Blood Pressure and Heart Rate Variability in Patients on Conventional or Sodium-Profiling Hemodialysis

Abstract

Background. Autonomic nervous system dysfunction and dialysate sodium (Na) concentration are believed to play a role in the pathogenesis of hemodialysis-related hypertension. The present study was undertaken to determine whether increases in blood pressure in hemodialysis patients are associated with changes in heart rate variability (HRV), a measure of the autonomic nervous system function, and long-term exposure to increased dialysate Na concentration. Methods. Baseline clinical, biochemical data and HRV of patients undergoing increased Na profiling dialysis (High-Na, n = 9) and on conventional treatment (Control, n = 11) were compared with those obtained after one year of study. Results. After one year, the mean predialysis systolic blood pressure (SBP) increased in seven patients of the High-Na and in five of the Control group, and decreased or remained unchanged in the remaining subjects. Initial HRV was significantly higher in patients with increased SBP, and it increased further in these patients after one year. At the end of the study, post-dialysis plasma Na, osmolality, and weight gains were significantly higher in the High-Na group. No significant correlation, however, was found between individual changes in intradialytic sodium elimination and the alterations in blood pressure. Conclusion. These data suggest that the dialysate sodium concentration, a most important determinant of interdialytic weight gain and fluid balance, is only partly correlated with long-term changes in blood pressure. An increased blood pressure over time may develop in a subset of hemodialysis patients with higher HRV, suggestive of increased sympathetic activity.

Keywords:

• autonomic nervous system
• blood pressure
• dialysis hypotension
• heart rate variability
• sodium dialysate profiling

INTRODUCTION

Dialysis-related hypertension is a significant predictor of all-cause mortality and cardiovascular survival rate in patients with end stage renal disease. The worldwide experience suggests a clear link between elevated blood pressure and increased cardiovascular mortality.[1],[2] Acute hypotension, the most important cause of intradialytic morbidity, occurs in 20–35% of patients, especially in those with ischemic heart disease, reduced left ventricular systolic function, and diabetes mellitus. Increasing the dialysate Na concentration, a frequently used technique to reduce the frequency and the severity of intradialytic hypotensive episodes and of other adverse effects, may lead to greater interdialytic weight gain and alterations in the blood pressure.[3],[4] An elevation in plasma tonicity during the high-Na period provides a greater driving force for plasma refilling with reduced incidence of decreased blood pressure episodes. Most clinical studies on Na profiling report that the decrease in the number of hypotensive episodes is associated with increases in thirst, body weight, interdialytic hypertension, and a positive Na balance.[3],[4] The increased blood pressure is assumed to be the result of an expanded extracellular volume with elevated cardiac output due to the positive salt balance and an increased peripheral resistance caused by the Na-dependent elevation in intracellular calcium.[5] Most studies on increasing the dialysate Na profiling were conducted over short-term periods, with insufficient data on the long-term effects of this therapeutic modality.

Dysfunction of the autonomic nervous system is believed to play an important role in the pathogenesis of both hemodialysis-associated hypotension and increased blood pressure.[6–8] Heart rate variability (HRV) is a powerful noninvasive clinical tool for the assessment of the autonomic nervous system function. Several studies have shown a reduced HRV in patients undergoing chronic hemodialysis, especially in those with frequent hypotensive episodes. These data were interpreted as evidence for a diminished efficiency of the autonomic control of cardiovascular function, resulting in hemodynamic instability and hypotension.[6],[8] There is, however, limited information on the correlation of the effects of various therapeutic maneuvers designed to improve patients' hemodynamic stability, such as that of increasing dialysis Na concentration, with HRV.

Thus, the first objective of the present study was to determine whether the long-term variations in blood pressure are related to changes in heart rate variability (HRV), a measure of the autonomic nervous system function. The second objective of the study was to assess whether these alterations are related to increased Na dialysate concentration. Therefore, clinical data and HRV measurements were obtained in a group of patients dialyzed against Na profiling for one year, and compared to those of patients in whom the Na dialysate concentration was maintained constant.

