www.elsevier.com/locate/jhep

Journal of Hepatology 51 (2009) 47–54

Pregnancy in women with known and treated Budd–Chiari syndrome:Maternal and fetal outcomesq

Pierre-Emmanuel Rautou1, Bernhard Angermayr2, Juan-Carlos Garcia-Pagan3,Rami Moucari1, Markus Peck-Radosavljevic2, Sebastian Raffa3, Jacques Bernuau1,

Bertrand Condat1, Michel Levardon4, Carine Yver4, Guillaume Ducarme4,Dominique Luton4, Marie-Helene Denninger5, Dominique Valla1,*, Aurelie Plessier1

1Pole des Maladies de l’Appareil Digestif, Service d’Hepatologie, Hopital Beaujon, AP-HP and INSERM U773,

Centre de Recherche Biomedicale Bichat Beaujon CRB3, Universite Paris 7-Denis-Diderot, Clichy, France2Department of Gastroenterology and Hepatology, Medizinische Universitat & AKH Wien, Vienna, Austria

3Hepatic Hemodynamic Laboratory, Liver Unit, Hospital Clinic, IDIBAPS (Institut de Investigacions Biomediques August Pi i Sunyer) and

Ciberehd (Centre de Investigacion en Red de Enfermedades Hepaticas y Digestivas), Barcelona, Spain4Service de Gynecologie et Obstetrique, Hopital Beaujon, Clichy, France

5Service d’Hematologie Biologique, Hopital Beaujon, Clichy, France

Background/Aims: Budd–Chiari syndrome (BCS) mainly affects women of childbearing age. We aimed to clarify

whether pregnancy, a thrombotic risk factor, should be contraindicated in patients with known and treated BCS.

Methods: A retrospective study of pregnancy in women with known and treated BCS.

Results: Sixteen women had 24 pregnancies. Nine women had undergone surgical or radiological treatment. Anticoag-

ulation was administered during 17 pregnancies. Seven fetuses were lost before gestation week 20. Deliveries occurred

between week 20 and 31 in two patients, week 32 and 36 in eleven and after week 37 in four. There was one stillbirth,

but 16 infants did well. Factor II gene mutation was a factor for a poor outcome of pregnancies. In two patients, symp-tomatic thrombosis recurred during pregnancy or postpartum. All patients were alive after a median follow-up of 34

months after the last delivery. Bleeding at delivery, although non-lethal, occurred only on anticoagulation therapy.

Conclusions: When known and treated BCS is well controlled, pregnancy should not be contraindicated as maternal out-

come, and fetal outcome beyond gestation week 20, are good. The risk-benefit ratio of anticoagulant therapy needs to be

further clarified. Patients should be fully informed of the persistent risks of such pregnancies.

� 2009 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.

Keywords: Preterm; Delivery; Miscarriage; Thrombophilia; Prothrombin Gene mutation

0168-8278/$36.00 � 2009 European Association for the Study of the Liver.

doi:10.1016/j.jhep.2009.02.028

Received 18 January 2009; accepted 10 February 2009; available online

16 April 2009

Associate Editor: C. Merkelq The authors who have taken part in this study declared that they do

not have anything to disclose regarding funding from industry orconflict of interest with respect to this manuscript.

* Corresponding author. Tel.: +33 140 875594; fax: +33 140 874426.E-mail address: dominique.valla@bjn.aphp.fr (D. Valla).Abbreviations: BCS, Budd–Chiari syndrome; LMWH, low molec-

ular weight heparin; TIPS, transjugular intrahepatic portosystemicshunt.

1. Introduction

Budd–Chiari syndrome (BCS) is a rare disordercaused by the obstruction of hepatic venous outflow,leading to sinusoidal congestion, ischemic injury to theliver and portal hypertension. The main mechanismsfor BCS is thrombosis of the hepatic veins, or of the ter-minal portion of the inferior vena cava [1,2]. Multiplerisk factors for venous thromboembolism, including

Published by Elsevier B.V. All rights reserved.

48 P.-E. Rautou et al. / Journal of Hepatology 51 (2009) 47–54

various prothrombotic conditions, are usually present inBCS patients [1,3–5].

Pregnancy is a crucial issue in women with BCS sincethe desire for pregnancy is increasingly expressed bythese young patients once their condition has greatlyimproved [6]. However, these women could be at riskof developing severe exacerbation of their disease duringpregnancy [7,8]. As available reports on pregnancy inwomen with known BCS are scarce, it is unclear whetheror not pregnancy should be contraindicated in patientswith known and treated BCS [9–16]. This prompted usto assess the fetal and maternal outcome in a group ofwomen with a desire for pregnancy and who becamepregnant after BCS had been recognized and treated.

2. Patients and methods

2.1. Definitions

Date of diagnosis of BCS corresponded to the first imaging proce-dure showing an obstructed venous outflow tract [2]. Rotterdam prog-nostic index was calculated as initially described [17]. Miscarriageswere defined as a spontaneous termination of pregnancy before 20weeks’ gestation.

