Arousal threshold modifies the effect of CPAP on executive function among individuals with obstructive sleep apnoea

Background Obstructive sleep apnoea (OSA) is associated with neurocognitive dysfunction. However, randomised trials evaluating the effects of continuous positive airway pressure (CPAP) on neurocognition in those without dementia do not show a benefit. We thus aimed to assess whether arousal threshold (ArTH) modifies the effect of CPAP on neurocognitive function. Methods We performed a secondary analysis of a randomised, sham-controlled trial (Apnea Positive Pressure Long-term Efficacy Study (APPLES); ClinicalTrials.gov: NCT00051363). ArTH was estimated from polysomnography using a translatable method. Neurocognitive outcomes included the Sustained Working Memory Test-Overall Mid-Day (SWMT-OMD) score (executive function, primary outcome), with the Pathfinder Number Test total time (attention) and Buschke Selective Reminding Test sum recall (learning and memory) as secondary outcomes. Generalised linear modelling assessed whether the effect of CPAP was modified by baseline ArTH (treatment×ArTH interaction). 833 participants with OSA (apnoea–hypopnoea index ⩾10 events·h⁻¹), available ArTH and outcomes were analysed (active CPAP n=437 and sham CPAP n=396). Results For executive function, the effect of CPAP treatment was modified by ArTH (p_interaction=0.042). Specifically, for every 1SD increase in ArTH, the SWMT-OMD score improved by 0.091 (95% CI 0.003–0.178) in active compared to sham CPAP at 6 months; at ArTH 1SD above the mean, SWMT-OMD improvements were nearly three times that in those with average ArTH (0.139 (95% CI 0.018–0.261) versus 0.053 (95% CI −0.034–0.140), respectively). No effect modification was observed for attention (p=0.311) or learning and memory (p=0.744). Conclusion In OSA, a higher ArTH is associated with greater improvements in executive function following CPAP therapy.