Polyneuropathy in kidney transplant recipients: a cross-sectional study in Groningen, the Netherlands

Abstract

Objectives

To determine the prevalence and clinical characteristics associated with polyneuropathy in kidney transplant recipients (KTRs).

Design

Cross-sectional study.

Setting

SENS study at the University Medical Center Groningen, the Netherlands, December 2021–May 2023.

Participants

KTR, participating in the ongoing TransplantLines Biobank and Cohort Study, ≥12 months post-transplantation.

Main outcome measures

Participants underwent a structured neurological assessment including history taking, neurological examination, quantitative sensory testing and nerve conduction studies. An expert panel classified participants into no/possible, probable/definite large fibre polyneuropathy or small fibre neuropathy. Large-fibre subtypes included axonal or demyelinating, pure sensory, pure motor and sensorimotor. To assess potential associations with clinical characteristics, logistic regression analysis was conducted.

Results

We included 160 KTRs with a mean age of 59.8±11.6 years at a median of 6.1 (95% CI 3.9 to 13.1) years post-transplantation, with 16 KTRs (10%) diagnosed with polyneuropathy before study inclusion. In total, 84 KTRs (53%) were identified with large fibre polyneuropathy and 7 KTRs (4%) with small fibre neuropathy. KTRs with large fibre polyneuropathy presented with either sensor-predominant polyneuropathy (40 KTR (48%)) or sensorimotor polyneuropathy (44 KTR (52%)). We found no neurophysiological characteristics of demyelination. Overall, 18% (95% CI 11% to 27%) of KTRs with large fibre polyneuropathy were asymptomatic. Higher age (OR=1.04 (1.01 to 1.08), p=0.01), male sex (OR=2.55 (1.19 to 5.60), p=0.02), diabetes (OR=5.58 (1.36 to 38.14), p=0.03) and elevated urea levels (OR=1.12 (1.04 to 1.23), p=0.01) were significantly associated with polyneuropathy in KTR.

Conclusions

In contrast with previous studies, axonal sensory or sensorimotor polyneuropathy is highly prevalent and often underdiagnosed in KTR. Next to higher age and male sex, it was independently associated with diabetes and higher urea levels. Further research is needed to reveal the aetiology and course of polyneuropathy in KTRs.

Trial registration number

NCT04664426.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The cohort consisted of a large representative sample of kidney transplant recipients (KTRs) who underwent extensive neurophysiological examination.

• An expert panel assigned all KTRs to a diagnostic class of polyneuropathy using a staged approach with an individual assessment followed by a plenary panel, prioritising specificity to minimise the number of false positives.

• The cross-sectional study design, chosen as a necessary starting point due to the lack of prior research on polyneuropathy in KTRs, allowed the identification of associations between polyneuropathy and clinical characteristics but limited conclusions on the underlying aetiology.

Introduction

Kidney transplantation is considered the best treatment for patients with end-stage kidney disease, improving survival significantly compared with dialysis or conservative treatment.¹ Medical outcomes of kidney transplantation have improved in the last decades, mainly because of innovations in immunosuppressant treatments.² A consequence of longer survival following kidney transplantation is the increased occurrence of complications and medical disorders resulting from kidney disease, the transplantation itself or prolonged exposure to immunosuppressive drugs, all of which can negatively impact quality of life.³ Compared with the general population, kidney transplant recipients (KTR) experience a reduced health-related quality of life.⁴ Clinically, signs and symptoms of polyneuropathy are perceived as very common in KTR, with polyneuropathy known to significantly impact quality of life in other patient populations.^{5 6} However, a systematic review and meta-analysis challenges this perception, reporting that neurological disorders affect only 8% of KTR, with a subgroup analysis showing that peripheral neuropathy was the most common neurological disorder, observed in only 2.4%.⁷ This discrepancy between clinical perception and empirical evidence questions the actual prevalence of polyneuropathy in KTR.

Although unknown, the underlying cause of polyneuropathy in KTRs is assumed to be multifactorial⁸; patients are experiencing a large variety of accompanying comorbidities next to the primary disease which initially led to kidney transplantation. Hence, symptoms of polyneuropathy are often attributed to the uraemic state pre-transplantation, the presence of diabetes mellitus or the treatment with calcineurin inhibitors tacrolimus and cyclosporine.9,12 Typically, distal nerve fibres are affected first and most prominently, gradually progressing proximally throughout the disease course.¹³ The diagnosis of polyneuropathy is not straightforward, and polyneuropathy can be categorised according to different principles. First, different nerve fibres can be affected; patients can present with sensory predominant, motor predominant or sensorimotor polyneuropathy.¹⁴ Second, axonal pathophysiology can be differentiated from demyelinating pathophysiology, which can guide the diagnostic evaluation and potential therapeutic intervention.¹⁵ Third, different nerve fibre types can be affected, providing information on the symptoms a patient will experience; small unmyelinated C fibres and lightly myelinated Aδ fibres transmit noxious and thermal information, whereas larger myelinated Aβ fibres are responsible for proprioceptive and vibratory information.¹³ The wide array of symptoms and signs include pain, tingling and altered sensation in a distal symmetric distribution.¹⁰ Especially in case of demyelination, symptoms can progress to motor complaints and muscle weakness.¹⁶

Thus far, the prevalence, type of polyneuropathy and underlying aetiology in this specific patient group remain poorly understood. In the current study, we therefore comprehensively investigated the prevalence, classification and subtypes of polyneuropathy in KTRs by employing thorough history taking and additional neurological examination including extensive nerve conduction studies (NCSs), which has often been lacking in previous studies. Furthermore, we explored the relationship of different clinical characteristics in the pretransplant and post-transplant period with polyneuropathy.

