Depression, heart rate variability, and exercise training in dialysis patients

Abstract

Background Functional limitations, altered cardiac autonomic activity, and psychological distress are known disorders in chronic hemodialysis (HD) patients, relating to increased morbidity and mortality. The aim of this study was to examine the influence of an exercise training program on emotional parameters and heart rate variability (HRV) indices, as well as to determine whether emotional stress contributes to autonomic dysfunction in these patients.

Methods Forty-four HD patients were randomly assigned into group A (24 patients, aged 46.3 ± 11.2 years), who participated in a 1-year intradialytic exercise training program and group B (20 patients, aged 45.8 ± 10.8 years), who were used as controls. At baseline and a year after, measures of HRV were obtained for the estimation of standard deviation of RR intervals, the mean square successive differences, percentage of RR intervals differing by more than 50 ms from the preceding RR interval (pNN50), and low to high frequency components. Emotional parameters (Beck Depression Inventory – BDI and Hospital Anxiety and Depression Scale – HADS) were also assessed by validated questionnaires. Moreover, all patients performed a spiroergometric study for the estimation of VO_2peak.

Results At baseline, all measurements were similar in the two groups and remained almost unchanged after a year in group B. After a year of training, VO_2peak increased from 16.79 ± 5.24 to 22.33 ± 4.90 ml/kg per min (P >0.001) in group A. Trained patients also showed an increase in standard deviation of RR intervals by 58.8% (P >0.001), the mean square successive differences by 68.1% (P >0.001), pNN50 by 23.5% (P >0.001), and a low to high frequency ratio by 17.3% (P >0.001). Finally, at the end of the study, group A showed a decrease in BDI score by 34.5% (P >0.001) and HADS by 23.9% (P >0.001). Canonical correlation revealed significant inverse correlation among depression (in BDI and HADS) and HRV indices before and after exercise training.

Conclusion Cardiac autonomic modulation seemed to be sensitive to the experience of persistent depression in HD patients. Significantly, exercise training reduced emotional distress and concomitantly improved HRV.

depression, dialysis, exercise training, heart rate variability

Introduction

Patients on hemodialysis (HD) suffer from complex health disorders, which lead to impaired physical function, psychological status and health-related quality of life, and poor prognosis [1,2]. Depression is common in these patients, with a prevalence ranging from 13 to 60% [1–3]. Carney et al. [4] reported that the levels of depression and anxiety in HD patients are comparable with those of psychiatric patients. Furthermore, it was reported that depression may be a treatable cause of mortality in HD patients [5]. Several factors have been suggested as potential mechanisms responsible for the deleterious effects of depression on prognosis and secondary outcomes in patients with chronic diseases. Amongst these, increased sympathetic tone, elevated catecholamine levels, increased platelet activation, inflammatory processes, and nonadherence to prevention and treatment regimens are the most serious factors [6,7]. Importantly, altered cardiac autonomic tone remains one of the most plausible explanations for the potential role of depression on cardiac mortality [8,9]. It is well known that increased sympathetic or decreased parasympathetic nervous system activity, as assessed by heart rate variability (HRV), is associated with malignant arrhythmias and sudden cardiac death, especially in coronary heart disease patients [10–12]. Recent studies showed a strong association between depression and low HRV in postinfarction patients [10,13,14]. Moreover, elderly patients suffering from a major depressive disorder were described to have a reduced HRV [15,16]. It is supported that HD patients have low HRV indices and increased incidence of cardiac arrhythmias [17–19].

Studies in the field of renal rehabilitation have reported that physical training increased HRV indices, as well as baroreflex sensitivity in HD patients [18,20,21]. Moreover, it was supported that exercise training (ET) leads to a significant improvement in their psychological status [1,2,22]. It is, however, unclear whether emotional stress contributes to excessive sympathetic or inadequate parasympathetic tone and whether this relationship is affected by the level of their aerobic capacity.

Hence, the objective of this study was to investigate the effects of ET on emotional parameters and HRV in HD patients and also to determine whether there is a relationship between depression and HRV indices. Another goal was to examine whether the possible dependence between depression and HRV are still apparent after a year of intradialytic ET.

