Brain Advance Access originally published online on May 5, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Brain, Vol. 127, No. 7, 1593-1605,
July 2004
© 2004 Guarantors of Brain
doi: 10.1093/brain/awh180
Skin denervation in type 2 diabetes: correlations with diabetic duration and functional impairments
Departments of 1 Pathology, 2 Neurology and 3 Internal Medicine, National Taiwan University Hospital and Departments of 4 Forensic Medicine and 5 Anatomy and Cell Biology, National Taiwan University College of Medicine, Taipei, Taiwan
Correspondence to: Dr Sung-Tsang Hsieh, Department of Anatomy and Cell Biology, National Taiwan University College of Medicine, 1 Jen-Ai Road, Sec. 1, Taipei 10018, Taiwan E-mail: sthsieh{at}ntumc.org
Received October 27, 2003. Revised February 29, 2004. Accepted March 3, 2004.
 |
Summary |
|---|
 Top Summary IntroductionMaterial and methodsResultsDiscussionAppendixReferences |
|---|
Sensory neuropathy is a prominent component of diabetic neuropathy. It is not entirely clear how diabetes influences skin innervation, and whether these changes are correlated with clinical signs and laboratory findings. To investigate these issues, we performed skin biopsies on the distal leg of 38 consecutive type 2 diabetic patients with sensory symptoms in lower limbs (25 males and 13 females, aged 56.2 ± 9.4 years) and analysed the correlations of intraepidermal nerve fibre (IENF) densities in skin with glycaemic status (duration of diabetes, HbA1C, and fasting and post-prandial glucose levels), and functional parameters of small fibres (warm and cold thresholds) and large fibres (vibratory threshold and parameters of nerve conduction studies). Clinically, 23 patients (60.5%) had signs of small-fibre impairment, and 19 patients (50.0%) had signs of large-fibre impairment. IENF densities were much lower in diabetic patients than in age- and gender-matched controls (1.794 ± 2.120 versus 9.359 ± 3.466 fibres/mm, P < 0.0001), and 81.6% (31/38) of diabetic patients had reduced IENF densities. IENF densities were negatively associated with the duration of diabetes (standardized coefficient: –0.422, P = 0.015) by analysis with a multivariate linear regression model. Abnormal results of functional examinations were present in 81.6% (warm threshold), 57.9% (cold threshold), 63.2% (vibratory threshold) and 49% (amplitude of sural sensory action potential) of diabetic patients. Among the three sensory thresholds, the warm threshold temperature had the highest correlation with IENF densities (standardized coefficient: –0.773, P < 0.0001). On nerve conduction studies in lower-limb nerves, there were abnormal responses in 54.1% of sural nerves, and 50.0% of peroneal nerves. Of neurophysiological parameters, the amplitude of the sural sensory action potential had the highest correlation with IENF density (standardized coefficient: 0.739, P < 0.0001). On clinical examination, 15 patients showed no sign of small-fibre impairment, but seven of these patients had reduced IENF densities. In conclusion, small-fibre sensory neuropathy presenting with reduced IENF densities and correlated elevation of warm thresholds is a major manifestation of type 2 diabetes. In addition, the extent of skin denervation increases with diabetic duration.
Key Words: Keywords: diabetic neuropathy; epidermal nerves; skin biopsy; ubiquitin; quantitative sensory testing; nerve conduction studies
Abbreviations: Abbreviations: 95% CI= 95% confidence interval; CMAP = compound muscle action potential; IENF density = intraepidermal nerve fibre density; NCV = nerve conduction velocity; PGP 9.5 = protein gene product 9.5; QST = quantitative sensory testing; SAP = sensory action potential
|
Introduction |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionAppendixReferences |
|---|
Diabetes mellitus is a common cause of peripheral nerve disorders in adults (Singleton et al., 2001
). Among different subtypes of diabetic neuropathy, sensory neuropathy is a frequent cause of painless injury, and is related to various complications including the necessity for the amputation of limbs (McNeely et al., 1995; Adler et al., 1999). Thus, early detection of sensory nerve impairment is an important issue and a great challenge in evaluating diabetic neuropathy (Dyck et al., 1999; Polydefkis et al., 2003). The perception of nociceptive stimuli begins with sensory receptors in the skin with unmyelinated sensory nerves terminating in the epidermis (Kennedy and Wendelschafer-Crabb, 1993; McCarthy et al., 1995; Hsieh et al., 2000). Evaluation of these sensory structures in the skin should be useful for diagnosis of diabetic sensory neuropathy. Epidermal nerves in the skin are readily demonstrated by immunohistochemistry with various neuronal markers, particularly protein gene product 9.5 (PGP 9.5), a ubiquitin C-terminal hydrolase (Wilson et al., 1988; Karanth et al., 1991; Kennedy and Wendelschafer-Crabb, 1993; Periquet et al., 1999; Griffin et al., 2001). Sensory nerve terminals in the epidermis of the skin degenerate after cutaneous denervation by mechanical or chemical insults, and epidermal nerve fibres are depleted earlier than their nerve trunks in the dermis (Hsieh et al., 2000; Rajan et al., 2003). Skin biopsy together with quantification of epidermal nerve fibres has therefore become a novel pathological approach to diagnose small-fibre sensory neuropathy (Kennedy and Said, 1999; Griffin et al., 2001).
Intraepidermal nerve fibre (IENF) densities are reduced in patients with impaired glucose tolerance and clinically overt diabetes (Levy et al., 1989; Kennedy and Wendelschafer-Crabb, 1996; Smith et al., 2001; Sumner et al., 2003). These results clearly suggest that epidermal nerves are affected in diabetes, and skin biopsy is useful in evaluating small-fibre sensory neuropathy of diabetes (Kennedy and Wendelschafer-Crabb, 1999; Polydefkis et al., 2001). Apparently, epidermal denervation is related to diabetic status. However, there is a lack of direct evidence regarding the correlation between epidermal denervation and diabetic parameters, such as diabetic duration and hyperglycaemic control.
