Despite known metabolic effects of second-generation antipsychotics (SGAs) on children and adolescents, comparative effects in youth with different diagnoses remain underreported. We compared differences in metabolic changes three months after starting treatment with SGAs in youth with bipolar disorder and with other psychotic and nonpsychotic disorders. Weight and metabolic differences among diagnostic groups before and three months after starting treatment with SGAs were compared in a naturalistic cohort of children and adolescents (14.9 +/- 3.0 years) diagnosed with bipolar disorder (n = 31), other psychotic disorders (n = 29), and other nonpsychotic disorders (n = 30), with no (35.6%) or very little (6.6 +/- 9.0 days) previous exposure to antipsychotics. Composite measurements of significant weight gain [weight increase > or = 5% at three months or increase > or = 0.5 in body mass index (BMI) z-score] and 'risk for adverse health outcome' (> or = 95(th) BMI percentile, or > or = 85(th) BMI percentile plus presence of one other obesity-related complication) were included. SGAs (risperidone, olanzapine, and quetiapine) were prescribed in comparable proportion among groups. Baseline weight and metabolic indices were not significantly different among diagnoses. Three months after starting treatment with SGAs, more than 70% patients had significant weight gain, BMI z-score increased in all diagnostic groups (p < 0.001 for all comparisons), total cholesterol increased in the bipolar (p = 0.02) and psychotic (p = 0.01) disorder groups, low-density lipoprotein cholesterol increased in the bipolar group (p = 0.02), and free T4 decreased in the psychotic disorder group (p = 0.05). More patients with bipolar disorder presented overweight plus > or = 1 obesity-related complication at follow-up. There are early weight gain and metabolic changes across diagnoses in youth treated with SGAs.