PATIENTS AND METHODS

The study was conducted in the Hemodialysis Unit of Hadassah Hebrew University Medical Center, in which about 100 chronic patients undergo maintenance therapy, and was approved by the hospital's ethic committee, as part of a project on dialysis-related alterations in blood pressure. Twenty patients that consented to participate were selected to take part in the study. The remaining patients were excluded because of chronic atrial fibrillation or frequent ventricular premature beats, debilitating illness with frequent hospital admissions, permanent pacemakers, young age (less than 25 years), refusal to participate in the study, or lack of cooperation. Before the study, the patients were observed during a five-week adjustment period to assess the dry weight, blood flow rate, and antihypertensive medication by the treating nephrologists. During the adjustment period, the patients' diet, with an emphasis on Na intake, was evaluated by a nephrological dietician. The patients were instructed to maintain a daily dietary Na intake of 150–180 mmol/24 h and were regularly reassessed by the dietician during the whole study period. The patients were divided into two groups:

• dialyzed using high Na dialysate profiling (High-Na, n = 9), and

• control group, undergoing conventional dialysis (control, n = 11),

and were followed up for one year. The patients underwent hemodialysis three times weekly, 3.5 to 4 hours each time. Hemodialysis was performed using the Gambro AK 200 delivery system. The mean urea reduction rate (% URR) in all patients was 65–75%. The dialysis was performed using a polysulphone high flux dialyzer (1.3 m²). The ultrafiltration rate (% body weight reduction) during dialysis sessions was adjusted according to the presumed dry weight. The dry weight was estimated regularly on the basis of physical examination and plain chest x-ray. The % weight gain was estimated according to the formula:

The mean interdialytic weight gain was calculated as the mean of interdialytic weight gains during the five weeks preceding the beginning of the study and the mean of the five weeks of the end of the study. The lowest weight a patient could tolerate without development of hypotension or symptoms as dizziness, weakness, nausea, or cramps was defined as “dry weight.” The patients were considered to be overhydrated when edema on physical examination and/or when cardiomegaly with pulmonary congestion or pleural effusion on chest x-ray were detected. In most patients, an upper abdominal ultrasound was performed to further exclude the presence of a congested liver or inferior vena cava dilatation.

At the beginning and at the end of the study periods, the patients were requested to perform 24 h urine collections. Mean urine output was defined as the mean of at least two urine collections. The fluid intake was defined as the sum of the mean urine output and of the mean interdialytic weight gain and was evaluated for the five-week period before and for the last five weeks of the survey.

The temperature of dialysate was kept constant at 36°C. The dialysate flow rate was kept at 500 mL/min. The dialysate contents were: chloride 108 mmol/L, bicarbonate 35 mmol/L, potassium 2 mmol/L, and calcium 1.5 mmol/L. During conventional dialysis, the Na concentration was kept constant at 140 mmol/L. During Na profiling (step-down) dialysis, the dialysate Na concentration was increased to 144 mmol/L and gradually decreased to 140 mmol/L toward the end of the session.

The intradialytic sodium elimination was calculated using a simple single-compartment mathematical model as previously described.[9],[10] According to the proposed model, the sodium elimination during dialysis (J_Na) can be expressed as:

where Q_F is the ultrafiltration rate, Q_B is the blood flow at the dialysis inlet, C_Na,d is the sodium concentration in the dialysate, C_Na,_e is the sodium concentration in the extracellular fluid, and D is the sodium dialysance. We used plasma sodium as representative of extracellular sodium concentration without correction for plasma water fraction and the Donnan effect, as previously shown.[9],[10]

Measurements of the blood pressure during dialysis, digitally determined every 20 minutes, were compared with those obtained manually using a sphygmomanometer and were recorded for each dialysis. During hypotensive episodes, blood pressure was measured by two different nurses and electronically recorded every 5–10 minutes. To determine the number of the hypotensive episodes, the patients' dialysis charts of the two months preceding the study and of the last two months of the study were carefully screened. A hypotensive episode was defined as a symptomatic reduction in the mean blood pressure greater than 20% of the basal predialytic value and requiring intravenous saline administration or necessitating temporary cessation of ultrafiltration or the dialysis.