In order to identify its prognostic factors, pregnancy outcome wasclassified into 2 categories: (a) favorable outcome, when live birthoccurred at 32 or more completed weeks of gestation, with a healthyinfant and no serious obstetrical complication (intrahepatic cholestasiswas not considered as a serious complication); and (b) poor outcome,otherwise.

2.2. Study group

The files of all female patients, aged 15–45 years, diagnosed withBCS between January 1985 and December 2005 at three referral cen-ters (Hopital Beaujon, Clichy, France; Medizinische Universitaet,Vienna, Austria; and Hospital Clinic, Barcelona, Spain) were retro-spectively analyzed. Among these women, only those who becamepregnant during the follow-up for a known and treated BCS wereincluded in the study. Thus, 4 patients in whom pregnancy occurredbefore BCS was diagnosed and treated were not included although itis acknowledged that BCS might have already been present at thattime. Three of these 4 patients had a diagnosis of BCS made at the timeof pregnancy or postpartum and are reported elsewhere [18]. Weexcluded from analysis pregnancies voluntarily interrupted and preg-nancies occurring after liver transplantation since these patients dono longer have BCS.

Some of the cases reported here have already been included in pre-vious publications [5,18–22].

2.3. Work-up and management

Patients were tested according to previously reported methods [23]for the following prothrombotic factors: primary myeloproliferativedisorders [24]; factor V R 506 Q mutation (factor V Leiden); G20210 A factor II gene mutation; deficiencies in protein C, protein S,or antithrombin (regarded as primary deficiencies only in conjunctionwith normal coagulation factor levels); paroxysmal nocturnal hemo-globinuria; and anti-phospholipid antibodies. In all patients, testingfor deficiencies in antithrombin, protein C and protein S was per-formed either before conception, or at least 6 months after the endof pregnancy and in the absence of oral contraception [25,26]. Oralcontraceptive use within the 3 months preceding diagnosis of BCSwas considered a thrombotic risk factor [27].

BCS was managed using anticoagulation therapy, radiologicalinterventions [venous angioplasty, transjugular intrahepatic portosys-temic shunt (TIPS)], surgical portosystemic shunting, and/or livertransplantation treatment as previously reported [19,20]. As mentionedabove, liver transplanted patients were not included in this study.Patients managed at our centers were fully informed of the scarcityof data on pregnancy in patients with BCS, and of possible maternaland fetal risks. Given the risk of embryopathy and fetal loss associatedwith coumarin and phenindione, low molecular weight heparin(LMWH) was substituted with vitamin K antagonists when the patientexpressed the intention to become pregnant [28], or as soon as preg-nancy was diagnosed otherwise. LMWH were continued at least 2months after delivery, and subsequently replaced by vitamin Kantagonists.

The study of the cohort conforms to the ethical guidelines of the1975 Declaration of Helsinki and was approved by the institutionalreview board Bichat-Claude-Bernard (CHU Bichat, Paris, France).

2.4. Statistical analysis

Quantitative variables were expressed as median with range, andcategorical variables as absolute and relative frequencies. Comparisonsbetween groups of quantitative and qualitative variables were per-formed using respectively Mann–Whitney and the Fisher exact tests.Comparison of laboratory data at the time of diagnosis and last deliv-ery were performed using Wilcoxon test. All tests were two-sided andused a significance level of 0.05. Data handling and analysis were per-formed with SPSS 12.0 (SPSS Inc., Chicago, IL).

3. Results

3.1. General characteristics

Sixteen patients had 24 pregnancies during follow-upafter BCS was diagnosed in one of the three referral cen-ters from 1985 to 2005. Median age at conception was33 years (range 21–41). Median time between diagnosisof BCS and conception was 57 months (range, 4–184).During the period preceding pregnancy, 9 out of the16 patients underwent 11 liver decompression proce-dures, a median of 65 months (range, 5–174) before con-ception. Procedures included angioplasty in 4 patients,TIPS in 2, and surgical portosystemic shunts in 5. Addi-tionally, 2 patients had ovulation induction therapyusing clomiphene, and none of them developed throm-botic or liver-related complications.

Twenty pregnancies occurred in 13 out of the 94 BCSwomen aged 15–45 years diagnosed from 1985 to 2005in Hopital Beaujon; 1 pregnancy occurred in 1 out ofthe 33 BCS women aged 15–45 years diagnosed in thesame period in Hospital Clinic; and 3 pregnanciesoccurred in 2 patients in Medizinische Universitaet(total number not available for the whole period). Six-teen pregnancies were managed in one of the three cen-ters participating in this study whereas eight others,occurring in 6 patients, were managed in other obstetrichospitals.

Among the 16 patients, at least one causal factor forhepatic vein obstruction, other than pregnancy, wasidentified in 14 women (88%). These causal factorsincluded oral contraceptive use in 63% (10 out of 16

P.-E. Rautou et al. / Journal of Hepatology 51 (2009) 47–54 49

patients), myeloproliferative disorders in 27% (4 out of15 patients tested), factor II gene mutation in 31% (4out of 13 tested), and factor V Leiden in 25% (4 outof 16 tested), as well as other, less common, factors.At diagnosis, the number of thrombosed hepatic veinswas 1, 2, 3, and unavailable in 3, 3, 10, and 1 patients.Inferior vena cava thrombosis was also present in 2women.