Methods

Patients and design

All 97 end-stage renal disease patients who underwent HD three times a week for 4 h, for at least 6 months, in the Renal Unit of the AHEPA University Hospital in Thessaloniki, Greece, were informed about the research objectives, methods, and procedures, and 64 of them volunteered to participate in this study. Patients included in the protocol fulfilled the following criteria: (i) no history, clinical signs, or symptoms of psychiatric, neurological, cardiologic, or pulmonary disorders; (ii) absence of diabetes mellitus; (iii) no significant electrolytic instability or undisciplined patients; (iv) no musculoskeletal limitation or other medical problems contraindicating participation in an ET program. Fifty of them met these criteria and were assigned to either an ET (group A) or to a sedentary control group (group B) through complete randomization. All patients had a baseline symptom-limited cardiopulmonary exercise test. In 24 patients of group A (one dropped out of the ET program), a final similar exercise test was carried out after the completion of the last training session. Five patients of group B were lost to follow-up. The remaining 20 patients had the same cardiopulmonary exercise test at the end of the study. Moreover, data on autonomic function and psychological parameters were obtained at baseline and the end of the study in both groups. Throughout the study, the type and dose of medications remained the same for all patients. None of the patients was on antidepressants or other psychotropic agents. The University Ethics Committee approved the study protocol and all patients provided written informed consent.

Measures

Cardiopulmonary exercise testing

At the beginning and at the end of the study, all patients underwent a symptom-limited graded exercise treadmill test under continuous electrocardiographic recording by a Cambridge Heart 12 ECG System Co (CH-2000; Cambridge Heart Inc, Bedford, Massachusetts, USA), on a nondialysis day, according to Bruce protocol. Expired air was collected in a facemask and analyzed on a breath-by-breath basis for quantification of oxygen consumption (VO₂), expired ventilation, and expiratory exchange ratio (VCO₂/VO₂) with the use of the MedGraphics Breeze Suite CPX Ultima ergospirometer device (Medical Graphics Corporation, Minnesota, USA). Peak values were determined from an average measured during the last 60 s when respiratory exchange ratio was greater than 1.0–1.10. VO_2peak was defined as the highest VO₂ obtained, characterized by a plateau of oxygen uptake (steady state), despite further increases in work rate.

Heart rate variability measurement

Measures of HRV were obtained by both time-domain and frequency-domain analyses of 24-h continuous ambulatory ECG recordings, through a three-channel Digital Ambulatory ECG Holter Recorder (GBI-3S; Galix Biomedical Instrumentation Inc., Miami Beach, Florida, USA) with electronic RAM memory, at the beginning and at the end of the study on a non-HD day. The WinTer Holter Analyzer software (GBI-3S), which has been tested and recognized by the FDA, was used for the analysis. Moreover, postprocessing was used to remove ectopic discharges from RR analysis, as only normal-to-normal RR intervals were used in the analysis of HRV indices. The participants were asked to avoid caffeine and alcoholic beverages, as well as any activity that could affect the heart rhythm, except for their daily activities, during the recording, to keep heart rate modulations as stable as possible, for a more accurate interpretation of the results of the frequency-domain analysis [23].

Briefly, measurements in the time domain included the mean RR interval for the 24 h period, the standard deviation of all RR intervals (SDNN), the mean square successive differences (MSSD), and the percentage of RR intervals differing by more than 50 ms from the preceding RR (pNN50). From the frequency-domain analysis performed on the entire 24 h period, the following indices were evaluated: low frequency (LF, frequencies with range 0.04–0.15 Hz) and high frequency (HF, frequencies with range 0.15–0.4 Hz) components, as well as the LF/HF ratio. The LF component is considered as a parameter that includes both sympathetic and vagal influences, whereas the HF component is considered as a marker of vagal activity [23]. In summary, the ratio LF/HF is a marker of sympathovagal balance [23].

