Association between oximetry-derived pulse-rate variability and stroke risk in patients with obstructive sleep apnoea

Abstract

Objectives: To evaluate whether pulse rate variability (PRV) derived from oximetry recordings during sleep studies is associated with the occurrence of stroke in patients with obstructive sleep apnoea (OSA).

Methods: Clinical and pulse oximetry data from 6,075 stroke-free OSA patients from the Pays de la Loire Sleep cohort were linked to heath administrative data for the identification of new stroke onset. Time and frequency domain parameters of PRV were extracted from the photoplethysmography signals’ normal-to-normal (NN) beat intervals. Stepwise Cox proportional hazard models were used to assess the association between PRV parameters and stroke risk.

Results: After a median follow-up of 6.56 years, 177 patients had received a diagnosis of stroke. After adjusting for confounders including anthropometric data, smoking habits, cardiovascular, metabolic and respiratory comorbidities, stroke risk was associated with the normalized low to high frequency power ratio (LF/HF) when analyzed as a continuous variable (HR [95%CI]=0.76 [0.59-0.96]; p=0.02) or divided in quartiles of the variable (p-trend=0.03). Further adjustments for CPAP adherence and OSA severity did not change the magnitude of the association. However, the association appeared stronger in patients with severe OSA (for lowest quartile LF/HF< ­0.65 : HR=2.39 [1.24-4.59]; p=0.0089) than in those with mild to moderate OSA (p=0.2 for interaction).

Conclusion: In patients with OSA, decreased sympathetic/parasympathetic tone assessed by oximetry derived LF/HF ratio was associated with a higher risk of stroke. Sleep oximetry-derived indices of PRV could help identify those patients with high risk of stroke.