HOW DO I . . .?

How do we transfuse blood components in cirrhotic patients

undergoing gastrointestinal procedures?

Sean G. Yates,1 Chakri Gavva,1 Deepak Agrawal,2 and Ravi Sarode1

The liver plays a pivotal role in hemostasis.

T
he incidence of chronic liver disease continues
Consequently, patients with cirrhosis frequently
to increase with an estimated 150,000 new
demonstrate abnormal coagulation profiles on routine
patients diagnosed each year in the United
laboratory tests. These tests mainly reflect decreased
States. Cirrhosis is present in 20% of these
procoagulant proteins. However, in cirrhosis, complex
patients at initial presentation.1,2 Although significant, this
changes also occur in anticoagulant and fibrinolytic
number may considerably underestimate the actual dis-
pathways. Recent evidence demonstrates that patients
ease prevalence with reports that one-third of patients
with cirrhosis exist in a state of hemostatic rebalance.
remain undiagnosed until the time of autopsy.3,4 The rate
Accordingly, routine tests inadequately represent
of morbidity and mortality associated with this population
hemostatic alterations in these patients. Unfortunately,
of patients is high, globally accounting for 1.03 million
these tests are regularly used to guide the transfusion of
deaths per year.5
blood components with the assumption that they will
The liver provides a vital role in maintaining normal
correct laboratory abnormalities and improve hemostasis
in a bleeding patient or prevent excessive bleeding
hemostasis as it synthesizes procoagulant and anticoagu-
following a procedure. With an absence of both accurate lant proteins, fibrinolytic and antifibrinolytic proteins, and
laboratory testing to assess hemostasis and evidence- thrombopoietin and degrades activated coagulation fac-
based guidelines to direct the transfusion of blood tors. As a consequence, significant abnormalities of hemo-
components, management of patients with cirrhosis stasis occur in cirrhotic patients. Conventional hemostasis
poses a significant challenge to clinicians. Therefore, we testing such as platelet (PLT) count, prothrombin time
developed multidisciplinary guidelines for the (PT), and partial thromboplastin time (PTT) are often
periprocedural transfusion of blood components in
patients with cirrhosis based on concurrent evidence and
ABBREVIATIONS: AASLD 5 American Association for the
personal experience at our medical center.
Study of the Liver Diseases; AT 5 antithrombin; INR 5
international normalized ratio; PCC(s) 5 prothrombin
complex concentrate(s); PT(s) 5 prothrombin time(s);
PTT 5 partial thromboplastin time; rFVIIa 5 recombinant
Factor VIIa; ROTEM 5 rotational thromboelastometry.

From the 1Department of Pathology, Division of Transfusion

Medicine and Hemostasis, and the 2Department of Internal
Medicine, Division of Division of Digestive and Liver Diseases,
University of Texas Southwestern Medical Center, Dallas, Texas.
Address reprint requests to: Ravi Sarode, MD, Director,
Division of Transfusion Medicine and Hemostasis, Department
of Pathology, UT Southwestern Medical Center, 5323 Harry
Hines Boulevard, Dallas, TX 75390; e-mail: Ravi.sarode@
UTsouthwestern.edu.
Received for publication October 13, 2015; revision
received December 2, 2015; and accepted December 18, 2015.
doi:10.1111/trf.13495
C 2016 AABB
V
TRANSFUSION 2016;56;791–798

Volume 56, April 2016 TRANSFUSION 791

YATES ET AL.