Depression measurement

The following psychological tests were administered to all patients at the beginning and at the end of the study by the same physician, who was not familiar with the patients.

1. The Beck Depression Inventory (BDI), as translated and standardized for the Greek population, which is a self-rating questionnaire for the assessment of the severity of depression [24]. It is composed of 21 items, each with a 0–3 grading system. Each item consists of four self-evaluative statements of increasing severity. A total score of 0–9 indicates no depression, 10–15 mild to moderate depression, 16–23 moderate to severe depression, and a score ≤ 24 indicates severe depression.

2. The Hospital Anxiety and Depression Scale (HADS), as translated in Greek, which is a self-administered questionnaire for assessing depression and anxiety of general hospital patients [25]. It is composed of seven items that are answered on a four-point Likert scale of 0–3 grading system (score range: 0–21). A cut-off ≤ 8 on the anxiety subscale identifies individuals with anxiety and depressive symptoms. HADS is a valid and internally consistent tool across 15 studies with Cronbach's a ranging from 0.68 to 0.93 for anxiety and 0.67–0.90 for depression measurements [26].

Exercise training program

Patients of group A attended 1-year ET sessions, which were conducted three times a week during the first 2 h of the HD treatment. Patients were required to attend at least 65% of all sessions over the 1-year period. All training sessions were supervised by a physician and two exercise trainers specialized in physical rehabilitation. The duration of each session was in between 60 and 90 min (according to each patient's ability) and the exercise intensity was prescribed on an individual basis and was monitored by rating the level of exertion using the Borg Rating of Perceived Exertion (RPE) scale. Patients’ heart rates during training were continuously monitored and their blood pressure was also measured every 15 min.

Each session consisted of a 5-min warm-up, a 30–60 min active cycling session with specific devices, which were adjusted to each patient's bed, a 20-min strengthening program, and a 5-min cool-down period. The devices used for cycling were supported by motor power and gave the possibility for passive training for the warm-up and cool-down period.

Regarding the aerobic part of the ET program, patients started with 30 min cycling at a desired workload with an exercise intensity of light to some-what hard (11–13 on the 6–20 RPE scale). The exercise prescriptions were adjusted periodically according to patients’ ability at a target RPE of 13 (some-what hard), which is approximately equivalent to working at 70% of VO_2peak. To increase workload, the duration of cycling was also gradually increased over time according to each patient's ability and finally reached an hour of active cycling.

At the beginning, the strengthening program consisted of two sets of exercises for lower limbs. The workload was then gradually increased to higher workloads by first increasing the number of repetitions per set (from eight to 12) and then the number of sets. Afterwards, exercises with therabands and free weights, applying round the limbs, were added.

Statistical analysis

Data are expressed as mean ± SD. Changes in variables within and between the groups at baseline and at the end of the study were evaluated by using two-way within-subjects analysis of variance (linear mixed model) with time and group being the independent variables. The χ² test was also used to compare the distributions of the BDI score among the two groups. The marginal homogeneity test was conducted to compare BDI score between baseline and follow-up measurements in each group. The study of the pattern of relationships among depression (in BDI and HADS) and HRV indices between baseline and follow-up measurements in group A was conducted using Pearson's correlation coefficient and canonical correlation analysis (CCA). CCA is a type of multivariate linear statistical analysis, first described by Hotelling [27], which is used in a wide range of disciplines to analyze the relationships between multiple independent and multiple dependent variables. CCA computes two derived (or ‘canonical variates') variables, X and Y, such that the correlation between X and Y is as large as possible. X is a weighted average of the independent variables (baseline), and Y is a weighted average of the dependent variables (follow-up). Following Cliff [28], the interpretations of the structure correlations between the covariates and the variates X, Y were used rather than the estimated weights. The proportion of the covariates that was explained by the canonical variates X and Y were also presented. The sampling theory for CCA is complex and assumes multivariate normality and large samples, therefore the use of CCA in this study was exploratory in nature [29]. Statistical analysis was performed using the Statistical Package for Social Sciences (SPSS Inc., Chicago, Illinois, USA), version 15.0 software for Windows. Statistical significance was accepted as P value less than 0.05.

