Abstract

Aims

A high prevalence of prolonged QT interval duration has been observed among haemodialysis (HD) patients. The aim of this cases series was to describe the association of various risk factors with total mortality and sudden cardiac death (SCD) in this population.

Methods and results

One hundred and twenty-two patients undergoing HD, [median: age 71.3 years [interquartile ratio (IQR) 62.9–76.6], HD duration 3.0 years (IQR 1.3–7.8) and 64.8% male], of which 37.7% with ischaemic cardiac disease, 41.8% with dilated cardiomyopathy (DC), 84.4% with hypertension, and 27.1% with diabetes, were studied. Median left ventricular ejection fraction (LVEF) was 60.0% (IQR 52–64) and left ventricular mass index (LVMI) was 147.3 g/m2 (IQR 128.0–179.9). QT interval duration corrected for heart rate (QTc) was measured by electrocardiogram Holter recording and considered prolonged when longer than 450 ms in men and 460 ms in women. Forty-four patients (36.0%) had a prolonged QTc. Female gender (P < 0.001) and DC (P = 0.018) were associated with a longer QTc, while LVEF (P = 0.012) was inversely related. During the study period (median follow-up 3.9 years), 51 patients died (41.8%), of whom 12 died for SCD. In multivariate analysis age at recruitment [HR = 1.07, 95% confidence interval (CI): 1.03–1.11, P < 0.001], prolonged QTc (HR = 2.16, 95% CI: 1.20–3.91, P = 0.011) and presence of DC (HR = 3.75, 95% CI: 1.01–7.00, P < 0.001) were independently associated with total mortality, while only a prolonged QTc (HR = 8.33, 95% CI: 1.71–40.48, P = 0.009) and increasing LVMI (HR = 1.01, 95% CI: 1.00–1.02, P = 0.022) were associated with SCD.

Conclusions

In a case series of HD patients, QTc was associated with total mortality and SCD. Further studies to test this hypothesis in a larger population are necessary.

What's new?
  • QTc duration is independently associated with total mortality in haemodialysis (HD) patients.

  • QTc duration is independently associated with sudden cardiac death in HD patients.

Introduction

Despite significant improvements in the haemodialysis (HD) techniques, patients with end-stage renal disease (ESRD) have a low life expectancy. Cardiovascular diseases are the main cause of mortality in these patients, accounting for more than half of the deaths with an incidence 10-fold greater than in the general population.1 Moreover the 3-year cumulative incidence of sudden cardiac death (SCD) is ∼7%,2 in comparison with 0.1–0.2% per year in the general population.3 In addition, the HD sessions may exert detrimental effects on the cardiovascular system, due to the rapid changes in electrolyte plasma levels, the fast reduction of plasma volume, and the fast correction of acidosis. Owing to these effects, the HD session may be considered an arrhythmogenic condition as shown by the high incidence of arrhythmias recorded during and immediately after the HD sessions.4,5 Indeed, different studies2,6 reported that the highest incidence of SCD has peak rates within 12 h after the first short interdialytic interval and during the long-weekend interval.

On the basis of these findings the identification of HD patients with higher arrhythmic risk, by using non-invasive SCD predictors, is extremely important. A prolongation of QTc interval is associated with a higher risk of mortality and SCD in cardiovascular diseases and even in healthy individuals7,8 and has been proposed as a possible risk factor also in HD patients. Different studies have shown that the HD sessions induce a prolongation of the QTc interval and that these acute alterations are mainly related to the HD-induced changes of calcium and potassium plasma concentration.5,9–11 However, most studies analysed the duration of ventricular repolarization in ESRD patients, but very few data are available on its prognostic value in this population. A retrospective study by Beaubien et al.12 showed that a greater QTc interval dispersion was associated with an increase of total and cardiovascular mortality in a population of HD and peritoneal dialysis patients. A recent study by Hage et al.,13 performed in patients evaluated for renal transplantation, reported that QTc interval prolongation was an independent predictor of mortality.

The aim of this case series of chronic HD patients was to describe the various risk factors for total mortality and SCD, in particular, to assess the prognostic value of the QTc interval duration, analysed by electrocardiogram (ECG) Holter monitoring.

Methods

We enrolled 177 consecutive patients on chronic HD treatment at the San Gerardo Hospital (Monza, Italy) from January 2005 to October 2010. In our Centre from January 2005 all HD patients routinely undergo ECG Holter recording and cardiac ultrasound once a year. Patients with complete bundle blocks (n = 28), permanent atrial fibrillation or atrial flutter (n = 21), and with cardiac pacemaker (n = 6) were excluded from the study as the presence of these conditions prevents the possibility of QT interval measurement. The remaining 122 patients were prospectively followed until September 2012. Data collection included clinical and demographic characteristics, record charts, and reports of cardiac ultrasound. The median follow-up time was 3.9 years (IQR: 2.4–6.2 years). All patients underwent three HD sessions per week, for a total of two short interdialytic intervals (2 days), and a long interdialytic interval (3 days).

Follow-up was performed by physicians responsible for the HD sessions. Deaths and cause of death were registered in an electronic clinical chart. All causes of death were revised by an expert cardiologist. Sudden cardiac death was defined as an unexpected death occurring within 1h of the onset of symptoms, as stated by American College of Cardiology/American Heart Association/European Society of Cardiology guidelines.14,15 All deaths registered as sudden deaths were confirmed by interviews with the patient's physician or family members.

Procedures were performed according to the Helsinki declaration for ethical treatment of human subjects and were approved by the local ethical committee.