severe, Child-Pugh B 5 intermediate severity, Child-Pugh

TABLE 1. Hemostatic alterations in chronic liver C 5 most severe).10 The splenic sequestration of PLTs and
disease
decreased liver production of thrombopoietin are the two
Anticoagulant Procoagulant
most common underlying mechanisms leading to throm-
Primary hemostasis
Thrombocytopenia " VWF bocytopenia in cirrhosis.11 Other factors that may contrib-
# ADAMTS-13 ute to a decline in PLT counts include the autoimmune
Secondary hemostasis
# FII, FV, FVII, F IX, FX, FXI # PC, PS, AT, a2-macroglobulin
destruction of PLTs12 and direct suppression of marrow
and heparin cofactor II PLT production by pathologic causes of cirrhosis such as
Hypo/Dysfibrinogenemia " FVIII hepatitis B, hepatitis C, or alcohol.13-15
Fibrinolysis
# TAFI, a2-antiplasmin # Plasminogen Previous studies have advocated the significance of
" t-PA levels " PAI-1 qualitative PLT dysfunction in cirrhosis.16 However, eval-
ADAMTS13 5 a disintegrin and metalloproteinase with a throm- uations of PLT function under flow conditions have dem-
bospondin type 1 motif, member 13; PAI-1 5 plasminogen onstrated that primary hemostasis remains intact.17
activator inhibitor-1; PC 5 protein C; PS 5 protein S; TAFI 5
thrombin-activatable fibrinolysis inhibitor; t-PA 5 tissue pathway Retention of PLT function may be explained by elevated
activator. circulating concentrations of von Willebrand factor
(VWF) with one study exhibiting VWF concentrations of
380, 500, and 760% in Child-Pugh A, Child-Pugh B, and
Child-Pugh C cirrhotic patients relative to 107% in
abnormal in patients with cirrhosis, suggesting an
healthy controls.17 Accompanying the increase in VWF
increased risk of bleeding. However, evidence from studies
concentrations is a concomitant decline in circulating
utilizing global hemostasis testing demonstrate that when
ADAMTS13, a metalloprotease enzyme that maintains a
stable, patients with cirrhosis are in a state of rebalanced
hemostasis despite abnormal routine tests.6 This hemo- normal distribution of VWF multimers through proteo-
static state is tenuous with minor clinical perturbations lytic digestion of ultralarge VWF.18 Together, these
potentially tipping a patient to either a hypocoagulable or changes result in an increased concentration of larger,
a hypercoagulable state. Consequently, patients with cir- more functional VWF multimers that augment PLT func-
rhosis may present with bleeding, thromboembolic tion and potentially offset thrombocytopenia.17,18
events, or in some cases both.7
Despite the inherent limitations of conventional Coagulation
hemostatic tests, they are frequently used by clinicians to
Cirrhosis is typically characterized by impaired hepatic
guide the transfusion of PLTs, plasma, and cryoprecipitate.
synthetic function, leading to a reduction in circulating
This practice persists although limited evidence demon-
levels of procoagulants (fibrinogen, Factor (F)II, FV, FVII,
strates the benefits of using arbitrary laboratory values to
F IX, FX, and FXI).19-22 Accompanying this decline in pro-
direct prophylactic or therapeutic transfusions.8 In this
coagulant proteins is a parallel decline in natural antico-
article, we describe hemostatic changes that characterize
agulant proteins (protein C, protein S, antithrombin [AT],
cirrhosis, discuss the limitations of routine laboratory test-
a2-macroglobulin, and heparin cofactor II).19-22 Among
ing frequently used to manage such patients, and provide
these anticoagulants, AT appears to have the most signifi-
our institution-specific, multidisciplinary guidelines for
cant influence on in vitro thrombin generation.23 Con-
the periprocedural transfusion of blood components in
centrations of AT decline in cirrhosis with measured
patients with cirrhosis.
activities of 43 6 14% relative to 109 6 10% in healthy
controls.24 This concomitant decline in anticoagulant
HEMOSTATIC CHANGES IN CIRRHOSIS proteins, specifically AT, accompanied by elevated levels
of FVIII may explain the ability of cirrhotic patients to
Procoagulant and anticoagulant changes of cirrhosis are
generate thrombin concentrations equivalent to healthy
summarized in Table 1.
persons.25,26
Dysfibrinogenemia is common in cirrhosis, occurring
PLTs in up to 70% of patients secondary to the addition of
Thrombocytopenia (PLTs < 150 3 109/L) may be observed excess sialic acid residues to fibrinogen, impairing its abil-
in as many as 76% of cirrhotic patients with an estimated ity to assemble properly.27,28 Dysfunctional fibrinogen
1% of patients demonstrating PLT counts of less than 50 may be offset by normal to increased fibrinogen levels in
3 109/L.9 The degree of thrombocytopenia directly corre- patients with Child-Pugh A cirrhosis and Child-Pugh B cir-
lates with the extent of liver disease as reflected by the rhosis demonstrating plasma concentrations of 210 mg/dL
Child-Pugh classification (a parameter often used to clas- (117% of normal) and 240 mg/dL (137% of normal),
sify the severity of liver disease: Child-Pugh A 5 least respectively.29

