Abstract
Background
To compare the effects of hemodialysis (HD) and peritoneal dialysis (PD) on peripheral nervous system function and frailty in non-diabetic end-stage renal disease (ESRD) patients.
Methods
This prospective, single-center, cross-sectional study consecutively enrolled 88 non-diabetic ESRD patients (HD: n = 45, PD: n = 43) receiving dialysis for ≥3 months at Bursa City Hospital between 1 May and 16 September 2025. All underwent neurological examination and nerve conduction studies. Frailty was assessed using the Fried Frailty Phenotype (FFP), classifying patients as robust (0), pre-frail (1–2) or frail (≥3). Multivariable linear regression identified independent correlates of frailty score.
Results
Median age was 57 (HD) versus 52 years (PD) (p = 0.077). HD patients had higher potassium (5.1 ± 0.7 vs 4.7 ± 0.6 mmol/L, p = 0.025) and ferritin (763 vs 406.5 ng/mL, p < 0.001) and a trend toward higher CRP (8.1 vs 3.2 mg/L, p = 0.075). PD patients showed better dialysis adequacy (Kt/V 1.9 vs 1.6, p < 0.001) and a trend toward higher sural sensory conduction velocity (54 vs 48 m/s, p = 0.052). Axonal polyneuropathy predominated (37.5%), without difference between modalities. Median FFP score was 1 in both groups (p = 0.705), but frailty-category distribution differed (frail: HD 20.0% vs PD 6.9%, p = 0.047). Univariately, CRP correlated with frailty in HD (rs = 0.31, p = 0.038); age (rs = 0.34, p = 0.041), hemoglobin (rs = 0.39, p = 0.014) and Kt/V (rs = −0.37, p = 0.027) correlated in PD. In multivariable regression, CRP remained an independent correlate (B = 0.017, 95% CI 0.001–0.033, p = 0.042).
Conclusion
Although median FFP scores were similar, frailty distribution and its correlates differed by modality – inflammation in HD, age, and dialysis adequacy in PD. These hypothesis-generating findings warrant larger multicenter evaluation.
Keywords: Hemodialysis, peritoneal dialysis, electromyography, uremic neuropathy, frailty, Fried Frailty Phenotype
PLAIN LANGUAGE SUMMARY
Kidney failure patients receiving dialysis often develop nerve damage and physical frailty, both of which reduce quality of life and increase the risk of falls, hospitalization, and death. It is not clear whether the two main forms of dialysis – hemodialysis (HD), which is given several times a week in a clinic, and peritoneal dialysis (PD), which is performed continuously at home – differ in their effects on these problems. In this prospective study of 88 non-diabetic dialysis patients at a single Turkish center, we found that nerve function was broadly similar between the two dialysis types, but the proportion of patients meeting the criteria for full frailty was almost four times higher in HD than in PD patients (20% vs 5%). Inflammation appeared to be linked to frailty in HD patients, whereas older age and lower dialysis efficiency were linked to frailty in PD patients. These findings suggest that frailty should be routinely assessed in dialysis clinics and that different strategies may be needed for the two treatment types, though larger studies are needed to confirm this.
KEY MESSAGES
What is known: Peripheral neuropathy and frailty are highly prevalent in dialysis patients and independently predict adverse outcomes, but comparative data on their occurrence and correlates between hemodialysis HD and PD are limited.
This study adds: In a prospective non-diabetic cohort, electrophysiological findings were broadly similar between HD and PD, yet the prevalence of frank frailty was approximately fourfold higher in HD (20.0% vs 5.4%), with inflammation independently associated with the frailty score.
Potential impact: Routine frailty assessment in dialysis care and modality-specific evaluation of interventions – targeting inflammation in HD and dialysis adequacy and age-related vulnerability in PD – merit prospective testing in larger, multicenter studies.
Introduction
Chronic kidney disease (CKD) represents a significant global health burden with profound implications for patient outcomes and healthcare systems worldwide. End-stage kidney disease (ESKD) affects approximately 90,000 patients receiving kidney replacement therapy in Turkey, with the majority treated by hemodialysis (HD), while peritoneal dialysis (PD) serves as an alternative treatment modality for approximately 4–10% of ESKD patients [1,2]. The choice between these dialysis modalities extends beyond simple efficacy considerations, encompassing quality of life and cardiovascular outcomes.
Neurological complications are among the most debilitating consequences of CKD, with peripheral neuropathy emerging as a predominant concern that significantly impacts patient quality of life and functional independence [3]. The prevalence of uremic neuropathy in dialysis patients ranges from 60% to 90%, making it one of the most common non-cardiovascular complications in this population [4]. Uremic neuropathy typically manifests as a distal symmetric sensorimotor polyneuropathy characterized by axonal degeneration with secondary segmental demyelination, involving accumulation of uremic toxins, electrolyte imbalances, chronic inflammation, and metabolic disturbances [4,5].
Electrophysiological studies serve as the gold standard for objective assessment of peripheral nerve function and monitoring disease progression in uremic neuropathy. Recent studies have demonstrated that peripheral neuropathy is already present in the predialysis stage of CKD, with electrophysiological studies being more sensitive than clinical examination for detecting subclinical nerve dysfunction [6,7]. Despite the clinical importance of peripheral neuropathy in dialysis patients, comparative studies examining the differential effects of HD versus PD on nerve function remain limited [8–10]. The continuous nature of PD therapy, as opposed to intermittent HD sessions, may provide more stable homeostatic conditions that could theoretically offer neuroprotective benefits [8].
Beyond peripheral neuropathy, frailty has emerged as another critical complication affecting dialysis patients. Frailty is defined as a medical syndrome characterized by diminished strength, endurance, and reduced physiological function, leading to increased vulnerability to stressors and adverse health outcomes [11,12]. This syndrome affects approximately one-third of CKD patients overall, with substantially higher rates in dialysis-dependent populations [13,14]. Zhang et al. [13] reported that 34.5% of CKD patients demonstrate signs of frailty, while an additional 39.4% present with prefrail symptoms.