The patients' baseline clinical characteristics and other initial data are listed in Table 1. Other initial data are shown in Tables 2 and 3. The routine medication in all patients included calcium, vitamin B supplements and vitamin D as needed, sevelamer hydrochloride (Renagel) in selected cases, and recombinant erythropoietin. All patients underwent routine echocardiography for determination of left ventricular function. One patient in the High-Na group and three in the control group were treated with angiotensin-converting enzyme (ACE) inhibitors (p = NS); two patients from each group were treated with calcium channel blockers (p = NS). Four patients in the High-Na group and five patients in the control group (p = NS) were treated with beta-blockers in the dialysis-free days. Antihypertensive medications were not changed during the year of the study.

Table 1 Patients' baseline characteristics

	High-Na (n = 9)	Control (n = 11)
Gender (M/F)	4/5	6/5
Age (y)	64.6 ± 12	56.3 ± 4.9
DM (no./%)	2 (22.2)	2 (18.2)
Duration of HD treatment (y)	4.33 ± 1.59	4.32 ± 1.50
History of hypertension (no. /%)	8 (88.9)	7 (63.6)
Ischemic heart disease (no. /%)	6 (66.7)	4 (36.4)
LV systolic dysfunction^† (no. /%)	4 (44.4)	2 (18.2)
Dry weight (kg)	74.6 ± 5.9	64.1 ± 5.1

*mean + SE

^†Left ventricular ejection fraction (LVEF) < 40%.

Table 2 Effect on increasing Na dialysate profiling on plasma biochemistry

	Baseline		End of study		p (baseline vs. end )
Dialysis modality	High-Na	Control	High-Na	Control	High-Na	Control
Na (mmol/L)-Pre-D	140.3 ± 0.8	139.9 ± 0.8	140.8 ± 0.9	140.6 ± 0.3	NS	NS
Na (mmol/L)- Post-D	140.6±0.5	140.5 ± 0.4	142.0 ± 0.5^†	140.1 ± 0.5	0.056	NS
Urea (mmol)-Pre-D	23.8 ± 2.2	21.4 ± 2.3	23.9 ± 1.5	25.3 ± 1.5	NS	NS
Urea (mmol/L)-Post-D	8.1 ± 1.1	8.0 ± 0.9	7.8 ± 0.9	6.7 ± 0.8	NS	NS
Creatinine (μmol/L),Pre-D	896 ± 78	840 ± 74	886 ± 70	882 ± 73	NS	NS
Creatinine (μmol/L),Post-D	444 ± 85	347 ± 37	372 ± 51	339 ± 25	NS	NS
Albumin (g/L)	38.8 ± 0.08	38.0 ± 0.07	37.4 ± 0.16	37.8 ± 0.09	NS	NS
Calcium (mmol/L)	2.30 ± 0.10	2.38 ± 0.03	2.30 ± 0.05	2.32 ± 0.04	NS	NS
Hemoglobin (g/dL)	11.6 ± 0.3	11.3 ± 0.2	11.7 ± 0.2	11.6 ± 0.2	NS	NS
Osmolality (mosm /kg H2O)-, pre-D	320.2 ± 2.6	320.0 ± 1.6	322.0 ± 2.6	323.0 ± 1.4	NS	NS
Osmolality (mosm /kg H2O)-, post-D	304.5 ± 2.3	302.5 ± 1.5	307.3 ± 2.3^‡	301.7 ± 1.4	NS	NS

*Pre-dialysis values.

