The use of certain atypical antipsychotics has been associated with metabolic disturbances. We have assessed the evolution of body weight and glycaemia during a 6-month randomized comparative trial of amisulpride and olanzapine. Three hundred and seventy-seven adult patients with schizophrenia were randomized to either amisulpride (200-800 mg/day) or olanzapine (5-20 mg/day) for 6 months. Body weight and fasting blood glucose were measured. Both treatments showed comparable antipsychotic activity. Weight gain over the study was significantly greater (P=0.0004) in the olanzapine group (3.9+/-5.3 kg) than in the amisulpride group (1.6+/-4.9 kg). Mean fasting blood glucose increased in the olanzapine group by 4.42 mg/dl to a mean maximum value (118+/-38 mg/dl). In the amisulpride group, mean glucose levels fell by 2.82 mg/dl. The difference between groups was significant at 2 (P=0.0066) and 6 months (P=0.017). One olanzapine-treated patient was diagnosed with diabetes. Metabolic changes in the amisulpride group were restricted to patients with high body mass index at inclusion. At doses that provide comparable control of psychosis, treatment with olanzapine was associated with greater increase in weight and blood glucose compared with amisulpride. This should be taken into account in assessing risk-benefit of treatment options in schizophrenia.