Skin innervation in uremia. The skin biopsy specimens from a control (A, C, and E) and a patient with uremia (B, D, and F) were immunostained with protein gene product 9.5 (PGP 9.5, 1: 1000; UltraClone, Isle of Wight, England). Original magnifications ×200 (A and B) and ×400 (C-F). A, In the skin of a control subject at a low magnification, there are prominent subepidermal nerve plexuses (SNP) between the epidermis (Epi) and dermis (Derm). Intraepidermal nerve fibers (arrows) are observed but require a higher magnification to resolve the details. B, In the skin of a patient with uremia, intraepidermal nerve fibers are nearly depleted and only a small fragment of SNP is seen. C, In this control subject, an intraepidermal nerve fiber ascends from the subepidermal nerve, which shows linear dense immunoreactivities at the border of the epidermis and dermis. D, In the patient with uremia, the epidermis becomes denervated and the SNP is almost absent and some nerve fibers become fragmented, indicating degeneration. E, In the dermis of the control skin, individual nerves in the nerve bundles show intense linear immunoreactivity. F, The nerve fibers in the dermis of the patient with uremia exhibit swelling and segmented features of degeneration. Bar = 50 μm for A and B and 25 μm for C to F.
The relationship between intraepidermal nerve fiber (IENF) densities and duration of chronic kidney disease. The IENF density is negatively correlated with the duration of renal disease (P = .03). The solid line is the regression line. The dotted lines are 95% confidence intervals.
Comparison of abnormal rates on large- and small-fiber examinations in late-stage chronic kidney disease. The graph shows the proportion of abnormal results of intraepidermal nerve fiber (IENF) density on skin biopsy of the leg (67.5%), sympathetic skin response (SSR) in the sole (62.5%), R-R interval variability (RRIV) at rest or during deep breathing (53.6%), and sural (20%) and peroneal (32.5%) nerves of nerve conduction studies.
The relationship between skin innervation and autonomic function test results. The intraepidermal nerve fiber (IENF) density was positively correlated with the R-R interval variability (RRIV) at rest (P = .009). The solid line is the regression line. The dotted lines are 95% confidence intervals.
Chao C, Wu V, Tan C, Wang Y, Tseng M, Wu P, Lin Y, Lin W, Wu K, Hsieh S. Skin Denervation and Its Clinical Significance in Late-Stage Chronic Kidney Disease. Arch Neurol. 2011;68(2):200-206. doi:10.1001/archneurol.2010.372
To investigate the skin innervation and its clinical significance in late-stage chronic kidney disease (CKD).
National Taiwan University Hospital, Taipei, Taiwan.
Forty consecutive nondiabetic patients with late-stage CKD (14 female and 26 male; mean [SD] age, 60.7 [12.3] years), including 2 cases with stage 3 CKD, 6 with stage 4 CKD, and 32 with stage 5 CKD, ie, end-stage kidney disease.
Clinical evaluation of neurological deficits, nerve conduction study, autonomic function tests, and a 3-mm-diameter skin biopsy specimen taken from the distal leg.
Main Outcome Measures
Quantitation of epidermal innervation, parameters of nerve conduction study, R-R interval variability, and sympathetic skin response.
Clinically, 21 patients (52.5%) were symptomatic with paresthesia over the limbs or autonomic symptoms. The intraepidermal nerve fiber (IENF) density was markedly reduced in patients with CKD compared with age- and sex-matched controls (mean [SD], 2.8 [2.0] vs 8.6 [2.8] fibers/mm; P < .001). Skin denervation was observed in 27 patients (67.5%). Fifteen patients (37.5%) had abnormalities on nerve conduction studies, and 29 patients (72.5%) had abnormal results on autonomic function tests. By analysis with multiple regression models, the IENF density was negatively correlated with the duration of renal disease (P = .02). Additionally, the R-R interval variability at rest was linearly correlated with the IENF density (P = .02) and the absence of sympathetic skin responses at the soles was associated with reduced IENF density (P = .03).
Small-fiber sensory and autonomic neuropathies constitute the major form of neuropathy in late-stage CKD. Furthermore, skin denervation was associated with the duration of renal disease.