3.2. Anticoagulation

Anticoagulant therapy was administered during 17 ofthe 24 pregnancies. Heparin therapy was used during 15pregnancies (unfractionated heparin in 1 and LMWH in14 pregnancies) after the substitution of vitamin Kantagonists to LMWH or heparin took place beforethe conception of 9 pregnancies, and a median of 4.5weeks (range, 1.3–7.3 weeks) after the conception of 6pregnancies. In the last two pregnancies, only vitaminK antagonists were used because miscarriages occurredat 6 and 9 gestation week, before the switch to LMWHcould be made.

No anticoagulant therapy was given during 7 preg-nancies. In 4 of them (occurring in 2 patients with amyeloproliferative disease), antiplatelet therapy wasadministered. After BCS diagnosis, these two patientsreceived anticoagulation therapy, and were treated withsurgical portosystemic shunting. Following this treat-ment, symptoms disappeared and vitamin K antagonistswere replaced by antiplatelet therapy. Pregnancy hap-pened several months later. The last 3 pregnanciesoccurred in 2 women managed without antithrombotictherapy. In one of them (Patient 8 in Table 6), anticoag-ulation was stopped after TIPS placement and disap-pearance of symptoms. Then, she became pregnanttwice. Her second pregnancy was complicated withsymptomatic TIPS obstruction. In the other womanwithout antithrombotic therapy during pregnancy, theonly identified risk factor for BCS was a history of oralcontraceptive use. She had never received anticoagula-tion therapy, but had a TIPS placement 5 years beforepregnancy.

Table 1

Gestational course and perinatal complications in 24 pregnancies occurring du

Weeks gestation Number ofpregnancies

Mode of delivery B(m

Ectopic pregnancies(3 and 7 weeks)

2 Salpingectomy

<20 5 Vaginal 520–31 2 Caesarean 1 11

Vaginal 132–36 11 Caesarean 7 25

Vaginal 4

P37 4 Vaginal 4 31

3.3. Gestational course and perinatal complications

All 16 patients that became pregnant had compen-sated liver disease at the time of conception. Gestationalcourse is detailed in Table 1. Seven out of the 24 preg-nancies terminated with fetal loss before gestation week20 (including two patients with ectopic pregnancy).

Out of the 17 pregnancies reaching gestation week 20,there were 9 vaginal deliveries and 8 caesarean sections.Two pregnancies resulted in very preterm births (i.e.between weeks 20 and 31 of completed gestation) of onestillbirth and one infant with low birth weight. Eleven preg-nancies resulted in moderately preterm births (i.e. betweenweeks 32 and 36 of completed gestation) of 11 infants witha favorable outcome, although 3 were small for gestationalage and 1 had severe respiratory distress syndrome.Finally, 4 pregnancies gave 4 infants at 37 or more com-pleted gestation weeks, all with a favorable outcome.

Maternal complications during pregnancy occurredin 9 women, including intrahepatic cholestasis recurringin all pregnancies in 3 patients; intrauterine hematomain 3 (Table 2); preeclampsia in 1; placenta praevia in1; and reversible exanthema of lower limbs with lym-phadenopathy of unknown origin in 1.

As detailed in Table 2, bleeding also occurred in 4women after the end of pregnancy. Among these 4 bleed-ings, one was related to ectopic pregnancy (fetal loss at 8gestation week) and the other 3 occurred following deliv-eries between gestation weeks 33 and 36. Two of these 4women had hematoma requiring surgical evacuation orprolonged hospitalization after salpingectomy (ectopicpregnancy) or caesarean section; in both cases, LMWHhad been continued or anti-Xa activity was still in thera-peutic ranges at delivery despite previous discontinuation.The remaining 2 women had vaginal bleeding after cae-sarean or vaginal delivery. Among children born to moth-ers with bleeding, 2 infants did well while the remainingone had a severe respiratory distress syndrome.

Pregnancy outcome was classified as favorable orpoor in 12 instances respectively. Among the baselinefactors presented in Table 3, only factor II gene muta-tion was associated with poor outcome of pregnancies.

ring the follow-up of BCS.

irth weightedian; range) (g)

Infant condition

30 - 1 low birth weight- 1 stillbirth attributed to anticoagulation

50 (2030–2900) - 7 healthy- 3 small for gestational age- 1 severe respiratory distress syndrome

50 (3000–3300) All healthy

Table 2

Data in six BCS women with bleeding during pregnancy or postpartum.

