J Am Geriatr Soc. 2024 Mar 4. doi: 10.1111/jgs.18825. Online ahead of print.
ABSTRACT
BACKGROUND: Delirium is common in older inpatients, causing distress, cognitive decline, and death. Current therapies are unsatisfactory, limited by lack of efficacy and adverse effects. There is an urgent need for effective delirium treatment. Sleep wake cycle is disturbed in delirium; endogenous Melatonin is perturbed, and exogenous Melatonin is a safe and effective medication for sleep disorders. This study aims to determine the effect of oral Melatonin 5 mg immediate release (IR) nightly for five nights on the severity of delirium in older (≥65 years) medical inpatients.
METHODS: This was a double-blinded, randomized controlled trial in general internal medicine units of a tertiary teaching hospital. Older inpatients with Confusion Assessment Method positive, hyperactive or mixed delirium within 48 h of admission or onset of in-hospital delirium were included. The primary outcome was change in delirium severity measured with the Memorial Delirium Assessment Scale (MDAS). A previous pilot trial showed 120 participants randomized 1:1 to Melatonin or Placebo would provide 90% power to demonstrate a 3-point reduction in the MDAS.
RESULTS: One hundred and twenty participants were randomized, 61 to Melatonin 5 mg and 59 to Placebo. The medication was well tolerated. The mean MDAS improvement was 4.9 (SD 7.6) in the Melatonin group and 5.4 (SD 7.2) in the Placebo group, p-value 0.42, a non-significant difference. A post-hoc analysis showed length of stay (LOS) was shorter in the intervention group (median 9 days [Interquartile Range (IQR) 4, 12] vs. Placebo group 10 [IQR 6, 16] p-value = 0.033, Wilcoxon Rank Sum test).
CONCLUSIONS: This trial does not support the hypothesis that Melatonin reduces the severity of delirium. This may be due to no effect of Melatonin, a smaller effect than anticipated, an effect not captured on a multidimensional delirium assessment scale, or a type II statistical error. Melatonin may improve LOS; this hypothesis should be studied.
PMID:38438279 | DOI:10.1111/jgs.18825