, if at all possible, a transurethral cath-
eter should be placed. The clinician, with his or her index and middle finger,
should displace the urethra laterally and partially incise the cephalad por-
tion of the symphysis with a scalpel blade or Kelly clamp [71]. Goodwin
and colleagues [126] presented a case series in which emergency symphysiot-
omy was performed in three patients in an effort to preserve fetal life after
approximately 12, 13, and 23 minutes. All infants subsequently died because
of severe anoxic insult. Goodwin suggests that, because of operator inexpe-
rience and maternal morbidity, the role of emergency symphysiotomy
remains unclear. Furthermore, they state that because the procedure takes
at least 2 minutes from the time a decision is made, it should be initiated
within 5 to 6 minutes of delivery of the fetal head.

Hysterotomy
The use of hysterotomy or an upper-segment uterine incision allows
either more direct pressure or cephalic replacement. More direct pressure
can achieve shoulder rotation or directly dislodge the anterior shoulder
for vaginal delivery. Cephalic replacement can facilitate abdominal delivery
[102]. The use of hysterotomy or an upper-segment uterine incision is by no
means always effective, and tragic consequences have been described [126].

Maneuvers to avoid
While no good evidence exists regarding the role of fundal pressure in
shoulder dystocia, fundal pressure applied in the setting of shoulder dystocia
has been reported to further press the shoulder on the pelvic brim and
increase intrauterine pressure, thereby increasing the risk of permanent neu-
rologic injury and orthopedic damage [102,127]. Hankins [128] published
a case report involving lower thoracic spinal cord injury with permanent
neurological injury when fundal injury was applied in an attempt to relieve
shoulder dystocia. The ACOG Practice Bulletin on shoulder dystocia [3]
reports that ‘‘fundal pressure may further worsen impaction of the shoulder
and also may resulting uterine rupture.’’ Therefore, it seems reasonable to
avoid fundal pressure with shoulder dystocia.
Any nuchal cord, if unable to be reduced over the fetus’ head, should not be
cut and clamped if at all possible. Iffy and Varandi [129] report a series of five
 SHOULDER DYSTOCIA: AN UPDATE 521

cases of cerebral palsy in infants where shoulder dystocia was recognized only
after interruption of a nuchal cord. The delay in delivery in that series ranged
from 3 to 7 minutes. Flamm [130] reports a case in which a tight nuchal cord
was encountered during a severe shoulder dystocia and was not clamped or
cut. He proposed that if the cord was severed, the infant ‘‘might have suffered
permanent neurologic injury or died before birth.’’ Stallings and colleagues [6]
speculate that, even in the face of shoulder dystocia with a nuchal cord, some
cord circulation may continue and that severing the cord may contribute to fe-
tal hypoxia and hypotension during the time it takes to resolve the dystocia.

Postpartum management
Shoulder dystocia is among the four most common causes of medical lit-
igation [131] and has been estimated to account for up to 11% of obstetric
claims. Following all complicated deliveries, measurements of umbilical
cord blood gases must be obtained, a discussion with the patient and family
must be held, and the events of the delivery must be documented by all care-
team members involved. Parents are usually traumatized by the events and
they deserve complete, immediate, and accurate information regarding the
delivery, the maneuvers used, and the rationale behind management [102].
If a brachial plexus injury is present, the clinician should not speculate
regarding the cause.
Acker [132] recommends that a shoulder dystocia intervention form
should include the following information:
When and how the dystocia was diagnosed
Progress of labor (active phase and second stage)
Position and rotation of the infant’s head
Presence of episiotomy
Anesthesia required
Estimation of force of traction applied
Order, duration, and results of maneuvers used
Duration of shoulder dystocia
Documentation of adequate pelvimetry before initiating labor induction
or augmentation
Neonatal and obstetric impressions of the infant after delivery
Information given to gravida that shoulder dystocia had occurred
Unfortunately, recent publications [5,63,133] have noted incomplete doc-
umentation in the majority of shoulder dystocia cases. A legal case with in-
adequate documentation can be difficult to defend.

Neonatal sequelae of shoulder dystocia

McFarland and colleagues [15] found that fetal and maternal morbidity
increases with number of maneuvers employed to resolve shoulder dystocia.
522 GOTTLIEB & GALAN

In regards to recent literature, many papers are using fetal injury (namely,
brachial plexus injury) as an endpoint as opposed to using shoulder dystocia
as the study endpoint. Fetal injuries associated with shoulder dystocia
include brachial plexus injury, fracture of the clavicle or humerus, and,
rarely, hypoxic injury or neonatal death. Reports of brachial plexus injury
during deliveries complicated by shoulder dystocia vary from 4% to 40%
[1,56,63,70,81,84,134–138], although case-control studies report an 18- to
21-fold increase in the risk of brachial plexus injury among infants with
birth weight greater than 4500 g [139–141]. The obstetrical literature reports
less than 10% of Erb’s palsies are permanent [63,135–137], although persis-
tent injury may be more common in birth weights over 4500 g [142] and in
infants of diabetic mothers [24]. Pediatric and orthopedic literature reports
permanent injury in up to 15% to 25% of cases [143,144].

Brachial plexus injury

Benjamin [143] provides an excellent review of the characteristics of
brachial plexus injuries. Damage to spinal nerves C5-C6 leads to Erb’s or
Erb-Duchenne palsy (80% of brachial plexus injuries). The classic posture is
a result of paralysis or weakness in the shoulder muscles, the elbow flexors,
and the forearm supinators. The affected arm hangs down and it is
internally rotated, extended, and pronated. Oftentimes, the C7 nerve is also in-
volved, causing loss of innervation to the forearm, wrist, and finger extensors.
The loss of extension causes the wrist to flex and the fingers to curl updthe
‘‘waiter’s tip’’ position. Phrenic nerve injury with resulting diaphragmatic
paralysis may be present due to damage to the C4 segment. Avulsion of
C8-T1 causes Klumpke’s palsy, which is characterized by weakness of the
triceps, forearm pronators, and wrist flexors leading to a ‘‘clawlike’’ paralyzed
hand with good elbow and shoulder function. Upper-arm function differenti-
ates Klumpke’s palsy from Erb’s palsy. Unfortunately, only 40% of
Klumpke’s palsies resolve by 1 year of life [73]. An associated Horner’s
syndrome with sensory deficits on the affected side, contraction of the pupil,
and ptosis of the eyelid is caused by cervical sympathetic nerve injury. Com-
plete brachial plexus injury, or Erb-Klumpke palsy, involving C5-T1, is char-
acterized by a flail, paralyzed arm without sensation or reflexes. Brachial
plexus injury occurs regardless of the number and type of maneuvers used
[5,63,145] and does not appear predictable before delivery [8,88].
Excessive traction applied at the time of delivery can cause injury to the bra-
chial plexus. Allen and colleagues [112] performed an in vivo study looking at
the force applied during delivery and demonstrated a significant difference in
the peak delivery force applied in routine versus shoulder dystocia deliveries.
The clinical utility of this information remains unknown. Birth injury is not
the only cause of brachial plexus injury. A significant proportion (34%–
47%) of brachial plexus injuries are not associated with shoulder dystocia
[3]. In fact, 4% occur after cesarean birth [84,146,147]. Aside from excessive
 SHOULDER DYSTOCIA: AN UPDATE 523

traction, other causes of injury include the normal forces of labor and delivery
[73], a compressive effect of the symphysis pubis against the brachial plexus,
and abnormal intrauterine pressures arising from uterine anomalies, such as
an anterior lower uterine segment leiomyoma, an intrauterine septum, or a bi-
cornuate uterus [72,148,149]. Performance of electromyeolography soon after
delivery (within 24–48 hours) can help determine the timing of brachial plexus
injury. Electromyelographic evidence of muscular denervation normally re-
quires 10 to 14 days to develop. Its finding in the early neonatal period, there-
fore, strongly suggests an insult predating delivery [150–152]. No matter the
cause, care of the newborn with brachial plexus injury should involve a multi-
disciplinary approach including pediatrics, pediatric neurology, physical ther-
apy, and possible referral to a brachial pleuxus injury center. The care plan
should be clearly communicated with the parents.

Fracture
Orthopedic fractures almost invariably heal with simple supportive ther-
apy and do not lead to permanent disability [153,154]. One investigator even
calls clavicular fracture ‘‘benign’’[154]. While clavicular fracture often
occurs in the absence of shoulder dystocia [155], the incidence of fracture
of the clavicle at the time of shoulder dystocia ranges from approximately
3% to 9.5% [5,6,15] with increasing risk with greater birth weight [5,155].

Fetal mortality
The reported incidence of perinatal death attributed to shoulder dystocia
ranges from zero to 2.5% [84,156,157]. Rouse and colleagues [84] conclude
that ‘‘although shoulder dystocia may result in perinatal death, this happens
rarely and would not serve as a reasonable justification, at least in pregnan-
cies of nondiabetic women, for cesarean delivery based on the ultrasono-
graphic diagnosis of macrosomia.’’

Maternal sequelae of shoulder dystocia

A study of 236 shoulder dystocias reported an 11% rate of postpartum
hemorrhage and a 3.8% rate of fourth-degree lacerations [111]. These were
independent of type of maneuver or maneuvers employed to resolve the dys-
tocia. Other maternal complications that have been reported include vaginal
and cervical lacerations, and bladder atony [71]. It should be noted that ‘‘he-
roic’’ measures, such as the Zavanelli maneuver and symphysiotomy, are
often associated with significant risk of maternal morbidity [124,158].

Training for shoulder dystocia

A team-oriented approach is necessary for management of shoulder
dystocia. A formalized activation system, good leadership, and good
524 GOTTLIEB & GALAN

organization of team members, with each member well trained in the

management of obstetric emergencies, helps facilitate a smooth delivery
of the fetus [102]. While there is no evidence available that training for
the management of shoulder dystocia improves neonatal outcome [108],
it seems intuitive that ‘‘skill drills’’ would help increase preparedness of
all team members. Deering and colleagues [159] published a report in
which residents were block-randomized by year-group to a training ses-
sion on shoulder dystocia management that used an obstetric birthing
simulator or to a control group with no specific training. They found
that trained residents had significantly higher scores in all evaluation cat-
egories including timeliness of their interventions, performance of their
maneuvers, and overall performance. Crofts and colleagues [160] devel-
oped a mannequin for training and found that the management of shoul-
der dystocia improved following training. Specifically, they found
a reduction in the head-to-body delivery duration, and the maximum ap-
plied delivery force. However, these reductions in delivery duration and
applied force did not reach statistical significance.

Antenatal counseling
As there is no accurate method to predict which pregnancies will experi-
ence shoulder dystocia, antenatal counseling should be individualized for
each patient. Ideally, this should be an ongoing discussion throughout the
antenatal course and should include discussion of any history of shoulder
dystocia with or without birth injury, estimate of current fetal weight com-
pared with previous infants’ birth weights, gestational age, the presence of
maternal glucose intolerance and/or diabetes, and history of severe perineal
trauma with any subsequent incontinence. Depending on the results of that
discussion, a conversation regarding elective cesarean delivery, induction of
labor, expectant management, and operative vaginal delivery should take
place. Respecting a patient’s autonomy is of paramount importance and, ul-
timately, in the setting of history of (or significant risk factors for) shoulder
dystocia, either vaginal or cesarean delivery is a reasonable option.

Summary
Shoulder dystocia, in the final analysis, remains somewhat enigmatic. The
rarity of its incidence leads to many of the ancillary problems associated
with the event: the difficulty of arriving at a definition all practitioners
can accept, the inability to predict it, and the elusiveness of a univocal
management plan. Key factors in successfully managing shoulder dystocia
include constant preparedness, a team approach, and appropriate documen-
tation. Future directions include further research on accurate prediction of
macrosomia and regarding ‘‘skill drills’’ and training with birth simulators.
 SHOULDER DYSTOCIA: AN UPDATE 525

References
[1] Gross TL, Sokol RJ, Williams T, et al. Shoulder dystocia: a fetal-physician risk. Am J
Obstet Gynecol 1987;156:1408–18.
[2] Pollack RN, Buchman AS, Yaffe H, et al. Obstetrical brachial plexus palsy: pathogenesis,
risk factors, and prevention. Clin Obstet Gynecol 2000;43:236–46.
[3] American College of Obstetricians and Gynecologists. Shoulder dystocia. ACOG practice
bulletin clinical management guidelines for obstetrician-gynecologists. Number 40,
November 2002. Obstet Gynecol 2002;100:1045–50.
[4] Gemer O, Bergman M, Segal S. Labor abnormalities as a risk factor for shoulder dystocia.
Acta Obstet Gynecol Scand 1999;78:735–6.
[5] Gherman RB, Ouzounian JG, Goodwin TM. Obstetric maneuvers for shoulder dystocia
and associated fetal morbidity. Am J Obstet Gynecol 1998;178(6):1126–30.
[6] Stallings SP, Edwards RK, Johnson JWC. Correlation of head-to-body delivery intervals in
shoulder dystocia and umbilical artery acidosis. Am J Obstet Gynecol 2001;185:268–74.
[7] Caughey AB, Sandberg PL, Zlatnik MG, et al. Forceps compared with vacuum: rates of
neonatal and maternal morbidity. Obstet Gynecol 2005;106:908–12.
[8] Gherman RB, Ouzounian JG, Satin AJ, et al. A comparison of shoulder dystocia-
associated transient and permanent brachial plexus palsies. Obstet Gynecol 2003;102:
544–8.
[9] Robinson H, Tkatch S, Mayes DC, et al. Is maternal obesity a predictor of shoulder
dystocia? Obstet Gynecol 2003;101(1):24–7.
[10] Gherman RB, Chauhan S, Ouzounian JG, et al. Shoulder dystocia: the unpreventable
obstetric emergency with empiric management guidelines. Am J Obstet Gynecol 2006;
195:657–72.
[11] Gurewitsch ED, Johnson E, Hamzehzadeh S, et al. Risk factors for brachial plexus injury
with and without shoulder dystocia. Am J Obstet Gynecol 2006;194:486–92.
[12] Mehta SH, Bujold E, Blackwell SC, et al. Is abnormal labor associated with shoulder
dystocia in nulliparous women? Am J Obstet Gynecol 2004;10:1604–9.
[13] Mehta SH, Blackwell SC, Bujold E, et al. What factors are associated with neonatal injury
following shoulder dystocia? J Perinatol 2006;26:85–8.
[14] Cheng YW, Norwitz ER, Caughey AB. The relationship of fetal position and ethnicity with
shoulder dystocia and birth injury. Am J Obstet Gynecol 2006;195:856–62.
[15] McFarland MB, Langer O, Piper JM, et al. Perinatal outcome and the type and number of
maneuvers in shoulder dystocia. Int J Gynaecol Obstet 1996;55:219–24.
[16] Nassar AH, Usta IM, Khalil AM. Fetal macrosomia (R4500 g): perinatal outcome of 231
cases according to mode of delivery. J Perinatol 2003;23(2):136–41.
[17] Weeks JW, Pitman T, Spinnato II. Fetal macrosomia: Does antenatal prediction affect
delivery route and birth outcome? Am J Obstet Gynecol 1995;173(4):1215–9.
[18] Belfort MA, Dildy GA, Saade GR, et al. Prediction of shoulder dystocia using multivariate
analysis. Am J Perinatol 2007;24:5–10.
[19] Cohen BF, Penning S, Ansley D, et al. The incidence and severity of shoulder dystocia
correlates with a sonographic measurement of asymmetry in patients with diabetes. Am
J Perinatol 1999;16(4):197–201.
[20] Spong CY, Beall M, Rodrigues D, et al. An objective definition of shoulder dystocia:
prolonged head-to-body delivery intervals and/or the use of ancillary obstetric maneuvers.
Obstet Gynecol 1995;86:433–6.
[21] Beall MH, Spong C, McKay J, et al. Objective definition of shoulder dystocia: a prospective
evaluation. Am J Obstet Gynecol 1998;179(4):934–7.
[22] Gherman RB. Shoulder dystocia: an evidence-based evaluation of the obstetric nightmare.
Clin Obstet Gynecol 2002;45:345–62.
[23] Wolf H, Hoeksma AF, Oei SL, et al. Obstetric brachial plexus injury: risk factors related to
recovery. Eur J Obstet Gynecol Reprod Biol 2000;88:133–8.
526 GOTTLIEB & GALAN

[24] Nesbitt TS, Gilbert WM, Herrchen B. Shoulder dystocia and associated risk factors with
macrosomic infants born in California. Am J Obstet Gynecol 1998;179(2):476–80.
[25] Yeo GS, Lim YW, Yeong CT, et al. An analysis of risk factors for the prediction of shoulder
dystocia in 16,471 consecutive births. Ann Acad Med Singapore 1995;24:836–40.
[26] Mazouni C, Porcu G, Cohen-Solal E, et al. Maternal and anthropomorphic risk factors for
shoulder dystocia. Acta Obstet Gynecol Scand 2006;85:567–70.
[27] Sanchez-Ramos L, Bernstein S, Kaunitz AM. Expectant management versus labor induc-
tion for suspected fetal macrosomia: a systematic review. Obstet Gynecol 2002;100:
997–1002.
[28] Boulet SL, Salihu HM, Alexander GR. Mode of delivery and the survival of macrosomic
infants in the United States, 1995–1999. Birth 2006;33(4):278–83.
[29] Chauhan SP, Grobman WA, Gherman RA, et al. Suspicion and treatment of the macroso-
mic fetus: a review. Am J Obstet Gynecol 2004;193:332–46.
[30] Athukorala C, Crowther CA, Willson K, et al. Women with gestational diabetes mellitus in
the ACHOIS trial: risk factors for shoulder dystocia. Aust N Z J Obstet Gynaecol 2007;47:
37–41.
[31] Boulet SL, Alexander GR, Salihu HM. Secular trends in cesarean delivery rates
among macrosomic deliveries in the United States, 1989 to 2002. J Perinatol 2005;25:
569–76.
[32] Gonen R, Bader D, Ajami M. Effects of a policy of elective cesarean delivery in cases of
suspected fetal macrosomia on the incidence of brachial plexus injury and the rate of
cesarean delivery. Am J Obstet Gynecol 2000;183(5):1296–300.
[33] Mahony R, Walsh C, Foley ME, et al. Outcome of second delivery after prior macrosomic
infant in women with normal glucose tolerance. Obstet Gynecol 2006;107(4):857–62.
[34] Gonen R, Spiegel D, Abend M. Is macrosomia predictable, and are shoulder dystocia and
birth trauma preventable? Obstet Gynecol 1996;88:526–9.
[35] American College of Obstetricians and Gynecologists. Fetal macrosomia. ACOG practice
bulletin clinical management guidelines for obstetrician-gynecologists. Number 22.
Washington, DC. American College of Obstetricians and Gynecologists. 2000.
[36] Rydhstrom H, Ingemarsson I. The extremely large fetus: antenatal identification, risks and
proposed management. Acta Obstet Gynecol Scand 1989;68:59–63.
[37] Langer O, Berkus MD, Huff RW, et al. Shoulder dystocia: should the fetus weighing greater
than or equal to 4,000 grams be delivered by cesarean section? Am J Obstet Gynecol 1991;
165:831–7.
[38] Acker DB, Sachs BP, Friedman EA. Risk factors for shoulder dystocia in the average-
weight infant. Obstet Gynecol 1986;67:614–8.
[39] Okun N, Verma A, Mitchell BF, et al. Relative importance of maternal constitutional
factors and glucose intolerance of pregnancy in the development of newborn macrosomia.
J Matern Fetal Med 1997;6:285–90.
[40] Chauhan SP, Lutton PM, Bailey KJ, et al. Intrapartum clinical, sonographic, and parous
patients’ estimates of newborn birth weight. Obstet Gynecol 1992;79(6):956–8.
[41] Weiner Z, Ben-Schlomo I, Beck-Fruchter R, et al. Clinical and ultrasonographic weight
estimation in large for gestational age fetus. Eur J Obstet Gynecol Reprod Biol 2002;
105(1):20–4.
[42] Alsulyman OM, Ouzounian JG, Kjos SL. The accuracy of intrapartum ultrasono-
graphic fetal weight estimation in diabetic pregnancies. Am J Obstet Gynecol 1997;
177:503–6.
[43] Smith GC, Smith MF, McNay MB, et al. The relationship between fetal abdominal circum-
ference and birth weight: findings in 3512 pregnancies. Br J Obstet Gynaecol 1997;104:
186–90.
[44] McLaren RA, Puckett JL, Chauhan SP. Estimators of birth weight in pregnant women
requiring insulin: a comparison of seven sonographic modes. Obstet Gynecol 1995;85:
565–9.
 SHOULDER DYSTOCIA: AN UPDATE 527

[45] Klaij FAV, Geirsson RT, Nielsen H, et al. Humerospinous distance measurements: accu-
racy and usefulness for predicting shoulder dystocia in delivery at term. Ultrasound Obstet
Gynecol 1998;12:115–9.
[46] Verspyck E, Goffinet F, Hellot MF, et al. Newborn shoulder width: a prospective study of
2222 consecutive measurements. Br J Obstet Gynaecol 1999;106(6):589–93.
[47] Cohen B, Penning S, Major C, et al. Sonographic prediction of shoulder dystocia in infants
of diabetic mothers. Obstet Gynecol 1996;88:10–3.
[48] Elliott JP, Garite TJ, Freeman RK, et al. Ultrasonographic prediction of fetal macrosomia
in diabetic patients. Obstet Gynecol 1982;60(2):159–62.
[49] Winn HN, Holcomb W, Shmway JB, et al. The neonatal bisacromial diameter: a prenatal
sonographic evaluation. J Perinat Med 1997;25(6):484–7.
[50] Ren-Ing L, Fong-Ming C, Bor-Lin Y, et al. Predicting birth weight by fetal upper-arm
volume with use of three-dimensional ultrasonography. Am J Obstet Gynecol 1997;177:
632–8.
[51] Schild RL, Fimmers R, Hansmann M. Can 3D volumetric analysis of the fetal upper arm
and thigh improve conventional 2D weight estimates? Ultraschall Med 1999;20(1):31–7.
[52] Schild RL, Fimmers R, Hansmann M. Fetal weight estimation by three-dimensional ultra-
sound. Ultrasound Obstet Gynecol 2000;16(5):445–52.
[53] Poggi SH, Stallings SP, Ghidini A, et al. Intrapartum risk factors for permanent brachial
plexus injury. Am J Obstet Gynecol 2003;189:725–9.
[54] Lewis DF, Edwards MS, Asrat T, et al. Can shoulder dystocia be predicted? Preconceptive
and prenatal factors. J Reprod Med 1998;43(8):654–8.
[55] Dildy GA, Clark SL. Shoulder dystocia: risk identification. Clin Obstet Gynecol 2000;
43(2):265–82.
[56] Acker DB, Sachs BP, Friedman EA. Risk factors for shoulder dystocia. Obstet Gynecol
1985;66:762–8.
[57] Modanlu HD, Komatsu G, Dorchester W, et al. Large-for-gestational-age neonates:
anthropometric reasons for shoulder dystocia. Obstet Gynecol 1982;60:417–23.
[58] McFarland MB, Tryloich CG, Langer O. Anthropometric differences in macrosomic
infants of diabetic and nondiabetic mothers. J Matern Fetal Med 1998;7(6):292–5.
[59] Buchanan TA, Kjos SL, Montoro MN, et al. Use of fetal ultrasound to select metabolic
therapy for pregnancies complicated by mild gestational diabetes. Diabetes Care 1994;
17:275–83.
[60] Langer O, Rodriguez DA, Xenakis EM, et al. Intensified versus conventional management
of gestational diabetes. Am J Obstet Gynecol 1994;170:1036–46.
[61] Crowther CA, Hiller JE, Moss JR, et al. Australian Carbohydrate Intolerance Study in
pregnant women (ACHOIS) Trial Group. Effect of treatment of gestational diabetes
mellitus on pregnancy outcomes. N Engl J Med 2005;352:2477–86.
[62] Benedetti TJ, Gabbe SG. Shoulder dystocia. A complication of fetal macrosomia and
prolonged second stage of labor with midpelvic delivery. Obstet Gynecol 1978;52(5):
526–9.
[63] Baskett TF, Allen AC. Perinatal implications of shoulder dystocia. Obstet Gynecol 1995;
86:14–7.
[64] Bofill JA, Rust OA, Devidas M, et al. Shoulder dystocia and operative vaginal delivery.
J Matern Fetal Med 1997;6(4):220–4.
[65] Demissie K, Rhoads G, Smulian JC, et al. Operative vaginal delivery and neonatal
and infant adverse outcomes: population based retrospective analysis. BMJ 2004;329:
24–9.
[66] Ventura SJ, Martin JA, Curtin SC, et al. CDC’s national vital statistics reports: births: final
data for 1999. Natl Vital Stat Rep 2001;49:1–98.
[67] Gardella C, Taylor M, Benedetti T, et al. The effect of sequential use of vacuum and forceps
for assisted vaginal delivery on neonatal and maternal outcomes. Am J Obstet Gynecol
2001;185:896–902.
528 GOTTLIEB & GALAN

[68] Kjos SL, Henry OA, Montoro M, et al. Insulin-requiring diabetes in pregnancy: a random-
ized trial of active induction of labor and expectant management. Am J Obstet Gynecol
1993;169(3):611–5.
[69] O’Leary JA, Leonetti HB. Shoulder dystocia: prevention and treatment. Am J Obstet
Gynecol 1990;162:5–9.
[70] Ginsberg NA, Moisidis C. How to predict recurrent shoulder dystocia. Am J Obstet
Gynecol 2001;184:1427–30.
[71] Gherman RB. Shoulder dystocia: prevention and management. Obstet Gynecol Clin North
Am 2005;32:297–305.
[72] Gherman RB, Goodwin TM, Ouzounian JG, et al. Brachial plexus palsy associated with
cesarean section: an in utero injury? Am J Obstet Gynecol 1997;177(5):1162–4.
[73] McFarland M, Hod M, Piper JM, et al. Are labor abnormalities more common in shoulder
dystocia? Am J Obstet Gynecol 1995;173(4):1211–4.
[74] Lurie S, Levy R, Ben-Arie A, et al. Shoulder dystocia: could it be deduced from the labor
partogram? Am J Perinatol 1995;12(1):61–2.
[75] Spellacy WN, Miller S, Winegar A, et al. Macrosomia: maternal characteristics and infant
complications. Obstet Gynecol 1985;66:158–61.
[76] Hope P, Breslin S, Lamont L, et al. Fatal shoulder dystocia: a review of 56 cases reported to
the Confidential Enquiry into Stillbirths and Deaths in Infancy. Br J Obstet Gynaecol 1999;
105(12):1256–61.
[77] Perlow JH, Morgan MA, Montgomery D, et al. Perinatal outcome in pregnancy compli-
cated by massive obesity. Am J Obstet Gynecol 1992;167(4 Pt 1):958–62.
[78] Campbell MK, Ostbye T, Irgens LM. Post-term birth: risk factors and outcomes in a
10-year cohort of Norwegian births. Obstet Gynecol 1997;89:543–8.
[79] Johnstone NR. Shoulder dystocia: a study of 47 cases. Aust N Z J Obstet Gynaecol 1979;19:
28–31.
[80] Hassaan AA. Shoulder dystocia: risk factors and prevention. Aust N Z J Obstet Gynaecol
1988;28:107–9.
[81] El Madany AA, Jallad KB, Radi FA, et al. Shoulder dystocia: anticipation and outcome.
Int J Gynaecol Obstet 1990;34:7–12.
[82] Geary M, McParland P, Johnson H, et al. Shoulder dystocia: Is it predictable? Eur J Obstet
Gynecol Reprod Biol 1995;62:15–8.
[83] Bahar AM. Risk factors and fetal outcome in cases of shoulder dystocia compared with
normal deliveries of similar birthweight. Br J Obstet Gynaecol 1997;104(1):121–2.
[84] Rouse DJ, Owen J, Goldenberg RL, et al. The effectiveness and costs of elective cesarean
delivery for fetal macrosomia diagnosed by ultrasound. JAMA 1996;276(18):1480–6.
[85] Rouse DJ, Owen J. Prophylactic cesarean delivery for fetal macrosomia diagnosed by
means of ultrasonographyda Faustian bargain? Am J Obstet Gynecol 1999;181(2):
332–8.
[86] Gherman RB, Ouzounian JG, Goodwin TM. Brachial plexus palsy: an in utero injury? Am
J Obstet Gynecol 1999;180(5):1303–7.
[87] Paradiso G, Grañana N, Maza E. Prenatal brachial plexus paralysis. Neurology 1997;49(1):
261–2.
[88] Donnelly V, Foran A, Murphy J, et al. Neonatal brachial plexus palsy: an unpredictable
injury. Am J Obstet Gynecol 2002;187:1209–12.
[89] Sacks DA, Chen W. Estimating fetal weight in the management of macrosomia. Obstet
Gynecol Surv 2000;55(4):229–39.
[90] Combs CA, Singh NB, Khoury JC. Elective induction versus spontaneous labor after
sonographic diagnosis of fetal macrosomia. Obstet Gynecol 1993;81:492–6.
[91] Friesen CD, Miller AM, Rayburn WF. Influence of spontaneous or induced labor on
delivering the macrosomic fetus. Am J Perinatol 1995;12:63–6.
[92] Leaphart WL, Meyer MC, Capeless EL. Labor induction with a prenatal diagnosis of fetal
macrosomia. J Matern Fetal Med 1997;6:99–102.
 SHOULDER DYSTOCIA: AN UPDATE 529

[93] Gonen O, Rosen DJD, Dolfin Z, et al. Induction of labor versus expectant management in
macrosomia: a randomized study. Obstet Gynecol 1997;89:913–7.
[94] Herbst MA. Treatment of suspected fetal macrosomia: a cost-effective analysis. Am J
Obstet Gynecol 2005;193:1035–9.
[95] Lurie S, Insler V, Hagay Z. Induction of labor at 38–39 weeks of gestation reduces the
incidence of shoulder dystocia in gestational diabetic patients class A2. Am J Perinatol
1996;13:293–6.
[96] American College of Obstetricians and Gynecologists. Pregestational diabetes mellitus.
ACOG practice bulletin clinical management guidelines for obstetrician-gynecologists.
Number 60. Washington, DC. American College of Obstetricians and Gynecologists. 2005.
[97] Common deviations from accepted standards. In: Goldsmith LS, editor. Medical malprac-
tice: guide to medical issues. New York: Bender; 1990. p. 103.
[98] Screening the neonatal casesdapplication of medical and legal principles to specific cases.
In: Volk MD, Morgan MD, editors. Medical malpractice handling obstetric and neonatal
cases. Colorado Springs (CO): Shepard’s/McGraw Hill; 1994. p. 85–100.
[99] Hilty RB. Shoulder dystocia. In: Donn SM, Fisher CW, editors. Risk management
techniques in perinatal and neonatal practice. Armonk (NY): Future Publishing; 1996.
p. 311–21.
[100] Walters L. Evaluating the brachial plexus birth injury case. Trial 1998;34:65–8.
[101] Beall MH, Spong CY, Ross MG. A randomized controlled trial of prophylactic maneuvers
to reduce head-to-body delivery time in patients at risk for shoulder dystocia. Obstet
Gynecol 2003;102:31–5.
[102] Kwek K, Yeo GSH. Shoulder dystocia and injuries: prevention and management. Curr
Opin Obstet Gynecol 2006;18:123–8.
[103] Poggi SH, Allen RH, Patel CR, et al. Randomized trial of McRoberts versus lithotomy
positioning to decrease the force that is applied to the fetus during delivery. Am J Obstet
Gynecol 2004;191:874–8.
[104] Wood C, Ng KH, Hounslaw D, et al. Timedan important variable in normal delivery.
J Obstet Gynaecol Br Commonw 1973;80:295–300.
[105] Ouzounian JG, Korst LM, Ahn MO, et al. Shoulder dystocia and neonatal brain injury:
significance of the head-shoulder interval. Am J Obstet Gynecol 1998;S76:244.
[106] Allen RH, Rosenbaum TC, Ghidini A, et al. Correlating head-to-body delivery intervals
with neonatal depression in vaginal births that result in permanent brachial plexus injury.
Am J Obstet Gynecol 2002;187:839–42.
[107] Allen RH, Bankoski BR, Butzin CA, et al. Comparing clinician-applied loads for routine,
difficult and shoulder dystocia deliveries. Am J Obstet Gynecol 1994;171(6):1621–7.
[108] Gherman RB, Tramont J, Muffley P, et al. Analysis of McRoberts’ maneuver by x-ray
pelvimetry. Obstet Gynecol 2000;95:43–7.
[109] Gonik B, Zhang N, Grimm M. Prediction of brachial plexus stretching during shoul-
der dystocia using a computer simulation model. Am J Obstet Gynecol 2003;189(4):
1168–72.
[110] Buhimschi CS, Buhimschi IA, Malinow A, et al. Use of McRoberts’ position during deliv-
ery and increase in pushing efficiency. Lancet 2001;358:470–1.
[111] Gherman RB, Goodwin TM, Souter I, et al. The McRoberts’ maneuver for the alleviation
of shoulder dystocia: How successful is it? Am J Obstet Gynecol 1997;176:656–61.
[112] Allen R, Sorab J, Gonik B. Risk factors for shoulder dystocia: an engineering study of
clinician-applied forces. Obstet Gynecol 1991;77:352–5.
[113] Gherman RB, Ouzounian JG, Incerpi MH, et al. Symphyseal separation and transient
femoral neuropathy associated with McRoberts’ maneuver. Am J Obstet Gynecol 1998;
178(3):609–10.
[114] Gurewitsch ED, Donithan M, Stallings SP, et al. Episiotomy versus fetal manipulation in
managing severe shoulder dystocia: a comparison of outcomes. Am J Obstet Gynecol 2004;
191:911–6.
530 GOTTLIEB & GALAN