The clinical significance of frailty extends far beyond its high prevalence. Frail patients demonstrate dramatically increased mortality risk compared to non-frail individuals, with hazard ratios ranging from 1.94 to 2.87 across various studies [13–15]. Additionally, frailty is strongly associated with increased hospitalization rates, with frail dialysis patients showing 1.5- to 2-fold higher risk of hospital admission [15]. The Fried Frailty Phenotype (FFP), based on five domains (unintentional weight loss, self-reported exhaustion, weak grip strength, slow walking speed, and low physical activity), remains one of the most widely validated assessment approaches and has been applied in both general dialysis and CKD populations [12].
The relationship between dialysis modality and frailty remains underexplored. Given that HD and PD differ fundamentally in their physiological impact – with HD causing intermittent hemodynamic stress and rapid solute shifts, while PD provides continuous, gentler clearance – it is plausible that these modalities may differentially affect frailty development and progression. However, comparative data examining frailty outcomes between HD and PD patients remain limited, representing an important knowledge gap.
Therefore, the aims of this study were: (1) to compare electrophysiological findings between HD and PD patients; (2) to assess and compare frailty status between these modalities; and (3) to explore the associations between frailty and biochemical parameters in each dialysis population. We hypothesized that different dialysis modalities may affect frailty through distinct pathophysiological mechanisms, although we anticipated that any such associations would require multivariable adjustment and prospective confirmation.
Methods
Study design and participants
This was a prospective, cross-sectional, single-center study conducted at the Department of Nephrology, Bursa City Hospital. Consecutive adult patients meeting the inclusion criteria and attending the outpatient HD unit or the PD outpatient clinic between 1 May 2025 and 16 September 2025 were invited to participate. No formal matching procedure was performed; the similarity in group sizes (HD = 45, PD = 43) reflects the relative size of the two outpatient programs during the study window, and no a priori matching for age, sex, or dialysis vintage was applied. Patient flow is summarized in Figure 1 (STROBE diagram): of 142 patients initially screened, 54 were excluded (32 with diabetes mellitus, 8 with other causes of peripheral neuropathy, 5 with active malignancy, 3 with alcohol use disorder, and 6 who declined consent), leaving 88 patients in the final analysis.
Figure 1.
STROBE flow diagram of patient enrollment (supplied as a separate image file). CAPD, continuous ambulatory peritoneal dialysis; ESRD, end-stage renal disease; FFP, fried frailty phenotype; HD, hemodialysis; NCS, nerve conduction studies; PD, peritoneal dialysis.
Inclusion criteria were age ≥18 years, regular dialysis treatment for ≥3 months, and no history of pre-dialysis neuropathy. Exclusion criteria were diabetes mellitus, other diseases known to cause peripheral neuropathy, alcohol use disorder, and active malignancy. Diabetic patients were excluded because diabetic and uremic neuropathy share overlapping electrophysiological phenotypes and would not be reliably separable in this cross-sectional design; we acknowledge that this restricts generalizability to the wider dialysis population, in which diabetes is the leading cause of ESKD, and discuss the implications below.
This study was approved by the Bursa City Hospital Clinical Research Ethics Committee (Decision No: 2025-7/9, 9 April 2025) and conducted in accordance with the Declaration of Helsinki. All participants provided written informed consent.
Clinical and neurological assessment
A comprehensive clinical history was obtained from all participants regarding symptoms of peripheral neuropathy. The duration of dialysis treatment was recorded for each patient. Neurological symptoms were systematically evaluated using a structured clinical questionnaire that assessed the presence or absence of neuropathic symptoms, including distal numbness, paresthesias, burning sensations, nocturnal pain, muscle cramps, restless leg symptoms, and functional limitations. Patients reporting one or more symptoms were considered symptomatic for peripheral neuropathy.
A detailed neurological examination was performed by the same neurologist, including assessment of muscle strength, deep tendon reflexes, sensory function, and gait pattern.
Laboratory assessments
Laboratory parameters measured included hemoglobin, potassium, calcium, transferrin saturation, ferritin, CRP, phosphorus, albumin, and total cholesterol. In HD patients, blood samples were collected immediately before the first HD session of the week, after the long interdialytic interval (typically Monday morning), under standardized fasting conditions. In PD patients, fasting morning samples were obtained during a routine outpatient clinic visit. All samples were processed in the central laboratory of Bursa City Hospital using identical assays for both groups.
Dialysis adequacy (Kt/V) calculation
Dialysis adequacy was assessed by Kt/V. In the HD group, single-pool Kt/V was calculated using the second-generation Daugirdas formula, based on pre- and post-dialysis blood urea nitrogen, ultrafiltration volume, dialysis duration, and post-dialysis weight. In the PD group, weekly Kt/V was calculated as the sum of dialysate Kt/V (urea clearance from a 24-h drained dialysate sample) and renal Kt/V (residual urinary urea clearance from a 24-h urine collection), normalized to total body water estimated by Watson’s formula. All PD patients in this cohort were on continuous ambulatory peritoneal dialysis (CAPD); no patient was on automated peritoneal dialysis (APD) during the study window.
Electrophysiological evaluation
Nerve conduction studies were performed using a Natus Nicolet Viking Quest EMG device at the Neurology Department, Bursa City Hospital. The evaluation protocol included bilateral median and ulnar motor and sensory nerve conduction studies from both upper extremities, and peroneal motor, tibial motor, and sural sensory nerve conduction studies from at least one lower extremity. All measurements were conducted with patients in the supine position at room temperature above 32 °C.
Surface silver electrodes were used for compound muscle action potential recording and stainless steel ring electrodes for sensory nerve recording. Stimulation was applied using bipolar surface electrodes. Distal latency was measured as the time interval from stimulus delivery to the initial deflection of the action potential. Amplitude was measured as the voltage difference between baseline and the negative peak. Conduction velocities were calculated for all evaluated nerve segments.
Electrophysiological diagnostic criteria and neuropathy classification
Peripheral neuropathy was diagnosed using standard electrophysiological criteria. Compound muscle action potential (CMAP) and sensory nerve action potential (SNAP) amplitude, latency, and conduction velocity were analyzed. Neuropathy patterns were categorized as follows:
Axonal Neuropathy: Reduced CMAP and SNAP amplitudes with preserved conduction velocities (>75% of lower limit of normal) and distal latencies (<130% of upper limit of normal).
