Cardiovascular diseases are among the most common causes of death worldwide. While classic risk factors such as high blood pressure or diabetes have been well studied, the biological age of the heart offers a new perspective for the early detection of health risks. In the current study, an AI-supported model was validated that estimates the ‘ECG age’ of a person and thus identifies potential deviations from chronological age.
The analysis is based on long-term data from a German population study with over 20 years of follow-up. The AI model was originally trained on Brazilian ECG data and has now been successfully applied to a European cohort. A strong correlation between predicted and actual biological age of the heart was found.
Increased risk of Cardiovascular diseases detected
The study shows that people whose ECG age exceeds their chronological age by more than eight years have a significantly higher risk of cardiac arrhythmia, heart failure and increased mortality. By taking consecutive ECGs into account, it was also possible to further increase the accuracy of the prognosis.
The mortality risk increased from 1.43 to 1.65 when not just a single ECG, but several ECGs recorded over a longer period of time were used for the analysis. This shows that changes in the ECG over time allow a more accurate prediction of mortality risk.
Potential for personalised prevention
The results underline the potential of AI-supported analyses for the early identification of patients with an increased cardiovascular risk. In the long term, this method could be integrated into routine health checks in order to recognise people at risk at an early stage and initiate preventive measures.
‘Our study shows that artificial intelligence is able to recognise subtle changes in the ECG that indicate accelerated heart ageing. This could open up new possibilities for personalised medicine and help to prevent cardiovascular diseases at an early stage,’ explains first author Philip Hempel, a member of the “Biosignal Processing” working group at the Institute of Medical Informatics at the University Medical Center Göttingen.
‘We have especially focussed on making the AI systems explainable. By integrating classic ECG parameters into our analyses, we combine AI systems with traditional, evidence-based medicine. In this way, doctors can not only benefit from the additional insights, but also understand which specific characteristics - whether from the data or the classic parameters - have led to a particular diagnosis. This promotes trust in new technologies and supports well-founded, individualised patient care,’ says Hempel.
The study was conducted under the direction of the University Medical Centre Göttingen in collaboration with the German Centre for Cardiovascular Research (DZHK). The co-operation between various locations in Germany and the support of international partners in Uppsala (Sweden) and Brazil was a key factor in the success of the project. The investigation is based on data from the SHIP study (Study of Health in Pomerania), a comprehensive long-term study of the population in northern Germany.
Original publication:
Hempel, P., Ribeiro, A.H., Vollmer, M. et al. Explainable AI associates ECG aging effects with increased cardiovascular risk in a longitudinal population study. npj Digit. Med. 8, 25 (2025). https://doi.org/10.1038/s41746-024-01428-7
Scientific Contact:
Philip Hempel, University Medical Centre Göttingen, philip.hempel[at]med.uni-goettingen.de
