Level of evidence assessment per review outcome.

IntroductionMelatonin supplements and melatonin receptor agonists are linked to reduced delirium in the Intensive Care Unit (ICU) which we hypothesised may affect the length of stay (LOS) in ICU or in hospital. In this review, we identified and critically appraised the literature on the effect of exogenous melatonin and melatonin receptor agonists on the ICU and/or hospital LOS among adults admitted to the ICU.MethodsSix electronic databases and three trial registries were searched for randomised controlled trials (RCTs). Screening, risk of bias assessment, quality appraisal, and level of evidence assessment were conducted and cross-checked by two reviewers independently. Meta-analyses with disease-specific subgroups were conducted to assess the mean difference in LOS for exogenous melatonin and melatonin receptor agonists compared with a placebo.ResultsTwenty RCTs were reviewed with 14 having a low risk of bias. For ICU LOS (18 studies) there was significant statistical heterogeneity (I2 = 73%); compared with placebo the 95% prediction interval for the mean difference was −3.18 and 1.39 days. For hospital stay (12 studies, I2 = 79%) the 95% prediction interval ranged from −6.68 to 3.52. Removing two statistical outliers, and correcting for publication bias, there was no overall statistically significant difference in mean ICU LOS (p-value = 0.298) or mean hospital LOS (p-value = 0.456). The subgroup analysis found statistically significant improvements for those who underwent coronary artery bypass graft surgery (ICU LOS −0.47 days, 95% CI: −0.78 to −0.16, p-value = 0.003); and patients with COVID-19 (hospital LOS −3.90 days, 95% CI: −6.28 to −1.51, p-value = 0.001).ConclusionThere was a very low certainty of evidence that melatonin and melatonin receptor agonists were associated with reductions in ICU and hospital LOS in ICU patients overall. However, further research is needed for surgical patients and those with pneumonia.