While second-generation antipsychotics (SGAs) have had benefits over earlier antipsychotic treatments, their use has been associated with reports of weight gain. Body mass index (BMI) has been studied in clinical trials with limited comparison between drugs. To investigate the impact of each SGA on the risk of weight increase in an adult population. Using a national electronic medical records database, a naturalistic impact of SGAs on BMI was evaluated. Patients (aged >/=18 y) receiving a prescription for an antipsychotic drug between January 1995 and March 2004 were identified. An adverse event was defined as at least a 7% increase in BMI from baseline within one year of antipsychotic prescription and a post-increase BMI of at least 25 kg/m(2). A total of 9394 patients were identified, with 1514 cases of increased BMI after initial prescription. Risperidone (OR 1.39; 95% CI 1.16 to 1.66), quetiapine (OR 1.36; 95% CI 1.13 to 1.64), and olanzapine (OR 1.76; 95% CI 1.50 to 2.07) were significantly more likely to cause BMI increase compared with first-generation antipsychotics (FGAs). Aripiprazole (OR 0.72; 95% CI 0.36 to 1.46), ziprasidone (OR 0.68; 95% CI 0.39 to 1.18), and clozapine (OR 1.01; 95% CI 0.56 to 1.81) were less likely to induce weight gain compared with FGAs. This study provides a foundation for understanding how SGAs impact weight gain in a naturalistic, as opposed to a clinical trial, setting and provides evidence that there are differential risks of weight gain between SGAs. Because of negative long-term health effects of weight gain, physicians need to take all factors into consideration when recommending pharmaceutical therapy for patients with severe mental illness.