792 TRANSFUSION Volume 56, April 2016

 TRANSFUSION IN CIRRHOSIS

Patients with cirrhosis frequently demonstrate poor Finally, while cirrhotic patients with abnormally pro-
nutritional status or malabsorption of fat-soluble vitamins longed coagulation test values frequently demonstrate
by the terminal ileum secondary to obstruction of the bili- bleeding complications, a significant proportion may
ary tract.30 Collectively, these findings lead to vitamin develop thrombotic complications. Large epidemiologic
K deficiency and compromise the synthesis of vitamin studies have shown a higher incidence of thrombotic
K–dependent factors (FII, FVII, F IX, FX, proteins C and S). complications in cirrhotic patients relative to the general
population with portal vein thrombosis occurring in 8% to
Fibrinolysis 26% of patients with end-stage liver disease.39-41 More-
over, cirrhosis has been identified as an independent risk
Derangements in fibrinolysis accompany cirrhosis. Patients
factor for the development of both unprovoked deep vein
generally exhibit diminished levels of fibrinolysis inhibitors
thrombosis and pulmonary emboli.42
such as a2-antiplasmin and thrombin-activatable fibrinoly-
sis inhibitor6,21,28 in addition to increased levels of tissue
plasminogen activator.28 However, studies provide little evi- JUSTIFICATION FOR RESTRICTIVE
dence of increased fibrinolytic activity, likely explained by a TRANSFUSION PRACTICES
reduction in plasminogen concentrations and elevated
While other forms of portal hypertension-related bleeding
plasminogen activator inhibitor levels generated by endo-
occur (i.e., portal hypertensive gastropathy) variceal
thelial cells.21,31 bleeding not only constitutes the largest number of pre-
sentations to the emergency department but are also
associated with the highest mortality.43 Factors that con-
LIMITATIONS OF CONVENTIONAL
tribute to its incidence include local vascular abnormal-
COAGULATION TESTING
ities and increased portal pressure.44 Little to no
The PT and PTT were developed to diagnose deficiencies relationship has been demonstrated between variceal
in individual procoagulant proteins in patients with bleed- bleeding and the coagulopathy associated with cirrho-
ing diatheses and for monitoring anticoagulation thera- sis.19,44,45 Attempting to correct abnormal laboratory val-
pies. Despite their intended purpose, routine hemostasis ues with therapeutic transfusions may paradoxically
tests often influence the clinical management of patients exacerbate bleeding by further increasing intravascular
with cirrhosis. These tests are regularly prolonged in cir- volumes and pressures.46
rhosis and primarily reflect a decline in the plasma con- Restrictive transfusion policies are now widely used in
centrations of procoagulants. Although sensitive to liver transplantation and have led to significant reductions
procoagulants, routine tests do not reflect parallel changes in intraoperative blood loss and transfusion require-
in concentrations of natural anticoagulants. Consequently, ments.47,48 By decreasing central venous pressures (and
conventional tests present an incomplete picture of the consequently portal pressures) through intraoperative
hemostatic changes occurring in cirrhosis. Furthermore, phlebotomy, 79% of patients at a single center received no
these abnormal tests fail to predict bleeding outcomes in perioperative blood products during liver transplanta-
several clinical scenarios including patients with esopha- tion.49 The success of restrictive transfusion strategies calls
geal varices or after liver biopsy or those undergoing liver into question prophylactically transfusing blood products
transplantation.8 Similarly, correction of prolonged test for less invasive procedures and supports the concept of
values through the administration of recombinant FVIIa rebalanced hemostasis. The significant complications
(rFVIIa) or plasma does not improve the incidence or associated with the transfusion of blood products that
severity of bleeding in cirrhosis.32,33 include transfusion-related acute lung injury, transfusion-
Rotational thromboelastometry (ROTEM) is increas- associated circulatory overload, anaphylactic reactions,
ingly used in perioperative settings, largely in liver and hemolytic reactions, and transfusion-transmitted infec-
cardiac surgeries but also in trauma and obstetrics.34-37 tions should also be considered.
ROTEM differs from conventional coagulation studies in
that it utilizes whole blood as opposed to plasma and
BEFORE TRANSFUSION
evaluates the rate and strength of clot formation and dis-
solution. Therefore, ROTEM may demonstrate increased Managing comorbidities
sensitivity to physiologic variations in coagulation, antico- Before transfusing blood components, underlying clinical
agulation, and fibrinolysis and provide insight into the conditions that contribute to bleeding outcomes should
potential risks of bleeding or thrombosis. Although its role be evaluated. Notably, bacterial infections place patients
in cirrhosis has not been established, ROTEM indices at a greater risk for gastrointestinal bleeding.50-52 As such,
have been reported to correspond with the extent of liver patients at risk for bacterial infections should receive
disease and coagulopathy.38 prophylactic antibiotics to optimize hemostatic function.

Volume 56, April 2016 TRANSFUSION 793

YATES ET AL.

