Original Research

Family history of postpartum hemorrhage is a risk

factor for postpartum hemorrhage after vaginal
delivery: results from the French prospective
multicenter Haemorrhages and Thromboembolic
Venous Disease of the Postpartum cohort study
François Anouilh, MS, CM; Claire de Moreuil, MD, PhD; Christophe Tre
mouilhac, MD;
Matthieu Jacquot, MD; Gilles Salnelle, MD; Violaine Bellec, MD; Nathalie Touffet, MD;
Caroline Cornec, CM; Matthieu Muller, MD; Pierre-François Dupre
, MD; Charles Bellot, MD;
Karine Morcel, MD, PhD; Delphine Le Joliff, MD; Guillaume Drugmanne; Elodie Gelebart; Sandy Lucier;
Emmanuel Nowak, PhD; Line Bihan, MS; Francis Couturaud, MD, PhD; Ce
cile Tromeur, MD, PhD;
Emmanuelle Le Moigne, MD, PhD; Brigitte Pan-Petesch, MD

BACKGROUND: Postpartum hemorrhage is a major component of −2.14) and a personal transfusion history (adjusted odds ratio, 1.90;
perinatal morbidity and mortality that affects young women worldwide and 95% confidence interval, 1.23−2.92) were significantly associated with
is still often unpredictable. Reducing the incidence of postpartum hemor- postpartum hemorrhage. The use of oxytocin during labor was also a risk
rhage is a major health issue and identifying women at risk for postpartum factor for postpartum hemorrhage (adjusted odds ratio, 1.24; 95% confi-
hemorrhage is a key element in preventing this complication. dence interval, 1.06−1.44). Inversely, smoking during pregnancy and
OBJECTIVE: This study aimed to estimate postpartum hemorrhage intrauterine growth restriction were associated with a reduced risk for
prevalence after vaginal delivery and to identify postpartum hemorrhage postpartum hemorrhage (adjusted odds ratio, 0.76; 95% confidence inter-
risk factors. val, 0.63−0.91, and 0.34; 95% confidence interval, 0.13−0.87,
STUDY DESIGN: Unselected pregnant women ≥16 years of age respectively).
admitted to 1 of 6 maternity wards in Brittany (France) for vaginal birth CONCLUSION: In addition to classical risk factors, this study identified
after 15 weeks of gestation were recruited in this prospective, multicenter a family history of postpartum hemorrhage and personal transfusion his-
cohort study between June 1, 2015, and January 31, 2019. Postpartum tory as new characteristics associated with postpartum hemorrhage after
hemorrhage was defined as blood loss ≥500 mL in the 24 hours following vaginal delivery. The association of postpartum hemorrhage with a family
delivery. Independent risk factors for postpartum hemorrhage were deter- history of postpartum hemorrhage suggests a hereditary hemorrhagic phe-
mined using logistic regression. Missing data were imputed using the Mul- notype and calls for genetic studies. Identifying women at risk for postpar-
tivariate Imputation by Chained Equations method. tum hemorrhage is a key element of being prepared for this complication.
RESULTS: Among 16,382 included women, the postpartum hemor-
rhage prevalence was 5.37%. A first-degree family history of postpartum Key words: family history, postpartum hemorrhage, prospective cohort,
hemorrhage (adjusted odds ratio, 1.63; 95% confidence interval, 1.24 risk factors, vaginal delivery

Introduction women worldwide.1,2 PPH is commonly prolonged labor, and chorioamniotitis.3

P ostpartum hemorrhage (PPH) is a

major component of severe mater-
nal morbidity and mortality, affecting
defined as a blood loss from the genital
tract of ≥500 mL regardless of the deliv-
ery mode.1,3,4 This complication occurs
A personal history of PPH is also strongly
associated with PPH.7 Other risk factors
related to medical practice, such as the
in about 5% of births in the absence of administration of oxytocin during labor
accurate loss measurement and in or delivery by cesarean delivery, have also
Cite this article as: Anouilh F, de Moreuil C, Tr
emouil-
around 10% if the blood loss is objec- been identified in numerous studies.3,8,9
hac C, et al. Family history of postpartum hemorrhage is
a risk factor for postpartum hemorrhage after vaginal tively measured.4 Women experiencing Nevertheless, recent prospective studies
delivery: results from the French prospective multicenter PPH are prone to blood transfusion, on the topic are scarce. Therefore, there is
Haemorrhages and Thromboembolic Venous Disease of hemorrhagic shock, infertility second- a need for large and well-conducted epi-
the Postpartum cohort study. Am J Obstet Gynecol MFM ary to hysterectomy, and postpartum demiologic studies to update PPH preva-
2023;5:101062.
depression.4,5 PPH can also affect lence and its risk factors.
2589-9333/$36.00 breastfeeding behavior and the maternal Because of its significant impact on
© 2023 The Author(s). Published by Elsevier Inc. This is care of the newborn.6 maternal health, reducing the incidence
an open access article under the CC BY-NC-ND license Uterine atony, the most common cause of PPH is a major health issue. Identify-
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
of PPH, is associated with high multipar- ing women at risk for PPH is a key ele-
http://dx.doi.org/10.1016/j.ajogmf.2023.101062
ity, multiple pregnancy, hydramnios, ment in preventing this complication,10

