Global benefit-risk assessment of antidepressants: venlafaxine XR and fluoxetine

Abstract

Background: Quantifying efficacy and safety differences between drugs is difficult because rigorous statistical methods to assess benefit and risk simultaneously are lacking.

Methods: Global benefit-risk (GBR) analysis of clinical trial data was used retrospectively to compare venlafaxine extended release (XR) and fluoxetine. Of 301 outpatients with moderate to severe depression given venlafaxine XR 75-225 mg/day (n=100), fluoxetine 20-60 mg/day (n=103), or placebo (n=98) for up to 8 weeks, 295 qualified for analysis. Primary efficacy variables were Hamilton Rating Scale for Depression (HAM-D) remission (final on-therapy score <or=7) and Clinical Global Impressions (CGI) response (very much or much improved).

Results: Using remission as outcome, the relative gain of venlafaxine XR was 1.78 vs. fluoxetine (P<0.01) and 2.04 vs. placebo (P<or=0.005). HAM-D remission benefit exceeded risk for significantly more patients given venlafaxine XR (34%) than with fluoxetine (19%, P=0.023) or placebo (19%, P=0.017). The odds ratios [95% confidence interval (CI)] that patients would be in the benefit > risk category were 2.1 (1.1-4.0) and 2.2 (1.1-4.3) for venlafaxine XR vs. fluoxetine and placebo, respectively. For CGI response, relative gains of venlafaxine XR were 1.39 (P<0.01) and 1.45 (P<0.01) vs. fluoxetine and placebo; benefit exceeded risk in 66, 53, and 52% of patients given venlafaxine XR, fluoxetine, and placebo (P=0.041 vs. venlafaxine XR), respectively.

Conclusions: GBR analysis can be applied to a wide array of efficacy and safety data to form statistical tests of clinically meaningful treatment comparisons. In this comparison, the GBR assessments on response and remission significantly favored venlafaxine XR.