Similarly, patients with renal impairment exhibit an Moreover, a study evaluating thrombin generation in
increased incidence of bleeding, frequently related to ure- cirrhotic patients with a median PLT count of 39 3 109/L
mia and volume overload.53,54 Dialysis is beneficial (range, 16 3 109264 3 109/L) who received prophylactic
because it corrects uremia while also addressing volume PLT transfusions demonstrated a marginal increase in PLT
overload, thereby decreasing the risk of bleeding. count (52 3 109/L [range, 19 3 109291 3 109/L];
p < 0.001), after transfusion.58 Despite PLT counts of less
than 50 3 109/L, the median thrombin generation varia-
BLOOD COMPONENT TRANSFUSION
bles of patients were within normal range before the
IN CIRRHOSIS
transfusion of PLTs. Moreover, patients exhibited endoge-
PLTs nous thrombin potentials above the upper limit of normal
Although the American Association for the Study of the despite thrombocytopenia and demonstrated a minimal
Liver Diseases (AASLD) suggests prophylactic transfusion increase in endogenous thrombin potentials (1,066 [685-
of PLTs with a PLT count of not more than 50 3 109/L, 2036 [vs. 1084 [273-1618]; p value not significant) after
limited high-quality evidence exists to support this prac- PLT transfusions. Collectively these findings along with
tice.55 These recommendations are principally based on the introduction of additional volumes that may exacer-
either retrospective data or consensus panel agreement bate bleeding argue against the use of prophylactic PLT
and were not developed with the input of experts in the transfusions.
field of transfusion medicine and hemostasis.
PLT counts often guide the transfusion of PLTs, yet University of Texas Southwestern (UTSW)
this measure does not account for the possible contribu- multidisciplinary guidelines: PLTs
tion of PLTs sequestered in the spleen or the presence of A summary of institutional guidelines is given in Table 2. In
larger, more functional VWF multimers augmenting PLT cirrhotic patients, we discourage preprocedure PLT transfu-
function. An elegant in vivo study by Aster56 demonstrated sions strictly to achieve a given threshold. While limited
that in patients with splenomegaly, up to 90% of radiola- data for a PLT threshold exist, a retrospective study evaluat-
beled PLTs underwent sequestration within minutes of ing patients undergoing transjugular liver biopsy demon-
transfusion. After epinephrine is administered, these strated that no hemorrhage-related complications occurred
radiolabeled PLTs redistributed to the peripheral circula- when PLT counts were more than 30 3 109/L.59 Therefore,
tion. We propose that a similar redistribution of PLTs in patients with cirrhosis and splenomegaly and who are
occurs after the endogenous release of epinephrine in scheduled to undergo high-risk procedures, we discourage
response to a hemostatic challenge encountered during preprocedure PLT transfusions to achieve a particular num-
an invasive procedure or bleeding. Thus, peripheral PLT ber unless counts are less than 30 3 109/L in which case
counts may not represent the actual number of PLTs avail- one dose (equivalent to single-donor apheresis PLTs or
able at the time of a hemostatic challenge. Furthermore, five-pooled whole blood–derived PLTs) of intraprocedural
achieving PLT counts set forth by many guidelines is often PLTs may be given, as the peripheral count is unlikely to
difficult, with as little as 10% of transfused PLTs remaining increase significantly or be maintained for a meaningful
in the circulation of patients with splenomegaly.56 These amount of time (Table 2). Consequently, we do not recom-
poor posttransfusion PLT counts result in repeat transfu- mend evaluating PLT increments after transfusions.
sions, leading to excessive intravascular volumes with
increased portal pressure and bleeding risk, and may Plasma
delay invasive procedures. In one survey, approximately one-third of all plasma uti-
A recent randomized clinical trial involving patients lized in the hospital setting was attributed to patients with
with cirrhosis assessed the use of a thrombopoietin recep- hepatobiliary diseases including cirrhosis.60 This practice
tor agonist (Eltrombopag) to increase PLT counts before contradicts recommendations made by organizations
elective invasive procedures.57 Both the control and the such as the AASLD that recognize limitations of conven-
treatment groups demonstrated a median baseline PLT tional coagulation tests (PT and international normalized
count of 40 3 109 L. The study was prematurely termi- ratio [INR]) and discourage using arbitrary values to influ-
nated due to an increased incidence of thromboembolic ence the transfusion of plasma.55 Furthermore, this prac-
events in six patients in the treatment group with five of tice ignores evidence demonstrating that although plasma
them demonstrating PLT counts of more than 200 3 109 may either transiently shorten or rarely normalizes PTs, it
L. These findings suggest attempts to increase PLT counts has an insignificant impact on thrombin generation.61
may place patients at an increased risk of thrombosis. The preceding observation may explain the limited effi-
Notably, higher preprocedure PLT counts achieved by the cacy of plasma transfusions in cirrhosis.62 Indeed, intrao-