Small-diameter sensory nerves are presumably responsible for detection of thermal stimuli to the skin, and thermal thresholds are altered in diabetic subjects (Dyck et al., 2000). Quantitative sensory testing examines the function of the thermal stimuli-detecting system, but provides limited information regarding the pathology of small-diameter nerves. It remains unanswered whether changes in IENF densities parallel changes in thermal thresholds in diabetic patients (Kennedy et al., 1996; Hirai et al., 2000; Yasuda et al., 2000; Sumner et al., 2003).
Sensory neuropathies can be classified as large-fibre type and small-fibre type according to clinical signs. In addition to clinical assessment, nerve conduction studies are performed to evaluate large-fibre neuropathy, and skin innervation can be used as a parameter of small-fibre sensory neuropathy (Sumner et al., 2003). Diabetic neuropathy is conceptually considered a mixed neuropathy of both large- and small-fibre types (Thomas and Tomlinson, 1993). It would be intriguing to investigate whether large- and small-diameter nerves are affected in diabetes to similar degrees, and how clinical assessments and laboratory diagnoses are related.
To address the issues of small-diameter sensory nerve degeneration and the extent of large- and small-fibre neuropathies in human diabetes, we studied diabetic patients by evaluating skin innervation, measuring sensory thresholds with quantitative sensory testing, and comparing results from nerve conduction studies.
|
Material and methods |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionAppendixReferences |
|---|
Patients and control subjects
The study population exclusively consisted of diabetic patients with sensory symptoms referred to an out-patient neuropathy clinic of National Taiwan University Hospital, Taipei, Taiwan between January 2000 and August 2001. Symmetric sensory symptoms in the foot with a graded stocking pattern of distribution were the prominent feature of all these patients. All patients were able to ambulate without clinical weakness or signs of injury. Neurological examinations followed routine procedures, including detailed examinations of sensations to hot, cold, vibratory and kinaesthetic stimuli. Analysis of these signs was modified from several clinical forms, including the Neuropathy Symptom Score, Neuropathy Disability Score and Total Neuropathy Score (Dyck et al., 1980
; Cornblath et al., 1999), and results are listed in Table 1. Painful neuropathy was defined as burning or shooting pains disturbing sleep and work activities (Eaton et al., 2003). No medication was given for painful neuropathy before neurophysiological, psychophysical and pathological evaluations. Hypoglycaemic or insulin-induced neuropathy was excluded by (i) a history of hypoglycaemic events and cerebral manifestations; and (ii) the absence of muscle weakness (Thomas and Tomlinson, 1993). All patients completed an initial evaluation that revealed no history of familial neuropathy or toxin exposure, and normal results on blood tests including electrolytes, liver function, complete blood count, thyroid function tests, vitamin B12 and folic acid levels, and serum protein electrophoresis with immunofixation analysis.
|
A diagnosis of type 2 diabetes was based on the revised American Diabetes Association recommendations (American Diabetes Association, 1997
). These patients were regularly followed-up and received oral hypoglycaemic agents and/or insulin. Patients usually visited the diabetic clinic once every 1–3 months for adjustment of diabetic control by regular measurements of fasting and 2 h post-prandial glucose levels and glycated haemoglobin (HbA1C). Data of hyperglyacemic control 3 years before the skin biopsy were analysed for patients with diabetic duration longer than 3 years. For those patients with diabetic duration shorter than 3 years, all glycaemic data before the skin biopsy were collected for analysis.
There were 38 diabetic patients (25 males and 13 females), aged 56.2 ± 9.4 years (range, 36–72) with a diabetic duration of 10.55 ± 7.20 years (range, 1–23). The wide range reflected the various sources of referral and the diversity of patients. Subjects in the control group were recruited from a previously described cohort (Pan et al., 2001a) matched by gender and age (stratified by decade), including 25 males and 13 females, aged 55.5 ± 11.0 years (range, 31–73).
Skin biopsy
A skin specimen of 3 mm in diameter was taken with a biopsy punch from the lateral side of the distal leg under 2% lidocaine local anaesthesia (Chien et al., 2001). All subjects tolerated the procedure with no obvious signs of discomfort. No suturing was required, and the wounds were covered with a piece of gauze. Wound healing took 7–10 days, similar to a typical abrasion wound. The Ethics Committee of National Taiwan University Hospital had approved this study. Informed consent was obtained from each patient before the skin biopsy.
Immunohistochemistry
For immunohistochemistry on microtome sections (Hsieh and Lin, 1999), skin tissues were fixed with 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS) pH 7.4 for 48 h. Sections at 50 µm were quenched with 1% H2O2, blocked with 5% normal goat serum, and incubated with rabbit antiserum to PGP 9.5 (UltraClone, UK, diluted 1 : 1000 in 1% normal serum/Tris) at 4°C for 16–24 h. After rinsing in Tris, sections were incubated with biotinylated goat anti-rabbit immunoglobulin G at room temperature for 1 h, followed by incubation with the avidin–biotin complex (Vector, Burlingame, CA) for another hour. The reaction product was demonstrated with chromogen SG (Vector), and counterstained with eosin (Sigma, St Louis, MO).