September 2023 AJOG MFM 1

Original Research

AJOG MFM at a Glance relatives, and personal medical, surgical,

and obstetrical history) and all data on
Why was this study conducted? induction of pregnancy and pathologies
This study aimed to evaluate the prevalence of postpartum hemorrhage (PPH) and medications used during preg-
after vaginal delivery in a large prospective cohort of unselected pregnant nancy, delivery, and the postpartum
women who delivered at 6 French maternity wards in the Brittany district and period. Continuous variables (eg, age,
to identify factors independently associated with PPH, including family history. body mass index [BMI] before preg-
Key findings nancy, term at delivery) were catego-
In this French prospective study, the prevalence of primary PPH was 5.37%. rized using standard clinical definitions.
Fourteen independent risk factors for PPH were identified, including a family Categorical variables (eg, parity) were
history of PPH in first-degree relatives and a history of transfusion. redefined according to clinical rele-
vance.
What does this add to what is known? Uterine pathology included uterine
The association between PPH and a family history of PPH suggests an inherited malformations, myomas, and fibroids.
hemorrhagic phenotype and paves the way for further genetic studies. Intrauterine growth restriction (IUGR)
was defined according to the 2013
French College of Obstetricians and
Gynecologists guidelines as an esti-
especially to refer women to appropriate This study did not include women
mated fetal weight below the 10th per-
antenatal facilities and to adapt resour- who gave birth at home without subse-
centile using a locally accepted curve
ces and medical practices. quent support in a hospital. Women for
(from the Association des Utilisateurs
Vaginal delivery is the most common whom blood loss during delivery were
de Dossiers Informatis
es en P
ediatrie,
mode of delivery in developed coun- not objectively measured or recorded
Obst
etrique et Gyn
ecologie study) asso-
tries.11−13 However, there is no effective were also excluded. For this report, the
ciated with clinical signs of pathology.14
preventive strategy based on the identi- population was restricted to vaginal
Abnormal placental insertion included
fication of PPH risk factors for women deliveries.
low-lying placentas and placenta previa.
delivering vaginally. We hypothesized
Macrosomia was defined as a newborn
that features not yet studied, such as Main outcome weight ≥4000 grams.
family history, may also influence the The primary outcome was the occur- Data on all medications used during
occurrence of PPH and would be rence of primary PPH, defined as blood pregnancy were collected, but only 3
important for PPH risk assessment. loss ≥500 mL within the first 24 hours therapeutic classes were studied here,
This study aimed to evaluate the following delivery. Measurement of namely anticoagulant, antiaggregant,
prevalence of PPH after vaginal delivery blood loss was systematically performed and antidepressant.
in a large prospective cohort of unse- with a graduated collector bag routinely Anemia was defined as any record of
lected pregnant women who delivered used in the recruiting centers, and its a predelivery hemoglobin concentration
in 6 French maternity wards in the Brit- volume was systematically reported in of <11 g/dL during the first and third
tany district and to identify factors the medical records. If the bleeding trimesters or <10.5 g/dL during the sec-
independently associated with PPH, occurred when the woman was in the ond trimester of pregnancy. Thrombo-
including family history. maternity ward, the losses were quanti- cytopenia was defined as a predelivery
fied by weighing the sanitary pads. platelet count <150 £ 109/L.
Material and Methods The secondary outcome was the The assigned level of the maternity
Participants and study population identification of factors associated with ward was according to a French classifi-
The Hemorrhages and Thromboem- PPH. cation system of maternity wards based
bolic Venous Disease of the Postpartum on neonatal and pediatric resuscitation
(HEMOTHEPP) study is a multicenter, Variables resources.15 Level 1 corresponds to a
prospective cohort study conducted in 6 Clinical data were collected by midwives ward without facilities for nonroutine
maternity wards in Brittany (Brest Uni- and obstetricians during antenatal care neonatal care, level 2 maternity hospi-
versity Hospital, Keraudren Private visits, labor, delivery, and the postpar- tals have a neonatal care unit, and level
Polyclinic [Brest], Landerneau Hospital, tum period in the obstetric files. These 3 centers have an onsite neonatal inten-
Morlaix Hospital, Quimper Hospital, data were collected through a personal, sive care and resuscitation unit.
Carhaix Hospital) between June 1, structured interview by trained clinical
2015, and January 31, 2019. All women research assistants after delivery. Col- Details of patient’s consent and
≥16 years of age who were admitted for lected variables included maternal data ethics approval
delivery after 15 weeks of gestation (sociodemographic characteristics, fam- All pregnant women approached in 1 of
(WG), were asked to participate in this ily history of PPH and of venous throm- the 6 maternity wards participating in
study. boembolism [VTE] among first-degree the HEMOTHEPP study were informed

2 AJOG MFM September 2023

 Original Research

FIGURE Statistical analysis

Flow chart of the study Descriptive statistics were used to sum-
marize the women’s characteristics.
Continuous variables were presented as
the median with interquartile range
(IQR) and categorical variables as
counts and proportion.
Missing data were replaced using the
Multivariate Imputation by Chained
Equations (MICE) method for which 10
iterations were completed and 20 data
sets were imputed.16−18
Univariable logistic regression analy-
sis was used to assess the association
between PPH and each potential risk
factor. Unadjusted odds ratios (ORs)
were calculated and reported with their
95% confidence intervals (CIs). In addi-
tion to univariable logistic regression, 4
clinically relevant interactions were
tested, namely transfusion history and
personal history of PPH, parity and
oxytocin use during labor, instrumental
delivery and type of perineal injuries,
and IUGR and smoking during preg-
nancy. To determine independent pre-
dictors of PPH, a multivariable analysis
HEMOTHEPP, Hemorrhages and Thromboembolic Venous Disease of the Postpartum; PPH, postpartum hemorrhage.
was performed. Informative variables
Anouilh. Family history of postpartum hemorrhage is a risk factor for postpartum hemorrhage. Am J Obstet Gynecol
MFM 2023.
with a P value <.15 in the univariable
analysis were included in a stepwise
backward elimination logistic regression
model. Adjusted ORs (aORs) and their
95% CIs were calculated with an adjust-
about the study by midwives or obstetri- collected for the purpose of the study ment at the maternity ward level (level
cians during one of their medical visits without any blood sampling. For each 1 to 3).
at the hospital and had to give their oral recruiting center, a record of the con- Furthermore, a sensitivity analysis
consent to participate in the study. Each sent process was held by the investiga- was performed, corresponding to the
of them received a written information tors of the study for each woman who univariable and multivariable analyses
note (included in the supplemental delivered during the inclusion period of without imputation of missing data.
material) about the aims of the study, the study with details on the date of oral Statistical significance was considered
the conduct of the study, the optional consent, the date of inclusion in the with a bilateral alpha risk of 5%
blood sampling, the way it would be study, and the possibility of withdrawal. (P<.05). All statistical tests were per-
used for research purposes, especially For the women who were not included formed using R, version 4.1.2, and R
genetic testing, and the possibility to get in the study, information about the rea- Studio, version 1.4.1106 (R Core Team,
the results of the genetic testing if so sons for noninclusion was recorded too. Vienna, Austria).
wished. Information about the confi- A patient’s permission for publica-
dentiality and the use of collected medi- tion was obtained through the oral con- Results
cal data was also detailed. Written sent to participate in the study in Characteristics of the population
consent was necessary only for the compliance with the Good Clinical Among the 22,033 women who gave
blood collection intended for genetic Practice protocol and the Declaration of birth in 1 of the 6 maternity wards dur-
analyses. Because the blood sampling at Helsinki principles. The study protocol ing the study period, 20,238 (91.9%)
entry in the delivery room was not was approved by the CCTIRS and CCP women were included in the HEMO-
mandatory, a woman could consent Ouest. The CNIL gave its authorization THEPP study. Six women were
orally to participate in the study and on April 30, 2015, and the study was excluded because of a lack of informa-
have her clinical and biologic data registered on ClinicalTrials.gov. tion on PPH. From the 20,232 women