^†p < 0.02; ^‡p < 0.05 vs. control.

Abbreviations: Pre-D = pre-dialysis, Post-D = post-dialysis.

Table 3 Baseline and end-of-study fluid balance and number of hypotensive episodes in High-Na and control patients

	Baseline		End of study		p (baseline vs. end)
Dialysis modality	High-Na	Control	High-Na	Control	High-Na	Control
Weight gain (kg)	2.06 ± 0.24	2.18 ± 0.212	2.37 ± 0.23	2.31 ± 0.21	0.009	NS
% Weight gain	2.76 ± 0.23	3.71 ± 0.31	3.25 ± 0.26	3.71 ± 0.31	0.018	NS
Diuresis (ml)	344 ± 146	186 ± 86	222 ± 106	27 ± 19	0.045	0.046
Fluid intake (mL/24h)	2400 ± 233	2386 ± 195	2600 ± 227	2281 ± 208	0.012	NS
Hypotensive episodes (no./week)	1.50 ± 0.14	0.41 ± 0.19	0.86 ± 0.17	0.52 ± 0.26	0.056	NS
Pre-dialysis SBP (mm Hg)	130 ± 7	137 ± 7	137 ± 1	137 ± 9	NS	NS
Pre-dialysis DBP (mm Hg)	74 ± 3	80 ± 5	74 ± 2	74 ± 4	NS	NS
Pre-dialysis pulse pressure (mm Hg)	56 ± 5	56 ± 6	63 ± 7	63 ± 8	NS	NS

Abbreviations: SBP = systolic blood pressure, DBP = diastolic blood pressure.

*p < 0.001 vs. control.

During the year of the study, the number of all cause hospital admissions and of those due to cardiovascular diseases were monitored for each patient.

Definitions of Clinical Variables

Ischemic heart disease was defined by documentation of prior myocardial infarction or of coronary interventions, the presence of significantly abnormal Q waves on a 12 lead electrocardiogram, symptomatic angina pectoris, or a thallium perfusion scan suggestive of myocardial ischemia. Left ventricular systolic dysfunction was defined as a decrease of left ventricular ejection fraction (LVEF) to less than 40% by echocardiographic measurements. The mean pre- and post-dialysis systolic and diastolic blood pressure and the pulse pressure values were calculated as the mean of three weekly measurements averaged for the five weeks preceding and the last five weeks of the study.

Biochemical Determinations

Biochemical determinations were performed using Vitros Test methodology (Ortho-Clinical Diagnostics, Johnson & Johnson, Raritan, New Jersey, USA). Plasma sodium was determined using Vitros slides (i.e., dry multilayered analytical elements that use direct potentiometry for measurements of sodium ions). Plasma osmolality was determined using a Fiske freezing-point osmometer.

Determination of HRV

Heart rate variability was evaluated in all patients, before and at the end of the study. 24-hour Holter ECG monitoring was performed on the day of the mid-week dialysis session. Marquette Series 8500 Holter recorders were used to record a modified V4 or V5 lead and modified V1 lead. The 24-hour Holter cassette tapes were analyzed on a Marquette MARS analysis system. Files of R-R intervals with their annotations were then transferred to a PC. Four time domain variables were calculated for each patient:

1. mean RR: the mean of mean five-minute RR intervals,

2. SD: the mean of five-minute RR standard deviations,

3. SDANN: the standard deviation of mean five-minute RR intervals, and

4. RMSSD: the mean of five-minute root-mean square of differences of successive RR intervals.

Five-minute epochs of RR intervals as a function of time were used for power spectrum analysis. A 16^th order autoregressive (AR) power spectrum analysis of R-R intervals was performed using the Levinson- Durbin algorithm to solve the Yule-Walker equations.[11]

Three frequency bands were used: VLF (very low frequency), 0–0.05 Hz; LF (low frequency), 0.05–0.2 Hz; and HF (high frequency), 0.2–0.4 Hz. The average spectral amplitudes for each frequency band and for every epoch of 5 minutes were determined. Averages of time and frequency domain variables were calculated for the whole 24 hour period.