Quantification of epidermal innervation
Epidermal innervation was quantified following established protocols, and slides were coded to ensure that measurements were blinded (Chiang et al., 1998). PGP 9.5-immunoreactive nerve fibres in the epidermis of each section were counted at a magnification of 40x with an Olympus BX40 microscope (Tokyo, Japan) through the depth of the entire section. Each individual nerve fibre with branching points inside the epidermis was counted as one. For epidermal nerve fibres with branching points in the dermis, each individual nerve fibre was counted separately. The length of the epidermis along the upper margin of the stratum corneum in each section was measured using Image-Pro PLUS software (Media Cybernetics, Silver Spring, MD). IENF density was derived and expressed as fibres/mm. For each tissue, there were 48–50 sections after sectioning, and all sections were labelled sequentially. Every fourth section of each tissue sample was immunostained. The mean of epidermal nerve fibre densities of the stained sections was defined as the IENF density of that tissue specimen. In the distal leg, normative values of IENF density from our laboratory were 11.16 ± 3.70, 5.88 and 4.2 fibres/mm (mean ± SD, the 5th percentile value, and the 1st percentile value, respectively) for those aged <60 years, and 7.64 ± 3.08, 2.50 and 2.2 fibres/mm for those aged 60 years. These values are similar to those of McCarthy et al. with the same staining methods and quantitation criteria (McCarthy et al., 1995, 1998; Chien et al., 2001). The cut-off point of IENF density was 5.88 and 2.50 fibres/mm according to the respective age group, and skin denervation was defined as an IENF density lower than the cut-off value for that patient.
Quantitative sensory testing
We performed quantitative sensory testing (QST) with a Thermal Sensory Analyzer and Vibratory Sensory Analyzer (Medoc Advanced Medical System, Minneapolis, MN) to measure sensory thresholds of warm, cold and vibratory sensations. The facilities and procedures were described previously (Yarnitsky and Ochoa, 1991; Pan et al., 2001a, 2003). Maintenance procedures followed the manufacturer’s suggestions and were done every 4 weeks, including filling water and adjustment of pump voltage. The analyser was calibrated with two methods every 2 months: (i) using the default test programs of the sensory analyser, including the patient response unit test programme and the thermode test program; and (ii) measuring the surface temperature of the thermode stimulator by attaching an MC 8700 type digital thermometer (Exacon A/S, Roskilde, Denmark) to the thermode.
The stimulator was applied to the skin of the foot dorsum for thermal and vibratory stimuli. The examiner explained the procedures to the subjects, and the subjects underwent several trials to become familiar with the test.
We used two testing strategies: the method of limits and the method of level. Results of these two algorithms were correlated; correlation coefficients were 0.89–0.92 among different sensory modalities (Lin et al., 1998). The method of level is independent of reaction time. Briefly, the sensory analyser delivered a stimulus of constant intensity set by the algorithm. The intensity of the next stimulus was either increased or decreased by a fixed ratio according to the response of the subject, i.e. whether or not the subject had perceived the stimulus. Such procedures were repeated until a predetermined difference in intensity was reached. The mean intensity of the final two stimuli was the threshold for the level method. Thermal thresholds were expressed as a warm threshold temperature and cold threshold temperature. Vibratory thresholds were measured with similar algorithms, and are expressed in micrometres.
We have established normative data for Taiwanese using descriptive functions, including percentiles, of the statistical software SPSS (Chicago, IL). This database was set up initially in 1996, and has been updated every 6 months. Currently, there are data on >400 normal Taiwanese subjects in the database. A staff neurologist had examined each subject of the control database to ensure the absence of neurological symptoms and signs. Systemic diseases were excluded by laboratory examinations including plasma glucose levels and kidney and liver functions. We previously compared the normative values of this database with those of other databases, and no significant ethnic differences were detected (Yarnitsky and Ochoa, 1991; Yarnitsky and Sprecher, 1994; Yarnitsky, 1997; Lin et al., 1998). The normative values of this database are listed in the Appendix. Cut-off values were defined as the 95th percentile value (for warm threshold and vibratory threshold) and the 5th percentile value (for cold threshold). Thresholds beyond the cut-off value were considered abnormal.
Nerve conduction studies
Nerve conduction studies were performed with a Nicolet Viking IV Electromyographer (Madison, WI) in all patients following standardized methods recommended by the Consensus Development Conference on Standardized Measures in Diabetic Neuropathy (Consensus Development Conference, 1992; Pan et al., 2003). Amplitudes of the sensory action potential (SAP) and compound muscle action potential (CMAP) as well as the nerve conduction velocity (NCV) were recorded for analysis. Studied nerves included sural, peroneal and median (motor and sensory) nerves. Criteria for abnormality of each individual nerve were based on the lower limits of amplitudes and conduction velocities of SAP and CMAP (Pan et al., 2001b).
Statistical analysis
Numerical variables are expressed as the mean ± SD, and compared with t tests if the data followed a Gaussian distribution. In the first step, graphic evaluation of the correlations between variables was analysed with the slope of the regression line, including the 95% confidence interval (95% CI), using GraphPad Prism (GraphPad Software, San Diego, CA). The correlation was explored further with multiple linear regression analysis using SPSS software. To evaluate the influence of diabetes on skin innervation, IENF density was set as the dependent variable. Age, gender and diabetic parameters (including diabetic duration, fasting and post-prandial glucose levels, and HbA1C) were independent variables. To understand whether sensory thresholds were correlated with IENF densities in diabetic patients, IENF density, age, gender and significant diabetic parameters from the previous analysis were independent variables. Each sensory threshold was a dependent variable. In addition to entering all independent variables for analysis, forward and backward stepwise linear regressions were applied. The significance of each model was determined by R2 (goodness-of-fit for the model), and the significance of each variable in the tested model was judged by the coefficient for that variable (standardized coefficients ß, t, and P). Results were considered significant at P 
0.05.
|
Results |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionAppendixReferences |
|---|
Clinical features of diabetic patients
All 38 patients had constant, symmetric paraesthesia in the lower limbs but with various distributions: limited to the toes in 13 patients (34.2%), involving the entire foot in 18 patients (47.4%), affecting the foot and the distal leg in six patients (15.8%), and affecting the leg and foot in one patient (2.6%) (Table 1). None had motor weakness as assessed by neurological examinations. Ankle jerk was absent in 11 patients (28.9%); and ankle jerk and knee jerk were absent in nine of those patients (23.7%). Generalized areflexia was noted in seven patients (18.4%). Nine patients (23.7%) had painful neuropathy, such as burning and shooting pain. Patients without painful neuropathy tended to have distribution of sensory symptoms limited to the toes compared with patients with painful neuropathy (P < 0.0072). Painful neuropathy was not related to diabetic duration (P = 0.54).
