Sleep-promoting and circadian interventions for prevention of postoperative delirium in adult surgical patients: a systematic review and meta-analysis

Background Sleep and circadian disruption may contribute to postoperative delirium (POD), but previous reviews often pooled sedative drugs, melatonergic agents, and broad delirium-prevention bundles. Objective To evaluate whether sleep-promoting and circadian interventions reduce POD in adult surgical patients. Methods We searched PubMed/MEDLINE, Embase, CENTRAL, Web of Science, Scopus, CINAHL, and PsycINFO to 31 July 2026 for randomized trials of sleep-promoting or circadian interventions in adult surgical patients. Eligible interventions included melatonin, ramelteon, orexin receptor antagonists, bright-light/circadian interventions, sensory-environment strategies, and multicomponent or behavioral sleep-promoting protocols. Primary sedative, anesthetic, antipsychotic, or analgesic interventions were excluded. Risk ratios (RRs) were pooled using random-effects models. Core analyses were restricted to validated or clinically accepted delirium assessment methods. Results Twenty-nine studies were included; 27 trials involving 2,788 participants contributed to the dichotomous meta-analysis and 2 studies were summarized narratively. Overall, interventions were associated with reduced POD risk (RR 0.62, 95% CI 0.49–0.78; p < 0.001; I² = 41.8%). The association attenuated in Core studies (RR 0.70, 95% CI 0.52–0.93; p = 0.018; I² = 49.7%; k = 18). Melatonergic agents showed an association in broad analysis (RR 0.66, 95% CI 0.50–0.88), but not in Core melatonergic trials (RR 0.76, 95% CI 0.54–1.08) or after excluding balanced dexmedetomidine-background trials (RR 0.83, 95% CI 0.57–1.22). Light/circadian interventions showed a numerically strong but small-study signal (RR 0.24, 95% CI 0.09–0.64; k = 3). Non-pharmacological Core interventions showed a consistent association (RR 0.44, 95% CI 0.21–0.93; I² = 14.7%). Conclusions Sleep-promoting and circadian interventions were associated with lower POD risk, but evidence was heterogeneous, mostly low or very low certainty, and sensitive to delirium ascertainment. Larger trials using standardized delirium and sleep measurements are needed. Systematic Review Registration https://www.crd.york.ac.uk/PROSPERO/view/CRD420261467303 , identifier CRD420261467303.