Patient Age atpregnancy(years)

Risk factorsfor BCS

Anticoagulationduringpregnancy

Mode ofdelivery

Date ofbleeding

Anticoagulationmanagementat delivery

Bleeding site Therapy forbleeding

1 31 Oralcontraceptiveuse,factor II genemutation,protein Cdeficiency

LMWHintroducedbefore 5 weeksgestation

Vaginal First trimesterof pregnancy

Intrauterinehematoma

Bedridden

2 23 Oralcontraceptiveuse,factor IIgene mutation

LMWHintroducedbeforeconception

Caesarean 18 gestationweeks

Intrauterinehematomaandvaginalbleeding

Bedridden

11 gestationweeks

Intrauterinehematoma

Bedridden

3 42 Oralcontraceptiveuse,FactorV Leiden

LMWHintroducedbeforeconception

Caesarean Earlypostpartum

Anti-Xaactivitybelowtherapeuticranges(<0.06 U/mL)

Moderatevaginal bleedingfor 3 weeks

None

4 40 Oralcontraceptiveuse

LMWHintroducedbefore 5weeks gestation

Caesarean 5 dayspostpartum

Anti-Xaactivitystill intherapeuticranges(>0.22 U/mL)

Parietalhematoma

- LMWHdosereduced

- 18 dayshospitalization

5 38 ProteinS deficiency

LMWHintroducedbeforeconception

Salpingectomy 5 dayspostpartum

LMWHcontinued

Parietalhematoma

- LMWHinterruption

- Surgicalevacuation

6 27 None LMWHintroducedbeforeconception

Vaginal 8 dayspostpartum

Notavailable

Acutevaginalbleeding

LMWHinterruption

Abbreviation: LMWH, low molecular weight heparin.

50 P.-E. Rautou et al. / Journal of Hepatology 51 (2009) 47–54

The outcome of pregnancies was not influenced by thedate of conception, the center of management, other riskfactors of BCS, clinical, laboratory or other radiologicaldata, prognostic indices, or interventional therapyreceived for BCS. Among the 5 women who had 2 ormore pregnancies in the follow-up for BCS, one had 2pregnancies with a favorable outcome (as defined inthe Section 2) whereas outcome of the last pregnancywas poor in the 4 other women. Among the latter, onepatient had 3 pregnancies with a poor outcome whilethe other 3 women had initially 1 or 2 pregnancies witha favorable outcome each and, then, a pregnancy with apoor outcome.

Pregnancy and maternal outcome according to antico-agulation received during pregnancy are presented inTable 4. The only difference was that all 6 pregnanciescomplicated by bleeding occurred in women receivinganticoagulation. One pregnancy was complicated by intra-

uterine hematoma and then postpartum bleeding. Thedate of conception and the center of management of preg-nancies were not related with the occurrence of bleedings.

3.4. Impact of pregnancies on vascular disease in the

mother

As shown in Table 5, serum ALT level and Rotter-dam BCS index were significantly higher at diagnosisthan at delivery. In most women, clinical, laboratoryand imaging features were not recorded within a shorttime prior to conception. Three of the 16 patients devel-oped pregnancy-related thrombosis with a good mater-nal outcome. In all these three cases, fetal outcomewas poor. The main characteristics of the thromboticepisodes are presented in Table 6.

The 16 women were followed-up for a median of 34months (range, 2–119) after the last delivery. At the

Table 3

Baseline factors associated with pregnancy outcome in 24 pregnancies in 16 women with known and treated BCS.

Favorable outcome (n = 12) Poor outcome (n = 12) p

Clinical manifestations

Age at BCS diagnosis – years 27 ± 7 28 ± 6 0.84Age at conception – years 31 ± 6 34 ± 5 0.21Date of conception – median (range) 25/10/2001 (11/11/1985–23/09/2006) 20/09/1999 (23/05/1993–05/07/2006) 0.854Follow-up of pregnancy in 1 of the 3 study centers 8/12 (67) 8/12 (67) 1.00Ascites 7/12 (58) 10/12 (83) 0.37Encephalopathy 0/12 (0) 1/12 (8) 1.00Abdominal pain 4/9 (44) 5/11 (45) 1.00Oesophageal varices 9/10 (90) 6/10 (60) 0.30

Serum levels

ALT (� ULN) (available in 24) 1.6 ± 1.2 11.3 ± 32.9 0.80Bilirubin (lmol/L) (available in 23) 21 ± 13 24 ± 15 0.79Prothrombin time (%) (available in 24) 75 ± 16 67 ± 18 0.32Albumin (g/L) (available in 16) 42 ± 2 37 ± 8 0.25Plasma creatinine (lmol/L) (available in 23) 66 ± 17 88 ± 56 0.32

Prognostic indices (PI)Child–Pugh PI (available in 16) 6.4 ± 1.1 7.0 ± 1.5 0.54Rotterdam PI (available in 23) 0.69 ± 0.56 1.12 ± 0.67 0.21Clichy PI (available in 16) 4.26 ± 0.49 4.70 ± 0.72 0.11

Radiological data

Hepatomegaly 10/12 (83) 9/12 (75) 1.00Splenomegaly 9/12 (75) 10/12 (83) 1.00Atrophy/hypertrophy of liver lobes 5/9 (42) 4/11 (36) 0.65IVC obstruction 1/12 (8) 3/12 (25) 0.59Normal hepatic veins (0/1/2/3) 9/ 1/ 1/ 0 9/ 2/ 1/ 0 0.87Intrahepatic venous collaterals 8/11 (73) 9/11 (82) 1.00Portosystemic venous collaterals 6/11 (55) 2/11 (18) 0.18

