Higher 24-hour systolic blood pressure was significantly associated with increased deep learning-predicted ECG-age (β 0.038 years per 1 mmHg increase; 95% CI 0.009-0.067).
Observational (n=781)
Is ambulatory blood pressure and heart rate associated with deep learning-predicted electrocardiographic age in adults?
Higher 24-hour ambulatory blood pressure and heart rate are independently associated with increased deep learning-derived electrocardiographic age, suggesting ECG-age reflects cumulative vascular and autonomic stress.
Effect estimate: β 0.038 (95% CI 0.009-0.067)
Abstract Introduction Deep learning–predicted electrocardiographic age (ECG-age), captures biological aging of the heart and predicts cardiovascular events beyond traditional risk factors. Whether ECG-age reflects 24-hour hemodynamic and autonomic burden, specifically ambulatory blood pressure (ABP) and heart rate (HR), remains unclear. Because ABP—especially nocturnal BP—is a strong marker of vascular and autonomic dysregulation, examining its relationship with ECG-age may clarify whether ECG-derived biological aging mirrors daily cardiovascular load. Methods We analyzed 781 adults (mean age 60 ± 7 years; 53.1% female) from the KoGES Ansan–Anseong cohort with valid resting ECGs and 24-hour ABP monitoring. ECG-age was estimated using a validated deep neural network trained on 1.5 million ECGs. The difference between ECG-age and chronological age (Δage) was evaluated as a continuous measure and as decelerated vs accelerated aging groups. ABP was recorded every 30 minutes (daytime) and 60 minutes (nighttime). Mean 24-hour, daytime, and nighttime systolic (SBP) and diastolic BP (DBP), and HR were computed. Associations between ABP/HR and ECG-age were assessed using multivariable linear regression adjusting for age, sex, BMI, diabetes, and prevalent cardiovascular disease. Sensitivity analyses excluded participants with CVD. Results Higher ECG-age was significantly associated with elevated 24-hour ABP and HR. Each 1 mmHg increase in mean SBP corresponded to 0.038 years higher ECG-age (95% CI 0.009–0.067). Each 1 mmHg increase in mean DBP corresponded to 0.076 years higher ECG-age (95% CI 0.040–0.113). Associations were consistent across daytime and nighttime periods, with similar effect sizes. Mean ambulatory HR was also positively associated with ECG-age (β = 0.070 years per bpm, 95% CI 0.033–0.108). In logistic models, DBP and HR significantly predicted accelerated aging status, whereas SBP did not. Sensitivity analyses excluding cardiovascular disease cases yielded comparable results, confirming robustness. Conclusion ECG-age, a deep learning–derived marker of electrophysiologic aging, is independently associated with higher 24-hour blood pressure and heart rate. These findings indicate that ECG-age reflects cumulative vascular and autonomic stress across the circadian cycle and may serve as a scalable biomarker of early cardiovascular aging relevant to sleep and cardiometabolic physiology. Support (if any)
Hernandez et al. (Fri,) conducted a observational in General population (n=781). Ambulatory blood pressure and heart rate was evaluated on Association between 24-hour ambulatory systolic blood pressure and ECG-age (β 0.038, 95% CI 0.009-0.067). Higher 24-hour systolic blood pressure was significantly associated with increased deep learning-predicted ECG-age (β 0.038 years per 1 mmHg increase; 95% CI 0.009-0.067).