Results

Demographic, social, and clinical characteristics

Information on demographic and clinical variables, including sex and age, dialysis prescription, hypertension, history of associated diseases, education, employment, and family status of the 44 HD patients, who completed the study, are presented in Table 1.

No significant differences were found for sex, age, clinical variables, hematological, and biochemical measurements between the two groups at baseline and a year after (Tables 1 and 2). Throughout the study, the level of Hb, plasma urea, creatinine, and electrolytes remained almost constant for both groups. There were no adverse effects associated with the training sessions. All patients of group A attended 82% of the exercise session (range 68–92%) and these 24 patients were reevaluated.

Table 1

Demographic, social, and clinical characteristics of the trained (group A) and untrained (group B) HD patients

	Group A (n =24)	Group B (n = 20)
Age (years)	46.3 ± 11.2	45.8 ± 10.9
Sex (male/female)	14/10	12/8
Years on HD	6.1 ± 4.6	6.3 ± 4.9
Frequency of HD (times/week)	3	3
Duration of HD (h)	4	4
Education (years)	10.2 ± 3.4	10.3 ± 3.5
Employment status
Unemployed	4	2
Part-time	9	8
Full-time	11	10
Family status
Married	13	12
Single	7	6
Divorced	4	2
Systolic BP (mmHg)	142.3 ± 15.1	143.3 ± 15.4
Diastolic BP (mmHg)	85.1 ± 9.4	85.3 ± 9.5
Patients on erythropoietin	20	18

Values are mean ± SD, unless otherwise noted. BP, blood pressure; HD, hemodialysis.

Effects of exercise training on physical capacity

There was no statistical difference in the baseline values of VO_2peak and exercise time between the two groups (P = 0.508, Table 3). After a year of training, significant improvements occurred in cardiorespiratory fitness level of group A, as VO_2peak increased by 24.8% (P >0.001) and exercise time (ET) by 61.4% (P >0.001). There was no change in the physical capacity of the controls.

Effects of exercise training on heart rate variability indices

There was no statistical difference in the baseline values of SDNN, MSSD, pNN50, LF, HF, and LF/HF between the two groups (Table 3). After a year of training, the SDNN was increased by 58.8% (P >0.001), MSSD by 68.1% (P >0.001), and pNN50 by 23.5% (P >0.001). From the frequency-domain analysis, the LF/HF was increased by 17.3% (P >0.001). The HRV indices were not significantly changed in group B by the end of the study.

Effects of exercise training on psychological indices

At baseline, there was no statistical difference in the values of BDI score and distribution, as well as HADS score between the two groups (Tables 3 and 4). At the end of the 1-year training program, patients of group A showed a significant decrease in the total score of BDI depression by 34.5% (P >0.001, Table 3), whereas significant differences were observed on the distributions of BDI scores in groups A and B (P =0.001, Table 4). Marginal homogeneity test showed a statistically significant difference between baseline and follow-up only for the trained group (P >0.001 for group A and P = 0.083 for group B, Table 4). The difference in group A is because of changes in ‘mildly depressed', 33.3% (two of six) became ‘not depressed', changes in ‘moderately depressed', 50% (three of six) became ‘not depressed’ and 33.3% (two of six) became ‘mildly depressed', and changes in ‘severely depressed', 100% (12 of 12) became ‘moderately depressed'. Hence, the number of depressed trained patients statistically decreased according to their depressive morbidity. In addition, the HADS score was decreased by 23.5% (P >0.001, Table 3).