treatment group did not improve bleeding outcomes rela- perative plasma transfusions are associated with adverse
tive to controls.57 outcomes in liver transplantation.49,63 Again, additional

794 TRANSFUSION Volume 56, April 2016

 TRANSFUSION IN CIRRHOSIS

bleeding cirrhotic patients.45,64 Early studies evaluating

TABLE 2. Transfusion guidelines for cirrhotic patients with congenital hypofibrinogenemia demon-
patients undergoing gastrointestinal procedures
at University of Texas Southwestern Medical Center strated little hemostatic benefit when increasing fibrino-
at Dallas gen levels to more than 100 mg/dL.65 Furthermore,
Component Recommendation previous studies of patients with massive hemorrhage
PLT transfusions for mild Transfusing PLTs with have shown that hemostatic defects arise with fibrinogen
to moderate risk counts > 30 3 109 L not concentrations of less than 100 g/dL.66 Consequently,
procedures* recommended.
PLT transfusions for PLTs may be warranted consensus committees advocate maintaining a minimum
high-risk procedures† during procedure with fibrinogen concentration of 100 mg/dL during periopera-
PLT counts  30 3 109 L21
tive bleeding or massive blood loss.67,68
(for percutaneous liver
biopsy  50 3 109 L21) Fibrinogen should be supplemented through the use
Plasma transfusions in INR  2.5, no plasma of cryoprecipitate, which has a higher concentration of
patients undergoing a indicated
fibrinogen per unit volume relative to plasma. Accord-
procedure or actively INR  2.6, give 10 mg IV
bleeding vitamin K and plasma ingly, there is less risk of volume overload with cryopreci-
transfusion at 10 mL/kg pitate transfusion.69
(provided that fibrinogen
is > 100)‡
Cryoprecipitate With fibrinogen UTSW multidisciplinary guidelines: cryoprecipitate
concentrations < 100 mg/dL,
Most conventional hemostasis tests (i.e., PT and PTT)
transfuse one dose of
cryoprecipitate§ have a limited sensitivity to hypofibrinogenemia unless
RBC transfusion in Transfuse only with fibrinogen concentrations are significantly reduced
hemodynamically stable Hb < 7 g/dL
(<75 mg/dL in our laboratory). Fibrinogen concentrations
patients without
cardiovascular disease should be evaluated in all cirrhotic patients. In patients
* Moderate-risk procedures include the following: diagnostic scheduled to undergo an invasive procedure with or with-
endoscopies with or without biopsy, endoscopic interventions out active bleeding and fibrinogen concentrations of not
for gastrointestinal bleeding such as clip placement or cauteri-
more than 100 mg/dL, we recommend transfusing one
zation or epinephrine injection, elective variceal banding, rou-
tine screening colonoscopy (polyps up to 1 cm in size can be adult dose of cryoprecipitate (10 prepooled units; Table 2).
biopsied or removed with cold snare).
† High-risk—procedures involving significant disruption of mucosa
such as endoscopic ultrasound with biopsy or fine-needle aspi-
Red blood cells
ration, endoscopic retrograde cholangiopancreatography with Patients with gastrointestinal bleeding often require red
sphincterotomy, snare polypectomy (for polyps > 1 cm in size),
blood cell (RBCs) transfusions. Recent studies have ques-
and endoscopic dilation.
‡ If INR does not correct with IV vitamin K, it is likely due to sig- tioned the traditional practice of transfusing RBCs to
nificant liver dysfunction or significant hypofibrinogenemia, in achieve hemoglobin (Hb) of 9 to 10 g/dL. A study evaluat-
which case the transfusion of cryoprecipitate should be
ing outcomes in patients with upper gastrointestinal
considered.
§ One dose 5 10 units. bleeding classified patients into two categories: a restric-
tive transfusion group (Hb < 7 g/dL) and a liberal transfu-
sion group (Hb < 9 g/dL). In Child-Pugh Class A and B
cirrhotic patients, the restrictive transfusion group dem-
volume introduced through plasma transfusions may
onstrated a significant reduction in both bleeding and
increase bleeding risks by raising intravascular pressures.
mortality at 6 weeks compared to the liberal transfusion
group.46 No difference was observed in mortality or bleed-
UTSW multidisciplinary guidelines: plasma ing between Child-Pugh Class C transfusion groups.46 Lib-
Our institutional guidelines discourage administering eral transfusion groups also demonstrated a significant
plasma before or during procedures in cirrhotic patients increase in the mean hepatic venous pressure gradient
with an INR of not more than 2.5 (Table 2). If the INR is not exhibited in restrictive transfusion groups, a finding
more than 2.5, we recommend 10 mg of intravenous (IV) that may explain the higher rates of bleeding with liberal
vitamin K and checking fibrinogen. In our experience vita- transfusions.46 It has also been suggested that RBC trans-
min K alone and/or cryoprecipitate for fibrinogen of less fusion may counteract the splanchnic vasoconstrictive
than 100 mg/dL corrects INR to less than 2.5. However, if response after hypovolemic hemorrhage, resulting in a
there is no correction or the patient is actively bleeding rebound increase in splanchnic blood flow and pressure.
we recommend transfusing plasma (10 mL/kg). Collectively, these changes may inhibit hemostasis.46

