Comparative effects of hypnotic agents on sleep architecture and respiratory outcomes in obstructive sleep apnea: A systematic review and network meta-analysis

Aim: This network meta-analysis of randomized controlled trials (RCTs) aimed to investigate which hypnotics are associated with the most favorable sleep architecture and respiratory outcomes in adults with obstructive sleep apnea. Methods: Primary outcomes included total sleep time (TST) and apnea–hypopnea index (AHI) during TST. Other outcomes were rapid eye movement (REM) sleep time, latency to persistent sleep (LPS), wake after sleep onset (WASO), sleep efficiency (SE), AHI during non-REM or REM sleep, mean peripheral oxygen saturation (SpO₂) during TST, mean SpO₂ nadir during TST, arousal index (AI), all-cause discontinuation, adverse event-related discontinuation, and incidence of individual adverse events. Effect sizes with 95% confidence intervals were calculated. Results: This systematic review included 32 RCTs (n = 1871, average age = 51.60 years, 62.52% male, mean AHI = 23.60). Our network meta-analysis evaluated brotizolam, daridorexant, eszopiclone, flurazepam, lemborexant, nitrazepam, ramelteon, temazepam, triazolam, zaleplon, zolpidem, zopiclone, and placebo. Compared with placebo, lemborexant increased TST, REM sleep time, and SE and decreased LPS and WASO, whereas both daridorexant and zolpidem increased TST and SE and decreased WASO. These three medications demonstrated respiratory safety and discontinuation profiles similar to those of placebo. Eszopiclone increased TST and SE and decreased LPS, WASO, AHI during TST, and AI, but its effects on LPS, WASO, AHI during TST, and AI disappeared in the sensitivity analysis, excluding continuous positive airway pressure titration studies. Conclusion: Our network meta-analysis identified different effects of various hypnotics on sleep architecture and respiratory parameters; however, the lack of data prevented a formal synthesis of subjective outcomes. Therefore, these results should be interpreted with caution in clinical practice.