Statistical Analysis

Statistical analysis was performed using the SPSS 10.0 for Windows statistical package. The Mann-Whitney U test was used to compare HRV measurements in the High-Na group with those in the control patients. Patients' clinical data are presented as mean ± SE. HRV data are given as median and interquartile range. The Wilcoxon signed ranks test (paired) was used to compare initial and end-of-study HRV measurements of the two subgroups. Student's t-test and χ² test was used when appropriate. p values < 0.05 were considered significant. Logistic regression analysis (stepwise forward conditional procedure) was performed to select determinants of an increase in predialysis systolic blood pressure. The model included the dialysis modality (High-Na or standard) and HRV parameters. Heart rate variability time and frequency domain variables were selected for logistic regression analysis in relation to the dependence between them. Variables with correlation coefficients greater than 0.75 with other variables were excluded, with preference given to time domain variables. The final choice of HRV variables included the baseline RR, SD, RMSSD, SDANN, and LF/HF ratio, and the end of study SD, RMSSD, and SDANN.

RESULTS

Clinical and Demographic Characteristics

Patients' initial clinical characteristics and other baseline data are shown in 1–3. There were no significant differences between groups with regard to gender, age, duration of hemodialysis therapy, dry weight, biochemical parameters, fluid balance, and blood pressure. Each group included two diabetic patients. While the proportion of patients with ischemic heart disease, history of hypertension, and left ventricular systolic dysfunction was higher in the High-Na group, it was not significantly different from that in the Control group.

Effects of High-Na Dialysis on Plasma Biochemistry and on Fluid Balance

Plasma biochemistry, fluid balance, and the number of hypotensive episodes at the beginning and at the end of the study are shown in Tables 2 and 3. The initial biochemical data, the residual diuresis, and the interdialytic weight gain were similar in both groups.

At the end of the study, the post-dialysis plasma sodium level and osmolality were significantly higher in the high Na group. In both groups, the daily urine output decreased at the end of the study year. The interdialytic fluid intake and weight gain, however, significantly increased only in High-Na patients.

During the screening period, intradialytic hypotensive episodes were recorded in eight patients of the High-Na and in 5 of the control groups. One patient from High-Na and three from the control group suffered from muscular cramps during dialysis. The weekly number of hypotensive episodes, initially higher in patients selected to increased Na dialysate profiling protocol, was similar in both groups after one year. During the year of study, the number of all-cause hospital admissions and the number of admissions due to cardiovascular diseases were comparable in both groups (data not shown in tables).

Blood Pressure, Intradialytic Sodium Elimination, and Heart Rate Variability in High-Na and Control Hemodialysis Patients

After one year, the mean predialysis systolic, diastolic, and the pulse pressure values were not different in the High-Na patients group than in the control group (see Table 3). A closer examination of the individual measurements, however, shows that at the end of the study, the predialysis systolic blood pressure had markedly increased in seven patients from the High-Na and in five (p = NS) from the control group (see Table 4). Hypertension in chronic kidney disease and in hemodialysis patient is predominantly systolic.[1],[10] Therefore, predialysis systolic blood pressure was selected as being the most representative of the average blood pressure in stable hemodialysis patients. The mean increase over one year in predialysis SBP in the above patients (the increased SBP group, n = 12) was 14.67 + 1.83 mm Hg, with no differences between those treated with increased Na in the dialysate and those receiving standard dialysis. In patients with increased SBP (n = 12), the end of study mean SBP was 149 ± 6 mmHg as compared with 119 ± 9 mmHg in patients in whom SPB decreased or remained unchanged (n = 8, p < 0.05)). In two patients from the High-Na group and in four from the control group, the predialysis SBP decreased by a mean of 17.33 ± 4.57 mm Hg. In the remaining two patients, both from control group, there was no change in SBP. It is worthwhile to note that the baseline (initial) measurements in patients with increased SBP (n = 12) were similar to those found in patients with decreased or unchanged SBP (n = 8; see Table 4). There were no other significant differences between patients with increased SBP as compared with those with decreased SBP with regard to age, comorbidities, duration of dialysis baseline, or end of study biochemical data.

