Long-term changes in QT interval in hemodialysis patients

Abstract

Cardiovascular disease, particularly sudden cardiac death, remains a major challenge in dialysis patients whose hearts experience ongoing structural, electrophysiological, and neurohormonal stress. Among these alterations, QT interval prolongation—reflecting an imbalance in autonomic regulation—is closely associated with arrhythmic risk. Yet, its long-term trajectory in hemodialysis patients has not been systematically examined. In this study, we evaluated heart rate–corrected QT (QTc) intervals at 1, 4, 7, and 10 years after hemodialysis (HD) initiation in 80 patients. ECGs were obtained immediately before HD sessions, and QTc was calculated using the Bazett formula. Although an overall increasing trend was observed, a statistically significant prolongation was detected only at 10 years by Dunnett-type multiple comparison. These findings provide new insight into the gradual nature of QTc progression in HD patients and may contribute to improved cardiovascular risk stratification in clinical practice.

Main text and discussion

Patients on hemodialysis (HD) are in a particularly poor condition regarding cardiovascular (CV) risk, including sudden cardiac death (SCD). These patients experience continuous structural and neurohormonal stress, predisposing them to arrhythmia and heart failure. Beta-blockers, angiotensin-converting enzyme inhibitors, and angiotensin receptor blockers are widely recognized to reduce CV mortality in this vulnerable population. Although the renin-angiotensin-aldosterone system and the autonomic nervous system are intricately interconnected [1], CV death in HD patients appears to be strongly influenced by autonomic imbalance, as most SCDs were preceded by bradyarrhythmia [2]. A previous study [3] linked sympathetic overactivity to SCD in HD patients with left ventricular hypertrophy, focusing on heart rate variability (HRV) as an index of autonomic tone. The QT interval, spanning from the onset of ventricular depolarization to the end of repolarization, may also serve as an indicator of sympathetic–parasympathetic balance.

In a retrospective study of 102 HD patients [4], we evaluated the heart rate-corrected QT (QTc) interval at 1, 4, and 7 years after HD initiation. Patients receiving thrice-weekly 4-h HD sessions and with available pre-dialysis ECG recordings at these time points were included. QT intervals were corrected for heart rate using the Bazett formula. As a comparison cohort, age-matched individuals with normal renal function who were followed for lifestyle-related diseases or routine health screening and had at least two ECG recordings obtained more than four years apart were enrolled as controls. In both groups, any patient who had heart rate <57 beats per minute (bpm) or >103 bpm, non-sinus rhythms, or any instances of extrasystoles in their ECG reports were excluded to ensure accurate QTc assessment. QTc at 4 and 7 years were significantly longer than at 1 year, whereas no QTc prolongation was observed in the control group during a mean follow-up of six years.

Of the 102 patients, 60 had ECGs at 10 years after HD initiation. To examine long-term changes, the remaining 42 patients were observed using existing clinical records for up to 10 years; during this period, 9 died and 7 were lost to follow-up due to transfer. Among these, 6 were further excluded based on the predefined ECG exclusion criteria, consistent with our previous protocol [4]. Consequently, 80 patients were eligible for inclusion.

This study was approved by the Ethics Committee of Shibukawa Clinic (Approval Number: 005) on April 6, 2024, and adhered to the principles of the Declaration of Helsinki. Owing to the retrospective observational design, the requirement for written informed consent was waived by the Ethics Committee; however, informed consent was obtained from participants or, in the case of deceased participants, from their bereaved family members when required.

In this cohort, 38% were female. The mean age at 1 year after HD initiation was 56 ± 13 years, and diabetes was the primary cause of HD in 30% of patients. Their mean QTc intervals at 1, 4, 7, and 10 years after HD initiation were 436.2, 441.6, 440.1, and 450.2 msec, respectively (Figure 1). There appears to be an increasing trend over time, but a statistically significant rise was observed only at 10 years after HD initiation by Dunnett-type multiple comparison. In contrast, the 22 excluded patients were older (67 ± 13 years), had a higher prevalence of diabetes (45%), and included fewer females (27%). Although these patients had initially satisfied all inclusion criteria and underwent outpatient ECG evaluations at 1, 4, and 7 years, their subsequent exclusion due to death, transfer, or ECG-based criteria suggests increased clinical vulnerability during long-term follow-up.

Figure 1.

QTc interval changes after hemodialysis initiation.

Mean ± standard error of QTc interval at 1, 4, 7, and 10 years after HD initiation are shown. An increasing trend in QTc interval was observed over time. Dunnett-type multiple comparison revealed a significant difference between 1 year and 10 years after HD initiation (p < 0.001). QTc: heart rate–corrected QT; HD: hemodialysis; SE: standard error.

Several factors contribute to QT prolongation in HD patients. Myocardial fibrosis or remodeling due to hypertension, volume overload, or secondary cardiomyopathy, as well as ischemia from macro- or microcirculatory disturbances, can delay cardiac depolarization. Holter monitoring in HD patients has consistently shown reduced HRV [5], and sympathetic overactivity with vagal withdrawal has been linked to left ventricular mass [6], reflecting a reduction in autonomic flexibility typical of end-stage kidney disease. The progressive QTc increase in our study may therefore reflect an accumulation of structural, electrophysiological, and neurohormonal abnormalities over time.

Studies across different stages of chronic kidney disease have demonstrated that QTc lengthens progressively from stage 2 to stage 5 [7]. Along with our findings in the dialysis stage, these results suggest that QTc prolongation develops as kidney function declines, depending on both its severity and duration. A Japanese nationwide cohort study also demonstrated a significant association between QT prolongation and CV mortality within one year [8]. Given the reported relationship between NT-proBNP and QTc [9], periodic monitoring of QTc may provide valuable information for the ongoing management and risk stratification of dialysis patients.

In hemodialysis patients, QT interval duration is strongly influenced by rapid dialysis-related electrolyte shifts—particularly reductions in serum potassium and magnesium and fluctuations in calcium levels—and previous studies [10,11] have consistently demonstrated significant post-dialysis QTc prolongation, largely reflecting these acute metabolic shifts rather than intrinsic myocardial electrophysiological remodeling. To reduce such confounding effects and better capture baseline autonomic and myocardial influences on ventricular repolarization, ECGs in the present study were systematically obtained before dialysis sessions. This standardized timing likely improved the reliability and clinical interpretability of longitudinal QTc assessment, particularly for evaluating long-term cardiovascular risk and disease progression.

Our findings have several limitations. First, this study lacked a contemporary control group and detailed longitudinal data on electrolyte levels and oral medication use. Therefore, we cannot exclude the possibility that significant QTc prolongation might also occur over a longer 10-year period even in individuals with normal renal function. Second, survivorship bias is possible due to selection effects. The clinical differences observed between the 80 included patients and the 22 excluded patients were plausible, suggesting that our analytic cohort may represent relatively healthier HD patients. Importantly, the observation that a significant QTc increase was detected even among long-term survivors may further support the presence of a progressive trend toward QTc prolongation in HD patients. Third, regarding statistical methodology for longitudinal analysis, although linear mixed-effects models are often recommended, several factors made it difficult to reliably specify an appropriate functional form, including the limited number of measurement time points and the possibility that QTc changes may not follow a strictly linear pattern but instead involve periods of relative stability followed by phases of increase. Accordingly, we adopted a multiple comparison strategy to detect time-dependent differences without imposing strict parametric assumptions.

Funding Statement

The author(s) reported there is no funding associated with the work featured in this article.

Disclosure statement

The authors have no conflicts of interest to declare.

Data availability statement

All data generated or analyzed during the study are included in this article. Further enquiries can be directed to the corresponding author.