Nerve Conduction Studies and Prediction of Mortality in Hemodialysis Patients

Abstract

Background. The electrophysiological aspects of uremic neuropathy have been studied extensively, but never for prediction of mortality. In order to assess the parameters of nerve conduction study (NCS) as predictors of mortality in hemodialysis patients, a post hoc analysis of a prospective observation study was made. Methods. We examined conventional electrophysiological parameters (motor nerve conduction velocity [MCV], terminal latency [TL], and F wave latency of the peroneal nerve, as well as sensory nerve conduction velocity [SCV] of the sural nerve) in 75 nondiabetic patients. Hemodialysis modality (bicarbonate dialysis and biocompatible membranes), Kt/V, comorbidity (ischemic heart disease and congestive heart failure), and clinical and laboratory parameters were also evaluated. Survival was analyzed using the Cox proportional hazard model. Results. SCV was significantly higher (t-test, p < 0.01) in the group of patients treated with polysulfone compared to those using cuprophane membranes. On the other hand, MCV significantly correlated with Kt/V (Pearson, r = 0.388; p < 0.01). Multivariate Cox regression revealed only MCV as a significant predictor of mortality in this group of hemodialysis patients (HR = 0.92; CI (0.86–0.99); p < 0.05). Conclusion. Only MCV was a significant mortality risk predictor among NCS parameters. This parameter correlates significantly with dialysis dose. SCV was related to the use of biocompatible membranes, indicating the complexity of polyneuropathy in dialysis patients.

Keywords:

• hemodialysis
• retrospective study
• survival analysis
• polyneuropathy
• nerve conduction study

INTRODUCTION

Polyneuropathy in chronic renal failure has been a well-known problem for a long time.[1] It is a distal, sensorimotor polyneuropathy, pathologically characterized by segmental demyelination and remyelination as well as axonal degeneration.[2] The electrophysiological aspects of uremic neuropathy have been studied extensively. The involvement of peripheral motor and sensory nerves, at least at the subclinical level, is a nearly constant finding in chronic renal failure.[3] Although most hemodialysis patients show electrophysiological evidence of uremic neuropathy, no remarkable electrophysiological change in uremic neuropathy during hemodialysis treatment was recognized in several investigations.[4] On the other hand, peripheral neuropathy may deteriorate during both hemodialysis and continuous ambulatory peritoneal dialysis (CAPD), more severely in hemodialysis patients.[5] Polyneuropathy has been associated with age and duration of dialysis treatment.[6] The most sensitive electrophysiological parameter of polyneuropathy is F wave latency, which may also be used as a parameter of dialysis adequacy.[6] Dialysis with polysulfone membranes acutely improved SCV more than with cuprophane membranes.[7] Clinical, morphological, and morphometric indications of recovery from uremic neuropathy were observed in transplant recipients.[8] However, to the best of our knowledge, NCS parameters in the prediction of morality of hemodialysis patients have not been studied yet.

Therefore, a post hoc analysis of a prospective observation study was made in order to evaluate NCS parameters in the prediction of mortality in hemodialysis patients.

PATIENTS AND METHODS

This post-hoc analysis of an observational study included a cohort of 75 patients (40 male and 35 female) on maintenance hemodialysis in the Dialysis Unit of the Clinical Center of Serbia.[9] The design of the study was approved by the Ethics Committee of the Clinical Center.

All patients were advocated to complete NCS studies regardless of whether they had clinical findings of polyneuropathy or not. The investigation included patients who had undergone NCS in the period from January 1, 1995, to December 31, 2001, and who had been on hemodialysis for up to 213 months (66.9 ± 48.8).

Baseline measurements were made during the first calendar year after the patients entered the study (January 1 to December 31) and involved clinical, nerve conduction study and monthly laboratory examinations. Many patients in this cohort did not undergo repeated nerve conduction studies, so only baseline measurements were analyzed. Diabetic patients were excluded. Patients with systemic diseases showed no significant difference in NCS parameters compared to the other patients, so they were included in this analysis.

All patients were monitored until their death (36 patients), kidney transplantation (5 patients), departure from the center (4 patients), or drop out of the study (3 patients), or until the end of the study on December 31, 2004 (15 men and 12 women).

NCS parameters included motor nerve conduction velocity (MCV), terminal latency (TL), F wave latency of the peroneal nerve, and sensory nerve conduction velocity (SCV) of the sural nerve. F wave latencies were adjusted for the patient's height.

Kt/V was calculated using the second generation Daugirdas formula[10]:

where R is post-dialysis/pre-dialysis blood urea nitrogen, t is dialysis hours, UF is pre-post dialysis weight change, and W is post-dialysis weight.

During the baseline period, 48 patients were on i.v. and oral vitamin B supplements, and no one used alpha-lipoic acid or other medicaments that may influence NCS. Most of them had started such therapy years ago and continued to take it during the follow-up period.