September 2023 AJOG MFM 3

Original Research

in the study population, 16,382 (81.0%)

TABLE 1
women with a vaginal delivery were
Baseline characteristics of the study population
included in this report (Figure).
Study population Vaginal deliveries Table 1 shows the characteristics of
Variables n=20,232 (100%) n=16,382 (81.0%) the women included in this study.
General characteristics The median (IQR) maternal age was
30.6 (27.4−34.0) years, and the median
Age (y)
(IQR) BMI was 23.0 (20.0−26.0) kg/m2.
<20 268 (1.3%) 226 (1.4%) Of the included women, 3187 (19.0%)
20−34 15,731 (78%) 12,969 (79%) women were ≥35 years of age, and 1794
≥35 4233 (21%) 3187 (19%) (11.0%) women had obesity. In addi-
tion, 6770 (41.0%) women were primip-
BMI before pregnancy
arous, whereas 865 (5.3%) women had
<18.5 1286 (6.6%) 1122 (7.1%) already had ≥3 children. A history of
18.50−24.99 11,361 (58%) 9576 (60%) cesarean delivery was present in 897
≥25 4357 (22%) 3380 (21%) (5.5%) women, and 481 (3.0%) had a
personal history of PPH.
≥30 2587 (13%) 1794 (11%)
A family history of PPH in first-
Missing 641 (3.2%) 510 (3.1%) degree relatives was found in 802
Blood group O 8647 (45%) 7037 (45%) (5.4%) women, and 1167 (7.7%) women
Missing 897 (4.4%) 708 (4.3%)
had a family history of VTE in first-
degree relatives. Among the women
Non-European origin 1371 (7.0%) 1021 (6.5%) with a personal history of transfusion,
Missing 765 (3.8%) 605 (3.7%) 83 (31.8%) had a personal history of
Smoking during pregnancy 4736 (25%) 3828 (25%) PPH.
Conception was assisted in 669
Missing 1232 (6.1%) 995 (6.1%)
(4.2%) women, and multiple pregnancy
Personal and family history occurred in 191 (1.2%) women. Of note,
PPH family history 961 (5.2%) 802 (5.4%) 3828 (25.0%) women smoked during
Missing 1819 (9.0%) 1449 (8.8%) pregnancy. Gestational diabetes mellitus
and antepartum bleeding were observed
VTE family history 1436 (7.7%) 1167 (7.7%)
in 11.0% and 7.3% of pregnant women,
Missing 1489 (7.4%) 1188 (7.3%) respectively.
VTE personal history 207 (1.1%) 163 (1.0%) Induction of labor occurred in 24.0%
Missing 589 (2.9%) 459 (2.8%) of cases. During labor, 14,014 (87.0%)
women made use of epidural anesthesia,
Hemorrhagic disease 42 (0.2%) 28 (0.2%)
and 2889 (18.0%) women had an instru-
Missing 642 (3.2%) 504 (3.1%) mental delivery. Finally, 1242 (7.6%)
Transfusion history 360 (1.8%) 261 (1.6%) newborns had macrosomia.
Missing 675 (3.3%) 525 (3.2%)
Prevalence of postpartum
Chronic hypertension 145 (0.7%) 101 (0.6%)
hemorrhage
Missing 639 (3.2%) 502 (3.1%) Among 16,382 vaginal deliveries, 880
Diabetes 71 (0.4%) 41 (0.3%) were complicated by PPH, that is, an
Missing 641 (3.2%) 502 (3.1%) overall PPH prevalence of 5.37% (95%
CI, 5.03−5.73).
Dysthyroidism 539 (2.8%) 412 (2.6%)
Missing 641 (3.2%) 502 (3.1%) Risk factors for postpartum
Uterine pathology 63 (0.3%) 40 (0.3%) hemorrhage in the univariable
Missing 654 (3.2%) 507 (3.1%) analysis
The results of the univariable analysis,
Obstetrical history
after imputation of missing data, are
Cesarean delivery history 2101 (11%) 897 (5.5%) presented in Table 2. The percentage of
Missing 224 (1.1%) 186 (1.1%) missing data ranged from 0% to 8.8%,
except for labor duration that reached
(continued)
25.7%. A large number of variables that