Clinically, sensory neuropathy according to signs on neurological examinations was divided into small-fibre type (impairment of at least one sensation to pinprick, warm stimulus and cold stimulus), and large-fibre type (impairment of at least one kinaesthetic sensation). Twenty-three patients (60.5%) had signs of small-fibre impairment, and 19 patients (50.0%) had signs of large-fibre impairment. Ten patients (26.3%) had no clinical sign of either type: eight of these had symptoms limited to the toes, and the other two patients had symptoms in the foot.
Skin innervation in diabetic patients
In normal skin, typical epidermal nerve fibres immunoreactive for PGP 9.5 ascended through the epidermal–dermal border, occasionally with a varicose appearance (Fig. 1A). PGP 9.5 (+) nerves formed dense subepidermal nerve plexuses in control subjects. In the epidermis of diabetic patients, the abundance of epidermal nerves was markedly reduced. In many patients, the skin was completely denervated (Fig. 1B).
|
In the dermis of normal skin, individual dermal nerves with dense immunoreactivities in linear patterns were grouped together as nerve bundles (Fig. 1C). The immunoreactive pattern of dermal nerves became fragmented in the dermis of diabetic patients, suggesting nerve degeneration (Fig. 1D).
Quantitative pathology in diabetic skin
To quantify the pathology of epidermal innervation, we compared IENF densities of diabetic patients with those of age- and gender-matched normal subjects serving as controls. IENF densities of diabetic patients were significantly lower than those of age- and gender-matched control subjects (1.794 ± 2.120 versus 9.359 ± 3.466 fibres/mm, P < 0.0001, Fig. 2), with 81.6% (31 out of 38) of diabetic patients having IENF densities lower than the 5th percentile value of the norm (Fig. 3).
|
|
To explore further the influence of diabetes on skin innervation, we plotted the relationship between IENF densities and diabetic parameters (Fig. 4). IENF densities were negatively correlated with diabetic duration (slope = –0.1220 ± 0.052, 95% CI = –0.2272 to –0.0169, P = 0.024, Fig. 4A). Other parameters (HbA1C, fasting and post-prandial glucose levels) were not linearly correlated with IENF densities (Fig. 4B–D).
|
Because age and gender influence IENF densities in normal subjects (McArthur et al., 1998
), we performed a multiple linear regression analysis (Table 2). Diabetic duration was significantly associated with IENF density after age and gender were controlled for, and had a standardized coefficient of –0.422 (P = 0.015). The standardized coefficient of post-prandial glucose level with IENF density was 0.367, but this value did not reach statistical significance (P = 0.056).
|
Thermal thresholds in diabetes
To understand the functional correlations of cutaneous nerve degeneration, sensory thresholds of various modalities in diabetic patients were compared with those of age- and gender-matched control subjects. With regard to the functions of small-diameter sensory nerves, diabetic patients had abnormal thresholds to warm and cold stimuli. Warm threshold temperatures of the foot dorsum (the method of level) were significantly higher in diabetic patients than in control subjects (44.43 ± 4.26 versus 37.58 ± 1.60°C, P < 0.0001, Fig. 5A). In summary, 81.6% of diabetic patients had elevated warm thresholds (Fig. 3). A similar trend was observed by using the method of limits, with higher warm threshold temperatures in diabetes than in control subjects (45.94 ± 3.87 versus 39.25 ± 2.32°C, P < 0.0001).
|
Diabetic patients also had reduced cold threshold temperatures (the method of level) in the foot dorsum (22.66 ± 11.02 versus 30.31 ± 0.79°C, P = 0.001, Fig. 5B). The abnormal rate of altered cold thresholds in the diabetic group was 57.9% (Fig. 3). Cold threshold temperatures measured by the method of limits were also markedly reduced in diabetic patients compared with normal subjects (20.07 ± 11.02 versus 28.50 ± 1.88°C, P < 0.0001).
Large-fibre neuropathy in diabetes
To evaluate the function and physiology of large-diameter nerves, we compared the results of vibratory thresholds and nerve conduction studies between diabetic and control subjects. Vibratory thresholds (the method of level) were higher in diabetic patients than in control subjects (25.34 ± 33.93 versus 4.10 ± 1.64 µm, P < 0.0001, Fig. 5C), with 63.2% of diabetic patients having elevated vibratory thresholds (Fig. 3). Similarly, vibratory thresholds using the method of limits were elevated in diabetic patients compared with control subjects (29.24 ± 34.00 versus 6.55 ± 2.81 µm, P < 0.0001).
