The prognostic value of electrocardiograms (ECGs) has been reported in predialysis patients but not in incident hemodialysis patients with overhydration and electrolyte disturbances, both of which potentially affect ECG results. We performed a retrospective multicenter cohort study involving incident hemodialysis patients and examined whether ECG parameters immediately before hemodialysis initiation can predict subsequent cardiovascular disease (CVD) using Cox proportional hazards models. We explored potential effect modifications by several electrolytes on the predictive power of ECG abnormalities. Among the 618 enrolled patients, 16%, 10%, 46%, and 22% showed a PR interval ≥ 200 ms, QRS interval ≥120 ms, QTc interval ≥ 450/460 ms (male/female), and left ventricular hypertrophy (LVH) by voltage criteria, respectively. Over a median 3-year follow-up, 19% and 16% of the patients developed atherosclerotic and nonatherosclerotic CVD, respectively. The Cox regression model results revealed that the sum of the number of abnormalities in PR, QRS, and QT intervals was a significant risk factor for nonatherosclerotic CVD (hazard ratios (HRs) [95% confidence interval (CI)]: 1.58 [1.24-2.01] per number of abnormalities). The predictive value of LVH for atherosclerotic CVD was attenuated over time. At up to 36 months, although the proportional hazards assumption was met, LVH was significantly associated with atherosclerotic CVD (HR [95% CI]: 1.89 [1.15-3.11]). The adjusted HR was particularly high (HR [95% CI]: 4.02 [1.68-9.60]) among patients who were in the lowest tertile of serum magnesium levels (P for interaction = 0.04). PR, QRS, and QT prolongation additively predicted nonatherosclerotic CVD, while LVH predicted atherosclerotic CVD in the short term.
Keywords: PR prolongation; QRS prolongation; QT prolongation; left ventricular hypertrophy.