Depression-free days as a summary measure of the temporal pattern of response and remission in the treatment of major depression: a comparison of venlafaxine, selective serotonin reuptake inhibitors, and placebo.

Abstract

This article develops and applies depression-free days (DFDs) as a summary measure of the temporal pattern of response and remission in a comparison of venlafaxine (a dual-action serotonin-norepinephrine reuptake inhibitor) with selective serotonin reuptake inhibitors (SSRIs) and placebo. Weekly data on the 17-item Hamilton Rating Scale for Depression (HAM-D-17) from 2046 patients with DSM-III-R/IV-established moderate-to-severe major depression, participating in 1 of 8 randomized, double-blind, controlled studies that compared venlafaxine with an SSRI (fluoxetine, paroxetine, or fluvoxamine) or with both placebo and an SSRI, were used to estimate DFDs. Maximum DFDs were imputed to maintained HAM-D-17 scores < or = 7 (asymptomatic depression) over time, minimum DFDs to persistent HAM-D-17 scores > or = 15 (acutely symptomatic depression), and prorated DFDs to intermediate HAM-D-17 scores. A secondary construct was developed to test sensitivity to a less stringent upper threshold of acutely symptomatic depression (HAM-D-17 score > or = 22). Using a tertiary construct, sensitivity to a more stringent lower threshold representing elimination of residual symptoms was also evaluated. The construct validity of the primary and the secondary DFDs measures was assessed in terms of their correlation with sustained low clinical global severity of illness (scores of 1 or 2 on the Clinical Global Impressions-Severity of Illness scale). for each construct, DFDs were compared across the 3 treatment groups and corresponding effect sizes were generated. Overall, sustained low clinical global severity of illness was associated with 38.3 median (interquartile range, 29.8 to 44.2) DFDs relative to 5.7 (interquartile range, 0 to 20.6) median DFDs associated with nonsustained low clinical global severity; similar differences emerged in terms of sustained asymptomatic depression. The venlafaxine group (N = 851) experienced a median of 18.8 (interquartile range, 0.4 to 34.6) DFDs compared with a median of 13.6 (interquartile range, 0 to 29.8) DFDs in the SSRI group (N = 749) and 7.4 (interquartile range, 0 to 26.2) DFDs in the placebo group (N = 446) (p <.0001 overall; venlafaxine vs. SSRIs, p =.0015, effect size = 0.2; venlafaxine vs. placebo, p <.0001, effect size = 0.4; and SSRIs vs. placebo, p =.0007, effect size = 0.2). The secondary and tertiary DFDs constructs yielded similar, albeit narrower, differences in all comparisons. The construct of DFDs was found to be a useful summary measure of sustained remission. Active treatments were associated with more DFDs than placebo, and venlafaxine with more DFDs than SSRIs, consistent with corresponding differences in sustained remission.