Prothrombotic disorders – n (%)Myeloproliferative disorders 5/12 (42) 3/11 (27) 0.67Factor V Leiden 3/12 (25) 3/12 (25) 1.00Prothrombin gene mutation 1/10 (10) 6/10 (60) 0.05

Protein C deficiency 0/12 (0) 3/12 (25) 0.22Protein S deficiency 2/11 (18) 1/12 (8) 0.59Oral contraceptive use – n (%) 10/12 (83) 9/12 (75) 1.00

Treatment before conception – n (%)Angioplasty 4/12 (33) 5/12 (42) 1.00TIPS 2/12 (17) 1/12 (8) 1.00Surgical portosystemic shunt 4/12 (33) 5/12 (42) 1.00

Data are means ± SD or frequency (%). Variables reaching a significance level of 0.05 are in bold.Abbreviation: ULN, upper limit of normal values.None of the patients had portal or splenic or mesenteric vein thrombosis at diagnosis, antithrombin III deficiency, paroxysmal nocturnal hemo-

globinuria or anti-phospholipid antibodies syndrome.

P.-E. Rautou et al. / Journal of Hepatology 51 (2009) 47–54 51

end of follow-up, 14 women were alive and had notneeded further interventional therapy for BCS, whereasthe remaining 2 women had undergone invasive therapiesfor BCS after pregnancy: one had TIPS recanalization(Patient 8, Table 6) and the other had liver transplanta-tion, 73 months after delivery, because of symptomaticthrombosis of a surgical portosystemic shunt.

4. Discussion

A majority of the patients affected with BCS in Wes-tern countries are women of child-bearing age [1,17,29].Many of these young women insistently express their

desire for pregnancy. However, several previouslyreported observations suggest that pregnancy in BCSwomen could cause liver disease to deteriorate. Preg-nancy was reported to induce the development of ascitesin several women with known BCS [9,12,16]. Moreover,in older series of patients in India, women presentingwith BCS during pregnancy or postpartum, i.e. withan unknown and untreated BCS, had a poor outcomewith a 1-yr mortality rate around 50% [7,8]. Finally,the risk of recurrence of hepatic vein thrombosis inwomen with recognized BCS who become pregnant islikely increased by high levels of female sex hormones.Indeed, pregnancy is a risk factor for developing BCSin women with underlying prothrombotic conditions

Table 4

Pregnancy and maternal outcome in 24 pregnancies occurring during the follow-up of BCS according to anticoagulation received.

LMWH introduced beforeconception

p LMWH introduced before5 weeks gestation

p Anticoagulation (heparinor oral) during pregnancy

p

Yes (n = 9) No (n = 15) Yes (n = 14) No (n = 10) Yes (n = 17) No (n = 7)

Favorable outcome of pregnancya 4 (44) 8 (53) 1.00 7 (50) 5 (50) 1.00 8 (47) 4 (57) 1.00Favorable maternal outcomeb 7 (78) 12 (80) 1.00 11 (79) 8 (80) 1.00 14 (82) 5 (71) 0.61Bleeding 4 (44) 2 (13) 0.15 6 (43) 0 0.03 6 (35) 0 0.13

Seven bleedings occurred in 6 pregnancies.Data are presented as the number of patients (percentage).a Live birth occurring at 32 or more completed weeks of gestation, with a healthy infant and no serious obstetrical complication (intrahepatic

cholestasis was not considered as a serious complication).b Absence of any maternal complication, particularly any preeclampsia or thrombotic or liver-related events.

52 P.-E. Rautou et al. / Journal of Hepatology 51 (2009) 47–54

[7,8,18]. Furthermore, oral contraceptive use has beenshown to increase the risk of BCS with odds ratiosabout 2.5 as compared to non-users [27,30].

The present study clearly shows that the maternaloutcome is currently good in women becoming pregnantafter the diagnosis of BCS was made and initiation of itstreatment. All mothers were alive at a median follow-upof 34 months after last delivery and only one of themrequired liver transplantation after 73 months follow-up. It is tempting to explain this good maternal outcomeby improvement in BCS management including treat-ment for underlying conditions, anticoagulant therapy,and efficient decompressive interventions [20,31], as wellas by careful follow-up and management of pregnancyand delivery. There are little data in the literature allow-ing for a valid comparison with our present findings.Great caution is required when comparing our Euro-pean patients seen after 1985 with Indian patients man-aged between 1960 and the early nineties. Twelvepregnant Indian women with symptoms of a chronicform of BCS for a mean duration of 52 months, butwho did not receive anticoagulation and likely nodecompressive therapy prior to pregnancy, had a 17%fatality rate during hospitalization [8]. In another seriesof Indian women, the one-year mortality rate was 50%but these women suffered BCS revealed by pregnancy[7,8]. Therefore, currently, it seems that modern overallmanagement of BCS has allowed an increasing numberof young female patients to achieve an improved condi-tion allowing them to consider pregnancy and gothrough this pregnancy without markedly and immedi-ately jeopardizing their lives.