Comorbidities and treatments

Several comorbidities have been collected: ischaemic heart disease (IHD), diabetes mellitus (DM), hypertension, dilated cardiomyopathy (DC), valvular heart disease, dyslipidaemia, and ischaemic cerebral disease (ICD).

Ischaemic heart disease included previous myocardial infarction or coronary revascularization procedures, such as angioplasty or coronary bypass graft. Hypertension was defined as systolic blood pressure equal or greater than 140 mmHg and/or diastolic blood pressure equal or greater than 90 mmHg before the beginning of HD session or when antihypertensive drugs were administered.

Dilated cardiomyopathy was defined as left ventricular dilation (left ventricular end-diastolic diameter >55 mm) at ultrasound examination.

Valvular heart disease was determined as the presence of echocardiography documentation of moderate-to-severe mitral or aortic stenosis and/or regurgitation.

Dyslipidaemia was considered in the presence of clinical history or when low-density lipoprotein cholesterol was equal or greater than 130 mg/dL.

Ischaemic heart disease was defined as one or more documented previous stroke or transient ischaemic attack.

Treatments with renin–angiotensin system inhibitors, alpha-blockers, amiodarone, antiplatelets, beta-blockers, calcium-channel blockers, digoxin, statins, and oral anticoagulants have been taken into consideration.

Laboratory tests

The following haematochemical variables were obtained from the mean of pre-dialytic values recorded during the last month before the recruitment: haemoglobin (g/dL), albumin (g/L), phosphate (mmol/L), potassium (mmol/L), total (mmol/L) and ionized calcium (mmol/L), and parathormone (pmol/L).

Cardiac ultrasound examination

Cardiac ultrasound examination was performed in all patients during the mid-week dialysis interval. The echocardiograms were obtained in the standard precordial positions using digital echocardiography equipment (Aloka ProSound SSD Alpha 10). We followed the recommendations for standard measurements from M-mode echocardiograms.16 Instantaneous measurements were made from three cardiac cycles and the average values of the following parameters were obtained: septal wall thickness at end-diastole, left ventricular posterior wall thickness at end-diastole, and left ventricular internal diameter at the end of diastole. Left ventricular mass (LVM) was calculated according to the formula modified by Devereux16 using the American Society of Echocardiography convention and indexed for body surface area [LVM index (LVMI)]. Left ventricular ejection fraction (LVEF) was calculated by 2D measurements for volume calculations using biplane method of disks (modified Simpson's rule) in apical four-chamber and apical two-chamber views and indexed for body surface area.

Electrocardiographic Holter recordings and analysis

A 24 h ECG Holter monitoring was recorded in each subject by using a portable battery-operated three-channel Holter recorder. The digitized three-channel ECG signals were processed by the Synescope Holter analysis software (Sorin Group Company), which sampled the 24 h recording into 2880 templates obtained by 30 s time intervals. To improve the signal-to-noise ratio, one median complex was computed every 6 s from the consecutive sinus beats; then the five median beats within each 30 s template were averaged to obtain single representative PQRST complexes for each of the 2880 templates. For each template, an algorithm automatically measured the QT and the RR interval (ms). The programme also provided the mean of QT intervals corrected for heart rate (HR) according to the Bazett's formula (QTc) in pre-specified time periods: (i) 4h of HD session (HD-QTc), (ii) 4h after HD session (post-HD-QTc), and (iii) the remaining 16h of recording (basal QTc). QTc was considered prolonged when equal to or greater than 450 ms in men, and 460 ms in women.17

Statistical analysis

Demographic and clinical characteristics were described by using median and interquartile range for skewed continuous variables and percentages for qualitative variables.

Linear regression was applied to identify factors related to an increase of basal QTc duration, first as univariate and subsequently as a multivariate model (this latter considered only the factors significantly associated with basal QTc duration at a univariate analysis). P values were considered significant when <0.20 in the univariate analysis and when <0.05 in the multivariate analysis.

The distribution of QTc duration in different subgroups was represented and tested by means of box-plot and analysis of variance (ANOVA) regression model, while proportion of events were compared between normal and prolonged QTc by means of test for proportions.

The crude cumulative incidence of death was estimated for total mortality, according to Kaplan–Meier, and for SCD competing risks, by using the Aalen–Johansen estimator. Comparison of crude cumulative incidence was performed by log-rank test and by Grey's test, for total mortality and for SCD, respectively. Observation time was censored at 30 September 2012 for patients who did not present an event before.

To investigate the effect of a prolonged basal QTc on the overall hazard of death and cause-specific hazard of SCD, a Cox regression model was used, adjusting for the clinically relevant variables and confounding factors: age and duration of HD therapy at enrolment, IHD, DC, DM, LVMI, LVEF, plasma potassium, and phosphate. All the relevant variables and confounding factors included in the multivariate analysis were previously tested in a univariate analysis and excluded from the multivariate analysis if P > 0.20. Owing to the high number of significant factors at the univariate analysis, a backward selection was applied to the multivariate model to identify only those mainly related to the increase of death and SCD risk. The assumption of proportional hazards was assessed by means of graphical check on the log cumulative hazard for each covariate, and major departures were not detected. Results of the Cox model are expressed in terms of estimated hazard ratios (HRs) and 95% confidence intervals (95% CIs). All computations were carried out using SAS, version 9.2 (SAS Institute).

Results

Clinical and demographic characteristics of the studied population are described in Table 1.

Table 1

Characteristics of the 122 patients at enrolment.