Cryoprecipitate UTSW multidisciplinary guidelines: RBCs

No consensus exists regarding appropriate transfusion When managing variceal bleeding, the AASLD recom-
thresholds for repletion of fibrinogen in nonbleeding or mends maintaining a Hb level of at least 8 g/dL.43

Volume 56, April 2016 TRANSFUSION 795

YATES ET AL.

However, in accordance with recently published evidence- patients with chronic liver disease.76 Despite these differ-
based AABB guidelines,70 we recommend transfusing ences that exist between acute and chronic disease, the
RBCs in hemodynamically stable patients without cardiac ability of laboratory tests to identify patients at risk for
disease only when the Hb level is less than 7 g/dL (Table bleeding remains limited. Accordingly, restrictive transfu-
2). Only the smallest volumes of blood necessary to sion guidelines that restrict utilization of blood compo-
achieve a Hb level of 7 g/dL should be used as transfusion nents should be used in both patient populations.
of additional volume may raise portal pressures and In conclusion, these guidelines were successfully
increase bleeding risk. implemented within the preceding year. Faculty members
from both gastroenterology and anesthesiology depart-
Use of other hemostatic agents ments ensure adherence to the guidelines while the trans-
Kcentra (CSL Behring, King of Prussia, PA) is a Food and fusion medicine and hemostasis service prospectively
Drug Administration–approved four-factor prothrombin assist with transfusion support of patients. Additional
complex concentrate (PCC) licensed for the reversal of efforts should be made to standardize transfusion prac-
vitamin K antagonists in patients who are bleeding or tices and advance research in cirrhosis. Furthermore,
undergoing an urgent surgical procedure.64 The PCC is high-quality prospective clinical trials are needed to char-
a plasma-derived lyophilized concentrate of vitamin acterize the hemostatic profile of cirrhotic patients with
the aim of improving clinical management of this growing
K–dependent factors (FII, FVII, F IX, FX, PC, and PS).71
patient population.
Currently, limited data exist regarding the safety and effi-
cacy of PCCs in cirrhosis. A previous study evaluating an
earlier preparation of PCCs (with a lower concentration of CONFLICT OF INTEREST
FVII) demonstrated effective hemostasis in patients with
The authors have disclosed no conflicts of interest.
liver disease.64,72 However, this study only enrolled 21
patients and did not contain a control arm of cirrhotic
patients who did not receive PCCs. Furthermore, previous REFERENCES
case reports have documented thrombotic complications
1. Scaglione S, Kliethermes S, Cao G, et al. The epidemiology of
in patients with liver disease receiving PCCs.73 Until more
cirrhosis in the United States: a population-based study.
data are available, we do not recommend administering
J Clin Gastroenterol 2015;49:690-6.
PCCs in bleeding cirrhotic patients or those undergoing a
2. Kim Y, Ejaz A, Tayal A, et al. Temporal trends in population-
procedure.
based death rates associated with chronic liver disease and
A rFVIIa (NovoSeven, Novo Nordisk A/S, Bagsvaerd,
liver cancer in the United States over the last 30 years. Can-
Denmark) is approved for the control of bleeding in
cer 2014;120:3058-65.
patients with congenital FVII deficiency, hemophilia A or
3. Giarelli L, Melato M, Laurino L, et al. Occurrence of liver cir-
B with inhibitors, or Glanzmann’s thrombasthenia.74 A
rhosis among autopsies in Trieste. IARC Sci Publ 1990;37-43.
randomized controlled trial of Child-Pugh Class B and
4. Graudal N, Leth P, Mårbjerg L, et al. Characteristics of cirrho-
Class C cirrhotic patients admitted with an upper gastro-
sis undiagnosed during life: a comparative analysis of 73
intestinal bleed demonstrated correction in INR from a
undiagnosed cases and 149 diagnosed cases of cirrhosis,
mean 2.0 to 1.3 and from 2.1 to 1.5 in those receiving 600
detected in 4929 consecutive autopsies. J Intern Med 1991;
and 300 mg/kg rFVIIa, respectively. However, there was no
230:165-71.
difference in acute bleeding, prevention of rebleeding 5. Tsochatzis EA, Bosch J, Burroughs AK. Liver cirrhosis. Lancet
within 24 hours, or 5-day mortality in patients who 2014;383:1749-61.
received rFVIIa relative to placebo.74 There was also an 6. Tripodi A, Mannucci PM. The coagulopathy of chronic liver
increased rate of arterial thromboembolic events in the disease. N Engl J Med 2011;365:147-56.
group receiving rVIIa.74 Again, until more data are avail- 7. Tripodi A, Primignani M, Mannucci PM. Abnormalities of
able establishing the benefits and risks of rFVIIa in cir- hemostasis and bleeding in chronic liver disease: the para-
rhotic patients, we do not advocate the use of rFVIIa in digm is challenged. Intern Emerg Med 2010;5:7-12.
patients with cirrhosis. 8. Segal JB, Dzik WH. Paucity of studies to support that abnor-
mal coagulation test results predict bleeding in the setting of
Acute liver disease invasive procedures: an evidence-based review. Transfusion
While acute and chronic liver disease share the same over- 2005;45:1413-25.
all characteristics of rebalanced hemostasis, significant 9. Afdhal N, McHutchison J, Brown R, et al. Thrombocytopenia
differences between distinct forms of liver disease do associated with chronic liver disease. J Hepatol 2008;48:
exist. In acute liver failure, the decrease in procoagulant 1000-7.
and anticoagulant proteins is generally more marked,75 10. Giannini EG, Botta F, Borro P, et al. Application of the platelet
and the elevation in INR is often greater in comparison to count/spleen diameter ratio to rule out the presence of

