We included three studies (total n=289; 2 RCTs zotepine vs clozapine; 1 RCT zotepine vs clozapine vs risperidone (at 4 mg, 8 mg doses) vs remoxipride. All studies were of limited methodological quality. When zotepine was compared with clozapine, it was clozapine that was found to be more effective in terms of global state (n=59, 1 RCT, RR No clinically significant response 8.23 CI 1.14 to 59.17). Mental state scores also favoured clozapine (n=59, 1 RCT, MD average score (BPRS total, high = poor) 6.00 CI 2.17 to 9.83) and there was less use of antiparkinson medication in the clozapine group (n=116, 2 RCTs, RR 20.96 CI 2.89 to 151.90). In the comparison of zotepine and risperidone, mental state scoring found no significant difference between the groups (vs 4 mg: n=40, 1 RCT, MD average endpoint score (BPRS total, high=poor) 1.40 CI -9.82 to 12.62; vs 8 mg: n=40, 1 RCT, MD -1.30 CI -12.95 to 10.35) and use of antiparkinson medication was equivocal (vs 4 mg: n=40, 1 RCT, MD 1.80 CI -0.64 to 4.24; vs 8 mg: n=40, 1 RCT, MD 2.50 CI -0.05 to 5.05). Finally, when zotepine was compared with remoxipride, again no effect was found for mental state (n=58, 1 RCT, MD average endpoint score (BPRS total, high=poor) 5.70 CI -4.13 to 15.53) and there was no significant difference between the two groups in terms of use of antiparkinson medication (n=49, 1 RCT, RR 0.97 CI 0.41 to 2.29).
Data on important other outcomes such as other adverse events, service use or satisfaction with care, quality of life were not available.