4 AJOG MFM September 2023

 Original Research

TABLE 1 were studied as potential risk factors for

Baseline characteristics of the study population (continued) PPH were associated with PPH. Of
note, among the original variables col-
Study population Vaginal deliveries
Variables n=20,232 (100%) n=16,382 (81.0%)
lected in this study, neither a personal
history of VTE, nor medications of
PPH history 623 (3.1%) 481 (3.0%) interest were associated with PPH in
Missing 390 (1.9%) 311 (1.9%) the univariable analysis. Moreover,
Parity none of the interactions tested were sig-
nificant. Therefore, none were included
0 8,645 (43%) 6,770 (41%)
in the multivariable analysis.
1−2 10,415 (52%) 8701 (53%)
≥3 1116 (5.5%) 865 (5.3%) Risk factors for postpartum
Missing 56 (0.3%) 46 (0.3%) hemorrhage in the multivariable
analysis
Ongoing pregnancy
The results of the multivariable analysis
Assisted conception 978 (5.0%) 669 (4.2%) adjusted for the maternity ward level
Missing 658 (3.3%) 529 (3.2%) are presented in Table 3.
Multiple pregnancy 433 (2.1%) 191 (1.2%) A total of 16 variables were signifi-
cantly associated with PPH, with 14
Missing 16 (0.1%) 12 (0.1%)
being identified as risk factors for PPH
Medication during pregnancy and 2 being protective against PPH.
Anticoagulant 92 (0.5%) 54 (0.3%) Independent risk factors for PPH
Antiaggregant 107 (0.5%) 74 (0.5%) previously identified in other studies
and confirmed in this study were BMI
Antidepressant 149 (0.7%) 117 (0.7%)
class (overweight and obesity), non-
Pregnancy complications European origin, personal history of
Anemia 477 (2.5%) 381 (2.5%) PPH, nulliparity, multiple pregnancy,
Missing 1077 (5.3%) 884 (5.4%) abnormal placental insertion (previa
and low-lying placenta), preeclampsia,
Antepartum bleeding 1453 (7.6%) 1125 (7.3%) term at delivery >41 WG, use of oxyto-
Missing 1075 (5.3%) 881 (5.4%) cin during labor, retained placenta, per-
Premature delivery threat 831 (4.3%) 644 (4.2%) ineal injuries (episiotomy and/or
Missing 1075 (5.3%) 881 (5.4%)
perineal tears), and macrosomia.
Independent risk factors for PPH
IUGR 425 (2.2%) 245 (1.6%) specifically identified in this study were
Missing 1075 (5.3%) 881 (5.4%) a first-degree family history of PPH
Gestational hypertension 341 (1.8%) 341 (1.8%) (aOR, 1.63; 95% CI, 1.24−2.14) and a
personal transfusion history (aOR, 1.90;
Missing 1075 (5.3%) 881 (5.4%)
95% CI, 1.23−2.92).
Gestational diabetes mellitus 2259 (12%) 1668 (11%) In contrast, smoking during preg-
Missing 1075 (5.3%) 881 (5.4%) nancy and IUGR were associated with a
Hydramnios 97 (0.5%) 63 (0.4%) reduced risk for PPH (aOR, 0.76; 95%
CI, 0.63−0.91 and aOR, 0.34; 95% CI,
Missing 1075 (5.3%) 881 (5.4%)
0.13−0.87, respectively).
Previa or low-lying placenta 318 (1.7%) 193 (1.2%)
Missing 1075 (5.3%) 881 (5.4%) Sensitivity analysis
Preeclampsia 332 (1.7%) 157 (1.0%) The sensitivity analysis carried out on
the initial data without imputation of
Missing 1075 (5.3%) 881 (5.4%)
missing data showed similar results in
Placental abruption 175 (0.9%) 114 (0.7%) the univariable analysis (Supplemental
Missing 1075 (5.3%) 881 (5.4%) Table 1). The multivariable analysis,
Thrombocytopenia 546 (2.9%) 438 (2.8%) performed with data from 10,303
women (62.9%), is depicted in Supple-
Missing 1075 (5.3%) 881 (5.4%)
mental Table 2. The results of this anal-
(continued) ysis were also similar to those of the
main analysis except for 3 variables

September 2023 AJOG MFM 5

Original Research

TABLE 1 (overweight, non-European origin, and

Baseline characteristics of the study population (continued) premature delivery threat).
Study population Vaginal deliveries
Variables n=20,232 (100%) n=16,382 (81.0%)
Discussion
Principal findings
Labor and delivery In this French prospective study of
Term more than 16,000 women who gave
<37 WG 1390 (6.9%) 854 (5.2%) birth vaginally, the prevalence of pri-
mary PPH was 5.37%.
37−41 WG 14,545 (72%) 12,002 (73%)
Fourteen independent risk factors for
>41 WG 4297 (21%) 3526 (22%) PPH were identified. Interestingly, a
Induction of labor 4852 (24%) 3922 (24%) family history of PPH in first-degree
Missing 142 (0.7%) 101 (0.6%) relatives and a transfusion history were
identified as risk factors for PPH,
Epidural anesthesia 15,686 (78%) 14,014 (87%)
whereas other known risk factors, such
Missing 240 (1.2%) 214 (1.3%) as macrosomia, were confirmed. Con-
Labor duration (h) versely, smoking during pregnancy and
<4 6275 (47%) 5899 (48%) IUGR were associated with a reduced
risk for PPH. Of note, a family and per-
4 to <8 3651 (27%) 3435 (28%)
sonal history of VTE and the studied
8 to <12 2710 (20%) 2342 (19%) medications were not associated with
12 to <16 571 (4.3%) 432 (3.5%) PPH.
≥16 85 (0.6%) 64 (0.5%)
Results
Missing 6940 (34.3%) 4210 (25.7%)
Comparing our results with those of
Oxytocin during labor 7709 (39%) 6226 (39%) previous published studies may be lim-
Missing 648 (3.2%) 495 (3.0%) ited by the heterogeneity in the defini-
Instrumental delivery − 2889 (18%) tion of PPH and in the inclusion
criteria, especially the delivery mode. In
Missing − 0 our study, the prevalence of PPH after
Prophylactic administration of oxytocin 17,251 (88%) 15,466 (97%) vaginal delivery was 5.37%. In another
Missing 610 (3.0%) 490 (3.0%) large-scale, French prospective cohort
Retained placenta − 1798 (11%)
study conducted between 2004 and
2006 that included all modes of delivery,
Missing − 557 (3.4%) PPH occurred in 4.8% of deliveries
Type of perineal injuries when the outcome was clinically
None − 5752 (36%) defined.19
The main characteristics of pregnant
Episiotomy without tearing − 1969 (12%)
women included in our study, namely
Episiotomy and tear − 405 (2.6%) age, BMI, parity, cesarean delivery his-
Tear without episiotomy − 7732 (49%) tory, and proportion of multiple preg-
Missing − 524 (3.2%) nancies, were comparable with that in
the French population as described in
Macrosomia 1645 (8.2%) 1242 (7.6%)
the 2016 National Perinatal Survey.20
Missing 124 (0.6%) 108 (0.7%) In addition, we highlighted that a
Maternity level family history of PPH in first-degree
1 2060 (10%) 1659 (10%) relatives was a risk factor for PPH after
vaginal delivery (aOR, 1.63). Only 2
2 12,460 (62%) 10,276 (63%)
other studies have investigated this
3 5712 (28%) 4447 (27%) parameter as a potential risk factor for
Maternity level: 1, no facilities for nonroutine neonatal care; 2, with a neonatal care unit; 3, with an onsite neonatal intensive care PPH. These studies, carried out retro-
and resuscitation unit.
spectively, showed contradictory results.
BMI, body mass index; IUGR, intrauterine growth restriction; PPH, postpartum hemorrhage; VTE, venous thromboembolism; WG,
weeks of gestation. The Swedish register-based cohort
Anouilh. Family history of postpartum hemorrhage is a risk factor for postpartum hemorrhage. Am J Obstet Gyne- study attributed 18% of excessive PPH
col MFM 2023. to maternal genetic factors among
women who gave birth vaginally.21

