1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion and conclusions
  7. Author contributions
  8. References


Neuropathy can lead not only to impaired function but also to sensory sensitization. We aimed to link reduced skin nerve fibre density in different levels to layer-specific functional impairment in neuropathic pain patients and tried to identify pain-specific functional and structural markers.


In 12 healthy controls and 36 patients with neuropathic pain, we assessed clinical characteristics, thermal thresholds (quantitative sensory testing) and electrically induced pain and axon reflex erythema. At the most painful sites and at intra-individual control sites, skin biopsies were taken and innervation densities in the different skin layers were assessed. Moreover, neuronal calcitonin gene-related peptide staining was quantified.


Perception of warm, cold and heat pain and nerve fibre density were reduced in the painful areas compared with the control sites and with healthy controls. Warm and cold detection thresholds correlated best with epidermal innervation density, whereas heat and cold pain thresholds and axon reflex flare correlated best with dermal innervation density. Clinical pain ratings correlated only with epidermal nerve fibre density (r = 0.38, p < 0.05) and better preserved cold detection thresholds (r = 0.39, p < 0.05), but not with other assessed functional and structural parameters.


Thermal thresholds, axon reflex measurements and assessment of skin innervation density are valuable tools to characterize and quantify peripheral neuropathy and link neuronal function to different layers of the skin. The severity of small fibre neuropathy, however, did not correspond to clinical pain intensity and a specific parameter or pattern that would predict pain intensity in peripheral neuropathy could not be identified.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion and conclusions
  7. Author contributions
  8. References

Clinical characteristics of neuropathy comprise reduced sensory and motor function; however, only a minority of the patients with central or peripheral nervous system lesions develop pain. The mechanisms that determine whether peripheral nerve lesions lead to a pure loss of function or are accompanied by chronic pain have not been identified. In order to clarify which mechanisms are involved, structural and functional patterns of small fibre function in neuropathic pain patients have been investigated using quantitative sensory testing (QST) (Dyck et al., 2000; Jaaskelainen et al., 2005; Rolke et al., 2006; Aasvang et al., 2008; Freynhagen et al., 2008; Maier et al., 2010), quantification of axon reflex responses (Novak et al., 2001; Bickel et al., 2002; Kramer et al., 2004b) and skin biopsies (Oaklander, 2001; Polydefkis et al., 2002; Sorensen et al., 2006; Devigili et al., 2008; Oaklander, 2008; Vlckova-Moravcova et al., 2008; Ho et al., 2009). Early studies (Oaklander, 2001; Polydefkis et al., 2002; Sorensen et al., 2006) suggested that reduced epidermal innervation density might predict neuropathic pain intensity; however, recent studies challenged these findings. Although intra-epidermal fibre density (assessed histologically) is an objective marker of small fibre neuropathy (Devigili et al., 2008; Oaklander, 2008; Ho et al., 2009; Cruccu et al., 2010), no consistent correlation of epidermal nerve fibre density and pain intensity has been detected (Shun et al., 2004; Devigili et al., 2008; Landerholm et al., 2010; Boyette-Davis and Dougherty, 2011). Thus, the mechanisms underlying the development of pain in neuropathy remain unclear.

What's already known about this topic?
  • Neuronal dysfunction is a prerequisite of neuropathic pain, but it is unclear how the degree of structural and functional neuronal impairment is linked to the intensity of neuropathic pain.
What does this study add?
  • Impaired skin innervation reduced neuronal functions (thermal thresholds, axon reflex erythema and electrically induced pain) in a layer-specific way, but was not significantly correlated with the levels of neuropathic pain.

Therefore, we clinically investigated patients with neuropathic pain and assessed different functional (QST, axon reflex erythema) and structural parameters (skin biopsy). Innervation density was quantified in the epidermis, the sub-epidermal plexus and in dermal layers. We aimed to identify significant correlations between neuronal function and innervation density in different skin layers and specifically between pain intensity and nerve fibre pathology. We hypothesized that changes most relevant to pain development should correlate with the clinical pain level.

2. Methods

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion and conclusions
  7. Author contributions
  8. References

2.1 Patients and healthy controls

Patients (n = 36) and controls (n = 13) were investigated in the clinics of Anaesthesiology (Mannheim, University of Heidelberg, Germany) and Neurology (University Medical Center, Mainz, Germany) after they had given their informed consent with approval of the ethics committees of the universities of Mainz and Mannheim. Patients with chronic neuropathic pain of at least 3/10 visual analogue scale (VAS) for more than 3 months were recruited in the outpatient pain clinics. Patients with impaired haemostasis, patients with an impaired wound healing at the biopsy site to be expected and pregnant women were excluded.

2.2 Pain assessment

All patients filled in the Brief Pain Inventory Short Form (Daut et al., 1983) in the outpatient clinics. We did not assess pain ratings by a diary. Pain intensity was rated using an 11-point scale [0–10 numeric pain rating (NRS)], and peak pain intensity in the last 4 weeks was analysed for the correlations.

2.3 Clinical investigation

Each patient was examined by a physician in order to assess neuropathic symptoms as well as signs and symptoms of neuropathic pain, such as sensory gain (allodynia: camel's hair brush 0.5, Somedic, Sollentuna, Sweden; hyperalgesia to pin-prick stimuli, toothpick) and sensory loss of function (tactile hypoesthesia, hypalgesia to pin-prick stimuli). The most painful skin areas were identified and used as test sites. For suitable locations, local anaesthetic cream (EMLA®, AstraZeneca, Soedertaelje, Sweden) was applied topically under occlusion for 2 h, covering the most painful areas, and pain ratings were assessed at 60-min interval.

