Antipsychotic drugs for the acute treatment of patients with a first episode of schizophrenia: a systematic review with pairwise and network meta-analyses.

Background: The first episode of schizophrenia is a pivotal phase of this debilitating illness. Which drug to use remains controversial without a summary of all direct or indirect comparisons of drugs. We did a systematic review with pairwise and network meta-analyses of efficacy and tolerability.

Methods: We searched MEDLINE, Embase, PsycINFO, Cochrane Library, PubMed, Biosis, and ClinicalTrials.gov for randomised controlled trials of antipsychotics for the acute treatment of first-episode schizophrenia, published up to Nov 17, 2016. Our primary outcome was overall change in symptoms. Secondary outcomes were change in positive and negative symptoms, categorical response to treatment, study dropout for any reason and for inefficacy of treatment, use of drugs to treat parkinsonian symptoms, weight gain, sedation, increase in prolactin release, overall functioning, and quality of life. We did the meta-analyses with a random-effects model to calculate standardised mean differences (SMDs) or odds ratios (ORs) with 95% CIs.

Findings: We identified 19 relevant randomised controlled trials of 12 antipsychotic drugs that involved 2669 participants. 13 of the studies presented data on the primary outcome. For overall reduction of symptoms, amisulpride (SMD -0·37, 95% CI -0·61 to -0·14), olanzapine (-0·25, -0·39 to -0·12), ziprasidone (-0·25, -0·48 to -0·01), and risperidone (-0·14, -0·27 to -0·01) were significantly more efficacious than haloperidol, but the evidence was very low to moderate quality. Amisulpride was superior for reduction of symptoms to quetiapine (SMD -0·25, 95% CI -0·50 to -0·01). Olanzapine was superior to haloperidol and risperidone for reduction of negative symptoms. Several second-generation antipsychotics were superior to haloperidol in terms of all-cause discontinuation. Olanzapine was associated with at least one use of drugs to treat parkinsonian symptoms and quetiapine with less akathisia than haloperidol, aripiprazole, risperidone, and olanzapine, but, again, evidence was very low to low quality. Molindone was superior to risperidone, haloperidol, and olanzapine in terms of weight gain, and superior to risperidone in terms of increase in prolactin release.

Interpreation: Haloperidol seems to be a suboptimum treatment option for acute treatment of first-episode schizophrenia, but we found little difference between second-generation antipsychotics. The evidence was generally of low quality and the numbers of patients for each drug were small. Thus, the choice of treatment should be guided primarily by side-effects.

Funding: German Federal Ministry of Education and Research.