[115] Dandolu V, Jain NJ, Hernandez E, et al. Shoulder dystocia at noninstrumental vaginal
delivery. Am J Perinatol 2006;23(7):439–44.
[116] Ramsey PS, Rain KD, Field CS. Shoulder dystocia: rotational maneuvers revisited.
J Reprod Med 2000;45:85–8.
[117] Gurewitsch ED, Kim EJ, Yang JH, et al. Comparing McRoberts’ and Rubin’s maneuvers
for initial management of shoulder dystocia: an objective evaluation. Am J Obstet Gynecol
2005;192:153–60.
[118] Barnum CG. Dystocia due to the shoulders. Am J Obstet Gynecol 1945;50:439–42.
[119] Thompson KA, Satin AJ, Gherman RB. Spiral fracture of the radius: an unusual case of
shoulder dystocia-associated morbidity. Obstet Gynecol 2003;102:36–8.
[120] Poggi SH, Spong CY, Allen RH. Prioritizing posterior arm delivery during severe shoulder
dystocia. Obstet Gynecol 2003;101(5):1068–72.
[121] Bruner JP, Drummond SB, Meenan AL, et al. All-fours maneuver for reducing shoulder
dystocia during labor. J Reprod Med 1998;43(10):922–4.
[122] Kovavisarach E. The ‘‘all-fours’’ maneuver for the management of shoulder dystocia. Int
J Gynaecol Obstet 2006;95(2):153–4.
[123] Borell U, Fernstrom I. A pelvimetric method for assessment of pelvic mouldability. Acta
Radiol 1957;47:365–70.
[124] O’Leary JA. Cephalic replacement for shoulder dystocia: present status and future role of
Zavanelli maneuver. Obstet Gynecol 1993;82:847–50.
[125] Sandberg EC. The Zavanelli maneuver: 12 years of recorded experience. Obstet Gynecol
1999;93:312–7.
[126] Goodwin TM, Banks E, Lynnae K, et al. Catastrophic shoulder dystocia and emergency
symphystiotomy. Am Jour Obstet Gynecol 1997;177:463–4.
[127] Buhimschi CS, Buhimschi IA, Malinow AM, et al. The effect of fundal pressure manoeuvre
on intrauterine pressure in the second stage of labour. Br J Obstet Gynaecol 2002;109:
520–6.
[128] Hankins GD. Lower thoracic spinal cord injuryda severe complication of shoulder dysto-
cia. Am J Perinatol 1998;15(7):443–4.
[129] Iffy L, Varadi V. Cerebral palsy following cutting of the nuchal cord before delivery. Med
Law 1994;13:323–30.
[130] Flamm BL. Tight nuchal cord and shoulder dystocia: a potentially catastrophic combina-
tion. Obstet Gynecol 1999;94(5):853.
[131] Mavroforou A, Koumantakis E, Michalodimitrakis E. Physicians’ liability in obstetric and
gynecology practice. Med Law 2005;24:1–9.
[132] Acker DB. A shoulder dystocia intervention form. Obstet Gynecol 1991;78(1):150–1.
[133] Deering S, Poggi S, Hodor J, et al. Evaluation of residents’ delivery notes after a simulated
shoulder dytocia. Obstet Gynecol 2004;104:667–70.
[134] Hopwood HG. Shoulder dystocia: fifteen years’ experience in a community hospital. Am
J Obstet Gynecol 1982;144:162–6.
[135] Morrison JC, Sanders JR, Magann EF, et al. The diagnosis and management of dystocia of
the shoulder. Surg Gynecol Obstet 1992;175:515–22.
[136] al-Najashi S, al-Suleiman SA, el-Yahia A, et al. Shoulder dystociada clinical study of 56
cases. Aust N Z J Obstet Gynaecol 1989;29:129–32.
[137] Keller JD, Lopez-Zeno JA, Dooley SL, et al. Shoulder dystocia and birth trauma in gesta-
tional diabetes: a five year experience. Am J Obstet Gynecol 1991;165:928–30.
[138] Gonik B, Hollyer VL, Allen R. Shoulder dystocia recognition: differences in neonatal risks
for injury. Am J Perinatol 1991;8:31–4.
[139] McFarland LV, Raskin M, Daling JR, et al. Erb/Duchenne’s palsy: a consequence of fetal
macrosomia and method of delivery. Obstet Gynecol 1986;68:784–8.
[140] Ecker JL, Greenburg JA, Norwitz ER, et al. Birth weight as a predictor of brachial plexus
injury. Obstet Gynecol 1997;89:643–7.
 SHOULDER DYSTOCIA: AN UPDATE 531

[141] Perlow JH, Wigton T, Hart J, et al. Birth trauma. A five-year review of incidence and
associated perinatal factors. J Reprod Med 1996;41:754–60.
[142] Kolderup LB, Laros RK Jr, Musci TJ. Incidence of persistent birth injury in macrosomic
infants: association with mode of delivery. Am J Obstet Gynecol 1997;177:37–41.
[143] Benjamin K. Part 1. Injuries to brachial plexus: mechanisms of injury and identification of
risk factors. Adv Neonatal Care 2005;5(4):181–9.
[144] Kirjavainen M, Remes V, Peltonen J, et al. Long-term results of surgery for brachial plexus
birth palsy. J Bone Joint Surg 2007;89:18–26.
[145] Nocon JJ, McKenzie DK, Thomas LJ, et al. Shoulder dystocia: an analysis of risks and
obstetric maneuvers. Am J Obstet Gynecol 1993;168:1732–9.
[146] Gilbert WM, Nesbitt TS, Danielsen B. Associated factors in 1611 cases of brachial plexus
injury. Obstet Gynecol 1999;93:536–40.
[147] Graham EM, Forouzan I, Morgan MA. A retrospective analysis of Erb’s palsy cases and
their relation to birth weight and trauma at delivery. J Matern Fetal Med 1997;6:1–5.
[148] Cone LN. Mechanisms of brachial plexus lesions. Clin Neurol Neurosurg 1993;73:S24–9.
[149] Dunn DW, Engle WA. Brachial plexus palsy: intrauterine onset. Pediatr Neurol 1985;1(6):
367–9.
[150] Koenigsberger MR. Brachial plexus palsy at birth: intrauterine or due to delivery trauma?
Ann Neurol 1980;8:228.
[151] Mancias P, Slopis JM, Yeakley JW, et al. Combined brachial plexus injury and root
avulsion after complicated delivery. Muscle Nerve 1994;17:1237–8.
[152] Peterson GW, Bohr TW. Neonatal ‘‘obstetric’’ palsy, a ‘‘pre-existing condition?’’: two case
reports. Muscle Nerve 1995;18:1031.
[153] Nadas S, Reinberg O. Obstetric fractures. Eur J Pediatr Surg 1992;2:165–8.
[154] Turnpenny PD, Nimmo A. Fractured clavicle of the newborn in a population with a high
prevalence of grand-multiparity: analysis of 78 consecutive cases. Br J Obstet Gynaecol
1993;100(4):338–41.
[155] Hsu TY, Hung FC, Lu YJ, et al. Neonatal clavicular fracture: clinical analysis of incidence,
predisposing factors, diagnosis and outcome. Am J Perinatol 2002;19(1):17–21.
[156] Lewis DF, Raymond RC, Perkins MP, et al. Recurrence rate of shoulder dystocia. Am
J Obstet Gynecol 1995;172:1369–71.
[157] Modanlou HD, Dorchester WL, Thorosian A, et al. Macrosomiadmaternal, fetal, and
neonatal implications. Obstet Gynecol 1980;55:420–4.
[158] Goodwin TM, Banks E, Millar LK, et al. Catastrophic shoulder dystocia and emergency
symphysiotomy. Am J Obstet Gynecol 1997;177(2):463–4.
[159] Deering S, Poggi S, Macedonia C, et al. Improving resident competency in management of
shoulder dystocia with simulation training. Obstet Gynecol 2004;103:1224–8.
[160] Crofts JF, Attilakos G, Read M, et al. Shoulder dystocia training using a new training man-
nequin. BJOG 2005;112(7):997–9.
 Obstet Gynecol Clin N Am
34 (2007) 533–543

Diabetic Ketoacidosis in Pregnancy

Jason A. Parker, MDa, Deborah L. Conway, MDa,*
a
Department of Obstetrics and Gynecology, Division of Maternal-Fetal Medicine,
The University of Texas Health Science Center in San Antonio,
7703 Floyd Curl Drive, San Antonio, TX 78229, USA

Diabetic ketoacidosis (DKA) is a disease process involving numerous

pathophysiologic changes that can be markedly exaggerated in the pregnant
state. Episodes of DKA are infrequent in the general population and even
more so in pregnancy. Nonetheless, DKA may represent a life-threatening
event for the mother and her fetus. The incidence of DKA and factors unique
to pregnancy are discussed in this article, along with the effects of the disease
process on pregnancy. Clinical presentation, diagnosis, and treatment modal-
ities are covered in detail to offer data to improve maternal and fetal outcome.
The reported incidence of DKA outside of pregnancy ranges from 4.6 to
8 episodes per 1000 patients annually [1]. The overall incidence of DKA in
pregnancies complicated by diabetes is difficult to ascertain. Numerous re-
view articles and retrospective studies have found the incidence to range
from 1% to 10%, and the overall prevalence of DKA during pregnancy
and fetal loss associated with DKA have fallen significantly in recent years.
This trend is likely secondary to prenatal counseling (with a goal of optimal
glucose control before pregnancy) and improved understanding and man-
agement of the acute event. Cousins [2] reported the incidence of DKA dur-
ing pregnancy to be 9.3% in a group of 1508 patients studied between 1965
and 1985. More recent retrospective studies by Rodgers and Rodgers [3] and
Cullen and colleagues [4] found an incidence of DKA in pregnancy of 1% to
2%. In a case series by Kilvert and colleagues [5], the reported incidence of
DKA among 635 pregnant patients who had pregestational type 1 diabetes
mellitus was 1.73%. A larger, more recent case series by Schneider and col-
leagues [6] reported the incidence of DKA in pregnant patients to be 1.2%
among women who received insulin for diabetes control during their gesta-
tion. Four of the 11 patients who developed DKA on insulin therapy were

* Corresponding author. Department of Obstetrics and Gynecology, UTHSCSA, 7703

Floyd Curl Drive, San Antonio, TX 78229.
E-mail address: conway@uthscsa.edu (D.L. Conway).

0889-8545/07/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ogc.2007.08.001 obgyn.theclinics.com
534 PARKER & CONWAY

considered to have gestational diabetes mellitus, accounting for an incidence

of 0.45%. The occurrence of DKA in pregnancies complicated by gesta-
tional diabetes mellitus is rare [7], and when it is encountered, the possibility
of unrecognized pre-existing diabetes should be strongly considered. The
vast majority of cases of DKA occur in patients whose pregnancy is compli-
cated by pre-existing diabetes mellitus, particularly those who are prone to
DKA before pregnancy [8]. Although the use of insulin in a strict manner to
prevent occurrence has helped to lower the incidence of DKA in pregnancy,
it appears that a significant number of cases may occur in individuals who
have previously undiagnosed diabetes. In one study, 30% of cases of
DKA occurred in women who did not have known diabetes [9].
The overall incidence of fetal and maternal mortality secondary to DKA in
pregnancy is limited to data from case series. The true incidence of maternal
mortality is unknown, though it is historically reported as 5% to 15% [10].
Like the overall incidence of DKA in pregnancy, the maternal mortality ap-
pears to be declining. A study by Drury and colleagues [11] found the mater-
nal mortality rate to be less than 1% among 13 episodes of DKA experienced
in 600 consecutive pregnancies. Fetal loss rates, however, are much higher.
The study by Drury and colleagues [11] reported a fetal mortality rate of
85%. Montoro and colleagues [9] reported a 35% incidence of fetal demise
in women who had type 1 diabetes mellitus who presented with DKA. The
fetal mortality rate was even higher (57%) in the one third of patients for
whom the episode of DKA was their first diagnosis of diabetes. A more recent
study by Cullen and colleagues [4] found a lower fetal loss rate of 9%.
Specific adaptations that occur during pregnancy place the gravid dia-
betic woman at risk for episodes of DKA. Pregnancy is a relatively diabeto-
genic state. It is well known that overall insulin resistance and lipolysis are
increased during normal pregnancy; the increased lipolysis contributes to
the ‘‘accelerated starvation’’ and propensity toward ketone body formation
during pregnancy [12]. Hormones such as human placental lactogen (HPL),
growth hormone, prolactin, and progesterone play key roles in insulin sen-
sitivity and are discussed in further detail later. Another key adaptation of
pregnancy that contributes to a propensity to DKA involves the intricate
link between the renal and respiratory systems. Increases in minute ventila-
tion at the alveolar level place the pregnant woman in a state of respiratory
alkalosis. At the renal level, this is compensated for by increased excretion
of bicarbonate, a key metabolic buffer. This state of ‘‘compensated respira-
tory alkalosis’’ during pregnancy plays its role by decreasing the pregnant
woman’s ability to buffer ketone acids present in the serum during episodes
of DKA.

Pathophysiology
Proficiency and effectiveness in diagnosing and treating DKA necessitates
a thorough understanding of the pathophysiology that underlies this disease
 DIABETIC KETOACIDOSIS IN PREGNANCY 535

process. It should become clear from the following description that the path-
ophysiology of DKA feeds on itself: ‘‘the worse it gets, the worse it gets.’’ In
short, DKA is a state of inadequate insulin action (absolute lack, as in type
1 diabetes mellitus, or relative lack, as can occur in type 2 diabetes mellitus),
resulting in perceived hypoglycemia at the level of target cells (adipose, mus-
cle, and liver tissue). It is essential to keep in mind that most of the clinical
hallmarks of DKA (hyperglycemia, hypovolemia, ketosis, and acidosis) are
the result of an exaggerated counter-regulatory response to the perceived
hypoglycemia, which sets off a cascade effect that becomes apparent in the
clinical presentation and laboratory findings. Insulin counter-regulatory
hormones such as glucagon are released into the circulation in response to
cellular hypoglycemia, causing gluconeogenesis and glycogenolysis to be-
come disinhibited at the level of the liver. Therefore, the hyperglycemia
in DKA originates from three sources: (1) a high availability of glucose
precursors due to glucagon- and epinephrine-driven lipolysis (glycerol)
and muscle breakdown (amino acids); (2) a breakdown of glycogen stores;
and (3) a decreased peripheral uptake of glucose, caused by insulin lack
and made worse by increased counter-regulatory hormones. The increased
insensitivity to insulin results in decreased adipocyte storage of free fatty acids,
now present in the circulation in high amounts due to increased lipolysis.
These increased fatty acids undergo oxidation and are converted to ketoacids
by the liver (3-b-hydroxybutyrate and acetoacetate). The ketoacid acetoace-
tate may undergo decarboxylation and conversion to acetone, and can often
present clinically as a fruity odor from the patient’s breath [13]. The increased
levels of ketone bodies, combined with the buildup of lactic acid from periph-
eral hypoperfusion, result in the metabolic acidosis seen with DKA.
The intravascular hyperglycemia is just as important pathophysiologi-
cally as the intracellular hypoglycemia. High levels of glucose within the
circulation serve as an osmotic reservoir resulting in diuresis, leading to pro-
found hypovolemia and dehydration and further exacerbating the hypergly-
cemia and the acidosis. The ensuing hypovolemia stimulates the release of
other counter-regulatory stress hormones such as catecholamines, growth
hormone, and cortisol while enhancing the release of glucagon [14].
Some hormones that are increased during normal pregnancy have also
been found to play a role in the pathophysiology of DKA. HPL, which is
unique to pregnancy, serves as a counter-regulatory hormone for protection
against the hypoglycemic state. HPL can be seen in increased levels along
with glucagon in patients who have DKA. Prolactin is increased during
pregnancy and acts as a counter-regulatory hormone. The previously men-
tioned release of catecholamines, growth hormone, and cortisol along with
HPL and prolactin acts on insulin-sensitive tissues to produce alternative
substrates for energy use during DKA [13]. Like glucagon, these hormones
also serve to increase insulin resistance at the cellular level.
Electrolyte abnormalities are present in DKA and can be well understood
through pathophysiology. Serum sodium and potassium levels can become
536 PARKER & CONWAY

grossly abnormal during episodes of DKA as a result of the osmotic diure-

sis. Ketoacids also play a role in decreasing these serum electrolyte levels.
The electrolyte salts containing sodium and potassium may become bound
to anions from ketoacids and be excreted in the urine [13]. Potassium levels
tend to be high because of protein breakdown and inhibited entry of potas-
sium into cells due to insulin lack. Therefore, low potassium levels obtained
in the management of DKA indicate severe hypokalemia. Sodium levels
may be elevated, normal, or decreased. High levels suggest severe dehydra-
tion, whereas low levels may be real or a result of ‘‘pseudohyponatremia’’
associated with high serum levels of triglycerides.

How pregnancy affects this disease

When assessing a patient who has a particular medical illness during
pregnancy, one must question the effects that the pregnancy has on the dis-
ease process in particular and vice versa. Several factors are unique to preg-
nancy and to how the gravid state affects the development of DKA.
Pregnancy is a state of accelerated starvation. Glucose is readily absorbed
by the placenta and fetus as a source of energy. The metabolic hallmarks of
pregnancy include fasting hypoglycemia and hyperinsulinemia. These
states, in combination with constant acquisition of glucose by the fetus
and placenta, place the gravid patient at risk for cellular hypoglycemia.
Pregnancy also has a noted effect on insulin activity. Pregnancy is a rela-
tively insulin-resistant state, and this insulin resistance increases throughout
gestation. Several hormones such as HPL, cortisol, and prolactin are ele-
vated during pregnancy and serve to antagonize the effects of insulin at
the cellular level even further. Progesterone levels increase during preg-
nancy and serve to antagonize the effects of insulin by decreasing gastroin-
testinal motility, effectively resulting in the promotion of hyperglycemia
[15]. The respiratory changes in pregnancy previously mentioned lead to
a state of respiratory alkalosis. A compensatory increase in renal excretion
of bicarbonate results in a lower buffering capacity in the gravid state.
These changes make the pregnant diabetic patient more susceptible to
DKA by altering the ability to buffer ketoacids. This decreased buffering
capacity also places pregnant patients who have diabetes at risk for the de-
velopment of DKA at lower serum levels of hyperglycemia then those seen
in nonpregnant patients [14]. High levels of human chorionic gonadotropin
have been associated with emesis, placing a strain on the already hypogly-
cemic state of pregnancy. Dehydration from emesis leads to a resultant in-
crease in release of stress hormones that, as mentioned previously, have
insulin antagonistic effects. Due to the release and effect of these insulin
antagonistic hormones, any event that leads to stress at the physiologic
level places a pregnant patient who has diabetes at risk for the development
of DKA.
 DIABETIC KETOACIDOSIS IN PREGNANCY 537

How this disease affects pregnancy

One of the most profound effects that a disease can have on a pregnancy
outcome involves fetal loss. Current literature from the last decade supports
a fetal loss rate of approximately 9% [4,14]. Although the exact mechanism
remains unknown, several pathophysiologic aspects of DKA probably con-
tribute to fetal loss. Fetal status during correction of DKA is based on lim-
ited case reports demonstrating fetal heart rate tracings that are concerning
for fetal distress and on animal models of DKA [16,17]. Decreased uteropla-
cental blood flow almost certainly plays a major role. The massive osmotic
diuresis that takes place during DKA leads to intravascular volume deple-
tion, culminating in a decrease in and redistribution of maternal cardiac out-
put, resulting in decreased uteroplacental blood flow. Along with the
osmotic effects of DKA, maternal acidemia itself can reduce uteroplacental
perfusion, as can maternal cardiac arrhythmias resulting from potassium
disturbances [18].
Another proposed mechanism involves the fetal response to the meta-
bolic derangements of DKA. Glucose and ketoacids are readily transported
across the placenta to the fetus. Studies in ewes have demonstrated that in-
creased maternal ketoacids and hyperglycemia can lead to lactic acidosis
and hypoxia in the fetuses [19,20]. This acidosis is thought to occur by sev-
eral pathways: (1) fetal hyperglycemia resulting in fetal osmotic diuresis and
hypovolemia, contributing to lactic acidosis; (2) fetal hyperinsulinemia caus-
ing an increased fetal oxygen demand; and (3) increased affinity of maternal
hemoglobin for oxygen due to decreased 2,3-DPG levels, lowering the
amount of oxygen available to the fetus [21].
Limited literature exists regarding long-term outcomes on surviving
fetuses exposed to episodes of DKA in utero. Two studies suggest that there
may be an association between levels of ketoacids during pregnancy and
mental outcome. Stehbens and colleagues [22] showed an association be-
tween lower-level IQ scores and elevated ketoacid levels of pregnant patients
who had diabetes. Another study related specifically to maternal levels of
b-hydroxybutyrate during the third trimester showed an association with
decreased mental development scoring during the second year of life [23].
Although these findings are worrisome, their results should not be construed
as an indication to expedite delivery of a fetus before the woman who has
DKA is sufficiently stabilized and the pathophysiology is corrected.

Risk factors for diabetic ketoacidosis

Many studies have been performed on precipitating causes of DKA in
hopes to better recognize and prevent the onset of this disease process.
Rodgers and Rodgers [3] looked at variables associated with DKA in preg-
nancy by retrospectively reviewing admissions of affected patients over
a 10-year period. These data were then compared with the existing literature
538 PARKER & CONWAY

regarding DKA in pregnancy, for a total of 64 cases. The most common pre-
cipitating event was emesis from any cause, accounting for 42% of DKA
cases in their study. The second most common precipitating event was use
of b-sympathomimetics, and when combined with emesis, these events
accounted for 57% of episodes of DKA in this study. Other contributing
variables included infection, poor patient compliance, insulin pump failure,
undiagnosed diabetes, and physician management errors. A total of 80% of
DKA episodes in this study could be attributed to b-sympathomimetics,
emesis, poor compliance, and physician management errors. In a similar
study, Montoro and colleagues [9] found that poor patient compliance
was the most common variable inciting episodes of DKA. Cessation of in-
sulin use in the study population accounted for 35% of DKA episodes,
whereas infection accounted for 20%. Based on these studies, seven general
factors can be associated with precipitating the onset of DKA during preg-
nancy: emesis, infection, poor compliance/noncompliance, insulin pump
failure, use of b-sympathomimetics, use of corticosteroids, and poor physi-
cian management. Given these adverse effects of tocolysis with b-sympatho-
mimetics, magnesium sulfate is the recommended agent for tocolysis of
preterm labor in pregnancies complicated by diabetes or in the setting of
DKA. Corticosteroid therapy also poses a similar risk when administered
in an effort to increase pulmonary lung maturity and decrease intraventric-
ular hemorrhage in the anticipation of preterm delivery. Nonetheless, corti-
costeroids should not be withheld from women who have diabetes out of
fear of potential DKA. Rather, the physician should have concern and an-
ticipation for the onset of DKA (or worsening of its course) and plan ac-
cordingly. This anticipation of DKA may involve admitting a diabetic
woman who is to receive steroids to a unit in which frequent assessment
of maternal and fetal condition can be made and initiating an insulin drip
to control blood glucose levels.

Clinical presentation
DKA has classic clinical findings, none of which are pathognomonic of
the disease process but still raise a high level of suspicion for its presence.
The diagnosis of DKA is best made by ascertaining the patient’s symptoms
and findings on physical examination and by confirming the diagnosis with
laboratory studies. Patients suffering episodes of DKA generally present
with hyperventilation, altered mental status, weakness, dehydration, vomit-
ing, and polyuria. As previously discussed, the conversion of acetoacetate to
acetone by way of decarboxylation can lead to a fruity odor that is apparent
on the patient’s breath. Hyperventilation occurs as a response to ketoacids
in the body and is an attempt to decrease overall pH in the blood stream by
removing carbon dioxide by way of respiratory means. Altered mental sta-
tus is also an effect of the buildup of ketoacids and represents the effects an
 DIABETIC KETOACIDOSIS IN PREGNANCY 539

acidic environment has on the brain itself. Infrequently, the level of acidosis
can be so severe that patients may be completely obtunded. Vomiting, dehy-
dration, and polyuria are related to the osmotic diuresis that takes place
during episodes of DKA. Vomiting can be a response to this diuresis or,
as is discussed later, an inciting event.
The laboratory findings seen in DKA can be used to help confirm a correct
diagnosis of the disease. Findings of hyperglycemia, acidosis, and ketonemia
are generally seen in all cases of DKA [24]. Plasma glucose levels are usually
well over 300 mg/dL, but episodes of DKA in pregnancy can be seen at much
lower glucose levels in pregnancy. Blood glucose levels less than 200 mg/dL
have even been reported in some cases of DKA during pregnancy [4]. Acido-
sis is present and can be confirmed by arterial blood gas revealing a pH less
than 7.30. An anion gap is present along with this acidosis because the acido-
sis is caused by unmeasured anions: ketoacids and lactic acid. Arterial blood
gas findings also reveal an elevated base deficit that is consistent with a pri-
mary acidosis. Serum ketones and urine ketones are present in patients expe-
riencing episodes of DKA. Alterations in sodium and potassium levels can be
observed. Potassium may appear to be within normal limits on laboratory
results; however, it is likely that the total body potassium is decreased and
the patient is hypokalemic. Serum bicarbonate levels are decreased, often
to less than 15 mEq/L. Blood urea nitrogen and creatinine levels are elevated
due to dehydration and possibly renal failure. Phosphate levels may be
decreased as a result of binding to the anions of ketoacids in serum.

Treatment
The cornerstones of the treatment of DKA are aggressive fluid replace-
ment and insulin administration while ascertaining which precipitating fac-
tors brought about the current episode of DKA, and then treating
accordingly to mitigate those factors. The effects that DKA has on preg-
nancy make incorporating the mother and her fetus in the plan of care a ne-
cessity. Some of the fetal effects of DKA may be only transient and wholly
dependent on maternal condition, whereas maternal effects can be long-
standing depending on severity. Resolution and treatment of DKA in the
mother often leads to correction of the fetal physiologic response to the dis-
ease process. Except for the special circumstance of how to handle fetal sur-
veillance during an episode of DKA, it is helpful to keep in mind that
pregnancy itself does not alter the management of DKA. In other words,
recommendations for volume replacement and correction of hyperglycemia
and electrolyte disturbances are the same regardless of whether a person is
pregnant. Knowing this is helpful as we communicate with colleagues from
different disciplines in the care of these high-risk women.
Effective treatment of DKA in pregnancy requires an organized and mul-
tifaceted approach to correct physiologic abnormalities in the mother and
540 PARKER & CONWAY

secondarily in her fetus. Initial assessment regarding diagnosis of DKA

should be made promptly and an organized plan set in motion. This plan
should call on the talents and resources of a multidisciplinary team, which
can include the patient’s primary obstetric care provider, a perinatologist,
an intensive care specialist, an endocrinologist or general internist, and
skilled obstetric and intensive care nursing support.
Intravenous (IV) hydration with 0.9% sodium chloride is the recommen-
ded initial fluid replacement of choice. Marked hypovolemia can be assumed
with fluid deficits of at least 4 to 10 L. Estimating fluid deficit may be diffi-
cult, but Carroll and Yeomans [13] recommend calculating 100 mL/kg of
body weight when determining overall fluid deficit. Isotonic normal saline
should be administered as 1000 to 2000 mL per hour for 1 to 2 hours.
This aggressive administration has multiple effects. First, it immediately
increases tissue perfusion by increasing the markedly depleted intravascular
volume. Second, glucose values are decreased through hemodilution and
through increased renal loss of glucose when renal perfusion is improved.
After the first 1 to 2 hours, fluids are administered at a rate of 250 to 500
mL/h, with a long-term goal of correcting 75% of fluid deficit over
a 24-hour period [13]. Isotonic normal saline may be continued until glucose
values are less than 250 mg/dL, at which point administration of an IV so-
lution with 5% dextrose is begun. If hypernatremia develops during admin-
istration of isotonic normal saline, then 0.45% normal saline should be
administered. After hyponatremia is corrected, one may choose to adminis-
ter 0.45% normal saline during the remainder of volume resuscitation. An
indwelling bladder catheter should be placed to monitor hourly urine output
in response to treatment.
Along with aggressive fluid replacement, administration of insulin is
a requisite to correcting the disease process of DKA, by eliminating the per-
ceived intracellular hypoglycemia that drives the exaggerated counter-regu-
latory response. The amount of insulin required to correct and reverse the
process of DKA is large, and many algorithms exist for its dosing; however,
IV administration is preferred. Regular insulin should be administered IV as
a 0.1 U/kg bolus followed by 0.1 U/kg/h. This dosage should place the ini-
tial bolus and maintenance insulin level at about 10 U. If glucose levels do
not fall by 50 to 75 mg/dL over the first hour, then the hourly infusion rate
should be doubled.
Electrolyte abnormalities, particularly hypokalemia, are also addressed.
Every patient in whom DKA is suspected should have a comprehensive met-
abolic profile drawn for laboratory evaluation and repeated frequently when
the diagnosis is confirmed. Before administering potassium by way of IV,
adequate renal function must be established. If urine output is adequate af-
ter IV fluid administration has begun, then appropriate replacement of po-
tassium deficit may begin. Although serum levels of potassium may appear
normal, total body potassium is typically low. In addition, insulin adminis-
tration aimed at correction of hyperglycemia causes intracellular shifts of
 DIABETIC KETOACIDOSIS IN PREGNANCY 541

potassium, thus worsening an already present potassium deficit. Fluid ad-

ministration also serves as a sieve on potassium stores as ketoacids become
bound to potassium and are excreted in the urine. Potassium should be ad-
ministered IV and may be replaced with 20 mEq/L of potassium chloride
solution. Phosphate levels may appear abnormally low on serum chemistries
and may be replaced in conjunction with potassium by administering 10 to
20 mEq/L of potassium phosphate for each 10 to 20 mEq/L of potassium
chloride [25]. Anticipation of potassium deficit, on average, can be 5 to
10 mEq/kg of body weight. Serum potassium levels should be maintained
at around 4 to 5mEq/L.
The need for replacement of other electrolytes, including magnesium and
calcium, is debatable. Studies regarding replacement of these electrolytes
have not confirmed a specific benefit, and some investigators believe that se-
rum phosphate need not be replaced unless values fall below 1.0 mg/dL [26].
Much debate surrounds administration of bicarbonate for correction of
the metabolic acidosis in cases of DKA. Administration of bicarbonate
may cause a delay in the correction of ketoacids in the bloodstream [3].
There has also been a suggestion that administration of bicarbonate can
be associated with an ‘‘overshoot’’ alklosis or worsening acidemia secondary
to increased partial pressure of carbon dioxide (PCO2) [26]. On the other
hand, fetal benefits may come from maternal bicarbonate replacement.
According to Lobue and Goodlin [17], administration of bicarbonate to
the mother resulted in resolution of fetal heart rate abnormalities including
late decelerations and absent beat-to-beat variability. It is concerning, how-
ever, that rapid correction of maternal pH and PCO2 levels with bicarbonate
administration could lead to elevated fetal levels of PCO2, thus having a det-
rimental effect on the fetal ability to maintain adequate oxygen transfer [13].
Other investigators suggest replacement of bicarbonate only in cases of
severe acidosis during which pH is less than 5.0 to 7.0 mEq/L [14,25,27].
Several pitfalls exist during the treatment of DKA that the physician
should be aware of. A common error is to discontinue or decrease volume
therapy inappropriately after glucose levels normalize. Acidemia may still
be present despite correction of hyperglycemia, and restoration of circulat-
ing volume is critical to its resolution. Without continued and adequate fluid
replacement, the possibility for recurrence of DKA exists. To avoid this mis-
take, the estimated fluid deficit should be calculated based on body weight,
as described earlier, and fluid replacement should be continued until this cal-
culated deficit has been remedied.
The desire to discontinue insulin therapy may also exist after hyperglyce-
mia has been corrected, but discontinuation also could result in worsening
or continued presence of DKA. It should be remembered that it is the intra-
cellular hypoglycemia, not the serum level of glucose, that determines the
level of counter-regulatory hormone activity and that this intracellular hy-
poglycemia is what drives the DKA process. Correction of the acidemia
present in DKA takes much longer than correction of hyperglycemia;
542 PARKER & CONWAY

therefore, insulin should be continued at a basal infusion rate of 1 to 2 U/h

after normoglycemia is established [28]. Furthermore, early discontinuation
of IV insulin in favor of intermittent subcutaneous injections may prolong
complete resolution of DKA. IV insulin therapy should not be discontinued
until after the first subcutaneous dose of regular insulin is administered [28].
Electronic fetal heart rate monitoring is recommended for gestational
ages greater than 24 weeks [14]. Fetal heart rate abnormalities, however,
are to be expected in the acute phases of an episode of DKA. It is critical
that no intervention on fetal behalf occur unless the mother’s condition is
stable enough to withstand the rigors of delivery, particularly by cesarean
section. Maternal mortality can be the result of operative intervention
before full stabilization of the mother.

Summary
It is fortunate that episodes of DKA are rare in pregnancy. When pres-
ent, however, DKA can represent a life-threatening emergency for mother
and fetus. Most cases of DKA occur in patients who have diabetes existing
before pregnancy. Several adaptations place the gravid patient who has di-
abetes at risk for development of DKA. The obstetrician must be aware of
several precipitating events that can serve as a catalyst for the onset of
DKA. No substitute exists for adequate history and physical examination
in the diagnosis of DKA, and subsequent confirmation can be obtained
with the hallmark laboratory findings of hyperglycemia, acidosis, and keto-
nemia. Treatment involves aggressive fluid management, insulin administra-
tion, and the identification and treatment of precipitating causes. Care
should be taken to stabilize and treat the mother first because most fetal
heart rate abnormalities subside after correction of maternal hypovolemia
and acidosis.

Acknowledgments
The authors would like to thank Dr. Ashley Parker for her assistance
with reviewing and summarizing the literature referenced in this article.