Methods

Study population

The present study was part of the SENS (Sensory Neuropathy Scores) study at the University Medical Center Groningen (UMCG), the Netherlands (ClinicalTrials.gov identifier: NCT04664426). KTRs at least 12 months post-transplantation were invited to participate following simple random sampling after participating in the TransplantLines Biobank and Cohort Study at the UMCG.¹⁷ Data were collected between December 2021 and May 2023. KTRs were eligible to participate if they were at least 18 years of age, able to understand the Dutch language and capable of following the instructions during the neurological testing. Reasons for exclusion from participation were bilateral amputation of upper or lower limbs, bilateral limb injury (including an arteriovenous dialysis shunt), bilateral metal implants in the limbs, pacemaker or implantable cardioverter defibrillator, previous diagnosis with a mononeuropathy including carpal tunnel syndrome, pregnancy, restart of dialysis due to transplant failure, previous treatment with chemotherapy or use of mind-altering drugs. For included patients, demographic and clinical data were extracted from medical records. Clinical examinations were performed during a single study visit at the outpatient clinic of the UMCG. All deceased donor kidneys were allocated by Eurotransplant in accordance with the legislation of the member states, European Union regulations and Eurotransplant’s ethical guidelines. Living donor procedures were performed in compliance with the Dutch Transplantation Law. All living donors were residents of the Netherlands and provided written informed consent after receiving adequate information on the risks of the procedure and sufficient time to consider their decision. Donors were explicitly informed of their right to withdraw consent at any time prior to donation. All donor candidates are screened for potential coercion or financial inducement, and procedures were not carried out in cases of doubt. Donor anonymity and confidentiality were strictly maintained, where appropriate. The manuscript was prepared following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.¹⁸

The SENS study was designed to address two primary research questions. The first, reported in the present manuscript, aimed to investigate the prevalence, classification and subtypes of polyneuropathy in KTR, as well as to explore associations with clinical characteristics. The second research question focused on validating a screening tool for diagnosing polyneuropathy in KTRs and has been reported separately.¹⁹ Both research questions were addressed using data from the SENS study cohort and followed the same methodological approach for classifying polyneuropathy.

Patient and public involvement

This study was initiated in response to KTRs frequently reporting symptoms of polyneuropathy during outpatient follow-up appointments after transplantation, despite the absence of a formal diagnosis of polyneuropathy. Unpublished data from a cohort of 650 KTRs in the TransplantLines Biobank and Cohort Study at the UMCG revealed that 36% of KTRs reported experiencing tingling sensations in hands or feet, as assessed by the MTSOSD-59R questionnaire.^{17 20} At the commencement of the study, we conducted preliminary visits with four participants, who were considered patient representatives, to collect feedback on the sequence of tests, feasibility and the information provided beforehand. Dissemination to the public will include conference presentations, press releases and plain language summaries in patient-focused journals.

Neurological and neurophysiological assessment

The first part of the neurological assessment involved taking a structured history focusing on the presence of common complaints of polyneuropathy: bilateral pain in the feet, allodynia of the feet, tingling or numbness in the feet, weakness, ataxia and similar symptoms in the hands.

Subsequently, a standardised neurological examination was performed, which consisted of assessment of sensory perception in the feet at the level of toes, insteps and ankles, quantified by testing pinprick and light touch perception with a neuropen (Neuropen, Owen Mumford, Oxfordshire, UK). Pinprick was evaluated by testing the ability to discriminate the sharp end of a neurotip from the blunt end. To assess light touch perception, participants were asked if they could feel a bent 10 g monofilament on their skin. Position sense was tested by passively flexing or extending the big toe. Vibration perception was tested using a hand-held biothesiometer (Bio-thesiometer, Bio Medical Instrument, Ohio, USA) with an applicator button that vibrates at a frequency of 120 hertz and a vibration amplitude ranging from 0 to 25.5 microns of motion. Normal values were derived from a previous study conducted on healthy subjects undergoing screening for potential kidney donation.²¹ Additionally, proximal and distal muscular strength of the arms and legs was evaluated. The strength of the foot dorsiflexors, hip flexors, elbow flexors and three-point grip was tested with a hand-held dynamometer (CIT Technics, Haren, the Netherlands). Lastly, Achilles and patellar tendon reflexes were tested bilaterally and scored as normal, reduced or absent. In case of discrepancies between the two sides, the better side was taken as the reference. History taking and neurological examination were carried out by a well-trained researcher.