Demyelinating Neuropathy: Preserved amplitudes with slowed conduction velocities (<75% of lower limit of normal) and/or prolonged distal latencies (>130% of upper limit of normal).
Mixed Neuropathy: Coexistence of both axonal and demyelinating features.
Neuropathy patterns were evaluated for length-dependent distal symmetric sensorimotor polyneuropathy, which preferentially affects the longest nerve fibers in a distal-to-proximal distribution.
Frailty assessment
Frailty was assessed using the FFP as originally defined by Fried et al. in the Cardiovascular Health Study cohort [12]. The FFP was selected because it is among the most extensively validated frailty instruments in CKD and dialysis populations [11,13,14], it is based on objective physical measurements rather than clinician gestalt; and it captures the physical phenotype of frailty most closely linked to the muscular and energetic disturbances of uremia. The Clinical Frailty Scale and the Edmonton Frail Scale, which rely, respectively, on subjective clinical judgment and on geriatric inpatient assessment, were considered less suitable for the present comparative dialysis study.
Each of the five FFP criteria was scored 0 (absent) or 1 (present):
Weight Loss: Unintentional weight loss of ≥4.5 kg in the past year or ≥5% of body weight from the previous year.
Exhaustion: Self-reported exhaustion identified by two questions from the CES-D scale, positive if experienced 3–4 days or more per week.
Low Physical Activity: Weekly kilocalorie expenditure in the lowest 20% by gender (men <383 kcal/week, women <270 kcal/week), estimated using the short version of the Minnesota Leisure Time Activity Questionnaire [16] as in the original Cardiovascular Health Study.
Slow Walking Speed: Time to walk 4.5 meters, adjusted for gender and height, in the slowest 20%.
Weakness: Grip strength measured with a Jamar dynamometer (average of three measurements), adjusted for gender and body mass index, in the lowest 20%.
The total FFP score was calculated by summing the five criteria, yielding a score from 0 to 5. Patients were classified using the conventional Fried cutoffs as robust (score 0), pre-frail (score 1–2) or frail (score ≥3).
Sample size calculation
An a priori sample size calculation was performed using G*Power based on the chi-square goodness-of-fit framework, anchored on previously reported effect sizes for the comparison of frailty distributions between dialysis cohorts [17]. Assuming an effect size of w = 0.386, α = 0.05 and statistical power (1 − β) = 0.95 with df = 1, the required total sample size was 88 patients (critical χ2 = 3.84; non-centrality parameter λ = 13.11; actual power = 0.952). The study was planned and conducted to meet this target.
Statistical analysis
The normality of continuous variables was assessed using the Shapiro-Wilk test. According to the results of the normality test, variables that conformed to normal distribution were expressed as mean ± standard deviation, and variables that did not conform to normal distribution were expressed as median (Q1–Q3) values; categorical variables were expressed as n (%). In comparisons between two groups, the independent samples t-test was used in case of conformity to normal distribution and the Mann–Whitney U test in case of non-conformity. Chi-square and Fisher’s exact tests were used for categorical variables. The relationships between frailty score and biochemical parameters were examined using Spearman’s correlation coefficient.
To identify factors independently associated with frailty score, multivariable linear regression was performed using variables with p < 0.25 in univariate analyses. CRP and ferritin were considered for log-transformation given their skewed distributions. Variance inflation factors were inspected to detect multicollinearity (threshold VIF < 5). The model was pre-specified to respect the events-per-variable rule given the modest sample size. The categorical distribution of frailty (robust/pre-frail/frail) between HD and PD was compared by Chi-square test.
For all statistical analyses, SPSS (IBM Corp. Released 2017. IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY: IBM Corp.) was used and the type I error rate was set at 5%.
Results
Patient flow and characteristics
Of 142 patients screened, 54 were excluded for the reasons detailed in Figure 1. The final analytic sample comprised 88 patients: 45 HD and 43 PD. The median age was 57 years (range: 24–75) in the HD group and 52 years (range: 19–77) in the PD group, with no significant difference between groups (p = 0.077). Gender distribution was similar between groups (p = 0.544), with male predominance in both HD (64.4%) and PD (58.1%) groups (Table 1).
Table 1.
Demographic, clinical, and laboratory characteristics of hemodialysis and peritoneal dialysis patients.
| Parameter | Total (n = 88) | HD (n = 45) | PD (n = 43) | p Value |
|---|---|---|---|---|
| Age (years) | 54 (43.25–65.75) | 57 (45.5–69) | 52 (42–61) | 0.077ᵃ |
| Gender – n (%) | ||||
| Male | 54 (61.4%) | 29 (64.4%) | 25 (58.1%) | 0.544ᵇ |
| Female | 34 (38.6%) | 16 (35.6%) | 18 (41.9%) | |
| Hemoglobin (g/dL) | 10.9 (9.73–11.78) | 11.3 (10–11.8) | 10.7 (9.6–11.4) | 0.162ᵃ |
| Potassium (mmol/L) | 4.9 ± 0.7 | 5.1 ± 0.7 | 4.7 ± 0.6 | 0.025ᶜ |
| Transferrin saturation (%) | 26.5 (21–37.91) | 26.3 (22–35.88) | 27 (20–39) | 0.828ᵃ |
| Ferritin (ng/mL) | 542 (264–963.5) | 763 (416.5–1315) | 406.5 (238.75–542.5) | <0.001ᵃ |
| CRP (mg/L) | 5.1 (1.68–18.55) | 8.1 (2–20.5) | 3.2 (1.45–9.85) | 0.075ᵃ |
| Phosphorus (mg/dL) | 5.2 ± 1.3 | 5.3 ± 1.5 | 5.1 ± 1.2 | 0.540ᶜ |
| Kt/V | 1.8 (1.53–1.92) | 1.6 (1.41–1.8) | 1.9 (1.74–2.44) | <0.001ᵃ |
| Dialysis duration (months) | 43 (17–63) | 46 (23.5–72) | 23 (15–59.5) | 0.289ᵃ |
| Albumin (g/L) | 37.6 ± 4.9 | 38.6 ± 4.9 | 36.7 ± 4.7 | 0.073ᶜ |
| Total cholesterol (mg/dL) | 162 (139.5–210) | 151 (129–177.5) | 170 (154.25–218) | 0.005ᵃ |
| Frailty score (FFP) | 1 (0–2) | 1 (0–2) | 1 (0–1) | 0.705ᵃ |
Data expressed as mean ± SD, median (Q1–Q3), or n (%).
