OBJECTIVES: To compare the efficacy of antipsychotics in the treatment of negative symptoms of schizophrenia. METHODS: Studies were identified by searching for randomized, double-blind, placebo-controlled trials reporting data on efficacy of antipsychotics on negative symptoms in schizophrenia. A structured literature review was conducted in the following databases: Pubmed, The Cochrane Central Register of Controlled Trials (CENTRAL), Proquest Health and Medical Complete, Science Citation Index Expanded, and Current Contents Connect. The outcome measure was the mean change in negative symptoms score from baseline to end-point on the PANSS-N, SANS, and BPRS retardation factor (BPRS-R) scales. As we obtained data from different assessment instruments to be combined, we chose a standardized statistic - Cohen's d - which is a type of standardized mean difference. The Der-Simonian and Laird random effects model was used to synthesize data. All calculations were made using STATA 8.2, with 95 % confidence intervals and significance level of 0.05. RESULTS: A total of 46 homogeneous trials (Q=45.18, df=50, p=0.667, I2=0%) were included; 4 with amisulpride, 3 with ziprasidone, 9 with olanzapine, 8 with risperidone, 11 with quetiapine, 8 with haloperidol, 2 with zotepine, and 1 with chlorpromazine. Ranked effect sizes (Cohen's d) favoring active treatment against placebo were for amisulpride; 0.52 (p<0.001), ziprasidone; 0.50 (p<0.001), olanzapine; 0.43 (p<0.001), risperidone; 0.40 (p<0.001), quetiapine; 0.36 (p<0.001), haloperidol; 0.34 (p<0.001), zotepine; 0.31 (p=0.11), and chlorpromazine; 0.12 (p=0.532). CONCLUSION: The efficacy of antipsychotics on negative symptoms seems to be drug-dependent. Amisulpride and ziprasidone showed better overall effects than the other antipsychotics.