References
[1] Kitabchi AE, Umpierrez GE, Murphy MB, et al. Management of hyperglycemic crises
in patients with diabetes. Diabetes Care 2001;24:131–53.
[2] Cousins L. Pregnancy complications among diabetic women: review, 1965-1985. Obstet
Gynecol Surv 1987;42:140–9.
[3] Rodgers BD, Rodgers DE. Clinical variables associated with diabetic ketoacidosis during
pregnancy. J Reprod Med 1991;36:797–800.
[4] Cullen MT, Reece EA, Homko CJ, et al. The changing presentations of diabetic ketoacidosis
during pregnancy. Am J Perinatol 1996;13:449–51.
 DIABETIC KETOACIDOSIS IN PREGNANCY 543

[5] Kilvert J, Nicholson HO, Wright AD. Ketoacidosis in diabetic pregnancy. Diabet Med 1993;
10:278–81.
[6] Schneider M, Umpierrez GE, Ramsey RD, et al. Pregnancy complicated by diabetic ketoa-
cidosis: maternal and fetal outcomes. Diabetes Care 2003;26:958–9.
[7] Pitteloud N, Binz K, Caulfield A, et al. Ketoacidosis during gestational diabetes: case report.
Diabetes Care 1998;21:1031–2.
[8] Kent LA, Gall GV, Williams G. Mortality and outcome of patients with brittle diabetes and
recurrent ketoacidosis. Lancet 1994;344:778–81.
[9] Montoro MN, Myers VP, Mestman JH, et al. Outcome of pregnancy in diabetic ketoacido-
sis. Am J Perinatol 1993;10:17–20.
[10] Gabbe SG, Mestman JH, Hibbard LT. Maternal mortality in diabetes mellitus: an 18-year
survey. Obstet Gynecol 1976;48:549–51.
[11] Drury MI, Greene AT, Stronge JM. Pregnancy complicated by clinical diabetes mellitus:
a study of 600 pregnancies. Obstet Gynecol 1977;49:519–22.
[12] Laffel L. Ketone bodies: a review of physiology, pathophysiology and application of moni-
toring to diabetes. Diabetes Metab Res Rev 1999;15:412–26.
[13] Carroll M, Yeomans ER. Diabetic ketoacidosis in pregnancy. Crit Care Med 2005;33:
S347–53.
[14] Moore TR. Diabetes in pregnancy. In: Creasy RK, Resnik R, Iams JD, editors. Maternal-
fetal medicine. 5th edition. Philadelphia: Saunders; 2004. p. 1031–2.
[15] Kamalakannan D, Baskar V, Barton DM, et al. Diabetic ketoacidosis in pregnancy. Post-
grad Med J 2003;79:454–7.
[16] Hughes AB. Fetal heart rate changes during diabetic ketosis. Acta Obstet Gynecol Scand
1987;66:71–3.
[17] Lobue C, Goodlin RC. Treatment of fetal distress during diabetic keto-acidosis. J Reprod
Med 1978;20:101–4.
[18] Bard H, Fouron JC, Demuylder X, et al. Myocardial function and hemoglobin oxygen affin-
ity during hyperglycemia in the fetal lamb. J Clin Invest 1986;78:191–5.
[19] Miodovnik M, Skillman CA, Hertzberg V, et al. Effect of maternal hyperketonemia in
hyperglycemic pregnant ewes and their fetuses. Obstet Gynecol 1986;154:394–401.
[20] Miodovnik M, Lavin JP, Harrington DJ, et al. Cardiovascular and biochemical effects of
infusion of beta hydroxybutyrate into the fetal lamb. Obstet Gynecol 1982;144:594–600.
[21] Ditzel J, Standl E. The oxygen transport system of red blood cells during diabetic ketoacido-
sis and recovery. Diabetologia 1975;11:255–60.
[22] Stehbens J, Baker GL, Kitchell M. Outcome at ages 1, 3, and 5 years of children born to
diabetic women. Am J Obstet Gynecol 1977;127:408–13.
[23] Rizzo T, Metzger BE, Burns WJ, et al. Correlations between antepartum maternal metabo-
lism and child intelligence. N Engl J Med 1991;325:911–6.
[24] Chauhan SP, Perry KG Jr. Management of diabetic ketoacidosis in the obstetric patient.
Obstet Gynecol Clin North Am 1995;22:143–55.
[25] Ramin KD. Diabetic ketoacidosis in pregnancy. Obstet Gynecol Clin North Am 1999;26:
481–9.
[26] Adrogue HJ, Madias NE. Management of life-threatening acid-base disorders: first of two
parts. N Engl J Med 1998;338:26–34.
[27] Winkler C, Lowell D, et al. Endocrine emergencies. In: Dildy G, Belfort M, Saade G, editors.
Critical care obstetrics. 4th edition. Malden (MA): Blackwell; 2004. p. 420–35.
[28] Whiteman V, Homko CJ, Reece EA. Management of hypoglycemia and diabetic ketoacido-
sis in pregnancy. Obstet Gynecol Clin North Am 1996;23:87–107.
 Obstet Gynecol Clin N Am
34 (2007) 545–553

Amniotic Fluid Embolism

Irene Stafford, MD*, Jeanne Sheffield, MD
Department of Obstetrics & Gynecology, University of Texas Southwestern
Medical Center, 5323 Harry Hines Blvd., Dallas, TX 75390-9032, USA

Amniotic fluid embolism (AFE) is a catastrophic syndrome occurring

during labor and delivery or immediately postpartum. Although presenting
symptoms may vary, common clinical features include shortness of breath,
altered mental status followed by sudden cardiovascular collapse, dissemi-
nated intravascular coagulation (DIC), and maternal death. It was first rec-
ognized as a syndrome in 1941, when two investigators described fetal mucin
and squamous cells during postmortem examination of the pulmonary vas-
culature in women who had unexplained obstetric deaths [1]. Since then,
many studies, case reports, and series have been published in an attempt
to elucidate the etiology, risk factors, and pathogenesis of this mysterious
obstetric complication.
The incidence of AFE has been reported in the range of 1 in 8000 to
80,000 deliveries [2]. Two recent large population-based cohort studies
have demonstrated the rate of AFE to be 14.8 and 6.0 per 100,000 in mul-
tiparous and primigravid deliveries, respectively [3,4]. The true incidence is
unclear because this syndrome is difficult to identify and the diagnosis re-
mains one of exclusion, with possible under-reporting of nonfatal cases.
There have also been discrepancies in the published maternal mortality rates
associated with AFE. In a well-defined United States national registry exam-
ining 46 cases of AFE within a 5-year span, maternal mortality rates were
reported at 61%, with a neurologically intact maternal survival rate of
15% [5]. Investigators from the United Kingdom report a maternal mortal-
ity rate of 37% in their registry of AFE, with 93% of survivors remaining
neurologically intact [6]. Other retrospective studies reporting from data-
bases derived from hospital charts have reported maternal mortality rates
between 13% and 26%, with normal maternal outcome in 87% of survivors
[3,4]. Fetal outcome is poor when AFE occurs before delivery. The fetal

* Corresponding author.
E-mail address: istaff@parknet.pmh.org (I. Stafford).

0889-8545/07/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ogc.2007.08.002 obgyn.theclinics.com
546 STAFFORD & SHEFFIELD

survival rate approaches 40%, though with 29% and 50% of surviving
neonates developing neurologic abnormalities [5,6].
Although the United States national registry did not find any maternal
demographic risk factors for AFE, they found that 70% of cases occurred
during labor, 19% were recorded during cesarean section, and 11% of cases
occurred immediately following vaginal delivery [5]. Other studies have also
found an increased frequency of AFE in women who underwent cesarean
delivery, with rates between 20% and 60% [4,7]. Approximately 50% of
these cases were associated with fetal distress, suggesting that amniotic fluid
embolus and associated hypoxia preceded cesarean delivery. This interpreta-
tion is supported by studies from the United Kingdom in which only one of
the five cesarean deliveries in the registry was performed before the diagno-
sis of AFE [6]. Rupture of membranes was a consistent finding among 78%
of women in the United States AFE registry, with onset of symptoms occur-
ring within 3 minutes of amniotomy in 11% of cases [5]. Another study
found maternal age (mean age, 33 years) and multiparity (mean parity,
2.6) to be associated with AFE [4]. Conflicting data have been reported
on multiple gestation. The frequency of twin gestation in the national
AFE registry was not increased from baseline population estimates but
found to be approximately threefold higher in one retrospective analysis [4].
In the large cohort study examining the association between AFE and the
induction of labor along with other risk factors, AFE was found in twice as
many women who underwent medical induction of labor. This association
was even stronger for fatal cases (odds ratio, 3.5). Increased rates of AFE
were also found in women who had placenta previa, placental abruption,
cervical lacerations, or uterine rupture and in women who underwent oper-
ative vaginal delivery [3]. Although eclampsia was also strongly associated
with AFE in this study, no risk factor has been consistently substantiated
in the literature.

Etiology and pathogenesis

The mechanism of disease for AFE is poorly understood. Early studies
describe the histologic presence of amniotic fluid components in lung tissue
during postmortem examination in obstetric patients who had unexplained
death [1]. This finding was followed by reports of amniotic fluid debris
found in maternal circulation in fatal and nonfatal cases of AFE [8,9]. Con-
ventional wisdom describes the efflux of amniotic fluid components into ma-
ternal vasculature as driven by a pressure or electrochemical gradient by
way of lacerations in the lower uterine segment, endocervical vessels, and
placental site [9]. Plugging of the cervical vasculature by amniotic fluid ele-
ments has been described, although the mechanism by which this leads to
AFE is unclear. In addition, elements of amniotic fluid have been isolated
in blood and sputum of pregnant women who did not have clinical evidence
of AFE [10,11].
 AMNIOTIC FLUID EMBOLISM 547

Amniotic fluid contains various concentrations of fetal squamous epithe-

lial cells, lanugo hair, vernix, mucin, zinc coproporphyrin, prostaglandins,
and platelet activating factor. One possible mechanism of disease includes
the effect of direct procoagulants found in amniotic fluid on maternal sys-
tems. The presence of vasoactive substances, such as platelet activating fac-
tor, in the placenta and amniotic fluid has been shown to cause increased
vascular permeability; bronchoconstriction; platelet aggregation; recruit-
ment of leukotrienes, cytokines, and thromboxanes; and the cascade of
prostaglandin production [12]. In one small study examining the effect of au-
tologous fetal membranes on the coagulation profile in pigs, findings were
significant for decreased platelets, fibrinogen, and antithrombin III.
Although these laboratory abnormalities are consistent with AFE, the syn-
drome of AFE could not be elicited in this study [13]. Similar studies involv-
ing primates also failed to model the syndrome despite procoagulant effects
of autologous amniotic fluid [14–16]. Currently, there is no suitable animal
model for amniotic fluid embolus secondary to the limitations of autologous
amniotic fluid.
Laboratory testing for the fetal antigen sialyl Tn has shown some diag-
nostic value with AFE [17–19]. Sialyl Tn is a fetal antigen present in meco-
nium and amniotic fluid detected most accurately with the TKH-2
monoclonal antibody [19,20]. In a small Japanese case series, seven of nine
women who had the diagnosis of AFE had elevated serum levels of fetal an-
tigen compared with control subjects. In addition, special imunohistochem-
ical stains for the presence of fetal antigen in lung tissue were positive in
women who had a history of AFE [18]. An anaphylactic or complement ac-
tivation reaction to sialyl Tn may explain the mechanism of disease. In one
small series, complement activation was found along with high levels of
sialyl Tn. Levels of complement C3 and C4 were twofold to threefold lower
than normal [17]. When these markers were used for evaluation of anesthe-
sia-induced allergic anaphylaxis, however, similar results were found [17,21].
An alternative immunologic mechanism for AFE involves the possibility
of anaphylaxis with massive mast cell degranulation, independent of anti-
gen-antibody–mediated classic anaphylaxis. In early studies, immunohisto-
chemical staining in postmortem cases of AFE revealed elevated numbers
of mast cells in the pulmonary vasculature [22]. Tryptase has been examined
as a factor involved in anaphylaxis because it is specific to mast cells and has
a longer half-life than histamine. In one study using serum tryptase and uri-
nary histamine concentrations as markers for mast cell degranulation, no
difference was found between women who had a history of AFE compared
with control subjects [17]. Other investigators, however, found elevated
tryptase levels in women who had AFE, but these values were compared
with nonpregnant control subjects [23,24]. Of note, in some cases when com-
plement is involved in classic antibody-antigen anaphylaxis, mast cell
degranulation can occur [25]. The studies evaluating serum tryptase levels
in AFE cases did not simultaneously measure complement levels [23,24].
548 STAFFORD & SHEFFIELD

Clinical presentation
Although AFE typically occurs during labor and delivery or immediately
postpartum, rare cases of AFE have been reported after midtrimester termi-
nation, transabdominal amniocentesis, trauma, and saline amnioinfusion
[26–30]. Classic presenting symptoms of AFE include respiratory distress,
altered mental status, profound hypotension, coagulopathy, and death [2].
Historical studies have described the presenting symptom as primarily respi-
ratory distress, whereas other studies describe the most common presenting
symptom before delivery to be altered mental status. Seizure or seizure-like
activity was reported as the initial symptom in 30% of patients involved in
the United States national registry, followed by dyspnea (27%), fetal brady-
cardia (17%), and hypotension (13%) [5]. Over 50% of postpartum patients
who had AFE presented with postpartum hemorrhage secondary to coagul-
opathy [5]. Other signs and symptoms include nausea, vomiting, fever, chills,
and headache. Diagnostic criteria used for the United States and the United
Kingdom registries for AFE are listed in Box 1.
Due to the vast overlap of the symptomatology of AFE with other dis-
ease states, consideration for the differential diagnosis of AFE is warranted.
A differential diagnosis for possible AFE is shown in Box 2.
Clinical features of AFE include profound cardiovascular changes.
According to the United States national registry, all patients who had
AFE experienced hypotension. Most women (93%) had some level of pul-
monary edema or adult respiratory distress syndrome along with hypoxia
[5]. One explanation for these findings includes the possibility of severe
bronchospasm related to the presence of fetal elements in the pulmonary
vasculature; however, only 15% of patients were found to have broncho-
spasm [5]. Transesophageal echocardiograpy and pulmonary artery cathe-
ters have demonstrated transiently elevated pulmonary artery pressures in
cases of AFE along with left ventricular dysfunction, supporting the notion
that these pulmonary findings are consistent with cardiogenic shock. There
have also been reports of isolated right ventricular dysfunction with high

Box 1. Diagnostic criteria for amniotic fluid embolism

Acute hypotension and/or cardiac arrest
Acute hypoxia diagnosed by dyspnea, cyanosis, and/or
respiratory arrest
Coagulopathy or severe clinical hemorrhage in the absence of
other explanations
All of these occurring during labor, cesarean delivery, or dilation
and evacuation or within 30 minutes postpartum with no other
explanation for the findings
 AMNIOTIC FLUID EMBOLISM 549

Box 2. Differential diagnosis for women presenting

with possible amniotic fluid embolism
Pulmonary thromboembolism
Transfusion reaction
Hemorrhage
Air embolism
Anaphylaxis
High spinal anesthesia
Placental abruption
Peripartum cardiomyopathy
Eclampsia
Myocardial infarction
Septic shock
Uterine rupture

right-sided pressures and tricuspid regurgitation [14,31–37]. In the United

States registry, all but two patients experienced cardiac arrest or serious car-
diac arrhythmia, with 50% of these events occurring within 5 minutes of
symptom onset [5]. Myocardial hypoxic injury may be related to decreased
cardiac output and impaired filling, resulting in decreased coronary artery
perfusion. Dilation of the right ventricle with subsequent leftward displace-
ment of the interventricular septum may also contribute to myocardial dys-
function [31]. Initially, pulmonary and systemic pressures may be elevated.
Although the etiology of these changes is unclear, small studies have reported
vasoconstrictive effects of amniotic fluid in animal models [33,38]. This vaso-
constriction is often followed by profound hypotension and shock, most
likely resulting from cardiogenic or obstructive causes as described earlier.
After initial survival, hypoxia relates more to noncardiogenic shock,
whereby severe alveolar-capillary membrane leak leads to increase pulmo-
nary edema and decreased oxygenation [14]. In the presence of DIC, hem-
orrhagic shock may further complicate the management of the patient
who has AFE.
DIC is a common feature of AFE. According to the United States regis-
try for AFE, 83% of patients demonstrated laboratory abnormalities or
clinical findings consistent with DIC, regardless of mode of delivery. Onset
was variable, with 50% of cases occurring within 4 hours of presentation,
often within 20 to 30 minutes of symptom onset [5]. The presence of clotting
factors in amniotic fluid has been linked with the possible activation of the
clotting cascade in the pulmonary vasculature of affected women [38,39].
Additional data report that increased levels of plasminogen activator inhib-
itor–1 antigen in amniotic fluid may become active in maternal circulation,
leading to consumptive coagulopathy [40]. Within the national registry,
550 STAFFORD & SHEFFIELD

75% of patients who presented with hemorrhage and isolated coagulopathy

died despite appropriate aggressive management.

Management
Currently, there are no proven laboratory tests that confirm the diagnosis
of AFE. Most events occur in an unpredictable manner and have variable
presentation. The initial management goal includes rapid maternal cardio-
pulmonary stabilization with prevention of hypoxia and maintenance of
vascular perfusion. In cases of refractory hypotension, vasopressors may
be necessary. Central monitoring for cardiovascular status may assist in
these endeavors. Eighty-seven percent of patients in the national AFE
registry suffered cardiac arrest. Of these, 40% occurred within 5 minutes
from symptom onset. The most common dysrhythmia was found to be elec-
trochemical dissociation, followed by bradycardia and ventricular tachy-
cardia or fibrillation [5]. Ionotropes may need to be added to improve
myocardial function. Initial laboratory data should include complete blood
count, arterial blood gas, electrolytes, and a coagulation profile. A tryptase
level is available at some hospitals, in addition to TKH-2 monoclonal anti-
body to fetal mucin. With or without evidence of hemorrhage as a presenting
symptom, blood products should be ordered expeditiously in anticipation of
profound bleeding and DIC. Uterine artery embolization and recombinant
factor VII have been used in cases of severe coagulopathy resistant to con-
ventional blood and product replacement [41–43].
Transthoracic or transesophageal echocardiography is often necessary to
evaluate cardiac function and to guide treatment, along with a 12-lead ECG.
When ischemia or infarction is suspected, cardiac isoenzymes and troponins
should be obtained. A chest radiograph should be ordered to evaluate the
possibility of pulmonary edema and cardiac enlargement. Diuretics may
be used with caution for pulmonary edema.
Other case reports have described the use of continuous hemodiafiltra-
tion, extracorporeal membrane oxygenation, and intra-aortic balloon coun-
terpulsation in cases of AFE [44–46]. In one report, early transesophageal
echocardiogram demonstrating severe pulmonary vasoconstriction and cor
pulmonale led to successful rescue using cardiopulmonary bypass [45].
According to the national registry, 70% of patients were in labor when
AFE occurred. When fetuses are undelivered, the fetal mortality rate ap-
proaches 20% [47]. Of the surviving fetuses recorded in the registries,
30% were severely acidotic, with a 12% perinatal mortality rate [5,6]. In
cases of cardiac arrest, administration of all cardiac support measures,
including medications used in resuscitation, should be without delay. The
patient can be placed in the left lateral decubitus position before chest com-
pressions to avoid compression of the inferior vena cava by the gravid
uterus. In cases in which asystole or malignant arrhythmia is present for
greater than 4 minutes, perimortum cesarean delivery should be considered
 AMNIOTIC FLUID EMBOLISM 551

[48]. Uterine evacuation after unsuccessful resuscitation may not only be

therapeutic for the mother but also improve neonatal outcome [48,49].
Intact fetal survival has been shown to be possible when delivery is accom-
plished within 5 minutes of maternal cardiac arrest [48].
Significant maternal morbidity is associated with AFE. Over 75% of pa-
tients in the United Kingdom registry required intensive care management,
with an average length of stay of 5.2  9.7 days among survivors. An aver-
age of 34 U of blood products was required in these patients [6]. In the
United States AFE registry, only 15% of patients who had cardiac arrest
survived neurologically intact [5]. Other sequelae include liver hematoma,
renal and multisystem failure, and ischemic encephalopathy. There are no
data to support recurrence risk for subsequent pregnancies in women who
survive [37].
Overall morbidity and mortality of AFE has improved with early recog-
nition of the syndrome and improved resuscitative efforts involving multiple
disciplines of medicine. In cases recorded within the United Kingdom regis-
try, women who survived AFE had a shorter time frame between symptom
onset and treatment (41.5 minutes versus 108 minutes) [6,50].
Although there are many new research developments in this field, the eti-
ology and the pathogenesis of AFE remain unclear. Currently, there is no
‘‘gold standard’’ diagnostic test. AFE remains a diagnosis of exclusion de-
pendent on rapid bedside evaluation and judgment. Ideal management in-
cludes prompt evaluation of and intervention for each of the pathologic
events found in this complex obstetric condition.

References
[1] Steiner PE, Lushbaugh C. Maternal pulmonary embolism by amniotic fluid as a cause
of obstetric shock and unexplained death in obstetrics. JAMA 1941;117:1245–54.
[2] Morgan M. Amniotic fluid embolism. Anaesthesia 1979;34:20–32.
[3] Kramer MS, Rouleau J, Baskett TF, et al. Amniotic-fluid embolism and medical induction
of labour: a retrospective, population-based cohort study. Lancet 2006;368(9545):1444–8.
[4] Gilbert WM, Danielson B. Amniotic fluid embolism: decreased mortality in a population-
based study. Obstet Gynecol 1999;93(6):973–7.
[5] Clark SL, Hankins GDV, Dudley DA, et al. Amniotic fluid embolism: analysis of the
national registery. Am J Obstet Gynecol 1995;172:1158–69.
[6] Tuffnell DJ. United Kingdom amniotic fluid embolism register. BJOG 2005;112(12):1625–9.
[7] Lau G, Chui PP. Amniotic fluid embolism: a review of 10 fatal cases. Singapore Med J 1994;
35:180–3.
[8] Gross PBE. Pulmonary embolism by amniotic fluid: report of three cases with a new diag-
nostic procedure. Surg Gynecol Obstet 1947;85:315–20.
[9] Resnik R, Swartz WH, Plummer MH, et al. Amniotic fluid embolism with survival. Obstet
Gynecol 1976;47:295–8.
[10] Clark SL, Pavlova Z, Greenspoon J, et al. Squamous cells in the maternal pulmonary circu-
lation. Am J Obstet Gynecol 1986;154:104–6.
[11] Lee W, Ginsburg KA, Cotton DB, et al. Squamous and trophoblastic cells in the maternal
pulmonary circulation identified by invasive hemodynamic monitoring during the peripar-
tum period. Am J Obstet Gynecol 1986;155:999–1001.
552 STAFFORD & SHEFFIELD

[12] Karetsky M, Ramirez M. Acute respiratory failure in pregnancy. An analysis of 19 cases.

Medicine 1998;77:41–9.
[13] Petroianu GA, Toomes LM, Maleck WM, et al. Administration of autologous fetal mem-
branes: effects on the coagulation in pregnant mini-pigs. Pediatric Crit Care Med 2000;1:65–71.
[14] Clark SL. New concepts of amniotic fluid embolism: a review. Obstet Gynecol Surv 1990;45:
360–8.
[15] el Maradny E, Kanayama N, Halim M, et al. Endothelin has a role in early pathogenesis of
amniotic fluid embolism. Gynecol Obstet Invest 1995;40:14–8.
[16] Stolte L, van Kessel H, Seelen J, et al. Failure to produce the syndrome of amniotic fluid em-
bolism by infusion of amniotic fluid and meconium into monkeys. Am J Obstet Gynecol
1967;98:694–7.
[17] Benson MD, Kobayashi H, Silver RK, et al. Immunologic studies in presumed amniotic fluid
embolism. Obstet Gynecol 2001;97(4):510–4.
[18] Oi H, Kobayashi H, Hirashima Y, et al. Serological and immunohistochemical diagnosis of
amniotic fluid embolism. Semin Thromb Hemost 1998;24(5):479–84.
[19] Hiroshi K, Hidekazu OOI, Hiroshi H, et al. Histological diagnosis of amniotic fluid embo-
lism by monoclonal antibody TKH-2 that recognizes NeuAc alpha 2-6GaINAc epitope.
Hum Pathol 1997;28(4):428–33.
[20] Kobayashi H, Ohi H, Terao T. A simple, noninvasive, sensitive method for the diagnosis
of amniotic fluid embolism by monoclonal antibody TKH-2 that recognizes NeuAc alpha
2-6GalNAc. Am J Obstet Gynecol 1993;168(3):848–53.
[21] Harboe T, Benson MD, Oi H, et al. Cardiopulmonary distress during obstetrical anaes-
thesia: attempts to diagnose amniotic fluid embolism in a case series of suspected allergic
anaphylaxis. Acta Anaesthesiol Scand 2006;50(3):324–30.
[22] Fineschi V, Gambassi R, Gherardi M, et al. The diagnosis of amniotic fluid embolism:
an immunohistochemical study for the quantification of pulmonary mast cell tryptase. Int
J Legal Med 1998;111:238–43.
[23] Nishio H, Matsui K, Miyazaki T, et al. A fatal case of amniotic fluid embolism with elevation
of serum mast cell tryptase. Forensic Sci Int 2002;126(1):53–6.
[24] Farrar SC, Gherman RB. Serum tryptase analysis in a woman with amniotic fluid embolism.
A case report. J Reprod Med 2001;46(10):926–8.
[25] Benson MD. A hypothesis regarding complement activation and amniotic fluid embolism.
Med Hypothesis 2007;68(5):1019–25.
[26] Ray BK, Vallejo MC, Creinin MD, et al. Amniotic fluid embolism with second trimester
pregnancy termination: a case report. Can J Anesth 2004;51:139–44.
[27] Hassart TH, Essed GG. Amniotic fluid embolism after transabdominal amniocentesis. Eur
J Obstet Gynecol Reprod Biol 1983;16:25–30.
[28] Maher JE, Wenstrom KD, Hauth JC, et al. Amniotic fluid embolism after saline amnioinfu-
sion: 2 cases and a review of the literature. Obstet Gynecol 1994;83:851–4.
[29] Judich A, Kuriansky J, Engelberg I, et al. Amniotic fluid embolism following blunt abdom-
inal trauma in pregnancy. Injury 1998;29(6):475–7.
[30] Rainio J, Penttila A. Amniotic fluid embolism as cause of death in a car accidentda case
report. Forensic Sci Int 2003;137(2–3):231–4.
[31] McDougall RJ, Duke GJ. Amniotic fluid embolism syndrome: case report and review.
Anaesth Intensive Care 1995;23:735–40.
[32] Clark SL. Hemodynamic alterations associated with amniotic fluid embolism: a reappraisal.
Am J Obstet Gynecol 1985;151:617–21.
[33] Goetz KL, Wang BC, Madweb JB, et al. Cardiovascular, renal and endocrine responses to
intravenous endothelin in conscious dogs. Am J Physiol 1988;255:1064–8.
[34] Koegler A, Sauder P, Marof A, et al. Amniotic fluid embolism: a case with noncardiogenic
pulmonary edema. Intensive Care Med 1994;20:45–6.
[35] Girard P, Mal H, Laie JJF, et al. Left heart failure in amniotic fluid embolism. Anesthesiol-
ogy 1986;64:262–5.
 AMNIOTIC FLUID EMBOLISM 553

[36] Shechtman M, Ziser A, Markovits R, et al. Amniotic fluid embolism: early findings of trans-
esophageal echocardiography. Anesth Analg 1999;89:1456–8.
[37] Moore J, Baldisseri MR. Amniotic fluid embolism. Crit Care Med 2005;33(10):279–85.
[38] Hankins GDV, Snyder RR, Clark SL, et al. Acute hemodynamic and respiratory effects of
amniotic fluid embolism in the pregnant goat model. Am J Obstet Gynecol 1993;168:
1113–30.
[39] Lockwood CJ, Bach R, Guha A, et al. Amniotic fluid contains tissue factor, a potent initiator
of coagulation. Am J Obstet Gynecol 1991;165:1335–41.
[40] Estelles A, Gilabert J, Andres C, et al. Plasminogen activator inhibitor type 1 and type 2 and
plasminogen activators in amniotic fluid during pregnancy. Thromb Haemost 1990;64:
281–5.
[41] Lim Y, Loo CC, Chia V, et al. Recombinant factor VIIa after amniotic fluid embolism and
disseminated intravascular coagulopathy. Int J Obstet Gynecol 2004;87:178–9.
[42] Goldszmidt E, Davies S. Two cases of hemorrhage secondary to amniotic fluid embolus
managed with uterine artery embolization. Can J Anaesth 2003;50:917–21.
[43] Prosper SC, Goudge CS, Lupo VR. Recombinant factor VIIa to successfully manage dis-
seminated intravascular coagulation from amniotic fluid embolism. Obstet Gynecol 2007;
109:524–5.
[44] Kaneko Y, Ogihara T, Tajima H, et al. Continuous hemodiafiltration for disseminated
intravascular coagulation and shock due to amniotic fluid embolism: report of a dramatic
response. Intern Med 2001;40:945–7.
[45] Stanten RD, Iverson LI, Daugharty TM, et al. Amniotic fluid embolism causing catastrophic
pulmonary vasoconstriction: diagnosis by transesophageal echocardiogram and treatment
by cardiopulmonary bypass. Obstet Gynecol 2003;102(3):496–8.
[46] Hsieh YY, Chang CC, Li PC, et al. Successful application of extracorporeal membrane
oxygenation and intraaortic balloon counterpulsation as lifesaving therapy for a patient
with amniotic fluid embolism. Am J Obstet Gynecol 2000;183:496–7.
[47] Johnson TR, Abbasi IA, Urso PJ. Fetal heart rate patterns associated with amniotic fluid
embolus. Am J Perinatol 1987;4:187–90.
[48] Morris JA, Rosenbower TJ, Jurkovich GJ, et al. Infant survival after cesarean section for
trauma. Ann Surg 1996;223:481–8.
[49] Moise KJ, Belfort MA. Damage control for the obstetric patient. Surg Clin North Am 1997;
77:835–52.
[50] Tuffnell DJ. Amniotic fluid embolism. Curr Opin Obstet Gynecol 2003;15(2):119–22.
 Obstet Gynecol Clin N Am
34 (2007) 555–583

Trauma in Pregnancy
Michael V. Muench, MDa,b,*,
Joseph C. Canterino, MDa,b
a
Department of Obstetrics, Gynecology and Reproductive Sciences, University of Medicine and
Dentistry of New Jersey, Robert Wood Johnson School of Medicine, 125 Paterson Street,
New Brunswick, NJ 08901, USA
b
Jersey Shore University Medical Center, Neptune, NJ 07753, USA

Trauma is the most common cause of nonobstetric morbidity and mor-

tality in pregnancy and complicates at least 6% to 7% of all pregnancies
[1–5]. According to the Centers for Disease Control and Prevention, trauma
is the leading cause of death in women 35 years or younger [6]. Maternal
death rates from trauma may be noted as high as 10% to 11% [7,8]. Death
to the fetus is reported to be even higher than death of the mother from
traumatic injuries. With trauma, fetal mortality is as high as 65%, from
placental abruption, direct fetal injury, unexplained fetal loss, maternal
shock, disseminated intravascular coagulation, and other causes [9].
Largely because of the increase in size of the developing fetus and uterus,
the risk of trauma to the mother and fetus increases as pregnancy prog-
resses. There is a 10% to 15% risk of maternal or fetal injury from trauma
during the first trimester, 32% to 40% in the second trimester, and 50% to
54% during the third trimester [10,11]. Motor vehicle crashes cause most in-
juries, but domestic violence, penetrating trauma, and head injuries are also
frequently seen. It has been estimated that motor vehicle collisions occur
during a pregnancy in about 2% of all live births in the United States, or
79,000 children are exposed in utero to a police-reported crash [4,12].
One of the unique characteristics of pregnancy is that relatively minor
injuries can be life threatening for the mother and the developing fetus.
Anatomic and physiologic changes in pregnancy can mask or mimic injury,
making diagnosis of trauma-related problems difficult. To the physician,
these features represent a unique challenge because care must be provided

* Corresponding author. Department of Obstetrics, Gynecology and Reproductive

Sciences, University of Medicine and Dentistry of New Jersey, Robert Wood Johnson School
of Medicine, 125 Patterson Street, Room 2150, New Brunswick, NJ 08901.
E-mail address: mvmuench@comcast.net (M.V. Muench).

0889-8545/07/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ogc.2007.06.001 obgyn.theclinics.com
556 MUENCH & CANTERINO

for two patients. Many physicians are overwhelmed and intimidated in the
management of these patients. However, familiarity with normal anatomical
and physiologic changes, mechanisms of injuries, and maternal trauma
assessment skills enhance the physician’s ability to care for the mother
and her unborn child.