2.4 Quantitative sensory testing (QST)

Thresholds for warmth, cold, heat pain and cold pain were investigated and intensity ratings at thresholds were obtained using a standard QST device (Thermotest Somedic, Horby, Sweden). The baseline temperature was 32 °C, the slope of temperature change was 1 °C/s and the thermode size was 12.5 cm2. The matching contralateral skin sites were tested as intra-individual controls. For those patients with systemic polyneuropathy, a proximal test location on the same limb was used as the control site. Tests on symptomatic and control skin were performed in randomized order. Intensity ratings were obtained for the cold- and heat-pain test (Kelly et al., 2005).

2.5 Electrically induced axon reflex flare

A pair of adhesive surface electrodes (3 × 10 mm) was attached in parallel at a distance of 2 mm at the centre of the QST test sites. Through these electrodes, electrical stimulation (1 Hz, 0.5 ms duration) was applied via a constant current stimulator (Digitimer DS-7, Welwyn Garden, UK). Current intensity was stepwise increased at 3-min interval (2.5, 5, 10, 15 and 20 mA). Pain ratings were assessed after the new current intensity had been applied and before switching to the next stimulus intensity.

Superficial blood flow was quantified using a laser-Doppler imager (LDI; Moor, Axminster, UK). LDI scans (256 × 256 pixels; scan resolution: 4 pixels/s; distance: 50 cm to the site of electrical stimulation) were recorded at baseline and at 1.5-min interval during the electrical stimulation (two scans per current intensity). Blood flux was calculated for each pixel by means of intensity of the Doppler shift of the backscattered laser light (arbitrary perfusion units; PU). Thereby, a two-dimensional map of the superficial blood flow was generated. Area and intensity of the neurogenic flare reaction were analysed offline by dedicated software (MLDI 5.3, Moor). Flare area was determined by the total number of pixels in which flux values exceeded the mean flux by two standard deviations from the baseline picture.

2.6 Skin biopsies

Skin biopsies were taken from the most painful skin sites identified during clinical investigation and from control sites after sensory testing by a physician under local anaesthesia (2% lidocaine). Skin biopsies containing epidermal and dermal compartments were harvested using a 3-mm-diameter skin punch (Stiefel Laboratorium, Offenbach, Germany). Biopsies were immediately fixed by immersion into cold 4% paraformaldehyde (Sigma, Deisenhofen, Germany) for at least 4 h and then transferred into 1% phosphate buffered saline (PBS, Sigma) containing 0.01 NaN3 (sodium azide, Sigma). The fixed tissue samples were stored at 4 °C and shipped on ice directly to the analysis site.

2.7 Immunohistochemistry

Upon receipt, each biopsy was rinsed with fresh PBS and cryoprotected by immersion overnight in 30% sucrose in PBS. Biopsies were subsequently placed in optimal cutting temperature (OCT) compound, frozen and sectioned by cryostat. Serial sections of 14 μm were quick thaw mounted in consecutive order and rotated across a 10-slide series so that approximately 15–20 sections taken at equal intervals throughout each biopsy were obtained per slide. Selected slides were then processed for immunofluorescent detection utilizing anti-PGP9.5 [PGP] (rb polyclonal, 1:800; UltraClone Ltd., Isle of Wight, England), anti-200kD neurofilament protein [NF] (rb polyclonal, 1:800; Millipore, Billerica, MA, USA) and anti- calcitonin gene-related peptide (CGRP) (gp polyclonal, 1:400; Peninsula Labs, San Carlos, CA, USA) alone and in combination, and counterstained with DAPI (Sigma, St. Louis, MO, USA). Secondary antibodies directed against the appropriate species of primary antibody were produced in donkey or goat and conjugated either with Cy3 (1:500 dilution, Jackson ImmunoResearch Laboratories, West Grove, PA, USA) for red immunofluorescence or Alexa 488 (1:250–500, Molecular Probes Inc., Eugene, OR, USA) for green immunofluorescence. Detailed procedures for double labelling and image capture are described in prior publications (Pare et al., 2001; Albrecht et al., 2006).

