Background: Intrahepatic cholestasis of pregnancy is associated with adverse perinatal outcomes, but the association with the concentration of specific biochemical markers is unclear. We aimed to quantify the adverse perinatal effects of intrahepatic cholestasis of pregnancy in women with increased serum bile acid concentrations and determine whether elevated bile acid concentrations were associated with the risk of stillbirth and preterm birth.

Methods: We did a systematic review by searching PubMed, Web of Science, and Embase databases for studies published from database inception to June 1, 2018, reporting perinatal outcomes for women with intrahepatic cholestasis of pregnancy when serum bile acid concentrations were available. 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, and that reported bile acid concentrations and perinatal outcomes. Studies at potential higher risk of reporter bias were excluded, including case reports, studies not comprising cohorts, or successive cases seen in a unit; we also excluded studies with high risk of bias from groups selected (eg, a subgroup of babies with poor outcomes were explicitly excluded), conference abstracts, and Letters to the Editor without clear peer review. We also included unpublished data from two UK hospitals. We did a random effects meta-analysis to determine risk of adverse perinatal outcomes. Aggregate data for maternal and perinatal outcomes were extracted from case-control studies, and individual patient data (IPD) were requested from study authors for all types of study (as no control group was required for the IPD analysis) to assess associations between biochemical markers and adverse outcomes using logistic and stepwise logistic regression. This study is registered with PROSPERO, number CRD42017069134.

Findings: We assessed 109 full-text articles, of which 23 studies were eligible for the aggregate data meta-analysis (5557 intrahepatic cholestasis of pregnancy cases and 165 136 controls), and 27 provided IPD (5269 intrahepatic cholestasis of pregnancy cases). Stillbirth occurred in 45 (0·83%) of 4936 intrahepatic cholestasis of pregnancy cases and 519 (0·32%) of 163 947 control pregnancies (odds ratio [OR] 1·46 [95% CI 0·73-2·89]; I²=59·8%). In singleton pregnancies, stillbirth was associated with maximum total bile acid concentration (area under the receiver operating characteristic curve [ROC AUC]) 0·83 [95% CI 0·74-0·92]), but not alanine aminotransferase (ROC AUC 0·46 [0·35-0·57]). For singleton pregnancies, the prevalence of stillbirth was three (0·13%; 95% CI 0·02-0·38) 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 [HR] 2·35 [95% CI 0·52-10·50]; p=0·26), 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]; p<0·0001).

Interpretation: The risk of stillbirth is increased in women with intrahepatic cholestasis of pregnancy and singleton pregnancies when serum bile acids concentrations are of 100 μmol/L or more. Because most women with intrahepatic cholestasis of pregnancy have bile acids below this concentration, they can probably be reassured that the risk of stillbirth is similar to that of pregnant women in the general population, provided repeat bile acid testing is done until delivery.

Funding: Tommy's, ICP Support, UK National Institute of Health Research, Wellcome Trust, and Genesis Research Trust.