To assess small fibre function, quantitative sensory testing (QST) was performed unilaterally. The QST protocol included warm and cold temperature threshold testing according to the method of levels which was performed using a TSA-II-NeuroSensory Analyzer (Medoc, Ramat Yishay, Israel).²² Testing was initially conducted on the dorsum of the foot (S1 dermatome) and subsequently on the thenar eminence (C6 dermatome) using a thermode (surface 30×30 mm) attached to the skin with elastic Velcro tape. The thermode baseline temperature was set at 32°C. Prior to testing, skin temperature was verified to be above 31°C. The obtained thresholds were compared with published normative values.²³ The QST was deemed abnormal if at least one extremity (ie, foot or hand) was abnormal.

To identify large fibre function deficits, NCSs were conducted using a Synergy EDX (Cephalon, Nørresundby, Denmark) with surface stimulation and recording, adhering to standard techniques. Sensory nerve conduction was assessed unilaterally on the same side of the body as the assessment of QST of the ulnar and sural nerves, and the assessment of motor nerve conduction of the ulnar, peroneal and tibial nerves. Additionally, the soleus H-reflex was recorded. During the examinations, the skin temperature of hands and feet was closely monitored to ensure it remained above 31°C. Sensory nerve action potential (SNAP) or compound muscle action potential (CMAP) amplitudes, nerve conduction velocity and distal motor latency for CMAPs were determined. SNAP amplitudes were measured from the first negative to the next positive peak, and CMAP amplitudes were measured from baseline to the first negative peak. An assessment of a nerve was classified as abnormal if at least one of these parameters was abnormal.^{24 25} For peripheral motor and sensory NCSs, normal values of the UMCG clinical neurophysiology department were used. For the H-reflex, normal values for the latency of Schimsheimer et al were applied.²⁶ Neuropathy was classified as demyelinating if nerve conduction velocities were less than 35 m/s in the arms and less than 30 m/s in the legs.²⁷

QST and NCSs were carried out by neurophysiology technologists, under the supervision of a neurologist or clinical neurophysiologist.

The standardised operating procedure of the complete neurologic and neurophysiological assessment used for this study can be found in the appendix (see online supplemental appendix A).

Primary outcome

All participants were assigned to a diagnostic class of polyneuropathy using a staged approach.²⁸ An expert panel consisting of a neuromuscular specialist and clinical neurophysiologist (GD), a neurophysiology specialist (FL) and a neurologist (HRM) initially classified all patients through an individual assessment. Any discrepancies in the diagnostic classifications were subsequently discussed in a plenary session, where a final diagnosis was assigned to each patient. The classification of polyneuropathy by experts was based on a five-category system: (1) no polyneuropathy, (2) possible polyneuropathy, (3) probable large fibre polyneuropathy, (4) definite large fibre polyneuropathy and (5) small fibre neuropathy. Subsequently, the classification was transformed, whereby the categories of ‘no’ and ‘possible’ polyneuropathy were defined as ‘no polyneuropathy’, the categories of ‘probable’ and ‘definite’ large fibre polyneuropathy were defined as ‘polyneuropathy’, and small fibre neuropathy was handled as a distinct category. The detailed protocol for the classification can be found in online supplemental appendix B.

To identify sensory and motor abnormalities in KTRs affected by large fibre polyneuropathy (ie, ‘probable’ and ‘definite’ large fibre polyneuropathy), NCS results of the sural, tibial and peroneal nerve were assessed. Sensor-predominant polyneuropathy was defined as KTRs presenting with solely abnormal sural nerve amplitudes, while motor-predominant polyneuropathy was defined as KTRs with normal sural nerve amplitudes presenting with abnormal tibial nerve amplitudes, with peroneal nerve abnormalities only considered relevant if tibial nerve amplitudes were also abnormal. KTRs with sensorimotor polyneuropathy exhibited both sensory and motor abnormalities, characterised by a combination of criteria for sensor-predominant and motor-predominant polyneuropathy.

Additionally, to determine the percentage of KTRs with polyneuropathy before study inclusion, electronic medical records were reviewed. Diagnoses confirmed by a neurologist or documented in the participant’s medical history as having been diagnosed were considered.

Clinical characteristics

To assess the association of polyneuropathy in KTRs with clinical characteristics at the time of study inclusion and the peritransplant period, electronic medical records were used. We retrieved information on the primary disease before kidney transplantation, dialysis prior to transplantation, the date of transplantation, the type of transplantation (ie, kidney or combined pancreas and kidney transplantation), donor type (ie, living or deceased), smoking status, alcohol intake, information on the immunosuppressive regimen, and presence of diabetes defined as fasting plasma glucose higher or equal to 7.0 mmol/L, haemoglobin A1c (HbA1c) above or equal to 48 mmol/L or the use of antidiabetic medication.²⁹ Additionally, laboratory values from the most recent outpatient follow-up at the time of study inclusion were collected, including urea, estimated glomerular filtration rate (eGFR), creatinine, creatinine clearance, albumin, cystatin C and urinary protein excretion.