Table 2

Hematological and biochemical data of the trained (group A) and untrained (group B) HD patients at baseline and at follow-up

	Group A			Group B
	Baseline	Follow-up	P value	Baseline	Follow-up	P value
Hb (g/dl)	11.2 ± 1.1	11.3 ± 1.2	0.6	11.4 ± 1.2	11.2 ± 1.3	0.6
Urea (mg/d)	185.3 ± 29.1	188.2 ± 27.3	0.3	186.3 ± 28.7	187.5 ± 29.1	0.8
Creatinine (mg/dl)	12.5 ± 2.8	12.9 ± 3.1	0.5	12.8 ± 3.4	12.5 ± 3.9	0.9
K (mEq/l)	5.8 ± 0.8	5.6 ± 0.7	0.2	5.7 ± 0.9	5.8 ± 0.6	0.3
Na (mEq/l)	141.2 ± 3.6	140.9 ± 3.8	0.2	140.5 ± 3.8	141.4 ± 3.9	0.4
Ca (mEq/l)	8.9 ± 1.1	8.7 ± 1.0	0.9	8.8 ± 1.4	8.9 ± 1.3	0.6
P (mg/dl)	6.5 ± 1.6	6.6 ± 1.7	0.6	6.7 ± 1.8	6.5 ± 1.6	0.6

Values are mean ± SD. Ca, calcium; Hb, hemoglobulin; HD, hemodialysis; K, potassium; Na, sodium; P, phosphorus.

Table 3

VO_2peak, HRV, BDI, and HADS data of the trained (group A) and untrained (group B) HD patients at baseline and at follow-up

	Group A (n =24)	Group B (n =20)	Groups A versus B
	Baseline	Follow-up	F-B P value	Baseline	Follow-up	F-B P value	Baseline P value	Follow-up P value
VO2peak (ml/kg per min)	16.79 ± 5.24	22.33 ± 4.90	>0.001	15.66 ± 4.10	15.33 ± 3.79	0.508	0.434	[ 0.001
SDNN (ms)	103.54 ± 17.55	164.47 ± 27.84	>0.001	97.25 ± 25.53	95.90 ± 26.10	0.734	0.340	[ 0.001
MSSD (ms)	23.25 ± 6.85	39.08 ± 8.35	>0.001	22.10 ± 6.63	21.90 ± 6.52	0.888	0.577	[ 0.001
pNN50 (ms)	3.36 ± .76	4.15 ± .87	>0.001	3.16 ± .71	2.96 ± .88	0.001	0.370	[ 0.001
LF/HF	1.39 ± .35	1.63 ± .34	>0.001	1.54 ± .39	1.40 ± .35	0.005	0.203	[ 0.001
BDI	22.29 ± 6.71	14.61 ± 4.15	>0.001	22.30 ± 6.81	22.10 ± 6.24	0.730	0.997	[ 0.001
HADS	10.63 ± 2.60	8.08 ± 2.19	>0.001	10.40 ± 2.50	10.35 ± 2.13	0.840	0.773	[ 0.001

Values are mean ± SD. BDI, Beck Depression Inventory Questionnaire; F–B, follow-up–baseline; HADS, Hospital Anxiety and Depression Scale Questionnaire; HD, hemodialysis; HF, high-frequency power; HRV, heart rate variability; LF, low-frequency power; MSSD, mean square successive differences; pNN50, percentage of RR intervals differing by more than 50 ms from the preceding RR; SDNN, standard deviation of RR intervals.

Table 4

Data of the psychological profile of the trained (group A) and untrained (group B) HD patients at baseline and at follow-up and results of hypothesis tests

			BDI score at baseline
		Not depressed	Mildly depressed	Moderately depressed	Severely depressed
Group	A	N	0	6	6	12
			%	0.0	25.0	25.0	50.0
	B	N	0	6	4	10
		%	0.0	30.0	20.0	50.0
			BDI score at follow-up
Group	A	N	5	6	13	0
		%	20.8	25.0	54.2	0.0
	B	N	0	3	7	10
		%	0.0	15.0	35.0	50.0

Comparison between baseline and follow-up for group A: P >0.001 (marginal homogeneity test). Comparison between baseline and follow-up for group B: P =0.083 (marginal homogeneity test). Comparison between groups A and B at baseline: P =0.896 (χ² test). Comparison between groups A and B at follow-up: P >0.001 (χ² test). BDI, Beck Depression Inventory Questionnaire; HD, hemodialysis.