796 TRANSFUSION Volume 56, April 2016

 TRANSFUSION IN CIRRHOSIS

oesophageal varices in patients with cirrhosis: a validation 29. de Maat MP, Nieuwenhuizen W, Knot EA, et al. Measuring
study based on follow-up. Dig Liver Dis 2005;37:779-85. plasma fibrinogen levels in patients with liver cirrhosis. The
11. Hayashi H, Beppu T, Shirabe K, et al. Management of throm- occurrence of proteolytic fibrin (ogen) degradation products
bocytopenia due to liver cirrhosis: a review. World J Gastro- and their influence on several fibrinogen assays. Thromb
enterol 2014;20:2595-605. Res 1995;78:353-62.
12. Kajihara M, Kato S, Okazaki Y, et al. A role of autoantibody- 30. Kujovich JL. Hemostatic defects in end stage liver disease.
mediated platelet destruction in thrombocytopenia in Crit Care Clin 2005;21:563-87.
patients with cirrhosis. Hepatology 2003;37:1267-76. 31. Lisman T, Leebeek FW, Mosnier LO, et al. Thrombin-activat-
13. Sing GK, Prior S, Fernan A, et al. Hepatitis B virus differen- able fibrinolysis inhibitor deficiency in cirrhosis is not asso-
tially suppresses myelopoiesis and displays tropism for ciated with increased plasma fibrinolysis. Gastroenterology
immature hematopoietic cells. J Virol 1993;67:3454-60. 2001;121:131-9.
€ O. Excess alcohol consumption
14. Latvala J, Parkkila S, Niemela 32. Lodge JP, Jonas S, Jones RM, et al. Efficacy and safety of
is common in patients with cytopenia: studies in blood and repeated perioperative doses of recombinant factor VIIa in
bone marrow cells. Alcohol Clin Exp Res 2004;28:619-24. liver transplantation. Liver Transpl 2005;11:973-9.
15. Olariu M, Olariu C, Olteanu D. Thrombocytopenia in 33. Planinsic RM, van der Meer J, Testa G, et al. Safety and effi-
chronic hepatitis C. J Gastrointestin Liver Dis 2010;19:381-5. cacy of a single bolus administration of recombinant factor
16. Tripodi A, Mannucci PM. Abnormalities of hemostasis in VIIa in liver transplantation due to chronic liver disease.
chronic liver disease: reappraisal of their clinical significance Liver Transpl 2005;11:895-900.
and need for clinical and laboratory research. J Hepatol 34. Roullet S, Pillot J, Freyburger G, et al. Rotation thromboelas-
2007;46:727-33. tometry detects thrombocytopenia and hypofibrinogenae-
17. Lisman T, Bongers TN, Adelmeijer J, et al. Elevated levels of von mia during orthotopic liver transplantation. Br J Anaesth
Willebrand Factor in cirrhosis support platelet adhesion despite 2010;104:422-8.
reduced functional capacity. Hepatology 2006;44:53-61. 35. Rugeri L, Levrat A, David JS, et al. Diagnosis of early coagula-
18. Feys HB, Canciani MT, Peyvandi F, et al. ADAMTS13 activity tion abnormalities in trauma patients by rotation thrombe-
to antigen ratio in physiological and pathological conditions lastography. J Thromb Haemost 2007;5:289-95.
associated with an increased risk of thrombosis. Br J Haema- 36. Shore-Lesserson L, Manspeizer HE, DePerio M, et al. Throm-
tol 2007;138:534-40. boelastography-guided transfusion algorithm reduces trans-
19. Lisman T, Porte RJ. Rebalanced hemostasis in patients with fusions in complex cardiac surgery. Anesth Analg 1999;88:
liver disease: evidence and clinical consequences. Blood 312-9.
2010;116:878-85. 37. Armstrong S, Fernando R, Ashpole K, et al. Assessment of
R
20. Gale A. Current understanding of hemostasis. Toxicol Pathol coagulation in the obstetric population using ROTEMV
2011;39:273-80. thromboelastometry. Int J Obstet Anesth 2011;20:293-8.
21. Caldwell SH, Hoffman M, Lisman T, et al. Coagulation disor- 38. Tripodi A, Primignani M, Chantarangkul V, et al. The coagul-
ders and hemostasis in liver disease: pathophysiology and opathy of cirrhosis assessed by thromboelastometry and its
critical assessment of current management. Hepatology correlation with conventional coagulation parameters.
2006;44:1039-46. Thromb Res 2009;124:132-6.
22. Wada H, Usui M, Sakuragawa N. Hemostatic abnormalities 39. Weeder PD, Porte RJ, Lisman T. Hemostasis in liver disease:
and liver diseases. Semin Thromb Hemost. 2008;34:772-8. implications of new concepts for perioperative manage-
23. Butenas S, van’t Veer C, Mann KG. “Normal” thrombin gen- ment. Transfus Med Rev 2014;28:107-13.
eration. Blood 1999;94:2169-78. 40. Tsochatzis EA, Senzolo M, Germani G, et al. Systematic
24. Chandler WL, Dawson KL, Ruegger MC, et al. Patients with review: portal vein thrombosis in cirrhosis. Aliment Pharma-
cirrhosis show a relative increase in thrombin generation col Ther 2010;31:366-74.
that is correlated with lower antithrombin levels. Blood Coa- 41. Ogren M, Bergqvist D, Bjorck M, et al. Portal vein thrombo-
gul Fibrinolysis 2014 [Epub ahead of print]. sis: prevalence, patient characteristics and lifetime risk: a
25. Tripodi A, Salerno F, Chantarangkul V, et al. Evidence of nor- population study based on 23796 consecutive autopsies.
mal thrombin generation in cirrhosis despite abnormal con- World J Gastroenterol 2006;12:2115-9.
ventional coagulation tests. Hepatology 2005;41:553-8. 42. Roberts SE, Goldacre MJ, Yeates D. Trends in mortality after
26. Hollestelle MJ, Geertzen HG, Straatsburg IH, et al. Factor hospital admission for liver cirrhosis in an English popula-
VIII expression in liver disease. Thromb Haemost 2004;91: tion from 1968 to 1999. Gut 2005;54:1615-21.
267-75. 43. Garcia-Tsao G, Sanyal AJ, Grace ND, et al. Prevention and
27. Francis JL, Armstrong DJ. Fibrinogen-bound sialic acid levels management of gastroesophageal varices and variceal hem-
in the dysfibrinogenaemia of liver disease. Haemostasis orrhage in cirrhosis. Hepatology 2007;46:922-38.
1982;11:215-22. 44. Garcia-Tsao G, Bosch J. Management of varices and
28. Roberts LN, Patel RK, Arya R. Haemostasis and thrombosis variceal hemorrhage in cirrhosis. N Engl J Med 2010;362:
in liver disease. Br J Haematol 2010;148:507-21. 823-32.

Volume 56, April 2016 TRANSFUSION 797

YATES ET AL.