Statistical analyses

Continuous variables with a normal distribution were described as mean±SD or as median (IQR) in case of a skewed distribution. Histograms and Q/Q plots were used to assess the distribution of continuous variables. Categorical variables were displayed as frequencies with percentages. To evaluate the representativeness of the SENS cohort of KTR, baseline characteristics were compared with those of a KTR cohort from the TransplantLines Biobank and Cohort Study, using unpaired t-tests for normally distributed data, Mann-Whitney U test for skewed data and Pearson χ² tests for categorical data.¹⁷

Prevalence of the different polyneuropathy categories and impaired fibre type were given as percentages with corresponding 95% CIs. Clinical characteristics were compared between KTRs with and without polyneuropathy using unpaired t-tests for normally distributed data, Mann-Whitney U test for skewed data and Pearson χ² tests for categorical data.

To identify possible associations between clinical characteristics at time of study inclusion and the pretransplant and post-transplant period with the dichotomous outcome polyneuropathy, logistic regression analyses were performed. First, potential clinical associations were tested in univariable logistic regression analyses. Second, multivariable logistic regression analyses were performed using a backward selection procedure (p value to enter ≤0.05). Time since transplantation was force-entered into the model to adjust for any potential bias related to the timing of the transplant. To prevent collinearity, representatives for kidney function (ie, urea, eGFR, creatinine, creatinine clearance, cystatin C) were not included simultaneously in the multivariable regression model; instead, the most significant variable in univariable regression analyses was included in the multivariable model. Since longstanding diabetes is predominantly known to cause polyneuropathy, post-transplant diabetes mellitus was excluded from the dichotomous variable for diabetes.³⁰ For logistic regression, dialysis vintage was set to 0 months for KTRs with a pre-emptive transplantation. Potential effect modification by age and sex was assessed by including interaction terms in the different regression models. The quality of the final multivariable model was evaluated using the area under the receiver operator characteristic (ROC) curve to determine its discriminatory value, and a calibration plot to assess the agreement between observed and predicted probabilities of polyneuropathy. To indicate the proportion of variance in polyneuropathy explained by the final multivariable model, Nagelkerke R² was calculated.

An a priori sample size calculation was conducted based on the primary aim of the SENS study, the validation of a clinical score for diagnosing polyneuropathy in KTRs (ClinicalTrials.gov identifier: NCT04664426). It was estimated that 15 participants would be sufficient to detect the anticipated correlation with 90% power and a significance level of 0.05. Additionally, a prevalence-based estimate, grounded in the general population prevalence of polyneuropathy at 7.0%,³¹ suggested a requirement for 205 participants. However, an interim analysis revealed a significantly higher observed prevalence (50.3%), leading to a revised estimate of 30 participants. Ultimately, after completing the remaining planned study visits, a final sample size of 169 KTRs was attained.

SPSS V.29 for Windows (IBM) and R Statistical Software V.4.0.5 (R Foundation for Statistical Computing, Vienna, Austria) were used for statistical analyses. P values below 0.05 were considered statistically significant.

Results

At the initiation of the study inclusion, 574 KTRs were approached for potential participation in the SENS study. As shown in figure 1, a total of 203 KTRs were eventually assessed for eligibility, of whom 34 KTRs were excluded based on the exclusion criteria. In addition, nine KTRs were excluded from the study because of insufficient information for accurate classification by the expert panel, leaving 160 KTRs eligible for analyses.

Figure 1. STROBE flow chart of participant inclusion. ICD, implantable cardioverter-defibrillator; ICU: intensive care unit; STROBE, Strengthening the Reporting of Observational Studies in Epidemiology.

The characteristics of the study population are summarised in table 1. Mean age was 59.8±11.6 years (range: 23–79 years), 107 (67%) were male and the median time since transplantation was 6.1 (3.9 to 13.1) years. At the time of inclusion, 16 (10%) KTRs had received the diagnosis of polyneuropathy in the past.

Table 1. Clinical characteristics of study population.

	KTR
N	160
Age, years	59.8±11.6
Sex, n (%) male	107 (66.9)
Height, cm	176.2±8.8
Weight, kg	81.3±14.3
Time since transplantation, years	6.1 (3.9–13.1)
Primary disease, n (%)
Unknown	26 (16.3)
Glomerulonephritis	45 (28.1)
Pyelonephritis	13 (8.1)
Cystic kidney disease	37 (23.1)
Other congenital/hereditary disease	5 (3.1)
Hypertensive nephropathy	14 (8.8)
Diabetic nephropathy	11 (6.9)
Other multisystem diseases	7 (4.4)
Other	2 (1.3)
Diabetes, n (%)	36 (22.5)
Type 1 diabetes	5 (3.1)
Type 2 diabetes	12 (7.5)
PTDM	19 (11.9)
Pre-emptive transplantation, n (%)	67 (41.9)
Dialysis prior to transplantation*, n (%)	93 (58.1)
Haemodialysis	55 (34.4)
Peritoneal dialysis	46 (28.8)
Dialysis vintage, months	25 (9–42)
Type of transplantation
Kidney	156 (97.5)
Combined kidney and pancreas	4 (2.5)
Donor type, n (%)
Living donor	97 (60.6)
Deceased donor†	63 (39.4)
Donation after brain death	33 (20.6)
Donation after circulatory death	29 (18.1)
Alcohol intake‡, units/week	1.2 (0–6.3)
Smoking status§, n (%)
Never	79 (54.5)
Ex-smoker	57 (39.3)
Current	9 (6.2)
Immunosuppression
Calcineurin inhibitor use, n (%)	137 (85.6)
Tacrolimus use, n (%)	118 (73.8)
Tacrolimus dosage, mg	3.0 (2.0–4.0)
Whole blood trough concentration tacrolimus, µg/L	5.1±1.4
Cyclosporine use, n (%)	19 (11.9)
Cyclosporine dosage, mg	160.5±63.6
Whole blood trough concentration cyclosporine, µg/L	76.0 (54.5–97.5)
Prednisolone use, n (%)	155 (96.9)
Third immunosuppressive
MMF, n (%)	121 (75.6)
Everolimus, n (%)	10 (6.3)
Azathioprine, n (%)	7 (4.4)
Other, n (%)	3 (1.9)
Laboratory values
Urea, mmol/L	10.2±5.8
eGFR, mL/min/1.73 m2	52.4±17.5
Creatinine, µmol/L	130.9±46.8
Creatinine clearance¶, mL/min	72.4±25.5
Albumin**, g/L	42.8±3.0
Cystatin C††, mg/L	1.59±0.53
Urinary protein excretion, g/24 hours	0.17 (0.13–0.28)