Relationships between baseline and follow-up measurements, among psychological and heart rate variability indices

Significant correlations exist between baseline and follow-up measurements mainly for group A (Tables 5 and 6). BDI and HADS scores were inversely correlated with all other variables merely for group A (Table 5). Moreover, for group A, CCA showed one significant pair of canonical variates X and Y. The correlation between X and Y (canonical correlation) was 0.998. Variate X accounted for the 77.5% of the variance of baseline measurements and variate Y accounted for the 52.4% of the variance of follow-up measurements (Table 7). Examination of the correlations between the covariates and the variates X, Y showed negative correlation between BDI, HADS (minus sign) and SDNN, LF/HF, VO_2peak, MSSD, pNN50 (plus sign), at both baseline and follow-up measurements. At baseline, BDI, MSSD, SDNN, pNN50, and VO_2peak had the largest correlations with the variate X, whereas at follow-up, VO_2peak, pNN50, BDI, and HADS had the largest correlations with Y. These findings indicate a significant negative correlation between psychological and HRV indices which after exercise appeared stronger between BDI, HADS (minus sign) and VO_2peak, pNN50 (plus sign). For group B, CCA showed two significant pairs of canonical variates X1, Y1 and X2, Y2. The correlation between X1 and Y1 was 0.999. The variate X1 accounted for 39% of the variance of baseline measurements and the variate Y1 accounted for 35% of the variance of follow-up measurements (Table 7). Examination of the correlations between the covariates and the variates X1, Y1 showed positive correlation among SDNN, MSSD, and pNN50, at both baseline and follow-up measurements, while the correlations for the other covariates were negligible. The correlation between X2 and Y2 was 0.998. The variate X2 accounted for the 28% of the variance of baseline measurements and the variate Y2 accounted for the 26% of the variance of follow-up measurements (Table 7). Examination of the correlations between the covariates and the variates X2, Y2 showed positive correlation between BDI and HADS, at both baseline and follow-up measurements, while the correlations for the other covariates were negligible. At baseline, BDI, MSSD, SDNN, pNN50, and VO_2peak had the largest correlations with the variate X, whereas at follow-up, VO_2peak, pNN50, BDI, and HADS had the largest correlations with Y. The findings for group B showed strong correlation either within HRV indices or within psychological indices but negligible correlation between them.

Table 5

Pearson's correlation coefficients between baseline and follow-up measurements for group A

Baseline/follow-up	VO2peak	SDNN	LF/HF	MSSD	pNN50	BDI	HADS
VO2peak	0.937b	0.611b	0.590b	0.468a	0.789b	–0.846b	–0.689b
SDNN	0.877b	0.648b	0.548b	0.556b	0.890b	–0.835b	–0.728b
LF/HF	0.619b	0.429a	0.880b	0.490a	0.555b	–0.531b	–0.637b
MSSD	0.797b	0.467a	0.472a	0.385	0.971b	–0.789b	–0.649b
pNN50	0.752b	0.444a	0.415a	0.353	0.984b	–0.744b	–0.607b
BDI	–0.942b	–0.487a	–0.597b	–0.411a	–0.797b	0.915b	0.769b
HADS	–0.733b	–0.397	–0.602b	–0.435a	–0.608b	0.710b	0.870b

BDI, Beck Depression Inventory Questionnaire; HADS, Hospital Anxiety and Depression Scale Questionnaire; HD, hemodialysis; HF, high-frequency power; LF, low-frequency power; MSSD, mean square successive differences; pNN50, percentage of RR intervals differing by more than 50 ms from the preceding RR; SDNN, standard deviation of RR intervals. ^aCorrelation is significant at the 0.05 level (two-tailed). ^bCorrelation is significant at the 0.01 level (two-tailed).