2.8 Data analysis and statistics

Immunostained sections were analysed with an Olympus Optical Provis AX70 microscope equipped with conventional fluorescence filters (Cy3: 528–553 nm excitation, 590–650 nm emission; Cy2/Alexa488: 460–500 nm excitation, 510–560 nm emission; DAPI/Hoechst: 375–400 nm excitation, 450–475 emission). Fluorescent images were collected with a high-resolution camera (DKC-ST5, Sony, Montvale, NJ, USA) interfaced with Northern Eclipse (Empix Imaging, Mississauga, ON, Canada), Photoshop (Adobe, San Jose, CA, USA) and NeuroLucida (MicroBrightField, Colchester, VT, USA) software. Intra-epidermal, sub-epidermal and dermal nerve fibre counts were obtained from three complete sections per specimen and are presented as averages ± standard error of the mean. Procedures for generating image montages and quantifying epidermal, sub-epidermal and upper dermal innervation density are described in prior publications (Albrecht et al., 2006; Petersen et al., 2010). Epidermal innervation was based on those immunolabelledprofiles located entirely within the epidermis. Sub-epidermal innervation was limited to those immunolabelled profiles that were immediately subjacent to and in contact with the basement membrane (i.e., the sub-epidermal plexus). Upper dermal innervation was limited to those immunolabelled profiles in the dermis to a depth of ∼75 μm deep to the epidermis but not including those profiles in the sub-epidermal plexus. The term ‘labelled profiles’ in the sub-epidermis and upper dermis include, but do not specifically distinguish between, profiles that are likely individual axons and possibly terminals and such that may contain a few axons (i.e., small nerves). The densities of epidermal, sub-epidermal and dermal innervation are expressed as the number of profiles per millimetre of epidermal length. In contrast to published guidelines that recommend a 50-μm-section thickness (Lauria et al., 2010), our preference is 14 μm, which provides more sections and slides per biopsy in order to increase options for multi-molecular immunolabelling combinations and archiving sections for future retrospective analyses (Bowsher et al., 2009; Rice et al., 2010). Therefore, our absolute values of epidermal innervation density are lower as compared with the values published on the basis of 50-μm sections. CGRP and neurofilament staining were evaluated semi-quantitatively. Recognizing limitations to this approach [16], proportions of CGRP and NF immunolabelled profiles relative to total innervation labeled with anti-PGP were rated using a 6-point rating scale: from 0 to 5, where ‘0’ meant staining is absent, ‘1' = depleted (∼25% of normal), ‘2' = diminished (∼50% of normal), ‘3' = limited (∼75% of normal), ‘4' = normal, ‘5' = exuberant. Photographs of the analysed specimens were grouped according to the staining levels into libraries to assure comparability (Fig. 1)


Figure 1. PGP 9.5 staining (red fluorescence; blue fluorescence: DAPI) of a representative specimen exemplifying the locations of intra-epidermal, sub-epidermal and upper dermal nerve fibres marked with red arrows.

Download figure to PowerPoint

Differences between patient groups and diseased vs. contralateral controls were calculated using Mann–Whitney tests. Fisher's exact tests were used to compare percentages. Spearman's rank order correlations were calculated between innervation density or staining proportion and functional data. Differences were considered statistically significant at values of p < 0.05. Since this is an exploratory hypothesis-generating study, statistics were intentionally not controlled for multiple testing.

3. Results

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion and conclusions
  7. Author contributions
  8. References

3.1 Patients

Nineteen patients (8 women, 11 men; mean age 61.4 years) suffered from painful localized neuropathy (peripheral nerve trauma, n = 13; complex regional pain syndrome II (CRPS II), n = 3, post-herpetic neuralgia, n = 3). Seventeen patients (6 women, 11 men; mean age 60.8 years) had painful symmetric neuropathy (diabetes, n = 6; chemotherapy-induced, n = 4; amyloidosis, n = 3; idiopathic small fibre polyneuropathy, n = 4). Patients with localized nerve pathology had higher peak pain ratings (7.2 ± 0.4) than patients with systemic neuropathy (5.4 ± 0.4; numeric rating scale, p < 0.01). The temporal pattern of pain did not differ between the patient groups as all patients reported constant pain or several attacks per day. About half of the patients with symmetric neuropathy (9 out of 17) and localized neuropathy (10 out of 19) had additional attacks in addition to their constant pain. No significant differences were found in the analgesic treatment of the patients with opioids (53% vs. 57%; symmetric vs. localized), gabapentin/pregabalin (53% vs. 57%), non-steroidal anti-inflammatory drugs (37% vs. 43%) and antidepressants (16% vs. 20%). Topical application of local anaesthetic cream reduced spontaneous pain by more than 50% in six out of nine patients (from median 6 to 2 VAS) and left it unchanged in three others (from median 6 to 6 VAS). There were no obvious differences between the two groups concerning temporal pattern of pain or medication. Healthy controls (five women, seven men) were 32.8 ± 2.3 years old.

3.2 Quantitative sensory testing

In both patient groups, impaired perception of thermal detection and heat pain thresholds was observed at the symptomatic skin sites compared with the intra-individual control site (Table 1, control vs. diseased site: ‘vs. dis.’). Compared with patients with systemic neuropathy, patients with localized nerve lesions had significantly better preserved detection of warm and cold stimuli, and also better preserved heat pain detection at the diseased skin site (Table 1, localized vs. systemic neuropathy: ‘loc./syst.’). In contrast, temperature thresholds at the control sites did neither differ significantly between the two patient groups nor between patients and healthy controls (tests not shown in the table).

Table 1. Results of quantitative sensory testing [temperatures in °C (relative to 32 °C for warm and cold thresholds], ratings on numeric rating scale 0–10, median (interquartile range)] – data of the patient groups (localized neuropathy vs. systemic neuropathy)
 Healthy controlsControl siteDiseased site
Localizedvs. loc. dis.Systemicvs. syst. dis.Localizedloc./syst.Systemic
  1. Mann–Whitney U-tests (*p < 0.05, **p < 0.01, ***p < 0.005) were used to compare patients' control versus diseased site for localized (‘vs. loc. dis’) and systemic (‘vs. syst. dis’) neuropathy patients. At the diseased sites values were compared between the two patients groups (‘loc./syst.’) by the same test. n.d., not done.