^*Eight KTRs received both haemodialysis and peritoneal dialysis prior to transplantation.

^†Missing data (n): 1.

^‡Missing data (n): 12.

^§Missing data (n): 15.

^¶Missing data (n): 5.

^**Missing data (n): 1.

^††Missing data (n): 9.

eGFR, estimated glomerular filtration rate; KTR, kidney transplant recipients; MMF, mycophenolate mofetil; PTDM, post-transplant diabetes mellitus.

To ensure the representativeness of our cohort, we compared baseline characteristics with those of the KTR in the TransplantLines cohort, which were not included in the SENS study, a total number of 1473 KTRs (online supplemental appendix D, table 1). We observed a slight but statistically significant difference in mean age between the two groups (SENS cohort: 59.8±11.6 years vs TransplantLines cohort: 57.8±14.0 years, p=0.04). No other baseline characteristics showed statistically significant differences between the groups.

Polyneuropathy classification

The results of the different components of the neurological assessment can be found in online supplemental appendix C.

In total, 97 KTRs (61%) experienced symptoms of polyneuropathy, most frequently tingling sensation and numbness of the feet (33% and 33%, respectively). Sensory testing was abnormal in 108 KTRs (68%), with the pinprick testing being most often abnormal (62%). Achilles tendon reflexes were reduced in 16% and absent in 29% of KTRs. Patellar tendon reflexes were reduced in 14% and absent in 2% of KTRs. Overall, 31 KTRs (20%) presented with general muscle weakness (ie, in upper and lower limbs). Furthermore, 7 KTRs (4%) showed solely distal muscle weakness in the legs and 3 KTRs (2%) exhibited distal muscle weakness in the arms combined with general muscle weakness in the legs. QST was abnormal in 114 of 158 KTRs (72%). Of these 114 KTR, 69 showed warm sensation abnormalities, 3 cold sensation abnormalities and 42 both warm sensation and cold sensation abnormalities. In total, 92 KTRs (58%) were classified as presenting with abnormal NCSs with the sural nerve most frequently being abnormal: the age-adjusted percentage of KTR with an abnormal sural nerve was 66% for subjects above and 57% for subjects under 60 years of age. The H-reflex was absent in 34% of KTRs.

The decision process of the expert panel is described in online supplemental appendix B, figure 1.

An overview of the categories of polyneuropathy in KTRs is provided in table 2. Of the total 160 included KTR, 84 KTRs (53% (45 to 60)) were classified as having polyneuropathy and 7 KTRs (4% (2 to 9)) were categorised as having small fibre neuropathy.

Table 2. Classification of PNP in KTRs.

			KTRs, n	Prevalence (%), 95% CI
No PNP	No PNP		24	15.0% (10.0% to 21.7%)
	Possible clinical PNP		45	28.1% (21.5% to 35.9%)
PNP	Large fibre PNP	Probable large fibre PNP	26	16.3% (11.1% to 23.1%)
		Definite large fibre PNP	58	36.3% (28.9% to 44.3%)
	Probable small fibre neuropathy		7	4.4% (1.9% to 9.2%)

KTRs, kidney transplant recipients; PNP, polyneuropathy.

A post hoc power analysis revealed that the current study population of 160 participants would have 80% power at a two-sided alpha of 0.05 to yield an accuracy of the prevalence estimation of 3.9% around the point of the estimate of 53% (84 KTRs were classified as having polyneuropathy). This means that with 95% confidence, the prevalence is between 49.1% and 56.9%.

An overview of a comparison of the characteristics between the groups with and without polyneuropathy is provided in table 3 and across the different polyneuropathy categories in online supplemental appendix D, table 2.

Table 3. Comparison of clinical characteristics between KTRs with and without PNP.

