Study Overview

• Objective: To create and validate AIRE-HTN, an AI-ECG model capable of predicting new-onset hypertension and stratifying patients' risk for hypertension-related adverse events.

• Methodology: The model was trained on over 1.1 million ECGs from approximately 190,000 patients at Beth Israel Deaconess Medical Center (BIDMC) between 2014 and 2023. It was internally validated on 19,423 BIDMC patients and externally validated using data from 65,610 participants in the UK Biobank (UKB).

Key Findings

• Predictive Performance: AIRE-HTN achieved a concordance index (C-index) of 0.70 in both internal and external cohorts, indicating consistent performance in predicting incident hypertension.

• Risk Stratification: The model effectively stratified risk for various hypertension-related adverse outcomes, including:

  • Cardiovascular death (Hazard Ratio [HR] per standard deviation: 2.24)

  • Heart failure (HR: 2.60)

  • Myocardial infarction (HR: 3.13)

  • Ischemic stroke (HR: 1.23)

  • Chronic kidney disease (HR: 1.89)

• Additive Value: AIRE-HTN provided significant incremental predictive value over traditional clinical risk factors, with a continuous net reclassification index of 0.44 for BIDMC and 0.32 for UKB.

Limitations

• Hypertension Definition: The study relied on ICD codes to define hypertension at BIDMC, which may not align perfectly with current clinical guidelines.

• Validation Constraints: The model was not validated against ambulatory blood pressure monitoring, and the UKB cohort's relatively healthy population may limit the generalizability of the findings to more diverse populations.

Conclusion

AIRE-HTN demonstrates potential as a tool for early detection of hypertension and risk stratification for related adverse outcomes, offering incremental value beyond conventional clinical risk factors. Its integration into clinical practice could enhance preventive strategies and patient management.

Jun 19, 2025