Pooling 14,126 participants from the Framingham Heart Study, MESA, and the Cardiovascular Health Study, investigators found that a positive composite ECG-AI screen was linked to a 10- to 20-fold higher risk of developing incident HF vs a negative screen, and improved near-term risk prediction when added to PREVENT-HF.

Why It Matters To Your Practice

• ECG-AI offers a scalable way to flag patients who may look “low/moderate risk” clinically but are at markedly higher HF risk based on ECG patterns of systolic/diastolic dysfunction.

• In this pooled analysis (run July–Sept 2025 on NHLBI BioDataCatalyst), incident HF occurred in 7.7% of participants, underscoring the value of earlier risk stratification in routine populations.

Clinical Benefits

• Potential to prioritize follow-up for a small, high-risk subgroup: 11.9% screened positive on the composite ECG-AI model (vs 25.1% with PREVENT-HF ≥10%).

• Adding ECG-AI to PREVENT-HF improved 1-directional net reclassification: 0.086–0.125 at the 10% PREVENT-HF threshold and 0.327–0.403 at the 20% threshold (across 1–10 years).

• May support targeted prevention workflows (e.g., tighter BP control, diabetes management, weight loss, sleep apnea evaluation, and earlier echo/labs when appropriate) for those most likely to convert to HF.

Managing Risks

• Expect false positives/clinical ambiguity: ECG-AI flags risk, not a diagnosis—confirmatory evaluation (history/volume status, natriuretic peptides when indicated, echocardiography) may be needed.

• Plan for downstream utilization: positive rates for ECG-AI DD were 11.1% and composite 11.9%, which could increase demand for imaging and follow-up if deployed broadly.

• Communicate clearly with patients: frame results as “higher risk” to motivate prevention, while avoiding labeling as Heart Failure (CHF) without clinical criteria.

The Bottom Line

• In a pooled cohort of 14,126, a positive composite ECG-AI screen identified patients with 10–20x higher HF risk vs negative screens and improved near-term reclassification when added to PREVENT-HF—suggesting a practical pathway for population-level HF risk stratification and targeted prevention.