0
0
An artificial intelligence–driven electrocardiogram (ECG) risk estimation model, AIRE-HTN, has been developed to predict the onset of hypertension and assess associated cardiovascular risks. The model has demonstrated the ability to outperform traditional clinical markers in identifying individuals at risk for adverse outcomes, according to a study published in JAMA Cardiology.
Study Methodology
The AIRE-HTN model was developed and validated through a large-scale prognostic cohort study conducted in secondary care settings. Researchers used a derivation cohort from Beth Israel Deaconess Medical Center in Boston, which included 1,163,401 ECGs from 189,539 patients (mean age 57.7 years, 52.1% women, 64.5% White).
To validate the model externally, 65,610 ECGs from a UK-based volunteer cohort were analyzed. The UK cohort had an average age of 65.4 years, with 51.5% women and 96.3% White individuals. Both cohorts included individuals without baseline hypertension to evaluate the model’s predictive accuracy.
Key Findings
The AIRE-HTN model predicted incident hypertension with a C-index of 0.70 in both cohorts. Patients in the highest quartile of AIRE-HTN scores had a fourfold higher risk of developing hypertension compared to those with lower scores (P < .001).
Importantly, the model maintained predictive accuracy in individuals with normal baseline blood pressure and ECGs, as well as those without left ventricular hypertrophy, underscoring its robustness.
AIRE-HTN also emerged as an independent predictor of significant cardiovascular outcomes:
- Cardiovascular death: Hazard ratio (HR) 2.24 per 1-SD increase in score
- Heart failure: HR 2.60
- Myocardial infarction: HR 3.13
- Ischemic stroke: HR 1.23
- Chronic kidney disease: HR 1.89
Implications for Practice
The authors emphasized the potential of AIRE-HTN to enhance hypertension screening and prevention programs. “Enhanced predictability could influence surveillance programs and primordial prevention,” they wrote, citing the biological plausibility of their findings.
The model showed a continuous net reclassification improvement over traditional clinical markers, with indices of 0.44 in the derivation cohort and 0.32 in the UK cohort.
Limitations and Future Directions
While promising, the study noted limitations. In one cohort, hypertension was defined using International Classification of Diseases (ICD) codes, which may lack alignment with current guidelines. Additionally, the model’s performance in diverse populations and clinical settings remains untested, and findings were not validated against ambulatory blood pressure monitoring standards.
Further studies are needed to explore AIRE-HTN’s applicability in broader clinical settings and among diverse demographic groups.
Funding and Disclosures
The study was led by Dr. Arunashis Sau and Joseph Barker from Imperial College London’s National Heart and Lung Institute. It was supported by the British Heart Foundation and other research grants. Some authors disclosed receiving personal fees, grants, and advisory fees unrelated to the study.
As AI continues to revolutionize medicine, tools like AIRE-HTN highlight the transformative potential of integrating technology into patient care to prevent and mitigate chronic conditions.