Neuropathy is common in chronic kidney disease (CKD).1,2 In addition to motor symptoms, sensory complaints are common and diverse in uremic neuropathies, ranging from paresthesia to neuropathic pain.1- 3 Some of these features, such as heat, burning, paradoxical heat, coldness, and freezing, suggest the presence of small-fiber sensory neuropathy (SFSN).4,5 Previous diagnoses of uremic neuropathies were mainly based on abnormal findings in nerve pathology tests, nerve conduction studies (NCSs), and elevated vibration thresholds.6- 8 Seminal studies established axonal degeneration with secondary demyelination as the major pathology of uremic neuropathy.9- 11 In addition to ultrastructural pathology, Krishnan et al12- 14 demonstrated alterations of axonal excitability in end-stage kidney disease (ESKD). All those studies indicated large-fiber neuropathies in end-stage CKD. The spectrum of uremic neuropathies, particularly the issue of SFSN, which affect small myelinated and unmyelinated nerve fibers terminating in the skin, however, has received little attention in the past. Skin biopsy for measuring the intraepidermal nerve fiber (IENF) density is a standard approach for diagnosing SFSN.15- 21 To our knowledge, skin innervation has never been extensively studied in CKD.
Chronic kidney disease, characterized by progressive deterioration of renal functions and related metabolic abnormalities, usually precedes the development of ESKD. The increasing prevalence of CKD and ESKD is a challenge to health care resources worldwide,22 for example, in the United States,23 Japan,24 and Taiwan.25 Therefore, CKD and its related neuropathic complications have become a major health problem. Despite neuropathies being major complications of late-stage CKD,13 clinical correlations of skin innervation with uremic parameters remain unknown. Investigations using skin biopsy specimens should provide new insights into the patterns of neuropathies in late-stage CKD.
In the current study, we used skin biopsies and conventional neurophysiological tests to explore the spectrum of neuropathies in late-stage CKD, including (1) the degree of small-fiber sensory nerve degeneration, (2) the clinical significance and risk factors for skin denervation, and (3) the patterns of neuropathic deficits, ie, large fibers vs small fibers, in patients with late-stage CKD.
Patients with CKD were recruited from the outpatient nephrologic and neurologic clinics and dialysis center of National Taiwan University Hospital, Taipei, and its Yun-Lin Center, Douliou, Yunlin County, southwestern Taiwan, from January 2006 to December 2007. The diagnosis and staging of CKD were based on renal function studies. Patients had to be regularly followed up, and patients with ESKD had to receive regular dialysis therapy if needed. The simplified Modification of Diet in Renal Disease Study equation was used to determine the estimated glomerular filtration rate (eGFR) (eGFR, mL/min/1.73 m2 = 186.3 × [Serum Creatinine Level in Milligrams per Deciliter]−1.154 × Age−0.203 × [0.742 for Women]). The CKD classifications were defined according to the US National Kidney Foundation in 200226: stage 3 CKD, eGFR = 30 to 59 mL/min/1.73 m2; stage 4 CKD, eGFR = 15 to 29 mL/min/1.73 m2; and stage 5 CKD, eGFR <15 mL/min/1.73 m2. The inclusion criteria were (1) persistent renal function impairment for longer than 3 months and (2) moderate to severe reduction in the eGFR (≥CKD stage 3) or well-established renal failure. The duration of renal disease was defined as the interval from the first occasion of abnormal renal function to the time of recruitment. Neurological evaluation included sensations mediated by small (pinprick and thermal) and large (vibration) sensory nerves. The pinprick, thermal, and vibratory sensations were tested with a monofilament (10 g), ice water (4°C, for coldness), and hot water (45°C, for warmth), and 128-Hz tuning fork (for vibration), respectively. Patients were classified as having small-fiber neuropathy if there was reduced sensation to pinprick or thermal stimuli. Large-fiber neuropathy was defined according to reduced vibratory sensation, absence of ankle jerk, or weakness of the toes or feet. Data on previous medical diseases, including diabetes mellitus, autoimmune diseases, chronic infections, drug and toxin exposure, alcoholism, malignancies, vitamin B12 level, erythropoietin use and its total dose in the past 6 months, and other systemic disease, were recorded. To avoid comorbidities that might affect the peripheral nerve system or confound the clinical evaluations, patients with dementia or other neurological diseases, diabetes mellitus, autoimmune syndromes, a malignancy, toxin exposure, vitamin B12 deficiency, or alcoholism were excluded. Each patient underwent skin biopsy, NCS, and autonomic function tests. The procedures of immunohistochemical staining, quantitation of IENF density, NCS, autonomic function tests, and statistical analysis are detailed in the eAppendix. The Ethics Committee of National Taiwan University Hospital approved this study. Informed consent was obtained from each subject before the procedures.