In nerve conduction studies, both amplitude and conduction velocity of motor and sensory nerves in diabetic patients were significantly reduced compared with age- and gender-matched control subjects (Table 3). Consistent with previous findings, a trend for higher abnormal rates was noted in lower-limb nerves versus upper-limb nerves, and in sensory nerves versus motor nerves (Albers et al., 1996). Although this difference was small due to the small sample size, the findings indicated that the current study group was not a biased population.
|
Parallel changes in IENF density with sensory thresholds and neurophysiological parameters
To investigate whether sensory thresholds paralleled IENF densities in diabetes, each sensory threshold was plotted against the IENF density (Fig. 6). All sensory thresholds were linearly correlated with IENF densities with a slope of –1.226 ± 0.225 (95% CI = –1.683 to –0.768, P < 0.0001) for the warm threshold (Fig. 6A), 2.375 ± 0.771 (95% CI = 0.8113–3.939, P = 0.0039) for the cold threshold (Fig. 6B), and –6.633 ± 2.458 (95% CI = –11.630 to –1.640, P = 0.0106) for the vibratory threshold (Fig. 6C).
|
Because diabetic duration was significantly associated with IENF densities, the correlation between sensory thresholds and IENF density was refined using multiple linear regression analysis (Table 4). After controlling for diabetic duration, age and gender, IENF densities were linearly correlated with all these sensory thresholds, particularly the warm threshold of the foot dorsum (standardized coefficient: –0.773 for the method of level, and –0.779 for the method of limits, P < 0.0001). A similar analysis was performed for parameters of nerve conduction studies (Table 5). Among these, sural SAP amplitude had the highest standardized coefficient (0.739, P < 0.0001).
|
|
Comparison of clinical and laboratory assessments
To understand the clinical value of skin biopsy, we compared the relationship between clinical and laboratory assessments (Table 6). We employed objective parameters as criteria of laboratory neuropathy: reduced IENF density for small-fibre neuropathy, and reduced sural SAP amplitude and/or NCV for large-fibre neuropathy. A substantial proportion of patients (seven out of 15) had subclinical small-fibre neuropathy on laboratory assessment: two from the subgroup of 10 patients without clinical signs, and five of all patients with signs of large-fibre impairment only.
|
We then examined the data to see whether laboratory parameters were related to symptoms (Table 6). The biopsy sites were at symptomatic areas of seven patients (Table 1). In the 10 patients without clinical signs, eight had symptoms limited to toes, and symptoms of the other two patients were limited to the foot. For patients with symptoms extending to the leg, IENF densities were much lower than those with symptoms limited to the toes and feet (P = 0.0116), as was the sural SAP amplitude (P = 0.0034) (Table 7). The presence of neuropathic pain, however, was not correlated with IENF density or sural SAP amplitude (P = 0.1137 and 0.1962, respectively) (Table 7).
|
|
Discussion |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionAppendixReferences |
|---|
The present study demonstrates significant skin denervation in type 2 diabetes, occurring in 81% of the current study group. Epidermal denervation was associated with pathological evidence of dermal nerve degeneration. The degree of epidermal denervation was negatively associated with diabetic duration. In this group of diabetic patients, IENF densities in the leg were highly correlated with warm thresholds of the foot; the high proportion of skin denervation together with functional impairments indicates that small-fibre sensory neuropathy is a major manifestation of diabetic neuropathy.
Epidermal denervation in diabetic skin
Skin biopsy has become a new approach to investigate small-fibre sensory neuropathy (Kennedy and Said, 1999; Polydefkis et al., 2003). The current study focused on skin innervation and parallel changes in psychophysical parameters. Characteristic findings in diabetic skin include marked denervation of the epidermis, a decrease in subepidermal nerve abundance, and the presence of fragmented dermal nerves, indicating ongoing nerve degeneration (Hsieh et al., 2000; Pan et al., 2003). These pathological hallmarks clearly suggest that nerve degeneration is responsible for epidermal denervation and that small-diameter sensory neuropathy is a major manifestation of diabetic neuropathy. This finding extends previous observations of small-fibre sensory neuropathy in diabetes; for example, epidermal nerve densities and lengths are reduced in type 1 diabetes (Kennedy and Wendelschafer-Crabb, 1996). The reduction in epidermal nerve density is an early event in diabetes, as demonstrated in recent studies on subjects with impaired glucose tolerance and clinical diabetes (Smith et al., 2001; Sumner et al., 2003). This is in contrast to another report indicating that epidermal innervation is increased in the early stages of type 1 diabetes (Properzi et al., 1993). Despite quantitative changes in epidermal innervation, several issues remain to be explored: for example, direct evidence of dermal nerve degeneration as a cause of epidermal denervation, relationships to hyperglycaemic control, and functional correlations of epidermal denervation. The current study suggests that morphometric changes in epidermal innervation may be related to the progression of neuropathy, as recently demonstrated in acrylamide neurotoxicity and painful neuropathy (Ko et al., 2002; Lauria et al., 2003).
Diabetes exerts a profound influence on sensory neurons; potential mechanisms include vascular, metabolic and immunological defects (Said et al., 2003). Neurological manifestations in most patients with diabetic neuropathy exhibit a stocking or glove-stocking pattern, suggesting that nerve terminals may be early targets of diabetic neuropathy, and skin biopsy can demonstrate lesions in the distalmost parts of nerve terminals. Similar approaches have been reported for diseases in which sensory nerves are predominantly affected, e.g. leprosy (Facer et al., 1998).
Skin biopsy and diagnosis of small-fibre sensory neuropathy
This technique of skin biopsy with PGP 9.5 immunohistochemistry has been demonstrated by ultrastructural studies to label the terminal portions of both small myelinated and unmyelinated nerves in the epidermis (Kennedy and Wendelschafer-Crabb, 1993; Hilliges et al., 1995; Hsieh et al., 1996, 2000). Two issues merit discussion regarding the pathology and biology of sensory nerves in the skin. First, IENF densities recorded by the current quantitation method include normal, pre-degenerating and regenerating epidermal nerves. For normal and pre-degenerating epidermal nerves, there is a significant correlation between changes at the light microscopic level and those under electronic microscopy (Hsieh et al., 2000). Our group previously has demonstrated pathological evidence of pre-degenerating epidermal nerves: swelling and an increased number of branches (Ko et al., 2002). Future ultrastructural studies are required to delineate the relative proportion of normal versus pre-degenerating epidermal nerves, and to investigate the significance of these nerve pathologies (Lauria et al., 2003). In contrast, there are no correlated light and electron microscopic findings in regenerating nerves. Determining this will require further immunohistochemical analyses with different neuronal markers of regenerating nerves, such as growth-associated protein 43 (Woolf et al., 1990).