Patients characteristics Median (nIQR (%) 
Age (years) 71.3 62.9–76.6 
Haemodialysis duration (years) 3.0 1.3–7.8 
Gender (male) 79 64.8 
Comorbidities (%)a   
 Ischaemic heart disease 46 37.7 
 Diabetes mellitus 33 27.1 
 Hypertension 103 84.4 
 Dilated cardiomyopathy 51 41.8 
 Valvular heart disease 29 23.8 
 Dyslipidaemia 22 18.0 
 Ischaemic cerebral disease 18 14.8 
 Atrial fibrillation (paroxysmal and/or persistent) 51 41.8 
 Haemodialysis modality   
 BHD (bicarbonate dialysis) 106 86.9 
 HDF (haemodiafiltration) 16 13.1 
 K+ in dialysis bath (mmol/L)   
 2 58 47.5 
 3 61 50.0 
 4 2.5 
Drugsb   
 Angiotensin system inhibitors 50 41.0 
 Alpha-blockers 29 23.7 
 Amiodarone 34 27.9 
 Antiplatelets agents 69 56.6 
 Beta-blockers 47 38.5 
 Calcium-channel blockers 42 34.4 
 Digoxin 4.1 
 Statin 36 29.5 
 Oral anticoagulant therapy 17 13.9 
Haematochemical parameters   
 Haemoglobin (g/dL) 10.6 10.1–11.2 
 Haematocrit (%) 32.7 31.3–34.4 
 Albumin (g/L) 0.39 0.36–0.42 
 Phosphate (mmol/L) 1.39 1.19–1.65 
 Potassium (mmol/L) 4.9 4.6–5.2 
 Total calcium (mmol/L) 2.2.3 2.15–2.33 
 Ionized calcium (mmol/L) 1.2 1.1–1.2 
 Parathormone (n = 121) (pmol/L) 26.7 17.0–41.8 
Cardiac ultrasound parameters   
 Left ventricular ejection fraction (n = 120) (%) 60.0 52.0–64.0 
 Left ventricular mass index (n = 119) (g/m2147.3 128.0–179.9 
Patients characteristics Median (nIQR (%) 
Age (years) 71.3 62.9–76.6 
Haemodialysis duration (years) 3.0 1.3–7.8 
Gender (male) 79 64.8 
Comorbidities (%)a   
 Ischaemic heart disease 46 37.7 
 Diabetes mellitus 33 27.1 
 Hypertension 103 84.4 
 Dilated cardiomyopathy 51 41.8 
 Valvular heart disease 29 23.8 
 Dyslipidaemia 22 18.0 
 Ischaemic cerebral disease 18 14.8 
 Atrial fibrillation (paroxysmal and/or persistent) 51 41.8 
 Haemodialysis modality   
 BHD (bicarbonate dialysis) 106 86.9 
 HDF (haemodiafiltration) 16 13.1 
 K+ in dialysis bath (mmol/L)   
 2 58 47.5 
 3 61 50.0 
 4 2.5 
Drugsb   
 Angiotensin system inhibitors 50 41.0 
 Alpha-blockers 29 23.7 
 Amiodarone 34 27.9 
 Antiplatelets agents 69 56.6 
 Beta-blockers 47 38.5 
 Calcium-channel blockers 42 34.4 
 Digoxin 4.1 
 Statin 36 29.5 
 Oral anticoagulant therapy 17 13.9 
Haematochemical parameters   
 Haemoglobin (g/dL) 10.6 10.1–11.2 
 Haematocrit (%) 32.7 31.3–34.4 
 Albumin (g/L) 0.39 0.36–0.42 
 Phosphate (mmol/L) 1.39 1.19–1.65 
 Potassium (mmol/L) 4.9 4.6–5.2 
 Total calcium (mmol/L) 2.2.3 2.15–2.33 
 Ionized calcium (mmol/L) 1.2 1.1–1.2 
 Parathormone (n = 121) (pmol/L) 26.7 17.0–41.8 
Cardiac ultrasound parameters   
 Left ventricular ejection fraction (n = 120) (%) 60.0 52.0–64.0 
 Left ventricular mass index (n = 119) (g/m2147.3 128.0–179.9 

For continuous variables median and interquartile range are reported; frequencies are reported for qualitative variables.

aTwo or more comorbidities may coexist.

bTwo or more drugs may coexist.

Ventricular repolarization duration and its predictors

QTc interval distribution by gender is shown in Figure 1. As expected, basal, HD, and post-HD-QTc were longer in women than in men. QTc did not change substantially between the pre-specified times. A prolonged basal QTc was present in 44 patients (36%) of whom 17 were females.

Figure 1

Distribution of basal QTc, HD-QTc, and post-HD-QTc duration by gender. The box includes observation from the 25th to 75th percentiles (the line and the square indicate the median and mean), whiskers are minimum and maximum values.

Figure 1

Distribution of basal QTc, HD-QTc, and post-HD-QTc duration by gender. The box includes observation from the 25th to 75th percentiles (the line and the square indicate the median and mean), whiskers are minimum and maximum values.

Univariate analysis of the determinants of basal QTc duration showed a direct relation with LVMI (P = 0.003) and an inverse relation with LVEF (P < 0.001), albumin plasma value (P = 0.014), and total calcium (P = 0.112). Female gender (P = 0.014), presence of IHD (P = 0.007), presence of hypertension (P = 0.169), presence of DC (P < 0.001), and amiodarone therapy (P = 0.036) were associated with a longer basal QTc, while alpha-blocker therapy (P = 0.011) and digoxin (P = 0.128) were associated with a shorter basal QTc duration. The multivariate linear regression model showed that the only independent predictors of basal QTc duration were gender (P < 0.001), LVEF (P = 0.020), and DC (P = 0.010) (Table 2). On an average, females had a 25 ms longer basal QTc compared with males; presence of DC was related to a 16 ms longer QTc duration; every 10% of LVEF reduction a 8 ms increase in QTc duration has been observed.