Table 5

Laboratory data at the time of diagnosis and last delivery in 16 BCS patients

n At d

ALT (� ULN) 14 1.4 (Bilirubin (lmol/L) 11 16 (7Prothrombin time (%) 14 71 (2Albumin (g/L) 6 43 (3Plasma creatinine (lmol/L) 9 61 (3Rotterdam prognostic index 10 1.1 (

Data are median (range).Abbreviation: ULN, times upper limit of normal.

In the present series, thrombotic events occurringduring pregnancy or postpartum were uncommon andrelatively easy to manage successfully (Table 6). It is dif-ficult, however, to ascertain firmly the role of antithrom-botic therapy in this relatively favorable outcome of thethrombotic complications. None of the 4 pregnancies onantiplatelet agents (in 2 women with myeloproliferativedisease) was complicated by thrombosis; 2 of 17 preg-nancies on anticoagulation therapy were complicatedby portal vein thrombosis; one of 3 pregnancies man-aged without any antithrombotic agents was compli-cated by TIPS obstruction. These limited data appearto support administration of antithrombotic therapywithout answering the issue of the type of antithrom-botic therapy to be chosen.

Bleeding occurred more often after caesarean sectionsthan after vaginal deliveries. Caesarean section is other-wise known to be associated with an increased risk ofthromboembolic disease [32] and could be hazardousin patients with portal hypertension due to large pelvicvenous collaterals and postoperative ascites [33]. There-fore, in the absence of a compelling indication for cae-sarean section, vaginal delivery should be encouragedin BCS patients.

Anticoagulation was associated with a higher fre-quency of bleedings since all 7 bleedings in this series(including all 3 intrauterine hematomas), occurred inwomen receiving anticoagulation. In two pregnanciescomplicated by severe bleeding, anti-Xa activity was stillin therapeutic ranges when caesarean section was per-formed. Thus, a closer management of anticoagulation

who became pregnant during follow-up.

iagnosis At delivery p

0.4–115.5) 0.5 (0.1–2.0) 0.03–43) 9 (3–46) 0.299–100) 84 (28–104) 0.072–46) 33 (23–47) 0.147–78) 50 (37–70) 0.120.03–2.4) 0.04 (0.01–1.1) 0.03

Table 6

Data in 3 BCS women with pregnancy-related thrombosis.

Patient Age atpregnancy(years)

Riskfactorsfor BCS

Anticoagu-lationduringpregnancy

Mode ofdelivery

Number ofgestationweeks atend ofpregnancy/infantcondition

Date ofdiagnosisof thrombosis

Manifestation/site ofthrombosis

Triggeringfactor

Treatmentforthrombosis/outcome

5 38 ProteinS deficiency

LMWHintroducedbeforeconception

Salpin-gectomy

8 weeks/ectopicpregnancy

8 days aftersalpingectomy

Ascites/portaland mesentericveins (partial)

- HIT- Infection

of aparietalhematoma

Medicaltherapy/ascitesdisappearance

7 32 None LMWHintroducedbefore 5weeksgestation

Vaginal 22 weeks/intrauterinefetal death

17 weeksgestation

Asymptomatic/obstruction ofa previoussurgicalmesenterico-cavalshunt

None None

8 36 Oralcontraceptiveuse, factor IImutation

None Caesarean 19 weeks/intrauterinefetal death

19 weeksgestation

Ascites/TIPSobstruction

None Recanalization/ascitesdisappearance

Patient 5 in this table is the same patient as patient 5 in Table 2.Abbreviations: HIT, heparin-induced thrombocytopenia; LMWH, low molecular weight heparin.

P.-E. Rautou et al. / Journal of Hepatology 51 (2009) 47–54 53

and a window of discontinuation around delivery coulddecrease this bleeding risk.

Contrasting with the good overall maternal outcomeobserved in the present series, fetal outcome remains amatter of concern. The 29% rate (95% confidence inter-val, 11–47%) of pregnancy loss before gestation week 20(including 2 ectopic pregnancies) was somewhat higherthan the corresponding figure in a healthy woman pop-ulation of similar age and period of observation (16.2%;95% confidence interval, 12.7 to 19.7%) [34]. Only twocases of pregnancy loss before gestation week 20 havebeen previously reported in BCS women [12,15]. Amongpregnancies reaching gestation week 20 in our series,preterm delivery was common (76%; 95% confidenceinterval, 56–97%). However, except for one stillbirth,the fetal outcome was relatively good, which is consis-tent with literature data in women with known BCS[9–11,13,14,16]. Out of 10 previously reported cases,there were 9 moderately preterm or term deliveries withhealthy infants and one early preterm delivery of aninfant who died 2 days later.