Anatomic and physiologic changes in pregnancy

Numerous changes take place in the cardiovascular system during
pregnancy (Table 1). Beginning in the eighth week of pregnancy, physiologic
changes start to appear. Progesterone-related smooth-muscle relaxation
leads to a significant decrease in the total peripheral resistance. At 10 to
12 weeks’ gestation, blood pressure gradually declines, reaching its nadir
around 28 weeks’ gestation. The systolic and diastolic pressures have
decreased by 5 to 15 mm Hg at this point during the pregnancy. During
the third trimester, blood pressure gradually increases, returning to nearly
prepregnancy readings. These effects are also seen in central venous pressure
as it slowly drops 9 mm Hg to about 4 mm Hg in the third trimester [12].
Progesterone is not the only pregnancy-related hormone that affects

Table 1
Hemodynamic changes during pregnancy
Change during normal Normal range during
Physiology pregnancy pregnancy
Systolic blood pressure Decreases by an average of 110–110 mm Hg
5–15 mm Hg
Diastolic blood pressure Decreases by 5–15 mm Hg 50–70 mm Hg
Mean arterial pressure Decreases by 10 mm Hg 80 mm Hg
Central venous pressure Slightly decreases or no 2–7 mm Hg
change
Heart rate Increases by 10–15 beats/ 75–95 beats/min
min
5
System vascular resistance Decreases by 10%–15% 1200–1500 dynes/sec/cm
Pulmonary vascular Decreases by 10%–15% 55–100 dynes/sec/cm 5
resistance
Cardiac output Increases by 30%–50% 6–7 L/min at rest; 10 L/min
with stress
Cardiac index Increases 4.0–4.5
Pulmonary capillary wedge Decreases 6–9 mm Hg
pressure
Oncotic pressure Decreases 16–19 mm Hg
Blood volume Increases by 30%–50% 4500 mL
Red blood cell volume Increases by 30% d
Hematocrit Decreases 32%–34%
White blood cell count May increase 5000–15,000/mm3
Electrocardiogram Flat or inverted T waves in d
leads III, V1, and V2; Q
waves in leads III and
aVF
 TRAUMA IN PREGNANCY 557

pregnancy. There is an increase in alpha-receptors within the myometrium

stimulated by estrogen. This results in a rise in heart rate of 10 to 15 beats
per minute above baseline. Cardiac output increases to 30% to 50% above
normal during the second trimester. In labor, there is an additional increase
in cardiac output as each uterine contraction results in blood transfer from
the uterus back into circulation. Finally, after delivery of the fetus and
placenta, maximal cardiac output is achieved as the contracted uterus
auto-transfuses the majority of blood it receives back into circulation.
This is usually the most critical period. Cardiac output remains elevated
at third-trimester values for the first 2 postpartum days, and then it slowly
declines to prepregnant values over the next 2 weeks. Blood volume
increases by 50% [13] mainly from plasma as red blood cell volume increases
only about 30%. A further expansion is seen in multiple gestations. The net
effect on the hematologic system is a dilution anemia, the so-called ‘‘physi-
ologic anemia’’ of pregnancy. The average hematocrit of pregnancy is 32%
to 34%. Nearly all coagulation factors increase throughout pregnancy
(Table 2). The net effects of these pregnancy-induced changes are an increase
of procoagulants and a reduction in fibrinolysis, thus creating a hypercoag-
ulable state. This hypercoagulable state is a double-edged sword, protecting
against hemorrhage at the time of delivery but increasing the risk of
thromboembolism.
These hyperdynamic and hypervolemic adaptations help the pregnant
patient tolerate the increase in the metabolic demands of the fetus and the
expected hemorrhage of childbirth. The average estimated blood losses for
a term vaginal delivery and cesarean section are approximately 500 mL
and 1000 mL, respectively. This amount of hemorrhage typically results in
no change in pulse, blood pressure, or other hemodynamic parameters.
The mechanism by which the maternal systemic blood pressure is preserved
is the result of vasoconstriction of the uteroplacental and splanchnic
circulation. During the prenatal period, this reflex reduces perfusion to
the uterus and places the fetus in harm’s way to save the mother. It is this
mechanism that leads many health care providers into a false sense of
security. Because of vasoconstriction of the uterine arteries, the fetus often

Table 2
Changes in coagulation during pregnancy
Coagulation factor Change during normal pregnancy
Fibrinogen Increases (normal range 300–600 mg/dL;
3.0–6.0 g/L)
Factors I, II, V, VII, X, XII Increases
Prothrombin time Decreases by 20%
Partial thromboplastin time Decreases by 20%
Protein S Decreases
Protein C Minimally increases
Plasminogen activator inhibitor-1,-2 Increases (fibrinolytic activity may not be affected)
558 MUENCH & CANTERINO

shows signs of distress (the fifth vital sign in obstetrics) before an alteration
in the maternal hemodynamic parameters. The first maternal signs of distress
may not occur until hemorrhage of 1500 to 2000 mL, a precarious time
because the mother’s condition rapidly deteriorates when blood loss is over
2500 mL. Hemodynamics of the mother are also affected by maternal posi-
tion. The uterus grows from 70 g to 1000 g, and the entire uterofetoplacental
unit averages 4500 g at term. During pregnancy, when the mother is placed in
the supine position, the uterofetoplacental unit compresses the inferior vena
cava. The result is decreased venous return and preload, and subsequently re-
duced cardiac output. This diminished cardiac output may result in significant
hypotension, which often results in vaso-vagal–type symptoms.
The respiratory system undergoes numerous changes during pregnancy
(Table 3). The pregnancy-related increase in blood volume leads to capillary
engorgement of the mucosa throughout the respiratory tract, causing
swelling of the nasal and oral pharynx, larynx, and trachea. This is
compounded by mucosal edema [12]. The end results are difficulty with
nasal breathing, epistaxis, and voice changes [14,15]. These changes may
be significantly exacerbated by a mild upper respiratory tract infection, fluid
overload, oncotic pressure, or the edema associated with preeclampsia.
Thus, leading to a severely compromised airway [14,16,17].
Beyond anatomical changes, there are also changes in respiratory
physiology. These changes are adaptations to the increasing metabolic
demands and oxygen delivery to the fetus. Oxygen consumption increases
by 15% to 20% during pregnancy. Progesterone stimulates the medullary
respiratory center, resulting in hyperventilation and respiratory alkalosis.
The renal tubules are able to metabolically compensate for some of these
effects by excreting bicarbonate. However, a slight alkalemia remains. The
hyperventilation also results in a decrease in the PCO2 to a level of 27 to
32 mm Hg in the pregnant patient. The tidal volume and minute ventilation

Table 3
Anatomical physiological changes in the respiratory system during pregnancy
Physiology or system Change during normal pregnancy
Upper airway Increased edema; capillary engorgement
Diaphragm Displaced 4 cm cephalad
Thoracic anteroposterior diameter Increases
Risk of aspiration Increases
Respiratory rate Slightly increases in the first trimester
Oxygen consumption Increases 15%–20% at rest
Partial pressure of carbon dioxide Decreases (normal range: 27–32 mm Hg)
Partial pressure of oxygen Increases (normal range: 100–108 mm Hg)
Minute ventilation Increases 40%
Tidal volume Increases 40% (normal: 600 mL)
Minute ventilation Increases 40% (normal: 10.5 L/min)
Functional residual capacity Decreases 20%–25%
2,3-Diphosphoglycerate Increases
 TRAUMA IN PREGNANCY 559

increase about 40% as the respiratory rate returns to baseline. There is

a gradual 4-cm elevation of the diaphragm and increase in the thoracic
anteroposterior diameter during the pregnancy. This contributes to a 20%
to 25% decrease in functional residual capacity. These changes and
increasing levels of 2,3-diphosphoglycerate help to facilitate oxygen release
to the fetus. Unfortunately, this process leaves the pregnant patient with
diminished oxygen reserve and buffering capacity. In clinical practice, this
translates into rapid hypoxia of the mother when respiratory stress is
introduced and an inability to compensate for the ensuing acidosis. Fetal
oxygenation remains constant provided the maternal PaO2 remains above
60 mm Hg. Below this PaO2 level, fetal oxygenation drops precipitously.
When fetal oxygen saturation drops by half, the so-called ‘‘diving reflex’’
shunts fetal blood flow away from the liver and abdominal organs to the
heart and brain, thereby exposing other organ systems to hypoxic injury
[18].
In addition to the cardiovascular and pulmonary systems, other organ
systems also undergo significant changes (Table 4). In the abdominal cavity
there is compartmentalization and cephalad displacement of intraabdominal
organs. There is gradual growth and stretching of the abdomen and
peritoneal cavity. This appears to desensitize the peritoneum to irritation
in the pregnant patient. Because of these changes, a physical examination
for abdominal tenderness, rebound, and guarding may find none of these
signs despite the presence of significant injury. Hormonal effects of
pregnancy become evident as progesterone decreases gastrointestinal
motility and relaxes smooth-muscle tone. The gravid uterus causes a shift
in the position of the stomach, which changes the angle of the gastroesoph-
ageal junction, resulting in incompetence of the gastroesophageal pinchcock
mechanism [19]. These effects place the pregnant patient at risk for regurgi-
tation and pulmonary aspiration. Lower esophageal sphincter tone

Table 4
Anatomical physiological changes in the abdomen and genitourinary system during pregnancy
Physiology or system Change during normal pregnancy
Intraabdominal organs Compartmentalization and cephalad displacement
Gastrointestinal tract Decreased gastric emptying; decreased motility;
increased risk of aspiration
Peritoneum Small amounts of intraperitoneal fluid normally
present; desensitized to stretching
Musculoskeletal system Widened symphysis pubis and sacroiliac joints
Kidneys Mild hydronephrosis (right O left)
Renal blood flow Increases by 60%
Glomerular filtration rate Increases by 60%
Serum creatinine Decreases (normal 0.6–0.7 mg/dL (50–60 mmol/L))
Serum urea nitrogen Decreases (normal 3–3.5 mg/dL (1.1–1.2 mmol/L))
Bicarbonate Decreases (normal 19–25 mEq/L)
560 MUENCH & CANTERINO

decreases, allowing gastric reflux and heartburn during pregnancy [20].

Labor itself also decreases gastric emptying [21,22]. Therefore, the pregnant
patient is at increased risk for silent regurgitation, active vomiting, and
aspiration during general anesthesia or impaired consciousness [23].
Within the genitourinary system, the pelvic uterus becomes a lower
abdominal organ by approximately 12 weeks’ gestation. Before 12 weeks,
the small size and pelvic location of the uterus make it relatively resistant
to injury [24]. After the pregnant uterus becomes abdominal, the location
predisposes it to injury from blunt or penetrating abdominal trauma.
Perhaps more importantly, uteropelvic blood flow increases markedly
during pregnancy. By pregnancy’s end, this dramatic increase in pelvic
blood flow increases the likelihood of appreciable hemorrhage in the circum-
stances of uterine injury or pelvic trauma [2,8]. Uterine rupture is relatively
uncommon, occurring in less than 1% of pregnant trauma victims, and is
generally associated with severe direct abdominal impact [25,26]. Fetal death
frequently occurs with uterine rupture, whereas maternal death takes place
in 10% of cases of traumatic uterine rupture [18].
The bladder is displaced anteriorly and superiorly by the uterus,
effectively becoming an intraabdominal organ and more susceptible to
injury. The renal pelves and ureters become dilated from the influence of
progesterone and direct compression of the uterus on the ureters [27]. The
right ureter is more dilated than the left in the second and third trimester.
This occurs under the influence of the sigmoid colon causing rotation of
the uterus to the right. The pressure of the gravid uterus on the ureter
may obstruct or impede urine outflow. Thus ultrasound evidence of a dilated
right renal pelvis is not uncommon [27]. The increase in blood volume dur-
ing the pregnancy increases renal blood flow by about 60%, leading to an
increase in glomerular filtration rate. The end effect is a significantly reduced
serum urea nitrogen to less than 10 and serum creatinine by about half (0.8
mg/dL). Therefore, a relatively ‘‘normal’’ serum urea nitrogen and
creatinine may reflect a seriously compromised renal function.

Trauma pathophysiology and management in pregnancy

The mechanisms of injury and death are composed of multiple categories,
including blunt and penetrating trauma, burns, electrocution, falls, and
assaults. Many attempts have been made to identify factors that predict
outcomes from maternal trauma, but few have been identified. Hypotension
seems intuitive and seems predictive in some studies, but other studies have
not validated these data [9,28–31]. It appears that initial maternal acidosis
may be a useful indicator [28,32–34], while initial pulse, white blood cell
count, hemoglobin, oxygen saturation, and other physiologic or laboratory
values are not useful [35–38]. Fetomaternal hemorrhage has also not been
shown to be predictive of fetal outcome [38–40]. However, fetomaternal
hemorrhage has been associated with uterine contractions and an increased
 TRAUMA IN PREGNANCY 561

risk of preterm labor [36,41]. With all these factors influencing the fetus, it is
not surprising that the most commonly observed complications of all types
of maternal trauma are preterm labor, spontaneous abortion, and placental
abruption [7,34,42]. These complications are thought to be secondary to
intramyometrial bleeding and disruption of the uterine–placental interface.
Intramyometrial bleeding is known to cause contractions by a mechanism
that involves the activation of thrombin, lysosomal enzymes, cytokines,
and prostaglandins [43,44]. Fortunately, in approximately 90% of cases,
as intramyometrial bleeding subsides, contractions also diminish [45].
Penetrating injuries, burns, and electric shock, which are less common
than blunt traumatic injury, may involve other mechanisms of pathophysi-
ology. These mechanisms may take the form of cytokines and inflammatory
mediators typically seen in systemic inflammatory response [46]. In the
following sections, general management and specific management strategies
are discussed.

Prehospital care
Paramedics and first responders should seek information regarding
pregnancy from female patients of childbearing age because there are
specific issues related to the traumatized pregnant patient. Care must be
undertaken during the initial assessment because, as previously stated, vital
signs and patient symptoms may not reflect the underlying injuries to the
patient and fetus. General standard guidelines for trauma patients apply
to the pregnant patient with some modification. Extrication should be
performed in normal fashion with spinal immobilization being employed
for most patients, especially those with blunt force trauma. Placing the
patient on a backboard with a 15 angle to the left is a pregnancy-specific
intervention to avoid compression of the vena cava by the uterus and
resultant hypotension. This technique must be employed in all patients
beyond 20 weeks’ gestation. Failure to employ this procedure can result
in a 30% decrease in cardiac output and possible maternal death from
decreased perfusion of vital organs. The use of towels or blankets placed
under the backboard is quick, easy, and effective. Supplemental oxygen
by nasal cannula or facemask should be given as soon as possible and
considered routine. Two large-bore intravenous catheters should be placed
and 1 to 2 L of resuscitative fluids initiated. The bolus of fluid may allow
for continued perfusion of the uterine placental unit and prevent mild
hypovolemia not noted in the vital signs.
Gestational age can be approximated by the size of the gravid uterus
(Fig. 1) or by the history obtained from the patient. Fetal viability is
extremely likely if the uterine fundal height is between the umbilicus and
xyphoid process. It is important to relay this information to the hospital
or trauma center. This simple task can allow for the obstetrical and neonatal
teams to be mobilized before the patient arrives at the trauma center or to
562 MUENCH & CANTERINO

Fig. 1. Uterine size in weeks’ gestation.

provide consultation for the medics in the field if needed. It is important to

direct the transport to a proper hospital that can care for both the mother
and a premature neonate if delivery is necessary. This has led some
emergency medical service systems to designate pregnancy as an indication
for transport to a trauma facility. Although most patients are unlikely to
need the resources of the trauma center, prehospital findings of tachycardia
(heart rate O110 beats/min), chest or abdominal pain, loss of consciousness,
and third-trimester gestation are associated with an increased need for
services available at the trauma center [38]. In the event that prehospital
transfusion is required, O-negative blood should be used whenever possible.
Emergency medical services that still use military antishock trousers should
be aware that it is contraindicated to inflate the abdominal portion of this
device for pregnant women. Not only can this maneuver result in reduced
uterine perfusion but it also can increase the cardiac workload.

General management
The pregnant trauma patient is best cared for using a team approach. The
emergency physician should involve the trauma surgeon and maternal fetal
medicine specialist or obstetrician early in the care of these patients. The
clinician should perform all necessary tests and procedures on the pregnant
 TRAUMA IN PREGNANCY 563

woman that are indicated, including radiologic imaging, intubation, central

venous access, ultrasonographic evaluations, and even diagnostic peritoneal
lavage (with use of a periumbilical approach or open lavage technique).
Because the most common cause of fetal death is maternal death, efforts
to assess fetal well-being are secondary to resuscitation of the pregnant
woman. However, the well-being of the fetus may represent the most
accurate measurement of maternal health. The patient may have experi-
enced a significant loss of blood, but arterial pressure often remains stable
due to the increase in blood volume during pregnancy and the shunting of
blood flow away from the uterus. This is the normal physiological response
of the pregnant woman to stressdself-preservation at the expense of the
fetus. For this reason, fluid management is important. Waiting for maternal
signs of hypotension will result in fetal compromise and distress. Early
transfusion of blood products may assist in providing volume and improve
oxygen-carrying capacity. Fetal heart monitoring can be useful to guide the
adequacy of fluid resuscitation because fetal heart rate abnormalities may be
the first sign of maternal hypovolemia. For this reason, the fetal heart rate
has been considered the ‘‘fifth vital sign’’ in obstetrics.
The primary survey varies little in the pregnant trauma patient compared
with the nonpregnant patient (Fig. 2). During pregnancy, the risk of
aspiration increases and monitoring of adequate oxygenation by pulse
oximetry is important. Further, because hypoxia results in fetal distress
and maternal oxygen reserve is significantly diminished, early endotracheal
intubation may be considered. During the first and early second trimester,
the woman may be tachypneic, but later in pregnancy other causes of
respiratory compromise must be considered. If a chest tube thoracostomy
is required, it needs to be placed one or two intercostal spaces higher than
usual to avoid diaphragmatic injury. If rapid sequence intubation is
required, lower dosages of succinylcholine are required because pseudocho-
linesterase levels decrease in pregnancy [47]. Both nondepolarizing and
depolarizing paralytics cross the placenta. Therefore a flaccid, apneic infant
may result.
A rapid but thorough secondary survey must include evaluation of the
pregnancy. Great care and precision are needed in performing the abdomi-
nal examination because the normal physiologic stretching of the abdominal
cavity may mask signs of significant peritoneal injury. Findings consistent
with injuries to the liver or spleen include upper abdominal pain, referred
shoulder pain, sudden onset of pain, and elevated liver transaminases. A
focus assessment sonographic trauma (FAST) scan should be performed
for intraabdominal hemorrhage [48,49]. Direct peritoneal lavage using an
open (direct) technique is feasible during pregnancy and appears to be
without any specific pregnancy-related complications [30,33,50]. The uterus
should be palpated carefully because tenderness and contractions may be
overlooked. The top of the fundus should be marked to evaluate the
possibility of concealed abruption as noted by an increasing fundal height.
564 MUENCH & CANTERINO

Fig. 2. Maternal trauma algorithm. CPR, cardiopulmonary resuscitation; FAST SCAN, focus
assessment sonographic trauma scan; IV, intravenous; KB, Kleihauer–Betke.

A sterile speculum examination is vital in the evaluation of the pregnant

trauma patient. Fluid within the vaginal vault may be difficult to
differentiate, but the use of nitrazine paper for a blue color change and
the presence of ferning on microscopic examination aids in distinguishing
alkaline amniotic fluid from urine. Vaginal bleeding may be present, indicat-
ing the possibility of placental abruption, uterine rupture, pelvic fracture
with vaginal injury, or other injuries. The cervix should be visually inspected
for evidence of dilation and effacement. A bimanual examination, which is
sometimes overlooked, is an integral part of the secondary survey.
 TRAUMA IN PREGNANCY 565

Cardiotocographic monitoring needs to be initiated in the emergency

department as soon as possible, preferably on arrival after the secondary
survey and FAST scan, because uterine contractions or irritability may
subside with time. All pregnant women at 20 weeks’ gestation or longer
should have cardiotocographic monitoring for 2 to 6 hours after a traumatic
injury [51]. Monitoring times should be increased in those with contractions,
abdominal pain, or significant maternal injury. An ultrasound of the fetus
and placenta can be performed after the FAST scan or incorporated as
part of an obstetrical/FAST trauma scan. Fetal ultrasound evaluation
should include position of the fetus and heart rate, gestational age
assessment, biophysical profile, fetal middle cerebral artery Doppler peak
velocity for anemia, and evaluation of placenta for abruption. Unfortu-
nately, ultrasound has a low sensitivity for detecting placental abruption
(50%) [52]. However, the positive predict value is high. Ultrasound findings
suggestive of placental abruption are (1) retroplacental hematoma
(hyperechoic, isoechoic, hypoechoic), (2) preplacental hematoma (gelatin-
like appearancedshimmering effectdof the chorionic plate with fetal move-
ment), (3) increased placental thickness and echogenicity, (4) subchorionic
collection, and (5) marginal collection.
The fetus may also be screened for acute anemia by Doppler ultrasound
of the middle cerebral artery. This may identify fetal anemia before
cardiotocographic monitoring indicates distress. In cases of penetrating
trauma, it is important to evaluate the placenta as it relates to the site of
injury. Visualization of streaming indicates placental vessel injury likely
needing immediate delivery.
Laboratory testing in the pregnant trauma patient should include hemoglo-
bin, hematocrit, coagulation studies, typing and cross matching, and a gross
inspection of the urine. Prenatal laboratory tests may be helpful if original
prenatal laboratory results are not available. A serum bicarbonate level, blood
gas analysis, or lactate level may be considered in severe trauma, as some
evidence suggests that maternal acidosis may be linked to fetal outcome
[28,32,33]. A fibrinogen level that is normal in a nonpregnant patient may
be abnormal for pregnancy and may be an early indicator of placental abrup-
tion with a consumptive coagulapathy. A Kleihauer–Betke test should be con-
sidered in all trauma patients because it may be an indicator of the severity of
uterine–placental trauma present and an indicator of those patients at risk of
preterm labor [36]. An Rh-negative patient with a positive test should be
treated with Rh-immune globulin (300 mg initially and an additional 300 mg
for each 30 mL of estimate whole fetal blood) to reduce the risk of isoimmu-
nization. The Rh-negative patient with significant trauma with a positive Klei-
hauer–Betke test, should have repeat Kleihauer–Betke testing and additional
antibody screening (Coombs testing) 24 to 48 hours after the trauma. A neg-
ative antibody screen indicates the need for additional Rh-immune globulin.
Diagnostic studies should be obtained in the pregnant trauma patient for
the same indications as in nonpregnant patients. No study to date has
566 MUENCH & CANTERINO

shown any increase in teratogenicity for a fetus exposed to less than 10 rad
or 100 mGy. Growth restriction, microcephaly, and mental retardation can
occur with high-dose radiation well above that used in medical imaging [53].
The fetus is most at risk for central nervous system effects from 8 to 15
weeks and the threshold appears to be at least 20 to 40 rad or 200 to 400
mGy. The American College of Obstetricians and Gynecologists (ACOG)
has published recommendations for diagnostic imaging during pregnancy
[54]. They state that a 5-rad or 50-mGy exposure to the fetus is not associ-
ated with any increased risk of fetal loss or birth defects. Radiation dosages
by study are listed in Table 5. The fetal radiation dose without shielding is
30% of that to the mother. Mandatory shielding of the fetus decreases
exposure further and should be performed for all studies except for pelvic
and lumbar spine films and CT scans. If multiple diagnostic radiographs
are performed, then consultation with a radiologist or radiation specialist
should be considered to calculate estimated fetal dose as recommended by
the ACOG. This is extremely important when radiation exposure
approaches 5 to 10 rad or 50 to 100 mGy.

Perimortum cesarean section

In cases of maternal cardiac arrest with potential fetal viability, perimor-
tem cesarean section should be performed when resuscitative measures have

Table 5
Radiation exposure to a unshielded uterus/fetus
Uterine radiation dose Uterine radiation dose in
Imaging study in rads milligray units (mGy)
Plain film studies
Abdomen (AP) 0.133–0.92 1.33–9.2
Abdomen (PA) 0.064–0.3 0.64–3
Cervical spine Undetectable Undetectable
Chest (AP) 0.0003–0.0043 0.003–0.043
Chest (PA) !0.001 !0.01
Femur (AP) 0.0016–0.012 0.016–0.12
Hip (AP) 0.01–0.21 0.1–2.1
Pelvis (AP) 0.142–2.2 1.42–22
Full spine (AP) 0.154–0.527 1.54–5.27
Lumbar spine (AP) 0.031–4.0 0.31–40
Thoracic spine (AP) !0.001 !0.01
Computed tomography
Upper abdomen 3.0–3.5 30–35
Entire abdomena 2.8–4.6 28–46
Head !0.05 !0.5
Pelvisa 1.94–5.0 19.4–50
Thorax 0.01–0.59 0.1–5.9
Shielding reduces exposure by 30%.
Abbreviations: AP, anteroposterior; PA, posteroanterior.
a
Depends on trimester.
 TRAUMA IN PREGNANCY 567

failed. The best outcomes occur if the infant is delivered within 5 minutes of
maternal cardiac arrest. This means the decision to operate must be made
and surgery begun by 4 minutes into the arrest [30,55–57]. The latest
reported survival was of an infant delivered 22 minutes after documented
maternal cardiac arrest [58]. Several factors must be considered when
deciding whether to undertake perimortem cesarean section [55,59–62].
These include estimated gestational age (EGA) of the fetus and the resources
of the hospital. The ability to salvage a fetus under ideal circumstances
(availability of all skilled personnel and a controlled setting) may range
from 23 to 28 weeks’ EGA. If the fetus is known to be 23 weeks’ EGA
and the institution’s nursery has never had a newborn of this EGA survive,
perimortem cesarean section is probably not indicated for the sake of the
fetus, but may improve maternal circulation by increasing cardiac return.
Before 23 weeks’ gestational age, delivery of the fetus may not improve
maternal venous return. Therefore aggressive maternal resuscitation is the
only indicated intervention. There has been at least one reported case of
complete maternal and fetal recovery after a prolonged arrest at 15 weeks’
gestation [63].

Blunt trauma
Blunt trauma during pregnancy may be the result of motor vehicle
accidents, accidental falls, and violence. Different mechanisms of maternal
injury occur in pregnant women with blunt abdominal trauma compared
with injuries to their nonpregnant counterparts [28]. Because the gravid
uterus changes the relative location of abdominal contents, transmission
of force may be altered in the pregnant abdomen. Due to increased
vascularity during pregnancy, splenic and retroperitoneal injury and
hematomas are more frequent in pregnant victims of blunt abdominal
trauma [64,65]. Up to 25% of pregnant women with severe blunt trauma
manifest hemodynamically significant hepatic or splenic injuries [66].
Conversely, bowel injury is less frequent [45,67].
Pelvic fractures are another concern during pregnancy and may result in
significant retroperitoneal bleeding [68]. Management is unchanged from the
nonpregnant patient, with consideration for associated injuries of the
bladder, urethra, or rectosigmoid. The presence of a pelvic fracture is not
an absolute contraindication for vaginal delivery. A safe vaginal delivery
can be performed provided the pelvic architecture is not substantially
disrupted and the old fracture is stable [51].
The manifestations of the trauma on the pregnancy may be placental
abruption, preterm labor, or late-onset growth restriction. The underlying
cause for each of these is the extent of placental injury. The placenta does
not contain elastic tissue and thus does not have the capacity to expand
and contract. In contrast, the uterus contains elastic tissue and can react
to acceleration–deceleration forces by changing its shape, in turn generating
568 MUENCH & CANTERINO

very high intrauterine pressures. This produces a shearing effect on the

placental attachment with resultant separation from the uterine wall, and
is the most likely mechanism for the abruption in blunt trauma [69,70].
Placental abruption is present in up to 40% of women with severe maternal
trauma. However, clinically evident abruption occurs in approximately 1%
to 5% of women with minor trauma as well when associated with decelera-
tion and/or uterine-directed force [71]. Fetal death is the result of placental
abruption in 50% to 70% of cases from motor vehicle collisions [72]. Severe
maternal injuries results in fetal death 20% to 40% of the time, and fetal
death from uterine rupture or direct fetal injury accounts for less than
10% of the cases [4]. However, maternal death still ranks as the number
one cause of fetal death.
Direct fetal injuries and fractures complicate less than 1% of cases of
severe blunt abdominal trauma in pregnant women. The reasons for the
low fetal injury rate are the protective nature of the maternal soft tissues,
uterus, and amniotic fluid, the mandatory use of seat belts and shoulder
restraints, and the presence of airbags as standard equipment in automo-
biles. Most such cases of fetal injury occur during late pregnancy in gravidas
[73,74]. Fetal brain and skull injuries may be more common in cases with
fetal head engagement in which maternal pelvic fracture occurs [75,76].
Deceleration injury to the unengaged fetal head may also occur [77].
It is believed that airbags, in conjunction with proper seat belt use,
affords the best protection to the pregnant woman and her unborn child.
The National Highway Traffic Safety Administration does not consider
pregnancy as an indication for deactivation of air bags [78]. They
recommend, however, that airbags be disconnected if vehicle occupants
cannot position themselves with their sternum (or uterine fundus) at least
10 in back from the center of the airbag cover. This is because, with frontal
airbag deployment, the cushion expands at a speed of about 125 mph toward
an individual [79]. Consequently, a person within this expansion zonedthat
is, within 10 in from the steering wheel hub or airbag cushiondis at con-
siderable risk for injury. Placental abruption, as well as fetal or uterine in-
jury, is a potential complication of airbag impact with the gravid abdomen
because of the proximity of the gravid uterus to the rapidly and forcefully
deploying airbag [80]. Relatively minor accidents with no maternal injury
have resulted in severe fetal injury. Data on airbag safety is based only on
case reports [69,81] and a small case series of 30 patients [82]. No conclu-
sive large-scale data exists and until such data suggests harm, the use of
airbags during pregnancy is recommended.

Penetrating trauma
Penetrating trauma in pregnancy is usually the result of gunshot or knife
wounds. Other causes are much less frequent. Gunshot wounds are more
common than knife wounds. The maternal death rate from gunshot wounds
 TRAUMA IN PREGNANCY 569

to the abdomen occurs in 3.9% compared with 12.5% of nonpregnant

victims. The death rate from abdominal stab wounds is also lower for preg-
nant women compared to nonpregnant victims. The reduction in mortality
stems from the anatomical changes induced by pregnancy. Visceral organs
are displaced superiorly by the uterus, which results in the so-called ‘‘protec-
tive effects’’ of the uterus. Thereby visceral injuries are less common during
pregnancy as well [83]. However, when penetrating trauma involves the upper
abdomen, a pregnant woman is more likely to sufer a viseceral injury than if
she were not pregnant. In these cases, the small bowel is more frequently in-
jured, especially during the third trimester. Upper abdominal entry is also the
most common site of abdominal stab wounds during pregnancy. The uterus
and fetus are at increased risk for direct injury as they grow cephalad. Fetal
injuries complicate 66% of gunshot injuries to the uterus [84]. Fetal mortality
ranges from 40% to 70% in cases of penetrating trauma and generally results
from either premature delivery or direct fetal injury by the missile [84]. Stab
wounds to the abdomen are less common than gunshot wounds in the preg-
nant patient and are less likely to result in fetal death. The disparity probably
results from the protective effect of the large muscular uterus on visceral or-
gans. Gunshot wounds cause transient shock waves and cavitations as they
impart their kinetic energy to the high-density tissues of the body, thus caus-
ing more severe injury patterns than low-velocity knife wounds.
Several key factors need to be considered in the management of penetrat-
ing abdominal trauma in the pregnant patient. Traditionally, in the
nonpregnant patient, the universal recommendation is immediate surgical
exploration of these injuries. However, the pattern of organ injury changes
with gestational age. For the pregnant patient with penetrating trauma,
management has become more controversial. Management options include
immediate surgical exploration, diagnostic peritoneal lavage, laparoscopy,
contrast-enhanced CT scanning, local wound exploration, and observation.
Penetrating trauma to the upper abdomen is associated with an increased
risk for maternal bowel injury and operative management is indicated
[85]. In the lower abdomen, the uterus seems to provide some protection
from missile injury and a more individualized approach may be more
appropriate. If the entrance wound of the bullet is below the uterine fundus,
and the bullet remains in the body of the uterus, the incidence of visceral
injury is less than 20% [2]. However, the fetus has a higher incidence of
injury from direct trauma to the uterus. Therefore an individualized
approach for conservative management in lower abdominal injuries is
needed, balancing maternal and fetal concerns [66]. Pregnant patients with
anterior abdominal stab wounds below the level of the uterine fundus are
the best candidates for conservative management [85]. However, the
physician should have a very low threshold for surgical exploration in which
conservative management is considered. Diagnostic peritoneal lavage, fistu-
logram, and ultrasound all may be used in the conservative management of
stable lower-abdominal penetrating injury during pregnancy.
570 MUENCH & CANTERINO

During exploratory laparotomy for the evaluation and management of

penetrating trauma during pregnancy, the uterus should receive careful
inspection with as little traction or twisting of the uterus as possible because
this may decrease blood flow to the fetus. Delivery of the fetus is rarely
necessary unless there is direct perforating injury to the uterus or fetal death.
In cases of uterine injury, care should be individualized to reflect the type of
injuries present, the gestational age of the fetus, and the maternal and fetal
prognosis if undelivered. Delivery of the fetus by cesarean section may be
required if the gravid uterus prevents surgical exposure for repair of
maternal injuries or in the presence of nonreassuring fetal status. In cases
of fetal death, it is often possible and preferred to attempt vaginal delivery
by induction of labor.
Fetal evaluation should include heart-rate tracing with an ultrasound
examination of the fetus (ie, biophysical profile, middle cerebral artery
peak systolic velocity Doppler). In a clinically stable fetus and mother
with suspecting uterine injury, Kleihauer–Betke testing or flow cytometry
for fetomaternal hemorrhage should be considered due to the possibility
of placenta bed disruption. Finally, the decision for conservative or
operative management should be made to ensure the best outcome for the
mother and the fetus.

Electric shock
The incidence of fetal injury after electric injury to the mother is not
known, but injuries appear to be rare during pregnancy. When electrical
injury does happen, it involves both direct and indirect mechanisms. The
direct damage is caused by the actual effect that the electric current has
on various body tissues (eg, the myocardium) or by the conversion of
electrical to thermal energy that is responsible for various types of burns.
Indirect injuries tend to be primarily the result of severe muscle contractions
caused by electrical injury. In general, the type and extent of an electrical
injury depends on the intensity (amperage) of the electric current and resis-
tance of the conducting material. Thus, exposure of different parts of the
body to the same voltage will generate a different current (and by extension,
a different degree of damage) because resistance varies significantly among
various tissues [86]. The least resistance is found in amniotic fluid, nerves,
blood, mucous membranes, and muscles; the highest resistance is found in
bones, fat, and tendons. Skin has intermediate resistance.
The spectrum of injury from accidental electrical shock for the mother
ranges from a transient unpleasant sensation after exposure to low-intensity
current to sudden death due to cardiac arrest. Fortunately for most preg-
nant women, electrical shock from low-voltage current, such as that used
in North America (110 V), results in no or minimal adverse effects on the
mother. In most cases, the current travels hand to hand and not hand to
foot, avoiding the uterus and is unlikely to acutely affect the fetus. This
 TRAUMA IN PREGNANCY 571

may not be the case in hand-to-foot transmission. When electrical current

traverses through the uterus, there is a high incidence of fetal death even
when the woman has no adverse symptoms after the event. In these cases,
fetal death may be immediate or might not become apparent until several
hours after injury [87]. Other fetal complications, including growth
restriction, abruption, and abortion, have been reported following electrical
shock [88–95]. Due to publication bias, reports of adverse outcomes are
more often published than reports of normal outcomes. Consequently, the
literature does not reflect the usual outcome of contact with low-voltage
current [96].
Although fetal and obstetric surveillance is recommended following
electrical injury, there is no evidence that any form of monitoring or
treatment has a direct effect on outcome. Recommendations for fetal
monitoring after electric shock have been published [97]. No fetal monitor-
ing is required before 20 weeks’ gestation. During the second half of
pregnancy, fetal ECG is recommended if it has not been performed earlier.
Maternal ECG and the monitoring of fetal heart rate and uterine activity
are recommended for 24 hours if the injury involved loss of consciousness,
abnormal maternal ECG results, or known maternal cardiovascular illness.
If a fall resulted from the electrical shock, then fetal and uterine monitoring
is indicated for 2 to 4 hours, which is the same as for patients with blunt
trauma. The fetus should have an ultrasound evaluation 2 weeks after the
incident for fetal well-being.