Table 2

Results from the univariate and multivariate regression linear models on basal QTc duration, adjusted by gender

 Univariate analysis Multivariate analysis
 
Variables P value β (SE) P value 95% CI 
Age at recruitment 0.673    
Haemodialytic duration 0.298    
Gender (F) 0.014 24.9 (6.1) <0.001 (13.1; 36.8) 
Ischaemic heart disease 0.007 2.8 (6.2) 0.654 (−9.5; 15.0) 
Diabetes mellitus 0.675    
Hypertension 0.169 0.8 (7.9) 0.924 (−15.0; 16.5) 
Dilated cardiomyopathy <0.001 16.2 (6.1) 0.010 (4.0; 28.3) 
Dyslipidaemia 0.621    
Ischaemic cerebral disease 0.779    
Bicarbonate dialysis 0.371    
Angiotensin system inhibitors 0.604    
Alpha-blockers 0.011 −9.9 (6.6) 0.138 (−22.9; 3.2) 
Amiodarone 0.036 10.9 (6.1) 0.078 (−1.9; 23.1) 
Antiplatelets 0.797    
Beta-blockers 0.444    
Calcium antagonists 0.932    
Digoxin 0.128 −15.3 (13.5) 0.259 (−12.1; 11.4) 
Statin 0.977    
Oral antithrombotic therapy 0.818    
Haemoglobin (g/dL) 0.436    
Haematocrit (%) 0.659    
Albumin (g/dL) 0.014 −6.6 (6.6) 0.322 (−19.6; 6.5) 
Phosphate (mg/dL) 0.980    
Potassium (mmol/L) 0.384    
Calcium total (mg/dL) 0.112 −5.3 (4.4) 0.229 (−13.9; 3.4) 
Calcium ionized (mmol/L) 0.748    
Parathormone (n = 121) (pg/mL) 0.916    
Left ventricular ejection fraction (%) <0.001 −0.8 (0.3) 0.020 (−1.4; −0.1) 
Left ventricular mass index (g/m20.003 0.1 (0.1) 0.089 (−0.0; 0.3) 
 Univariate analysis Multivariate analysis
 
Variables P value β (SE) P value 95% CI 
Age at recruitment 0.673    
Haemodialytic duration 0.298    
Gender (F) 0.014 24.9 (6.1) <0.001 (13.1; 36.8) 
Ischaemic heart disease 0.007 2.8 (6.2) 0.654 (−9.5; 15.0) 
Diabetes mellitus 0.675    
Hypertension 0.169 0.8 (7.9) 0.924 (−15.0; 16.5) 
Dilated cardiomyopathy <0.001 16.2 (6.1) 0.010 (4.0; 28.3) 
Dyslipidaemia 0.621    
Ischaemic cerebral disease 0.779    
Bicarbonate dialysis 0.371    
Angiotensin system inhibitors 0.604    
Alpha-blockers 0.011 −9.9 (6.6) 0.138 (−22.9; 3.2) 
Amiodarone 0.036 10.9 (6.1) 0.078 (−1.9; 23.1) 
Antiplatelets 0.797    
Beta-blockers 0.444    
Calcium antagonists 0.932    
Digoxin 0.128 −15.3 (13.5) 0.259 (−12.1; 11.4) 
Statin 0.977    
Oral antithrombotic therapy 0.818    
Haemoglobin (g/dL) 0.436    
Haematocrit (%) 0.659    
Albumin (g/dL) 0.014 −6.6 (6.6) 0.322 (−19.6; 6.5) 
Phosphate (mg/dL) 0.980    
Potassium (mmol/L) 0.384    
Calcium total (mg/dL) 0.112 −5.3 (4.4) 0.229 (−13.9; 3.4) 
Calcium ionized (mmol/L) 0.748    
Parathormone (n = 121) (pg/mL) 0.916    
Left ventricular ejection fraction (%) <0.001 −0.8 (0.3) 0.020 (−1.4; −0.1) 
Left ventricular mass index (g/m20.003 0.1 (0.1) 0.089 (−0.0; 0.3) 

QTc interval and mortality

During the study period, 51 patients died. Death occurred in 25 of the 78 patients with a normal basal QTc duration (32.0%) and in 25 of the 44 subjects with prolonged QTc (59.1%), with a significant difference (P = 0.004; Table 3). Cardiovascular non-sudden death occurred in 5.1% of patients with normal QTc and in 9.1% of those with prolonged QTc interval (P = 0.396). Also SCD was observed more frequently in patients with prolonged QTc than in those with normal ventricular repolarization (22.7 vs. 2.6%, P < 0.001; Table 3). Among the 12 SCD, 9 occurred during the long interdialytic interval, while the remaining 3 in the first short interval.