	PNP*	No PNP	P value
N	84	69
Age, years	62.9±10.4	56.7±12.2	0.002
Sex, n (%) male	66 (78.6)	36 (52.2)	0.002
Time since transplantation, years	6.6 (3.8–12.5)	5.3 (4.0–14.2)	0.87
Primary disease, n (%)			0.68
Unknown	10 (11.9)	14 (20.3)
Glomerulonephritis	26 (31.0)	18 (26.1)
Pyelonephritis	7 (8.3)	6 (8.7)
Cystic kidney disease	18 (21.4)	17 (24.6)
Other congenital/hereditary disease	2 (2.4)	3 (4.4)
Hypertensive nephropathy	7 (8.3)	7 (10.1)
Diabetic nephropathy	9 (10.7)	1 (1.5)
Other multisystem diseases	4 (4.8)	2 (2.9)
Other	1 (1.2)	1 (1.5)
Diabetes, n (%)	24 (28.6)	10 (14.5)	0.10
Type 1 diabetes	5 (6.0)	0
Type 2 diabetes	9 (10.7)	2 (2.9)
PTDM	10 (11.9)	8 (11.6)
Pre-emptive transplantation, n (%)	30 (35.7)	35 (50.7)	0.17
Dialysis prior to transplantation, n (%)	54 (64.3)	34 (49.3)	0.13
Haemodialysis	40 (47.6)	14 (20.3)
Peritoneal dialysis	20 (23.8)	22 (31.9)
Dialysis vintage, months	25 (10–45)	23 (9–37)	0.88
Donor type, n (%)			0.34
Living donor	49 (58.3)	46 (66.7)
Deceased donor	35 (41.7)	23 (33.3)
Donation after brain death	17 (20.3)	14 (20.3)
Donation after circulatory death	19 (22.6)	8 (11.6)
Alcohol intake, units/week	0.6 (0–4.5)	0.8 (0–6.2)	0.84
Smoking status, n (%)			0.09
Never	33 (46.5)	44 (65.7)
Ex-smoker	34 (47.9)	19 (28.4)
Current	4 (5.6)	4 (6.0)
Immunosuppression
Calcineurin inhibitor use, n (%)	73 (86.9)	57 (82.6)	0.41
Tacrolimus use, n (%)	59 (70.2)	53 (76.8)	0.50
Tacrolimus dosage, mg	3.0 (1.8–4.0)	3.0 (2.0–4.0)	0.38
Whole blood trough concentration tacrolimus, µg/L	5.3±1.6	4.9±1.2	0.21
Cyclosporine use, n (%)	14 (16.7)	4 (5.8)	0.12
Cyclosporine dosage, mg	150.0±67.9	187.5±47.9	0.63
Whole blood trough concentration cyclosporine, µg/L	72.0 (48.8–100.8)	80.0 (77.3–86.5)	0.90
Prednisolone use, n (%)	83 (98.8)	65 (94.2)	0.24
Third immunosuppressive			0.92
MMF, n (%)	62 (73.8)	53 (76.8)
Everolimus, n (%)	7 (8.3)	3 (4.4)
Azathioprine, n (%)	4 (4.8)	3 (4.4)
Other, n (%)	3 (3.6)	1 (1.4)
Laboratory values
Urea, mmol/L	11.7±6.7	8.3±4.0	0.002
eGFR, mL/min/1.73 m2	49.4±18.2	56.4±15.9	0.01
Creatinine, µmol/L	139.6±49.6	118.8±40.2	0.01
Creatinine clearance, mL/min	70.9±27.2	74.7±22.9	0.32
Albumin, g/L	42.3±3.3	43.4±2.5	0.19
Cystatin C, mg/L	1.71±0.61	1.45±0.38	0.01
Urinary protein excretion, g/24 hours	0.17 (0.14–0.28)	0.17 (0.12–0.23)	0.88

P values <0.05 in bold indicate statistical significance.

^*Including solely KTRs with large fibre PNP.

eGFR, estimated glomerular filtration rate; KTRs, kidney transplant recipients; MMF, mycophenolate mofetil; PNP, polyneuropathy; PTDM, post-transplant diabetes mellitus.

KTRs with large fibre polyneuropathy presented with sensor-predominant polyneuropathy (40 KTRs (48% (37% to 59%))) and sensorimotor polyneuropathy (44 KTRs (52% (41% to 63%))) (see table 4 and figure 2). None of the patients demonstrated neurophysiological characteristics of demyelination.

Table 4. Affected nerve fibre type in KTRs with large fibre PNP.

	Probable large fibre PNP		Definite large fibre PNP		Total
	n	% (95% CI)	n	% (95% CI)	n	% (95% CI)
Sensor-predominant PNP	23	88.5% (68.7% to 97.0%)	17	29.3% (18.5% to 42.9%)	40	47.6% (36.7% to 58.7%)
Motor-predominant PNP	0	0	0	0	0	0
Sensorimotor PNP	3	11.5% (3.0% to 31.3%)	41	70.7% (57.1% to 81.5%)	44	52.4% (41.3% to 63.3%)

Sensory abnormalities were defined as abnormal sural nerve amplitudes, motor abnormalities were defined as abnormal tibial nerve and peroneal nerve amplitudes.

KTR, kidney transplant recipient; PNP, polyneuropathy.