Because the distribution of sensory symptoms in our patients followed a pattern of length-dependent neuropathy, we followed the guideline of skin biopsy.27 A 3-mm-diameter skin punch was taken from the distal leg 10 cm proximal to the lateral malleolus process under local anesthesia with lidocaine, 2%. Sections 50 μm perpendicular to the dermis were immunostained with antisera to protein gene product 9.5 (PGP 9.5, 1: 1000; UltraClone, Isle of Wight, England). Epidermal innervation was quantified and expressed as the number of fibers per millimeter of epidermal length. In the distal leg, normative values from our laboratory (mean [SD] and 5th percentile) of IENF densities were 11.2 [3.7] and 5.8 fibers/mm for subjects younger than 60 years and 7.6 [3.1] and 2.5 fibers/mm for subjects 60 years and older.28 The cutoff point of IENF density was 5.9 and 2.5 fibers/mm in these 2 age groups, respectively. For comparison, data on age- and sex-matched subjects in the control group were retrieved from a previously described cohort.29
In total, 40 patients with CKD (14 female and 26 male; mean [SD] age 60.7 [12.3] years; range, 33-79 years) were enrolled. There was no significant difference in ages and severity of CKD between female and male patients. The stages of CKD and modes of dialytic therapy are listed in Table 1. The underlying etiologies of CKD included idiopathic glomerular nephritis (25 cases), tubulointerstitial nephropathy (1), hypertension (3), urolithiasis (2), renal tumor (1), polycystic kidney (1), or toxicity of drugs (7). The mean (SD) duration of renal diseases was 7.1 (7.3) years (range, 0.3-35 years). Nineteen patients had hypertension under regular medical control. All patients could independently ambulate without obvious weakness.
As for sensory symptoms, 13 patients had persistent numbness or tingling; in 10, this was confined to the lower limbs, while 3 had a stocking-glove distribution. Fourteen patients (35%) had autonomic symptoms: orthostatic hypotension in 7 cases, constipation in 6, and impotence in 4. In total, 21 patients were symptomatic with either sensory symptoms or autonomic dysfunctions, while 19 were asymptomatic. Based on neurological examinations, 8 patients had small-fiber dysfunctions, 6 patients had large-fiber dysfunctions, 13 patients had small- and large-fiber dysfunction, and 13 patients did not have signs of either type deficit. In total, 27 patients (67.5%) had clinical signs of peripheral nerve impairment (Table 1).
To explore the pathology of small-diameter sensory nerves, we performed skin biopsies on these patients. In the control skin, IENFs with a typical varicose appearance arose from the subepidermal nerve plexuses (Figure 1A and C). In contrast, IENFs were markedly reduced in the skin of patients with uremia (Figure 1B and D). Dermal nerve fibers exhibited dense and linear immunoreactivities in the control skin (Figure 1E). In patients with uremia, dermal nerve fibers were fragmented (Figure 1F), consistent with nerve degeneration.30
The IENF densities in patients with CKD were significantly lower than those of age- and sex-matched control subjects (mean [SD], 2.8 [2.0] vs 8.6 [2.8] fibers/mm; P < .001). The IENF densities were reduced in 27 patients (67.5%): 5 in CKD stage 4 and 22 in CKD stage 5. Among these patients, 11 (57.9%) were asymptomatic and 16 (76.2%) were symptomatic. There was no significant difference in IENF densities between the symptomatic and asymptomatic subgroups (mean [SD], 2.6 [2.1] vs 3.0 [1.8] fibers/mm; P = .50).