Burns
Burns sustained during pregnancy have been reported as increasing the
mortality and morbidity of both mother and infant. The extent of injury
and treatment is determined by body surface area and depth of injury (Table
6). Burns mainly consist of two groups, minor burns and major burns. The
pregnant patient with a minor burn (!10% of the total body surface area)
often does not require hospitalization and it rarely presents a threat to
maternal or fetal well-being [98]. However, when a major burn is present,
management is more challenging. The pregnant woman who has a major

Table 6
Characteristics of burn injury according to depth
Depth Description
Superficial Moist red wound that blanches with rapid refill
Superficial dermal Pale dry wound with slow color return after blanching
Deep dermal injuries Mottled cherry-red wound that does not blanch; damage
within the capillaries in the deep dermal plexus
Full thickness Dry leathery or waxy hard wound that does not blanch;
may be mistaken for unburnt skin in appearance
572 MUENCH & CANTERINO

burn is subject to all of the serious complications that occur in the nonpreg-
nant woman with a burn, including cardiovascular instability, respiratory
distress, sepsis, and renal and liver failure. The greatest risk occurs when
the total body surface area burned is over 60% [99]. With improvement in
the overall survival of burn patients, pregnant women with burns also stand
a better chance of survival. The best chance of fetal survival occurs when the
mother survives and remains free of severe complications, such as sepsis,
hypotension, hypoxia, and death.
The overall treatment of a burn patient is unchanged by pregnancy. The
basic principles of management include support of respiratory function and
stabilization of airway injury, fluid and electrolyte management, infection
control, nutritional support, eschar debridement, wound coverage with
autografts, and the prevention and treatment of any complications.
Inhalation injuries are known to increase the mortality rate in burn
victims and are highly problematic. Pregnant women with facial burns
should be monitored carefully for breathing difficulties. Inspection via
bronchoscopy may be necessary and intubation may be required if the
patient is not adequately oxygenated. Dyspnea and wheezing may develop
when overwhelming irritation is present, but often are not seen during the
first 12 to 48 hours after injury. The avoidance of hypoxia is most
important, and early oxygen therapy is always advised. Continuous
pulse-oximetry is helpful in assessing oxygenation. Bronchodilators and
assisted ventilation may be necessary. Corticosteroids and prophylactic
antibiotics have not been shown to be effective adjunctive therapies in the
treatment of respiratory complications.
Carbon monoxide is frequently inhaled in a closed fire and freely crosses
the placenta. Because fetal hemoglobin has a higher affinity for binding
carbon monoxide, the effects may be more pronounced in the fetus than
in the adult. Exposure to carbon monoxide in utero may affect cardiac
development and may produce fetal cardiac edema. Oxygen is the treatment
of choice, and ventilation with 100% oxygen will reduce the half-life of
carboxy-hemoglobin from 4.5 hours to approximately 50 minutes [100].
The second challenge in the pregnant burn patient is fluid loss. Fluid
losses are the greatest in the first 12 hours after the injury. Fluid shifts
may result in decreased uteroplacental circulation. These result in acute
ischemic changes in the placenta and may lead to fetal hypoxia and acidosis.
Even if the burned area is only 15% of total body surface area, sufficient
fluid loss may occur for the patient to become hypovolemic. According to
the Parkland formula, the fluid requirement in the first 24 hours postburn
is 4 mL/kg body weight per percent of body surface area burned [101].
One half of the calculated fluids are given in the first 8 hours and the rest
in the next 16 hours. In pregnancy, total body surface area is increased.
The pregnant burn patient requires additional fluid resuscitation beyond
amounts seen in nonpregnant individuals, rendering the Parkland formula
inaccurate. It is important to maintain normal maternal hemodynamics
 TRAUMA IN PREGNANCY 573

and adequate urine output. Invasive central hemodynamic monitoring may

be necessary in cases involving cardiac or pulmonary compromise or in
situations with inadequate urine output. If urine output remains low, renal
dose dopamine (1–2 mg/kg/min) should be given. Interstitial edema can be
expected to resolve within a few days, as noted by profound diuresis.
Both hyponatremia and hypokalemia can result from a serious burn injury
and its mismanagement. Hyponatremia is often the result of dilution effects
of intravenous fluids as well as fluid losses through the gastrointestinal and
genitourinary tracts and directly through the wound. Hypokalemia also can
result from chronic potassium losses through the wound. Frequent
monitoring is needed and the deficits should be corrected with potassium.
Other major concerns are nutrition and infection. The hypermetabolic
state for the pregnant mother is amplified after a burn injury. Adequate
nutritional support is essential during this period. Early enteral nutrition
is vital in the management of the pregnant burn patient. A nasogastric
tube is required if the burn is greater than 20% of the total body surface
area. These burn patients often develop an ileus. The increased metabolic
state of the patient also has an impact on the risk of infection. For the
burn patient, infection is one of the most devastating complications.
Septicemia and respiratory infections account for the majority of all deaths
in burn patients and their fetuses. The use of prophylactic antibiotics is
controversial, and treatment should be based on blood cultures and
sensitivities.
The risk of preterm labor increases with increasing total body surface
area burned. The best way to reduce the risks of preterm labor and fetal
demise is to maximize the health of the mother by preventing hypovolemia,
sepsis, hypoxia, and electrolyte imbalances. When preterm labor occurs,
treatment is considered controversial. The use of tocolytic therapy may be
considered when the total body surface area burned is less than 30% to
40% and the estimated gestational age is between 24 and 32 weeks, as
long as fetal monitoring is reassuring [102]. Corticosteroids to enhance fetal
lung maturity should be given because of the risk of premature delivery. The
mode of delivery in the pregnant burn patient is decided by obstetrical
indications. Vaginal deliveries are possible even in cases of extensive perineal
burns, and grossly infected perineal burns seem to have no effect on
neonatal survival. When a full-thickness perineal burn occurs, the tissue
loses its elasticity and an episiotomy might be required. Cesarean delivery
may be performed over a burned abdomen when obstetrically indicated.

Spinal cord injuries

The initial management of a spinal cord injury focuses on stabilization of
the neck and airway maintenance [103]. When multiple injuries are present,
it is important to rule out internal hemorrhage. The pregnant patient with
acute spinal injury is treated the same as the nonpregnant patient and
574 MUENCH & CANTERINO

should receive intravenous methylprednisolone within 8 hours of the injury

and continued for 24 hours. This is associated with significant improvement
in motor and sensory function 6 months after the injury [104]. It is also
important to avoid maternal hypotension so as to maintain uterine blood
flow and reduce the risk of secondary ischemic damage in the evolving lesion
of the cord.
The management of a spinal cord injury in pregnancy depends upon the
site, extent, and duration of the lesion. Complete transection of the cord is
associated with neurogenic shock, cardiovascular instability, and autonomic
hyperreflexia. Neurogenic shock develops with the blockade of the sympa-
thetic autonomic function by the cord injury, and is characterized by a dom-
inance of parasympathetic autonomic system. The patient typically develops
hypotension and bradycardia with decreased cardiac output, and warm dry
skin leading to loss of heat and hypothermia. Fetal distress may follow.
These effects from neurogenic shock generally last from 1 to 3 weeks. Fluid
management should be guided by central venous monitoring, as the typical
signs of hypovolemia may not be present. Positive inotropic agents, such as
dopamine and dobutamine (1–5 mg/kg/min), may be needed to enhance car-
diac output, perfusion pressure, and renal perfusion. These agents appear to
be safe in pregnancy because they do not reduce uterine perfusion and are
not associated with teratogenic effects of the fetus.
Autonomic hyperreflexia or dysreflexia, which occurs in up to 85% of
patients with a spinal cord injury above the level of the splanchnic
autonomic outflow (T5/6), is caused by unregulated sympathetic nervous
system activity. The occurrence of autonomic hyperreflexia may signal the
resolution of the period of neurogenic shock. The initiating stimulus is
below the level of the spinal injury and may be the result of labor, a full
bladder, or a distended rectum or bowel. This stimulus results in a paroxys-
mal release of catecholamines. The patient may experience severe
hypertension, tachycardia, reflex baroreceptor-mediated bradycardia,
throbbing headaches, flushing, skin blotching, sweating, piloerection, nasal
obstruction, chest pain, nausea, tremor, and feelings of anxiety. Serious
complications, including convulsions, permanent neurological deficit,
intracerebral hemorrhage and death, may occur. Management involves
interruption of the reflex arc with regional anesthesia or inhalation anesthe-
sia [105]. If anesthesia is not available, then intravenous labetalol or
nitroprusside, diphenhydramine, hydralazine, diazepam, and guanethidine
can be used. However, careful titration of these medications is required to
avoid hypotension.
Finally, apart from complications listed above, these patients also have
long-term needs. Special attention is needed to avoid urinary-tract
infections, constipation, pressure sores, thrombosis, and repeat episodes of
autonomic hyperrflexia. The mode and timing of delivery should be in
accordance with the usual obstetric indications, unless the pregnancy
impedes proper monitoring and care. The obstetric outcome is usually
 TRAUMA IN PREGNANCY 575

good. Where the patient has sustained permanent injury, most of the above
factors should be taken into consideration with a subsequent pregnancy.

Traumatic brain injury

Management of traumatic brain injury depends on the type and severity
of injury. The initial management of the head-injured gravida focuses on
maintaining ventilatory and circulatory function, cerebral blood flow, and
normal physiologic functions of the mother and fetus. Guidelines for the
management of severe traumatic brain injury have been published [106]
and the role of neurosurgical intervention is unchanged by pregnancy.
Early postinjury hypoxia and hypotension greatly increase morbidity and
mortality in traumatic brain injury patients. Accordingly, hypotension and
hypoxemia should be closely monitored and treated. Patients with a Glas-
gow Coma Score less than 9 and who are unable to maintain their airway
or who remain hypoxemic despite supplemental oxygen require endotra-
cheal intubation. When possible, the bed should be kept elevated at 30 to
reduce the intracranial pressure and the mean arterial blood pressures
should be maintained above 90 mm Hg through intravenous fluid manage-
ment. These therapies should help maintain cerebral perfusion pressure at
more than 70 mm Hg [107].
Brain edema and elevated intracranial pressure develops in 40% of
patients with severe traumatic brain injury. High or uncontrolled intracra-
nial pressure is the most common causes of death and neurologic disability
after severe traumatic brain injury. The main objective of intracranial
pressure monitoring is to maintain adequate cerebral perfusion and
oxygenation. It is also a means for guiding therapy. Intracranial pressure
monitoring is reserved for those individuals with severe traumatic brain
injury (ie, Glasgow Coma Score 8 or less) or an abnormal head CT scan
[106]. Treatment should be initiated at an intracranial pressure threshold
of 20 to 25 mm Hg. Treatment options consist of hyperventilation, chemical
agents, hypothermia, and neurosurgical intervention.
Hyperventilation reduces cerebral blood flow and therefore decreases
intracranial edema. In the nonpregnant patient, aggressive hyperventilation,
defined as PaCO2 of 25 or less, has previously been the cornerstone in the
management of severe trauma brain injury. However, aggressive hyperven-
tilation may cause cerebral ischemia by further reducing cerebral blood flow
without decreasing intracerebral edema and be associated with poorer
neurological outcomes [108]. In the pregnant patient, normal PaCO2 is 32
mm Hg. Thus, hyperventilation may require levels of carbon dioxide that
are extremely low. It has been suggested that extreme hypocapnia causes
direct uterine vasoconstriction, possibly leading to fetal hypoxia. Recent
evidence suggests hyperventilation rather than hypocapnia leads to a reduc-
tion in uterine blood flow through a mechanical reduction in venous return
and subsequent decrease in cardiac output [45]. For these reasons, the
576 MUENCH & CANTERINO

effective range for hyperventilation is reduced in pregnancy and caution

must be exercised [45].
The mainstay of chemical agents for the reduction of intracranial
pressure is mannitol, as hypertonic saline and steroids have had conflicting
results [109]. In the nonpregnant patient, mannitol is routinely used to
reduce intracranial pressure in traumatic brain injury patients with intracra-
nial hypertension. Mannitol has a beneficial effect on maternal intracranial
pressure, cerebral perfusion pressure, cerebral blood flow, brain metabolism,
and short-term neurologic outcome. However, it may adversely affect the
fetus. Indirectly, the osmotic diuresis can result in volume deficits in the
mother and hypoperfusion of the placenta. Mannitol forces free water
from the fetus and amniotic fluid to the mother, resulting in oligohydrami-
nos, contraction of fetal blood volume, cyanosis, and fetal bradycardia. For
these reasons, the use of mannitol during pregnancy should be restricted.
Hypothermia has also been proposed as a treatment of elevated intracranial
pressure, especially in those with hyperthermia [110]. The effects of
hypothermia on a fetus are unknown and this treatment should be
considered experimental. The physician is faced with a situation where
treatments to help the mother may be contraindicated for the fetus. This
is why some have advocated delivery of the fetus or termination to allow
for maternal treatment. There may be benefit of craniotomy in patients
with traumatic brain injuries when the intracranial pressure is refractory
to conventional treatment [111]. There have also been reports of better
outcomes with early craniotomy in traumatic brain injury when the
Glasgow Coma Scale is below 8 [112]. Although there are no case reports
of its use during pregnancy, craniotomy has been used during pregnancy
for the resection of brain tumors. Craniotomy may be the only treatment
available for elevated intracranial pressure after severe traumatic brain
injury that allows for the prolongation of pregnancy.

Domestic violence
Domestic violence is common during pregnancy and affects up to 20% of
all pregnancies [113]. It may be the leading cause of trauma in pregnancy. A
pregnant woman is more likely to suffer domestic abuse than preeclampsia.
Therefore, for physicians, diagnosing domestic abuse may be more crucial
than diagnosing a placental abruption. Domestic violence may increase
during pregnancy and lead to increased emergency room evaluations and
antepartum and postpartum admissions [114,115]. The abuser tends to focus
the attack on the abdomen, breast, and genitals. The effects of domestic
abuse on the fetus typically depend on the severity of placental injury. These
effects range from preterm delivery, preterm labor, growth restriction, and
low birth-weight as the severity of placental injury decreases. The first
step in treating domestic abuse is identification. Simple screening question-
naires have been developed to identify patients at risk (Box 1) [113]. The
 TRAUMA IN PREGNANCY 577

Box 1. Examples of screening questions for domestic violence

The Massachusetts Medical Society Committee on Violence
single-question screena
 ‘‘At any time, has a partner hit, kicked, or otherwise hurt
or threatened you?’’
Three-question abuse assessment screenb
 ‘‘Within the last year, have you been hit, slapped, kicked
or otherwise physically hurt by someone?’’
 ‘‘Since you’ve been pregnant, have you been hit, slapped,
kicked, or otherwise physically hurt by someone?’’
 ‘‘Within the last year, has anyone forced you to have sexual
activities?’’
The SAFE questionsc
 Stress/safetyd‘‘Do you feel safe in your relationship?’’
 Afraid/abusedd‘‘Have you ever been in a relationship where
you were threatened, hurt, or afraid?’’
 Friends/familyd‘‘Are your friends or family aware that you
have been hurt? Could you tell them, and would they be able
to give you support?’’
 Emergency pland‘‘Do you have a safe place to go and the
resources you need in an emergency?’’

From aMassachusetts Medical Society Committee on Violence. Partner vio-

lence: how to recognize and treat victims of abuse. Waltham (MA): Massachusetts
Medical Society; 1996; bMcFarlane J, Parker B, Soeken K, Bullock L. Assessing for
abuse during pregnancy: severity and frequency of injuries and associated entry
into prenatal care. JAMA 1992;267:3176; and cAshur, ML. Asking about domestic
violence: SAFE questions. JAMA 1993;269:2367.

most effective strategies for identifying domestic violence are screening

questionnaires followed by in-person interviews by highly trained individ-
uals [116]. A heightened index of suspicion and a concise screening tool
may afford the emergency physician the unique opportunity to identify,
intervene, and prevent reoccurrence of domestic violence. If domestic
violence is suspected, consultation with social services should not be
delayed. For a comprehensive review of domestic violence, see the article
by Gunter elsewhere in this issue.

Summary
Trauma is the leading nonobstetric cause of maternal mortality, with the
majority of injuries occurring from motor vehicle accidents. The basic tenets
of trauma evaluation and resuscitation should be applied in maternal
578 MUENCH & CANTERINO

trauma. It is important to understand the mechanism of injury, as well as

the anatomical and physiological changes present in pregnancy. Failure to
do so may have a significant impact on maternal hemodynamics and the
fetus. Therefore aggressive resuscitation of the mother is the best manage-
ment for the fetus. Care must be taken to keep the patient in the left lateral
decubitus position to avoid compression of the inferior vena cava and
resultant hypotension. Radiographic studies should not be avoided, but
rather used with care. Noninvasive diagnostics, such as abdominal ultraso-
nography, should be used when available. Cardiotocographic monitoring of
a viable gestations (O20 weeks’ gestation) should be initiated as soon as
possible in the emergency department to evaluate fetal well-being because
fetal well-being is often the best indicator of maternal health. Seemingly
minor injuries can result in placental abruption. Therefore monitoring is
required for at least 2 to 4 hours after any trauma. Longer monitoring is
needed if contractions or unstable maternal hemodynamics are present.
Kleihauer–Betke testing should be considered in all cases of blunt trauma
to determine the risk of preterm labor and placental injury. Rh-negative
mothers should receive Rh-immune globulin administration to reduce the
risk of Rh immunization. While routine cesarean section is not warranted,
even in patients requiring laparotomy, urgent cesarean section should be
considered if fetal distress is present, or if the presence of the fetus is contrib-
uting to maternal instability. For best fetal outcomes, perimortem cesarean
section should be undertaken within 5 minutes of maternal circulatory
arrest. Screening for domestic violence, particularly in patients with
repeated injuries, should be undertaken and appropriate interventions
made when identified. Finally, trauma centers and emergency rooms should
have protocols in place that address the unique situations for trauma
occurring during pregnancy. These protocols should include input from
all specialists involved in this multidisciplinary emergency.

References
[1] El-Kady D, Gilbert WM, Anderson J, et al. Trauma during pregnancy: an analysis of
maternal and fetal outcomes in a large population. Am J Obstet Gynecol 2004;190:
1661–8.
[2] Vaizey CJ, Jacobson MJ, Cross FW. Trauma in pregnancy. Br J Surg 1994;81:
1406–15.
[3] Warner MW, Salfinger SG, Rao S, et al. Management of trauma during pregnancy. ANZ J
Surg 2004;74:125–8.
[4] Hyde LK, Cook LJ, Olson LM, et al. Effect of motor vehicle crashes on adverse fetal
outcomes. Obstet Gynecol 2003;102:279–86.
[5] Drost TF, Rosemurgy AS, Sherman HF, et al. Major trauma in pregnant women:
maternal/fetal outcome. J Trauma 1990;30:574–8.
[6] National vital statistics reports. March 7, 2005, Volume 53, Issue 1.
[7] Weiss H, Songer T, Fabio A. Fetal deaths related to maternal injury. JAMA 2001;286(15):
1863–8.
 TRAUMA IN PREGNANCY 579

[8] Esposito T. Trauma during pregnancy. Emerg Med Clin North Am 1994;12(1):167–96.
[9] Ali J, Yeo A, Gana T, et al. Predictors of fetal mortality in pregnant trauma patients.
J Trauma 1997;42(5):782–5.
[10] Curet M, Schermer C, Demarest G, et al. Predictors of outcome in trauma during
pregnancy: identification of patients who can be monitored for less than 6 hours. J Trauma
2000;49(1):18–25.
[11] Fort A, Harlin R. Pregnancy outcome after noncatastrophic maternal trauma during
pregnancy. Obstet Gynecol 1970;35(6):912–5.
[12] Norwitz ER, Robinson JN, Malone FD, et al. Critical care obstetrics. In: Clark SL, Cotton
DB, Hankins GDV, editors. Critical care obstetrics. 4th edition. Malden (MA): Blackwell
Scientific; 2004. p. 19–42.
[13] Pritchard JA. Changes in blood volume during pregnancy and delivery. Anesthesiology
1965;26:393–9.
[14] Jouppila R, Jouppila P, Hollmen A. Laryngeal oedema as an obstetric anaesthesia
complication: Case reports. Acta Anaesthesiol Scand 1980;24:97–8.
[15] Kuczkowski KM, Reisner LS, Benumof JL. Airway problems and new solutions for the
obstetric patient. J Clin Anesth 2003;15:552–63.
[16] Farcon EL, Kim MH, Marx GF. Changing Mallampati score during labour. Can J Anaesth
1994;41(1):50–1.
[17] Brimacombe J. Acute pharyngolaryngeal oedema and pre-eclamptic toxaemia. Anaesth
Intensive Care 1992;20:97–8.
[18] Pearlman M, Tintinalli J. Evaluation and treatment of the gravida and fetus following
trauma during pregnancy. Obstet Gynecol Clin North Am 1991;18(2):371–81.
[19] Vanner RG. Mechanisms of regurgitation and its prevention with cricoid pressure. Int
J Obstet Anesth 1993;2:207–15.
[20] O’Sullivan G, Scrutton M. NPO during labor: is there any scientific validation? Anesthesiol
Clin North America 2003;21:87–98.
[21] Carp H, Jayaram A, Stoll M. Ultrasound examination of the stomach contents of
parturients. Anesth Analg 1992;74:683–7.
[22] Cheek TG, Gutsche BB. Pulmonary aspiration of gastric contents. In: Hughes SC,
Levinson G, Rosen MA, editors. Shnider and Levinson’s anesthesia for obstetrics. 4th
edition. Philadelphia: Lippincott Williams & Wilkins; 2002. p. 391–405.
[23] Cohen SE. The aspiration syndrome. Clin Obstet Gynaecol 1982;9:235–54.
[24] Cohen SE, Cohen SE. Safety of lap-belt restraint for pregnant victims of automobile
collisions. N Engl J Med 1971;284:632–6.
[25] Van Hook JW, Gei AF, Pacheco LD, et al. Trauma in pregnancy. In: Clark SL, Cotton DB,
Hankins GDV, editors. Critical care obstetrics. 4th edition. Malden (MA): Blackwell Sci-
entific; 2004. p. 484–505.
[26] Cunningham FG, Gant NF, Leveno KJ, et al. Williams Obstetrics. 22nd edition. New
York: McGraw-Hill; 2005.
[27] Lapinsky S, Kruczynski K, Slutsky A. Critical care in the pregnant patient. Am J Respir
Crit Care Med 1995;152:427–55.
[28] Shah KH, Simons RK, Holbrook T, et al. Trauma in pregnancy: maternal and fetal
outcomes. J Trauma 1998;45:83–6.
[29] Baerga-Varela Y, Zietlow S, Bannon M, et al. Trauma in pregnancy. Mayo Clin Proc 2000;
75:1243–8.
[30] Esposito T, Gens D, Smith L, et al. Trauma during pregnancy. Arch Surg 1991;126:1073–8.
[31] Kissinger D, Rozycki G, Morris J, et al. Trauma in pregnancy. Predicting pregnancy
outcome. Arch Surg 1991;126:1079–86.
[32] Hoff W, D’Amelio L, Tinkoff G, et al. Maternal predictors of fetal demise in trauma during
pregnancy. Surg Gynecol Obstet 1991;172(3):175–80.
[33] Scorpio R, Esposito T, Smith G, et al. Blunt trauma during pregnancy: factors affecting
fetal outcome. J Trauma 1992;32(2):213–6.
580 MUENCH & CANTERINO

[34] Rogers FB, Rozycki GS, Osler TM, et al. A multi-institutional study of factors associated
with fetal death in injured pregnant patients. Arch Surg 1999;134:1274–7.
[35] Muench MV, Baschat AA, Harman CR, et al. Elevated white blood cell count in maternal
trauma: does it predict placental abruption? Am J Obstet Gynecol 2003;189(6 Suppl 1):119.
[36] Muench MV, Baschat AA, Reddy UM, et al. Kleihauer–Betke testing is important in all
cases of maternal trauma. J Trauma 2004;57(5):1094–8.
[37] Schiff MA, Holt VL. The injury severity score in pregnant trauma patients: predicting
placental abruption and fetal death. J Trauma 2002;53(5):946–9.
[38] Goodwin T, Breen M. Pregnancy outcome and fetomaternal hemorrhage after
noncatastrophic trauma. Am J Obstet Gynecol 1990;162(3):665–71.
[39] Connolly A, Katz V, Bash K, et al. Trauma and pregnancy. Am J Perinatol 1997;14(6):
331–5.
[40] Dahmus M, Sibai B. Blunt abdominal trauma: are there any predictive factors for abruptio
placentae or maternal-fetal distress? Am J Obstet Gynecol 1993;169(4):1054–9.
[41] Muench MV, Harman C, Reddy UM, et al. In maternal trauma, Kleihauer–Betke testing
predicts preterm labor. Am J Obstet Gynecol 2001;185(6):S109.
[42] Weiss HB. The epidemiology of traumatic injury-related fetal mortality in Pennsylvania,
1995–1997: the role of motor vehicle crashes. Accid Anal Prev 2001;33:449–54.
[43] Phillippe M, Chien E. Intracellular signaling and phasic myometrial contractions. J Soc
Gynecol Investig 1998;5:169–77.
[44] Lavin JP, Polsky SS. Abdominal trauma during pregnancy. Clin Perinatol 1983;10:423–38.
[45] Pearlman MD, Tintinallli JE, Lorenz RP. Blunt trauma during pregnancy. N Engl J Med
1991;323:1609–13.
[46] Brun-Buisson C. The epidemiology of the systemic inflammatory response. Intensive Care
Med 2000;26:S64–74.
[47] Schneider R. Muscle relaxants. In: Walls R, editor. Emergency airway management.
Philadelphia: Lippincott Williams & Wilkins; 2000. p. 121–8.
[48] Ingeman J, Plewa M, Okasinski R, et al. Emergency physician use of ultrasonography in
blunt abdominal trauma. Acad Emerg Med 1996;3(10):931–7.
[49] Ma O, Mateer J, DeBehnke D. Use of ultrasonography for the evaluation of pregnant
trauma patients. J Trauma 1996;40(4):665–8.
[50] Rothenberger D, Quattlebaum F, Zabel J, et al. Diagnostic peritoneal lavage for blunt
trauma in pregnant women. Am J Obstet Gynecol 1977;129(5):479–81.
[51] American College of Obstetricians and Gynecologists. Obstetric aspects of trauma
management. Educational Bulletin No. 251. September, 1998.
[52] Harrison SD, Ngheim HV, Shy K. Uterine rupture with fetal death following blunt trauma.
AJR Am J Roentgenol 1995;165:1452.
[53] Berlin L. Radiation exposure and the pregnant patient. AJR Am J Roentgenol 1996;167:
1377–9.
[54] American College of Obstetricians and Gynecologists Committee Opinion #299:
Guidelines for diagnostic imaging during pregnancy; 2004.
[55] Katz VL, Dotters DJ, Droegemueller W. Perimortem cesarean delivery. Obstet Gynecol
1986;68(4):571–6.
[56] Morris J, Rosenbower T, Jurkovich G, et al. Infant survival after cesarean section for
trauma. Ann Surg 1996;223(5):481–91.
[57] Towery R, English P, Wisner D. Evaluation of pregnant women after blunt injury.
J Trauma 1993;35(5):731–6.
[58] Lopez-Zeno JA, Carlo WA, O’Grady JP, et al. Infant survival following delayed postmor-
tem cesarean delivery. Obstet Gynecol 1990;76(5 Pt 2):991–2.
[59] Weber CE. Postmortem cesarean section: review of the literature and case reports. Am
J Obstet Gynecol 1971;110(2):158–65.
[60] DePace NL, Betesh JS, Kotler MN. ‘Postmortem’ cesarean section with recovery of both
mother and offspring. JAMA 1982;248(8):971–3.
 TRAUMA IN PREGNANCY 581

[61] Lanoix R, Akkapeddi V, Goldfeder B. Perimortem cesarean section: case reports and
recommendations. Acad Emerg Med 1995;2(12):1063–7.
[62] Strong TH Jr, Lowe RA. Perimortem cesarean section. Am J Emerg Med 1989;7(5):489–94.
[63] Selden BS, Burke TJ. Complete maternal and fetal recovery after prolonged cardiac arrest.
Ann Emerg Med 1988;17(4):346–9.
[64] Flick RP, Bofill JA, King JC. Pregnancy complicated by traumatic diaphragmatic rupture.
A case report. J Reprod Med 1999;44:127–30.
[65] Icely S, Chez RA. Traumatic liver rupture in pregnancy. Am J Obstet Gynecol 1999;180:
1030–1.
[66] Kuhlmann RS, Cruikshank DP. Maternal trauma during pregnancy. Clin Obstet Gynecol
1994;37:274–93.
[67] Pearlman MD, Tintinalli JE, Lorenz RP. A prospective controlled study of outcome after
trauma during pregnancy. Am J Obstet Gynecol 1990;162:1502–7.
[68] Leggon RE, Wood G, Craig MS, et al. Pelvic fractures in pregnancy: factors influencing
maternal and fetal outcomes. J Trauma 2002;53(4):796–804.
[69] Schultze PM, Stamm CA, Roger J. Placental abruption and fetal death with airbag
deployment in a motor vehicle accident. Obstet Gynecol 1998;92(4 Pt 2):719.
[70] Fries MH, Hankins GDV. Motor vehicle accident associated with minimal maternal
trauma but subsequent fetal demise. Ann Emerg Med 1989;18:301–4.
[71] Pearlman MD, Viano D. Automobile crash simulation with the first pregnant crash test
dummy. Am J Obstet Gynecol 1996;175(4 Pt 1):977–81.
[72] Weintraub AY, Leron E, Mazor M. The pathophysiology of trauma in pregnancy: a review.
J Matern Fetal Neonatal Med 2006;19(10):601–5.
[73] Evrard JR, Sturmer WQ, Murray EJ. Fetal skull fracture from an automobile accident. Am
J Forensic Med Pathol 1898;10:232–4.
[74] Hartl R, Ko K. In utero skull fracture: case report. J Trauma 1996;41:549–52.
[75] Alley JR Jr, Yahagi Y, Moncure MM, et al. A case of in utero fetal brain trauma after
motor vehicle collision. J Trauma 2003;55(4):782–5.
[76] Palmer JD, Sparrow OC. Extradural haematoma following intrauterine trauma. Injury
1994;25:671–3.
[77] Weyerts LK, Jones MC, James HE. Paraplegia and congenital fractures as a consequence of
intrauterine trauma. Am J Med Genet 1992;43:751–2.
[78] National conference on medical indications for air bag disconnection. George Washington
University Medical Center. Final report; 1997.
[79] Pearlman MD, Klinich KD, Schneider LW, et al. A comprehensive program to improve
safety for pregnant women and fetuses in motor vehicle crashes: a preliminary report.
Am J Obstet Gynecol 2000;182:1554–64.
[80] Bjornstig U, Haraldsson PO, Polland W, et al. Awareness of the risk of air bag-associated
injuries. Lakartidningen 2002;99:3022–6.
[81] Fusco A, Kelly K, Winslow J. Uterine rupture in a motor vehicle crash with airbag
deployment. J Trauma 2001;51:1192–4.
[82] Metz TD, Torri D, Abbott JT, et al. Uterine trauma in pregnancy after motor vehicle
crashes with airbag deployment: a 30-case series. J Trauma 2006;61(3):658–61.
[83] Lavery J, Staten-McCormick M. Management of moderate to severe trauma in pregnancy.
Obstet Gynecol Clin North Am 1995;22(1):69–90.
[84] Buchsbaum H. Penetrating injury of the abdomen. In: Buchsbaum H, editor. Trauma in
pregnancy. Philadelphia: WB Saunders; 1979. p. 82–100.
[85] Awwad J, Azar G, Seoud M, et al. High velocity penetrating wounds of the gravid uterus:
review of 16 years of civil war. Obstet Gynecol 1994;83(2):259–64.
[86] Cooper MA. Emergent care of lightning and electrical injuries. Semin Neurol 1995;15:
268–78.
[87] Fish R. Electric shock. Part I: physics and pathophysiology. J Emerg Med 1993;11:
309–12.
582 MUENCH & CANTERINO

