Long-term effectiveness, safety, and liver stiffness dynamics of PBC treatment with obeticholic acid in real-world

Abstract

Background & aims: Several studies have assessed the short-term effectiveness and safety of obeticholic acid (OCA) in the real-world setting. We aimed to extend knowledge on the real-world effectiveness and safety of OCA treatment by expanding sample size and follow-up, and by exploring changes in liver stiffness measurement (LSM) over time.

Methods: The RECAPITULATE project involves centres belonging to the "Italian PBC registry" and/or the "Club Epatologi Ospedalieri" PBC working group. Effectiveness was evaluated as biochemical response according to POISE and normal range (NR) criteria (normal alkaline phosphatase/alanine aminotransferase/bilirubin). Safety was assessed as the incidence of de novo/worsening pruritus and discontinuation rate/causes. Available LSMs were also captured.

Results: We included 747 patients from 66 Italian centres: mean age 58 years; female/male 88%/14%; median follow-up 24 months [IQR 12-42]; 28% with cirrhosis, and 14% with autoimmune hepatitis (AIH)/PBC overlap syndrome. Probabilities of POISE and NR response increased from baseline to 57% and 20%, respectively, by the 42^nd month. The probabilities of response were lower in patients with cirrhosis (p = 0.02 and p = 0.004 for POISE and NR), but not different between patients with AIH/PBC and pure PBC (p = 0.8). Overall, 130 patients (17%) discontinued treatment, mainly due to pruritus (36.9%), while 28.5% did so after developing hepatic events. The discontinuation rate was higher in patients with cirrhosis (p <0.001). LSM was available in 573 patients (∼77%), of whom 255 had multiple measurements. LSM variation over time differed based on the attainment of POISE biochemical response (expected mean annual variation -0.48 [-0.78, -0.19] in responders vs. +0.33 [-0.07, 0.73] in non-responders, respectively, p <0.001).

Conclusions: Our findings confirm the effectiveness and safety profiles of OCA in the medium/long term and demonstrate that biochemical response is associated with the change in LSM over time.

Impact and implications: After the conditional approval of OCA for the treatment of PBC, the main confirmatory study failed to demonstrate OCA's ability to reduce liver-related events, leading the EMA to revoke the drug's marketing authorization. The ensuing scientific debate highlights an urgent need for further evidence from real-world practice. In the largest real-world series of patients treated with OCA to date, we confirm that the drug's effectiveness and safety profiles are maintained over a medium-to-long follow-up period. Valuable data for the management of the drug in relevant subgroups of patients, such as those with cirrhosis and autoimmune hepatitis/PBC overlap syndrome, are also provided. Our original results on liver stiffness measurement variation over time suggest a favourable impact of OCA on fibrosis progression, particularly in patients achieving a biochemical response to the drug. Overall, these data provide important insights for clinicians managing patients with PBC and contribute to the ongoing scientific debate about the effectiveness/safety profile of this drug.

Keywords: cirrhosis; liver stiffness measurement; obeticholic acid; overlap syndrome; primary biliary cholangitis; real-world evidence.

Graphical abstract

Fig. 1

Probability of response to OCA treatment over time in the overall cohort (n = 747) according to the POISE and the normal range criteria. Numbers over the bars are the cumulative numbers of individuals responding according to POISE/NR criteria along with Aalen-Johansen estimates of the cumulative incidence over time, taking into account the competing risk of OCA discontinuation. NR, normal range; OCA, obeticholic acid.

Fig. 2

Variation of liver biochemistry over time (ALP, ALT, GGT, bilirubin). Y-axis values refer to ULN for ALP, ALT, GGT and to mg/dl for total bilirubin. The number over the x-axis represents the number of patients with available measurements. Points and ranges represent mean values +/- SD at each time point. p values for the variation of biochemical parameters over time were estimated with linear mixed models. Levels of significance p <0.05. ALP, alkaline phosphatase; ALT, alanine aminotransferase; GGT, gamma-glutamyltransferase; ULN, upper limit of normal.

Fig. 3

Probability of response to OCA treatment over time stratified by the presence of cirrhosis. Numbers over the bars are the cumulative numbers of individuals responding according to POISE/NR along with Aalen-Johansen estimates of the cumulative incidence over time, considering the competing risk of OCA discontinuation. Differences of response probabilities between patients with or without cirrhosis were evaluated using the log rank test. Levels of significance p <0.05. OCA, obeticholic acid.

Fig. 4

Discontinuation rates. Discontinuation rate (A) in the entire cohort, (B) by cirrhosis status, and (C) by AIH/PBC overlap status. Numbers over the bars are the cumulative numbers of subjects discontinuing OCA along with Kaplan-Meier estimates of the cumulative incidence over time. Differences of discontinuation probabilities between patients with or without liver cirrhosis and with or without AIH/PBC overlap were evaluated with log rank test. Levels of significance p <0.05. OCA, obeticholic acid; AIH/PBC, autoimmune hepatitis/primary biliary cholangitis.

Fig. 5

On-treatment variation of LSM in individuals with PBC treated with OCA in the entire population and divided based on the response to the drug. The variation of LSM over time was evaluated with linear mixed models. A p for interactions was obtained from these models to investigate a differential time trajectory of LSM between patients achieving or not biochemical response. Level of significance p <0.05. LSM, liver stiffness measurement; OCA, obeticholic acid; PBC, primary biliary cholangitis.