Antipsychotic-induced weight gain is an important issue in the treatment of psychotic illnesses, and affects 80% of individuals being treated with antipsychotic drugs. However, the true dimension of weight gain and many accepted 'facts' in this area remain unclear as most research has been conducted in short-term trials and has included individuals receiving prolonged antipsychotic treatment. This review aims to systematically and critically review the evidence on weight gain induced by the two leading second-generation antipsychotics (olanzapine and risperidone) and the most widely researched first-generation antipsychotic (haloperidol) in patients with chronic and first-episode psychotic disorders. Weight gain was 3- to 4-fold greater in studies that included young patients with limited previous exposure to antipsychotic agents in both short-term studies (7.1-9.2 kg for olanzapine, 4.0-5.6 kg for risperidone and 2.6-3.8 kg for haloperidol vs 1.8-5.4 kg, 1.0-2.3 kg and 0.01-1.4 kg, respectively, in studies that included patients with chronic psychotic disorders) and long-term trials (10.2-15.4 kg for olanzapine, 6.6-8.9 kg for risperidone and 4.0-9.7 kg for haloperidol vs 2.0-6.2 kg, 0.4-3.9 kg and -0.7 to 0.4 kg, respectively). The same disparity was observed regarding the proportion of patients increasing their baseline weight by > or =7% (the cut-off for clinically significant weight gain). Recent studies carried out in young patients with first-episode psychosis (FEP), along with methodological artefacts in studies of chronic populations, suggest that the magnitude of weight gain reported by much of the literature could in fact be an underestimation of the true magnitude of this adverse effect. Although antipsychotics present idiosyncratic patterns of weight gain, they may also generate similar absolute gains.