Figure 2. Sensorimotor PNP in kidney transplant recipients. Participants with small fibre neuropathy were excluded since they did not have any abnormalities in NCSs. NCSs, nerve conduction studies; PNP, polyneuropathy.

Role of history taking in diagnostic process of polyneuropathy

The frequency of history of symptoms can be found in online supplemental appendix C, table 1.

An exploration of the relation of the history of symptoms of polyneuropathy with the polyneuropathy classification reflecting the results of the additional assessments, that is, neurological examination and NCSs, is shown in figure 3. The findings indicate that 42% (31 to 54) of KTRs without polyneuropathy reported symptoms, and 18% (11% to 27%) of KTRs with large fibre polyneuropathy reported no symptoms.

Figure 3. Relation of history of symptoms with additional assessment. Bar plots indicating the relative frequency of number of symptoms in (A) KTRs without polyneuropathy and (B) KTRs with large fibre polyneuropathy. KTRs with small fibre neuropathy were excluded due to the diagnostic criteria requiring the presence of symptoms consistent with the clinical presentation of small fibre neuropathy (see online supplemental appendix B, criteria for small fibre neuropathy). History taking appears to be unreliable for the diagnosis of polyneuropathy due to discrepancies between the reported symptoms and objective findings. KTR, kidney transplant recipient.

Association of clinical characteristics with polyneuropathy

To identify possible associations of clinical characteristics in the pretransplant and post-transplant period and the outcome of large fibre polyneuropathy, logistic regression analyses were performed (detailed results see online supplemental appendix D).

The online supplemental appendix includes the univariable logistic regression analyses of clinical characteristics with polyneuropathy (online supplemental appendix D, table 3). In the final backward selection model, which was adjusted for time since transplantation, age (OR=1.04 (95% CI 1.01 to 1.08), p=0.01), male sex (OR=2.55 (95% CI 1.19 to 5.60), p=0.02), diabetes (OR=5.58 (95% CI 1.36 to 38.14), p=0.03) and urea (OR=1.12 (95% CI 1.04 to 1.23), p=0.01) remained as significant associations with polyneuropathy in KTRs (see online supplemental appendix D, table 4).

The quality of the final model was assessed, with results presented in online supplemental appendix D, figure 1. The area under the ROC curve was 0.77, indicating an acceptable discriminatory power for predicting polyneuropathy. The calibration curve closely followed the ideal calibration line indicating good agreement between predicted and observed probabilities with minimal systematic bias. The final model explained 30% of the variance in polyneuropathy among KTRs (Nagelkerke R²=0.30).

The cross-sectional study demonstrated that there was no statistically significant difference in the distribution of time since transplantation between KTRs with and without polyneuropathy (OR=0.99 (95% CI 0.95 to 1.04), p=0.75) (see figure 4). This finding may suggest that pretransplant factors may play a greater role in terms of polyneuropathy risk. While cross-sectional studies are not equipped to elucidate causal relationships, this observation offers some support for the hypothesis that kidney transplantation may serve as a protective factor against the onset and progression of polyneuropathy in patients with chronic kidney disease.

Figure 4. Overview of time since kidney transplantation. Histogram of time since transplantation for (A) the entire cohort of KTRs excluding cases of small fibre neuropathy (n=153, purple) and (B) divided according to the presence of large fibre polyneuropathy (PNP) (n=84, blue) and absence of PNP (n=69, orange). No association was observed between the time since transplantation and the presence of PNP. KTR, kidney transplant recipient.

Discussion

This is the first study to describe the prevalence of polyneuropathy in a large and representative cohort of KTRs. Polyneuropathy was highly prevalent in KTRs, mostly presenting as an axonal sensory or sensorimotor large fibre polyneuropathy, making this study the first to explore this neurological complication in KTRs. Higher age, male sex, diabetes and elevated urea levels were significantly associated with polyneuropathy in KTRs.

Our study indicates a high prevalence of large fibre polyneuropathy (53%) among KTRs, suggesting that polyneuropathy is under-recognised in this patient population, as only 10% had received the diagnosis before study participation. This diverged from a recent systematic review and meta-analysis, reporting polyneuropathy in 30% of the 8% of KTRs with neurological complications, corresponding to a prevalence of 2.4% of all KTRs.⁷ To contextualise, the Italian Longitudinal Study on Aging reported the prevalence of polyneuropathy in the general population at 7.0%,³¹ while the Rotterdam study indicated a prevalence of 4.0% in the Dutch population.³² In contrast, uraemic polyneuropathy in end-stage kidney disease shows prevalence estimates between 60% and 100%.³³ A similar observation is made in patients undergoing chemotherapy and patients with inflammatory bowel disease, both of whom exhibit significant underreporting of polyneuropathy.^{34 35} A first possible reason for underreporting of polyneuropathy could be that previous data were based exclusively on history of symptoms, whereas our study showed that symptoms alone are insufficient for diagnostic accuracy. However, most studies neither explore polyneuropathy in greater depth nor incorporate further examinations, such as NCSs, to support the diagnosis of polyneuropathy.836,39 Second, the under-reporting might suggest a slow onset and progression of polyneuropathy, making patients less likely to notice the development of symptoms. Third, when visiting the nephrologist, polyneuropathy may not be the primary concern for KTR, despite being associated with a significant impact on quality of life. Fourth, the underreporting might be due to the complexity of the diagnosis of polyneuropathy, evident from the large number of discrepant cases observed among individual assessments of the expert panel. Differences between assessment of ‘possible’ and ‘probable large fibre’ polyneuropathy particularly highlight the challenges in diagnosing borderline cases. These findings indicate that polyneuropathy in KTRs warrants more attention due to its substantial disease burden and significant impact on quality of life.⁴⁰