Table 4 Baseline and end of study predialysis systolic blood pressure (SBP, mm Hg) measurements in patients with increased or decreased SBP over the year of study

	Increased SBP			Decreased SBP
	High-Na (n = 7)	Control (n = 5)	All (n = 12)	High-Na (n = 2)	Control (n = 6)	All (n = 8)
Baseline	132 ± 8	139 ± 7	135 ± 6	125 ± 15	135 ± 12	132 ± 9
End of study	145 ± 8	155 ± 9	149 ± 6	111 ± 5^‡	122 ± 13	119 ± 9^†

*Two patients with unchanged SBP were also included in this subgroup.

^†p < 0.05, ^‡p = 0.068 (NS) vs. patients with increased SBP.

To determine whether the increase in mean predialysis systolic blood pressure observed in specific patients was associated with decreased sodium removal related to the higher dialysate-plasma sodium gradient, the initial and end-of-study J_Na, a measure of intradialytic sodium elimination, were determined in each patient. Figure 1 shows the correlation between individual changes in intradialytic sodium elimination (ΔJ_Na, mmol) during one year (baseline J_Na∼ end- of-study J_Na) and in SBP (ΔSBP, mmHg). After one year, a decreased J_Na, suggestive of sodium retention, was found in seven patients from the High-Na group, but also in four control patients. Increased J_Na, suggesting increased sodium elimination during dialysis, was found in two patients from the High-Na and in seven from the control groups, respectively. No correlation, however, was found between individual changes in J_Na and the variations in SBP.

Figure 1. Changes (baseline and end-of-study) in intradialytic sodium elimination (ΔJ_Na, mmol) and in predialysis systolic blood pressure (ΔSBP, mmHg) after one year in control (triangles), and in High-Na (squares) patients.

To determine whether the changes in the blood pressure observed in our patients were related to alterations in HRV, we compared the initial and the end-of-study 24 h HRV measurements from patients with increased SBP as compared with those with decreased or unchanged SBP. Baseline and end-of-study time domain HRV measurements in patients with increased SBP as compared with those with decreased SBP are shown in Figure 2. The frequency domain HRV measurements are shown in Figure 2. Initial HRV was significantly higher in patients in whom SBP increased during the study. After one year, the HRV differences between the patients with increased SBP and those with decreased or unchanged SBP were even more prominent. A trend toward an increase in HRV in patients with an elevation in SBP and a decrease in HRV in the remaining subjects was noted after one-year follow-up, as compared to the baseline measurements. The decreases in LF and HF in the patients with decreased or unchanged SBP were statistically significant.

Figure 2. Baseline and end-of- study time-domain HRV measurements (median and interquartile range) in patients with increased (n = 12) and decreased or unchanged (n = 8) SBP. Data are presented as box plots. The box stretches from the 25^th to the 75^th percentile. The median is shown with a small black square in the box. The range (the upper and the lower extreme values) is indicated by the whiskers. Time domain measurements: RR (msec) = mean of five-minute mean RR interval between normal beats, SD (msec) = mean of five-minute standard deviation of RRs; SDANN (msec) = standard deviation of five-minute mean normal RR intervals; RMSSD (msec) = mean of five-minute root-mean square of differences of successive RRs.

Logistic regression analysis showed that a higher end of study SD, a time-domain measurement of HRV, is the most significant predictor of the increase in predialysis systolic blood pressure (p = 0.026, odds ratio 1.379, 95% CI (confidence interval) 1.039–1.829).