Statistical Methods

The outcome variable in Cox regression was survival time in months. Predictor variables used in Cox regression were derived from data in the baseline period and at entry into the study.

Variables Obtained in the Baseline Period

NCS parameters were measured only once during the baseline period, and separate variables were used for MCV, TL, F-wave latency, and SCV. A separate binary variable for F wave latency was constructed to include all the patients with measured values below the normal and another 20 patients in whom F wave latency could not be measured.

Separate binary variables were constructed for bicarbonate and polysulfone membrane dialysis. Individual mean values of laboratory parameters, Kt/V, and hours of dialysis per week were calculated for each patient from monthly measurements in the 12-month baseline period, and a separate variable was constructed for each parameter.

Comorbidity was determined during the baseline period, and single binary predictor variables were used for ischemic heart disease and congestive heart failure.

Variables Obtained at Entry into the Study

The following predictor variables were determined when the patients entered the study: age (years), gender (male/female) and the time of dialysis (months). All binary variables were coded as 0/1 except gender (coded 1/2).

Pearson and Spearman correlation coefficients were used for detecting associations among predictor variables.

Univariate survival analysis was performed with the Cox proportional hazard model. The primary dependent variable was the time to death measured in months. Variables that were potential predictors of death in univariate analysis (p < 0.10) were tested in the multivariate Cox proportional hazard model using the stepwise method. Proportional hazards were tested by fitting time-varying models.

RESULTS

Demographic data for the 75 patients are presented in Table 1. Their age ranged from 24 to 76 years, and at the beginning of the study, they had been on hemodialysis for 3–213 months. Diabetic patients were excluded. The majority of the patients were dialyzed with bicarbonate solution thrice per week for four hours, and one-third of them used polysulfone membranes. High-flux membrane was used in the dialysis of only three patients. Duration on dialysis ranged from 3 to 213 months, and there were only nine patients with 3 to 12 months duration of dialysis.

Table 1 Demographic data, primary renal diagnosis, co-morbidity, and clinical parameters for the patients (N = 75)

Variable	Mean ± SD or percent (counts)
Age (years)	52.7 ± 12.2
Sex (male/female)	40/35
Time of dialysis (months)	66.9 ± 48.8
Primary renal diagnosis
Chronic glomerulonephritis	21.3 (16)
Hypertension	9.3 (7)
Polycystic kidney disease	12.0 (9)
Chronic interstitial nephropathy*	6.7 (5)
Renal calculi	4.0 (3)
Systemic diseases	6.7 (5)
Vesicoureteral reflux	4.0 (3)
Other	8.0 (6)
Unknown	28.0 (21)
Comorbidity
Ischemic heart disease	26.7 (20)
Congestive heart failure	41.3 (31)
Laboratory variables
Hemoglobin (g/L)	89.6 ± 11.5
Urea (mmol/L)	27.9 ± 4.4
Creatinine (μmol/L)	1091 ± 207
Calcium (mmol/)	2.28 ± 0.20
Phosphorus (mmol/L)	1.83 ± .0.33
Albumin (mmol/L)	38.0 ± 4.5
Kt/V (Daugirdas)	1.21 ± 0.23
Hours per week	11.4 ± 1.8
Bicarbonate dialysis	80.0 (60)
Polysulfone membrane	32.0 (24)
Vitamin B	48 (64)

*Chronic interstitial nephropathy includes Balkan endemic nephropathy, chronic interstitial nephritis, and chronic pyelonephritis.

NCS parameters are summarized in Table 2. MCV and TL were measured in all patients, while F wave latency could be obtained in 47 patients and SCV in 66 patients.

Table 2 Nerve conduction study

Variable	Normal values (range)	Mean ± SD or percent (counts)	Percent (counts of abnormal values)	N
Peroneal nerve
Motor conduction velocity (m/s)	50.7 (44–59)	41.1 ± 5.4	65 (49)	75
Terminal latency (ms)	4.9 (3.3–5.9)	4.1 ± 0.9	15 (11)	75
F wave latency (ms)	Dependent on height^†	53.6 ± 7.6	15 (7)	47
F wave latency (binary)	[See methods]	40.0 (27)		67
Sural nerve
Sensory conduction velocity (m/s)	50 (43–59)	47.3 ± 8.9	33 (22)	66

*Below the lower value of the range except for F wave latency (below the normal value).

^†Height, cm (normal value): 155 (45), 160 (46), 165 (47.5), 170 (49), 175 (50), 180 (51.5), 185 (53), 190 (54), 195 (55.5).

An assessment of the difference between the acetate and bicarbonate groups revealed no significant difference in NCS parameters. On the other hand, there was a significant difference between cuprophane and polysulfone membranes, but only in SCV (t-test, p < 0.01). It was not possible to test the effect of high-flux membranes because only three patients had used them.

