Frontiers in Medicine · Nov 10, 2026

Comparative Efficacy of Non-invasive Rehabilitation Therapies for Obstructive Sleep Apnea: A Systematic Review and Network Meta-analysis

By Xinyi Hu, Jialin Gan, Qixin Wang et al.

Objective: This network meta-analysis compared 16 active non-invasive interventions for obstructive sleep apnea (OSA) across seven domains: sleep apnea severity, nocturnal oxygenation, cardiopulmonary fitness, respiratory muscle strength, upper-airway function, neuropsychological health, and physical functional capacity. Methods: Six databases were searched for randomized controlled trials in adults with OSA. Random-effects frequentist network meta-analysis estimated mean differences for sleep apnea severity and Hedges' g for the other domains. Treatment ranking and evidence certainty were assessed using the surface under the cumulative ranking curve and the Confidence in Network Meta-Analysis framework, respectively. Results: Eighty-one reports representing 78 independent trials and 4,798 adults were included. High-intensity interval training produced the largest reduction in apnea–hypopnea index (mean difference −37.22 events/h, 95% confidence interval −48.45 to −25.99; low certainty), while respiratory muscle rehabilitation combined with positive airway pressure (PAP) and PAP alone were supported by high-certainty evidence. Aerobic exercise combined with PAP ranked highest for nocturnal oxygenation (Hedges' g 4.20, 95% confidence interval 1.65 to 6.74; low certainty), and PAP showed moderate-certainty benefit. Aerobic and resistance training improved cardiopulmonary fitness (Hedges' g 1.52, 95% confidence interval 0.53 to 2.52; moderate certainty) and neuropsychological health (Hedges' g −1.97, 95% confidence interval −2.89 to −1.05; low certainty). Respiratory muscle rehabilitation ranked highest for respiratory muscle strength (Hedges' g 3.28, 95% confidence interval 1.18 to 5.39; very low certainty) and upper-airway function (Hedges' g 1.51, 95% confidence interval 0.78 to 2.25; moderate certainty). For physical functional capacity, high-intensity interval training ranked highest, whereas aerobic and resistance training produced the most precise favorable estimate (Hedges' g 1.16, 95% confidence interval 0.09 to 2.22; low certainty). Conclusion: Effects were outcome-specific: PAP-based strategies favored sleep-related outcomes, respiratory muscle rehabilitation favored muscular and upper-airway outcomes, and exercise-based strategies favored cardiopulmonary, neuropsychological, and functional outcomes. Treatment selection should consider the target outcome and evidence certainty.

The paperJournal siteInterventions

AgingResearch.News