P4404Prognostic value of novel nocturnal oxygen saturation metrics in patients with obstructive sleep apnoea and high cardiovascular event risk

Abstract

Abstract Aim To determine prognostic value of novel oximetry-derived metrics in patients with obstructive sleep apnoea (OSA) and high cardiovascular (CV) event risk. Methods and results Ancillary study of the Sleep Apnoea CV Endpoints (SAVE) randomised controlled trial, which investigated the effects of CPAP treatment on secondary CV event outcomes in 2687 participants with moderate-severe OSA and established CV disease. Associations between baseline characteristics, standard sleep and pulse oximetry-derived metrics, and CV outcomes were determined in Cox proportional hazards regression models stratified for treatment allocation. Metrics evaluated included: oxygen desaturation index, time below 90% oxygen saturation (SpO2), average SpO2across recordings (mean SpO2), across episodic desaturation events (desaturation SpO2), and baseline interpolated through episodic desaturation events (baseline SpO2); duration and desaturation/resaturation time ratio of episodic desaturation events; and mean and standard deviation of pulse rate. No SpO2 metric was associated with the composite outcome. Mean and baseline SpO2 were negatively associated with heart failure (hazard ratio [HR] 0.81, 95% confidence interval [CI] 0.69–0.95; P=0.009 and 0.78, 0.67–0.90; P=0.001, respectively) and myocardial infarction risk (0.86, 0.77–0.95; P=0.003 and 0.81, 0.73–0.90; P<0.001, respectively) and marginally positively associated with stroke risk (1.09, 1.00–1.20; P=0.065 and 1.13, 1.02–1.26; P=0.020, respectively). Desaturation duration and desaturation/resaturation time ratio, with established risk factors, predicted heart failure (area under the receiver-operating-characteristic curve 0.86, 95% CI 0.79–0.93). Conclusions Higher mean and baseline SpO2 were associated with reduced risk of heart failure and myocardial infarction, and small increases in stroke rate. Desaturation shape may have predictive utility for future CV events. Clinical trial registration: clinicaltrials (NCT00738179).