To explore the risk factors associated with skin denervation in patients with late-stage CKD, we analyzed the correlation of IENF densities with age, body mass index, duration of renal disease or dialysis therapy, CKD stage, the use of erythropoietin, and the total dose of erythropoietin in the past 6 months separately by simple linear regressions and compared the IENF densities between different sexes and between patients with and without hypertension or dialytic therapy. The IENF density was negatively correlated with the duration of renal disease (standardized coefficient = −0.093; P = .03) (Figure 2), and male patients had lower IENF densities than female patients (mean [SD], 2.4 [1.8] vs 3.6 [2.1] fibers/mm; P = .048). There was also a trend between the IENF density and duration of dialysis in those patients with dialysis therapy (P = .049). The IENF density was not correlated with the use (P = .60) or the total dose (P = .78) of erythropoietin. This relationship was further confirmed by the multiple linear regression model. In this model, the IENF density was set as the dependent variable, and age, sex, body mass index, and renal disease duration were used as independent variables. The IENF density was still negatively correlated with the duration of renal disease (standardized coefficient = −0.099; P = .02; and R2 = 0.24 and P = .04 for the model).
On NCSs of the lower limbs, 13 patients (32.5%) had reduction of compound muscle action potentials or a slowing of the motor conduction velocities in the peroneal nerves. Eight patients (20%) had reduction in sensory action potentials or a slowing of sensory conduction velocities in the sural nerves. In the autonomic function tests, sympathetic skin responses (SSRs) were abnormal in 25 patients (62.5%): absent responses from the palms of 17 patients and from the soles of 25 patients. For 28 patients without arrhythmia, 15 (53.6%) had reduced R-R interval variability at rest or during deep breathing. In total, 15 patients (37.5%) had neurophysiological evidence of large-fiber neuropathies, and 11 of them had reduced IENF density. For autonomic function tests, 29 patients (72.5%) had abnormal results and IENF density was reduced in 22 of these patients. In summary, the autonomic function test results had the highest overall abnormal rate, similar to the abnormal rate of skin biopsy results but much higher than those of NCSs (Figure 3).
To investigate whether there were parallel changes between skin denervation and neurophysiologic dysfunctions in patients with CKD, we analyzed the relationship between the results of NCSs and autonomic function tests and IENF densities. The IENF densities had no correlation with results of the NCSs but were positively correlated with the R-R interval variability during rest (standardized coefficient = 0.304; P = .009) (Figure 4) in the simple linear regression model. The IENF densities were lower in patients without an SSR in the sole than in those with an SSR in the sole (P = .04). These correlations were further refined by the multiple regression model. In this model, the IENF density was set as the dependent variable, and age, sex, the duration of renal disease, and the amplitudes of the sural sensory action potentials and peroneal compound muscle action potentials with R-R interval variability at rest or SSR in the sole were set as independent variables. The results showed that the IENF density was still correlated with R-R interval variability during rest (P = .02) and was associated with the SSR in the sole (P = .03) (Table 2).
The present study demonstrated important findings on skin denervation in late-stage CKD, including (1) significant skin denervation in patients with late-stage CKD, with 67.5% patients having reduced IENF densities, and (2) correlation of skin denervation with the duration of renal diseases and the abnormal autonomic function test results (R-R interval variability at rest and SSR in the sole). These observations indicate that small-fiber neuropathy including sensory and autonomic involvement is a major neurological deficit compared with large-fiber neuropathies in nondiabetic patients with CKD.