Secondly, encapsulated sensory receptors, such as Meissner corpuscles and Merkel cells in the human skin, are not labelled with PGP 9.5 immunohistochemistry (Hsieh et al., 2000). It is an open question whether receptors for conducting thermal stimuli are located exclusively in epidermal nerve terminals or in keratinocytes surrounding nerve terminals. This issue will require further studies with new markers against temperature-sensitive molecules, such as transient receptor potential channels (Clapham, 2003); this type of study may elucidate the relationship between clinical symptoms and skin innervation.
Skin innervation and hyperglycaemic control
The present study demonstrates that skin innervation was reduced with increased diabetic duration. This finding supports the concept that long-term impaired glucose metabolism is related to changes in structural organization and function of small-diameter sensory neurons (Perkins et al., 2001). Although previous studies indicated that epidermal nerve fibres and dermal nerve fibre lengths are reduced in diabetes (Kennedy et al., 1996), it has not been demonstrated whether skin denervation was related to diabetic parameters. Beyond diabetic duration, the current study showed no correlation between epidermal innervation and HbA1C levels. This indicates that mechanisms for skin denervation in diabetes may be complicated issues. For example, parameters in the blood are not necessarily correlated with changes at the tissue level, which directly contribute to nerve damage, such as oxidative stress (Feldman, 2003). Alternatively, HbA1C data of 3 years before skin biopsy were available in the current study, and this interval is short compared with a diabetic duration of 10.55 ± 7.20 years of the study population (Ziegler et al., 1991). The limited sample size is also a factor given the dispersed values of HbA1C in the current study group. These latter two issues are important for addressing the effects of glycaemic control (Dyck et al., 1999; Perkins et al., 2001). A larger scale prospective study is necessary, and skin biopsy offers the opportunity to evaluate the influence of diabetes on skin innervation longitudinally, as in a DCCT (Diabetes Control and Complications Trial) study on nerve conduction (Albers et al., 1995).
In addition to metabolic and vascular effects of diabetes, sensory neurons of the small type with their processes depend on various specific neurotrophins for survival and maintenance, particularly nerve growth factor (Snider, 1994; Lentz et al., 1999). In diabetic skin, the transcript level of nerve growth factor is increased (Diemel et al., 1999), while the protein content of nerve growth factor is reduced (Anand et al., 1996). Diabetes reduces the retrograde transport of nerve growth factor, and this effect can be ameliorated by exogenous replacement of the sonic hedgehog protein (Calcutt et al., 2003; Feldman, 2003). These findings suggest that the altered balance of nerve growth factor by modifications at the levels of transcription, translation and post-translation may underlie skin denervation in diabetes, and skin biopsy provides a new approach to test this hypothesis.
Pathology of skin innervation and functional changes
The present report documents parallel psychophysical changes with IENF densities. Previous studies on sural nerves of diabetics indicated that both myelinated and unmyelinated nerve densities were reduced in diabetes (Llewelyn et al., 1991; Malik et al., 2001). However, unmyelinated nerve densities of sural nerves were not correlated with parameters of nerve conduction studies or quantitative sensory testing (Malik et al., 2001). Several possibilities may underlie this discrepancy. These include the presence of autonomic nerves and difficulties in differentiating regenerating nerve sprouts from unmyelinated axons in sural nerve biopsies, which may have influenced the precise counting of unmyelinated nerves (Bickel et al., 2000; Malik et al., 2001).
In the current report, IENF densities, which reflect the abundance of skin innervation, clearly demonstrated linear correlations with thermal thresholds when diabetic duration and age were controlled for. Changes in thermal sensitivities are common in diabetic neuropathy (Dyck et al., 2000). Quantitative sensory testing evaluates the entire pathway of the sensory system, while skin biopsy provides complementary information about sensory nerve terminal changes in diabetic neuropathy.
Most studies on diabetic neuropathy have focused mainly on a certain subtype of neuropathy, and rarely compared the degree of involvement among various subtypes of neuropathy (Albers et al., 1995; Sands et al., 1997). The current report provides laboratory evidence that diabetic neuropathy encompasses both large-fibre and small-fibre neuropathies, and the abnormal rates of small-fibre tests were higher than those of large-fibre tests. Several possibilities may underlie this observation. Patients were at a relatively advanced stage of diabetes in the current series, because all of them had clinically overt diabetes and sensory symptoms. Alternatively, epidermal denervation is a rather early event for a subset of diabetic patients (Sumner et al., 2003). Nevertheless, parallel changes in epidermal innervation and thermal thresholds suggest that skin biopsy together with quantitative sensory testing could potentially play a role in screening and managing diabetic neuropathy. In the management of diabetic neuropathy, different strategies are necessary to target different components of clinical impairment, i.e. large versus small fibres. Given the high proportion of small-fibre sensory neuropathy in the current report, instructions on preventing painless injury due to the reduced sensitivity to thermal stimuli in diabetes need to be stressed.
|
Acknowledgements |
|---|
This work was supported by grants from the National Science Council, Taiwan (NSC 91-2320-B-002-126 and 92-2320-B-002-101).
|
Appendix |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionAppendixReferences |
|---|
|
|
References |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionAppendixReferences |
|---|
Adler AI, Boyko EJ, Ahroni JH, Smith DG. Lower-extremity amputation in diabetes. The independent effects of peripheral vascular disease, sensory neuropathy, and foot ulcers. Diabetes Care 1999; 22: 1029–35.