Factor II gene mutation, otherwise known to be asso-ciated with miscarriage and late fetal loss [35], was signif-icantly associated with a poor outcome. This findingsuggests that, in women affected by BCS, some underlyingprothrombotic disorders may place pregnancies at a par-ticularly high risk. Older age at conception tended to beassociated with a poor outcome of pregnancy (Table 3),in accordance with what has been observed in the generalpopulation [36]. Moreover, in these women with BCS, apregnancy with a poor outcome may follow previous

pregnancies with favorable outcome. Finally, factorsrelated to the liver disease per se could also play a rolein the outcome of pregnancy. Indeed, as shown in Table3, the outcome of pregnancies tended to be poor whenseverity of BCS at diagnosis was high. The impact of por-tal hypertension per se, and thus of portal decompressivetherapy, is difficult to analyze separately in this context.

In conclusion, with current treatment modalities andclose surveillance of BCS, maternal outcome is good inwomen with pregnancy occurring after BCS diagnosisand treatment. Therefore, BCS cannot be considered acontraindication to pregnancy in stable patients with awell controlled disease. However, fetal outcome remainsrelatively poor, which is at least partly related to underly-ing prothrombotic disorders. Patients should be accu-rately informed of fetal and maternal risk, of suchpregnancies. Although antithrombotic therapy may pre-vent recurrence of splanchnic vein thrombosis, theiractual beneficial effect during pregnancy in these womenstill needs to be clarified. In order to decrease the incidenceof bleeding, every effort should be made to prevent over-dosing of antithrombotic therapy, and a brief interruptionof anticoagulation at the onset of labor might be helpful.

Acknowledgements

We thank Dr. Noelle Bendersky, Alice Marot andMohamed Achahboun for their help identifying pa-tients, Beatrice Michel and Saida Roussin for their sec-retarial work.

Ciberehd is funded by Instituto de Salud Carlos III.

54 P.-E. Rautou et al. / Journal of Hepatology 51 (2009) 47–54

References

[1] Valla DC. The diagnosis and management of the Budd–Chiarisyndrome: consensus and controversies. Hepatology2003;38:793–803.

[2] Janssen HL, Garcia-Pagan JC, Elias E, Mentha G, Hadengue A,Valla DC. Budd–Chiari syndrome: a review by an expert panel. JHepatol 2003;38:364–371.

[3] Mohanty D, Shetty S, Ghosh K, Pawar A, Abraham P.Hereditary thrombophilia as a cause of Budd–Chiari syndrome:a study from Western India. Hepatology 2001;34:666–670.

[4] Janssen HL, Meinardi JR, Vleggaar FP, van Uum SH, HaagsmaEB, van Der Meer FJ, et al. Factor V Leiden mutation,prothrombin gene mutation, and deficiencies in coagulationinhibitors associated with Budd–Chiari syndrome and portal veinthrombosis: results of a case-control study. Blood2000;96:2364–2368.

[5] Deltenre P, Denninger MH, Hillaire S, Guillin MC, Casadevall N,Briere J, et al. Factor V Leiden related Budd–Chiari syndrome.Gut 2001;48:264–268.

[6] Valla DC. Budd–Chiari syndrome and veno-occlusive disease/sinusoidal obstruction syndrome. Gut 2008;57:1469–1478.

[7] Khuroo MS, Datta DV. Budd–Chiari syndrome following preg-nancy. Report of 16 cases, with roentgenologic, hemodynamic andhistologic studies of the hepatic outflow tract. Am J Med1980;68:113–121.

[8] Dilawari JB, Bambery P, Chawla Y, Kaur U, Bhusnurmath SR,Malhotra HS, et al. Hepatic outflow obstruction (Budd–Chiarisyndrome). Experience with 177 patients and a review of theliterature. Medicine (Baltimore) 1994;73:21–36.

[9] Powell-Jackson PR, Melia W, Canalese J, Pickford RB, PortmannB, Williams R. Budd–Chiari syndrome: clinical patterns andtherapy. Q J Med 1982;51:79–88.

[10] Vons C, Smadja C, Franco D, Valla D, Rueff B, Benhamou JP.Successful pregnancy after Budd–Chiari syndrome. Lancet1984;2:975.

[11] Marpeau L, Peraudeau-Daoud P, Castiel J, Pigne A, Parc R,Hecht Y, et al. Normal pregnancy after a Budd–Chiari syndrometreated by shunting: apropos of a case, review of the literature. JGynecol Obstet Biol Reprod (Paris) 1987;16:1045–1052.

[12] Huguet C, Deliere T, Ollvier JM, Levy VG. Budd–Chiarisyndrome with thrombosis of the inferior vena cava: long-termpatency of meso-caval and cavoatrial prosthetic bypass. Surgery1984;95:108–111.

[13] Coelho JC, Parolin MB, Matias JE. Successful twin pregnancyafter orthotopic liver transplantation. Arq Gastroenterol2002;39:246–247.

[14] Ramsey PS, Hay JE, Ramin KD. Successful pregnancy followingorthotopic liver transplantation for idiopathic Budd–Chiari syn-drome. J Matern Fetal Med 1998;7:235–237.