[88] Toongsuwan S. Post mortem caesarean section following death by electrocution. Aust N Z
J Obstet Gynaecol 1972;12:265–6.
[89] Yoong AF. Electrical shock sustained in pregnancy followed by placental abruption.
Postgrad Med J 1990;66(777):563–4.
[90] Peppler RD, Labranche FJ Jr, Comeaux JJ. Intrauterine death of a fetus in a mother
shocked by an electrical current: a case report. J La State Med Soc 1973;124(2):37–8.
[91] Steer RG. Delayed fetal death following electrical injury in the first trimester. Aust N Z J
Obstet Gynaecol 1992;32:377–8.
[92] Mehl LE. Electrical injury from tasering and miscarriage. Acta Obstet Gynecol Scand 1992;
71:118–23.
[93] Jaffe R, Fejgin M, Ben Aderet N. Fetal death in early pregnancy due to electric current.
Acta Obstet Gynecol Scand 1986;65(3):283.
[94] Fatovich DM. Electric shock in pregnancy. J Emerg Med 1993;11:175–7.
[95] Leiberman JR, Mazor M, Molcho J, et al. Electrical accidents during pregnancy. Obstet
Gynecol 1986;67(6):861–3.
[96] Einarson A, Bailey B, Inocencion G, et al. Accidental electric shock in pregnancy: a prospec-
tive cohort study. Am J Obstet Gynecol 1997;176(3):678–81.
[97] Fish RM. Electric injury. Part III: cardiac monitoring indications, the pregnant patient, and
lightning. J Emerg Med 2000;18:181–7.
[98] Mabogunje OA. Burns injuries during pregnancy: an African series. J Natl Med Assoc
1990;82:641–4.
[99] Polko LE, McMahon MJ. Burns in pregnancy. Obstet Gynecol Surv 1998;53(1):50–6.
[100] Smith G, Ledingham IM, Sharp GR, et al. Treatment of coalgas poisoning with oxygen at
two atmospheres pressure. Lancet 1962;1:816–9.
[101] Reiss G. Thermal injuries. In: Lopez-Viego MA, editor. The Parkland trauma handbook.
St. Louis (MO): Mosby; 1994. p. 389–412.
[102] Maghsoudi H, Samnia R, Garadaghi A, et al. Burns in pregnancy. Burns 2006;32(2):
246–50.
[103] Harris MB, Sethi RK. The initial assessment and management of the multiple-trauma
patient with an associated spine injury. Spine 2006;31(Suppl 11):S9–15.
[104] Tsutsumi S, Ueta T, Shiba K, et al. Effects of the Second National Acute Spinal Cord Injury
Study of high-dose methylprednisolone therapy on acute cervical spinal cord injuryd
results in spinal injuries center. Spine 2006;31(26):2992–6.
[105] Popov I, Ngambu F, Mantel G, et al. Acute spinal cord injury in pregnancy: an illustrative
case and literature review. J Obstet Gynaecol 2003;23(6):596–8.
[106] Bullock R, Chesnut RM, Clifton G, et al. Guidelines for the management of severe
traumatic brain injury. The Brain Trauma Foundation and the American Association of
Neurological Surgeons, Joint Section of Neurotrauma and Critical Care. Journal of
Neurotrauma 2007;24(Suppl 1):S1–104.
[107] Cremer OL, van Dijk GW, van Wensen E, et al. Effect of intracranial pressure monitoring
and targeted intensive care on functional outcome after severe head injury. Crit Care Med
2005;33:2207–13.
[108] Muizelaar JP, Marmarou A, Ward DJ, et al. Adverse effects of prolonged hyperventilation
in patients with severe head injury: a randomized control trial. J Neurosurg 1991;75:731–9.
[109] Deitch EA, Dayal SD. Intensive care unit management of the trauma patient. Crit Care
Med 2006;34(9):2294–301.
[110] Manno EM, Farmer JC. Acute brain injury: if hypothermia is good, then is hyperthermia
bad? Crit Care Med 2004;32:1489–95.
[111] Bullock MR, Chesnut R, Ghajar J, et al. Surgical management of traumatic parachymal
lesions. Neurosurgery 2006;58:S25–46.
[112] Okie S. Traumatic brain injury in the war zone. N Engl J Med 2005;352:2043–7.
 TRAUMA IN PREGNANCY 583

[113] Muench MV, Canterino JC. Physical abuse/assault and psychological trauma. In: Quirk
JG, Garry D, Figueroa R, editors. Trauma and pregnancy. New York: Cambridge Univer-
sity Press, in press.
[114] Rachana C, Suriaya K, Hisham AS, et al. Prevalence and complications of physical violence
during pregnancy. Eur J Obstet Gynecol Reprod Biol 2002;103:26–9.
[115] Randall T. Domestic violence intervention calls for more than treating injuries. JAMA
1990;264:939–40.
[116] Canterino JC, VanHorn LG, Harrigan JF, et al. Domestic abuse in pregnancy: a compari-
son of a self-completed domestic abuse questionnaire with a directed interview. Am J Obstet
Gynecol 1999;181(5Pt 1):1049–51.
 Obstet Gynecol Clin N Am
34 (2007) 585–597

Cardiopulmonary Resuscitation
in Pregnancy
Emad Atta, MDa, Michael Gardner, MD, MPHb,*
a
Department of Obstetrics and Gynecology, Medical College of Georgia,
1120 15th Street, Augusta, GA 30912, USA
b
Division of Maternal Fetal Medicine, Department of Gynecology and Obstetrics,
Emory University School of Medicine, 69 Jesse Hill Jr. Drive SE,
Atlanta, GA 30303, USA

Cardiac arrest in pregnant patients is an infrequent event that obstetri-

cians and critical care medicine practitioners will encounter in their careers.
Because outcomes depend on the underlying cause of the arrest and the
speed of resuscitation efforts, an understanding of basic resuscitation prin-
ciples and the specific challenges of an arrest in the pregnant woman is
required to achieve a successful outcome. During attempted resuscitation of
a pregnant woman, providers have two potential patients, the mother and
the fetus [1]. The best hope of fetal survival is maternal survival. For the crit-
ically ill patient who is pregnant, rescuers must provide appropriate resuscita-
tion, with consideration of the physiologic changes caused by pregnancy.
The true incidence of cardiac arrest during pregnancy is not known but
has been estimated to be about 1 in 30,000 pregnancies [2]. Most informa-
tion in the literature regarding cardiac arrest during pregnancy is in the
form of case reports and case series. Some of the important etiologic factors
causing cardiac arrest in the pregnant population are listed in Box 1. These
factors differ somewhat from causes of cardiac arrest in nonpregnant
patients. In the developed world, including the United States, the major
causes of maternal mortality, in order of decreasing frequency, are venous
thromboembolism, severe pregnancy-induced hypertension (pre-eclampsia
or eclampsia), sepsis, amniotic fluid embolism, hemorrhage, trauma, iatro-
genic causes including complications of anesthesia and drug errors or al-
lergy, and maternal heart disease [3,4].

* Corresponding author.
E-mail address: michael.gardner@emory.edu (M. Gardner).

0889-8545/07/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ogc.2007.06.008 obgyn.theclinics.com
586 ATTA & GARDNER

Box 1. Major causes of cardiac arrest during pregnancy

Venous thromboembolism
Pregnancy-induced hypertension
Sepsis
Amniotic fluid embolism
Hemorrhage
Placental abruption
Placenta previa
Uterine atony
Disseminated intravascular coagulation
Trauma
Iatrogenic causes
Medication errors or allergy
Anesthetic complications
Hypermagnesemia
Pre-existing heart disease
Congenital
Acquired

Amniotic fluid embolism (AFE), also called ‘‘anaphylactoid reaction of

pregnancy,’’ is a rare complication of late gestation and the immediate post-
partum period that deserves special mention. AFE, which occurs in an esti-
mated 1:8000 to 1:80,000 pregnancies has a high mortality rate, ranging
from 50% to 80%, and may rapidly precipitate cardiac arrest. In the anal-
ysis of the United States registry for AFE, cardiac arrest occurred in 87% of
cases [5,6].
Another increasingly important contributor to cardiac arrest in preg-
nancy is the increasing average age of pregnant women in the United States.
Because of personal choice or because of the effects of assisted reproductive
technologies; pregnancy in women 45 years and older is much more com-
mon than it was a generation ago. This change tends to add patients who
have chronic medical conditions that may be less prevalent in younger
women. These chronic conditions can lead to complications in the preg-
nancy and, albeit rarely, to cardiac arrest.

Physiologic changes of pregnancy and implications for resuscitation

The cardiovascular and respiratory changes that occur during pregnancy
are summarized here to emphasize their implications for resuscitation after
cardiac arrest (Box 2). Cardiac output increases by 30% to 50% by 32
weeks’ gestation [7]. Heart rate and resting oxygen consumption also are
 CARDIOPULMONARY RESUSCITATION IN PREGNANCY 587

Box 2. Physiologic changes in late pregnancy affecting

cardiopulmonary resuscitation
Respiratory
Increased ventilation
Increased oxygen demand
Reduced chest compliance
Reduced functional residual capacity
Cardiovascular
Incompetent gastroesophageal (cardiac) sphincter
Increased intragastric pressure
Increased risk of regurgitation

increased, whereas systemic vascular resistance and plasma oncotic pressure

decrease as compared with the nongravid state. Uteroplacental blood flow
increases during pregnancy so that the uterus receives up to 30% of cardiac
output, as compared with less than 2% in the nonpregnant state. After
spontaneous delivery, cardiac output increases by 60% to 80% of prelabor
values [8]. This increase is smaller after cesarean delivery (about 30% of
prelabor values), possibly because of the effects of anesthetics and blood
loss.
Another critically important physiologic factor that has an impact on the
effectiveness of cardiopulmonary resuscitation (CPR) and hemodynamic
support in pregnant patients is aortocaval compression by the gravid uterus
during the latter half of pregnancy [9,10]. In late pregnancy, the vena cava
may be obstructed completely in most women when in the supine position,
forcing venous return to flow through azygous lumbar and paraspinal veins.
About 10% of pregnant women manifest the supine hypotensive syndrome,
in which syncope, hypotension, and bradycardia occur when supine because
of aortocaval compression. Stroke volume and cardiac output increase by
25% to 30% when late-term pregnant patients move from supine to lateral
decubitus position.
The respiratory changes of pregnancy include increased minute ventila-
tion caused by the effects of progesterone on respiratory drive, increased
oxygen consumption, and a restrictive ventilatory defect caused by upward
displacement of the diaphragm. Arterial blood gases during late pregnancy
normally reflect a state of compensated respiratory alkalosis. The mechani-
cal effects of the gravid uterus and hypertrophied breasts result in reduced
functional residual capacity and reduced chest wall compliance. Reduced
functional residual capacity and increased oxygen consumption can lead
to precipitous oxygen desaturation if hypoventilation occurs [9].
Understanding the respiratory changes of pregnancy is essential during
the management of cardiac arrest [11,12]. These changes necessitate quick
588 ATTA & GARDNER

establishment of oxygenation and ventilation. Increased oxygen consump-

tion leads to increased rates of arterial oxygen desaturation in the parturient
who becomes apneic. Because of the hormonal and physical changes of
pregnancy, patients are at increased risk for difficult ventilation and failed
intubation. Increased levels of progesterone lead to delayed gastric empty-
ing, increasing the risk for aspiration during mask ventilation and intuba-
tion. Although many centers require that patients take nothing by mouth
during labor, patients often present in spontaneous labor after consuming
a large meal. Edema of the upper airway, increased breast size, and gener-
alized weight gain can delay the establishment of adequate ventilation and
intubation. It is essential that oxygenation and ventilation be restored expe-
ditiously while maintaining cricoid pressure. Moreover, it is imperative to
intubate the patient as soon as possible to maximize oxygenation and min-
imize the risk of aspiration [8]. This need for rapid intubation is a key dif-
ference between the pregnant women in cardiac arrest and nonpregnant
patients.

Differential diagnoses
The same reversible causes of cardiac arrest that occur in nonpregnant
women can occur during pregnancy, but providers should be familiar with
pregnancy-specific diseases and procedural complications. Obviously, pro-
viders should try to identify these common and reversible causes of cardiac
arrest in pregnancy during resuscitation attempts. Some possible causes of
cardiac arrest are discussed in this section.

Excess magnesium sulfate

Iatrogenic overdose is possible in women who have eclampsia and receive
magnesium sulfate, particularly if the woman becomes oliguric. Administra-
tion of calcium gluconate (1 ampoule or 1 g) is the treatment of choice for
magnesium toxicity [13]. Empiric calcium administration may be lifesaving.

Acute coronary syndromes

Pregnant women may experience acute coronary syndromes, typically in
association with other medical conditions. Because fibrinolytics are relatively
contraindicated in pregnancy, percutaneous coronary intervention is the re-
perfusion strategy of choice for ST-elevation myocardial infarction [14].

Pre-eclampsia/eclampsia
Pre-eclampsia/eclampsia develops after the twentieth week of gestation
and can produce severe hypertension and ultimately diffuse organ system
failure. If untreated it may result in maternal and fetal morbidity and mor-
tality. Uncontrolled blood pressures can lead to stroke and subsequent car-
diac arrest. Arrest during eclamptic seizures is relatively rare, particularly if
the seizures are treated adequately and maternal oxygenation is maintained.
 CARDIOPULMONARY RESUSCITATION IN PREGNANCY 589

Aortic dissection
Pregnant women are at increased risk for spontaneous aortic dissection.

Life-threatening pulmonary embolism and stroke

The successful use of fibrinolytic therapy for a massive, life-threatening
pulmonary embolism and ischemic stroke has been reported in pregnant
women [15].

Amniotic fluid embolism

Clinicians have reported successful use of cardiopulmonary bypass for
women who have life-threatening amniotic fluid embolism during labor
and delivery.

Trauma
Pregnant women are not exempt from the accidents and violence that
afflict much of society. Domestic violence also increases during pregnancy;
in fact, homicide and suicide are leading causes of mortality during preg-
nancy, and motor vehicle accidents cause more maternal deaths in the
United States than any other cause. Identification of the pregnancy early
in the resuscitative effort of the pregnant trauma patient is critical. This
statement may seem obvious, but because so many women in the United
States are overweight and because trauma victims often arrive to the hospi-
tal unconscious and alone, a pregnancy may not be readily apparent, even
after fetal viability. Therefore, the resuscitation team always must remember
the possibility that any woman of childbearing age (an age range, as earlier
noted, that is increasing) may be pregnant.

Cardiopulmonary resuscitation and advanced cardiac life support in

pregnant patients
In general, resuscitation algorithms during cardiac arrest are the same for
pregnant patients as for nonpregnant patients, but with some exceptions
[16]. Principal among the modifications for the late-term pregnant woman
are more aggressive airway management, attention to lateral displacement
of the uterus, caution in the use of sodium bicarbonate, and early consider-
ation of perimortem cesarean delivery (Box 3).
Cardiac output during CPR is estimated to be about 30% of normal, so
uteroplacental blood flow is reduced markedly during cardiac arrest even
with optimal performance of chest compressions [17]. CPR is performed
in the same way on pregnant patients as on nonpregnant patients, except
that in the second half of pregnancy an attempt to relieve aortocaval com-
pression in the supine position is essential to restoring effective circulation.
Rees and Willis [18] measured the force achieved with chest compressions
performed on a manikin in the decubitus position at various angles of
590 ATTA & GARDNER

Box 3. Specific difficulties in pregnant patients

Airway: patient needs to be inclined laterally for
Suction or aspiration
Removing dentures or foreign bodies
Inserting airways
Breathing
Greater oxygen requirement
Reduced chest compliance
More difficult to see rise and fall of chest
More risk of regurgitation and aspiration
Circulation: external chest compression is difficult because
Ribs are flared
Diaphragm is raised
Patient is obese
Breasts are hypertrophied
Supine position causes inferior vena cava compression by the
gravid uterus

inclination. The resuscitative force decreased from 67% of the rescuer’s

body weight with the manikin in the supine position to 36% in the full lat-
eral position. At an angle of 27 , the maximal possible resuscitative force
during CPR was 80% of that which could be achieved with in the supine
position. This study led to the development of the Cardiff resuscitation
wedge, a wooden frame inclined at a 27 angle and specifically designed
for performing CPR on pregnant patients. This apparatus may not always
be available in critical care units or in the emergency department. Obstetric
practitioners should ensure the availability of the wedge in the ICU, the op-
erating suite, and the emergency room as well as in labor and delivery suites,
because all these sites may be the locale of a code arrest in a pregnant pa-
tient. If one is faced with an arrest in a gravid woman beyond 20 weeks
and the wedge is not available, alternative maneuvers to relieve aortocaval
compression include manual displacement of the uterus to the left and
upward while the patient is supine, use of a wedge such as a bed sheet placed
under the right hip, use of a ‘‘human wedge’’ (one rescuer kneels on the floor
or another surface with the woman’s back positioned against the rescuer’s
thighs) [19]. All these measures can relieve some of the aortocaval
compression.
Available evidence suggests that defibrillation energy requirements do
not change significantly during pregnancy. In the only study to address
this question directly, Nanson and colleagues [20] assessed transthoracic
impedance, as measured by a defibrillator, in 45 women at term pregnancy
 CARDIOPULMONARY RESUSCITATION IN PREGNANCY 591

and repeated the measurements at 6 to 8 weeks after delivery in 42 of the

women. They found no significant difference in the mean transthoracic im-
pedance before or after delivery. The same defibrillation regimens recom-
mended in the advanced cardiac life-support (ACLS) algorithm for
appropriate cardiac arrhythmias, such as ventricular fibrillation or pulseless
ventricular tachycardia, are recommended for pregnant patients.
Supplemental oxygen should be administered at a concentration of 100%
during CPR in pregnant and nonpregnant patients. As mentioned previ-
ously, rapid control of the airway through performance of endotracheal in-
tubation early in the resuscitation effort is highly recommended. In addition
to the increased susceptibility to hypoxia for the mother and fetus, pregnant
patients also are at increased risk for aspiration of gastric contents caused
by delayed gastric emptying and reduced lower esophageal sphincter tone.
This risk may be exacerbated further by gastric distention from air insuffla-
tion during bag-mask ventilation.
The use of sodium bicarbonate to reverse metabolic acidosis during car-
diac arrest has been questioned; its role in managing maternal acidosis is
controversial also [21]. Animal studies suggest that bicarbonate crosses
the placenta poorly (although this finding may not be true in humans).
Rapid correction of maternal (but not fetal) acidosis could lead to reduced
compensatory hyperventilation and normalization of maternal PaCO2, which
could result in a concomitant increase in fetal PaCO2 and potential worsen-
ing of fetal acidosis. Available evidence, however, suggests that the fetus
may tolerate significant respiratory acidosis for short periods. Restoration
of effective maternal circulation, with subsequent correction of hypoxia, is
the most effective way to correct fetal acidosis during maternal cardiac
arrest.
There is little information regarding pharmacologic therapy during
ACLS in pregnant patients. The use of a-adrenergic agents theoretically
may reduce uteroplacental blood flow, but their actual clinical effect is un-
known. In general, the same protocols for pharmacologic management of
ACLS should be used in pregnant and nonpregnant patients with cardiac
arrest. The best chance for survival for the mother and fetus depends on
rapid resuscitation of the mother.

Perimortem cesarean delivery and outcomes

Physicians must decide whether to attempt emergent cesarean delivery in
the resuscitation of pregnant patients with cardiac arrest in whom initial re-
suscitative efforts are not immediately successful. Timing and speed of the
procedure are keys to optimizing outcome and limiting adverse neurologic
sequelae in survivors.
Cesarean delivery is one of the oldest surgical procedures in history, with
literature dating back to at least 800 BCE [22]. Before the twentieth century,
592 ATTA & GARDNER

however, the phrase ‘‘postmortem cesarean’’ would have been redundant,

because the procedure was never undertaken unless the mother was dead
or moribund.
Initially, the Roman decree (Lex Cesare, or law of Caesar) that unborn
infants should be separated from their mothers’ bodies was for purposes
of religious ritual rather than attempts for survival of either the newborn
or mother. Some infants did survive, and indeed, several mythological
and ancient historical figures were reported to have been born in this fash-
ion, including the Greek physician Asklepios, ‘‘from the womb of dead Kor-
onis.’’ During the late nineteenth and early twentieth centuries, case reports
began to arise of perimortem cesarean delivery successfully salvaging the
fetus, and the procedure began to be considered seriously as a legitimate
medical intervention. Well into the twentieth century, the salvage rate was
very low, and therefore authors on the subject advocated it only after all
other resuscitative measures had failed.
During the 1980s, several authors reported unexpected maternal recov-
eries after postmortem cesarean deliveries [23,24]. This experience suggested
that the procedure actually might improve, rather than worsen, a mother’s
chance of survival during a collapse.
Katz and colleagues [25] reviewed the medical literature about perimor-
tem cesarean deliveries that were reported through 1985. Of 188 surviving
infants, they identified 61 cases in which the data showed the time from
death of the mother to delivery of the infant. They found that 70% of sur-
viving neonates were delivered within 5 minutes of maternal death, and 93%
were delivered within 15 minutes. Some infants survived when delivery
occurred more than 21 minutes after maternal death, but the neurologic def-
icits in these infants were more frequent and more severe. Lopez-Zeno and
colleagues [26] reported a case of perimortem cesarean delivery after 22 min-
utes of CPR in a mother who developed cardiac arrest secondary to a fatal
gunshot wound. The infant survived and was described as clinically normal
at 18 months of age. Based on their findings, Katz and colleagues [25] rec-
ommended initiation of cesarean delivery within 4 minutes of maternal car-
diac arrest if circulation has not been restored and recommended fetal
delivery within 5 minutes. These recommendations have been supported
by other investigators and form the basis of the ‘‘4-minute rule.’’ Given
the number of reports of neonatal survival without adverse neurologic
sequelae when delivery occurred well after 5 minutes of maternal cardiac
arrest, this rule should not be taken as absolute. The outcomes of infants
delivered by perimortem cesarean delivery are summarized in Table 1.
Estimated gestational age is an important factor in predicting prognosis
for infants after perimortem cesarean deliveries. The threshold for expected
fetal viability may vary slightly between institutions but generally is consid-
ered to be around 24 weeks of gestation. If the gestational age cannot be de-
termined from the available medical or prenatal history, practical and rapid
methods of its assessment in the emergency setting include calculation based
 CARDIOPULMONARY RESUSCITATION IN PREGNANCY 593

Table 1
Outcomes of infants delivered by perimortem cesarean delivery
Time (in minutes) No. infants surviving % Surviving intact
0–5 45 98
6–15 18 83
16–25 9 33
26–35 4 25
36 þ 1 0
Data modified from Katz VL, Dotters DJ, Droegemueller W. Perimortem cesarean delivery.
Obstet Gynecol 1986;68:571–6, and Clark SL, Hankins GDV, Dudley DA, et al. Amniotic fluid
embolism: analysis of the National Registry. Am J Obstet Gynecol 1995;172:1158–69.

on the mother’s last menstrual period or measurement of fundal height. Be-

tween 20 and 36 weeks of gestation, the fetal age in weeks is approximated
by the distance in centimeters from the pubic symphysis to the top of the
uterine fundus when the mother is supine. A fundal height at the level of
the umbilicus corresponds to 20 weeks’ gestation. In the settings of multi-
parity, extreme obesity, abdominal distention from other causes, or intra-
uterine growth retardation, these methods of estimating gestational age
may be unreliable.
Although the primary goal of cesarean delivery in the perimortem period
has been survival of the fetus, the procedure also may have a role in saving
both the mother and infant. Because of the impact of aortocaval compres-
sion by the gravid uterus on the efficacy of CPR, delivery of the fetus
may improve maternal cardiac output significantly in addition to improving
survival of the fetus. In numerous case reports, emergent cesarean delivery
in the setting of apparently refractory maternal cardiac arrest has resulted in
survival of the infant and mother because of more effective resuscitation of
the mother after delivery. Finegold and colleagues [27] reported a case of
a previously healthy 35-year-old woman who was at 39 weeks’ gestation
and experienced cardiac arrest soon after rupture of membranes. Emergent
cesarean delivery was performed 15 minutes after the cardiac arrest, with
immediate recovery of maternal pulse and blood pressure. The mother
and infant survived and had normal neurologic function. Other investiga-
tors have described similar situations and favorable outcomes, although
such case reports cannot permit firm conclusions about whether cesarean
delivery improves maternal outcome in the setting of late-term cardiac
arrest.
If cardiac arrest occurs earlier in pregnancy, it is not known whether per-
formance of cesarean delivery to produce a previable fetus is beneficial to
maternal outcome. With significantly smaller fetal–placental mass, the he-
modynamic benefits to the mother would not be expected to be as significant
as later in pregnancy. In general, perimortem cesarean delivery is not re-
commended in cases with an estimated gestational age of less than 24 weeks,
and efforts should focus on optimizing resuscitation performance and
594 ATTA & GARDNER

restoration of spontaneous circulation to provide the best hope of recovery

for mother and fetus. In one case report, maternal and fetal survival
occurred after prolonged maternal cardiac arrest at 15 weeks’ gestation
secondary to accidental lidocaine overdose. CPR was performed for
22 minutes before return of spontaneous circulation. The patient recovered
neurologic function and had a normal, spontaneous vaginal delivery at
40 weeks’ gestation, delivering a neurologically normal infant.

Decision making for a perimortem cesarean delivery

The resuscitation team should consider several maternal and fetal factors
in determining the need for an emergency hysterotomy. Although the gravid
uterus reaches a size that will begin to compromise aortocaval blood flow at
approximately 20 weeks of gestation, fetal viability begins at approximately
24 to 25 weeks. Portable ultrasonography, available in many emergency de-
partments, may, in experienced hands, aid in determining gestational age as
well as placental location, fetal lie, and the presence of fetal cardiac activity.
The use of ultrasound should not delay the decision to perform emergency
hysterotomy, however, and may be impractical in the setting of maternal
cardiac arrest. If the gestational age is less than 20 weeks, urgent cesarean
delivery should not be considered, because a gravid uterus of this size is un-
likely to compromise maternal cardiac output significantly. A more difficult
decision may be when the mother’s gestational age is approximately 20 to
23 weeks. Performing an emergency hysterotomy to enable successful resus-
citation of the mother may be reasonable, although minimal data exist to
support the premise. Obviously, survival of the infant is unlikely at this
gestational age and should not be a factor in the decision-making process.
After 24 to 25 weeks, cesarean delivery may contribute to saving the life
of the mother and allows an attempt of resuscitation of the infant.
The critical point is that both mother and infant will be lost if blood flow to
the mother’s heart cannot be restored. Four to 5 minutes is the maximum time
rescuers have to determine if the arrest can be reversed by basic life support
and ACLS interventions. The rescue team need not wait for this time to elapse
before initiating emergency hysterotomy. Unfortunately, recent reports doc-
ument long intervals between an urgent decision for hysterotomy and actual
delivery of the infant, far exceeding the obstetric guideline of 30 minutes [28].
Establishment of intravenous access and an advanced airway typically re-
quires several minutes. In most cases the actual cesarean delivery cannot
proceed until after administration of intravenous medications and endotra-
cheal intubation. Resuscitation team leaders should activate the protocol for
an emergency cesarean delivery as soon as cardiac arrest is identified in the
pregnant woman. By the time the team leader is poised to deliver the baby,
intravenous access has been established, initial medications have been ad-
ministered, an advanced airway is in place, and the immediate reversibility
of the cardiac arrest has been determined.
 CARDIOPULMONARY RESUSCITATION IN PREGNANCY 595

Features of the cardiac arrest

The following features of the cardiac arrest can increase the infant’s
chance for survival:
 Short interval between the mother’s arrest and the infant’s delivery
 No sustained prearrest hypoxia in the mother
 Minimal or no signs of fetal distress before the mother’s cardiac arrest
 Aggressive and effective resuscitative efforts for the mother
 Cesarean delivery performed in a medical center with a neonatal ICU
The professional setting
 Are appropriate equipment and supplies available?
 Is emergency cesarean delivery within the rescuer’s procedural range of
experience and skills?
 Are skilled neonatal/pediatric support personnel available to care for the
infant, especially if the infant is not full term?
 Are obstetric personnel immediately available to support the mother
after delivery?
The technique of perimortem cesarean delivery requires speed and deci-
siveness. CPR must be continued during delivery, and the procedure should
not be delayed for attempts to obtain consent from next of kin. Most experts
agree that in the setting of maternal cardiac arrest, the doctrine of emer-
gency or implied consent applies, and the best interests of the child take pre-
cedence. Katz and colleagues [25] note that there have been no legal findings
of liability against physicians in the United States for performing a postmor-
tem cesarean delivery.

Summary
Successful resuscitation of a pregnant woman and survival of the fetus
require prompt and excellent CPR with some modifications in basic and
advanced cardiovascular life-support techniques. By the twentieth week of
gestation, the gravid uterus can compress the inferior vena cava and the
aorta, obstructing venous return and arterial blood flow. Rescuers can re-
lieve this compression by positioning the woman on her side or by pulling
the gravid uterus to the side. Defibrillation and medication doses used for
resuscitation of the pregnant woman are the same as those used for other
adults in pulseless arrest. Electric cardioversion during pregnancy has
been described in the literature and seems safe for the fetus. The physiologic
changes in pregnancy do not change defibrillation requirements for adult
defibrillation.
Rescuers should consider the need for perimortem cesarean delivery as
soon as the pregnant woman develops cardiac arrest, because rescuers
should be prepared to proceed with the hysterotomy if the resuscitation is
not successful within minutes.
596 ATTA & GARDNER

Although pregnancy and delivery in the United States usually are safe for
the mother and her newborn child, serious maternal complications, includ-
ing cardiac arrest, can and do occur in the prenatal, intrapartum, and post-
partum periods. The busy clinical obstetrician can expect to encounter this
complication in his or her career. It is incumbent on the obstetrician to be
aware of the special circumstances of resuscitation of the gravid woman
to assist emergency medicine and critical care physicians in reviving the
patient. Moreover, understanding the decision process leading to the perfor-
mance of a perimortem cesarean and the actual performance of the cesarean
delivery clearly are the responsibilities of the obstetrician.

References
[1] Morris S, Stacey M. Resuscitation in pregnancy. BMJ 2003;327:1277–9.
[2] Department of Health, Welsh Office, Scottish Office Department of Health, Department of
Health and Social Services, Northern Ireland. Why mothers die. Report on confidential
enquiries into maternal deaths in the United Kingdom 2000–2002. London (UK): The Sta-
tionery Office; 2004.
[3] Zaritsky A, Morley P. The evidence evaluation process for the 2005 International Consensus
Conference on cardiopulmonary resuscitation and emergency cardiovascular care science
with treatment recommendations. Circulation 2005;112:III-128–30.
[4] The 2005 International Consensus Conference on cardiopulmonary resuscitation and emer-
gency cardiovascular care science with treatment recommendations, hosted by the American
Heart Association in Dallas, Texas, January 23–30, 2005. Circulation 2005;112:IV-150–3.
[5] Clark SL, Hankins GDV, Dudley DA, et al. Amniotic fluid embolism: analysis of the
National Registry. Am J Obstet Gynecol 1995;172:1158–69.
[6] Stanten RD, Iverson LI, Daugharty TM, et al. Amniotic fluid embolism causing catastrophic
pulmonary vasoconstriction: diagnosis by transesophageal echocardiogram and treatment
by cardiopulmonary bypass. Obstet Gynecol 2003;102:496–8.
[7] Poole JH, Long J. Maternal mortalityda review of current trends. Crit Care Nurs Clin
North Am 2004;16:227–30.
[8] Malampalli A, Powner DJ, Gardner M. CPR and somatic support of the pregnant patient.
Crit Care Clin 2004;20:747–63.
[9] Doan-Wiggins L. Resuscitation of the pregnant patient suffering sudden death. In: Paradis
NA, Halperin HR, Nowak RM, editors. Cardiac arrest: the science and practice of resusci-
tation medicine. Baltimore (MD): Williams & Wilkins; 1997. p. 812–9.
[10] Kerr MG. The mechanical effects of the gravid uterus in late pregnancy. J Obstet Gynaecol
Br Commw 1965;72:513–29.
[11] Johnson MD, Luppi CJ, Over DC. Cardiopulmonary resuscitation. In: Gambling DR,
Douglas MJ, editors. Obstetric anesthesia and uncommon disorders. Philadelphia: WB
Saunders; 1998. p. 51–74.
[12] Whitty JE. Maternal cardiac arrest in pregnancy. Clin Obstet Gynecol 2002;45:377–92.
[13] Munro PT. Management of eclampsia in the accident and emergency department. J Accid
Emerg Med 2000;17:7–11.
[14] Dapprich M, Boessenecker W. Fibrinolysis with alteplase in a pregnant woman with stroke.
Cerebrovasc Dis 2002;13:290–4.
[15] Trukhacheva E, Scharff M, Gardner M, et al. Massive pulmonary embolism in pregnancy
treated with tissue plasminogen activator. Obstet Gynecol 2005;106:1156–8.
[16] American Heart Association in collaboration with International Liaison Committee
on Resuscitation. Guidelines 2000 for cardiopulmonary resuscitation and emergency
 CARDIOPULMONARY RESUSCITATION IN PREGNANCY 597

cardiovascular care: International Consensus on Science, part 8: advanced challenges in re-

suscitation: section 3: advanced challenges in ECC. Circulation 2000;102(Suppl I):I229–52.
[17] Cummins RO, Hazinski MF, Zelop CM. Cardiac arrest associated with pregnancy. In:
Cummins R, Hazinski M, Field J, editors. ACLSdthe reference textbook. Dallas
(TX): American Heart Association; 2003. p. 143–58.
[18] Rees GA, Willis BA. Bioimpedance measurement of cardiac output. Eur J Obstet Gynecol
Reprod Biol 1990;36:11–7.
[19] Goodwin AP, Pearce AJ. The human wedge. A manoeuvre to relieve aortocaval compression
during resuscitation in late pregnancy. Anaesthesia 1992;47:433–4.
[20] Nanson J, Elcock D, Williams M, et al. Do physiological changes in pregnancy change de-
fibrillation energy requirements? Br J Anaesth 2001;87:237–9.
[21] Bar-Joseph G, Ambramson NS, Jansen-McWilliams L, et al. Clinical use of sodium bicar-
bonate during cardiopulmonary resuscitationdis it used sensibly? Resuscitation 2002;54:
47–55.
[22] Weber CE. Postmortem cesarean section: review of the literature and case reports. Am J
Obstet Gynecol 1971;110:158–65.
[23] Strong TH Jr, Lowe RA. Perimortem cesarean section. Am J Emerg Med 1989;7:489–94.
[24] O’Connor RL, Sevarino FB. Cardiopulmonary arrest in the pregnant patient: a report of
a successful resuscitation. J Clin Anesth 1994;6:66–8.
[25] Katz VL, Dotters DJ, Droegemueller W. Perimortem cesarean delivery. Obstet Gynecol
1986;68:571–6.
[26] Lopez-Zeno J, Carlo W, O’Grady JP, et al. Infant survival following delayed postmortem
cesarean delivery. Obstet Gynecol 1990;76:991–2.
[27] Finegold H, Darwich A, Romeo R, et al. Successful resuscitation after maternal cardiac ar-
rest by immediate cesarean section in the labor room. Anesthesiology 2002;96:1278–80.
[28] Bloom SL, Leveno KJ, Spong CY, et al. Decision-to-incision times and maternal and infant
outcomes. Obstet Gynecol 2006;108(1):6–11.
 Obstet Gynecol Clin N Am
34 (2007) 599–616

Angiographic and Interventional Options

in Obstetric and Gynecologic
Emergencies
Filip Banovac, MDa,*, Ralph Lin, BSa,
Dimple Shah, MDb, Amy White, MDb,
Jean-Pierre Pelage, MD, PhDc, James Spies, MDa
a
Department of Radiology, Georgetown University Hospital, 3800 Reservoir Road NW,
Washington, DC 20007, USA
b
Georgetown University School of Medicine, 3800 Reservoir Road NW,
Washington, DC 20007, USA
c
Department of Vascular and Body Imaging, Hôpital Lariboisiere, 2, rue Ambroise-Pare,
75475, Paris Cedex 10, France

The role of endovascular techniques for diagnosis and treatment of obstet-

rical and gynecologic emergencies has evolved over last 3 decades. Obstetri-
cians and gynecologists, using surgical techniques and medical management,
can adequately manage most vascular emergencies. However, in certain sit-
uations, angiographic intervention can play a complementary role in patient
management. In this article, the authors describe the technique of pelvic
arterial embolization and the role of embolization and catheterization for
specific kinds of postpartum hemorrhage and gynecologic vascular emergen-
cies. In addition, the authors provide a brief review of the literature on pelvic
arterial embolization in the setting of postpartum hemorrhage . Finally, the
authors discuss some complications and short- and intermediate-term out-
comes of embolization, particularly as they affect fertility.