6 AJOG MFM September 2023

 Original Research

Conversely, the study conducted on 2

TABLE 2
register-based Scottish cohorts, evaluat-
Risk factors for postpartum hemorrhage in univariable analysis after impu-
ing 3 generations of women, did not
tation of missing data
find any increase in PPH risk among
Variables OR (95% CI) P valuea women whose mother or grandmother
General characteristics
had PPH.22 Interestingly, an Italian
genetic prospective monocenter cohort
Age (y) .59 study focusing on vaginal deliveries
<20 1.24 (0.73−2.11) reported that a polymorphism in the
20−34 1 promoter of tissue factor was protective
against PPH (OR, 0.79; 95% CI, 0.69
≥35 0.95 (0.78−1.13)
−0.90), supporting the hypothesis of a
BMI before pregnancy <.001 genetic involvement in PPH risk.23
<18.5 0.80 (0.59−1.10) We also found that a personal blood
18.50−24.99 1 transfusion history was significantly
associated with PPH. However, 68.2% of
≥25 1.25 (1.05−1.48)
these women did not report any previous
≥30 1.64 (1.35−1.99) PPH. It would be interesting to have fur-
Blood group O 1.05 (0.91−1.21) .50 ther information on transfusion circum-
Non-European origin 1.38 (1.08−1.76) .011 stances (spontaneous or provoked
hemorrhage, surgery, site of bleeding,
Smoking during pregnancy 0.72 (0.60−0.86) <.001
etc.). Moreover, a personal blood trans-
Personal and family history fusion history could be associated with
PPH family history 1.74 (1.36−2.24) <.001 an undiagnosed hereditary bleeding dis-
VTE family history 1.01 (0.77−1.31) .97 order such as Von Willebrand disease, a
frequent medical condition. Indeed, in
VTE personal history 1.13 (0.59−2.15) .71
our study, hemorrhagic diseases were
Hemorrhagic disease 0.67 (0.09−4.94) .69 rare, thus probably underreported and
Transfusion history 2.68 (1.86−3.86) <.001 underdiagnosed.
Chronic hypertension 2.14 (1.14−4.02) .018 Of note, the usual PPH risk factors
were confirmed in this study. Multiple
Diabetes 1.36 (0.42−4.41) .61
pregnancy and macrosomia by increasing
Dysthyroidism 1.44 (1.02−2.15) .054 uterine distension were significantly asso-
Uterine pathology 2.42 (0.95−6.19) .065 ciated with PPH, as in many previous
Obstetrical history studies, regardless of the mode of
delivery7,24,25 but specifically after vaginal
Cesarean delivery history 1.21 (0.92−1.59) .18
delivery.8,26,27 A clinical study carried out
PPH history 3.14 (2.41−4.09) <.001 in France and a systematic review and
Parity <.001 meta-analysis that included all modes of
0 1.38 (1.20−1.59)
delivery have found an up to 3-fold
increased risk for PPH among women
1−2 1 with PPH history.7,19 However, a PPH
≥3 1.02 (0.73−1.41) history is not prevalent among pregnant
Ongoing pregnancy women (only 5% of nonprimiparous in
our population). Primiparity was also sig-
Assisted conception 1.60 (1.20−2.13) .001
nificantly associated with PPH (aOR,
Multiple pregnancy 3.64 (2.47−5.35) <.001 1.39). Primiparity is now identified as a
Medication during pregnancy risk factor for PPH in high-resource
Anticoagulant 1.04 (0.33−3.43) .95 countries, instead of high multiparity as
in the past.3 Regarding pregnancy pathol-
Antiaggregant 0.49 (0.12−1.99) .32
ogies, preeclampsia was identified as a
Antidepressant 1.12 (0.52−2.42) .77 risk factor for PPH (aOR, 2.43). This
Pregnancy complications association was also reported in a Brittany
Anemia 0.90 (0.56−1.44) .66 district and a Norwegian cohort.26,28 Pla-
centa previa and/or low-lying placentas
(continued) were also risk factors for PPH in this

September 2023 AJOG MFM 7

Original Research

TABLE 2 study, which is similar to other

Risk factors for postpartum hemorrhage in univariable analysis after impu- reports.12,29,30 Finally, placental retention
tation of missing data (continued) was strongly associated with PPH in our
study (aOR, 8.49) and was already
Variables OR (95% CI) P valuea
reported in studies focusing on vaginal
Antepartum bleeding 1.40 (1.10−1.76) .005 deliveries.8,31 However, the identification
Premature delivery threat 1.30 (0.95−1.77) .10 of this major risk factor is not very useful
IUGR 0.37 (0.15−0.92) .033 for prevention because placental retention
occurs late during delivery.3
Gestational hypertension 1.64 (1.03−2.61) .036
In contrast, smoking during preg-
Gestational diabetes 1.29 (1.06−1.59) .013 nancy and IUGR were associated with a
mellitus reduced risk for PPH. The effect of
Hydramnios 0.86 (0.25−2.77) .80 smoking varies among previous pub-
Previa or low-lying placenta 2.22 (1.42−3.48) <.001 lished studies. Indeed, in a retrospective
cohort study, smoking was a risk factor
Preeclampsia 2.77 (1.75−4.38) <.001
for PPH,32 whereas others identified
Placental abruption 1.43 (0.72−2.85) .30 smoking to be a protective factor
Thrombocytopenia 1.20 (0.82−1.77) .34 against undiagnosed postpartum blood
Labor and delivery loss and PPH.8,33 A protective effect
was also reported in a systematic review
Term <.001
by Feduniv et al in 2020.30 A search for
<37 WG 1.22 (0.90−1.64) possible confounding factors would
37−41 WG 1 deserve further investigation. However,
>41 WG 1.48 (1.27−1.72) smoking should not be recommended
during pregnancy because of the many
Induction of labor 1.43 (1.23−1.66) <.001
other risks associated with this habit.
Epidural anesthesia 0.90 (0.75−1.11) .27 An association between IUGR and a
Labor duration (h) <.001 reduced risk for PPH was also reported.
<4 0.85 (0.69−1.04) One possible explanation for this find-
ing is that the placentas of fetuses with
4 to <8 1
IUGR are usually smaller than those of
8 to <12 1.23 (1.02−1.49) eutrophic fetuses, thereby reducing the
12 to <16 1.58 (1.10−2.26) insertion area of the placenta and
≥16 1.95 (0.89−4.27) exposing a smaller surface of the uterine
lining to hemorrhage.
Oxytocin during labor 1.56 (1.36−1.79) <.001
Instrumental delivery 1.72 (1.47−2.01) <.001 Clinical and research implications
Prophylactic administration 0.50 (0.36−0.69) <.001 Some characteristics related to labor and
of oxytocin delivery management are interesting to
Retained placenta 8.60 (7.43−9.95) <.00 recognize because they are modifiable.3
Type of perineal injuries <.001 First, we confirmed that the use of oxyto-
cin during labor is a risk factor for PPH.
None 1
This practice has been identified to
Episiotomy without tearing 1.93 (1.56−2.39) increase the risk for PPH and severe
Episiotomy and tear 3.68 (2.70−5.02) PPH, and a dose-dependent relationship
Tear without episiotomy 1.32 (1.12−1.56) has been demonstrated.3,28,34,35 Along
with this, a large, retrospective American
Macrosomia 1.77 (1.43−2.19) <.001
cross-sectional study described that oxy-
BMI, body mass index; CI, confidence interval; IUGR, intrauterine growth restriction; OR, odds ratio; PPH, postpartum hemor-
rhage; VTE, venous thromboembolism; WG, weeks of gestation
tocin administration increased women’s
P value of univariable logistic regression.
a OR for PPH when administered for lon-
Anouilh. Family history of postpartum hemorrhage is a risk factor for postpartum hemorrhage. Am J Obstet Gyne-
ger than 4 hours during spontaneous
col MFM 2023. labor and for longer than 7 hours for
labor induction.34 Indeed, its use during