Table 6

Pearson's correlation coefficients between baseline and follow-up measurements for group B

Baseline/follow-up	VO2peak	SDNN	LF/HF	MSSD	pNN50	BDI	HADS
VO2peak	0.785a	0.175	0.435	0.093	0.197	–0.262	–0.209
SDNN	0.106	0.892a	0.194	0.895a	0.820a	–0.083	0.041
LF/HF	0.031	0.068	0.129	0.015	-0.166	–0.218	–0.145
MSSD	0.081	0.753a	0.077	0.998a	0.935a	0.038	0.161
pNN50	0.174	0.733a	0.117	0.967a	0.953a	–0.008	0.108
BDI	–0.081	–0.085	0.054	0.027	0.046	0.992a	0.902a
HADS	–0.024	0.012	–0.057	0.103	0.064	0.911a	0.949a

BDI, Beck Depression Inventory Questionnaire; HADS, Hospital Anxiety and Depression Scale Questionnaire; HD, hemodialysis; HF, high-frequency power; LF, low-frequency power; MSSD, mean square successive differences; pNN50, percentage of RR intervals differing by more than 50 ms from the preceding RR; SDNN, standard deviation of RR intervals. ^aCorrelation is significant at the 0.01 level (two-tailed).

Table 7

Correlations between covariates and canonical variates, and variances explained by canonical variates X, Y for groups A and B

	Covariates
	VO2peak	SDNN	LF/HF	MSSD	pNN50	BDI	HADS
Correlations (group A)
Between baseline and canonical variate X	0.896	0.927	0.716	0.950	0.926	–0.942	–0.775
Between follow-up and canonical variate Y	0.901	0.447	0.630	0.377	0.922	–0.874	–0.721
Correlations (group B)
Between baseline and canonical variate X1	0.077	0.898	0.024	0.998	0.966	–0.02	0.057
Between follow-up and canonical variate Y1	0.052	0.760	0.066	0.998	0.924	–0.006	0.119
Between baseline and canonical variate X2	0.178	0.082	0.256	–0.049	–0.020	–0.992	–0.922
Between follow-up and canonical variate Y2	–0.003	0.063	–0.077	–0.05	–0.092	–0.988	–0.910

Canonical correlation (correlation between X and Y): 0.998 variance of covariates at baseline explained by canonical variate X: 0.775, variance of covariates at baseline explained by canonical variate Y: 0.772, variance of covariates at follow-up explained by canonical variate Y: 0.526, variance of covariates at follow-up explained by canonical variate X: 0.524. Canonical correlation (correlation between X1 and Y1): 0.999 variance of covariates at baseline explained by canonical variate X1: 0.39, variance of covariates at baseline explained by canonical variate Y1: 0.39, variance of covariates at follow-up explained by canonical variate Y1: 0.35, variance of covariates at follow-up explained by canonical variate X1: 0.35. Canonical correlation (correlation between X2 and Y2): 0.998 variance of covariates at baseline explained by canonical variate X2: 0.28, variance of covariates at baseline explained by canonical variate Y2: 0.28, variance of covariates at follow-up explained by canonical variate Y2: 0.26, variance of covariates at follow-up explained by canonical variate X: 0.26. BDI, Beck Depression Inventory Questionnaire; HADS, Hospital Anxiety and Depression Scale Questionnaire; HD, hemodialysis; HF, high-frequency power; LF, low-frequency power; MSSD, mean square successive differences; pNN50, percentage of RR intervals differing by more than 50 ms from the preceding RR; SDNN, standard deviation of RR intervals.

Discussion

The main findings of this study were the demonstration of the beneficial effects of 1-year ET intervention during dialysis on both mental and cardiac autonomic disturbances, as well as the disclosure of the dependence between their depression and cardiac autonomic dysfunction.

Dialysis patients usually display psychosocial problems such as depression, anxiety, and social withdrawal [1,2]. In this study, the prevalence of severe depression symptoms was high on baseline evaluation (14 of 44 patients) and was comparable with previous reports [3,5,30]. It is well known that severe depression is characterized by the presence of a depressed mood and a decreased interest in all activities. Moreover, several authors observed that depressive symptoms increase mortality risk in chronic kidney disease patients similar to the cardiac patients [28,31]. Carney et al. [32] found a positive relationship between depression and HD patients’ aerobic capacity. Juergensen et al. [33] reported that higher depression and anxiety scores in continuous ambulatory peritoneal dialysis patients were associated with higher complication rates in terms of hospitalization, peritonitis, and exit-site infections. Depression has both behavioral and direct pathophysiological effects. It was suggested that pathophysiological effects of depression involve at least the following three mechanisms affecting the morbidity and mortality of the patients: hypercortisolemia, impairment of platelet function, and cardiac autonomic dysfunction [6,9].

