Influence of Ischemia on Heart-Rate Variability in Chronic Hemodialysis Patients

Abstract

Background: Sudden cardiac death occurring in patients with end-stage renal disease (ESRD) may be related to poor autonomic function with a significant decreased heart-rate variability (HRV). In addition, coronary artery disease has a high prevalence in this population and accounts for 50% of deaths. In the present study, relationships between HRV and myocardial ischemic abnormalities revealed by myocardial scintigraphy (MS) were evaluated in 32 chronic hemodialysis patients. Methods: We prospectively studied 32 chronic hemodialysis patients. Each underwent MS and 24 h electrocardiography at baseline for analysis of time and frequency domain the day of dialysis. Three periods were analyzed: during dialysis session, the morning after (nondialytic period), and in a 24 h period. Patients were included in group 1 (seven women, 11 men; mean age: 62 ± 19 years) when MS revealed no ischemia, whereas patients were included in group 2 (seven women, seven men; mean age: 63.1 ± 20 years) when MS revealed ischemic lesions. Results: A student ± test revealed that during the nondialytic period, two important markers of HRV, percentage of delta RR > 50 ms (pNN50) (4.5 ± 4.04 in group 1 versus 1.7 ± 1.4 in group 2), and root mean square of delta RR (rMSSD) (27.7 ± 13.4 versus 19.7 ± 6.8) were significantly reduced in group 2 compared with values in group 1. No significant difference appears between the two groups for standard deviation of normal to normal intervals (SDNN), mean heart rate, and spectral analysis. Conclusion: Patients with ESRD and myocardial ischemia revealed by MS have reduced parasympathetic activity during the nondialytic period. Correlations between parameters of HRV and ischemic lesions revealed by MS have been shown for the first time.

• Ambulatory 24 h electrocardiogram
• Autonomic nervous system
• Hemodialysis
• Heart rate variability
• Myocardial scintigraphy

Introduction

Despite dialysis and renal transplantation, the mortality of patients with end-stage renal disease (ESRD) remains 3.5-fold higher than in the general population.[1&2] Cardiac disease is highly prevalent in this population and accounts for 48% of overall deaths.[2] Previous studies suggest that sudden cardiac death in patients awaiting kidney transplantation may be related to poor autonomic function.[3]

Recent studies have reported that heart rate variability (HRV) measures are reduced in chronic hemodialysis patients compared to healthy subjects.[8&9] Furthermore, some measures are independent predictors of cardiac death.[4&5]

Moreover, in this population, myocardial ischemia evaluated by thallium scintigraphy provided prognostic information identifying patients at risk for cardiac deaths.[6]

The aim of the present study was to determine the relationship between coronary ischemia [evaluated by myocardial scintigraphy (MS)] and HRV in patients with ESRD.

Patients and Methods

Population Studied

A total of 32 hemodialysis patients (HD) were included in this study between December 2001 and March 2002. They underwent conventional 4 h dialysis three times a week, either at the hemodialysis center, or in an autodialysis center, or at home. Patient characteristics are described in Table 1. Demographic characteristics were identical to those of other patients dialyzed. Patients were not included in this study if cardiac arrhythmia was present. The clinical research committee approved this study, and informed consent was obtained from all patients.

Table 1 Characteristics of the Population Study.

	Age	Sex	Nephropathy	Dialysis duration	Hypercholeste-rolemia	Cardiac events	Renal transplantation waiting
Group 1	62 ± 19	7 women 11 men	5 nephrosclerosis	3.8 ± 1.8	4/18	3.2 ± 1.8	4/18
			3 polycystic renal disease
			5 diabetic glomerulopathy
			1 myeloma
			2 glomerulonephritis
			2 other causes
Group 2	63.1 ± 20	7 women 7 men	4 diabetic glomerulopathy	4.3 ± 1.6	3/14	3 ± 1.8	3/14
			4 nephrosclerosis
			1 polycystic renal disease
			1 goodpasture disease
			2 glomerulonephritis
			2 other causes

Study Protocol

Analysis of HRV

A 24 h Holter recording was performed on the day of dialysis, and patients were required to write in a notebook details, particularly times, of their activities and medication taken. The electrocardiogram readings were recorded on a “Synesis” numeric Holter monitor (Ela Médical). The data were stored on a 20 Mo flash card; the analysis was performed using “Elatec” software (Ela Médical) at electrophysiologic lab at Dijon hospital. Beta blockers or ACE were stopped 1 week before the ECG recording.

