Significance of including a surrogate arousal for sleep apnea-hypopnea syndrome diagnosis by respiratory polygraphy

Abstract

Rationale: Respiratory polygraphy is an accepted alternative to polysomnography (PSG) for sleep apnea/hypopnea syndrome (SAHS) diagnosis, although it underestimates the apnea-hypopnea index (AHI) because respiratory polygraphy cannot identify arousals. Objectives: We performed a multicentric, randomized, blinded crossover study to determine the agreement between home respiratory polygraphy (HRP) and PSG, and between simultaneous respiratory polygraphy (respiratory polygraphy with PSG) (SimultRP) and PSG by means of 2 AHI scoring protocols with or without hyperventilation following flow reduction considered as a surrogate arousal. Methods: We included suspected SAHS patients from 8 hospitals. They were assigned to home and hospital protocols at random. We determined the agreement between respiratory polygraphy AHI and PSG AHI scorings using Bland and Altman plots and diagnostic agreement using receiver operating characteristic (ROC) curves. The agreement in therapeutic decisions (continuous positive airway pressure treatment or not) between HRP and PSG scorings was done with likelihood ratios and post-test probability calculations. Results: Of 366 randomized patients, 342 completed the protocol. AHI from HRP scorings (with and without surrogate arousal) had similar agreement with PSG. AHI from SimultRP with surrogate arousal scoring had better agreement with PSG than AHI from SimultRP without surrogate arousal. HRP with surrogate arousal scoring had slightly worse ROC curves than HRP without surrogate arousal, and the opposite was true for SimultRP scorings. HRP with surrogate arousal showed slightly better agreement with PSG in therapeutic decisions than for HRP without surrogate arousal. Conclusion: Incorporating a surrogate arousal measure into HRP did not substantially increase its agreement with PSG when compared with the usual procedure (HRP without surrogate arousal).