Objective: This article reviews and compares the tolerability profiles of atypical antipsychotics in adults with schizophrenia or bipolar disorder in randomized controlled trials (RCTs). Methods: A systematic search of the BIOSIS, Cochrane Central Register of Controlled Trials, EMBASE, MEDLINE, and PsycINFO databases was conducted for English-language papers and abstracts that reported RCTs comparing >= 2 atypical antipsychotics (aripiprazole, olanzapine, quetiapine, risperidone, ziprasidone) in the treatment of adult patients with schizophrenia or bipolar disorder. The search was completed in December 2007. A panel of 4 psychiatrists with expertise in the treatment of schizophrenia, schizoaffective disorder, or bipolar disorder identified potentially relevant outcomes that would reflect patients' subjective perception of adverse effects during treatment with antipsychotic medications. These outcomes were daily activities, anxiety or depression, bodily anxiety or restlessness, dizziness or nausea, extrapyramidal symptoms (EPS), sexual dysfunction, stiffness or tremor, tiredness or weakness, and weight gain. Data on these outcomes were extracted from relevant studies, and mixed treatment comparisons were conducted using Bayesian Markov Chain Monte Carlo simulation. Summary effect estimates (odds ratios [ORs] and 95% credible intervals [95% CrIs]) were calculated and reported in comparison with risperidone. Results: The search identified 2963 potentially relevant publications, of which 50 provided data on 48 RCTs comparing >= 2 atypical antipsychotics. Outcomes that were statistically significant compared with risperidone at the 5% level were decreased bodily anxiety or restlessness with quetiapine (OR = 0.506; 95% CrI = 0.290, 0.789), decreased EPS with quetiapine (OR = 0.441; 95% CrI = 0.129, 0.910), increased weight gain with olanzapine (OR = 2.139; 95% CrI = 1.764, 2.626), and decreased weight gain with ziprasidone (OR = 0.466; 95% CrI = 0.31.7, 0.657). Conclusion: The atypical antipsychotics had mixed tolerability profiles in this mixed treatment comparison, with some agents having significantly greater tolerability than others depending on the outcome assessed. (Clin Ther. 2009;31[Theme Issue]: 1345-1359) (C) 2009 Excerpta Medica Inc.