Postpartum hemorrhage
Postpartum hemorrhage is among the most common causes of maternal
morbidity and mortality. In the United States, postpartum hemorrhage
ranks among the top three causes of maternal death [1]. Bleeding of 500

* Corresponding author.
E-mail address: banovac@isis.georgetown.edu (F. Banovac).

0889-8545/07/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ogc.2007.06.004 obgyn.theclinics.com
600 BANOVAC et al

mL or more following a vaginal delivery, or 1000 mL or more following a ce-

sarean section, meets the definition of postpartum hemorrhage. Because ob-
jective measurement of blood loss can be difficult, a more qualitative
definition includes a clinical need for a transfusion or a 10% hematocrit
drop between admission and the postpartum period [2,3].
The first line of treatment for postpartum hemorrhage includes conserva-
tive measures, such as administration of uterotonic medications, laceration
repair, uterine packing, and correction of underlying coagulopathies. If
these measures failed, obstetricians would often attempt surgical ligation
of the arterial supply to the uterus, or they would perform a hysterectomy
with associated loss of fertility.
In the last 30 years, a new angiographic approach for treatment of
postpartum hemorrhage has emerged. Pelvic arterial embolization, after
emerging as a treatment option to control and prevent pregnancy-related
hemorrhage, has been established to be safe and effective [4–8]. In
appropriate circumstances, pelvic arterial embolization provides some
advantages in management of postpartum hemorrhage. In addition to
providing a high technical success rate, this angiographic approach, com-
pared to other options, also offers a greater likelihood of preserving
fertility.

Common causes of obstetric hemorrhage

Postpartum hemorrhage is frequently categorized as either early or
delayed onset. Early postpartum hemorrhage occurs within the first 24
hours after delivery while delayed postpartum hemorrhage is commonly
defined as bleeding after 24 hours but within 6 weeks after delivery. The
most common cause of early postpartum hemorrhage is uterine atony
(Fig. 1) [9,10], although genital tract lacerations can also cause significant
bleeding in the early postpartum period [11]. Delayed postpartum hemor-
rhage usually occurs because of retained placental fragments.
Arteriovenous malformations, invading trophoblastic tissue, complica-
tions related to evacuation of ectopic pregnancies, and instrumentation in
the peripartum period are also known causes of pregnancy-related hemor-
rhage. Congenital arteriovenous malformations are not a complication
of pregnancy and are discussed separately later in this article. However,
acquired arteriovenous malformations can be seen after uterine curettage
or after removal of an intrauterine device, and are an uncommon cause of
hemorrhage (Fig. 2) [12,13]. Cervical and abdominal ectopic pregnancies
can also be difficult to evacuate surgically and thus are occasionally associ-
ated with significant hemorrhage.
Early postpartum hemorrhage usually occurs in the immediate postpar-
tum period while the patient is still under supervision of the obstetric team
and thus management can start immediately. The usual clinical
management includes uterine packing, repair of visible lacerations,
 ANGIOGRAPHIC AND INTERVENTIONAL OPTIONS 601

Fig. 1. Angiographic appearance of uterine atony. (A) Usual appearance of uterine atony with-
out contrast extravasation. (B) Postembolization image with occlusion of the anterior division
of the internal iliac artery. After selective embolization of the uterine artery, large gelatin sponge
pledgets were placed into the anterior division of the internal iliac to control the bleeding from
the lower uterine segment vaginal and cervical branches.

administration of uterotonic drugs, and correction of underlying coagulo-

pathies. If these measures fail, more aggressive management with surgical
ligation of the arterial supply to the uterus and hysterectomy can be per-
formed. Angiographic treatments for early postpartum hemorrhage are
now also gaining acceptance and are discussed in some detail in this
text. Delayed postpartum hemorrhage most commonly occurs weeks after
delivery and most women are not under immediate medical supervision.
Initial management of delayed postpartum hemorrhage includes a clinical
evaluation and the usual clinical management described above. If these
measures fail, angiographic options are also available and will be
discussed.
Finally, intrapartum and postpartum hemorrhage is seen in some high-
risk populations with placentation abnormalities and during surgical man-
agement of ectopic pregnancies. Advanced imaging and serologic testing
have greatly enhanced early detection and management of these conditions.
Therefore, the incidence of life-threatening hemorrhage has been signifi-
cantly reduced in last few decades. Nonetheless, clinical and surgical man-
agement of placentation abnormalities can be difficult and angiographic
techniques now exist to assist the obstetric team in limiting blood loss. Like-
wise, during operative management of abdominal pregnancies and occasion-
ally cervical pregnancies, interventional and angiographic management can
offer additional options for decreasing the risk of massive bleeding. Some of
these options are discussed here.
602 BANOVAC et al

Fig. 2. Uterine and adnexal arteriovenous malformation. (A) Twenty-four–year–old woman

with recent miscarriage, dilatation, and curettage, complicated by severe bleeding and subse-
quent recurrent hemorrhage. Ultrasound examination showed hypervascular lesion in right cer-
vical area with a differential diagnosis of either cervical pregnancy or arteriovenous fistula.
Bilateral simultaneous anteroposterior arteriographic image reveals increased vascularity on
the right uterus with an early filling right ovarian vein (arrow). (B) Right uterine arteriogram,
early arterial phase in the left anterior oblique position. An enlarged cervicovaginal branch
(arrows) supplied a hypervascular nidus on subsequent images. The position correlated with
the cervical abnormality on duplex ultrasound (not shown). (C) Right uterine arteriogram in
the left anterior oblique position after embolization. The cervicovaginal branch is occluded
(arrow) and the arterial flow diminished. There is faint early opacification of veins in the right
adnexae for uncertain reasons. (D) Arteriogram of anterior division of right hypogastric artery
after embolization, revealing additional branches to the adnexae from other pelvic branches.
(E) Late phase of arteriogram showing markedly enlarged veins extending laterally from the
margin of the uterus into the adnexa. The uterine bleeding was controlled by the limited embo-
lization of the cervicovaginal branch. The larger asymptomatic portion of the vascular malfor-
mation was asymptomatic and not treated. (From Roth AR, Goodwin SC, Vedantham S, et al.
Management of gynecologic hemorrhage. In: Spies JB, Pelage JP, editors. Uterine artery embo-
lization. Philadelphia: Lippincott Williams and Wilkins; 2005. p. 155; with permission).
 ANGIOGRAPHIC AND INTERVENTIONAL OPTIONS 603

Angiographic techniques and strategies in obstetric hemorrhage

Early postpartum hemorrhage
The traditional approach to persistent and massive postpartum hemor-
rhage, if initial conservative clinical measures fail, has involved emergent
hysterectomy. Uterine artery ligation or internal iliac artery ligation, instead
of hysterectomy, is possible in women who wish to preserve their fertility.
While some investigators report very good results with bilateral internal iliac
artery ligation for controlling postpartum hemorrhage [14], others found
that internal iliac artery ligation for control of hemorrhage is often unsuc-
cessful, with success rates as low as 42% [15]. This may be attributable to
distal reconstitution of the internal iliac arteries in the setting of a markedly
hypervascular postpartum uterus [16]. Hysterectomy after a failed ligation
carries a higher morbidity than hysterectomy alone [15].
Angiographic techniques with embolization are available to contribute to
the overall management of early postpartum hemorrhage. The technique
was first described by Brown and colleagues [17] in 1979 and by Pais and
colleagues [16] in 1980. To this day, the embolization technique has
remained largely unchanged from these initial reports. The procedure is per-
formed by interventional radiologists in the angiographic suite. Fluoro-
scopic guidance is used to catheterize the anterior division of internal iliac
arteries with angiographic catheters and embolization is subsequently per-
formed. Subselective embolization of the uterine arteries [10] or vaginal
arteries [7] is performed whenever possible as each of these have been
separately reported as the most common source of bleeding. If the exact
source of bleeding cannot be identified, as sometimes occurs, empiric embo-
lization of the anterior division of the internal iliac arteries is performed
with pledgets of gelatin sponge or gelatin sponge slurry. A gelatin sponge
is the agent of choice because it causes a temporary arterial occlusion
with recanalization of blood flow within weeks. Microcoil embolization
alone for postpartum hemorrhage is not advocated because rich distal vas-
cular supply often reconstitutes the hypervascular gravid uterus [16,18]. Col-
lateral branches, such as the medial circumflex artery from the profunda
femoris and branches from the inferior epigastric artery, often reconstitute
the distal supply and bleeding often continues. For similar reasons, bilateral
embolization is often performed because bleeding can continue through
transpelvic vascular supply after unilateral embolization [19]. No prospec-
tive studies are available comparing unilateral and bilateral embolization,
but bilateral embolization is typically performed.

Delayed postpartum hemorrhage

From the technical standpoint, embolization for late or delayed postpar-
tum hemorrhage is the same as for early postpartum hemorrhage. Delayed
postpartum hemorrhage is often attributable to retained placental fragments
604 BANOVAC et al

with or without endometritis and occasionally genital tract lacerations. If

the bleeding persists after primary repair of the lacerations or curettage,
embolization is an alternative to surgical ligation or hysterectomy [20,21].
Again, reported embolization techniques vary (Fig. 3). Most investigators
used gelatin sponge pledgets for embolization. Pelage and colleagues [21]
selectively catheterized and embolized the uterine arteries in most cases,

Fig. 3. Thirty-three–year–old woman 3 weeks postpartum with intermittent vaginal bleeding.

(A) Selective uterine artery arteriograms using simultaneous injections through catheters placed
via left and right common femoral approach fail to demonstrate any active extravasation. (B) A
more thorough search for the source of bleeding resulted in a selective arteriogram of the right
ovarian artery, which demonstrated a pseudoaneurysm. (C) Gelatin slurry embolization of the
right ovarian artery through a coaxially placed microcatheter successfully amputated the flow to
the distal branches that were supplying the pseudoaneurysm. (D) Shortly after the procedure,
the hemorrhage continued and patient was brought back to the interventional radiology suite
for microcoil embolization. Coils were deposited into the right ovarian artery, effectively arrest-
ing antegrade flow. (From Baum S, Pentecost M. Abrams’ angiography: interventional radiol-
ogy. Philadelphia: Lippincott Williams and Wilkins;. 2005. p. 823; with permission).
 ANGIOGRAPHIC AND INTERVENTIONAL OPTIONS 605

while Feinberg and colleagues [22] reported nonselective gelatin sponge and
coil embolization of internal iliac arteries.

Embolization options for ectopic pregnancies

Embolization plays a very limited role in management of ectopic preg-
nancies. The diagnosis of ectopic pregnancy based on ultrasound findings,
elevated levels of the beta subunit of human chorionic gonadotrphin
(beta-HCG), and clinical presentation most often results in prompt medical
or operative treatment. Clinically significant hemorrhage is mostly avoided.
However, successful embolization has been reported in the context of overall
management for abdominal [4,5,23–25] and cervical [26,27] ectopic pregnan-
cies, mostly to reduce the operative blood loss. Using standard angiographic
techniques, vascular supply to the pregnancy is determined and selective gel-
atin sponge embolization is performed before operative management.
Badawy and colleagues [12] reviewed 11 reports totaling 21 cases of arterial
embolization for abdominal and cervical pregnancies and reported a 100%
success rate in controlling the hemorrhage. In the setting of hemorrhage
after the operative removal of the cervical ectopic pregnancy, successful selec-
tive embolization of the placental fragment has been reported [28]. Although
the role of angiography and embolization in the setting of abdominal and
cervical pregnancies is limited, consultation for perioperative embolization
can be considered in certain cases.

Embolization for placentation abnormalities

Placentation abnormalities can present a formidable clinical challenge.
Among them, placenta percreta is the most problematic because uterine rup-
ture and ensuing hemorrhage can occur. However placenta accreta and
increta also carry an increased risk of bleeding (Fig. 4). Miller and colleagues
[29] quantified the amount of blood loss during cesarean hysterectomy asso-
ciated with placenta accreta in a group of 62 patients. Estimated blood loss
exceeded 2000 mL in 41 patients, 5000 mL in 9 patients, 10,000 mL in 4
patients, and 20,000 mL in 2 patients.
Interventional radiologists have two principal management algorithms in
this setting. First, embolization either before or immediately after the cesar-
ean delivery can be performed. Arterial access can be obtained via the axil-
lary artery [5], thus limiting the exposure of the fetus to ionizing radiation.
Embolization can be performed quickly if significant hemorrhage occurs.
Alternatively, femoral artery catheterization can be obtained emergently
after delivery.
Second, temporary occlusion of both internal iliac arteries with angio-
plasty balloons or compliant occlusion balloons can be done after initial
catheterization via either femoral or axillary approach [30,31]. This portion
of the procedure is done in the angiographic suite before delivery. During
operative delivery, the balloons are left deflated. Then, while the patient is
606 BANOVAC et al

Fig. 4. Angiographic appearance of placenta accreta. (A) Early arterial phase. (B) Late arterial
phase. A typical pseudotumoral multifocal blush is seen.

in the operating room, balloons are inflated to occlude the blood flow. This
technique allows additional time to try to control the hemorrhage surgically.
Alternatively, with the catheters still in place, the patient can be transferred
to the interventional radiology suite for embolization. Either selective gela-
tin sponge embolization of the uterine arteries can be performed by inserting
a microcatheter coaxially [32], or nonselective gelatin sponge embolization
can be performed through the end hole of the balloon catheter [30,31].
Although most investigators advocate balloon placement before delivery,
the technique is somewhat controversial because a small prospective cohort
study [33] and a retrospective review [34] failed to demonstrate a benefit.
Therefore, the value of this technique is yet to be proven. Nonetheless, em-
bolization in the setting of placentation abnormalities has a role and has
been used in minimizing the operative blood loss during hysterectomy. As
described by Greenberg and colleagues [35], embolization can also be used
as an adjunct to hysteroscopic morcellation of placenta accreta. In a separate
report, embolization in the setting of placenta accreta has even been shown
to control the hemorrhage and preserve the uterus and fertility [36].

Fertility outcomes after embolization

Embolization for postpartum hemorrhage, placentation abnormalities,
or arteriovenous malformations (or as an adjunct in treatment of ectopic
pregnancies) invariably causes some degree of ischemia to the uterus. How-
ever, clinically significant ischemic injury to the uterus is extremely rare. Fer-
tility after embolization has not been thoroughly studied, although
a number of reports in both gynecologic and radiological literature suggests
favorable outcomes. Normal resumption of menses [8,11,24,37] and normal
pregnancies have been reported by a number of investigators [8,20,24,
38–41]. In a small group follow-up study (1–6 years), Stancato-Pasik and
 ANGIOGRAPHIC AND INTERVENTIONAL OPTIONS 607

colleagues [24] found that 92% of the patients resumed normal menses
within 2 to 5 months after embolization, without complications related to
embolotherapy. All three patients who wished to conceive gave birth to
full-term, healthy newborns. Similar results were reported by Ornan and
colleagues [42] who found that, after embolization for postpartum hemor-
rhage, all patients who wished to become pregnant were successful. Shim
and colleagues [41] followed 37 patients after embolization for postpartum
hemorrhage and found that 36 resumed normal menses and 9 became
pregnant.
The effects of prior embolization on potential complications during the
ensuing pregnancies have not been studied thoroughly; however, some
groups reported an increased rate of postpartum hemorrhage in those pa-
tients who had a prior embolization. Additionally, the effects on fetal devel-
opment are only sporadically reported. Although most investigators report
normal pregnancies after embolization, in utero death and preeclampsia
have been reported [43], without speculation on the attributable cause.

Pelvic embolization in gynecologic hemorrhage

Perhaps because hemmorhage in gynecologic conditions is so rare, the
use of embolotherapy in such cases has been studied less than the use of em-
bolotherapy for obstetric-related hemmorhage. However, the same general
techniques and principles apply, and these appear to be effective in most
cases. Most reports have been in the setting of gynecologic malignancy,
but case reports suggest similar results would be achieved for arteriovenous
malformation and other benign pathology, such as pelvic hemorrhage due
to trauma [44]. This section reviews applications of embolotherapy to con-
trol hemorrhage from gynecologic conditions, as opposed to obstetric-
related hemorrhage.

Causes of gynecologic hemorrhage

Malignancy
Among malignant causes of vaginal hemorrhage, carcinoma of the cervix,
endometrium, and choriocarcinoma are the most common causative tumors
[45]. Vaginal bleeding related to pelvic neoplasms is typically slow and inter-
mittent but persistent and poorly responsive to surgical intervention or ra-
diation therapy. The bleeding may be the result of invasion of small
vessels by tumor or the result of ulceration or necrosis of the tumors.
More massive bleeding may occur as tumors invade large vessels (Fig. 5).

Uterine vascular abnormalities

Uterine vascular abnormalities may be congenital or acquired. Congenital
arteriovenous malformations are quite rare and may have a complex set of
feeding arteries and draining veins. They may cause massive hemorrhage or
intermittent bleeding. Acquired lesions are more common and often are the
608 BANOVAC et al

Fig. 5. Massive hemorrhage from neoplastic erosion or radiation injury to left hypogastric
artery. (A) Initial bilateral hypogastric arteriogram revealing postoperative changes in the left
hypogastric artery, intact vessels on the right, without a clear site of bleeding. It was decided
to proceed with embolization on the right, using polyvinyl alcohol particles. After the emboli-
zation on the right, the patient suddenly became tachycardic. The blood pressure was main-
tained and the cause of the tachycardia was not immediately clear. It was decided to consider
termination of the procedure after a repeat arteriogram. (B) Repeat arteriogram reveals massive
bleeding from the left hypogastric artery stump. The anterior division of the right hypogastric
artery was occluded. (C) After embolization of the left bleeding site with gelatin sponge and
coils, with control of the bleeding. (From Roth AR, Goodwin SC, Vedantham S, et al. Manage-
ment of gynecologic hemorrhage. In: Spies JB, Pelage JP, editors. Uterine artery embolization.
Philadelphia: Lippincott Williams and Wilkins; 2005. p. 152; with permission).

result of uterine surgery, curettage, a retained placenta, or obstetrical trauma.

Delayed postpartum hemorrhage that does not resolve spontaneously may be
due to arterial injury. In one small series of 14 patients, those 3 patients treated
for delayed postpartum hemorrhage had uterine vascular abnormalities as the
underlying cause [21]. Similar findings have been noted in patients with
hemorrhage following uterine curettage or surgery [46]. These vascular
 ANGIOGRAPHIC AND INTERVENTIONAL OPTIONS 609

abnormalities include pseudoaneurysms, arteriovenous fistulas, and direct

vessel rupture.
The diagnosis of vascular malformations of the uterus is usually made
with color and duplex Doppler ultrasound revealing a blood-filled cystic
structure with swirling arterial flow (pseudoaneurysm); rapid arterial-to-
venous shunting, such as that seen in arteriovenous fistulas; or an intense
vascular tangle of vessels or arteriovenous malformations [47]. On duplex
Doppler ultrasound, arteriovenous fistulas and arteriovenous malforma-
tions demonstrate low-resistance, high-velocity arterial flow.

Embolization technique for gynecologic hemorrhage

The exact source of the bleeding is often not known before the emboliza-
tion procedure and survey arteriography is often needed before catheterizing
individual vessels. A pelvic arteriogram with injection of contrast in the
lower abdominal aorta is usually the first study performed. This may be sup-
plemented by injection into ovarian and inferior mesenteric vessels if no
bleeding site is identified. Using these initial studies as a guide, further selec-
tive arteriograms are performed to isolate the site of bleeding. If the patient
has not had a hysterectomy, then the uterine vessels are likely the source. In
the postoperative patient, other pelvic visceral vessels are often involved.
Although a variety of technical approaches have been used, there is a con-
sensus on the general approach to pelvic embolization [44,48]. An angio-
graphic catheter is advanced selectively into the branch in question, using
fluoroscopic guidance with a digital roadmap or image as a guide. This is
often done using a coaxial technique, in which a microcatheter is advanced
through an outer 5F catheter. The use of a microcatheter allows smaller
branches to be entered, which provides for a more targeted embolization.
Embolic material, usually polyvinyl alcohol particles, microspheres, or coils,
are injected in the bleeding vessel. The material is injected into the target
vessel until stasis of flow is confirmed angiographically. These materials
are best applied when bleeding is identified from small arterial branches
of less than a few millimeters in diameter.
Gelatin sponge is often used in larger vessels, often in conjunction with
coils. In general, particulate emboli penetrate farther into vessels than gela-
tin sponge and will occlude at or within the tumor. Gelatin sponge and coils
provide a more proximal occlusion of larger feeding vessels than that pro-
vided by the particulate materials. Also, gelatin sponge and coils allow
a more rapid occlusion of large vessels when bleeding is severe. Coils are of-
ten also used as a trap for the gelatin sponge and a ‘‘plug’’ in the vessels can
be created. This is particularly useful when the feeding vessels to the bleed-
ing site arise from a larger branch and the feeding vessels themselves cannot
be catheterized. Rather than extensively embolizing the large vessel and pos-
sibly endangering the perfusion of other important structures, a plug can be
created that crosses the origins of the vessels and seals them, while filling the
610 BANOVAC et al

lumen of the large vessel distal to the plug. This preserves the potential for
collateral vessel flow to the normal tissue below the site of occlusion.
The technique for vascular malformation embolization varies with the
size and extent of the abnormalities. Congenital malformations often are
complex and may be treated in some cases with the embolic materials men-
tioned above. However, these often must be supplemented with permanent
tissue adhesive or other liquid embolics [49,50]. For simple arteriovenous fis-
tulas and pseudoaneurysms, various combinations of embolic materials
have been employed, the choice depending the anatomic considerations in
each case [51].
The embolization process is monitored using video fluoroscopy. Ipsilat-
eral internal iliac angiography is repeated to exclude the possibility of addi-
tional feeding arteries, which occasionally become visible only after the
primary feeding artery is occluded. Embolization of the contralateral hypo-
gastric artery or its branches may be performed to decrease the likelihood of
cross-filling. This type of ‘‘prophylactic’’ embolization is usually only done
in patients with pelvic malignancy and often is more limited in extent to
minimize the chance of ischemic injury to the pelvic organs. If the patient’s
clinical condition suggests that the bleeding has not stopped, then additional
angiographic exploration is necessary to identify other potential sources of
blood.
The technique of embolization can be more complicated in gynecologic
bleeding than in the postpartum setting, particularly when the patient has
already had surgery, radiotherapy, or both. Normal anatomic relationships
are distorted and there may be atypical sources of blood supply to the bleed-
ing site. Thus, to be effective, the embolization may need to be more exten-
sive than is normally required for a typical postpartum embolization.

Outcome from embolotherapy for gynecologic hemorrhage

The limited published data on embolotherapy of pelvic malignancy sug-
gest that embolization can be both safe and effective in controlling hemor-
rhage secondary to pelvic malignancy. Early case reports from the 1970s
demonstrated the feasibility of the technique [52,53]. Several case series
were published in the 1980s. In 1981, Lang [54] reported the results of embo-
lization of 23 patients who had pelvic neoplasms. There was 100% immediate
cessation of bleeding. In a larger series, Pisco and colleagues [55] reported the
results of a series of 108 patients with hemorrhage from pelvic malignancy,
including 55 with gynecologic malignancy. They achieved complete hemor-
rhage control in 74 patients, partial control in 23, and no control in 11.
In 1993, Yamshita and colleagues [56] reported on 17 patients with cer-
vical cancer who developed malignancy-related massive hemorrhage. Active
extravasation from the uterine artery was found in only 2 patients, but neo-
vascularity was demonstrated in 12. Immediate control of bleeding was
achieved in 100%. However 7 patients (41.1%) had recurrent bleeding after
 ANGIOGRAPHIC AND INTERVENTIONAL OPTIONS 611

2 weeks, and 3 of those patients required repeat embolization. All patients

continued with planned therapy after control of hemorrhage. Other smaller
series have noted similar results [57]. There is also evidence that these pro-
cedures prolong survival in patients with advanced malignancy, with median
survival extended 4 to 6 months [58,59]. Long-term follow-up in patients
with hemorrhage from benign causes suggests long-term symptom control,
with one 7.5-year follow-up study [60].
Therapeutic outcomes from treatment of vascular malformations are also
quite good. Among 17 patients treated, Maleux and colleagues [51] noted
recurrence in only 1, with a mean follow-up of 18.8 months. Similar results
were noted by Ghai and colleagues [61]. In the long-term, this therapy ap-
pears durable. Jacobowitz and colleagues [62] treated 35 patients, each hav-
ing an average of 2.4 procedures. More than one procedure was often
needed to completely treat these complex lesions. With an average of 84
months follow-up, 83% of patients remained asymptomatic or significantly
improved.

Complications of embolization for peripartum and gynecologic

hemorrhage
Complications of angiography and embolization in the setting of peripar-
tum bleeding after angiographic intervention for gynecologic emergencies
are rare. From a technical standpoint, catheterization is straightforward
for a trained interventional radiologist because most women in this group
are young and free of atherosclerotic vascular disease. In the early evolution
of embolization in this setting, a concern about end-organ ischemia was
raised. Therefore in 1980, Pais and colleagues [16] histologically evaluated
a hysterectomy specimen after embolization and failed to find any ischemic
changes. Over time, it was found that generalized uterine ischemia is indeed
a very rare complication. However, uterine ischemic necrosis was reported
when very small (150–250 mm) polyvinyl alcohol particles were used along
with gelatin sponge pledgets [63]. Small particles can cause very distal embo-
lization and have potential to cause ischemic injury. When using gelatin
sponge alone, only one report of severe ischemic injury and uterine necrosis
has been reported [64].
While trying to elicit the effects and outcomes of embolization for post-
partum hemorrhage, Badawy and colleagues [12] reviewed 22 publications
from 1979 to 1999. In their review of 138 cases of postpartum hemorrhage
treated with embolization, they report a high technical success rate of 95%
for controlling the bleeding in the short term. Seven cases required an even-
tual hysterectomy. Other reported complications related to embolization for
either postpartum hemorrhage or gynecologic hemorrhage included tran-
sient fevers [11,65], transient buttock ischemia and lower extremity para-
sthesia [8], external iliac artery perforation [4], groin hematoma, pelvic
abscess formation [66], transient foot ischemia, bladder gangrene [12], and
612 BANOVAC et al

Table 1
Embolization for treatment of postpartum hemorrhage
Number
Investigators of patients Embolic material Complications
Abbas et al [43] 1 Gelatin sponge, Readmission, fever,
coil, PVA vaginal bleeding,
abdominal hematoma,
septic shock
Bakri and Linjawi [57] 3 Gelatin sponge, coil Femoral hematoma
Brown et al [17] 1 Gelatin sponge none
Chin et al [65] 2 Gelatin sponge, coil fever
Dubois et al [30] 2 Gelatin sponge None
Feinberg et al [22] 1 Gelatin sponge, coil none
Gilbert et al [6] 6 Gelatin sponge none
Greenwood et al [4] 6 Gelatin sponge, coil Transient buttock
ischemia, external
iliac perforation
Hansch et al [32] 5 Gelatin sponge, none
coil, PVA
Heffner et al [68] 3 Gelatin sponge None
Hsu and Wan [69] 2 Gelatin sponge None
Joseph et al [70] 2 Gelatin sponge, coil None
Merland et al [71] 16 Gelatin sponge, PVA None
Mitty et al [5] 7 Gelatin sponge, coil None
Pais et al [16] 1 Gelatin sponge, coil Uterine perforation,
fever
Pelage et al [10] 27 Gelatin sponge, PVA Repeat embolization,
hysterectomy
Pelage et al [21] 14 Gelatin sponge, none
n-butyl-
2-cyanoacrylate
Rosenthal and 2 Coil Failed embolization,
Colapinto [19] wound infection
Shweni et al [72] 4 Gelatin sponge None
Soncini et al [73] 14 Gelatin sponge, coil Hysterectomy, fever
Stancato-Pasik 12 Gelatin sponge None
et al [24]
Vegas et al [74] 27 Coil, PVA Hysterectomy, repeat
embolization,
hysterectomy,
vaginal fistula
Yamashita et al [11] 6 Gelatin sponge Fever
Yamashita et al [56] 15 Gelatin sponge, coil None
Yong and Cheung [75] 29 Not specified Cardiac arrest,
hysterectomy,
claudication, fever
Abbreviation: PVA, polyvinyl alcohol particles.
 ANGIOGRAPHIC AND INTERVENTIONAL OPTIONS 613

vesicovaginal fistula formation [67]. Primary embolization has also been

reported to technically fail with need for a hysterectomy [10]. A summary
of selected reports describing the technique and complications of emboliza-
tion in the setting of postpartum hemorrhage is presented in Table 1 [68–75].

Summary
Arterial embolization can play an important role in overall management of
obstetric and gynecologic vascular emergencies. A substantial body of litera-
ture from obstetric–gynecologic and radiological sources describes the embo-
lization techniques as safe and effective in achieving control of hemorrhagic
complications for postpartum hemorrhage and gynecologic emergencies.
Embolization avoids operative morbidity in patients who are usually
poor surgical candidates due to anemia and coagulopathies. Embolization
does not preclude surgical ligation or hysterectomy should embolization
fail and surgical approaches become necessary [76]. Additionally, new
angiographic techniques can serve as an important adjunct in the manage-
ment of ectopic pregnancies and placentation abnormalities. Obstetricians,
gynecologists, and interventional radiologists alike should be familiar with
these options to provide the most comprehensive care to patients.

