Development and validation of a scoring system to predict response to obeticholic acid in primary biliary cholangitis

Background & aims: Obeticholic acid (OCA) is the only licensed second-line therapy for primary biliary cholangitis (PBC). With novel therapeutics in advanced development, clinical tools are needed to tailor the treatment algorithm. We aimed to derive and externally validate the OCA response score (ORS) for predicting the response probability of individuals with PBC to OCA. Methods: We used data from the Italian RECAPITULATE (N = 441) and the IBER-PBC (N = 244) OCA real-world prospective cohorts to derive/validate a score including widely available variables obtained either pre-treatment (ORS) or also after 6 months of treatment (ORS+). Multivariable Cox regressions with backward selection were applied to obtain parsimonious predictive models. The predicted outcomes were biochemical response according to POISE (alkaline phosphatase [ALP]/upper limit of normal [ULN]<1.67 with a reduction of at least 15%, and normal bilirubin), or ALP/ULN<1.67, or Normal range criteria (NR: normal ALP, alanine aminotransferase [ALT], and bilirubin) up to 24 months. Results: Depending on the response criteria, ORS included age, pruritus, cirrhosis, ALP/ULN, ALT/ULN, GGT/ULN, and bilirubin. ORS+ also included ALP/ULN and bilirubin after 6 months of OCA therapy. Internally validated c-statistics for ORS were 0.75, 0.78, and 0.72 for POISE, ALP/ULN<1.67, and NR response, which raised to 0.83, 0.88, and 0.81 with ORS+, respectively. The respective performances in validation were 0.70, 0.72, and 0.71 for ORS and 0.80, 0.84, and 0.78 for ORS+. Results were consistent across groups with mild/severe disease. Conclusions: We developed and externally validated a scoring system capable to predict OCA response according to different criteria. This tool will enhance a stratified second-line therapy model to streamline standard care and trial delivery in PBC.