A Cochrane review included 15 studies with a total of 2 022 subjects. Compared to placebo, haloperidol was more effective at reducing manic symptoms, both as monotherapy (2 trials, 482 participants, WMD –5.85, 95% CI –7.69 to –4.00) and as adjunctive treatment to lithium or valproate (1 trial, 97 participants, WMD –5.20, 95% CI –9.26 to –1.14), and a smaller proportion of failed to complete treatment (2 trials, 484 participants, RR 0.74, 95% Cl 0.57 to 0.96). No significant differences between haloperidol and risperidone, olanzapine, carbamazepine, valproate, sultopride and zuclopentixol were found, but haloperidol was less effective than aripiprazole at week 12 (1 trial, 344 participants RR 1.45, 95% CI 1.22 to 1.73). Haloperidol was associated with less weight gain than olanzapine (RR 0.28, 95% CI 0.12 to 0.67), but with a higher incidence of tremor (RR 3.01, 95% CI 1.55 to 5.84) and other movement disorders.
Comment: The level of evidence is downgraded by the small total sample size and low quality of reporting of the included trials.