Several studies have shown an inverse relationship between depression and HRV in cardiac patients [8,9,10,13]. In contrast, Gehi et al. [34] did not find any association between depression and HRV in 873 outpatients with stable coronary artery disease. Moreover, it was supported that anxiety, but not depression, influenced negatively cardiac vagal modulation in postmyocardial infarction patients [35].

To our knowledge there has not been any study heretofore documenting an association between depression and cardiac autonomic dysfunction in chronic kidney disease patients. However, there are limited data available regarding alterations of HRV in HD patients, as well as the clinical impact of these changes. It was described that reduced HRV was associated with electrical instability of the myocardium and high incidence of ventricular arrhythmias [17–19]. Central and/or peripheral neuropathy, cardiac nerve fiber damage, dysfunction of myocardial pacemaker cells, uremia-related toxins, electrolyte abnormalities, anemia, and deconditioning are the main reasons for cardiac autonomic dysfunction in HD patients, including sympathetic overactivity and vagus reduction, leading to depressed HRV and baroreflex function [20,21,36].

This study provides modest support for reduction of depression in HD patients after a year of ET program during dialysis. This improvement was accompanied by reduction in the level of anxiety. These beneficial effects of ET have already been reported in HD patients [1,3,22,37]. Interestingly, it was shown that the most severely depressed HD patients were benefited to the greatest extent [3]. Moreover, the reduction of depression was accompanied by augmented objective and subjective indices of their health-related quality of life [3,22]. Our findings are concordant with several other studies, although not all, showing an association between physical capacity and depression changes after ET in HD patients [1,3,35,38].

The improvement in cardiorespiratory fitness level had a significant beneficial effect on cardiac autonomic outflow in HD patients. The exercise program during dialysis in this study led to a marked increase in time-domain and frequency-domain indices of HRV. The significant correlations among VO_2peak and SDNN and LF/HF ratio, preexercise, and postexercise, support this observation. A similar influence of ET on HRV indices, as well as on the arrhythmic events after ET on the nondialysis days was described earlier [18]. Furthermore, it is well known that ET improves HRV indices after ET in patients with coronary artery disease, chronic heart failure, diabetes, etc [12,39–41].

Importantly, the present data indicate that the reduction of BDI and HADS scores after ET may affect the autonomic nervous system. Thus, our findings provide evidence for the efficacy of ET in reducing the level of sympathetic response to mental stress in HD patients. Similarly, Blumenthal et al. [42] suggested that 16 weeks of aerobic exercise and stress management training improves cardiac autonomic and endothelial function through improvement in psychosocial adjustment in coronary heart disease patients. A similar beneficial effect of ET on sympathoadrenal response to mental stress in healthy individuals was described by Blumenthal et al. [43] in an earlier study. These findings raise the possibility that antidepressant clinical interventions can be effective in reducing traditional cardiovascular risk factors [8,14,44,45]. In contrast, Sothmann et al. [46] were not able to support the hypothesis that short-term ET alters the sympathetic nervous system activity and behavioral measures in middle-aged men exposed to an acute psychological challenge.

Study limitations

The main limitation is that the study has no long-term clinical follow-up. Therefore, it is not able to answer the question, whether treating depressive symptoms by a physical rehabilitation program reduces the cardiovascular mortality in HD patients. More long-term efforts are required to assess the effects of treating depression on cardiac risk factors, morbidity, and mortality in these patients.

Conclusion

In conclusion, our study has shown that there is a link between depression and HRV indices in HD patients. In addition, ET during dialysis was found to act effectively on both mental and cardiac autonomic disturbances.

Acknowledgements

Conflicts of interest: none.