Phenotyping of Heart Failure in CKD Using Electrocardiography Features.

Area of Science:

• Cardiology
• Nephrology
• Medical Informatics

Background:

• Predictive tools for heart failure (HF) in patients with chronic kidney disease (CKD) are limited.
• This study investigated the utility of standard electrocardiogram (ECG) features and heart rate variability (HRV) parameters in predicting de novo HF hospitalization in individuals with CKD.

Purpose of the Study:

• To determine if standard ECG parameters or HRV can predict incident HF hospitalization in CKD patients.
• To identify specific ECG features associated with increased HF hospitalization risk.

Main Methods:

• A cohort of 11,409 CKD patients from the NYU ECG database (2012-2021) was analyzed.
• Standard ECG features and HRV parameters were extracted from baseline ECGs prior to dialysis, ESKD, or transplant.
• LASSO-penalized Cox regression and Fine-Gray competing risk models were used to identify predictors of HF hospitalization.

Main Results:

• 880 patients (8%) experienced HF hospitalization. Models integrating ECG and clinical data showed strong discrimination (C-statistic 0.73-0.76).
• PR interval, corrected QT, and T axis on ECG were independently associated with higher HF hospitalization risk (p<0.01).
• History of arrhythmia, valvular disease, and diabetes were significant clinical predictors; HRV parameters were not independently associated with HF.

Conclusions:

• Standard ECG features, specifically PR interval, corrected QT, and T axis, are associated with HF hospitalization risk in CKD patients.
• ECG-derived HRV indices did not independently predict HF hospitalization in this cohort.