Dose-Response Meta-Analysis of Antipsychotic Drugs for Acute Schizophrenia

Objective: The dose-response relationships of antipsychotic drugs for schizophrenia are not well defined, but such information would be important for decision making by clinicians. The authors sought to fill this gap by conducting dose-response meta-analyses. Methods: A search of multiple electronic databases (through November 2018) was conducted for all placebo-controlled dose-finding studies for 20 second-generation antipsychotic drugs and haloperidol (oral and long-acting injectable, LAI) in people with acute schizophrenia symptoms. Dose-response curves were constructed with random-effects dose-response meta-analyses and a spline model. The outcome measure was total score reduction from baseline on the Positive and Negative Syndrome Scale or the Brief Psychiatric Rating Scale. The authors identified 95% effective doses, explored whether higher or lower doses than the currently licensed ones might be more appropriate, and derived dose equivalencies from the 95% effective doses. Results: Sixty-eight studies met the inclusion criteria. The 95% effective doses and the doses equivalent to 1 mg of oral risperidone, respectively, were as follows: amisulpride for patients with positive symptoms, 537 mg/day and 85.8 mg; aripiprazole, 11.5 mg/day and 1.8 mg; aripiprazole LAI (lauroxil), 463 mg every 4 weeks and 264 mg; asenapine, 15.0 mg/day and 2.4 mg; brexpiprazole, 3.36 mg/day and 0.54 mg; haloperidol, 6.3 mg/day and 1.01 mg; iloperidone, 20.13 mg/day and 3.2 mg; lurasidone, 147 mg/day and 23.5 mg; olanzapine, 15.2 mg/day and 2.4 mg; olanzapine LAI, 277 mg every 2 weeks and 3.2 mg; paliperidone, 13.4 mg/day and 2.1 mg; paliperidone LAI, 120 mg every 4 weeks and 1.53 mg; quetiapine, 482 mg/day and 77 mg; risperidone, 6.3 mg/day and 1 mg; risperidone LAI, 36.6 mg every 2 weeks and 0.42 mg; sertindole, 22.5 mg/day and 3.6 mg; and ziprasidone, 186 mg/day and 30 mg. For amisulpride and olanzapine, specific data for patients with predominant negative symptoms were available. The authors have made available on their web site a spreadsheet with this method and other updated methods that can be used to estimate dose equivalencies in practice. Conclusions: In chronic schizophrenia patients with acute exacerbations, doses higher than the identified 95% effective doses may on average not provide more efficacy. For some drugs, higher than currently licensed doses might be tested in further trials, because their dose-response curves did not plateau.