Warm threshold2.4 (3.3)2.0 (2.6) ** 3.2 (6.9) *** 4.5 (7.1) ** 11.2 (5.8)
Cold threshold−3.4 (3)−4.4 (3) * −3.1 (2.6) *** −6.4 (4.0) −10.7 (18.3)
Heat pain threshold45.0 (4.1)43.0 (2) * 42.3 (3) * 45.3 (4.6) * 47.3 (2.75)
Heat pain ratingn.d.6 (3) 7.5 (5) 8 (3) 6 (4)
Cold pain threshold10 (2)13.4 (11) 13.8 (18) 16.3 (11.7) ** 10.0 (6.3)
Cold pain ratingn.d.2 (6) 3.8 (7) 5 (8) 4 (6)

3.3 Cold allodynia, touch-evoked allodynia and punctate hyperalgesia

In 8 out of 19 patients with localized neuropathy, cold allodynia with cold pain thresholds >20 °C was detected, whereas only 2 out of 17 patients with systemic neuropathy showed cold allodynia. Similarly, brush-evoked allodynia was more abundant in patients with localized (12 out of 19) as compared with systemic neuropathy (4 out of 17; p = 0.03, Fisher's exact test). Punctate hyperalgesia was found in 12 out of 19 localized nerve lesion patients and in 7 out of 17 neuropathy patients (n.s.; Fisher exact test).

3.4 Electrically induced pain and axon reflex erythema

Electrical stimulation intensity dependently provoked pain and an axon reflex erythema in the vicinity of the stimulation electrodes. Current-evoked pain did not differ significantly between patient groups, neither did it differ significantly between diseased and control sites nor between patients and healthy controls. The area of axon reflex erythema was reduced at the diseased site compared with the control site in patients with localized (7 ± 1.7 cm2 vs. 13.6 ± 5.1 cm2, p < 0.01) and systemic neuropathic pain (4.1 ± 0.9 cm2 vs. 13.9 ± 3.9 cm2, p < 0.01). No significant differences were observed for the flare responses at the control sites of patients and healthy volunteers (17.9 ± 3.1 cm2).

3.5 Nerve fibre density

Intra-epidermal fibre density was reduced at the painful skin site in both patient groups compared with the intra-individual control site (see Table 2), whereas the reduction of sub-epidermal and dermal fibre density at the painful site was only significant in patients with systemic neuropathy. When comparing innervation densities between the patient groups at the painful site, reduced density was observed for the patients with systemic neuropathy that was statistically significant for the sub-epidermal innervation.

Table 2. Innervation density in skin biopsies (fibres/mm epidermis, mean ± standard error of the mean) and semi-quantitative staining of neuronal calcitonin gene-related peptide (CGRP) and neuronal neurofilament [median (quartiles), see Methods section] – data of the patient groups (localized neuropathy vs. systemic neuropathy)
Nerve fibre densityHealthy controlsControl siteDiseased site
Localizedvs. loc. disSystemicvs. syst. dis.LocalizedLoc./syst.Systemic
  1. Mann–Whitney U-tests (*p < 0.05, **p < 0.01) were used to compare patients′ control versus diseased site for localized (‘loc./loc.’) and systemic (‘syst./syst.’) neuropathy and the diseased sites of localized and system neuropathy (‘loc./syst.’). CGRP and neurofilament staining is given semi-quantitatively on a 6-point rating scale [0: absent, 1: depleted (∼25% of normal), 2: diminished (∼50% of normal), 3: limited (∼75% of normal), 4: normal, 5: exuberant].

Epidermal3.36 ± 0.622.28 ± 0.41 * 2.94 ± 0.79 ** 1.52 ± 0.37 0.77 ± 0.27
Sub-epidermal2.73 ± 0.461.84 ± 0.36 2.51 ± 1.01 * 1.37 ± 0.28 * 0.57 ± 0.20
Dermal8.67 ± 0.567.31 ± 0.77 8.52 ± 1.33 * 6.46 ± 0.63 4.62 ± 1.10
Neuronal CGRP
Epidermal0 (0–0)0 (0–1) 0 (0–1) 0 (0–0) 0 (0–0)
Sub-epidermal3 (2–3.5)2 (1–3) 2 (1–2) 2 (1–2) 1 (0–2)
Dermal4 (3–4)4 (2–4) 4 (4–4) * 2 (1–4) 2 (2–3)
Epidermal0 (0–0)0 (0–1) 1 (0–1) 0 (0–1) 0 (0–1)
Sub-epidermal3.5 (3–4)3 (1–3) * 3 (2–3) * 1 (0–2) 1 (1–2)
Dermal4 (4–4)3 (3–3) 3 (3–3) * 3 (1.5–3) 2 (1–2)

Innervation density at the control sites did not differ significantly between the two patient groups (tests not shown).

3.6 Neuronal staining for CGRP and neurofilament

At the painful sites, neuronal CGRP immunolabelling was significantly reduced in the dermal layers compared with the intra-individual control site for the systemic neuropathy patients (see Table 2). No significant differences in neuronal CGRP labelling were observed between the control sites of the two patient groups or between control sites and healthy volunteers (specimen in Fig. 2).