Albers JW, Kenny DJ, Brown M, Greene D, Cleary PA, Lachin JM, et al. Effect of intensive diabetes treatment on nerve conduction in the Diabetes Control and Complications Trial. Ann Neurol 1995; 38: 869–80.[CrossRef][ISI][Medline]
Albers JW, Brown MB, Sima AA, Greene DA. Nerve conduction measures in mild diabetic neuropathy in the Early Diabetes Intervention Trial: the effects of age, sex, type of diabetes, disease duration, and anthropometric factors. Tolrestat Study Group for the Early Diabetes Intervention Trial. Neurology 1996; 46: 85–91.
American Diabetes Association. Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 1997; 20: 1183–97.[ISI][Medline]
Anand P, Terenghi G, Warner G, Kopelman P, Williams-Chestnut RE, Sinicropi DV. The role of endogenous nerve growth factor in human diabetic neuropathy. Nature Med 1996; 2: 703–7.[CrossRef][ISI][Medline]
Bickel A, Butz M, Schmelz M, Handwerker HO, Neundorfer B. Density of sympathetic axons in sural nerve biopsies of neuropathy patients is related to painfulness. Pain 2000; 84: 413–9.[CrossRef][ISI][Medline]
Calcutt NA, Allendoerfer KL, Mizisin AP, Middlemas A, Freshwater JD, Burgers M, et al. Therapeutic efficacy of sonic hedgehog protein in experimental diabetic neuropathy. J Clin Invest 2003; 111: 507–14.[CrossRef][ISI][Medline]
Chiang HY, Huang IT, Chen WP, Chien HF, Shun CT, Chang YC, et al. Regional difference in epidermal thinning after skin denervation. Exp Neurol 1998; 154: 137–45.[CrossRef][ISI][Medline]
Chien HF, Tseng TJ, Lin WM, Yang CC, Chang YC, Chen RC, et al. Quantitative pathology of cutaneous nerve terminal degeneration in the human skin. Acta Neuropathol (Berl) 2001; 102: 455–61.[Medline]
Clapham DE. TRP channels as cellular sensors. Nature 2003; 426: 517–24.[CrossRef][Medline]
Consensus Development Conference. Proceedings of a consensus development conference on standardized measures in diabetic neuropathy. Electrodiagnostic measures. Neurology 1992; 42: 1827–9.[ISI][Medline]
Cornblath DR, Chaudhry V, Carter K, Lee D, Seysedadr M, Miernicki M, et al. Total neuropathy score: validation and reliability study. Neurology 1999; 53: 1660–4.
Diemel LT, Cai F, Anand P, Warner G, Kopelman PG, Fernyhough P, et al. Increased nerve growth factor mRNA in lateral calf skin biopsies from diabetic patients. Diabetes Med 1999; 16: 113–8.[CrossRef][ISI][Medline]
Dyck PJ, Sherman WR, Hallcher LM, Service FJ, O’Brien PC, Grina LA, et al. Human diabetic endoneurial sorbitol, fructose, and myo-inositol related to sural nerve morphometry. Ann Neurol 1980; 8: 590–6.[CrossRef][ISI][Medline]
Dyck PJ, Davies JL, Wilson DM, Service FJ, Melton LJ, 3rd, O’Brien PC. Risk factors for severity of diabetic polyneuropathy: intensive longitudinal assessment of the Rochester Diabetic Neuropathy Study cohort. Diabetes Care 1999; 22: 1479–86.
Dyck PJ, Dyck PJ, Larson TS, O’Brien PC, Velosa JA. Patterns of quantitative sensation testing of hypoesthesia and hyperalgesia are predictive of diabetic polyneuropathy: a study of three cohorts. Nerve growth factor study group. Diabetes Care 2000; 23: 510–17.[Abstract]
Eaton SE, Harris ND, Ibrahim S, Patel KA, Selmi F, Radatz M, et al. Increased sural nerve epineurial blood flow in human subjects with painful diabetic neuropathy. Diabetologia 2003; 46: 934–9.[CrossRef][ISI][Medline]
Facer P, Mathur R, Pandya SS, Ladiwala U, Singhal BS, Anand P. Correlation of quantitative tests of nerve and target organ dysfunction with skin immunohistology in leprosy. Brain 1998; 121: 2239–47.
Feldman EL. Oxidative stress and diabetic neuropathy: a new understanding of an old problem. J Clin Invest 2003; 111: 431–33.[CrossRef][ISI][Medline]
Griffin JW, McArthur JC, Polydefkis M. Assessment of cutaneous innervation by skin biopsies. Curr Opin Neurol 2001; 14: 655–9.[CrossRef][ISI][Medline]
Hilliges M, Wang L, Johansson O. Ultrastructural evidence for nerve fibers within all vital layers of the human epidermis. J Invest Dermatol 1995; 104: 134–7.[CrossRef][ISI][Medline]
Hirai A, Yasuda H, Joko M, Maeda T, Kikkawa R. Evaluation of diabetic neuropathy through the quantitation of cutaneous nerves. J Neurol Sci 2000; 172: 55–62.[CrossRef][ISI][Medline]
Hsieh ST, Choi S, Lin WM, Chang YC, McArthur JC, Griffin JW. Epidermal denervation and its effects on keratinocytes and Langerhans cells. J Neurocytol 1996; 25: 513–24.[CrossRef][ISI][Medline]
Hsieh ST, Lin WM. Modulation of keratinocyte proliferation by skin innervation. J Invest Dermatol 1999; 113: 579–86.[CrossRef][ISI][Medline]
Hsieh ST, Chiang HY, Lin WM. Pathology of nerve terminal degeneration in the skin. J Neuropathol Exp Neurol 2000; 59: 297–307.[ISI][Medline]
Karanth SS, Springall DR, Kuhn DM, Levene MM, Polak JM. An immunocytochemical study of cutaneous innervation and the distribution of neuropeptides and protein gene product 9.5 in man and commonly employed laboratory animals. Am J Anat 1991; 191: 369–83.[CrossRef][ISI][Medline]
Kennedy WR, Said G. Sensory nerves in skin: answers about painful feet? Neurology 1999; 53: 1614–5.
