Prognostic value of chronic ECG changes in patients undergoing hemodialysis

Patients on maintenance hemodialysis (HD) face high cardiovascular (CV) mortality risk. Although electrocardiograms (ECGs) are routinely performed, the prognostic significance of chronic ECG abnormalities in predicting mortality among HD patients remains unclear. We retrospectively analyzed patients with kidney failure receiving maintenance HD at a single center between January 2000 and March 2022. ECGs obtained within two years of the end of follow-up were independently reviewed by two physicians. Multivariable logistic regression was used to identify predictors of mortality, integrating clinical and ECG parameters. Model performance was assessed using calibration and discrimination statistics. Of 291 records, 149 were included (mean age 67.1 ± 13.2 years; 64.4% male; median dialysis vintage 47 months [IQR 25–87]). Common comorbidities included hypertension (90%), diabetes (54.4%), and coronary artery disease (42.4%). Frequent ECG findings were ST depression (33.8%), left axis deviation (32.9%), and poor R-wave progression (26.4%), with sinus rhythm in 94%. Overall mortality was 55%, including31.7% sudden cardiac deaths. The final model showed good calibration and excellent discrimination (AUC 0.80 ± 0.04, <i>P</i> &lt; 0.001). ECG abnormalities enhance mortality risk prediction in HD patients and should be validated in larger prospective cohorts. Patients who receive regular hemodialysis (HD) for kidney failure have a much higher risk of dying from heart-related problems. Doctors often use an electrocardiogram (ECG); a simple heart test that records the electrical activity of the heart, to look for heart disease or rhythm problems. However, it is not well known whether long-term ECG changes can help predict which HD patients are at higher risk of dying.In this study, we reviewed medical records from patients on maintenance HD between 2000 and 2022 at one dialysis center. We analyzed ECG results taken during the last two years of follow-up and combined them with other health data such as age, diabetes, and heart disease. Using statistical models, we identified ECG features that were linked with a higher risk of death. Among 149 patients, more than half died during the study period, and about one-third died suddenly from cardiac arrest. Certain ECG changes (such as ST depression, left axis deviation, and poor R-wave progression) were common and helped predict death risk when combined with other clinical factors. Our findings suggest that routine ECG testing could help identify HD patients at higher risk of death, allowing earlier and more personalized care.