Strengths and limitations:
A strength of this secondary analysis was the size of the original trial, which was considerably larger than any previous trial investigating the efficacy of UDCA in this population. The trial recruited women with a typical phenotype of ICP, with 24% of women having a baseline bile acid concentration ≥40 µmol/L, and the study findings are therefore likely to be generalisable to women with a similar clinical phenotype. The subgroups of baseline bile acid concentration analysed were based on clinically relevant thresholds that had previously demonstrated an association with differences in perinatal outcomes.2Interaction tests were used to test for differences in treatment effect, which limited the possibility that the effect size seen in any individual subgroup was over-interpreted. By understanding the size of a clinically meaningful reduction in itch severity (judged by women and clinicians), it was possible to consider whether or not any small differences identified were clinically relevant.
The original trial was not designed with a view to undertaking the present secondary analysis, and therefore no power calculation was undertaken to ensure that the study was sufficiently powered to find these effects. The study was limited by the size of certain subgroups, particularly when investigating women with baseline bile acid concentrations greater than 100 μmol/L (34 of 604 women). Although multiple analyses were undertaken, increasing the risk of false positive results, no significant differences were identified, mitigating the risk of false discovery.