Comparisons of antipsychotics with placebo can be biased by unblinding due to side effects. Therefore, this meta-analysis compared the efficacy of antipsychotics for acute schizophrenia in trials using barbiturates or benzodiazepines as active placebos.

Randomized controlled trials (RCTs) in acute schizophrenia with at least 3 weeks duration and comparing any antipsychotic with barbiturates or benzodiazepines were eligible. ClinicalTrials.gov, CENTRAL, EMBASE, MEDLINE, PsycINFO, PubMed, WHO-ICTRP as well as previous reviews were searched up to 9 January 2018. Two separate meta-analyses, one for barbiturates and one for benzodiazepines, were conducted using random-effects models. The primary outcome was response to treatment, and mean values of schizophrenia rating scales and dropouts were analyzed as secondary outcomes. This study is registered with PROSPERO (CRD42018086263).

Seven barbiturate-RCTs (number of participants n = 1736), and two benzodiazepine-RCTs (n = 76) were included in the analysis. The studies were published between 1960 and 1968 and involved mainly chronically ill patients. More patients on antipsychotics in comparison to barbiturates achieved a ‘good’ response (36.2% v. 16.8%; RR 2.15; 95% CI 1.36–3.41; I2 = 48.9) and ‘any’ response (57.4% v. 27.8%; RR 2.07; 95% CI 1.35–3.18; I2 = 68.2). In a single small trial (n = 60), there was no difference between antipsychotics and benzodiazepines on ‘any’ response (74.7% v. 65%; RR 1.15; 95% CI 0.82–1.62).

Antipsychotic drugs were more efficacious than barbiturates, based on a large sample size. Response ratios were similar to those observed in placebo-controlled trials. The results on benzodiazepines were inconclusive due to the small number of studies and participants.