Figure 2. PGP 9.5 (upper panel) and calcitonin gene-related peptide (CGRP)/PGP double staining (lower panels) of representative specimen. Epidermal fibre loss is evident in a patient (nerve fibres marked by open arrows) with traumatic nerve injury (upper left) as compared with the healthy control site (upper right). Some of the dermal fibres positive for PGP9.5 (green fluorescence, green arrows) also stain for CGRP (red fluorescence, yellow arrows; blue fluorescence: DAPI). Sub-epidermal CGRP staining appeared slightly reduced in traumatic nerve injury (lower left) as compared with contralateral control (lower right).

Download figure to PowerPoint

200 kD NF, a marker for myelinated fibres (Albrecht et al., 2006), was reduced in the sub-epidermal layer at the painful site of both patient groups and did not differ significantly in the epidermal layer. NF labelling at the control sites did not differ between patient groups, but was reduced in the dermal layers compared with healthy controls.

3.7 Correlations between structural and functional data

In symptomatic skin, epidermal and sub-epidermal nerve fibre density only correlated with preserved cold pain threshold (Spearman's r = 0.43, r = 0.44; p < 0.05) (Table 3, lower right numerals in the cells). When adding the control biopsies of the patients to the analysis (Table 3, upper left numerals in the cells), all temperature thresholds correlated with epidermal nerve fibre density. Sub-epidermal fibre density correlated best with cold pain threshold (r = 0.49; p < 0.01) and the deeper dermal innervation with heat pain thresholds (r = 0.42; p < 0.01; Fig. 3, Table 3). More interestingly, innervation density of deeper skin layers correlated best with pain thresholds, whereas the highest correlations to cold and warm detection were found for the superficial epidermal innervation density (Table 3).


Figure 3. Scatterplots of innervation densities and thermal thresholds are shown. Epidermal innervation density decreased with impaired warm detection (upper left, Spearman's rank correlation: r = −0.43; p < 0.01). Electrically induced pain (cumulative pain ratings to 2.5–20 mA; NRS = numeric rating scale) did not correlate with upper dermal innervation density assessed by PGP (lower left panel; Spearman's r = 0.04, n.s.), but with the specific neuronal CGRP staining in the dermis (lower right; r = 0.49; p < 0.005).

Download figure to PowerPoint

Table 3. Spearman's rank correlations between functional and structural markers of neuropathy and pain levels (peak pain in the last 4 weeks).Thumbnail image of

The area of the electrically evoked axon reflex flare correlated with CRPG staining of dermal (r = 0.4; p < 0.05), but not of epidermal or sub-epidermal fibres. Similarly, pain ratings to strong electrical pulses (up to 20 mA) also correlated with CGRP staining intensity of dermal fibres (r = 0.51; p < 0.01), but not epidermal innervation density (n.s.; Fig. 3, Table 3).

3.8 Correlations of functional and structural data to pain ratings

Among the pain patients, higher peak pain ratings correlated with preserved cold detection (Spearman's rank correlation: r = 0.39; p = 0.016) and higher epidermal nerve fibre density (r = 0.38; p = 0.026). Healthy controls had a higher innervation density and no pain; thus, merging the data of patients and controls would falsely suggest a negative correlation between pain ratings and innervation density (r = −0.28; p = 0.027) (Fig. 4A), between pain ratings and axon reflex flare (r = −0.36; p = 0.007) (Fig. 4B), and between pain and warm thresholds (r = 0.30; p = 0.019) (Fig. 4C).


Figure 4. Scatterplots showing numeric pain ratings (0–10) and epidermal fibre density (A), electrically induced axon reflex area (B), warm detection threshold (C) and cold detection threshold (D). In the group of pain patients (solid symbols), there was a positive correlation between pain and epidermal innervation density (Spearman's rank correlation: r = 0.38; p = 0.026) and between pain and cold detection threshold (r = 0.39; p = 0.016).

Download figure to PowerPoint

4. Discussion and conclusions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion and conclusions
  7. Author contributions
  8. References

In our study, functional and structural impairments of nerve fibres in chronic pain patients with localized and systemic neuropathies were quantified. In diseased skin, perception of warm, cold and heat pain was impaired. Moreover, the spatial extent of the axon reflex flare erythema and the epidermal innervation density were reduced compared with the intra-individual control site, or compared with healthy control subjects. Thus, functional and structural parameters of primary afferent nerve fibres confirmed neuropathy in the diseased skin. Impairment of cold and warm detection correlated best with the superficial epidermal nerve fibre density, while heat and cold pain thresholds, pain ratings to electrical pulses and the axon reflex erythema correlated with deeper layer innervation. Clinical pain intensity was found higher in those patients with a milder reduction of epidermal fibre density and better preserved cold thresholds; however, no pain-specific biomarker was identified in the patients.

4.1 Localized versus symmetric neuropathy

Recent studies have suggested that neuropathic pain patients should be grouped according to their symptom pattern rather than being classified according to the underlying aetiology or disease localization (Attal et al., 2008; Baron et al., 2009; Scholz et al., 2009; Maier et al., 2010). Moreover, phenotypic profiles based on sensory testing or pain descriptors might differentiate patients such that therapeutic prediction can be improved (Attal et al., 2011). In our study, we differentiated localized and symmetric neuropathic pain with the control sites being contralaterally for the localized pain and proximal thigh in distal symmetric neuropathies. Still we cannot exclude neuropathic changes at the control sites. Patients with localized pain had higher clinical pain ratings and more frequently showed cold- and touch-evoked allodynia. In patients with systemic neuropathy, we found a stronger impairment of temperature detection and heat pain, associated with a more profound fibre loss. This suggests that skin denervation might be protective against allodynia.

