Fifteen studies with 16 contrasts of antipsychotic drugs with placebo, and involving a total of 5,387 patients, were included in the review.
Death.
The pooled OR for all 15 trials was 1.54 (95% confidence interval, CI: 1.06, 2.23, P=0.02) in favour of placebo, with no significant heterogeneity among the studies.
In trials (n=3) using aripiprazole versus placebo, the risk difference was 0.01 (95% CI: -0.01, 0.03, P=0.20) in favour of placebo.
In trials (n=5) using olanzapine versus placebo, the risk difference was 0.01 (95% CI: -0.00, 0.07) in favour of placebo.
In trials (n=3) using quetiapine versus placebo, the risk difference was 0.02 (95% CI: -0.01, 0.05, P=0.22) in favour of placebo.
In trials (n=5) using risperidone, the risk difference was 0.01 (95% CI: -0.01, 0.02, P=0.33) in favour of placebo.
Drop-outs.
There was no significant difference in drop-outs between the drug and placebo groups, although there was significant heterogeneity among the trials and between drugs.
Other analyses.
The ad hoc analysis of death rates by length of exposure favoured placebo (pooled RR 1.65, 95% CI: 1.19, 2.29). The sensitivity analyses found no significant differences between groups.