Mann–Whitney U test.
Chi-square test.
Independent samples t-test.
Laboratory parameters
Serum potassium levels were significantly higher in HD patients (5.1 ± 0.7 vs 4.7 ± 0.6 mmol/L, p = 0.025). Median ferritin levels were significantly higher in the HD group compared to the PD group (763 vs 406.5 ng/mL, p < 0.001). C-reactive protein (CRP) levels showed a trend toward being higher in HD patients, though this did not reach statistical significance (8.1 vs 3.2 mg/L, p = 0.075). Serum albumin levels showed no significant difference (HD: 38.6 ± 4.9 g/L vs PD: 36.7 ± 4.7 g/L, p = 0.073). Total cholesterol levels were significantly higher in PD patients (170 vs 151 mg/dL, p = 0.005). PD patients had significantly higher Kt/V than HD patients (1.9 vs 1.6, p < 0.001).
Frailty score and categorical distribution
The median FFP score was 1 (Q1–Q3: 0–2) in both HD and PD groups, with no significant difference in the continuous score between dialysis modalities (p = 0.705).
However, when patients were classified using the conventional Fried cutoffs, the distribution of frailty categories differed significantly between HD and PD (p = 0.047, Table 2). In the HD group, 46.7% were robust, 33.3% pre-frail and 20.0% frail. In the PD group, 39.5% were robust, 53.4% pre-frail and 6.9% frail. The prevalence of frailty (FFP ≥ 3) was, therefore, approximately threefold higher in HD than in PD patients (20.0% vs 6.9%), while pre-frailty was more common in PD.
Table 2.
Distribution of frailty categories according to dialysis modality.
| Frailty group | Total (n = 88) | HD | PD | p Value |
|---|---|---|---|---|
| Robust/non-frail (FFP 0) | 38 (43.1%) | 21 (46.7%) | 17 (39.5%) | 0.047a |
| Pre-frail (FFP 1–2) | 38 (43.1%) | 15 (33.3%) | 23 (53.4%) | |
| Frail (FFP ≥ 3) | 12 (13.6%) | 9 (20.0%) | 3 (6.9%) |
Values in bold indicate statistical significance (p < 0.05).
aChi-square test. Conventional Fried cutoffs: robust (score 0), pre-frail (1–2), frail (≥3).
Univariate correlates of frailty
HD Group: In HD patients, frailty score showed a significant positive correlation with CRP levels (rs = 0.31, p = 0.038). No significant correlations were found between frailty score and age, hemoglobin, potassium, ferritin, phosphorus, Kt/V, dialysis duration, albumin, or total cholesterol levels (Table 3).
Table 3.
Correlations between frailty score and biochemical parameters.
| HD (n = 45) | PD (n = 43) | |||
|---|---|---|---|---|
| Parameter | rs | p | rs | p |
| Age | 0.17 | 0.248 | 0.34 | 0.041 |
| Hemoglobin | −0.13 | 0.412 | 0.39 | 0.014 |
| Potassium | 0.11 | 0.482 | −0.25 | 0.136 |
| Transferrin sat (%) | −0.29 | 0.053 | −0.07 | 0.689 |
| Ferritin | 0.10 | 0.508 | 0.02 | 0.899 |
| CRP | 0.31 | 0.038 | 0.04 | 0.828 |
| Phosphorus | 0 | 0.996 | 0.13 | 0.450 |
| Kt/V | 0.12 | 0.451 | −0.37 | 0.027 |
| Dialysis duration | 0.04 | 0.809 | 0.30 | 0.080 |
| Albumin | −0.20 | 0.188 | 0.05 | 0.787 |
| Total cholesterol | −0.21 | 0.151 | 0.19 | 0.266 |
rs = Spearman’s correlation coefficient. Significant correlations (p < 0.05) shown in bold.
PD Group: In PD patients, frailty score demonstrated significant positive correlations with age (rs = 0.34, p = 0.041) and hemoglobin (rs = 0.39, p = 0.014). A significant negative correlation was observed between frailty score and Kt/V (rs = −0.37, p = 0.027). No significant correlations were found with other parameters (Table 3).
Multivariable analysis of frailty predictors
Multivariable linear regression was performed in the combined cohort with frailty score as the dependent variable, including age, peritoneal/HD vintage, CRP, and transferrin saturation as candidate predictors (selected based on univariate p < 0.25 or established clinical relevance). CRP remained an independent correlate of frailty score (B = 0.017, 95% CI: 0.001–0.033, p = 0.042). Age, dialysis duration, and transferrin saturation were not independently associated with frailty (all p > 0.05). The model was statistically significant (F = 2.707, p = 0.038) but explained only a modest fraction of the variance in frailty score (adjusted R2 = 0.090), consistent with the cross-sectional design and modest sample size (Table 4).
Table 4.
Multivariable linear regression analysis for predictors of frailty score.
| Variable | B | 95% CI | p Value |
|---|---|---|---|
| Age | 0.008 | −0.014 to 0.030 | 0.463 |
| Dialysis duration (months) | 0.003 | −0.003 to 0.010 | 0.354 |
| CRP | 0.017 | 0.001 to 0.033 | 0.042 |
| Transferrin saturation (%) | −0.013 | −0.036 to 0.010 | 0.267 |
Model F = 2.707, p = 0.038. Adjusted R2 = 0.090. Significant predictors shown in bold.
Electrophysiological findings
Sensory and motor nerve conduction parameters (latency, amplitude, and conduction velocity) for median, ulnar, peroneal, and tibial nerves showed no significant differences between HD and PD groups (all p > 0.05). Sural nerve conduction velocity showed a trend toward being higher in PD patients (54 vs 48 m/s, p = 0.052). F-wave latencies were comparable between groups (p > 0.05). Detailed nerve conduction study results for all sensory and motor nerves in both dialysis groups are presented in Table 5.
