Comprehensive analysis of remission (COMPARE) with venlafaxine versus SSRIs

Abstract

Background: To compare venlafaxine and selective serotonin reuptake inhibitors (SSRIs; fluoxetine, sertraline, paroxetine, fluvoxamine, and citalopram) in the treatment of depression.

Methods and materials: Meta-analysis of 34 randomized, double-blind studies identified by a worldwide search of all research sponsored by Wyeth Pharmaceuticals through January 2007. Patients were treated with venlafaxine (n = 4191; mean dose 151 mg/day) or SSRIs (n = 3621); nine studies also included a placebo control group (n = 932). The primary outcome measure was intent-to-treat (ITT) remission rates (Hamilton Rating Scale for Depression </=7) at week 8.

Results: The overall difference in ITT remission rates was 5.9% favoring venlafaxine (95% confidence interval [CI]: .038-.081; p < .001). Based on this difference, the number needed to treat (NNT) to benefit is 17 (95% CI: 12-26). In the nine placebo controlled studies, the drug-placebo differences were 6% (.02-.09) for the SSRIs and 13% (.09-.16) for venlafaxine. For the specific SSRIs, the difference versus fluoxetine (mean dose = 37 mg/day; 20 studies) was significant (6.6% [95% CI: .030-.095]); smaller differences versus paroxetine (mean dose = 25 mg/day; eight studies; 5%), sertraline (mean dose = 127 mg/day; three studies; 3%), and citalopram (mean dose = 38 mg/day; two studies; 4%) were not significant. Attrition rates due to adverse events were higher with venlafaxine than with SSRI therapy, 11% and 9% respectively (p = .0011).

Conclusions: These results indicate that venlafaxine therapy is statistically superior to SSRIs as a class, but only to fluoxetine individually. The clinical significance of this modest advantage seems limited for the broad grouping of major depressive disorder. Nonetheless, an NNT of 17 may be of public health relevance given the large number of patients treated for depression and the significant burden of illness associated with this disorder.