An integrated analysis of liver safety data from orlistat clinical trials.

OBJECTIVE: Orlistat is an oral gastrointestinal lipase inhibitor and is indicated for treatment of obesity in combination with a hypocaloric diet. Post-marketing reports of adverse reactions revealed hints for possible drug-induced liver injury which has prompted changes to the product information. Orlistat's development program, involving over 30,000 patients, did not indicate a hepatic safety issue. METHODS: We analyzed liver function test data from randomized clinical trials of orlistat, using i) meta-analysis of published study safety data, ii) time-to-event analysis for individual patients, and iii) a novel and more sensitive method derived from the U.S. Food and Drug Administration's (FDA) evaluation of drug-induced serious hepatotoxicity (eDISH) technique. Over 10,000 subjects were included. RESULTS: The combined odds ratio from a simple summary-level fixed-effects meta-analysis of treatment-emergent abnormalities in serum alanine aminotransferase (ALT) (defined as greater than the upper level of normal for 2 successive measurements) was 1.09 (95% CI 0.93-1.28), and in total bilirubin 1.24 (95% CI 1.03-1.49). Part of the small apparent effect was due to longer exposure to orlistat than to placebo, on average. A patient-level display, adjusting for regression towards the mean, and Kaplan-Meier analysis of changes in ALT and bilirubin, taking account of different exposure, showed no significant difference between orlistat and placebo. This shows that there is no signal for hepatic damage in clinical studies of orlistat. CONCLUSION: While idiosyncratic liver injury following exposure to orlistat cannot be excluded, it is likely to be extremely rare.