The analysis of HRV was conducted in two steps: using a spectral approach, and according to frequency. HRV parameters (Table 2) were as follows: SDNN, pNN50, and rMSSD, and for the spectral analysis predictors: ptot, LnHF, LnLF, and the LF/HF ratio, which reflects the balance between sympathetic and vagal activity.

Table 2 Analyzed Heart Rate Variability (HRV) Measures and Definitions.

Variable	Definition
Time domain statistical measures
SDNN (ms)	Standard deviation of all RR intervals during 24 h
rMSSD (ms)	Root mean square successive difference of RR intervals during 24 h
pNN50 (%)	Percentage of successive pairs of RR intervals differing by 50 ms
Frequency domain geometric measures
Total power (ms^2)	Energy in power spectrum below 0.40 Hz
LF power (ms^2)	Energy in power spectrum between 0.04 and 0.15 Hz
HF power (ms^2)	Energy in power spectrum between 0.15 Hz and 0.40 Hz

The ECG recording was started before the hemodialysis session (predialytic period), continued throughout the dialysis (dialytic period), until the morning after (postdialytic period). The postdialytic period, which was analyzed, was late postdialytic period (4 h later dialysis period).

The Power Spectral Analysis of heart-rate variability (PSA) was estimated of a sequence of 500 RR intervals. Ectopics beat or artefacts that may affect the estimation of the power spectral densities of the heart rate variability, were automatically deleted from the data. The spectrum of the PSA was separated into a low-frequency (LF) band in the range of 0.04–0.15 Hz, and a high-frequency (HF) band in the range of 0.15–0.40 Hz. LF and HF components are regarded as specific markers of, respectively, sympathetic and parasympathetic activities.[7]

Myocardial Scintigraphy

Each of these patients had undergone MS in the course of the 3 months. All patients were asked to discontinue anti-ischemic drugs (beta-adrenergic blocking,agent calcium channels antagonists) at least 48 h prior to the stress test.

Induction of Ischemia

Three methods were used, and imaging was performed either after exercise testing, dipyridamole infusion alone, or intravenous dipyridamole infusion combined with a low-level standardized bicycle exercise.

Exercise

The exercise test was done on an ergometer bicycle in the upright position. Dipyridamole was administered intravenously at 0.56 mg/kg body weight over 4 min. An ECG was recorded before and after the injection.

Combined Dipyridamole/Exercise

When exercise was expected to be submaximal, a low-level bicycle exercise was performed (30 W during 4 min for women, and 40 W during 4 min for men) after dipyridamole infusion.

Acquisition and Analysis

A standard procedure was used to perform thallium-201 or Tc-99m-Sestamibi single photon emission computed tomoscintigraphy (SPECT) imaging after stress, as well as after reinjection. At peak exercise, or 3 min following dipyridamole infusion, 74 MBq of thallium-201 or 259 MBq of Tc-99m-Sestamibi were injected intravenously.

The initial images (stress) were obtained 10 min after injection of the tracer using a rotating gamma camera quipped with a low-energy, high-resolution collimator.

Rest images were obtained 4 h after stress, after reinjection of 37 MBq of thallium-201 or 777 MBq of Tc-99m-Sestamibi, using the same imaging technique. During the period between initial imaging and reinjection, the patient remained at rest.