Table 5.
Electromyographic findings in hemodialysis and peritoneal dialysis patients.
| Parameter | HD (n = 45) | PD (n = 43) | p Value |
|---|---|---|---|
| R Median Sensory | |||
| Latency (ms) | 2.7 (2.5–3.1) | 2.7 (2.3–3.2) | 0.809ᵃ |
| Amplitude (μV) | 17 (13.9–28.3) | 21 (11.8–29.1) | 0.570ᵃ |
| Velocity (m/s) | 46 (42–52) | 49 (40–56) | 0.422ᵃ |
| R Ulnar Sensory | |||
| Latency (ms) | 2.5 (2.1–2.7) | 2.2 (2.1–2.6) | 0.573ᵃ |
| Amplitude (μV) | 22.1 (11–28.6) | 18.9 (10–35.2) | 0.537ᵃ |
| Velocity (m/s) | 47.4 ± 8.9 | 48.4 ± 8.4 | 0.625b |
| R Sural Sensory | |||
| Latency (ms) | 2.4 (2–2.9) | 2.6 (2.2–3) | 0.177ᵃ |
| Amplitude (μV) | 13.6 (7.3–18.6) | 12.2 (8.4–19.2) | 0.577ᵃ |
| Velocity (m/s) | 48 (38.5–58.5) | 54 (47–63) | 0.052ᵃ |
| R Median Motor | |||
| Latency (ms) | 3.3 (2.9–3.5) | 3.2 (2.9–3.5) | 0.900ᵃ |
| Amplitude (mV) | 7.9 ± 2.7 | 8.0 ± 2.2 | 0.955b |
| Velocity (m/s) | 52 (47–55) | 52 (48–55) | 0.431ᵃ |
| R Ulnar Motor | |||
| Latency (ms) | 2.7 ± 0.4 | 2.6 ± 0.4 | 0.141b |
| Amplitude (mV) | 7.7 (5.9–8.5) | 8.0 (6.3–9.3) | 0.242ᵃ |
| Velocity (m/s) | 54.9 ± 8.3 | 57.2 ± 6.9 | 0.184b |
| R Peroneal Motor | |||
| Latency (ms) | 3.9 (3.4–4.6) | 3.9 (3.6–4.6) | 0.338ᵃ |
| Amplitude (mV) | 2.1 (1–3.2) | 2.2 (1.5–2.9) | 0.897ᵃ |
| Velocity (m/s) | 41 (36.3–48) | 44 (40–48) | 0.217ᵃ |
| R Tibial Motor | |||
| Latency (ms) | 4.4 (3.8–5.9) | 4.4 (3.8–5.1) | 0.580ᵃ |
| Amplitude (mV) | 5.0 ± 2.9 | 5.7 ± 3.0 | 0.323b |
| Velocity (m/s) | 40 (34–42) | 41 (35.5–43) | 0.289ᵃ |
| R Ulnar F-wave latency (ms) | 28.9 (26.2–30.9) | 27.8 (25.7–30.7) | 0.513ᵃ |
| R Tibial F-wave latency (ms) | 41.5 ± 14.6 | 42.4 ± 11.2 | 0.809b |
Data expressed as mean ± SD or median (Q1–Q3). ᵃMann–Whitney U test.
bIndependent samples t-test. R = right side.
Axonal neuropathy was the most common pattern (37.5%), followed by mixed axonal-demyelinating (23.9%), with demyelinating neuropathy being rare (4.5%). Normal nerve conduction studies were found in 33.3% of patients. Sensorimotor polyneuropathy was the predominant distribution pattern (55.7%), with pure sensory (8.0%) and pure motor (2.3%) neuropathy being less common. There were no significant differences in the prevalence or patterns of neuropathy between HD and PD patients (all p > 0.05) (Table 6).
Table 6.
Distribution of neuropathy patterns and types.
| Classification | Total (n = 88) | HD (n = 45) | PD (n = 43) | p Value |
|---|---|---|---|---|
| Neuropathy Pattern | ||||
| Axonal | 33 (37.5%) | 17 (37.8%) | 16 (37.2%) | 0.956a |
| Demyelinating | 4 (4.5%) | 3 (6.7%) | 1 (2.3%) | 0.617b |
| Mixed | 21 (23.9%) | 11 (24.4%) | 10 (23.3%) | 0.896a |
| Normal | 29 (33.3%) | 14 (31.1%) | 15 (35.7%) | 0.649a |
| Distribution | ||||
| Pure motor | 2 (2.3%) | 2 (4.4%) | 0 | 0.495b |
| Pure sensory | 7 (8.0%) | 2 (4.4%) | 5 (11.6%) | 0.261b |
| Sensorimotor | 49 (55.7%) | 27 (60.0%) | 22 (51.2%) | 0.404a |
| Normal | 28 (32.2%) | 14 (31.1%) | 14 (33.3%) | 0.825a |
Data expressed as n (%).
aChi-square test.
bFisher’s exact test.
Among the 12 (26.6%) HD patients and 14 (32.5%) PD patients who were clinically asymptomatic on history and examination, electrophysiological examination revealed abnormal findings in 9 (75%) and 11 (78.6%), respectively, indicating a high prevalence of subclinical neuropathy in both modalities.
Discussion
This single-center, cross-sectional study compared electrophysiological findings and frailty between HD and PD patients in a non-diabetic ESKD cohort. The main observations were: (i) the overall prevalence of electrophysiological abnormalities was high and broadly similar between modalities, with axonal polyneuropathy predominating; (ii) median FFP scores did not differ between HD and PD, but the categorical distribution of frailty did, with a markedly higher prevalence of frank frailty (FFP ≥ 3) in HD; (iii) the variables univariately associated with frailty differed between modalities (CRP in HD; age, hemoglobin, and Kt/V in PD); and (iv) in the combined multivariable model, CRP was the only independent correlate of frailty score, although the model’s explanatory power was modest.
