Accelerated electrocardiographic aging of the heart as a digital marker of brain aging, cognitive decline, and dementia risk
S H Cho, P S Yang, T H Hwang, H J Park, D H Kim, T H Kim, J S Uhm, H N Pak, S C You, H T Yu, B Y JoungAbstract
Background
Artificial intelligence (AI)–derived electrocardiographic aging (ECG-aging) is a non-invasive marker of biological cardiac aging linked to adverse cardiovascular outcomes. However, its association with brain aging, cognitive decline, and dementia risk remains unclear.
Methods
AI-predicted age from 12-lead ECGs (AI-ECG age) was estimated using the previously validated PROPHECG-Age deep-learning model, and AI-ECG age gap was calculated as the difference between AI-ECG age and chronological age in the UK Biobank (n=42,571; median follow-up, 3.0 years) and Severance cohorts (n=113,584; mean follow-up, 4.2 years). Participants were categorized as normal (age gap <7 years) or ECG-aged (≥7 years) groups. Cognitive function was compared between groups using general cognitive ability (GCA), derived from principal component analysis of 11 cognitive tests with longitudinal changes additionally assessed in the UK Biobank, and Montreal Cognitive Assessment (MoCA) domain and total scores in the Severance cohort. Dementia risk was compared between groups using survival analysis in both cohorts.
Results
In the UK Biobank, the ECG-aged group demonstrated significantly poorer cognitive performance, particularly in processing speed, executive function, memory, and fluid intelligence, and greater longitudinal decline in GCA (−0.11 vs −0.02 per 5-year interval). In the Severance cohort, MoCA domain and total scores were numerically lower but not significantly different between groups. Accelerated ECG-aging was associated with increased dementia risk in both cohorts (HR 2.56 [95% CI 1.01–6.51] in the UK Biobank; HR 1.19 [95% CI 1.05–1.34] in the Severance cohort).
Conclusions
Accelerated ECG-aging is associated with poorer cognitive performance, faster cognitive decline, and increased dementia risk. ECG-based biological aging may serve as a scalable digital marker of brain aging and neurodegenerative risk beyond cardiovascular aging.