Zwangerschapscholestase is een aandoening gekarakteriseerd door (intense) jeuk zonder huidafwijkingen (behalve krabeffecten) in combinatie met een verhoogde concentratie van random galzure zouten. Er bestaat een relatie tussen zwangerschapscholestase en het optreden van foetale complicaties zoals vroeggeboorte, meconium houdend vruchtwater en intra-uteriene vruchtdood (IUVD). Naast zwangerschapscholestase bestaat er ook cholestase in het kader van (pre-existente) leverziekte. In deze richtlijn wordt deze aandoening buiten beschouwing gelaten. Bij twijfel over de diagnose zwangerschapscholestase kan een consult bij de maag darm lever arts worden overwogen om de differentiaal diagnose behorend bij verhoogde galzure zouten nader uit te diepen.

Sinds de publicatie van de richtlijn cholestase in 2011 en modules 2018 zijn er een aantal relevante artikelen verschenen over dit onderwerp. Deze nieuwe artikelen roepen op tot beleidswijzigingen voor zwangeren met dit ziektebeeld. Omdat deze artikelen nog niet waren opgenomen in de landelijke richtlijn uit 2011 kan dit leiden tot praktijkvariatie. Deze praktijkvariatie kan mogelijk leiden tot substandard care.

In deze richtlijn wordt de behandeling van zwangerschapscholestase buiten beschouwing gelaten. Dit onderwerp wordt in een separate module behandeld.

Comparison induction of labor versus expectant management < 37 weeks gestational age.

Description of studies

Puljic (2015) conducted an observational cohort study for the optimal timing of labor in women with gestational cholestasis by reporting the risk of fetal and neonatal mortality for each week of gestational age. Data from 1,604,386 pregnant women were included. Based on the ICD-codes, it was determined that in 5,545 cases the pregnancy had been complicated by gestational cholestasis. Puljic (2015) reported the risk of fetal mortality per gestational week (calculated by the number of fetal mortality cases divided by the number of ongoing pregnancies in specific weeks) and the risk of neonatal death in the week following birth as a measure of risk of the birth. Furthermore, Puljic (2015) defined the risk of expectant management for a certain week (per 10,000 fetuses at risk) as the risk of fetal mortality in that week plus the risk of neonatal mortality in the following week (as a measure of risk of birth). Unfortunately, no further data were reported on the number of births per week and the number of fetal and neonatal mortality per week. The statistical analyzes are unclear.

Results

Foetal and neonatal mortality

Puljic (2015) reported that fetal mortality (per 10,000 ongoing pregnancies) between 34 and 40 weeks of gestation was higher in women with gestational cholestasis than in women without gestational cholestasis. From 36 weeks of gestation, the risk of death due to preterm birth (represented as neonatal deaths per 10,000 live births) in women with gestational cholestasis is lower than the risk of intra-uterine fetal demise. In other words,  the risk of expectant management from 36 weeks onwards is greater than the risks associated with preterm birth. For a comprehensive overview of the results, we refer to the evidence table.

Neonatal intensive care unit (NICU) admission

Puljic (2015) did not report the outcome NICU admission.

Duration of NICU admission

Puljic (2015) did not report the outcome duration of NICU admission.

Level of evidence of the literature

Because of the unclear/absent statistical analyzes, uncertainty about the number of occurrences of fetal and neonatal mortality and the number of women who gave birth at a given gestational age, it has been decided not to GRADE the level of evidence for the outcome measures 'fetal and neonatal mortality'.

The level of evidence for the outcome measures NICU admission and duration of admission could not be graded due to lack of data.

Comparison induction of labor versus expectant management > 37 weeks gestational age

Jain (2013) is a prospective randomized study that included 69 women with obstetric cholestasis. Group I (n=34) was planned for delivery at 37 weeks and in Group II (n=35) pregnancy was carried to 38 weeks under surveillance.

Foetal surveillance started at diagnosis (34 weeks or later) and included daily maternal record of foetal movements, and biophysical profile. The frequency of foetal monitoring was weekly before 36 weeks and biweekly after 36 weeks.  All study patients received treatment with ursodeoxycholic acid (300 mg twice daily). In patients with severe symptoms, the dose was further increased to 900 mg to a maximum of 1500 mg/day. Primary outcome of the study was correlation of foetal and neonatal complication rates to gestational age at delivery.