Kennedy WR, Wendelschafer-Crabb G. The innervation of human epidermis. J Neurol Sci 1993; 115: 184–90.[CrossRef][ISI][Medline]
Kennedy WR, Wendelschafer-Crabb G. Utility of skin biopsy in diabetic neuropathy. Semin Neurol 1996; 16: 163–71.[ISI][Medline]
Kennedy WR, Wendelschafer-Crabb G. Utility of the skin biopsy method in studies of diabetic neuropathy. Electroencephalogr Clin Neurophysiol Suppl 1999; 50: 553–9.
Kennedy WR, Wendelschafer-Crabb G, Johnson T. Quantitation of epidermal nerves in diabetic neuropathy. Neurology 1996; 47: 1042–8.
Ko MH, Chen WP, Hsieh ST. Neuropathology of skin denervation in acrylamide-induced neuropathy. Neurobiol Dis 2002; 11: 155–65.[CrossRef][ISI][Medline]
Lauria G, Morbin M, Lombardi R, Borgna M, Mazzoleni G, Sghirlanzoni A, et al. Axonal swellings predict the degeneration of epidermal nerve fibers in painful neuropathies. Neurology 2003; 61: 631–6.
Lentz SI, Knudson CM, Korsmeyer SJ, Snider WD. Neurotrophins support the development of diverse sensory axon morphologies. J Neurosci 1999; 19: 1038–48.
Levy DM, Karanth SS, Springall DR, Polak JM. Depletion of cutaneous nerves and neuropeptides in diabetes mellitus: an immunocytochemical study. Diabetologia 1989; 32: 427–33.[CrossRef][ISI][Medline]
Lin YH, Huang MH, Chang YC, Tai TY, Chen WH, Yang CC, et al. Quantitative sensory testing: normative values and its application in diabetic neuropathy. Acta Neurol Taiwan 1998; 7: 176–84.
Llewelyn JG, Gilbey SG, Thomas PK, King RH, Muddle JR, Watkins PJ. Sural nerve morphometry in diabetic autonomic and painful sensory neuropathy. A clinicopathological study. Brain 1991; 114: 867–92.
Malik RA, Veves A, Walker D, Siddique I, Lye RH, Schady W, et al. Sural nerve fibre pathology in diabetic patients with mild neuropathy: relationship to pain, quantitative sensory testing and peripheral nerve electrophysiology. Acta Neuropathol (Berl) 2001; 101: 367–74.[Medline]
McArthur JC, Stocks EA, Hauer P, Cornblath DR, Griffin JW. Epidermal nerve fiber density: normative reference range and diagnostic efficiency. Arch Neurol 1998; 55: 1513–20.
McCarthy B, Hsieh ST, Stocks EA, Hauer P, Macko C, Cornblath DR, et al. Cutaneous innervation in sensory neuropathies: evaluation by skin biopsy. Neurology 1995; 45: 1848–55.
McNeely MJ, Boyko EJ, Ahroni JH, Stensel VL, Reiber GE, Smith DG, et al. The independent contributions of diabetic neuropathy and vasculopathy in foot ulceration. How great are the risks? Diabetes Care 1995; 18: 216–9.[Abstract]
Pan CL, Lin YH, Lin WM, Tai TY, Hsieh ST. Degeneration of nociceptive nerve terminals in human peripheral neuropathy. Neuroreport 2001a; 12: 787–92.[CrossRef][ISI][Medline]
Pan CL, Yuki N, Koga M, Chiang MC, Hsieh ST. Acute sensory ataxic neuropathy associated with monospecific anti-GD1b IgG antibody. Neurology 2001b; 57: 1316–8.
Pan CL, Tseng TJ, Lin YH, Chiang MC, Lin WM, Hsieh ST. Cutaneous innervation in Guillain–Barré syndrome: pathology and clinical correlations. Brain 2003; 126: 386–97.
Periquet MI, Novak V, Collins MP, Nagaraja HN, Erdem S, Nash SM, et al. Painful sensory neuropathy: prospective evaluation using skin biopsy. Neurology 1999; 53: 1641–7.
Perkins BA, Greene DA, Bril V. Glycemic control is related to the morphological severity of diabetic sensorimotor polyneuropathy. Diabetes Care 2001; 24: 748–52.
Polydefkis M, Hauer P, Griffin JW, McArthur JC. Skin biopsy as a tool to assess distal small fiber innervation in diabetic neuropathy. Diabetes Technol Ther 2001; 3: 23–8.[CrossRef][Medline]
Polydefkis M, Griffin JW, McArthur J. New insights into diabetic polyneuropathy. J Am Med Assoc 2003; 290: 1371–6.
Properzi G, Francavilla S, Poccia G, Aloisi P, Gu XH, Terenghi G, et al. Early increase precedes a depletion of VIP and PGP-9.5 in the skin of insulin-dependent diabetics—correlation between quantitative immunohistochemistry and clinical assessment of peripheral neuropathy. J Pathol 1993; 169: 269–77.