Prevalence of subclinical neuropathy
Among clinically asymptomatic patients in our cohort, electrophysiological abnormalities were detected in 75% of HD and 78.5% of PD patients. These findings are consistent with previous studies. Tilki et al. [9] reported electrophysiological evidence of polyneuropathy in all 17 of their clinically silent dialysis patients, and Ezzeldin et al. [18] reported abnormal electrophysiology in 92.5% of clinically asymptomatic HD patients. Jovanovic et al. [19] found pathological neurophysiological parameters in 97% of CAPD patients, with subclinical neuropathy present in 35% of diabetic and 47% of non-diabetic patients. Collectively, these data support that subclinical neuropathy is common in both HD and PD patients and that electrophysiological evaluation is more sensitive than clinical examination. Whether subclinical electrophysiological abnormalities translate into clinically meaningful outcomes – such as falls, functional decline, or frailty progression – remains an open question that our cross-sectional design cannot resolve; this is an important direction for prospective longitudinal work.
Electrophysiological findings and modality comparison
Although the difference did not reach statistical significance, sural sensory nerve conduction velocity trended higher in PD patients (54 vs 48 m/s, p = 0.052). This is of interest because the sural nerve is one of the most sensitive early indicators of peripheral neuropathy [5,20]. Arnold et al. [8] previously demonstrated that PD patients had nerve excitability parameters similar to stage 4 CKD patients and significantly more normal patterns than HD patients, with pronounced pre- and post-dialysis fluctuations in HD but remarkable stability in PD, suggesting superior homeostatic control with continuous dialysis. Janda et al. [21] reported higher ulnar sensory amplitudes and shorter sural distal latencies in CAPD compared with HD patients despite similar overall polyneuropathy prevalence, and Tilki et al. [9] also found shorter ulnar sensory distal latency in CAPD patients. In our study, the trend toward better sural conduction velocity in PD patients alongside significantly lower ferritin, a trend toward lower CRP, and better potassium control is compatible with the hypothesis that the continuous nature of PD provides a more stable metabolic environment that may help preserve peripheral nerve function. However, our cross-sectional, single-center design cannot establish such a causal link.
Axonal polyneuropathy was the predominant pattern (37.5%) in both modalities, with similar prevalence (HD 37.8% vs PD 37.2%, p = 0.956). This pattern, observed in a non-diabetic cohort, is consistent with uremic toxins themselves being sufficient to cause significant axonal degeneration independent of diabetic complications. Although HD patients had significantly higher potassium and ferritin levels, the prevalence of neuropathy was similar between groups, suggesting that uremic neuropathy is multifactorial with contributions from uremic toxins, chronic inflammation, electrolyte imbalances, and other metabolic disturbances, with no single factor being solely determinant [22].
Frailty: Comparison of categorical distribution between HD and PD
Although median FFP scores were identical in HD and PD (median 1 in both groups, p = 0.705), the categorical distribution of frailty differed significantly (p = 0.047): the prevalence of frank frailty (FFP ≥ 3) was 20.0% in HD compared with 6.9% in PD, while pre-frailty was more common in PD (53.4% vs 33.3%). The discrepancy between the continuous score comparison and the categorical distribution is informative: the median is insensitive to the extremes of the distribution, and the higher proportion of HD patients in the upper tail of FFP scores is what drives the difference in categorical prevalence. This pattern is biologically plausible given the higher inflammatory burden and intermittent metabolic stress associated with HD, but it should be interpreted cautiously in a single-center, cross-sectional sample with modest size.
Modality-specific correlates of frailty: a cautious interpretation
In our cohort, the variables univariately associated with frailty differed between dialysis modalities: CRP in HD and age, hemoglobin and Kt/V in PD. In multivariable analysis of the combined cohort, CRP remained the only independent correlate of frailty score, although the model explained only 9.0% of the variance in frailty. These associations should be regarded as hypothesis-generating rather than confirmatory of distinct mechanistic pathways. Several considerations support a cautious interpretation. First, the study is cross-sectional and cannot establish temporal or causal relationships. Second, the sample size limits the power of stratified models. Third, frailty in our cohort was largely mild (median FFP score 1 in both groups), restricting the dynamic range of the outcome and the strength of associations that can be detected. Fourth, we did not perform formal direct correlation analyses between electrophysiological parameters and frailty scores in the combined cohort, an analysis, which would have helped clarify whether neuropathy contributes to frailty; this is a limitation acknowledged below.
With these caveats, the correlation of CRP with frailty in HD is consistent with the well-described association between systemic inflammation and physical frailty and with previous work showing higher inflammatory burden in HD [23]. Hendra et al. [24] similarly reported significantly higher CRP in frail HD patients (median 8.2 vs 3.0 mg/L). The intermittent nature of HD may contribute to a chronic inflammatory state through hemodynamic stress, rapid solute shifts, and complement activation from blood-membrane contact. The trend toward higher CRP and significantly higher ferritin in HD in our cohort is compatible with this mechanism. However, our analysis is cross-sectional, the correlation is modest in magnitude, and the multivariable model explained only a small fraction of the variance; the observation is therefore best framed as supportive of, but not proof of, an inflammation–frailty link in this population.
In PD patients, the positive association of age with frailty is consistent with frailty being an aging-related syndrome [13,25]; Chen et al. [25] similarly reported that frail PD patients were significantly older. The negative correlation between Kt/V and frailty may reflect the protective influence of adequate solute clearance, although reverse causation cannot be excluded – frail patients may have more constrained dialysis prescriptions or smaller body water volumes that mathematically inflate Kt/V. The positive correlation of hemoglobin with frailty was unexpected and contrasts with the literature reporting a positive association between anemia and frailty [26]. It may reflect the wider hemoglobin range in our PD group (8.3–16 g/dL), residual confounding by volume status, or chance given the small sample.
Our findings are broadly concordant with the comparative HD–PD literature in end-stage renal disease (ESRD). In a retrospective cohort of 151 ESRD patients, Zhao [27] reported significantly higher post-dialysis Kt/V (median 2.11 vs 1.73, p < 0.001) and better hemodynamic and metabolic control in PD compared with HD, consistent with our observations of higher Kt/V (1.9 vs 1.6, p < 0.001) and a more favorable inflammatory and electrolyte profile in PD.
