
MADRID, Spain — Researchers in Spain have developed an artificial-intelligence model that analyses electrical signals sent by implanted defibrillators to identify periods when some heart-failure patients may face a higher risk of a severe heart rhythm disturbance. The study, led by the Spanish National Centre for Cardiovascular Research (CNIC), focuses on people who already have an implantable cardioverter-defibrillator, a device designed to detect and treat dangerous arrhythmias. Rather than relying only on occasional clinical measurements, the system examines changes in electrical activity that the implanted devices transmit remotely to hospitals during routine follow-up. Researchers trained the model using recordings from 5,089 patients treated across 50 Spanish hospitals and then tested it in an independent group of 2,040 patients across seven centres. The model generates a risk score that updates with each new transmission. According to the CNIC research announcement published on November 19, the score rose in the months before ventricular fibrillation among patients who later experienced the condition, while it remained more stable in other patients. The findings were published in the American Heart Association journal *Circulation: Arrhythmia and Electrophysiology*.

The potential value is that the approach may help clinicians recognise changes in risk before a serious event occurs, using information that implanted devices are already collecting rather than requiring another implant. The researchers say the system could eventually be incorporated into remote-monitoring platforms to help care teams decide when a patient may need earlier review or a treatment adjustment. However, the work should not be interpreted as a proven warning system for everyone with heart disease. It was developed for a defined group of heart-failure patients who already carry implanted defibrillators, and further assessment is needed before its use becomes routine clinical practice. A predictive score would support medical judgement; it would not replace a cardiologist, guarantee that an arrhythmia can be prevented or diagnose a heart condition in the general population. The study’s scale, including its independent validation cohort, gives researchers a basis for further testing how reliably the signal changes can be used in everyday care. The research paper is identified by DOI 10.1161/CIRCEP.125.015093. HeadlinesXtra has also reported on AI’s growing role in biomedical research in its coverage of a biological AI model for drug discovery and genomic research.