DISCUSSION

Our data show that in 7 out of 9 patients from the High-Na group and in 5 of 11 control subjects, the blood pressure significantly increased during the year of study. Increased dialysate sodium profiling was associated with higher postdialysis plasma Na concentration, osmolality, and marked increases in interdialytic weight gain and fluid intake. In hemodynamically unstable patients, however, there was also a significant reduction in the number of hypotensive episodes. An increased weight gain and a higher blood pressure during High-Na dialysis, reported in several studies, is usually thought to be the consequence of a positive Na balance and volume expansion due to dialysate to plasma Na transfer.[3],[4] In our patients, regular assessment of their clinical status did not reveal overt signs of volume overload, while the mean dry weight did not change. Therefore, we presume that the volume status of patients from both groups was unchanged. Furthermore, assessment of sodium kinetics using a single compartment simple mathematical model did not show a significant correlation between changes in intradialytic Na elimination at the end-of-study and the alterations in the blood pressure (see Figure 1).

The relationship between sodium, volume excess, and hypertension has been explored by several studies. It has been shown that in salt-sensitive patients, volume control alone fails to reduce blood pressure or left ventricular hypertrophy associated with hypertension,[12–14] while the reduction of the dialysate Na concentration combined with dietary salt restriction over several weeks may result in a decrease in the blood pressure, without additional changes in the dry weight.[15] Moreover, it was shown that the volume removal during dialysis is not significantly affected by changes in the dialysis Na concentration.[16] These observations suggested that even though the fluid balance is efficiently controlled during increased Na dialysate profiling treatment, additional factors may contribute to the observed increase in the blood pressure. The lack of significant changes in the dry weight in both our patient groups over the study period may also support this hypothesis.

While there is general agreement that sodium profiling dialysis reduces the number of hypotensive episodes and intradialytic discomfort, there are contradictory reports regarding the effects of this modality on sodium balance and blood pressure. Thus, clinical and the laboratory signs of sodium accumulation were reported in several (but not all) studies.[3],[4],[9],[10],[17] Individualization of the sodium profiling with regard to the predialysis plasma sodium concentration was shown to improve fluid balance gain and dialysis-induced symptoms in specific patients.[10],[17] An effect on the blood pressure was noted only in patients with uncontrolled hypertension.[17] A recent study reviewing the effects of dialysate sodium concentration in a large group of hemodialysis patients reported a higher blood pressure in those dialyzing with dialysate sodium of 140 mmol/L or more. In this study, however, the use of a decreased dialysate sodium concentration was not necessarily associated with lowest blood pressure, especially in younger male patients.[18] A crossover trial comparing the effects of conventional dialysis with different time-averaged sodium dialysate concentration (TAC_Na) did not find significant correlations between TAC_Na and blood pressure, despite the strong effects on plasma sodium and interdialytic weight gain.[19]

The activation of the sympathetic nervous system, well recognized in chronic renal disease, is thought to be an important mechanism of hypertension in hemodialysis patients. Studies using intraneural direct recordings of muscle sympathetic nerve activity (MSNA) have shown an increased sympathetic discharge in patients with renal insufficiency and on chronic hemodialysis. These findings were interpreted as evidence for excessive activation of sympathetic activity in renal failure.[20],[21] Interestingly, 14 of 18 patients with increased MSNA described in one of these studies had hypertension.[20]

HRV determination, a frequently used method of assessment of autonomic nervous system function, was found to be decreased in hemodialysis patients, especially in those with coexistent systemic or heart diseases.[6] The reason for the apparent divergence between MSNA and HRV may be related to the fact that the former represents centrally generated post-sympathetic activity to the skeletal muscle circulation, a determinant of blood pressure, while the latter is the result of both sympathetic and parasympathetic, vagal-mediated output controlling the heart rate. Furthermore, there are regional differences in sympathetic activity in the human body. The majority of studies on HRV in hemodialysis patients focused on hemodynamic instability and hypotensive episodes during ultrafiltration. HRV has not been not properly studied with regard to hypertension in hemodialysis patients.