Initial imaging and reinjection data were processed to obtain transaxial sections by filtered back-projection (ramp filter combined with a Hamming window) without attenuation correction. These transaxial sections were reoriented to the three standard cardiac planes (short, horizontal long, and vertical long axes) to allow visual interpretation. The left ventricle was then divided into 13 segments. Visual analysis was performed, and an agreement was obtained from two experienced observers unaware of clinical and follow-up data.

Patients were included in group 1 if the scintigraphy revealed no ischemia either on the thallium test or on the injection of dipyridamole (normal imaging, persistent defects, that is, patients with no reinjection in any initially abnormal segments, or initial defects with reversibility limited at one segment), whereas patients were included in group 2 if the scintigraphy revealed ischemic lesions (patients who presented at least two adjacent abnormal myocardial segments with partial or total normalization at reinjection).

Statistical Analysis

All descriptive values were expressed as mean value ± SD.

Comparisons of continuous variables between the two groups were performed by student's unpaired t-test. A p value of less than 0.05 was considered to indicate significance.

Results

A total of 32 hemodialysis patients were included in this study: 18 patients in group 1 (seven women, 11 men; mean age: 62 ± 19 years) and 14 patients in group 2 (seven women, seven men, mean age: 63.1 ± 20 years).

Analysis of HRV

Dialysis Period

Table 3 summarizes values during the dialysis period. No significant difference was observed between the two groups for all HRV parameters. In particular, no difference was observed for LnLF (4.67 ± 1.22 versus 4.39 ± 1.39).

Table 3 Comparison HRV Measures Between the Two Groups During the Dialysis Period.

	Group 1 (MS −)	Group 2 (MS +)
SD (ms)	58 ± 29	61 ± 29
pNN50 (%)	3.8 ± 5.2	2.6 ± 4.7
rMSSD (ms)	24.5 ± 18	22.2 ± 15
Fc (puls/min)	71 ± 10	75 ± 9
LnLF	4.67 ± 1.22	4.39 ± 1.39
LnHF	3.88 ± 1.30	3.64 ± 1.07
LF/HF	3.40 ± 2.66	2.70 ± 2.25
Ptot (ms^2)	1003 ± 773	760 ± 528

Data represent mean ± SD.

p < 0.05.

Later Nondialysis Period

HRV data collected during the nondialysis period are summarized in Table 4. Time domain parameters (pNN50, rMSSD) were significantly lower in group 2 compared to group 1. Respectively, p = 0.01 for pNN50, and p = 0.03 for rMSSD.

Table 4 Comparison HRV Measures Between the Two Groups During the Later NonDialysis Period.

	Group 1 (MS −)	Group 2 (MS +)
SD (ms)	72 ± 28	57 ± 23
pNN50 (%)	6.7 ± 2.7	1.8 ± 1.4a
rMSSD (ms)	30.7 ± 21.5	18.8 ± 7.1a
Fc (puls/min)	77 ± 12	77 ± 10
LnLF	4.73 ± 0.90	4.04 ± 1.41
LnHF	4.21 ± 1.15	3.55 ± 0.89
LF/HF	3.09 ± 2.73	3.22 ± 3.77
Ptot (ms^2)	984 ± 656	810 ± 786

Data represent mean ± SD.

^ap < 0.05.

No significant difference between the two groups for lnLF, lnHF, and LF/HF values was demonstrated.

24 h Period

HRV data collected during the nondialysis period are shown in Table 5.

Table 5 Comparison of HRV Measures Between the Two Groups During the 24 Hours Period.

	Group 1 (MS −)	Group 2 (MS +)
SD (ms)	89 ± 32	81.4 ± 30.5
pNN50 (%)	4.5 ± 2.7	1.7 ± 1.4a
rMSSD (ms)	27.7 ± 13.4	19.7 ± 6.8a
Fc (puls/min)	74 ± 11	77 ± 10
lnLF	4.89 ± 1.20	4.29 ± 1.39
lnHF	4.17 ± 0.80	3.64 ± 0.82
LF/HF	3.37 ± 2.61	2.85 ± 2.61
Ptot (ms^2)	976 ± 592	862 ± 716

Data represent mean ± SD.