4.2 Innervation density

Epidermal fibre density was reduced in the symptomatic skin areas and thus proved to be an objective parameter for neuropathy (McArthur et al., 1998). We found a weak positive correlation between epidermal nerve fibre density and clinical pain ratings. Apparently, a more pronounced skin fibre loss does not necessarily lead to more intense pain, as also reported in a recent study on post-herpetic neuralgia (Petersen et al., 2010).

Assessment of skin innervation density obviously verifies the degree of small fibre neuropathy (Attal et al., 2008; Devigili et al., 2008; Oaklander, 2008; Ho et al., 2009), even though it does not seem to be a reliable biomarker of pain intensity in these patients.

4.3  CGRP staining

CGRP is a marker for peptidergic afferent nerve fibres. CGRP has also been shown to be involved in the generation of migraine pain, and mechanical (Khodorova et al., 2009) and heat hyperalgesia (Mogil et al., 2005). Moreover, CGRP has been shown to upregulate keratinocyte nerve growth factor expression (Dallos et al., 2006), a neurotrophin linked to nociceptor sensitization (Rukwied et al., 2010) and chronic pain (Lane et al., 2010). Therefore, CGRP staining might correlate with the clinical pain level or sensitization. Indeed, we found that heat pain thresholds correlated with higher neuronal CGRP staining in the sub-epidermal skin layer. This confirms that the CGRP-positive sub-population of C-fibres is particularly linked to heat pain thresholds and therefore to peripheral sensitization.

4.4 Layer-specific correlations of fibre density and sensory function

Reduced epidermal innervation density correlated with increased warm and cold detection thresholds, whereas reduced deeper dermal skin innervation density correlated with increased heat and cold pain thresholds, and smaller axon reflex erythema, as shown earlier (Bickel et al., 2009). It should be noted that among healthy volunteers, variations of the normal epidermal innervation density are not correlated with the corresponding changes of temperature thresholds (Selim et al., 2010), indicating that a massive reduction of innervation densities is required before functional impairment becomes apparent (Kennedy et al., 2010). Moreover, the younger age of our healthy controls limit a direct comparison to our pain patients. Electrically induced pain ratings rather correlated with the density of CGRP-positive dermal nerve fibres than with the pan-neuronal structural marker PGP 9.5, suggesting that CGRP-positive nociceptors play a prominent role in encoding the painfulness of strong electrical pulses. These results are in line with the intuitive concept that mild thermal stimuli should be detected by sensory endings in the uppermost skin layers, whereas strong stimuli would also penetrate deeper skin layers. Moreover, the present results confirm that the correlations between functional and structural data in the patients are coherent.

More severe neuropathy as assessed by innervation densities was linked to impaired sensory function. However, the link between neuropathy and pain – one of the primary objectives of this study – was found to be more complex. In our exploratory analysis, the correlations to pain were weak, such that controlling for multiple comparisons (see Table 3) would leave them non-significant. Moreover, the correlation between the level of ongoing neuropathic pain and better preserved epidermal innervation density and cold detection thresholds clearly speaks against a simple process that would simultaneously cause neuropathy and pain. Similar results were found following herniotomy (Aasvang et al., 2008) or mastectomy (Gottrup et al., 2000), with warm detection being better preserved in pain patients compared with operated control patients without pain.

In line with our results, studies on traumatic nerve injury patients verified impaired nerve fibre function (QST, axon reflex) and skin innervation density (Kalliomaki et al., 2011; Wildgaard et al., 2012). Most importantly, these studies did not detect differences in the structural or functional impairment between patients with and without pain.

Contrasting our results, diabetic neuropathy pain patients were reported to have lower epidermal innervation densities (Sorensen et al., 2006). Moreover, higher pain ratings correlated with more pronounced impairment of temperature sensitivity in advanced painful diabetic neuropathy patients (Kramer et al., 2004a). In these patients, diabetic polyneuropathy was advanced and pain scores were obtained as an average of 2 years, yielding a median pain level of only 3.5/10.

In conclusion, functional and structural neuronal markers in patients with neuropathic pain can be used to assess the degree of small fibre impairment. Coherent correlations between functional and structural markers indicate skin layer-specific sensory functions. While severity of small fibre impairment increases with decreased epidermal innervation density, the level of neuropathic pain was higher in patients with better preserved epidermal innervation. Thus, different mechanisms determine the severity of neuropathic degeneration of small fibres and the intensity of neuropathic pain. The analysis of neuronal markers, such as sodium channel subtypes in painful neuroma (Black et al., 2008), appears to be a more promising approach to characterize the mechanisms that turn neuropathy into neuropathic pain. However, given that the clinical pain level results from a complex neuronal interaction between primary afferent nociceptors, processing in spinal cord and brain with numerous counter-regulating circuits, it remains to be shown to which extent peripheral expression of specific functional markers on nociceptors contribute to the level of clinical pain.