Table 3

Causes of death in patients with normal and prolonged QTc interval

Cause of death Normal QTc (n = 78) Prolonged QTc (n = 44) Total (n = 122) 
n (%) n (%) n (%) 
Cardiovascular 6 (7.7) 14 (31.8) 20 (16.4) 
 Sudden 2 (2.6) 10 (22.7) 12 (9.8) 
 Non-sudden 4 (5.1) 4 (9.1) 8 (6.6) 
Non-cardiovascular 19 (24.3) 12 (27.3) 31 (25.4) 
 Sepsis 2 (2.6) 5 (11.4) 7 (5.7) 
 Neoplasia 4 (5.1) 3 (6.8) 7 (5.7) 
 Cachexia 10 (12.8) 3 (6.8) 13 (10.7) 
 Other 3 (3.8) 1 (2.3) 4 (3.3) 
Total events 25 (32.0) 26 (59.1) 51 (41.8) 
Cause of death Normal QTc (n = 78) Prolonged QTc (n = 44) Total (n = 122) 
n (%) n (%) n (%) 
Cardiovascular 6 (7.7) 14 (31.8) 20 (16.4) 
 Sudden 2 (2.6) 10 (22.7) 12 (9.8) 
 Non-sudden 4 (5.1) 4 (9.1) 8 (6.6) 
Non-cardiovascular 19 (24.3) 12 (27.3) 31 (25.4) 
 Sepsis 2 (2.6) 5 (11.4) 7 (5.7) 
 Neoplasia 4 (5.1) 3 (6.8) 7 (5.7) 
 Cachexia 10 (12.8) 3 (6.8) 13 (10.7) 
 Other 3 (3.8) 1 (2.3) 4 (3.3) 
Total events 25 (32.0) 26 (59.1) 51 (41.8) 

QTc duration differed according to outcome both in females and males (Figure 2). The ANOVA regression confirmed that QTc duration significantly differed according to patient outcome, especially in males (P < 0.001), while in females the low number of events did not allow to reach a statistical significance (P = 0.059). QTc interval was longer in patients who died than in survivors, especially patients who suffered of SCD tended to have the longest durations.

Figure 2

Distribution of QTc duration by gender and outcome. The box includes observation from the 25th to 75th percentiles (the line and the square indicate the median and mean), whiskers are minimum and maximum values.

Figure 2

Distribution of QTc duration by gender and outcome. The box includes observation from the 25th to 75th percentiles (the line and the square indicate the median and mean), whiskers are minimum and maximum values.

Figure 3 shows the cumulative incidence curves of total mortality (left) and SCD (right) in patients with prolonged and with normal basal QTc interval. Both total mortality and SCD were significantly higher in patients with prolonged QTc interval (P < 0.001).

Figure 3

Cumulative incidence of total mortality (A) and SCD (B) by QTc interval.

Figure 3

Cumulative incidence of total mortality (A) and SCD (B) by QTc interval.

Among the 44 patients with prolonged QTc, 18 had a marked prolongation (QTc ≥ 480 ms) and their outcome was strikingly poor, with 13 deaths observed, of whom 6 were SCD (33.3%). Of the remaining 26 patients with moderately prolonged QTc, 13 died, of whom 4 for SCD (15.4%).

The univariate Cox regression models showed that all the clinically relevant variables were related to the overall mortality, except for duration of HD (HR = 1.06). Prolonged QTc (HR = 2.72), older age (HR = 1.06), presence of IHD (HR = 1.53), DM (HR = 1.61), DC (HR = 3.34), and higher LVMI (HR = 1.01) were positively associated to an increase of mortality risk from any cause. Higher phosphate level (HR = 0.83), higher potassium level (HR = 0.60), and higher LVEF% (HR = 0.95) seemed to reduce the risk of death (Table 4). The cause-specific univariate models on SCD run on the same variables showed that age at recruitment (HR = 1.04), duration of HD (HR = 1.34), presence of DM (HR = 0.91), higher phosphate (HR = 0.80), and higher potassium level (HR = 0.60) were not associated to the risk of SCD. Prolonged QTc (HR = 11.90), presence of IHD (HR = 2.44), presence of DC (HR = 3.44), and higher LVMI (HR = 1.02) were positively related to an increasing SCD risk, while higher LVEF (HR = 0.93) was negatively related.

Table 4

Cox regression model for total mortality and for cause-specific SCD.

Variables Univariate analysis
 
Multivariate analysis*
 
Total mortality SCD Total mortality
 
SCD
 
P value P value HR 95% CI P value HR 95% CI P value 
Basal QTc (prolonged vs. normal) <0.001 0.001 2.16 (1.20–3.91) 0.011 8.33 (1.71–40.48) 0.009 
Age at recruitment (years) <0.001 0.241 1.07 (1.03–1.11) <0.001    
Duration of haemodialysis (≥3 vs. <3 years) 0.847 0.618       
Ischaemic heart disease (yes vs. no) 0.129 0.129       
Diabetes mellitus (yes vs. no) 0.117 0.886       
Dilated cardiomyopathy (yes vs. no) <0.001 0.044 3.75 (1.01–7.00) <0.001    
Phosphate (mmol/L) 0.152 0.427       
Potassium (mmol/L) 0.108 0.428       
Left ventricular ejection fraction (%) <0.001 0.003       
Left ventricular mass index (g/m20.010 <0.001    1.01 (1.00–1.02) 0.022 
Variables Univariate analysis
 
Multivariate analysis*
 
Total mortality SCD Total mortality
 
SCD
 
P value P value HR 95% CI P value HR 95% CI P value 
Basal QTc (prolonged vs. normal) <0.001 0.001 2.16 (1.20–3.91) 0.011 8.33 (1.71–40.48) 0.009 
Age at recruitment (years) <0.001 0.241 1.07 (1.03–1.11) <0.001    
Duration of haemodialysis (≥3 vs. <3 years) 0.847 0.618       
Ischaemic heart disease (yes vs. no) 0.129 0.129       
Diabetes mellitus (yes vs. no) 0.117 0.886       
Dilated cardiomyopathy (yes vs. no) <0.001 0.044 3.75 (1.01–7.00) <0.001    
Phosphate (mmol/L) 0.152 0.427       
Potassium (mmol/L) 0.108 0.428       
Left ventricular ejection fraction (%) <0.001 0.003       
Left ventricular mass index (g/m20.010 <0.001    1.01 (1.00–1.02) 0.022 

*Results from a backward selection.