Chappel (2012) is a semifactorial design randomized controlled trial (PITCH study) including n=125 women from 9 maternity units in the UK. Women with intrahepatic cholestasis of pregnancy (pruritus and raised maternal serum bile acids) or pruritus and raised alanine transaminase (> 100 IU/L) were recruited after 24 weeks' gestation (n=111). Included women were primarily randomized in a therapeutic trial comparing UDCA (n=56) with placebo (n=55). A subgroup of women from this cohort and an additional 14 women not included in the therapeutic trial was subsequently randomized to the early term delivery group (n=30; induction or delivery commenced between 37 + 0 and 37 + 6) or to the group expectant management (n=33; spontaneous labor awaited until 40 weeks or caesarean section as indicated, usually after 39 weeks' gestation). Primary outcome for the timing of delivery comparison was caesarean section.

Foetal and neonatal mortality

Jain (2013) reported no stillbirths in the study. Jain (2013) did not report the outcome neonatal death in an intention to treat analysis, however the outcome was reported for gestational age at delivery (table 1).  

Table 4.1 Gestation at delivery in correlation to neonatal death (Jain, 2013)

Chappel (2012) reported no stillbirths or neonatal deaths in either group. The outcome live birth (until discharge from hospital) occurred in 100% in both groups. 

Neonatal intensive care unit (NICU) admission

Jain (2013) did not report the outcome NICU admission in an intention to treat analysis.

Chappel (2012) reported 3% (n/N = 1/30) NICU admissions in the group early term delivery and 6% (n/N=2/33) in the group expectant management (RR 0.54, 95%CI 0.05 to 5.72).

Duration of NICU admission

Jain (2013) did not report the outcome duration of NICU admission.   

Chappel (2012) reported the mean duration of admission to neonatal unit. Mean duration of admission was 2 days in the group early term delivery and 6.5 days (sd 0.7) in the group expectant management. 

Meconium-stained amniotic fluid

Chappel (2012) reported meconium-stained amniotic fluid in 10% (n/N=3/30) in the group early term delivery and 18.2% (n/N=6/33) in the group expectant management (RR 0.71 95%CI 0.23 to 2.17). 

Level of evidence of the literature

The level of evidence regarding the outcome measures mortality, NICU admission, and meconium-stained amniotic fluid was downgraded by 3 levels because of study limitations (risk of bias, no intention to treat analysis, Jain (2013)) and number of included patients (imprecision) to very low.

Results - PICO 3

Description of studies

For this research question, only studies were found that performed univariate analysis of determinants, without correction for patient level confounding factors. One meta-analysis (Ovadia, 2019) and three retrospective cohort studies published after the search of the meta-analysis (Celik, 2019; Juusela, 2019; Guszczynska-Losy, 2020) were included in the description of the literature.

Ovadia (2019) is an individual patient data meta-analyses that included studies reporting perinatal outcomes for women with intrahepatic cholestasis of pregnancy when serum bile acid concentrations were available. A systematic review of the literature was performed by searching PubMed, Web of Science, and Embase databases for studies published from database inception to June 1, 2018. Inclusion criteria were studies defining intrahepatic cholestasis of pregnancy based upon pruritus and elevated serum bile acid concentrations, with or without raised liver aminotransferase concentrations. Eligible studies were case-control, cohort, and population-based studies, and randomised controlled trials, with at least 30 participants, that reported bile acid concentrations and perinatal outcomes. The regression analyses were not adjusted for patient level confounders (such as coexistent pre-eclampsia and gestational diabetes).    

Ovadia (2019) reported, based on data from the systematic review, 45/542 stillbirths in the group women with intrahepatic cholestasis of pregnancy and 519/164.601 stillbirths in the control group (OR 1.46, 95% CI 0.73 to 2.89).

The association between stillbirth and peak TBA concentrations for singleton pregnancies was reported based on individual patient data associations between serum biochemistry and adverse perinatal outcome for singleton pregnancies. ROC curves (AUC, 95%CI) were reported to assess the association between bile acids and adverse perinatal outcome. For the outcome stillbirth an AUC was reported for bile acids: Stillbirth n/N (%) 25/4.269 (1%) with an AUC of 0.83 (95%CI 0.74-0.92).