8 AJOG MFM September 2023

 Original Research

labor can desensitize myometrium recep-

TABLE 3
tors, thereby reducing the postdelivery
Risk factors for PPH in multivariable analysis, after imputation of missing
effects of oxytocin on uterine contractil-
data and adjustment on maternity level
ity.36 In addition, a genetic predisposition
Variables Adjusted OR (95% CI) P valuea in the oxytocin receptor gene could be
General characteristics
involved in the susceptibility to atonic
PPH.37 Questions remain about the inter-
BMI before pregnancy action between pathologies leading to the
<18.5 0.82 (0.59−1.13) .26 use of oxytocin and favoring PPH. Sec-
18.50−24.99 1 ond, prophylactic use of oxytocin in the
immediate postpartum period to promote
≥25 1.25 (1.04−1.49) .017
uterine tone (in line with international
≥30 1.71 (1.39−2.11) <.001 guidelines14,38,39) was not found to be
Non-European origin 1.35 (1.03−1.76) .030 protective. However, our analysis was
Smoking during pregnancy 0.76 (0.63−0.91) .004 limited by the widespread use of prophy-
lactic oxytocin in 95% of the patients in
Personal and family history
our cohort, preventing the comparison
PPH family history 1.63 (1.24−2.14) <.001 with women who did not receive it.
Transfusion history 1.90 (1.23−2.92) .004 Third, an episiotomy was found to be an
Obstetrical history independent risk factor for PPH, particu-
larly when associated with perineal tears
PPH history 2.51 (1.83−3.45) <.001
(aOR, 3.02). This association has already
Parity been reported many times13,26,31 and
0 1.39 (1.18−1.64) <.001 encourages professionals to use this prac-
1−2 1 tice sparingly.
≥3 1.04 (0.73−1.49) .81
Strengths and limitations
Ongoing pregnancy Our study has several strengths. First,
Multiple pregnancy 2.10 (1.34−3.30) .001 this was a prospective and multicenter
Pregnancy complications study design. The collection of data
from the medical file was supplemented
IUGR 0.34 (0.13−0.87) .025
by a direct interview in the maternity
Previa or low-lying placenta 1.76 (1.07−2.87) .027 room. Second, the multicenter design
Preeclampsia 2.43 (1.46−4.03) <.001 included all centers in a geographical
Labor and delivery area and the recruitment was exhaustive
and performed in a nonselective man-
Term
ner. This enabled us to represent a vari-
<37 WG 0.80 (0.57−1.13) .20 ety of women and practices and
37−41 WG 1 contributed to the study’s external
>41 WG 1.31 (1.11−1.55) .002
validity, which was also reinforced by
the similarities of our findings with
Oxytocin during labor 1.24 (1.06−1.44) .006 those of national observational studies
Retained placenta 8.49 (7.27−9.92) <.001 in France and other countries. Third, a
Type of perineal injuries consensus definition of PPH was used
that was in agreement with the World
None 1
Health Organization recommenda-
Episiotomy without tearing 1.62 (1.27−2.06) <.001 tions.1 PPH was assessed objectively by
Episiotomy and tear 3.02 (2.13−4.26) <.001 the standardized use of a graduated col-
Tear without episiotomy 1.25 (1.04−1.49) .015 lector bag.40 The volume of blood loss
was collected for each woman, thus lim-
Macrosomia 1.55 (1.22−1.96) <.001
iting classification errors. Fourth, miss-
BMI, body mass index; CI, confidence interval; IUGR, intrauterine growth restriction; OR, odds ratio; PPH, postpartum hemor-
rhage; WG, weeks of gestation ing data were compensated for by
a
P value of multivariable logistic regression adjusted for maternity ward level. multiple imputation, which distin-
Anouilh. Family history of postpartum hemorrhage is a risk factor for postpartum hemorrhage. Am J Obstet Gyne- guishes this work from other studies on
col MFM 2023. PPH that generally performed statistical
analyses only on complete observations.

