Strengths and limitations
To the best of our knowledge, this is the first systematic review and meta-analysis on this topic. Previous reviews of the literature studied considerably fewer pregnancies, were not available in English or were no systematic reviews 28-32. To ensure high quality of evidence and representativeness of the pregnant population with liver cirrhosis, we excluded case reports and case series.
Although in individual studies the quality of selection of cohorts and controls was generally high, selection bias in our study could exist, since included studies used different inclusion criteria. For example, some studies18, 24 excluded decompensated cirrhosis, leading to an underestimation of reported outcomes.
Before sensitivity analysis the meta-analysis showed significant heterogeneity between studies (reasons are mentioned below). The high heterogeneity of the analysis of cesarean section (I268%) cannot be only attributed to differences in known patient characteristics, as is demonstrated by the inability of our sensitivity analyses to lower heterogeneity to an acceptable level. Overall differences in clinical management differences in cesarean section rates between countries are more likely to underlie the heterogeneity. Specifically there may be distinct regional differences in the management of delivery in liver cirrhosis, where some local policies may favour cesarean section to prevent variceal hemorrhage during delivery, while other policies may favor vaginal delivery in order to avoid perioperative risks involved in abdominal surgery. 3, 33
In our meta-analysis it was not possible to perform subgroup-analyses based on diagnosis underlying cirrhosis or severity of cirrhosis, due to the small number of events and missing information in included studies, which could in future studies further allow individualized counseling and management.