To assess whether a threshold of total bile acid concentration associated with an increased

risk of stillbirth could be defined, Ovadia (2019) performed stepwise logistic regression analysis between bile acid categories at 20 µmol/L intervals. For singleton pregnancies, the prevalence of stillbirth was three (0.13%; 95% CI 0.02 to 0.38) of 2310 intrahepatic cholestasis of pregnancy cases in women with serum total bile acids of less than 40 µmol/L, versus four (0.28%; 0.08–0.72) of 1412 cases with total bile acids of 40–99 µmol/L (hazard ratio 2.35 95%CI 0.52 to 10.50), and versus 18 (3.44%; 2.05–5.37) of 524 cases for bile acids of 100 µmol/L or more (HR 30.50 [8.83–105.30]).

To establish whether risk of stillbirth in women with total bile acids of less than 100 µmol/L was increased compared with the background population risk, Ovadia (2019) used published data on the prevalence of national stillbirth as comparator groups in a weighted analysis. No increased stillbirth risk for women with singleton intrahepatic cholestasis of pregnancy were found in the IPD analysis with total bile acids of less than 40 µmol/L or 40–99 µmol/L when compared with the pooled national prevalence of stillbirth from 2000 (0.42%) or 2015 (0.33%).

ROC curves (AUC, 95%CI) were reported to assess the association between bile acids and adverse perinatal outcome. The outcome neonatal death was reported in 7/2888 (<1%) with an AUC of 0.62 (95%CI 0.38-0.86). However, no comparison for different concentrations of bile acids was reported.

ROC curves (AUC, 95%CI) were reported to assess the association between bile acids and adverse perinatal outcome. Area under the curve for NICU admission n/N= 789/4014 (20%) was reported (AUC 0.55 (95%CI 0.52-0.57). AUC was further reported for preterm birth (AUC 0.60 (0.58-0.63), meconium-stained amniotic fluid (AUC 0.62 (0.59-0.64), and Apgar score < 7 at 5 minutes (AUC 0.65 (0.58-0.71). No comparison for different levels of bile acids was reported.

Celik (2019) is a retrospective observational study including 370 patients diagnosed with ICP. The study evaluated patients in two groups according to serum bile acid level before UDCA (> 40 μmol/L or < 40 μmol/L).

Management of ICP:  timing of delivery in ICP patients was planned to be 37–38 weeks at the latest, immediate delivery with labour induction or caesarean section was planned in patients diagnosed with ICP after 38 weeks and later, while patients with severe cholestasis patients with higher bile acid levels (>40 μmol/L) were delivered at 34–37 weeks. Univariate logistic regression analysis was used to study the relation of cholestasis of pregnancy on perinatal outcomes. The study additionally reviewed foetal outcomes in ICP patients before or after 34 weeks gestation.

No cases of perinatal mortality were reported in the study in either group; n=0 in the group mild cholestasis (Total bile acids < 40 μmol/L), n=246) and n=0 in the group severe cholestasis (Total bile acids > 40 μmol/L, n=124).  

The outcome neonatal unit admission was reported in 15.4% (n/N= 38/246) in the group mild cholestasis (and 33% (n/N= 41/124) in the group severe cholestasis (OR 4.19, 95% CI 2.17 to 8.10). The outcome preterm delivery was reported in 13% in the group mild cholestasis and 51% in the group severe cholestasis (OR 4.67, 95% CI 2.79 to 7.82). The outcome meconium staining was reported in 7.7% in the group mild cholestasis and 25% in the group severe cholestasis (OR 3.98, 95% CI 2.14 to 7.40)

Juusela (2019) is a retrospective chart review of women with intrahepatic cholestasis of pregnancy (n=61). Inclusion criteria were women with singleton gestations with the diagnosis of ICP, with BA ≥10mol/L, and in whom diagnostic work-up ruled out other causes of liver disease. Exclusion criteria were women admitted with bile acids <10mol/L; with other hepatic diseases and diseases affecting liver function tests; multiple gestations; and those who had no follow-up, delivered at another institution, or incomplete medical records.

In none of the 61 patients diagnosed with ICP included in the study IUFD, stillbirth, abruption, or neonatal demise occurred.

