Objective: To examine the efficacy and overall tolerability of augmentation of standard antidepressants with atypical antipsychotic agents for treatment-resistant major depressive disorder. Data Sources: MEDLINE/PubMed, EMBASE, the Cochrane database, and program syllabi from major psychiatric meetings held since 2000 as well as a number of online clinical trial results registries were searched. Makers of atypical antipsychotic agents who do not maintain an online clinical study results registry were contacted directly. Study Selection: Double-blind, randomized, placebo-controlled clinical trials assessing adjunctive treatment of standard antidepressants with an atypical antipsychotic agent for treatment-resistant major depressive disorder were identified. Data Extraction: Data were extracted with the use of a pre-coded form. Data Synthesis: Data from 10 clinical trial reports involving a total of 1500 outpatients with treatment-resistant major depressive disorder were identified and combined using a randomeffects model. Patients randomized to adjunctive treatment with an atypical antipsychotic agent were more likely to experience remission (risk ratio [RR] = 1.75, p < .0001) or clinical response (RR = 1.35, p = .001) than patients who received adjunctive placebo. Pooled remission and response rates for the 2 treatment groups were 47.4% vs. 22.3% and 57.2% vs. 35.4%, respectively. Although there was no difference in overall discontinuation rates (p = .929) or the rate of discontinuation due to inefficacy (p = .133) between the 2 treatment groups, the rate of discontinuation due to adverse events was lower among placebo-treated patients (RR = 3.38, p < .0001). Conclusions: These results support the utility of augmenting standard antidepressants with atypical antipsychotic agents for treatment-resistant major depressive disorder. An obvious limitation of this work is the absence of data focusing on the use of aripiprazole and ziprasidone. Future short- as well as long-term studies comparing the efficacy, safety, and tolerability of this versus other adjunctive strategies are warranted.