September 2023 AJOG MFM 9

Original Research

The MICE method used is a particularly Moal, CM, and to the many students that con- National Perinatal Surveys. Trends in perinatal
efficient, recent statistical data process- tributed to the enrichment of the database. health in metropolitan France from 1995 to
Thanks to Jane Hanks, for her careful proof- 2016: results from the French National Perinatal
ing technique used when dealing with
reading. Surveys. J Gynecol Obstet Hum Reprod
random missing data. Its use allows a 2017;46:701–13.
qualitative improvement in the perfor- 12. Liu CN, Yu FB, Xu YZ, et al. Prevalence and
Supplementary materials
mance of the results with the estimates risk factors of severe postpartum hemorrhage:
Supplementary material associated with
being less biased and more precise.16,18 a retrospective cohort study. BMC Pregnancy
this article can be found in the online ver- Childbirth 2021;21:332.
However, our study has some limita-
sion at doi:10.1016/j.ajogmf.2023.101062. 13. Bonnet MP, Basso O, Bouvier-Colle MH,
tions. Obesity and non-European origin et al. Postpartum haemorrhage in Canada and
were uncommon in our population of References France: a population-based comparison. PLoS
pregnant women, which limits the gen- 1. WHO recommendations for the prevention One 2013;8:e66882.
eralizability of our findings. Some varia- 14. Carbonne B, Sentilhes L, Vayssiere C.
and treatment of postpartum haemorrhage.
bles, such as the dose of oxytocin used National Institutes of Health. 2012. Available at: [Intra uterine growth retardation: guidelines for
http://www.ncbi.nlm.nih.gov/books/NBK131 clinical practice − introduction]. J Gynecol
during labor, were poorly reported and Obstet Biol Reprod (Paris) 2013;42:868–9.
therefore could not be studied as poten- 942/. Accessed June 30, 2022.
15. Article R6123-39 - Code de la sant
e publi-
2. Morau E, Ducloy JC, Le Roux S, Weber P,
tial risk factors for PPH. For this reason, que. 2005. Le
gifrance. Available at: https://
Dreyfus M. [Maternal deaths due to haemor-
the use of oxytocin during labor was rhage in France 2013-2015]. Gynecol Obstet www.legifrance.gouv.fr/codes/article_lc/LEGI-
ARTI000006916836. Accessed January 11,
analyzed as a binary variable. Family Fertil Senol 2021;49:60–6.
3. Deneux-Tharaux C, Bonnet MP, Tort J. [Epi- 2023.
history of PPH was reported in a declar- 16. White IR, Royston P, Wood AM. Multiple
demiology of post-partum haemorrhage]. J
ative manner by the women, which imputation using chained equations: issues and
Gynecol Obstet Biol Reprod (Paris)
could have been affected by memory 2014;43:936–50.
guidance for practice. Stat Med 2011;30:377–
bias. However, PPH in mothers and sis- 4. Sentilhes L, Vayssiere C, Deneux-Tharaux C, 99.
ters of these young women were rela- et al. Postpartum hemorrhage: guidelines for 17. Zhang Z. Multiple imputation with multivari-
clinical practice from the French College of ate imputation by chained equation (MICE)
tively recent and well documented. Last, package. Ann Transl Med 2016;4:30.
although many variables were analyzed, Gynaecologists and Obstetricians (CNGOF): in
collaboration with the French Society of Anes- 18. Austin PC, White IR, Lee DS, van Buuren
we cannot exclude residual confusion thesiology and Intensive Care (SFAR). Eur J S. Missing data in clinical research: a tutorial on
biases because this was an observational Obstet Gynecol Reprod Biol 2016;198:12–21. multiple imputation. Can J Cardiol
2021;37:1322–31.
study. 5. Ricbourg A, Gosme C, Gayat E, Ventre C,
Barranger E, Mebazaa A. Emotional impact of 19. Dupont C, Rudigoz RC, Cortet M, et al.
severe post-partum haemorrhage on women [Frequency, causes and risk factors of postpar-
Conclusion and their partners: an observational, case-
tum haemorrhage: a population-based study in
In addition to known risk factors for matched, prospective, single-centre pilot study. 106 French maternity units]. J Gynecol Obstet
Eur J Obstet Gynecol Reprod Biol Biol Reprod (Paris) 2014;43:244–53.
PPH, this large, multicenter, prospective 20. Blondel B, Gonzalez L, Raynaud P, et al.
cohort study identified new features 2015;193:140–3.
Enqu^ete nationale p
erinatale Rapport 2016.Les
6. Green S, Ryckman KK, Anderson E, Radke
associated with PPH that is related to naissances et les e
tablissements, Situation et
S. All-cause severe maternal morbidity (SMM)
personal and family history of PPH.
evolution depuis 2010. EPOP
e; 2017. Available
and transfusion-only SMM are independently
Identifying risk factors for PPH during associated with a lower likelihood of exclusive at: https://www.epop
e-inserm.fr/wp-content/
uploads/2017/10/ENP2016_rapport_complet.
pregnancy could help to optimize medi- breastfeeding. Breastfeed Med 2022;17:758–
63. pdf. Accessed June 30, 2022.
cal care for pregnant women by guiding 21. Oberg AS, Hernande
z-Dia
z S, Frisell T,
7. Ende HB, Lozada MJ, Chestnut DH, et al.
the choice of the place of birth and by Risk factors for atonic postpartum hemorrhage:
Greene MF, Almqvist C, Bateman BT. Genetic
increasing the vigilance of professionals a systematic review and meta-analysis. Obstet contribution to postpartum haemorrhage in
in the delivery room. This also high- Gynecol 2021;137:305–23. Swedish population: cohort study of 466,686
8. Girault A, Deneux-Tharaux C, Sentilhes L, births. BMJ 2014;349:g4984.
lights the changes needed in specific 22. Sharp GC, Saunders PTK, Greene SA,
practices, such as oxytocin use during Maillard F, Goffinet F. Undiagnosed abnormal
postpartum blood loss: incidence and risk fac- Morris AD, Norman JE. Intergenerational trans-
labor. Finally, the association between tors. PLoS One 2018;13:e0190845. mission of postpartum hemorrhage risk: analy-
PPH and a family history of PPH sug- 9. Sheldon WR, Blum J, Vogel JP, et al. Post- sis of 2 Scottish birth cohorts. Am J Obstet
gests an inherited hemorrhagic pheno- Gynecol 2014;211:51.e1–7.
partum haemorrhage management, risks, and
maternal outcomes: findings from the World 23. Biguzzi E, Franchi F, Acaia B, et al. Genetic
type and paves the way for further background and risk of postpartum haemor-
genetic studies. & Health Organization Multicountry Survey on
rhage: results from an Italian cohort of 3219
Maternal and Newborn Health. BJOG
2014;121(Suppl1):5–13. women. Haemophilia 2014;20:e377–83.


