A comparison of therapeutic oxygen versus medical air for the treatment of central sleep apnea in infants with Prader-Willi syndrome; A proof-of-concept randomized crossover trial

Authors
Category Primary study
JournalSleep medicine
Year 2026
Background Central sleep apnea (CSA) is common in infants with Prader–Willi syndrome (PWS). Although supplemental oxygen is frequently used to manage CSA, its mechanism of action remains unclear; comparison with medical air may help determine whether therapeutic benefit is mediated by inspired oxygen or by flow-related effects on respiratory control. Study design and methods In this randomized crossover trial, growth hormone–naïve infants <2 years with genetically confirmed PWS and CSA (CAHI >5 events/hr) underwent an overnight polysomnogram receiving medical air and supplemental oxygen in randomized order. Treatments were blinded to scoring sleep technologists and interpreting physicians. The primary outcome was the difference in CAHI between supplemental oxygen and medical air. Secondary outcomes included sleep architecture, obstructive apnea–hypopnea index (OAHI), and gas exchange. Results Of 20 infants screened, 10 were randomized; 4 (40%) were male, with a median age of 6.0 months (IQR 4.2–13.2 months). Baseline median CAHI was 9.0 events/hr (IQR 6.3–10.3). Supplemental oxygen reduced CAHI compared with medical air (adjusted difference −7.7 events/hr, 95% CI −15.2 to −0.1; p = 0.048; adjusted means 1.8 vs 9.4 events/hr). Compared with medial air, supplemental oxygen increased oxygen saturation nadir (+8.2%, p < 0.001), and mean oxygen saturation (+2.1%, p = 0.009), and reduced the oxygen desaturation index (−12.3 events/hr, p = 0.046). Mean transcutaneous CO2 was modestly lower with oxygen (−2.2 mmHg, p = 0.024), with no clinically meaningful hypercapnia. Sleep architecture and obstructive indices were unchanged. Interpretation In comparison to medical air, the administration of supplemental oxygen reduces CSA and improves oxygenation in infants with PWS without adverse effects on CO2 or sleep architecture, supporting its use as first-line therapy in this population. © 2026 The Authors.
Epistemonikos ID: 9d836643cfd4dc4049a35ad4dd4f94697dddaef5
First added on: Jun 16, 2026