Our group has previously reported on frailty and peripheral neuropathy in HD patients (Şeker et al.) [28]. The present study extends that work along a different and complementary axis: rather than describing the clinical and electrophysiological correlates of frailty within HD alone, the current analysis adds a PD comparator group and explicitly compares modality-specific patterns of electrophysiological findings, frailty, and biochemical predictors. The novel contributions of the present manuscript are therefore (i) the head-to-head HD vs PD comparison of nerve conduction parameters in a non-diabetic cohort, (ii) the modality-stratified examination of frailty correlates, and (iii) the description of the categorical frailty distribution showing a higher prevalence of frank frailty in HD.
Clinical implications
Our findings, although hypothesis-generating, support the integration of frailty assessment into routine dialysis care and suggest that the prevalence of frank frailty may be substantially higher in HD than in PD. In that case, regular monitoring of inflammatory markers and strategies to reduce inflammatory burden in HD patients and optimization of dialysis adequacy together with attention to age-related vulnerabilities in PD patients would represent reasonable avenues for prospective evaluation. In both modalities, multimodal interventions – exercise, nutritional support, and anemia management – warrant prospective investigation. Chen et al. demonstrated that frailty status can remain stable or improve within 6 months in young and incident PD patients [25], emphasizing the value of early intervention.
Strengths and limitations
Strengths of our study include the prospective design, standardized electrophysiological assessment by a single neurologist, comprehensive biochemical evaluation, and the inclusion of both HD and PD groups in a single center, which reduces inter-laboratory and inter-operator variability.
Several limitations must be considered. First, the cross-sectional design precludes causal inference and any conclusions about temporal trajectories of frailty or neuropathy. Second, the sample size, although calculated a priori to be adequate for the primary comparison, restricts the statistical power of multivariable and stratified analyses; effect estimates should be interpreted as hypothesis-generating. Third, recruitment was restricted to a single tertiary center and to outpatient ambulatory patients, and the exclusion of patients with diabetes mellitus narrows the spectrum of frailty observed and limits generalizability to the wider dialysis population. The low overall frailty prevalence we observed (median FFP score 1 in both modalities) is consistent with these selection effects. Fourth, we did not perform a formal direct correlation analysis between electrophysiological findings and frailty score in the combined cohort, nor did we directly compare clinically asymptomatic patients with positive electrophysiological findings against those with both clinical and electrophysiological evidence of neuropathy. These analyses would have addressed the clinical relevance of subclinical electrophysiological abnormalities and should be prioritized in future, adequately powered studies. Fifth, frailty was assessed by a single instrument (FFP); convergent validation against other tools such as the Clinical Frailty Scale or the Edmonton Frail Scale was not performed. Finally, conventional nerve conduction studies, while standard, are less sensitive than nerve excitability testing for detecting subtle uremic effects on axonal function, and the latter was not available in our center.
Conclusion
In this single-center cross-sectional study of non-diabetic adults receiving HD or PD, the prevalence and severity of peripheral neuropathy were broadly similar between dialysis modalities, with axonal polyneuropathy as the predominant electrophysiological pattern. Although median FFP scores did not differ between groups, the categorical distribution of frailty did, with a higher prevalence of frank frailty in HD patients. The variables associated with frailty differed by modality in univariate analysis – CRP in HD and age, hemoglobin, and Kt/V in PD – and CRP remained an independent correlate of frailty score in the combined multivariable model, although the model explained only a modest fraction of the variance. These findings are hypothesis-generating and should be confirmed in larger, multicenter, longitudinal studies that examine the direct relationship between electrophysiological abnormalities and frailty trajectories. If confirmed, modality-specific interventions targeting inflammation in HD and dialysis adequacy and age-related vulnerabilities in PD may help prevent the progression of frailty in dialysis patients.
Acknowledgments
We thank all personnel of the Neurology and Nephrology clinics for their contributions to conducting this study. Mehmet Usta and Ayşe Şeker performed study planning, methodology, validation, formal analysis, investigation, data curation, and writing – original draft. Can Hüzmeli, Sinan Gönüllü, and Hatice Ortaç performed validation, formal analysis, writing – review and editing, and visualization and prepared the manuscript. All authors reviewed the manuscript.
Disclosure statement
The authors certify that there is no conflict of interest with any financial organization regarding the material discussed in the manuscript.
Data availability statement
The data that support the findings of this study are available from the corresponding author, Mehmet Usta, upon reasonable request.