The adjusted multiple Cox regression model on the significant factors, applying a backward selection, showed that the only predictors of total mortality were older age at recruitment (P < 0.001), the presence of DC (P < 0.001), and a prolonged basal QTc (P = 0.011), with a two-fold increase in the hazard of death for patients with prolonged QTc. Prolonged basal QTc (P = 0.009) and LVMI (P = 0.022) were the only significant predictor of SCD (P = 0.022; Table 4). A similar prognostic value for total mortality was observed when prolonged HD-QTc (HR = 2.87, P = 0.001) or post-HD-QTc duration (HR = 2.18, P = 0.017) were used in the same model. Only HD-QTc, but not post-HD-QTc was a significant predictor of SCD.

The expected 4-year total and SCD mortality was estimated based on the Cox models. Probability curves of total mortality were traced for patients with normal QTc and without DC, normal QTc and DC, prolonged QTc and without DC, and prolonged QTc and DC, while curves of SCD mortality were traced for patients with normal QTc and LVMI = 100 g/m2, normal QTc and LVMI = 200 g/m2, prolonged QTc and LVMI = 100 g/m2, and prolonged QTc and LVMI = 200 g/m2 (Figure 4). A higher probability of both total and SCD mortality was estimated when prolonged QTc was associated to the presence of DC or higher LVMI, respectively.

Figure 4

Estimated curves of probability of death and SCD according to different patient profiles.

Figure 4

Estimated curves of probability of death and SCD according to different patient profiles.

Discussion

The present study shows that in a population of patients with ESRD undergoing HD therapy: Previous studies showed that ESRD patients may develop ventricular repolarization duration abnormalities during or after the HD sessions.18–20 To our knowledge, only one study13 evaluated the QTc interval duration in a relatively large sample of patients with ESRD. Among 196 patients undergoing HD 84 (43%) showed a prolongation of the QTc interval, a prevalence similar to that observed in the present study.

  1. There is a high prevalence of subjects with prolonged ventricular repolarization (36%);

  2. Female gender, presence of DC, and decreased LVEF are independently related to an increase in QTc duration;

  3. QTc prolongation, age, and DC are independently associated with total mortality, while QTc prolongation and LVMI are associated with SCD.

The mechanisms underlying an alteration in ventricular repolarization duration in ERSD patients remains to be elucidated. The duration of ventricular action potential is controlled by different ionic currents, among which potassium and calcium currents play a major role. In uraemic patients, the HD session can induce an increase in QTc interval,10 which largely depends on calcium and potassium concentrations in the dialysis bath.5,9–11 A tighter control of the dialysis bath composition has been applied in our HD unit. Accordingly, we did not observe dramatic changes in plasma electrolytes concentrations during and after the HD sessions. This may partly explain the lack of HD-related QTc interval increase in our patients. Nevertheless, a high prevalence of QTc interval prolongation is still shown in our population. This observation indicates that, in addition to plasma electrolyte concentration, also non-conventional cardiovascular risk factors due to the uraemic condition may have importance in determining electrocardiographic alterations.

In the present study, besides gender, as expected,21 a significant predictor of QTc interval duration was the presence of DC and a negative inverse correlation between LVEF and QTc interval was also found. Thus, ESRD patients with a greater degree of cardiac damage may be more predisposed to an alteration in ventricular repolarization duration. It is possible that HD patients develop cardiac fibrosis and calcification, which may partly explain these alterations of ventricular repolarization.

In our population HD patients with a prolonged QTc interval had a significantly higher mortality than patients with normal QTc and mortality increased with more marked prolongation. Accordingly, prolonged QTc interval was independently associated with total mortality. Age and the presence of DC were also independently associated with total mortality and the presence of both QTc interval prolongation and cardiac dilatation was associated with a higher risk of death.

Few data are available on the prognostic value of QTc interval in ESRD patients. A retrospective study by Beaubien et al.12 showed that mortality was associated with a greater QTc interval dispersion, but not a longer QTc duration, in a population of dialysis patients. In a recent prospective study by Hage et al.,13 QTc interval prolongation was an independent predictor of mortality. However, this study was performed in a selected population of patients evaluated for renal transplantation, who underwent coronary angiography for suspected IHD. Unlike the study by Hage, the QTc interval duration was analysed in our study by using ECG Holter monitoring, thus allowing a long-term dynamic analysis and the measurements in different periods including HD session. Moreover, the availability of information on the causes of death allowed the assessment of the effect of QTc interval duration, adjusted for several other variables, not only on the risk of total mortality but also of SCD. Sudden death is common in ESRD and accounts for ∼25% of lethal events. In the present study, QTc interval prolongation and LVMI were the only factors significantly associated with SCD, and the presence of both QT interval prolongation and high LVMI was associated with a higher risk of sudden death. Ventricular hypertrophy in both acquired22,23 and inherited24 diseases may predispose to arrhythmias and sudden death. A previous study reported an association between an increased LVMI and sudden death in HD patients.25 QTc interval prolongation may favour torsade de pointes ventricular tachycardia that may degenerate into ventricular fibrillation, leading to SCD. Previous studies showed that high pre-HD session serum potassium levels (>6.0 mmol/L)2 and potassium dialysate concentrations (<2 mmol/L)26 were significantly associated with an increased risk of SCD. In the present study serum potassium levels were not independently associated with SCD and this may be partly explained by the approach of our dialysis centre, where modifiable risk factors possibly associated with SCD are strictly controlled. In fact, only one patient had high pre-dialysis kalaemia (6.1 mmol/L) and none was treated with potassium bath lower than 2 mmol/L.

