OBJECTIVES: To summarize rates of clinical remission and dropouts/withdrawal due to adverse drug reactions (ADRs) or lack of efficacy/effectiveness (LoE) of serotonin-norepinephrine reuptake inhibitors (SNRIs), selective serotonin reuptake inhibitors (SSRIs) and tricyclic antidepressants (TCAs). METHODS: We searched Medline, Embase, International Pharmaceutical Abstracts, and The Cochrane International Library from 1980 to 2005 for clinical trials using MeSH terms “serotonin norepinephrine reuptake inhibitors”, “selective serotonin reuptake inhibitors”, “tricyclic antidepressants”, “major depression”, and “clinical remission”. Accepted languages were English, French, German, Portuguese, and Spanish. A meta-analytic approach was used to synthesize outcome rates from published head-to-head clinical trials comparing two or more drugs (in therapeutic doses) from SNRIs, and/or SSRIs, and/or TCAs from 6 to 12 weeks of treatment. Remission was defined as a final score =<7 on the Hamilton Depression Rating Scale or =<12 in the Montgomery-Asberg Depression Rating Scale. Data were combined across arms of individual studies using a random effects model, producing point estimates with 95% confidence intervals. RESULTS: Data were gathered from 30 arms of 15 head-to-head trials of 2524 patients. TCAs had the highest overall remission rate (45.7%), followed by SNRIs (45.0%), and SSRIs (38.8%) (p>0.05 for TCAs versus SNRIs; p<0.001 for TCAs versus SSRIs; and p<0.001 for SNRIs versus SSRIs). When patients were categorized as inpatients (n=582) and outpatients (n=1613), SNRIs had the highest remission rate (52.0% for 144 inpatients and 49.3% for 559 outpatients). SNRIs had the lowest overall discontinuation rates (24.3%), followed by SSRIs (27.1%), and TCAs (34.9%). Rates of discontinuation due to ADRs and LoE were 9.8% and 5.2% for SNRIs, 7.3% and 7.0% for SSRIs, and 18.9% and 9.1% for TCAs, respectively. CONCLUSION: SNRIs have the highest efficacy/effectiveness remission rates (statistically significant for inpatients and outpatients), and the lowest overall dropout/withdrawal rates suggesting clinical superiority for this class in treating major depressive disorders.