References
[1] Kaunitz AM, Hughes JM, Grimes DA, et al. Causes of maternal mortality in the United
States. Obstet Gynecol 1985;65(5):605–12.
[2] Combs CA, Murphy EL, Laros RK Jr. Factors associated with postpartum hemorrhage
with vaginal birth. Obstet Gynecol 1991;77(1):69–76.
[3] Combs CA, Murphy EL, Laros RK Jr. Factors associated with hemorrhage in cesarean
deliveries. Obstet Gynecol 1991;77(1):77–82.
[4] Greenwood LH, Glickman MG, Schwartz PE, et al. Obstetric and nonmalignant gyneco-
logic bleeding: treatment with angiographic embolization. Radiology 1987;164(1):155–9.
[5] Mitty HA, Sterling KM, Alvarez M, et al. Obstetric hemorrhage: prophylactic and emer-
gency arterial catheterization and embolotherapy. Radiology 1993;188(1):183–7.
[6] Gilbert WM, Moore TR, Resnik R, et al. Angiographic embolization in the management of
hemorrhagic complications of pregnancy. Am J Obstet Gynecol 1992;166(2):493–7.
[7] Deux JF, Bazot M, Le Blanche AF, et al. Is selective embolization of uterine arteries a safe
alternative to hysterectomy in patients with postpartum hemorrhage? AJR Am J Roentgenol
2001;177(1):145–9.
[8] Chung JW, Jeong HJ, Joh JH, et al. Percutaneous transcatheter angiographic embolization
in the management of obstetric hemorrhage. J Reprod Med 2003;48(4):268–76.
[9] Dildy GA 3rd. Postpartum hemorrhage: new management options. Clin Obstet Gynecol
2002;45(2):330–44.
[10] Pelage JP, Le Dref O, Mateo J, et al. Life-threatening primary postpartum hemorrhage:
treatment with emergency selective arterial embolization. Radiology 1998;208(2):359–62.
[11] Yamashita Y, Takahashi M, Ito M, et al. Transcatheter arterial embolization in the manage-
ment of postpartum hemorrhage due to genital tract injury. Obstet Gynecol 1991;77(1):160–3.
[12] Badawy SZ, Etman A, Singh M, et al. Uterine artery embolization: the role in obstetrics and
gynecology. Clin Imaging 2001;25(4):288–95.
614 BANOVAC et al

[13] Kelly SM, Belli AM, Campbell S. Arteriovenous malformation of the uterus associated with
secondary postpartum hemorrhage. Ultrasound Obstet Gynecol 2003;21(6):602–5.
[14] Joshi VM, Otiv SR, Majumder R, et al. Internal iliac artery ligation for arresting postpartum
haemorrhage. BJOG 2007;114(3):356–61.
[15] Clark SL, Phelan JP, Yeh SY, et al. Hypogastric artery ligation for obstetric hemorrhage.
Obstet Gynecol 1985;66(3):353–6.
[16] Pais SO, Glickman M, Schwartz P, et al. Embolization of pelvic arteries for control of post-
partum hemorrhage. Obstet Gynecol 1980;55(6):754–8.
[17] Brown BJ, Heaston DK, Poulson AM, et al. Uncontrollable postpartum bleeding: a new ap-
proach to hemostasis through angiographic arterial embolization. Obstet Gynecol 1979;
54(3):361–5.
[18] Minck RN, Palestrant A, Cherny WB. Successful management of postpartum vaginal hem-
orrhage by angiographic embolization. Ariz Med 1984;41(8):537–8.
[19] Rosenthal DM, Colapinto R. Angiographic arterial embolization in the management of
postoperative vaginal hemorrhage. Am J Obstet Gynecol 1985;151(2):227–31.
[20] Pelage JP, Le Dref O, Jacob D, et al. Selective arterial embolization of the uterine arteries in
the management of intractable post-partum hemorrhage. Acta Obstet Gynecol Scand 1999;
78(8):698–703.
[21] Pelage JP, Soyer P, Repiquet D, et al. Secondary postpartum hemorrhage: treatment with
selective arterial embolization. Radiology 1999;212(2):385–9.
[22] Feinberg BB, Resnik E, Hurt WG, et al. Angiographic embolization in the management of
late postpartum hemorrhage. A case report. J Reprod Med 1987;32(12):929–31.
[23] Kerr A, Trambert J, Mikhail M, et al. Preoperative transcatheter embolization of abdominal
pregnancy: report of three cases. J Vasc Interv Radiol 1993;4(6):733–5.
[24] Stancato-Pasik A, Mitty HA, Richard HM 3rd, et al. Obstetric embolotherapy: effect on
menses and pregnancy. Radiology 1997;204(3):791–3.
[25] Cardosi RJ, Nackley AC, Londono J, et al. Embolization for advanced abdominal preg-
nancy with a retained placenta. A case report. J Reprod Med 2002;47(10):861–3.
[26] Lobel SM, Meyerovitz MF, Benson CL, et al. Preoperative angiographic uterine artery em-
bolization in the management of cervical pregnancy. Obstet Gynecol 1990;76(5 Pt 2):938–41.
[27] Suzumori N, Katano K, Sato T, et al. Conservative treatment by angiographic artery embo-
lization of an 11-week cervical pregnancy after a period of heavy bleeding. Fertil Steril 2003;
80(3):617–9.
[28] Martin JN Jr, Ridgeway LE 3rd, Connors JJ, et al. Angiographic arterial embolization and
computed tomography-directed drainage for the management of hemorrhage and infection
with abdominal pregnancy. Obstet Gynecol 1990;76(5 Pt 2):941–5.
[29] Miller DA, Chollet JA, Goodwin TM. Clinical risk factors for placenta previa-placenta
accreta. Am J Obstet Gynecol 1997;177(1):210–4.
[30] Dubois J, Garel L, Grignon A, et al. Placenta percreta: balloon occlusion and embolization
of the internal iliac arteries to reduce intraoperative blood losses. Am J Obstet Gynecol 1997;
176(3):723–6.
[31] Weeks SM, Stroud TH, Sandhu J, et al. Temporary balloon occlusion of the internal iliac
arteries for control of hemorrhage during cesarean hysterectomy in a patient with placenta
previa and placenta increta. J Vasc Interv Radiol 2000;11(5):622–4.
[32] Hansch E, Chitkara U, McAlpine J, et al. Pelvic arterial embolization for control of obstetric
hemorrhage: a five-year experience. Am J Obstet Gynecol 1999;180(6 Pt 1):1454–60.
[33] Levine AB, Kuhlman K, Bonn J. Placenta accreta: comparison of cases managed with and
without pelvic artery balloon catheters. J Matern Fetal Med 1999;8(4):173–6.
[34] Bodner LJ, Nosher JL, Gribbin C, et al. Balloon-assisted occlusion of the internal iliac arteries in
patients with placenta accreta/percreta. Cardiovasc Intervent Radiol 2006;29(3):354–61.
[35] Greenberg JA, Miner JD, O’Horo SK. Uterine artery embolization and hysteroscopic resection
to treat retained placenta accreta: a case report. J Minim Invasive Gynecol 2006;13(4):342–4.
 ANGIOGRAPHIC AND INTERVENTIONAL OPTIONS 615

[36] Alanis M, Hurst BS, Marshburn PB, et al. Conservative management of placenta increta
with selective arterial embolization preserves future fertility and results in a favorable out-
come in subsequent pregnancies. Fertil Steril 2006;86(5):e3–7.
[37] Descargues G, Mauger Tinlot F, Douvrin F, et al. Menses, fertility and pregnancy after ar-
terial embolization for the control of postpartum haemorrhage. Hum Reprod 2004;19(2):
339–43.
[38] Wang H, Garmel S. Successful term pregnancy after bilateral uterine artery embolization for
postpartum hemorrhage. Obstet Gynecol 2003;102(3):603–4.
[39] Casele HL, Laifer SA. Successful pregnancy after bilateral hypogastric artery ligation.
A case report. J Reprod Med 1997;42(5):306–8.
[40] Delotte J, Chevallier P, Benoit B, et al. Pregnancy after embolization therapy for uterine ar-
teriovenous malformation. Fertil Steril 2006;85(1):228.
[41] Shim JY, Yoon HK, Won HS, et al. Angiographic embolization for obstetrical hemorrhage:
effectiveness and follow-up outcome of fertility. Acta Obstet Gynecol Scand 2006;85(7):
815–20.
[42] Ornan D, White R, Pollak J, et al. Pelvic embolization for intractable postpartum hemor-
rhage: long-term follow-up and implications for fertility. Obstet Gynecol 2003;102(5 Pt 1):
904–10.
[43] Abbas FM, Currie JL, Mitchell S, et al. Selective vascular embolization in benign gyneco-
logic conditions. J Reprod Med 1994;39(7):492–6.
[44] Vedantham S, Goodwin SC, McLucas B, et al. Uterine artery embolization: an underused
method of controlling pelvic hemorrhage. Am J Obstet Gynecol 1997;176(4):938–48.
[45] Lewis E, Zornoza J, Jing BS, et al. Radiologic contributions to the diagnosis and manage-
ment of gynecologic neoplasms. Semin Roentgenol 1982;17(4):251–68.
[46] Haseltine FP, Glickman MG, Marchesi S, et al. Uterine embolization in a patient with post-
abortal hemorrhage. Obstet Gynecol 1984;63(3 Suppl):78S–80S.
[47] Flynn MK, Levine D. The noninvasive diagnosis and management of a uterine arteriove-
nous malformation. Obstet Gynecol 1996;88(4 Pt 2):650–2.
[48] Roth A, Goodwin SC, Vedantham S, et al. Embolization for management of gynecologic
hemorrhage. In: Spies J, Pelage JP, editors. Uterine artery embolization. Philadelphia: Lip-
pencott, Williams and Wilkins; 2005.
[49] Yakes WF, Luethke JM, Parker SH, et al. Ethanol embolization of vascular malformations.
Radiographics 1990;10(5):787–96.
[50] Coldwell DM, Stokes KR, Yakes WF. Embolotherapy: agents, clinical applications, and
techniques. Radiographics 1994;14(3):623–43 [quiz: 645–6].
[51] Maleux G, Timmerman D, Heye S, et al. Acquired uterine vascular malformations: radiolog-
ical and clinical outcome after transcatheter embolotherapy. Eur Radiol 2006;16(2):
299–306.
[52] Athanasoulis CA, Waltman AC, Barnes AB, et al. Angiographic control of pelvic bleeding
from treated carcinoma of the cervix. Gynecol Oncol 1976;4(2):144–50.
[53] Miller FJ Jr, Mortel R, Mann WJ, et al. Selective arterial embolization for control of hem-
orrhage in pelvic malignancy: femoral and brachial catheter approaches. AJR Am J Roent-
genol 1976;126(5):1028–32.
[54] Lang EK. Transcatheter embolization of pelvic vessels for control of intractable hemor-
rhage. Radiology 1981;140(2):331–9.
[55] Pisco JM, Martins JM, Correia MG. Internal iliac artery: embolization to control hemor-
rhage from pelvic neoplasms. Radiology 1989;172(2):337–9.
[56] Yamashita Y, Harada M, Yamamoto H, et al. Transcatheter arterial embolization of obstet-
ric and gynaecological bleeding: efficacy and clinical outcome. Br J Radiol 1994;67(798):
530–4.
[57] Bakri YN, Linjawi T. Angiographic embolization for control of pelvic genital tract hemor-
rhage. Report of 14 cases. Acta Obstet Gynecol Scand 1992;71(1):17–21.
616 BANOVAC et al

[58] Hendrickx P, Orth G, Grunert JH. Long-term survival after embolization of potentially le-
thal bleeding malignant pelvic tumours. Br J Radiol 1995;68(816):1336–43.
[59] Jenkins CN, McIvor J. Survival after embolization of the internal iliac arteries in ten patients
with severe haematuria due to recurrent pelvic carcinoma. Clin Radiol 1996;51(12):865–8.
[60] Hendrickx P, Orth G, Grunert J. Embolisation of bleeding pelvic lesions from benign
origindlong-term results. J Belge Radiol 1995;78(6):339–41.
[61] Ghai S, Rajan DK, Asch MR, et al. Efficacy of embolization in traumatic uterine vascular
malformations. J Vasc Interv Radiol 2003;14(11):1401–8.
[62] Jacobowitz GR, Rosen RJ, Rockman LB, et al. Transcatheter embolization of complex pel-
vic vascular malformations: results and long-term follow-up. J Vasc Surg 2001;33(1):51–5.
[63] Cottier JP, Fignon A, Tranquart F, et al. Uterine necrosis after arterial embolization for
postpartum hemorrhage. Obstet Gynecol 2002;100(5 Pt 2):1074–7.
[64] Chitrit Y, Yafy S, Pelage JP, et al. Amenorrhea due to partial uterine necrosis after uterine
artery embolization for control of refractory postpartum hemorrhage. Eur J Obstet Gynecol
Reprod Biol 2006;127(1):140–2.
[65] Chin HG, Scott DR, Resnik R, et al. Angiographic embolization of intractable puerperal he-
matomas. Am J Obstet Gynecol 1989;160(2):434–8.
[66] Choo YC, Cho KJ. Pelvic abscess complicating embolic therapy for control of bleeding cer-
vical carcinoma and simultaneous radiation therapy. Obstet Gynecol 1980;55(3 Suppl):
76S–8S.
[67] Behnam K, Jarmolowski CR. Vesicovaginal fistula following hypogastric embolization for
control of intractable pelvic hemorrhage. J Reprod Med 1982;27(5):304–6.
[68] Heffner LJ, Mennuti MT, Rudoff JC, et al. Primary management of postpartum vulvovagi-
nal hematomas by angiographic embolization. Am J Perinatol 1985;2(3):204–7.
[69] Hsu YR, Wan YL. Successful management of intractable puerperal hematoma and severe
postpartum hemorrhage with DIC through transcatheter arterial embolizationdtwo cases.
Acta Obstet Gynecol Scand 1998;77(1):129–31.
[70] Joseph JF, Mernoff D, Donovan J, et al. Percutaneous angiographic arterial embolization
for gynecologic and obstetric pelvic hemorrhage. A report of three cases. J Reprod Med
1994;39(11):915–20.
[71] Merland JJ, Houdart E, Herbreteau D, et al. Place of emergency arterial embolisation in ob-
stetric haemorrhage about 16 personal cases. Eur J Obstet Gynecol Reprod Biol 1996;65(1):
141–3.
[72] Shweni PM, Bishop BB, Hansen JN, et al. Severe secondary postpartum haemorrhage after
caesarean section. S Afr Med J 1987;72(9):617–9.
[73] Soncini E, Pelicelli A, Larini P, et al. Uterine artery embolization in the treatment and pre-
vention of postpartum hemorrhage. Int J Gynaecol Obstet 2007;96(3):181–5.
[74] Vegas G, Illescas T, Muñoz M, et al. Selective pelvic arterial embolization in the manage-
ment of obstetric hemorrhage. Eur J Obstet Gynecol Reprod Biol 2006;127(1):68–72.
[75] Yong SP, Cheung KB. Management of primary postpartum haemorrhage with arterial
embolisation in Hong Kong public hospitals. Hong Kong Med J 2006;12(6):437–41.
[76] Banovac F. Obstetrical hemorrhage. In: Abrams HL, Baum S, Pentecost MJ, editors.
Abram’s angiography: interventional radiology. Philadelphia: Lippincott Williams & Wil-
kins; 2006.
 Obstet Gynecol Clin N Am
34 (2007) 617–625

Liability in High-Risk Obstetrics

James M. Shwayder, MD, JD
Obstetrics, Gynecology and Women’s Health, University of Louisville School of Medicine,
3rd Floor, ACB, 550 S. Jackson Street, Louisville, KY 40202, USA

Liability issues have changed the obstetrical landscape. The 2006 American
College of Obstetricians and Gynecologists (ACOG) survey on professional
liability revealed significant practice changes as a result of insurance avail-
ability or affordability. According to the survey, 25.6% of surveyed physicians
decreased their number of high-risk obstetrical patients, while 7.2% quit
practicing obstetrics altogether. Furthermore, 28.5% of those who continue
to deliver patients reported increasing the number of cesarean sections,
with 26.4% not performing vaginal births after cesarean sections (VBACs).
Reducing liability risk requires an understanding of the prime reasons
physicians are sued and are limiting exposure in the main areas affecting
obstetrics [1].

Reasons physicians are sued

A recent plaintiff’s attorney’s article highlighted the major reasons physi-
cians get sued. The most common reason for suit is, predictably, that a med-
ical error injured a patient. Although frivolous lawsuits do occur, credible
plaintiff firms avoid such suits, as the cost of prosecuting a medical malprac-
tice suit ranges from $40,000 to over $200,000 [2]. In general, plaintiff firms
turn down over 90% of cases, either because expert review supports the
physician actions or because damages are insufficient to cover the costs of
litigation. It is for this latter reason that obstetrics, with its high exposure,
is a focus of many plaintiff firms. The average damages in a successful
suit with a neurologically impaired infant in $1,150,687 [1].
New paradigms in delivering obstetrical care, such as laborists and large
group practices, may actually expose physicians to greater liability. Patients
may be attended and delivered by physicians they have never seen. If prob-
lems occur during labor and delivery, the lack of a trusting relationship

E-mail address: james.shwayder@louisville.edu

0889-8545/07/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ogc.2007.08.003 obgyn.theclinics.com
618 SHWAYDER

raises questions of competency, and thus the specter of malpractice. This

problem is compounded if there is a lack of communication between the
primary obstetrician and the covering physician. Thus, the covering obste-
trician may be unaware of risk factors that might lower the threshold for
cesarean section.
A common source of malpractice suit is injuries associated with delayed
cesarean sections and difficult operative vaginal deliveries. Pritchard [2]
termed this ‘‘wishful obstetrics.’’ This refers to futile attempts at vaginal
delivery by allowing another 0.5 to 1 hour of pushing, or making one or
two more pulls with the vacuum.
Some hospitals are not equipped or staffed to respond to acute emergen-
cies. Thus, if a delivery delay from emergencies, such as a prolapsed cord,
placental abruption, or uterine rupture results in an adverse fetal outcome,
hospitals, as well as physicians, are exposed to significant liability. The
patient should be informed of the facility’s capabilities and, if appropriate,
offered delivery at a hospital better equipped for the patient’s condition.
Physicians who treat indigent patients are sued more frequently. Inner
city hospitals may be staffed with poorly trained physicians, thus placing
those patients at greater risk for harm. Residency training commonly entails
treating indigent patients. In combination with recognized communication
concerns and numerous ‘‘hand-offs,’’ there is greater chance for medical
error. Numerous suits result from a lack of appropriate communication
skills. Families often seek legal advice to learn what happened, rather
than because of a primary interest in money. When adverse outcomes occur,
physicians must discuss this with the patient and her family, addressing her
concerns and questions in a frank and open manner. A 1995 Journal of the
American Medical Association article reported that in many cases, while no
technical errors occurred, the practitioner generated enough misunderstand-
ing and anger to provoke a malpractice claim [3]. Furthermore, 85% of suits
were filed against 3% to 6% of doctors. The authors concluded that doctors
who are hurried, uninterested, or unwilling to listen to and answer questions
are at risk of suit, even if they practice quality medicine. Conversely, those
who are perceived as concerned, accessible, and willing to communicate are
sued far less.
Innovative early-intervention programs, such as the COPIC Insurance
Company of Colorado’s 3R’s, have demonstrated the effectiveness of
such programs in averting malpractice suits [4]. The 3R’s stand for ‘‘Recog-
nize-Respond-Resolve.’’ COPIC insurance may pay disability payments up
to $5,000, with additional reimbursement for out-of-pocket expenses up to
$25,000. In essence, the physician-patient relationship is preserved through
clear, candid communication regarding a treatment-related injury. The pa-
tient retains her right to pursue further legal action if she desires. The physi-
cian’s participation in the program does not limit their professional
coverage or raise their premiums. This program resulted in remarkably lower
payouts than traditional methods, where the average cost of paid claims was
 LIABILITY IN HIGH-RISK OBSTETRICS 619

over $250,000. As of December 2003, no cases had gone to litigation. Thus, in

effective programs, early and clear communication can reduce litigation
exposure.

Areas of litigation in obstetrics

Litigation centers on errors of omission or commission. Thus prime areas
for obstetrical litigation comprise the following:
1. Errors or omission in antenatal screening and diagnosis
2. Errors in ultrasound diagnosis
3. The neurologically impaired infant
4. Neonatal encephalopathy
5. Stillborn or neonatal death
6. Shoulder dystocia, with either brachial plexus injury or hypoxic injury
7. Vaginal birth after cesarean section
8. Operative vaginal delivery
9. Training programs (Resident supervision markedly impacts litigation
exposure. Increased used of nurse midwives and nurse practitioners
may increase ones liability exposure.)

Antenatal screening and diagnosis

Errors in antenatal screening and diagnosis are an increasing focus for lit-
igation. Such errors can lead to claims for wrongful birth, wrongful life, and
wrongful death. Although these specific claims may be prohibited by law in
some states, they serve as the basis for most suits relating to antenatal
diagnosis.
Wrongful birth is a claim for relief by the parents, who allege they
would have avoided conception or would have terminated a pregnancy if
they had been advised of the likelihood of giving birth to an impaired child
[5]. Classic examples are Tay-Sachs disease or cystic fibrosis. A wrongful
life claim is a cause of action by a special needs child, who claims damages
because he was conceived or not aborted because of the physician’s
negligence [6]. This cause of action is barred in most states. Wrongful death
is a cause of action arising when an otherwise normal pregnancy, which
has reached viability, is terminated because of a misdiagnosis [7]. An
example would be a misdiagnosis of renal agenesis resulting in pregnancy
termination.
ACOG now recommends offering antenatal screening for chromosomal
abnormalities to all pregnancy patients regardless of age [8]. In addition,
the broader availability of nuchal translucency screening establishes a stan-
dard of care in which most patients should be offered the opportunity for
first trimester screening. A physician failing to offer patients such diagnostic
testing is at risk for suit.
620 SHWAYDER

Ultrasound is routine in caring for obstetrical patients. Missed diagnosis

of fetal anomalies accounts for over a one-fourth of obstetrical malpractice
cases. Guidelines for proper performance of obstetric ultrasound examina-
tions have been established [9]. As such, this represents the standard of
care for obstetrical ultrasound. The best approach for obstetrical ultrasound
is to have properly trained or certified ultrasonographers performing
comprehensive studies on contemporary, well-maintained equipment, with
image interpretation by a qualified sonologist. Referral for consultation is
appropriate in confusing circumstances or when the acuity of the clinical
situation warrants enhanced evaluation and knowledge.
Failure to recommend further testing or procedures is an area of increasing
concern, particularly when evaluating possible genetic syndromes or chromo-
somal abnormalities. For example, isolated findings suggesting Down’s
syndrome, such as an intracardiac echogenic focus or minimal pyelectasis,
may be of no consequence. However, when multiple subtle findings are pres-
ent, the patient’s risk for chromosomal abnormalities should be recalculated
and, if appropriate, further testing recommended [10]. Prenatal screening for
genetic disease is also indicated in certain populations [11]. Failure to commu-
nicate these findings to the referring physician, however subtle, can place the
consultant at risk of suit.

Antepartum fetal assessment

High-risk pregnancies require antepartum fetal surveilalance [12]. Fetal
heart rate monitoring, ultrasound surveillance, amniotic fluid volumes,
Doppler studies, and cordocentesis are appropriate in pregnancies compli-
cated by conditions such as intrauterine growth restriction, twins, diabetes,
hypertension, severe preeclampsia, and sensitization, among others [13–15].
Guidelines for appropriate use establish an accepted standard of care. De-
viating from these guidelines requires substantiated decision making; other-
wise, physicians are at risk of a malpractice suit in the event of an adverse
outcome.

Intrapartum liability
Obvious liability lies with an adverse fetal or neonatal outcome. Intrapar-
tum management undergoes close scrutiny. The most devastating outcomes,
and thus costly awards, center on neurologically impaired infants and babies
with permanent neurologic deficits after shoulder dystocia.

Neurologically impaired infants

Almost 70% of neonatal encephalopathy is attributable to antepartum
events. However, 19% of newborns meet criteria for intrapartum hypoxia,
with 10% having a significant intrapartum event that may be associated
with intrapartum hypoxia. The ACOG Task Force on Neonatal
 LIABILITY IN HIGH-RISK OBSTETRICS 621

Encephalopathy and Cerebral Palsy determined four essential criteria that

must be met for a diagnosis of hypoxic encephalopathy [16]. These are:
1. Metabolic acidosis evidenced by an umbilical cord artery pH lower than
7 and a base deficit greater than 12 mmol/L;
2. Early onset neonatal encephalopathy in infants at more than 34 weeks
of gestation;
3. Cerebral palsy of the spastic quadriplegic or dyskinetic type; and
4. Exclusion of other etiologies.
This group also outlined criteria suggesting intrapartum timing including:
1. A sentinel hypoxic event;
2. Sudden and sustained bradycardia or absence of variability with persis-
tent, late, or variable decelerations after a sentinel hypoxic event;
3. Apgar scores of 0 to 3 beyond 5 minutes;
4. Multisystem involvement under 72 hours of age; or
5. Imaging with acute nonfatal cerebral abnormality.
These criteria are offered as mandatory findings to establish an intrapar-
tum hypoxic event leading to neonatal encephalopathy and, ultimately, ce-
rebral palsy. However, at least five jurisdictions have held that the criteria
are not dispositive, that is, not a final determination. The report can be ad-
mitted into evidence; however, all proffered opinions are subject to scrutiny
and cross-examination.
It is clear that careful attention to labor progress and fetal status, includ-
ing adequate documentation, enhances defensibility. Intrapartum fetal heart
rate changes must be recognized and responded to appropriately [17].
Prompt intervention and operative delivery, if indicated, minimize allega-
tions of negligence.

Shoulder dystocia with permanent palsy

Shoulder dystocia is an infrequent, and often unpredictable, nightmare
for the obstetrician [18]. However, the law evaluates whether the complica-
tion was foreseeable and, if not, whether appropriate maneuvers performed.
Recognized risk factors include a prior pregnancy complicated by shoulder
dystocia and resultant Erb’s palsy, macrosomia, and a midpelvic operative
delivery in fetuses with an estimated weight over 4000 grams [19]. An esti-
mated fetal weight over 5000 grams in nondiabetic pregnancies and over
4500 grams in diabetic pregnancies has been offered as justification for a pri-
mary cesarean section. Thus, a physician who overlooks the prior obstetrical
history, does not estimate the fetal weight in labor, or who pursues a midpel-
vic operative delivery in larger infants subjects him or herself to a claim of
negligence.
Controversy exists regarding the impact of active phase abnormalities or
second stage abnormalities on shoulder dystocia [20]. Although not reaching
significance, the incidence of shoulder dystocia is twice as frequent with
622 SHWAYDER

operative delivery in both diabetic (23.8% versus 12.0%) and nondiabetic

(13.3% versus 6.5%) patients. Thus, an operative delivery should be ap-
proached with care. Documentation of indications, fetal status, fetal posi-
tion, pelvic adequacy, number of pulls or pop-offs, and the immediate
neonatal status are critical to defense. A dictated operative note is also rec-
ommended, including all pertinent information, such as infant birth weight,
Apgar scores, cord gases, anesthesia, and estimated blood loss.
Several maneuvers are appropriate in the event of a shoulder dystocia
[21,22]. There is no required sequence of maneuvers, only that they be ap-
plied in an orderly and timely fashion [23]. Two specific maneuvers should
always be avoided: fundal pressure, which serves to further impact the an-
terior shoulder, and extreme lateral flexion of the spinal column and neck,
the presumed cause of stretch or avulsion injuries. A dictated or thoroughly
documented delivery note includes the aforementioned items, plus other
critical information [24]:
1. All providers present at the delivery
2. Note of the anterior shoulder
3. The time from recognition of the shoulder dystocia and delivery
4. All maneuvers used and in what order
5. Note if the infant moves all extremities after delivery
Documenting such information enhances defensibility of a shoulder dys-
tocia case.

Vaginal birth after cesarean section

Vaginal birth after cesarean section has come under great scrutiny. It is
a safe alternative in well-selected patients delivering in hospitals with appro-
priate resources [25]. However, recognized risks and the dire consequences
have prompted some states to impose practice guidelines for VBAC [26].
Physicians should document discussions of the risks and benefits of
VBAC and the hospital’s capabilities, with signed patient consent. Immedi-
ate physician availability and operative capabilities are required. If this can-
not be offered, then the patient should be transferred to a facility with these
capabilities.

Supervision of residents and advanced-care providers

Guidelines have been established for resident supervision, propagated
by the Centers for Medicare and Medicaid Services [27]. Attending super-
vision requires an understanding of the resident’s skill, training, and
knowledge. Thus, appropriate delegation of responsibility can occur while
keeping patient safety paramount. In most jurisdictions, the attending is
held responsible for actions of a resident under their supervision. Clear
and specific direction of expectations, communication, and documentation
 LIABILITY IN HIGH-RISK OBSTETRICS 623

are required. Documentation review and confirmation are attending

responsibilities that must comply with billing rules for supervised care
by residents.
Certified nurse midwives often have independent practice authority.
However, collaborative agreements may be required to independently
prescribe medications [28]. Written protocols, including scope of practice
and referral guidelines should be in place and carefully followed. Hospital
protocols and guidelines often dictate the level of supervision and consulta-
tion required. A physician employing a midwife is liable for any acts under
the doctrine of respondent superior. Vicarious liability occurs as it would for
an employer liable for the wrong of an employee if it was committed within
the scope of employment [29]. Thus, guidelines and protocols must be
followed to maintain defensibility of a case.

A primer in medical malpractice

A plaintiff must prove the following elements of negligence for a success-
ful malpractice suit:
1. A duty was owed to the patient by the health care provider,
2. There was a breach of that duty,
3. This breach was the proximate cause of the injury, and
4. The injury that resulted is compensable.
The duty to care for a patient is clear with an established physician-
patient relationship. However, such a relationship may exist if phone advice
is rendered or communication and advice is given via electronic communi-
cations [30]. The Emergency Medical Treatment and Active Labor Act
creates a physician-patient relationship in emergency situations. As such,
the physician is obligated to care for the patient until they are stable for
discharge or transfer to a better-suited facility.
Breach of the duty is breach of the standard of care. The standard of care is
how a similarly qualified practitioner would have managed the patient’s care
under the same or similar circumstances. Breach of duty is typically proven by
expert testimony. Testimony must be based on reliable and accepted scientific
principles [31]. ACOG recommends that expert testimony should withstand
peer review [32]. Breach of hospital protocols may be introduced to demon-
strate a deviation from the established standard of care. Thus, a physician’s
knowledge of such protocols is imperative in caring for patients and
preparing for successful case defense. Occasionally a breach of duty falls
under the doctrine of res ipsa loquitur, which holds but for the failure to
exercise due care the injury would not have occurred [29]. An example would
be a retained surgical instrument or operating on the wrong limb.
The breach must also be the proximate cause of the damages suffered by
the patient. In civil cases, such as medical malpractice, the level of proof re-
quired is by a preponderance of the evidence, that is, greater than or equal
624 SHWAYDER

to 51%. This is different from criminal cases requiring proof beyond a rea-
sonable doubt. Thus, if the breach resulted in at least a 51% likelihood of
the injury or outcome, then proximate cause can be proven.
Finally, the injury must be compensable, commonly called damages.
Damages are of three types: economic, noneconomic, and punitive. Eco-
nomic damages, also called ‘‘special damages,’’ compensate for the medical
costs of an injury, such as medical bills, rehabilitation, and loss of income.
Noneconomic damages, termed ‘‘general damages,’’ compensate for losses
that are not monetary, such as loss of consortium, loss of future fertility,
or pain and suffering. Punitive damages, termed ‘‘exemplary damages,’’
are awarded to punish a defendant for willful and wanton conduct, such
as sexual misconduct. The latter two categories are limited, or capped, in
many jurisdictions.
If a plaintiff’s case is successful and damages are awarded, each state or
jurisdiction has specific rules regarding responsibility for payment. If a state
follows joint and several liability, then each defendant is individually liable
for the entire award. Ultimately, they can seek reimbursement from the non-
paying parties, the right of subrogation. Proportional liability, also called
comparative negligence, can be pure or partial in nature. Proportional liabil-
ity allocates a portion of the blame to each defendant and, in certain cases,
the plaintiff. In pure comparative negligence, a plaintiff may receive recovery
even if their contribution to the injury is more than the defendant’s. How-
ever, the award is reduced by that percent contribution. In some states, the
plaintiff is barred from recovery if their contribution is more that 50%.
With partial comparative negligence, each defendant is responsible for the
portion of the damage award based on the allocated proportion of their fault.

Summary
This article has outlined the major causes of malpractice suits, focusing
on those in obstetrical practice. It has reviewed the prime areas in antepar-
tum and intrapartum care. Finally, understanding the basic elements of
medical malpractice allows a provider to better understand the nature of
a suit for medical negligence. The threat of a medical malpractice is ever
present in obstetrics. However, practicing contemporary, evidence-based
medicine, with compassion and excellent communication is the best way
to avoid alleged negligence. If a suit occurs, the best defense entails compre-
hensive documentation, particularly in recognized areas of risk.

References
[1] Wilson N, Strunk AL. Survey on professional liability. ACOG Clin Rev 2006;12(2):1, 13–6.
[2] Pritchard DJ. A Plaintiff Attorneys’ candid view of medical malpractice. Clin Perinatol 2005;
32:191–202.
[3] Hickson GB, Clayton EW, Entman SS, et al. Obstetricians’ prior malpractice experience and
patients’ satisfaction with care. JAMA 1994;272(20):1583–7.
 LIABILITY IN HIGH-RISK OBSTETRICS 625

[4] COPIC. 3Rs program showing proof of value of early communication. Copiscope 2001;104:
5–6.
[5] Keel v. Banach. So. 2d. vol. 624: Ala; 1993:1022.
[6] Kimble. Ala. Law. vol. 55: Ala; 1994:84.
[7] Lollar v. Tankersley. So. 2d. vol. 613: Ala; 1993:1249.
[8] American College of Obstetricians and Gynecologists. Screening for fetal chromosomal ab-
normalities. ACOG Practice Bulletin No. 77. Obstet Gynecol 2007;109:217–27.
[9] AIUM. AIUM practice guideline for the performance of obstetric ultrasound examinations.
Laurel (MD): AIUM; 2007.
[10] Benacerraf BR, Nadel A, Bromley B. Identification of second-trimester fetuses with autoso-
mal trisomy by use of a sonographic scoring index. Radiology 1994;193:135–40.
[11] ACOG. Prenatal and preconceptual carrier screening for genetic disease in individuals of
eastern European Jewish descent. Washington, DC: ACOG; 2004.
[12] American College of Obstetricians and Gynecologists. Antepartum fetal surveillance.
ACOG Practice Bulletin No. 9. October 1999.
[13] Badawi N, Kurinczuk JJ, Keogh JM, et al. Antepartum risk factors for newborn encepha-
lopathy: the Western Australian case-control study. BMJ 1998;317:1549–53.
[14] Low JA, Galbraith RS, Muir D, et al. Intrauterine growth retardation: a study of long-term
morbidity. Am J Obstet Gynecol 1982;142:670–7.
[15] Mari G, Roberts A, Detti L, et al. Perinatal morbidity and mortality rates in severe twin-twin
transfusion syndrome: results of the international amnioreduction registry. Am J Obstet
Gynecol 2001;185:708–15.
[16] ACOG, AAP. Neonatal encephalopathy and cerebral palsy. Defining the pathogenesis and
pathophysiology. Washington, DC: ACOG; 2003.
[17] American College of Obstetricians and Gynecologists. Intrapartum fetal heart rate monitor-
ing. ACOG Practice Bulletin No. 70. Obstet Gynecol 2005;106:1453–61.
[18] Langer O, Berkus MD, Huff RW, et al. Shoulder dystocia: should the fetus weighing O 4000
g be delivered by cesarean section? Am J Obstet Gynecol 1991;165:831–7.
[19] Christoffersson M, Kannisto P, Rydhstroem H, et al. Shoulder dystocia and brachial plexus
injury: a case-control study. Acta Obstet Gynecol Scand 2003;82:147–51.
[20] McFarland M, Hod M, Piper JM, et al. Are labor abnormalities more common in shoulder
dystocia? Am J Obstet Gynecol 1995;173:1211–4.
[21] Gherman RB, Ouzounian JG, Goodwin TM. Obstetric maneuvers for shoulder dystocia and
associated fetal morbidity. Am J Obstet Gynecol 1998;178(6):1126–30.
[22] Ramsey PS, Ramin KD, Field CS, et al. Shoulder dystocia. Rotational maneuvers revisited.
J Reprod Med 2000;45(21):85–8.
[23] Schifrin B. Shoulder dystocia and instrumental trauma. Paper presented at: American Con-
ference Institute: reducing the risk of obstetric malpractice. Orlando (FL), 2004.
[24] Deering S, Poggi S, Hodor J, et al. Evaluation of residents’ delivery notes after a simulated
shoulder dystocia. Obstet Gynecol 2004;104:667–70.
[25] Miller DA, Diaz FG, Paul RH. Vaginal birth after cesarean: a 10-year experience. Obstet
Gynecol 1994;84:255–8.
[26] Studnicki J, Remmel R, Campbell R, et al. The impact of legislatively imposed practice
guidelines on cesarean section rates: the Florida experience. Am J Med Qual 1997;12:62–8.
[27] CMS. Guidelines for teaching physicians, interns, and residents. Washington, DC: Centers
for Medicare & Medicaid Services; 2006.
[28] Colorado Revised Statutes. C.R.S. 12-36-101; 2003.
[29] Black’s Law Dictionary. St. Paul (MN): West Group; 1999.
[30] Shwayder J. Cybermedicine: what evil web we weave. Preventive Law J. U of Denver College
of Law. May 2001. p. 7–11.
[31] Daubert v. Merrell Dow Pharmaceuticals, Inc. U.S. vol. 509: U.S.; 1993:579.
[32] American College of Obstetricians and Gynecologists. Expert testimony. ACOG Committee
Opinion No. 374. Obstet Gynecol 2007;110:445–56.