45. Northup PG, Caldwell SH. Coagulation in liver disease: a mented with normal plasma: considerations on the efficacy
guide for the clinician. Clin Gastroenterol Hepatol 2013;11: of treatment with fresh-frozen plasma. Intern Emerg Med
1064-74. 2012;7:139-44.
46. Villanueva C, Colomo A, Bosch A, et al. Transfusion strat- 62. Yang L, Stanworth S, Hopewell S, et al. Is fresh-frozen
egies for acute upper gastrointestinal bleeding. N Engl J Med plasma clinically effective? An update of a systematic review
2013;368:11-21. of randomized controlled trials (CME). Transfusion 2012;52:
47. Massicotte L, Lenis S, Thibeault L, et al. Effect of low central 1673-86.
venous pressure and phlebotomy on blood product transfu- 63. Benson AB, Burton JR, Austin GL, et al. Differential effects of
sion requirements during liver transplantations. Liver Trans- plasma and red blood cell transfusions on acute lung injury
plant 2006;12:117-23. and infection risk following liver transplantation. Liver
48. Feng ZY, Xu X, Zhu SM, et al. Effects of low central venous Transplant 2011;17:149-58.
pressure during preanhepatic phase on blood loss and liver 64. Shah NL, Intagliata NM, Northup PG, et al. Procoagulant
and renal function in liver transplantation. World J Surg therapeutics in liver disease: a critique and clinical rationale.
2010;34:1864-73. Nat Rev Gastroenterol Hepatol 2014;11:675-82.
49. Massicotte L, Denault AY, Beaulieu D, et al. Transfusion rate 65. Levy JH, Welsby I, Goodnough LT. Fibrinogen as a therapeu-
for 500 consecutive liver transplantations: experience of one tic target for bleeding: a review of critical levels and replace-
liver transplantation center. Transplantation 2012;93:1276- ment therapy. Transfusion 2014;54:1389-405.
81. 66. Levy JH, Goodnough LT. How I use fibrinogen replacement
50. Goulis J, Armonis A, Patch D, et al. Bacterial infection is therapy in acquired bleeding. Blood 2015;125:1387-93.
independently associated with failure to control bleeding in 67. Stainsby D, MacLennan S, Thomas D, et al. Guidelines on
cirrhotic patients with gastrointestinal hemorrhage. Hepato- the management of massive blood loss. Br J Haematol 2006;
logy 1998;27:1207-12. 135:634-41.
51. Vivas S, Rodriguez M, Palacio MA, et al. Presence of bacterial 68. American Society of Anesthesiologists Task Force on Periop-
infection in bleeding cirrhotic patients is independently erative Blood Management. Practice guidelines for perioper-
associated with early mortality and failure to control bleed- ative blood management: an updated report by the
ing. Dig Dis Sci 2001;46:2752-7. American Society of Anesthesiologists Task Force on Periop-
52. Borzio M, Salerno F, Piantoni L, et al. Bacterial infection in erative Blood Management*. Anesthesiology 2015;122:241-
patients with advanced cirrhosis: a multicentre prospective 75.
study. Dig Liver Dis 2001;33:41-8. 69. Franchini M, Lippi G. Fibrinogen replacement therapy: a
53. Jalal DI, Chonchol M, Targher G. Disorders of hemostasis critical review of the literature. Blood Transfus 2012;10:23-7.
associated with chronic kidney disease. Semin Thromb 70. Carson JL, Grossman BJ, Kleinman S, et al. Red blood cell
Hemost 2010;36:34-40. transfusion: a clinical practice guideline from the AABB*.
54. Pavord S, Myers B. Bleeding and thrombotic complications Ann Intern Med Am 2012;157:49-58.
of kidney disease. Blood Rev 2011;25:271-8. 71. Sarode R, Milling TJ, Refaai MA, et al. Efficacy and safety of a
55. Rockey DC, Caldwell SH, Goodman ZD, et al. Liver biopsy. 4-factor prothrombin complex concentrate in patients on
Hepatology 2009;49:1017-44. vitamin K antagonists presenting with major bleeding: a
56. Aster RH. Pooling of platelets in the spleen: role in the randomized, plasma-controlled, phase IIIb study. Circula-
pathogenesis of “hypersplenic” thrombocytopenia. J Clin tion 2013;128:1234-43.
Invest 1966;45:645-57. 72. Lorenz R, Kienast J, Otto U, et al. Efficacy and safety of a
57. Afdhal NH, Giannini EG, Tayyab G, et al. Eltrombopag before prothrombin complex concentrate with two virus-
procedures in patients with cirrhosis and thrombocytopenia. inactivation steps in patients with severe liver damage. Eur J
N Engl J Med 2012;367:716-24. Gastroenterol Hepatol 2003;15:15-20.
58. Tripodi A, Primignani M, Chantarangkul V, et al. Global 73. Cederbaum AI, Blatt PM, Roberts HR. Intravascular coagula-
hemostasis tests in patients with cirrhosis before and after tion with use of human prothrombin complex concentrates.
prophylactic platelet transfusion. Liver Int 2013;33:362-7. Ann Intern Med 1976;84:683-7.
59. Wallace MJ, Narvios A, Lichtiger B, et al. Transjugular liver 74. Bosch J, Thabut D, Albillos A, et al. Recombinant factor
biopsy in patients with hematologic malignancy and severe VIIa for variceal bleeding in patients with advanced cir-
thrombocytopenia. J Vasc Interv Radiol 2003;14:323-7. rhosis: a randomized, controlled trial. Hepatology 2008;47:
60. Wells AW, Llewelyn CA, Casbard A, et al. The EASTR study: 1604-14.
indications for transfusion and estimates of transfusion 75. Munoz SJ, Stravitz RT, Gabriel DA. Coagulopathy of acute
recipient numbers in hospitals supplied by the National liver failure. Clin Liver Dis 2009;13:95-107.
Blood Service. Transfus Med 2009;19:315-28. 76. Munoz SJ, Reddy KR, Lee W, et al. The coagulopathy of acute
61. Tripodi A, Chantarangkul V, Primignani M, et al. Thrombin liver failure and implications for intracranial pressure moni-
generation in plasma from patients with cirrhosis supple- toring. Neurocrit Care 2008;9:103-7.

798 TRANSFUSION Volume 56, April 2016