Synergistic prognostic value of event-aligned heart rate surge and hypoxic burden for cardiovascular mortality in sleep apnoea

Background Obstructive sleep apnoea is a prevalent, heterogeneous disorder intrinsically linked to cardiovascular disease (CVD). Traditional diagnostic metrics such as the apnoea-hypopnoea index correlate poorly with incident CVD outcomes as they treat all events equally, failing to capture individual physiological severity. Sleep apnoea imposes two distinct burdens: hypoxaemic stress (hypoxic burden, HB) and intense autonomic sympathetic activation. While HB quantifies oxygen desaturation, the magnitude of the autonomic response—quantifiable non-invasively via the event-aligned heart rate (HR) surge at event termination—remains profoundly underutilised as a prognostic biomarker. Purpose To determine whether the computationally extracted HR surge provides independent prognostic value for long-term CVD mortality beyond established risk factors and HB, and to evaluate the synergistic cardiovascular risk profile of these two distinct physiological phenotypes. Methods In a community-based patient study of 4,790 adults from the Sleep Heart Health Study, we extracted the HR surge—defined as the maximum HR acceleration relative to a 60-second pre-event baseline—from polysomnography-synchronised electrocardiograms. HB was the total area under the desaturation curve. Associations with CVD mortality were assessed using sequential multivariable Cox proportional hazards models. Discrimination, reclassification, and clinical utility were evaluated using Harrell’s C-index, Net Reclassification Improvement (NRI), and decision curve analysis. Results Over a median 11.6-year follow-up, 334 CVD deaths occurred among the 4,790 community-dwelling participants (mean age 63.9 years, 53.8% female). HR surge and HB demonstrated minimal correlation (rho=0.08), indicating they capture distinct pathophysiological pathways. In fully adjusted models accounting for demographics, comorbidities, and standard sleep metrics, patients in the highest HR surge quintile exhibited significantly elevated CVD mortality risk (hazard ratio 3.06, 95% confidence interval [CI] 2.14–4.38, p<0.001). Strikingly, HB lost statistical significance when mutually adjusted. A 'double jeopardy' phenotype emerged: patients with high HB and high HR surge faced nearly a three-fold increased risk (HR 2.84, 95% CI 1.93–4.19; 10-year absolute risk 9.20%). Integrating HR surge into a base model improved discrimination (C-index 0.77 to 0.81, p<0.001) and continuous NRI (0.62). Conclusions The event-aligned HR surge is a robust, independent predictor of CVD mortality, capturing the deleterious sympathetic consequences of sleep apnoea. The minimal correlation between HR surge and HB reveals distinct autonomic and hypoxaemic endotypes. Combined, these biomarkers identify a high-risk phenotype, optimising cardiovascular risk stratification.