Meta-Analysis of Dropout Rates in SSRIs Versus Placebo in Randomized Clinical Trials of PTSD

Abstract

Meta-analysis was conducted to examine dropout predictors and differences between SSRIs and placebo in randomized clinical trials (RCTs) of PTSD. Studies systematically located were SSRI versus placebo double blind RCTs of PTSD DSM diagnosis published between 1991 and 2008. Fourteen RCTs (n = 2815) met the inclusion criteria with an average duration of 10.8 weeks. Dropout rates were: 331 of 1111 (29.8%) among placebo arm and 513 of 1704 (30.3%) among SSRI participants. Random effects modeling showed that the dropout rates of SSRIs and placebo did not differ (OR = 1.05, 95% CI = 0.82-1.34), although favored SSRIs among civilian traumas (OR = 2.52, 95% CI = 1.11-5.7). Mixed effects modeling showed dropout was predicted by mixed trauma in the placebo arms, and duration and mean dose across treatments. With the exception of civilian trauma, SSRIs dropout rates were slightly lower than those of placebo. Formulae are available to guide the prediction of dropout.