10. Rubio-Alvarez
n M, Arias-
A, Molina-Alarco 24. Kramer MS, Berg C, Abenhaim H, et al.
ACKNOWLEDGMENTS
Arias A,
Hern
andez-Martínez A. Development Incidence, risk factors, and temporal trends in
severe postpartum hemorrhage. Am J Obstet
We thank CIC1412 and Brest DRCI for their and validation of a predictive model for exces-
role in all aspects of the study. We also thank all sive postpartum blood loss: a retrospective, Gynecol 2013;209:449.e1–7.
the obstetricians, midwives, and anesthetists cohort study. Int J Nurs Stud 2018;79:114–21. 25. Lautredou M, Pan-Petesch B, Dupr
e PF,
involved in the study. Thanks to the midwifery 11. Blondel B, Coulm B, Bonnet C, Goffinet F, et al. Excessive gestational weight gain is an
independent risk factor for gestational diabetes
school of Brest (UBO), to its director, Anne Le Ray C. National Coordination Group of the

10 AJOG MFM September 2023

 Original Research

mellitus in singleton pregnancies: results from a 34. Erickson EN, Carlson NS. Predicting postpar- M
edecine Vasculaire et Pneumologie, Centre Hospitalier
French cohort study. Eur J Obstet Gynecol tum hemorrhage after low-risk vaginal birth by Universitaire Brest, Brest, France (Drs Moreuil, Coutur-
Reprod Biol 2022;275:31–6. labor characteristics and oxytocin administration. J aud, Tromeur, and Le Moigne); Service de Gyn
ecologie
26. Salomon C, de Moreuil C, Hannigsberg J, Obstet Gynecol Neonatal Nurs 2020;49:549–63. Obst
etrique, Centre Hospitalier Universitaire Brest, Brest,
et al. Haematological parameters associated 35. Graugaard HL, Maimburg RD. Is the France (Dr Tr
emouilhac, Ms Cornec, and Drs Dupr
e and
with postpartum haemorrhage after vaginal increase in postpartum hemorrhage after vaginal Morcel); Service de Gyn
ecologie Obst
etrique, CHIC de
delivery: results from a French cohort study. J birth because of altered clinical practice?: a reg- Quimper, Quimper, France (Drs Jacquot, Bellot, and Le
Gynecol Obstet Hum Reprod 2021;50:102168. ister-based cohort study. Birth 2021;48:338–46. Joliff); Service de Gyn
ecologie Obst
etrique, CH des Pays
27. Cortet M, Maucort-Boulch D, Deneux- 36. Balki M, Ramachandran N, Lee S, Talati C. de Morlaix, Morlaix, France (Drs Salnelle and Muller); Ser-
Tharaux C, et al. Severity of post-partum hem- The recovery time of myometrial responsive- vice de Gyn
ecologie Obst
etrique, Centre Hospitalier Priv
e
orrhage after vaginal delivery is not predictable ness after oxytocin-induced desensitization in de Brest - Keraudren, Brest, France (Dr Bellec); Service
from clinical variables available at the time post- human myometrium in vitro. Anesth Analg de Gyn
ecologie Obst
etrique, CH de Landerneau, Lander-
partum hemorrhage is diagnosed. J Obstet 2016;122:1508–15. neau, France (Dr Touffet); CIC1412, Institut National de
Gynaecol Res 2015;41:199–206. 37. Erickson EN, Krol KM, Perkeybile AM, la Sante et de la Recherche Medicale, Brest, France (Mr
28. Nyfløt LT, Sandven I, Stray-Pedersen B, Connelly JJ, Myatt L. Oxytocin receptor single Drugmanne, Ms Gelebart, Ms Lucier, Dr Nowak, and Ms
et al. Risk factors for severe postpartum hemor- nucleotide polymorphism predicts atony- Bihan); Centre de Ressources et de Comp
etence des
rhage: a case-control study. BMC Pregnancy related postpartum hemorrhage. BMC Preg- Maladies H
emorragiques, Centre de Ressources et de
Childbirth 2017;17:17. nancy Childbirth 2022;22:884. Comp
etence des Maladies H
emorragiques, H
emostase,
29. Mehrabadi A, Hutcheon JA, Lee L, Kramer 38. Prevention and management of postpar- Service H
ematologie, Centre Hospitalier Universitaire
MS, Liston RM, Joseph KS. Epidemiological tum haemorrhage: Green-top Guideline No. 52. Brest, Brest, France (Dr Pan-Petesch).
investigation of a temporal increase in atonic BJOG 2017;124:e106–49. Received Apr. 17, 2023; revised June 8, 2023;
postpartum haemorrhage: a population-based 39. Escobar MF, Nassar AH, Theron G, et al. accepted June 12, 2023.
retrospective cohort study. BJOG FIGO recommendations on the management of F.A. and C.d.M. contributed equally to this study.
2013;120:853–62. postpartum hemorrhage 2022. Int J Gynaecol This study was registered with ClinicalTrials.gov under
30. Feduniw S, Warzecha D, Szymusik I, Wiel- Obstet 2022;157(Suppl1):3–50. the title “Hemorrhages and Thromboembolic Venous Dis-
gos M. Epidemiology, prevention and manage- 40. Legendre G, Richard M, Brun S, Chancerel ease of the Postpartum (HEMOTHEPP)” with identifier
ment of early postpartum hemorrhage − a M, Matuszewski S, Sentilhes L. Evaluation by NCT02443610.
systematic review. Ginekol Pol 2020;91:38–44. obstetric care providers of simulated postpar- The authors report no conflict of interest.
31. Biguzzi E, Franchi F, Ambrogi F, et al. Risk fac- tum blood loss using a collector bag: a French The study was conducted in agreement with and with
tors for postpartum hemorrhage in a cohort of prospective study. J Matern Fetal Neonatal the institutional support from the University Hospital of
6011 Italian women. Thromb Res 2012;129:e1–7. Med 2016;29:3575–81. Brest without a specific grant. The University Hospital of
32. Contreras KR, Kominiarek MA, Zollinger Brest (the study sponsor) had no role in the design of the
TW. The impact of tobacco smoking on perina- trial, the collection, analysis, or interpretation of the data,
tal outcome among patients with gestational Author and article information the writing of the manuscript, or the decision to submit
diabetes. J Perinatol 2010;30:319–23. From the Ecole Universitaire de Maïeutique de Brest, UFR for publication. An academic steering committee
33. Feferkorn I, Badeghiesh A, Baghlaf H, Sant
e - Brest, Brest, France (Mr Anouilh); UMR 1304, assumed overall responsibility for all of these steps.
Dahan MH. The relation between cigarette GETBO, Universit
e de Bretagne Occidentale - Brest Corresponding author: Claire de Moreuil, D
epartement
smoking with delivery outcomes. An evaluation (France), Brest, France (Mr Anouilh, Drs de Moreuil, de M
edecine Interne, M
edecine Vasculaire et Pneumolo-
of a database of more than nine million deliver- Tr
emouilhac, Morcel, Couturaud, Tromeur, Le Moigne, gie, CHU Brest - Brest (France), France, MD, PhD.
ies. J Perinat Med 2022;50:56–62. and Pan-Petesch); D
epartement de M
edecine Interne, claire.demoreuil@wanadoo.fr

September 2023 AJOG MFM 11