Correlations among nerve conduction parameters were significant between motor conduction velocity and terminal latency (Pearson, r = −427; p < 0.01, N = 75) and between motor conduction velocity and sensory conduction velocity (r = 375; p < 0.01, N = 66). The F-wave latency correlated significantly with motor conduction velocity (r = −785; p < 0.01, N = 47) and terminal latency (r = 494; p < 0.01, N = 47). Among nerve conduction parameters, only F-wave latency and motor conduction velocity were considered as collinear.

MCV and F wave latency correlated significantly with Kt/V (r = 388; p < 0.01, N = 75 and r = −414; p < 0.01, N = 47, respectively). SCV velocity correlated significantly with serum phosphate (r = −260; p < 0.05, N = 66). There were no parameters considered as collinear.

Mortality rate was 48% (36/75) during the observation period. Potential predictors of death were selected by the univariate Cox proportional hazard model (p < 0.10; see Table 3). Among NCS parameters, only MCV was selected for further testing. To confirm the independent predictive power of patient mortality, the above-mentioned data from Table 3 were entered into the multivariate Cox proportional hazard model (see Table 4). Age and MCV were selected as independent predictors of all-cause mortality. The risk for patient mortality increases by 8% for every 1 m/s decrease of MCV when it was adjusted for other risk factors. Testing the previous model with the MCV binary variable (lower than mean value coded with 0 and higher with 1) revealed a significant difference in survival between these patients (see Figure 1; RR = 0.41, 95% CI (0.19–0.87), p = 0.02).

Table 3 Potential predictors of patient mortality (by univariate Cox proportional hazard model, p < 0.10; hazard ratio [95% confidence interval]/significance)

Variable	Hazard ratio (CI)/p
Age (years)	1.04 (1.01–1.07)/0.00
Congestive heart failure	2.02 (1.06–3.86)/0.03
Kt/V	0.22 (0.05–1.00)/0.05
Calcium	0.22 (0.05–0.99)/0.04
Phosphorus	2.45 (0.92–6.50)/0.07
Hours per week	0.82 (0.91–0.94)/0.00
Motor nerve conduction velocity	0.92 (0.86–0.99)/0.02

Hazard ratio = Exp (B), CI = 95% confidence interval of Exp (B), p = significance of coefficient.

Table 4 Predictors of all-cause mortality in the cohort (N = 75), multivariate Cox proportional hazard model

Variable	Hazard ratio (CI)	p
Age	1.04 (1.01–1.07)	0.00
Motor nerve conduction velocity	0.92 (0.86–0.99)	0.02

Hazard ratio = Exp (B), CI = 95% confidence interval of Exp (B), p = significance of coefficient.

Figure 1. Difference in survival between the patients who had MCV higher and lower than the mean value (multivariate Cox regression).

DISCUSSION

The present study examined the predictive power of NCS parameters on all-cause patient mortality. The results obtained showed that, among them, only MCV was a predictor of mortality. To the best of our knowledge, no studies have examined the predictability of different NCS parameters on all cause patient mortality.

We also found a significant correlation between MCV and Kt/V, which additionally emphasizes the importance of this NCS parameter. Our results confirmed the relation between F wave latency and Kt/V.[6] However, this sensitive electrophysiological parameter was not a predictor of mortality here. The relation between cuprophane and polysulfone membranes and SCV are in accordance with previous studies[7] and emphasize the possible clinical complexity of different nerve conduction study parameters. Nevertheless, there are more questions to be answered. Could dose of dialysis and the type of dialyzer membrane have a different effect on NCS parameters? In the case of a positive answer, both of them might be important for the treatment of polyneuropathy. In either case, further investigations are needed.

There were a few drawbacks in our study. It was a post-hoc analysis of an observational study, as there was no a priori hypothesis. The number of patients (i.e., 75) is relatively small for survival analysis, but the observation period is long (10 years), though 36 of our patients did not survive this period. In this way, survival analysis was realistic, and parameters that are usually significant in larger samples were also significant here.

In addition, an important shortcoming of this study was the low Kt/V. Although the target Kt/V was aimed to be higher than 1.2, almost half of the cohort could not reach this value because the dialyzer membrane was not of a satisfactory size due to restrictive economic factors, while the time of dialysis was short due to the overcrowded center. However, lower Kt/V values aggravate polyneuropathy and thus ease the estimation of this condition.

Our results showed that MCV was the only predictor of mortality among the measured NCS parameters. Although it is not the most sensitive parameter of polyneuropathy, it correlated significantly with the dialysis dose. In addition, the dose of dialysis and the dialyzer membrane might have a significant effect on nerve conduction study parameters, but further investigations are needed to confirm this hypothesis.

DECLARATION OF INTEREST

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

ACKNOWLEDGMENTS

Data presented in this study were obtained from the doctoral dissertation of Dr. Milan Stosovic, entitled “Analysis of different dialysis adequacy indices based on clinical parameters upon different dialysis doses in hemodialysis patients,” approved by the Scientific Committee, School of Medicine, University of Belgrade, IRB vote number 7/5, April 21, 1992.