Predicting "heart age" using electrocardiography.

Area of Science:

• Cardiology
• Biostatistics
• Preventive Medicine

Background:

• Assessing cardiovascular health is crucial, especially for individuals with symptoms but no diagnosed cardiac pathology.
• Patient understanding of cardiovascular risk can motivate lifestyle modifications.

Purpose of the Study:

• To develop and evaluate a Bayesian statistical model for predicting an individual's "heart age" using electrocardiogram (ECG) data.
• To assess the model's utility across diverse populations, including healthy individuals, those with risk factors, cardiac disease patients, and athletes.

Main Methods:

• A Bayesian statistical model was created to predict heart age based on resting 12-lead ECGs.
• The model was trained and validated on a dataset of 776 healthy individuals aged 20+ with no known risk factors.
• The model was secondarily applied to groups with cardiac risk factors, diagnosed cardiac disease, and highly endurance-trained athletes.

Main Results:

• In healthy non-athletes, predicted heart age generally aligned with chronological age.
• Approximately 75% of subjects with cardiac risk factors and nearly all patients with diagnosed heart disease showed higher predicted heart ages than their body ages.
• A majority of highly endurance-trained athletes also exhibited higher predicted heart ages, potentially indicating subclinical cardiac changes.

Conclusions:

• The developed heart age prediction model, based on ECG, shows promise in identifying individuals at higher cardiovascular risk.
• The model's ability to differentiate heart age from body age may aid in targeted interventions and lifestyle change encouragement.
• Further research may explore the model's clinical implications, particularly in athletes and those with subclinical cardiac conditions.