Interestingly, among the 12 SCD, 9 occurred during the long interdialytic interval, while the remaining 3 in the first short interval. A previous large study27 on more than 30 000 HD patients showed that total mortality is higher during the long interdialytic interval. In other two studies sudden death was significantly more frequent during the first 24 h of the first short interdialytic interval and during the last 24 h of the long interval, i.e. immediately before and immediately after the first weekly HD session.2,28

This study has some limitations. We were not able to assess whether the predictive value of QT interval prolongation is present also when measured from a single standard ECG, because a standard ECG at the same time of the ECG Holter was available only for some patients. Another limitation is represented by the inability to provide data on QRS duration from ECG Holter recordings. Although patients with bundle branch block have been excluded from the study, the influence of an even slight prolongation of QRS duration on the QT interval cannot be ruled out.

To test the hypothesis that QTc is an independent predictor of total mortality and SCD in HD population, further studies are necessary in larger samples. The findings of the present study, however, might have significant implications for the clinical management of ESRD patients. A prolonged QTc interval might be used for risk stratification of uraemic patients, particularly in those with alterations in left ventricular dimension and/or ventricular mass. Subjects undergoing chronic HD with prolonged ventricular repolarization duration should be carefully monitored. In these patients, low concentrations of calcium and potassium in the dialysis bath should be avoided and, since sudden death is particularly frequent during the long interdialytic interval,2,6 the schedule of HD sessions should be changed accordingly. Despite the higher risk of SCD, patients with ESRD have been usually excluded by trials which assessed the efficacy of ICDs.29 Therefore, it is not known whether this intervention could be effective in this population, and clinical trials to assess the efficacy of interventions to reduce mortality in ESRD patients undergoing HD are needed.

Conflict of interest: none declared.