References
- 1.Seyahi N, Koçyiğit İ, Eren N, et al. Current status of kidney replacement therapy in Türkiye: a summary of 2023 Turkish society of nephrology registry report. TJN. 2025;34(2):141–148. doi: 10.5152/turkjnephrol.2025.24978. [DOI] [Google Scholar]
- 2.Mehrotra R, Devuyst O, Davies SJ, et al. The current state of peritoneal dialysis. J Am Soc Nephrol. 2016;27(11):3238–3252. doi: 10.1681/ASN.2016010112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Arnold R, Issar T, Krishnan AV, et al. Neurological complications in chronic kidney disease. JRSM Cardiovasc Dis. 2016;5:2048004016677687. doi: 10.1177/2048004016677687. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Krishnan AV, Kiernan MC.. Neurological complications of chronic kidney disease. Nat Rev Neurol. 2009;5(10):542–551. doi: 10.1038/nrneurol.2009.138. [DOI] [PubMed] [Google Scholar]
- 5.Krishnan AV, Kiernan MC.. Uremic neuropathy: clinical features and new pathophysiological insights. Muscle Nerve. 2007;35(3):273–290. doi: 10.1002/mus.20713. [DOI] [PubMed] [Google Scholar]
- 6.Moorthi RN, Doshi S, Fried LF, et al. Chronic kidney disease and peripheral nerve function in the health, aging and body composition study. Nephrol Dial Transplant. 2019;34(4):625–632. doi: 10.1093/ndt/gfy102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Jasti DB, Mallipeddi S, Apparao A, et al. A clinical and electrophysiological study of peripheral neuropathies in predialysis chronic kidney disease patients and relation of severity of peripheral neuropathy with degree of renal failure. J Neurosci Rural Pract. 2017;8(4):516–524. doi: 10.4103/jnrp.jnrp_186_17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Arnold R, Pussell BA, Kiernan MC, et al. Comparative study to evaluate the effects of peritoneal and hemodialysis on peripheral nerve function. Muscle Nerve. 2016;54(1):58–64. doi: 10.1002/mus.25016. [DOI] [PubMed] [Google Scholar]
- 9.Tilki HE, Akpolat T, Coşkun M, et al. Clinical and electrophysiologic findings in dialysis patients. J Electromyogr Kinesiol. 2009;19(3):500–508. doi: 10.1016/j.jelekin.2007.10.011. [DOI] [PubMed] [Google Scholar]
- 10.Wanic-Kossowska M, Koczocik-Przedpelska J.. Myoneuropathy in patients with chronic renal failure treated with hemodialysis (HD) and intermittent peritoneal dialysis (IPD). II. Sensory and motor nerve conduction velocity in patients with chronic renal failure treated with intermittent peritoneal dialysis and hemodialysis. Pol Arch Med Wewn. 1996;95(3):237–244. [PubMed] [Google Scholar]
- 11.Clegg A, Young J, Iliffe S, et al. Frailty in elderly people. Lancet. 2013;381(9868):752–762. doi: 10.1016/S0140-6736(12)62167-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Fried LP, Tangen CM, Walston J, Cardiovascular Health Study Collaborative Research Group, et al. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56(3):M146–M156. doi: 10.1093/gerona/56.3.m146. [DOI] [PubMed] [Google Scholar]
- 13.Zhang F, Wang H, Bai Y, et al. Prevalence of physical frailty and impact on survival in patients with chronic kidney disease: a systematic review and meta-analysis. BMC Nephrol. 2023;24(1):258. doi: 10.1186/s12882-023-03303-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Chowdhury R, Peel NM, Krosch M, et al. Frailty and chronic kidney disease: a systematic review. Arch Gerontol Geriatr. 2017;68:135–142. doi: 10.1016/j.archger.2016.10.007. [DOI] [PubMed] [Google Scholar]
- 15.Karnabi P, Massicotte-Azarniouch D, Ritchie LJ, et al. Physical frailty and functional status in patients with advanced chronic kidney disease: a systematic review. Can J Kidney Health Dis. 2023;10:20543581231181026. doi: 10.1177/20543581231181026. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Taylor HL, Jacobs DR, Jr, Schucker B, et al. A questionnaire for the assessment of leisure time physical activities. J Chronic Dis. 1978;31(12):741–755. doi: 10.1016/0021-9681(78)90058-9. [DOI] [PubMed] [Google Scholar]
- 17.Hussien H, Siriteanu L, Nistor I, et al. The impact of frailty and severe cognitive impairment on survival time and time to initiate dialysis in older adults with advanced chronic kidney disease: a prospective observational cohort study. Cureus. 2024;16(7):e64303. doi: 10.7759/cureus.64303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ezzeldin N, Abdel Galil SM, Said D, et al. Polyneuropathy associated with chronic hemodialysis: clinical and electrophysiological study. Int J Rheum Dis. 2019;22(5):826–833. doi: 10.1111/1756-185X.13462. [DOI] [PubMed] [Google Scholar]
- 19.Jovanovic DB, Matanovic DD, Simic-Ogrizovic SP, et al. Polyneuropathy in diabetic and nondiabetic patients on CAPD: is there an association with HRQOL? Perit Dial Int. 2009;29(1):102–107. doi: 10.1177/089686080902900114. [DOI] [PubMed] [Google Scholar]
- 20.Laaksonen S, Metsärinne K, Voipio-Pulkki LM, et al. Neurophysiologic parameters and symptoms in chronic renal failure. Muscle Nerve. 2002;25(6):884–890. doi: 10.1002/mus.10159. [DOI] [PubMed] [Google Scholar]
- 21.Janda K, Stompór T, Gryz E, et al. Evaluation of polyneuropathy severity in chronic renal failure patients on continuous ambulatory peritoneal dialysis or on maintenance hemodialysis. Przegl Lek. 2007;64(6):423–430. [PubMed] [Google Scholar]
- 22.Issar T, Arnold R, Kwai NCG, et al. Relative contributions of diabetes and chronic kidney disease to neuropathy development in diabetic nephropathy patients. Clin Neurophysiol. 2019;130(11):2088–2095. doi: 10.1016/j.clinph.2019.08.005. [DOI] [PubMed] [Google Scholar]
- 23.Cesari M, Prince M, Thiyagarajan JA, et al. Frailty: an emerging public health priority. J Am Med Dir Assoc. 2016;17(3):188–192. doi: 10.1016/j.jamda.2015.12.016. [DOI] [PubMed] [Google Scholar]
- 24.Hendra H, Sridharan S, Farrington K, et al. Characteristics of frailty in haemodialysis patients. Gerontol Geriatr Med. 2022;8:23337214221098889. doi: 10.1177/23337214221098889. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Chen YT, Lai TS, Tsao HM, et al. Clinical implications of frailty in peritoneal dialysis patients - A prospective observational study. J Formos Med Assoc. 2024;123(2):248–256. doi: 10.1016/j.jfma.2023.07.005. [DOI] [PubMed] [Google Scholar]
- 26.Roshanravan B, Khatri M, Robinson-Cohen C, et al. A prospective study of frailty in nephrology-referred patients with CKD. Am J Kidney Dis. 2012;60(6):912–921. doi: 10.1053/j.ajkd.2012.05.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Zhao Y. Comparison of the effect of hemodialysis and peritoneal dialysis in the treatment of end-stage renal disease. Pak J Med Sci. 2023;39(6):1562–1567. doi: 10.12669/pjms.39.6.8056. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Şeker A, Usta M, Gönüllü S, et al. Frailty and peripheral neuropathy in hemodialysis patients: clinical and electrophysiological correlations. Ren Fail. 2025;47(1):2547305. doi: 10.1080/0886022X.2025.2547305. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author, Mehmet Usta, upon reasonable request.