Table 4.2 NICU admission, spontaneous preterm labour, respiratory distress syndrome, meconium stained amniotic fluid (Juusela, 2019)

Guszczynska-Losy (2020) is a retrospective observational study including patients with ICP including women with single, twin and triplet pregnancies (n=86). Patients were divided into 3 groups according to their fasting serum bile acid level (group I n = 60, 10 to 39.90 μmol/L; group II n = 20, 40-99.90 μmol /L; group III n = 6, TBA (total bile acids) ≥ 100.00 μmol/L).

The relation of serum TBA with adverse perinatal outcome was studied. In the examined group of pregnant women with cholestasis, 67 had a singleton pregnancy, 18 had a twin pregnancy and one had a triplet pregnancy.

Table 4.3 Perinatal outcomes of patients with mild and severe cholestasis (single, twin and triplet pregnancy)

Table 4.4 OR for adverse perinatal outcomes (TBA) in intrahepatic cholestasis of pregnancy in singleton pregnancy

*Reference category is TBA 0–39.9 μmol/L

As compared to women with intrahepatic cholestasis of pregnancy <40 μmol/L, in women with TBA > 40 μmol/L the study reported a higher rate of preterm births (OR 2.91, 95%CI 1.17 to 7.40), and neonatal admission to NICU (OR 2.60, 95%CI 1.06 to 6.38). No differences were reported in breathing disorders (OR 1.60, 95%CI 0.58 to 4.46) or presence of meconium in amniotic fluid (OR 1.70, 95%CI 0.33 to 6.47). No stillbirths were reported in either group (TBA <40 μmol/L, n/N 0/60; >40 μmol/L, n/N 0/26).

Conclusions

A systematic review of the literature was performed to answer the following questions:

PICO 1- In the decision for induction of labor in women with cholestasis, does the consideration of gestational age (stratified per week and in preterm versus term) and/or severity of cholestasis (bile acid concentration) reduce the risk of perinatal mortality and morbidity?

PICO 2 - What are the (un)favorable effects of induction of labor compared to expectant management in women with ICP? 

PICO 3 - Which factors are associated/related to unfavorable maternal and perinatal outcomes in women with intrahepatic cholestasis of pregnancy?

Relevant outcome measures

The guideline development group considered perinatal mortality and NICU admission as critical outcome measures for decision making; and APGAR score <7 at 5 minutes, meconium present in amniotic fluid, prematurity, and IRDS as important outcome measures for decision making.

A priori, the working group did not define the outcome measures listed above but used the definitions used in the studies.

The working group defined any statistical difference in perinatal mortality and maternal mortality as a minimal clinically (patient) important difference. For all other outcome measures, the GRADE default - RR<0.80 or RR>1.25 for dichotomous outcomes (Schünemann, 2013) and 0.5 standard deviation for continuous outcomes - was taken as a minimal clinically important difference.

Search and select (Methods)

The databases Medline (via OVID) and Embase (via Embase.com) were searched with relevant search terms until July 9^th, 2020. The detailed search strategy is depicted under the tab Methods. The systematic literature search resulted in 320 hits. Studies were selected based on the following criteria: systematic reviews (searched in at least two databases, detailed search strategy available and risk-of-bias assessment), randomized trials or observational studies in which patients with gestational cholestasis participated and in which a comparison was made between induction of labor or expectant management or different levels of bile acids. In addition, at least one of the above (relevant) outcome measures had to be included. The working group assessed the definition of pregnancy cholestasis used in the studies, level of bile acid was not an inclusion or exclusion criterium. 32 studies were initially selected based on title and abstract screening. After reading the full text, 25 studies were excluded (see the table with reasons for exclusion under the tab Methods), and 7 studies were included.

PICO 1.

No studies were included.

PICO 2

Three studies were included.

PICO 3

To study prognostic factors the preferred study design would be an RCT that assesses the effectiveness of a prognostic model (PICO 1). If there are no studies available with aforementioned design, studies are included that study prognostic factors by means of a multivariate analysis. Four studies were found that did not report a multivariate model describing the relation between prognostic factors and neonatal outcomes in pregnant women (singleton pregnancy) with cholestasis. However, these studies did report this association in univariate analyses. These studies are described below, in order to present an overview of less direct evidence. However, no GRADE-assessment of these studies is performed, since the minimum criteria for inclusion are not fulfilled.

Results - PICO 1

No studies were included.

Conclusion

Results- PICO 2

Three studies were included in the analysis of the literature. Important study characteristics and results are summarized in the evidence tables. The assessment of the risk of bias is summarized in the risk of bias tables.