In this analysis we examined studies of tricyclic antidepressants (TCAs) and selective serotonin reuptake inhibitors (SSRIs) to compare efficacy and drop-out rates. Frequency of reported side effects was also studied. Using Medline, we located 36 clinical trials of TCAs and SSRIs in a double-blind comparison. We performed a meta-analysis on these studies and on a subgroup of 21 studies that had more uniformly defined outcome criteria. The main outcome measures were efficacy for treatment completers and for the intention-to-treat group; drop-out rates due to adverse reactions and lack of efficacy; and reported side effects. Overall, the response rate to treatment for patients who completed a trial was 63.2% for SSRIs and 68.2% for TCAs (P = 0.038). For the intention-to-treat groups, these rates dropped to 48.0 and 48.6% (P, NS), respectively. Significantly more TCA-treated than SSRI-treated subjects dropped out due to either lack of efficacy or adverse reactions (30.0 vs. 24.7%, P = 0.01). Patients taking SSRIs experienced significantly more gastrointestinal problems and sexual dysfunction, whereas treatment with TCAs produced significantly more complaints of sedation, dizziness, and anticholinergic symptoms.