The results document a higher prevalence of SFSN based on skin biopsy results compared with that of the NCS-based large-fiber neuropathies (67.5% vs 37.5%) in late-stage CKD. Although there are studies on the prevalence of neuropathies in chronic renal failure, those data were based on neurophysiological studies or nerve biopsies.6,31- 33 The prevalence of large-fiber neuropathies might be underestimated in the present study because pathological examinations of large myelinated nerves (sural nerve biopsy or motor point biopsy) were not included.34 Pathologic examinations demonstrated axonal atrophy and secondary demyelination in uremic neuropathy.2 In ESKD, there were significant changes in axonal excitability.12,14 Only limited studies have examined SFSN; for example, in 1 report evaluating thermal thresholds in patients with ESKD, there was little evidence of small-fiber dysfunction compared with large-fiber deficits on NCSs.35 The low sensitivity of thermal thresholds for detecting SFSN might cause underestimation of the incidence of SFSN in ESKD.36,37 In contrast, the assessment of IENF density, with a sensitivity of 80% to 90%, provides direct evidence of SFSN according to the pathology of cutaneous nerve degeneration.15,17,18,20,38,39 Because other diseases causing neuropathies, such as diabetes, were excluded, our study strongly suggests the association between SFSN and CKD. Skin denervation is not only present in ESKD but also exists in the preuremic stage, indicating the early involvement of small-diameter sensory nerves. Additionally, among the 27 patients with skin denervation, 11 patients were free of neurological symptoms. Such pathologic evidence suggests that SFSNs are common in late-stage CKD.
The high frequency of autonomic dysfunctions (72.5%) and the association with IENF density in the current study strongly indicate that autonomic dysfunctions are an important component of small-fiber neuropathies in CKD. This prevalence of autonomic dysfunctions was similar to the frequency of skin denervation (67.5%) but was much higher compared with the abnormal rate (37.5%) of NCS results, suggesting a predilection for the pathogenesis of small-fiber injury in late-stage CKD. Autonomic dysfunctions in uremia predispose patients to arrhythmias and represent an independent risk factor for sudden cardiac death in uremia.40,41 Our findings provide a further association between the physiology of autonomic dysfunctions and the pathology of SFSN. Taken together, these findings suggest that SFSNs and autonomic neuropathies are the major forms of neuropathy in CKD.
An intriguing observation in the current report was the lower IENF density in male patients than in female patients. Previous studies on IENF density of normal subjects have indicated that (1) female control subjects had higher IENF density than male ones and (2) IENF density was reduced with aging.42 In the current report, there was no difference in the age and severity of CKD between male and female patients. A possible explanation for the sex difference might be due to higher IENF densities in women than in men at baseline.42,43
In conclusion, SFSN is a major component of neuropathy in late-stage CKD and skin biopsy provides pathological evidence of SFSN, especially for those with long duration of renal disease and autonomic dysfunctions. The combination of skin biopsies, NCSs, and autonomic function tests provides a comprehensive picture of neuropathies involving large vs small fibers and identifies previously overlooked SFSN in late-stage CKD.
Correspondence: Sung-Tsang Hsieh, MD, PhD, Department of Neurology, National Taiwan University Hospital, 7 Chung-Shan S Rd, Taipei 10002, Taiwan (email@example.com).
Accepted for Publication: September 28, 2010.
Author Contributions: Dr Hsieh had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Chao, V. C. Wu, and Hsieh. Acquisition of data: Chao, V. C. Wu, Tan, Wang, P. C. Wu, and Y. H. Lin. Analysis and interpretation of data: Chao, V. C. Wu, Tan, Tseng, W. M. Lin, K. D. Wu, and Hsieh. Drafting of the manuscript: Chao, Tan, and Hsieh. Critical revision of the manuscript for important intellectual content: V. C. Wu, Wang, Tseng, P. C. Wu, Y. H. Lin, W. M. Lin, K. D. Wu, and Hsieh. Statistical analysis: Chao, V. C. Wu, Tseng, and Hsieh. Administrative, technical, and material support: Tan, Wang, P. C. Wu, Y. H. Lin, W. M. Lin, and Hsieh. Study supervision: K. D. Wu and Hsieh.
Financial Disclosure: None reported.
Funding/Support: The work was supported by grant NHRI-EX98-9736NI from the National Health Research Institute, Taiwan, and grants NSC97-2320-B-002-042-MY3 and NCS98-2314-B002-102-MY3 from the National Science Council, Taiwan.
Online-OnlyMaterial: The eAppendix is available at http://www.archneurol.com.