Our study reveals that KTRs exhibited primarily axonal sensory or sensorimotor large fibre polyneuropathy, consistent with other electrophysiological studies.^{9 34 35 39 40} In the cohort of KTRs with large fibre polyneuropathy, we observed an age-associated risk of polyneuropathy in KTR. The same observation was made in the general population and is mainly linked to an age-related increase in chronic idiopathic axonal polyneuropathy (CIAP) and diabetic polyneuropathy.^{41 42} While the role of CIAP in KTRs remains unexplored, it has been linked to metabolic syndrome, which is highly prevalent in KTR, suggesting that CIAP may represent a subtype of polyneuropathy in this group.^{43 44} Additionally, polyneuropathy was significantly associated with male sex, a finding supported by cohort studies in the general population.⁴¹ Furthermore, we found an association of polyneuropathy with diabetes and elevated urea levels, which are both known as major risk factors for polyneuropathy.³² First, we observed in this study that nearly all KTRs with diabetic nephropathy as their primary disease exhibited polyneuropathy.³⁰ Second, higher urea levels have been shown to correlate with lower nerve excitability in dialysis patients and have been demonstrated in vitro to compromise neuronal structure and function.^{45 46} The lack of an association with time since transplantation suggests that the underlying aetiology might be found in the pretransplant period. No association was found between calcineurin inhibitor use or concentration and polyneuropathy. However, this might be due to a relatively uniform use of calcineurin inhibitors in KTR. While our data do not support a recommendation to discontinue calcineurin inhibitors in KTRs with polyneuropathy, they do not rule out a potential role. Longitudinal studies are needed to reveal the aetiology of large fibre polyneuropathy in this specific patient population and to further explore the probable role of co-occurring risk factors.⁴⁷

The strength of this study lies in the expert panel’s careful decision-making process to classify patients into different polyneuropathy groups, based on a comprehensive assessment procedure. Furthermore, participants were included through simple random sampling from the TransplantLines cohort since no information on polyneuropathic symptoms and signs was available. No major differences were found between baseline characteristics of KTRs in the SENS study and the TransplantLines Biobank and Cohort Study. Additionally, baseline characteristics including age and sex align with recent literature, suggesting that the SENS study population is representative of the broader KTR population.^{48 49} Although patients with bilateral arteriovenous dialysis shunts were excluded (N=2), 36 KTRs with a unilateral arteriovenous dialysis shunt were included in the current study, minimising the risk of selection bias concerning vascular function. We acknowledge the limitation of using QST, a psychophysical test, as primary additional examination for identifying small fibre neuropathy.^{13 50} Skin biopsy, the gold standard, was not included in this study protocol due to its invasive nature and logistical constraints.⁵¹ Although small fibre neuropathy was diagnosed in a small group of KTRs presenting with matching symptoms and signs, the group was too small for performing any analyses on an underlying aetiology. Whereas the history of symptoms was shown to be insufficient for diagnostic accuracy, it is plausible that experiencing symptoms of polyneuropathy may have led to an increased interest in participating in the current study. This phenomenon could potentially have resulted in an overestimation of the overall group of KTR experiencing symptoms. However, 42% of KTRs without polyneuropathy reported symptoms and 18% of KTRs with polyneuropathy did not report symptoms. Therefore, the risk of substantially overestimating the prevalence of polyneuropathy itself is likely limited. Furthermore, we acknowledge the limitation of not including oral glucose tolerance testing in the definition of diabetes, as HbA1c levels can be affected by altered erythrocyte turnover. Additionally, data on the effectiveness of dialysis treatment prior to transplantation were unavailable, limiting our ability to assess dialysis adequacy as a contributing factor. A potential risk factor for polyneuropathy in KTRs may be prior immunosuppressive treatment for the underlying kidney disease, such as glomerulonephritis. Unfortunately, our cohort was too small to detect such an association; however, this potential risk factor should be considered in future research.

The implications for clinical practice are significant, particularly in the context of enhancing early detection and management of polyneuropathy in KTRs. This study underscores the necessity for effective yet straightforward screening tools that can be used by transplant physicians to identify both symptomatic and asymptomatic patients who may benefit from further evaluation. To gain a better understanding of the onset of polyneuropathy, the course after transplantation and how polyneuropathy could be prevented in KTRs in the future, longitudinal studies are essential.

In conclusion, axonal sensory or sensorimotor polyneuropathy is highly prevalent and often underdiagnosed in KTRs. Next to higher age and male sex, it was predominantly associated with diabetes and higher urea levels. Further research is needed to reveal the aetiology and course of polyneuropathy in KTRs.

Data availability statement

Data are available on reasonable request.