The pseudoscience of lithium and suicide: Reanalysis of a misleading meta-analysis.

This article critically examines a recent meta-analysis that concluded lithium is ineffective in preventing completed suicide, identifying three key biases: inclusion of large studies designed to exclude suicidality (yielding zero suicide events), arbitrary exclusion of pre-2000 trials that showed lithium's benefit, and underreporting of suicide deaths in a recent placebo group. By reanalyzing the data with standard meta-analytic practices—excluding zero-event studies, including older trials, and accounting for all probable suicides—the article finds a statistically significant protective effect of lithium against suicide, with odds ratios ranging from 0.13 to 0.25 favoring lithium. The critique highlights how selective inclusion criteria and data handling can produce misleading conclusions, emphasizing that the totality of randomized clinical trial evidence supports lithium's anti-suicide efficacy. It also discusses the challenges of studying completed suicide as a rare outcome and warns against the misuse of meta-analysis as pseudoscience when it serves to confirm preconceived beliefs rather than objectively evaluate evidence.