Comparison of Findings from Published Weight Loss Trials for Orlistat to the Findings Used by the Food and Drug Administration (FDA)

Abstract

OBJECTIVES: The objective was to compare differences in weight loss data presented in published orlistat studies on orlistat to their corresponding studies submitted to the FDA. METHODS: This meta-­‐analysis compared one-­‐year weight loss data reported in six published orlistat 120 mg studies to data reviewed by the FDA in the New Drug Application (NDA). The primary dependent variables were the percentage of subjects achieving 5% and 10% weight loss. Prior to analysis, weight loss data was stratified into placebo and orlistat groups. Potential for bias was assessed with a funnel plot and by calculating Kendall’s tau. The a priori alpha level was 0.05. RESULTS: Corresponding FDA reviews were located for 6 published orlistat trials. The pooled odds ratio of published vs. FDA 5%weight loss data for the placebo arm was 2.18 (95% CI: 1.83 to 2.60; p < 0.001) and 1.95 (95% CI: 1.70 to 2.24; p < 0.001) for the orlistat arm. The pooled odds ratio of published vs. FDA for 10% weight loss data for the placebo arm was 2.25 (95% CI: 1.74 to 2.91; p < 0.001) and 2.20 (95% CI: 1.88 to 2.57; p < 0.001) for the orlistat arm. The p-­‐values for Kendall’s tau for the 5% and 10% weight loss data were 0.054 and 0.34, respectively. CONCLUSIONS: Published orlistat trials presented 5% and 10% weight loss data that were twice of that reported in the FDA-­‐reviewed trials, and there was potential for bias in the 5% weight loss data.