^ap < 0.05.

Overall variability was also linked to low activity of the sympathetic (lnHF) and parasympathetic (lnLF, pNN50, rMSSD) systems.

There is a statistically significant difference between the two groups in pNN50, rMSSD; these markers reflect, in part, the parasympathetic component of the autonomic nervous system.

For frequency domain variables, differences between the two groups were not significant, but variables were lower in group 2: lnHF: 4.17 ± 0.8 in group 1 versus 3.64 ± 0.82 in group 2; lnLF: 4.89 ± 1.2 in group 1 versus 4.29 ± 1.39 in group 2; LF/HF: 3.37 ± 2.61 in group 1 versus 2.85 ± 2.61 in group 2.

Discussion

Major Findings

In this study, we examined the relationship between coronary ischemia (evaluated by MS) and HRV in a specific population of end-stage renal disease.

HRV of Hemodialysis Patients

Many studies have reported that time domain measures and frequency domain measures are significantly reduced in HD patients compared with age-matched healthy subjects.[8&9]

According to Steinberg et al.,[10] ESRD is an independent factor for decreased HRV, and the predictor variables of cardiac diseases, diabetes mellitus, age,gender, and smoking, explained only a small portion of the HRV. In our series, we found no significant differences with or without cardiac risk factors. Although decreased HRV and myocardial ischemia revealed by MS are both prognostic predictors; to our knowledge, this is the first study to show the relationship between these data in HD patients (Tables 4 and 5).

A recent study has demonstrated the incremental prognostic value of SPECT thallium imaging over the five clinical risk factors in this population.[6] We also observed that HRV measures are different in each phase (dialysis period, later nondialysis period, and 24 h period). Between the two groups, no HRV measures are significantly different during the dialysis period, whereas rMSSD and pNN50 are significantly different in later nondialysis and 24 h periods.

The hemodynamic fluctuations in the dialysis period could be explained by many sources, and our population was also a mixture of diabetic and nondiabetic patients. Diabetic patients have a severely impaired autonomic nervous system compared to nondiabetic patients during nondialysis period, probably due to autonomic neuropathy related to the effects of both diabetic and uremic conditions.[11]

We obtained the same statistical differences between the two groups in the analysis of each phase after excluding diabetic patients of our series. Also, HRV measures in a 24 h period are an efficient evaluation of HRV.

Clinical Implications

In this study, in contrast with other recently published studies on HRV, we have shown that, for certain markers, there is a significant correlation with coronary disease detected by MS.

Thus, the study of HRV may be useful in orienting diagnosis in chronic renal failure and should be included in the assessment and monitoring of heart condition. It could also be useful as a marker after transplantation. However, these data need to be confirmed by further studies with larger cohorts.

Limitations

The difference between the two groups in the values of LF and HF is not significant, but, most values seem to be low. Recent studies have found significantly decreased values for LF in chronic renal failure, even though HF values and the LF/HF ratio were not significantly decreased, which suggests that the two nervous systems, but mainly the parasympathetic nervous system, were affected.

Moreover, there is no difference for diabetic patients in our study, whereas they are included in group 1 or 2. Others studies have demonstrated a relationship between diabetes and poor autonomic function, perhaps due to smaller numbers in our study compared to previous studies. Furthermore, it is unclear whether HRV is a stable measure in HD patients because HRV may be affected by the hemodialysis session. Earlier studies have reported that HRV analyzed from short-term ECG recordings improves immediately after a single hemodialysis session, due to an improvement of hypervolemia[12] and a removal of uremic toxins.[13] We cannot exclude that electrolytes in dialysate solution could influence HRV during the dialysis period and produce false results.

Conclusion

In this study, we confirmed a lower HRV in hemodialysis patients and we showed, for the first time, a relationship between parameters of HRV and ischemic lesions revealed by MS. Further studies must be conducted to confirm these results over a larger population, because therapeutic measures aimed at improving the HRV profile may be expected to favorably influence the high mortality rate.