[15] Ouwendijk RJ, Koster JC, Wilson JH, Stibbe J, Lameris JS, VisserW, et al. Budd–Chiari syndrome in a young patient withanticardiolipin antibodies: need for prolonged anticoagulanttreatment. Gut 1994;35:1004–1006.

[16] Martinelli P, Maruotti GM, Coppola A, Agangi A, Paladini D,Grandone E, et al. Pregnancy in a woman with a history of Budd–Chiari syndrome treated by porto-systemic shunt, protein Cdeficiency and bicornuate uterus. Thromb Haemost2006;95:1033–1034.

[17] Murad SD, Valla DC, de Groen PC, Zeitoun G, Hopmans JA,Haagsma EB, et al. Determinants of survival and the effect ofportosystemic shunting in patients with Budd–Chiari syndrome.Hepatology 2004;39:500–508.

[18] Rautou PE, Plessier A, Bernuau J, Denninger MH, Moucari R,Valla D. Pregnancy: a risk factor for Budd–Chiari syndrome? Gut2009;58:606–608.

[19] Langlet P, Escolano S, Valla D, Coste-Zeitoun D, Denie C, MalletA, et al. Clinicopathological forms and prognostic index in Budd–Chiari syndrome. J Hepatol 2003;39:496–501.

[20] Plessier A, Sibert A, Consigny Y, Hakime A, Zappa M,Denninger MH, et al. Aiming at minimal invasiveness as atherapeutic strategy for Budd–Chiari syndrome. Hepatology2006;44:1308–1316.

[21] Bachet JB, Condat B, Hagege H, Plessier A, Consigny Y, BelghitiJ, et al. Long-term portosystemic shunt patency as a determinantof outcome in Budd–Chiari syndrome. J Hepatol 2007;46:60–68.

[22] Moucari R, Rautou PE, Cazals-Hatem D, Geara A, Bureau C,Consigny Y, et al. Hepatocellular carcinoma in Budd–Chiarisyndrome: characteristics and risk factors. Gut 2008;57:828–835.

[23] Denninger MH, Chait Y, Casadevall N, Hillaire S, Guillin MC,Bezeaud A, et al. Cause of portal or hepatic venous thrombosis inadults: the role of multiple concurrent factors. Hepatology2000;31:587–591.

[24] Valla D, Casadevall N, Lacombe C, Varet B, Goldwasser E,Franco D, et al. Primary myeloproliferative disorder and hepaticvein thrombosis. A prospective study of erythroid colony forma-tion in vitro in 20 patients with Budd–Chiari syndrome. AnnIntern Med 1985;103:329–334.

[25] Comp PC, Thurnau GR, Welsh J, Esmon CT. Functional andimmunologic protein S levels are decreased during pregnancy.Blood 1986;68:881–885.

[26] Gonzalez R, Alberca I, Vicente V. Protein C levels in latepregnancy, postpartum and in women on oral contraceptives.Thromb Res 1985;39:637–640.

[27] Valla D, Le MG, Poynard T, Zucman N, Rueff B, Benhamou JP.Risk of hepatic vein thrombosis in relation to recent use of oralcontraceptives. A case-control study. Gastroenterology1986;90:807–811.

[28] Bates SM, Greer IA, Pabinger I, Sofaer S, Hirsh J. Venousthromboembolism, thrombophilia, antithrombotic therapy, andpregnancy: American College of Chest Physicians Evidence-BasedClinical Practice Guidelines (8th Edition). Chest 2008;133:844S–886S.

[29] Zeitoun G, Escolano S, Hadengue A, Azar N, El Younsi M,Mallet A, et al. Outcome of Budd–Chiari syndrome: a multivar-iate analysis of factors related to survival including surgicalportosystemic shunting. Hepatology 1999;30:84–89.

[30] Krass I. Budd–Chiari syndrome after oral contraceptives. Br MedJ 1968;1:708.

[31] Eapen CE, Velissaris D, Heydtmann M, Gunson B, Olliff S, EliasE. Favourable medium term outcome following hepatic veinrecanalisation and/or transjugular intrahepatic portosystemicshunt for Budd–Chiari syndrome. Gut 2006;55:878–884.

[32] Greer IA. Thrombosis in pregnancy: maternal and fetal issues.Lancet 1999;353:1258–1265.

[33] d’Alteroche L, Perarnau JM, Perrotin F, Bacq Y. Portal hyper-tension and pregnancy. Gastroenterol Clin Biol 2008;32:541–546.

[34] Mills JL, Simpson JL, Driscoll SG, Jovanovic-Peterson L, VanAllen M, Aarons JH, et al. Incidence of spontaneous abortionamong normal women and insulin-dependent diabetic womenwhose pregnancies were identified within 21 days of conception. NEngl J Med 1988;319:1617–1623.

[35] Robertson L, Wu O, Langhorne P, Twaddle S, Clark P, LoweGD, et al. Thrombophilia in pregnancy: a systematic review. Br JHaematol 2006;132:171–196.

[36] Maternal age and outcome of pregnancy. N Engl J Med1990;323:413–416.