Improved electrocardiographic detection of left ventricular hypertrophy

Area of Science:

• Cardiology
• Medical Diagnostics
• Biostatistics

Background:

• Left ventricular hypertrophy (LVH) is a significant indicator of cardiovascular disease.
• Accurate detection of LVH is crucial for timely intervention and management.
• Existing electrocardiogram (ECG) diagnostic algorithms have limitations in sensitivity and specificity.

Purpose of the Study:

• To develop and validate a novel multivariate linear regression (MLR) model for diagnosing LVH using ECG parameters.
• To compare the diagnostic performance of the new MLR model against established commercial ECG algorithms.
• To assess sex-specific differences in the diagnostic accuracy of LVH detection.

Main Methods:

• Utilized echocardiographic measurements and digital ECGs from 3378 participants in the Cardiovascular Health Study.
• Defined LVH based on echocardiographic left ventricular mass index (LVMI) thresholds for men and women.
• Developed sex-specific MLR equations using a learning set and validated them on a separate test set, comparing against commercial ECG algorithms.

Main Results:

• The developed MLR equations demonstrated superior diagnostic performance for LVH compared to three commercial computerized ECG algorithms in both men and women.
• The MLR model and commercial algorithms showed higher diagnostic sensitivity in men than in women at matched specificities.
• The new MLR equation reliably identified echocardiographically determined LVH.

Conclusions:

• The novel MLR equation provides a more reliable method for detecting LVH using ECGs than current commercial algorithms.
• Sex-specific differences in diagnostic sensitivity exist, with men generally showing higher sensitivity.
• This approach offers a promising tool for improving LVH diagnosis in clinical practice.