In our study, despite individual variations, baseline HRV was similar in the High-Na and control groups. The initial measurements in patients in whom the SBP increased (n = 12), however, were higher, and increased even further at the end of the study. In contrast, the HRV parameters of the patients with decrease or no change in SBP (n = 8) were significantly lower than those of the patients with increased SBP both at baseline and at the end of study. In patients with decreased or unchanged SBP, HRV actually decreased during the year of the study (see Figures 2 and 3). The most prominent differences were noted in LF, a spectral domain related to the action of the sympathetic nervous system, and in SD, a parameter representative of overall autonomic activity.[22] The end of study SD was also found to be the most important predictor of the increase in predialysis SBP in a logistic regression model. Previous studies have shown progressive increases in LF (and reduced HF) relative powers of HRV, as well as reduced responsiveness to standing up in patients with essential hypertension. A reduced variation of 24 hr changes in the normalized spectral profile has also been reported using Holter monitoring.[23] These changes were interpreted as an indication of a shift of the autonomic drive to sympathetic excitatory mechanism. Thus, in our patients as well as in those with essential hypertension, the increase in the blood pressure may be related to the hyperactivity of the sympathetic nervous system, leading to alterations in HRV.

Figure 3. Baseline and end-of- study frequency-domain HRV measurements (median and interquartile range), in patients with increased (n = 12) and decreased or unchanged (n = 8) SBP. Data are presented as box plots. The box stretches from the 25^th to the 75^th percentile. The median is shown with a small black square in the box. The range (the upper and the lower extreme values) is indicated by the whiskers. Frequency domain measurements: VLF (0–0.050 Hz) = very low frequency power; LF (0.05–0.2 Hz) = low frequency power; HF (0.2–0.4 Hz) = high frequency power, LF/HF = LF/HF ratio × 100. VLF, LF, HF (msec²/Hz), and LF/HF values were evaluated from consecutive five-minute epochs.

Clinical and experimental studies provide evidence in favor of altered sympathetic nervous system activity in salt-sensitive hypertension. Chronic high dietary salt intake may lead to a disturbance of central sympathetic inhibition followed by an enhanced peripheral sympathetic tone. HRV studies in patients with salt-sensitivity hypertension have shown an increased LF, especially at night. These findings were construed as evidence for sympathetic hyperactivity and /or blunted parasympathetic modulation in salt-sensitive hypertensive subjects.[24],[25] In the present study, there were no differences in HRV in patients with an increased SBP who were dialyzed using Na profiling as compared with those on standard therapy (data not shown). In addition, although seven of nine patients with higher SBP belonged to the High-Na group, the dialysate Na concentration alone did not predict the increase in the blood pressure in the logistic regression model that included HRV parameters. This apparent lack of correlation in our patients could be related to the blunting effect of the patients' age on HRV, well documented in salt-sensitive hypertension,[26] or to the relative small number of subjects included in the study. Other confounding factors, including gender, the presence of ischemic heart disease and/or diabetes mellitus, and antihypertensive medication, may all affect HRV.[6],[22]

We recognize several limitations to our study. First, the study included a relatively small number of patients with a broad spectrum of clinical manifestations and large range distribution of HRV. Second, due to the long duration and the natural course of illnesses in this specific population, the undertaking of a crossover study to strengthen our data was not applicable.

In summary, taken together, our data suggest that the dialysate sodium concentration, a most important determinant of intradialytic weight gain and fluid balance, is only partly correlated with long-term changes in blood pressure. An increased blood pressure over time may develop in a subset of hemodialysis patients with higher HRV, suggestive of increased sympathetic activity. These preliminary observations need to be confirmed by additional more extensive studies.

ACKNOWLEDGMENT

This study was partly sponsored by Hadassah Women's Health Initiative.