Clozapine is the most effective antipsychotic for patients with treatment-refractory schizophrenia, but many adverse effects are noted. Clinicians usually hesitate to switch from clozapine to other antipsychotics because of the risk of a re-emergence or worsening of the psychosis, although empirical studies are very limited. Zotepine, an atypical antipsychotic with a pharmacologic profile similar to clozapine, was found to be an effective treatment for patients with treatment-resistant schizophrenia in Japan. This 12-week study is the first prospective, randomized, and rater-blind study to investigate the efficacy and tolerability of switching from clozapine to zotepine. Fifty-nine patients with schizophrenia, who had taken clozapine for at least 6 months with a Clinical Global Impression-Severity score of at least 3, were randomly allocated to the zotepine and the clozapine groups. At the end of the study, 52 patients (88%) had completed the trial. The 7 withdrawal cases were all in the zotepine group. The final mean (SD) dose of zotepine and clozapine was 397.1 (75.7) versus 377.1 (62.5) mg/d, respectively. Patients in the zotepine group showed a significant increase in the Brief Psychiatric Rating Scale [mean (SD), 4.7 (8.7) vs -1.3 (6.3); P = 0.005], more general adverse effects as revealed by the Udvalg for Kliniske Undersogelser Rating Scale [mean (SD), 1.74 (3.9) vs -0.2 (2.8); P = 0.039], more extrapyramidal adverse effects as demonstrated by the Simpson and Angus Scale [mean (SD), 1.29 (3.5) vs 0.17 (2.1); P = 0.022], an increased use of propranolol (37.1% vs 0%, P < 0.0001) and anticholinergics (25.7% vs 0%, P = 0.008), and an increased level of prolactin (29.6 vs -3.8 ng/ mL, P < 0.0005), compared with the clozapine group. The results suggested that switching from clozapine to zotepine treatment should be done with caution.