Author contributions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion and conclusions
  7. Author contributions
  8. References

M. Schley, patient selection, investigation of subjects and patients, paper writing.

A. Bayram, investigated subjects and patients, paper commenting.

R. Rukwied, patient investigation, data interpretation, paper commenting.

M. Dusch, patient selection, patient investigation, paper commenting.

C. Konrad, patient selection, data interpretation, paper commenting.

J. Benrath, patient selection, data interpretation, paper commenting.

C. Geber, patient selection, investigation of patients, data interpretation, paper writing.

F. Birklein, patient selection, investigation of patients, data interpretation, paper writing.

B. Hägglöf, study design, data interpretation, paper commenting.

N. Sjögren, statistical analysis, data interpretation, paper commenting.

L. Gee, immunohistochemistry and data analysis.

P.J. Albrecht, study design, immunohistochemistry and data analysis, data interpretation, paper commenting.

F.L. Rice, study design, immunohistochemistry and data analysis, data interpretation, paper commenting.

M. Schmelz study design, data interpretation, paper writing.


  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion and conclusions
  7. Author contributions
  8. References
  • Aasvang, E.K., Brandsborg, B., Christensen, B., Jensen, T.S., Kehlet, H. (2008). Neurophysiological characterization of postherniotomy pain. Pain 137, 173181.
  • Albrecht, P.J., Hines, S., Eisenberg, E., Pud, D., Finlay, D.R., Connolly, M.K., Pare, M., Davar, G., Rice, F.L. (2006). Pathologic alterations of cutaneous innervation and vasculature in affected limbs from patients with complex regional pain syndrome. Pain 120, 244266.
  • Attal, N., Bouhassira, D., Baron, R., Dostrovsky, J., Dworkin, R.H., Finnerup, N., Gourlay, G., Haanpaa, M., Raja, S., Rice, A.S., Simpson, D., Treede, R.D. (2011). Assessing symptom profiles in neuropathic pain clinical trials: Can it improve outcome? Eur J Pain 15, 441443.
  • Attal, N., Fermanian, C., Fermanian, J., Lanteri-Minet, M., Alchaar, H., Bouhassira, D. (2008). Neuropathic pain: Are there distinct subtypes depending on the aetiology or anatomical lesion? Pain 138, 343353.
  • Baron, R., Tolle, T.R., Gockel, U., Brosz, M., Freynhagen, R. (2009). A cross-sectional cohort survey in 2100 patients with painful diabetic neuropathy and postherpetic neuralgia: Differences in demographic data and sensory symptoms. Pain 146, 3440.
  • Bickel, A., Heyer, G., Senger, C., Maihofner, C., Heuss, D., Hilz, M.J., Namer, B. (2009). C-fiber axon reflex flare size correlates with epidermal nerve fiber density in human skin biopsies. J Peripher Nerv Syst 14, 294299.
  • Bickel, A., Kramer, H.H., Hilz, M.J., Birklein, F., Neundorfer, B., Schmelz, M. (2002). Assessment of the neurogenic flare reaction in small-fiber neuropathies. Neurology 59, 917919.
  • Black, J.A., Nikolajsen, L., Kroner, K., Jensen, T.S., Waxman, S.G. (2008). Multiple sodium channel isoforms and mitogen-activated protein kinases are present in painful human neuromas. Ann Neurol 64, 644653.
  • Bowsher, D., Geoffrey, W.C., Nicholas, A.K., Carvalho, O.M., Haggett, C.E., Tedman, B., Mackenzie, J.M., Crooks, D., Mahmood, N., Twomey, J.A., Hann, S., Jones, D., Wymer, J.P., Albrecht, P.J., Argoff, C.E., Rice, F.L. (2009). Absence of pain with hyperhidrosis: A new syndrome where vascular afferents may mediate cutaneous sensation. Pain 147, 287298.
  • Boyette-Davis, J., Dougherty, P.M. (2011). Protection against oxaliplatin-induced mechanical hyperalgesia and intraepidermal nerve fiber loss by minocycline. Exp Neurol 229, 353357.
  • Cruccu, G., Sommer, C., Anand, P., Attal, N., Baron, R., Garcia-Larrea, L., Haanpaa, M., Jensen, T.S., Serra, J., Treede, R.D. (2010). EFNS guidelines on neuropathic pain assessment: Revised 2009. Eur J Neurol 17, 10101018.
  • Dallos, A., Kiss, M., Polyanka, H., Dobozy, A., Kemeny, L., Husz, S. (2006). Effects of the neuropeptides substance P, calcitonin gene-related peptide, vasoactive intestinal polypeptide and galanin on the production of nerve growth factor and inflammatory cytokines in cultured human keratinocytes. Neuropeptides 40, 251263.
  • Daut, R.L., Cleeland, C.S., Flanery, R.C. (1983). Development of the Wisconsin Brief Pain Questionnaire to assess pain in cancer and other diseases. Pain 17, 197210.
  • Devigili, G., Tugnoli, V., Penza, P., Camozzi, F., Lombardi, R., Melli, G., Broglio, L., Granieri, E., Lauria, G. (2008). The diagnostic criteria for small fibre neuropathy: From symptoms to neuropathology. Brain 131, 19121925.
