Antidepressant effect or bias? Systematic review and meta-analysis of studies using the forced swimming test.

Abstract

The forced swim test (FST) assesses antidepressant activity in rodents by measuring suppression of immobility. This study reviewed the literature to evaluate how experimental conditions, study quality, and bias influence antidepressant efficacy in the FST (PROSPERO: CRD42020200604). Systematic searches in Embase and MEDLINE (PubMed) identified 8247 relevant records. After being screened by two independent reviewers, 2588 records were included in the library. A random sample ( k  = 200) yielded 561 studies for meta-analysis. One reviewer extracted data, double-checked by a second; discrepancies were resolved by a third. Meta-analyses were conducted using a random-effects model (metafor R package) to estimate combined effect size (CES), 95% confidence intervals (CI), heterogeneity, and publication bias. Risk of bias was assessed via SYRCLE's tool and the CAMARADES checklist. Despite high inconsistency ( I ² = 81.5%), the global CES was large and significant [Hedges' g  = 1.66, 95% CI (1.53; 1.79), k  = 561, power > 80%], consistent across most subgroups. Small study effects and publication bias inflated CES estimates, especially in mice, while results in rats were more variable. Nonetheless, antidepressants consistently reduced immobility in mice across diverse conditions. In rats, findings were less consistent, though the most robust data showed a significant, dose-dependent antidepressant-like effect of imipramine in both species. However, publication bias and incomplete reporting compromise the accuracy of CES estimates and raise concerns about the validity of the FST literature. These findings highlight the need for more transparent reporting practices in FST-based antidepressant research.