References

1
U.S. Renal Data System.
USRDS 2008 annual data report: Atlas of chronic kidney disease and end-stage renal disease in the United States, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases
2008
Bethesda, MD
2
Genovesi
S
Valsecchi
MG
Rossi
E
Pogliani
D
Acquistapace
I
De Cristofaro
V
, et al.  . 
Sudden death and associated factors in a historical cohort of chronic haemodialysis patients
Nephrol Dial Transplant
 , 
2009
, vol. 
24
 (pg. 
2529
-
36
)
3
Priori
SG
Aliot
E
Blomstrom-Lundqvist
C
Bossaert
L
Breithardt
G
Brugada
P
, et al.  . 
Task force on sudden cardiac death of the European Society of Cardiology
Eur Heart J
 , 
2001
, vol. 
22
 (pg. 
1374
-
450
)
4
Gruppo Emodialisi E Patologie Vascolari
Multicentre, cross-sectional study of ventricular arrhythmias in chronically haemodialysed patients
Lancet
 , 
1988
, vol. 
332
 (pg. 
305
-
9
)
5
Genovesi
S
Dossi
C
Viganò
MR
Galbiati
E
Prolo
F
Stella
A
, et al.  . 
Electrolyte concentration during haemodialysis and QT interval prolongation in uraemic patients
Europace
 , 
2008
, vol. 
10
 (pg. 
771
-
7
)
6
Bleyer
AJ
Hartman
J
Brannon
PC
Reeves-Daniel
A
Satko
SG
Russel
G
Characteristics of sudden death in haemodialysis patients
Kidney Int
 , 
2006
, vol. 
69
 (pg. 
2268
-
73
)
7
Montanez
A
Ruskin
JN
Hebert
PR
Lamas
GA
Hennekens
CH
Prolonged QTc interval and risks of total and cardiovascular mortality and sudden death in the general population: a review and qualitative overview of the prospective cohort studies
Arch Intern Med
 , 
2004
, vol. 
164
 (pg. 
943
-
8
)
8
Zhang
Y
Post
WS
Blasco-Colmenares
E
Dalal
D
Tomaselli
GF
Guallar
E
Electrocardiographic QT interval and mortality: a meta-analysis
Epidemiology
 , 
2011
, vol. 
5
 (pg. 
660
-
70
)
9
Näppi
SE
Virtanen
VK
Saha
HHT
Mustonen
J
Pasternack
AI
QTc dispersion increases during haemodialysis with low-calcium dialysate
Kidney Int
 , 
2000
, vol. 
57
 (pg. 
2117
-
22
)
10
Genovesi
S
Rivera
R
Fabbrini
P
Dossi
C
Bonforte
G
Mircoli
L
Dynamic QT interval analysis in uraemic patients receiving chronic haemodialysis
J Hypertens
 , 
2003
, vol. 
21
 (pg. 
1921
-
6
)
11
Severi
S
Grandi
E
Pes
C
Badiali
F
Grandi
F
Santoro
A
Calcium and potassium changes during hemodialysis alter ventricular repolarization duration: in vivo and in silico analysis
Nephrol Dial Transplant
 , 
2008
, vol. 
23
 (pg. 
1378
-
86
)
12
Beaubien
ER
Pylypchuk
GB
Akhtar
J
Jay Biem
H
Value of corrected QT interval dispersion in identifying patients initiating dialysis at increased risk of total and cardiovascular mortality
Am J Kidney Dis
 , 
2002
, vol. 
39
 (pg. 
834
-
42
)
13
Hage
FG
de Mattos
AM
Khamash
H
Mehta
S
Warnock
D
Iskandrian
AE
QT prolongation is an independent predictor of mortality in end-stage renal disease
Clin Cardiol
 , 
2010
, vol. 
44
 (pg. 
361
-
6
)
14
Zipes
DP
Camm
AJ
Borggrefe
M
Buxton
AE
Chaitman
B
Fromer
M
, et al.  . 
ACC/AHA/ESC 2006 guidelines for management of patients with ventricular arrhythmias and the prevention of sudden cardiac death
Europace
 , 
2006
, vol. 
8
 (pg. 
746
-
837
)
15
Kong
MH
Fonarow
GC
Peterson
ED
Curtis
AB
Hernandez
AF
Sanders
GD
, et al.  . 
Systematic review of the incidence of sudden cardiac death in the United States
J Am Coll Cardiol
 , 
2011
, vol. 
57
 (pg. 
794
-
801
)
16
Lang
RM
Bierig
M
Devereux
RB
Flachskampf
FA
Foster
E
Pellikka
PA
, et al.  . 
Chamber Quantification Writing Group; American Society of Echocardiography's Guidelines and Standards Committee; European Association of Echocardiography
J Am Soc Echocardiogr
 , 
2005
, vol. 
18
 (pg. 
1440
-
63
)
17
Rautaharju
PM
Surawicz
B
Gettes
LS
AHA/ACCF/HRS recommendations for the standardisation and interpretation of the electrocardiogram
J Am Coll Cardiol
 , 
2009
, vol. 
53
 (pg. 
982
-
91
)
18
Morris
ST
Galiatso
E
Stewart
GA
Stuart
R
Rodger
C
Jardine
AG
QT dispersion before and after haemodialysis
J Am Soc Nephrol
 , 
1999
, vol. 
1
 (pg. 
160
-
3
)
19
Lörincz
I
Màtyus
A
Zilahi
Z
Kun
C
Karànyi
Z
QT dispersion in patients with end-stage renal failure and during haemodialysis
J Am Soc Nephrol
 , 
1999
, vol. 
10
 (pg. 
1297
-
302
)
20
Covic
A
Diaconita
M
Gusbeth-Tatomir
P
Covic
M
Bozetan
A
Ungureanu
G
, et al.  . 
Haemodialysis increases QTc interval but not QTc dispersion in ERSD patients without manifest cardiac disease
Nephrol Dial Transplant
 , 
2002
, vol. 
17
 (pg. 
2170
-
7
)
21
Stramba-Badiale
M
Priori
SG
Gender-specific prescription for cardiovascular diseases?
Eur Heart J
 , 
2005
, vol. 
26
 (pg. 
1571
-
2
)
22
Haider
AW
Larson
MG
Benjamin
EJ
Levy
D
Increased left ventricular mass and hypertrophy are associated with increased risk for sudden death
J Am Coll Cardiol
 , 
1998
, vol. 
32
 (pg. 
1454
-
9
)
23
Wachtell
K
Okin
PM
Olsen
MH
Dahlöf
B
Devereux
RB
Ibsen
H
, et al.  . 
Regression of electrocardiographic left ventricular hypertrophy during antihypertensive therapy and reduction in sudden cardiac death: the LIFE study
Circulation
 , 
2007
, vol. 
116
 (pg. 
700
-
5
)
24
O'Mahony
C
Elliott
P
McKenna
W
Sudden cardiac death in hypertrophic cardiomyopathy
Circ Arrhythm Electrophysiol
 , 
2012
 
[e-pub ahead of print] doi:10.1161/CIRCEP.111.962043
25
Paoletti
E
Specchia
C
Di Maio
G
Bellino
D
Damasio
B
Cassottana
P
, et al.  . 
The worsening of left ventricular hypertrophy is the strongest predictor of sudden cardiac death in haemodialysis patients: a 10 year survey
Nephrol Dial Transplant
 , 
2004
, vol. 
19
 (pg. 
1829
-
34
)
26
Pun
PH
Lehrich
RW
Honeycutt
EF
Herzog
CA
Middleton
JP
Modifiable risk factors associated with sudden cardiac arrest within hemodialysis clinics
Kidney Int
 , 
2011
, vol. 
79
 (pg. 
218
-
27
)
27
Foley
RN
Gilbertson
DT
Murray
T
Collins
AJ
Long interdialytic interval and mortality among patients receiving hemodialysis
N Engl J Med
 , 
2011
, vol. 
365
 (pg. 
1099
-
107
)
28
Bleyer
AJ
Hartman
J
Brannon
PC
Reeves-Daniel
A
Satko
SG
Russell
G
Characteristics of sudden death in hemodialysis patients
Kidney Int
 , 
2006
, vol. 
69
 (pg. 
2268
-
73
)
29
Cannizzato
LA
Piccini
JP
Patel
UD
Hernandez
AF
Device therapy in heart failure patients with chronic kidney disease
J Am Coll Cardiol
 , 
2011
, vol. 
58
 (pg. 
889
-
96
)