  • Dyck, P.J., Dyck, P.J., Larson, T.S., O'Brien, P.C., Velosa, J.A. (2000). 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 23, 510517.
  • Freynhagen, R., Rolke, R., Baron, R., Tolle, T.R., Rutjes, A.K., Schu, S., Treede, R.D. (2008). Pseudoradicular and radicular low-back pain – a disease continuum rather than different entities? Answers from quantitative sensory testing. Pain 135, 6574.
  • Gottrup, H., Andersen, J., Arendt-Nielsen, L., Jensen, T.S. (2000). Psychophysical examination in patients with post-mastectomy pain. Pain 87, 275284.
  • Ho, T.W., Backonja, M., Ma, J., Leibensperger, H., Froman, S., Polydefkis, M. (2009). Efficient assessment of neuropathic pain drugs in patients with small fiber sensory neuropathies. Pain 141, 1924.
  • Jaaskelainen, S.K., Teerijoki-Oksa, T., Forssell, H. (2005). Neurophysiologic and quantitative sensory testing in the diagnosis of trigeminal neuropathy and neuropathic pain. Pain 117, 349357.
  • Kalliomaki, M., Kieseritzky, J.V., Schmidt, R., Hagglof, B., Karlsten, R., Sjogren, N., Albrecht, P., Gee, L., Rice, F., Wiig, M., Schmelz, M., Gordh, T. (2011). Structural and functional differences between neuropathy with and without pain? Exp Neurol 231, 199206.
  • Kelly, K.G., Cook, T., Backonja, M.M. (2005). Pain ratings at the thresholds are necessary for interpretation of quantitative sensory testing. Muscle Nerve 32, 179184.
  • Kennedy, W.R., Vanhove, G.F., Lu, S.P., Tobias, J., Bley, K.R., Walk, D., Wendelschafer-Crabb, G., Simone, D.A., Selim, M.M. (2010). A randomized, controlled, open-label study of the long-term effects of NGX-4010, a high-concentration capsaicin patch, on epidermal nerve fiber density and sensory function in healthy volunteers. J Pain 11, 579587.
  • Khodorova, A., Richter, J., Vasko, M.R., Strichartz, G. (2009). Early and late contributions of glutamate and CGRP to mechanical sensitization by endothelin-1. J Pain 10, 740749.
  • Kramer, H.H., Rolke, R., Bickel, A., Birklein, F. (2004a). Thermal thresholds predict painfulness of diabetic neuropathies. Diabetes Care 27, 23862391.
  • Kramer, H.H., Schmelz, M., Birklein, F., Bickel, A. (2004b). Electrically stimulated axon reflexes are diminished in diabetic small fiber neuropathies. Diabetes 53, 769774.
  • Landerholm, S.H., Ekblom, A.G., Hansson, P.T. (2010). Somatosensory function in patients with and without pain after traumatic peripheral nerve injury. Eur J Pain 14, 847853.
  • Lane, N.E., Schnitzer, T.J., Birbara, C.A., Mokhtarani, M., Shelton, D.L., Smith, M.D., Brown, M.T. (2010). Tanezumab for the treatment of pain from osteoarthritis of the knee. N Engl J Med 363, 15211531.
  • Lauria, G., Hsieh, S.T., Johansson, O., Kennedy, W.R., Leger, J.M., Mellgren, S.I., Nolano, M., Merkies, I.S., Polydefkis, M., Smith, A.G., Sommer, C., Valls-Sole, J. (2010). European Federation of Neurological Societies/Peripheral Nerve Society guideline on the use of skin biopsy in the diagnosis of small fiber neuropathy. Report of a joint task force of the European federation of neurological societies and the peripheral nerve society. Eur J Neurol 17, 903909.
  • Maier, C., Baron, R., Tolle, T.R., Binder, A., Birbaumer, N., Birklein, F., Gierthmuhlen, J., Flor, H., Geber, C., Huge, V., Krumova, E.K., Landwehrmeyer, G.B., Magerl, W., Maihofner, C., Richter, H., Rolke, R., Scherens, A., Schwarz, A., Sommer, C., Tronnier, V., Uceyler, N., Valet, M., Wasner, G., Treede, R.D. (2010). Quantitative sensory testing in the German Research Network on Neuropathic Pain (DFNS): Somatosensory abnormalities in 1236 patients with different neuropathic pain syndromes. Pain 150, 439450.
  • McArthur, J.C., Stocks, E.A., Hauer, P., Cornblath, D.R., Griffin, J.W. (1998). Epidermal nerve fiber density: Normative reference range and diagnostic efficiency. Arch Neurol 55, 15131520.
  • Mogil, J.S., Miermeister, F., Seifert, F., Strasburg, K., Zimmermann, K., Reinold, H., Austin, J.S., Bernardini, N., Chesler, E.J., Hofmann, H.A., Hordo, C., Messlinger, K., Nemmani, K.V., Rankin, A.L., Ritchie, J., Siegling, A., Smith, S.B., Sotocinal, S., Vater, A., Lehto, S.G., Klussmann, S., Quirion, R., Michaelis, M., Devor, M., Reeh, P.W. (2005). Variable sensitivity to noxious heat is mediated by differential expression of the CGRP gene. Proc Natl Acad Sci U S A 102, 1293812943.
  • Novak, V., Freimer, M.L., Kissel, J.T., Sahenk, Z., Periquet, I.M., Nash, S.M., Collins, M.P., Mendell, J.R. (2001). Autonomic impairment in painful neuropathy. Neurology 56, 861868.
  • Oaklander, A.L. (2001). The density of remaining nerve endings in human skin with and without postherpetic neuralgia after shingles. Pain 92, 139145.
  • Oaklander, A.L. (2008). Mechanisms of pain and itch caused by herpes zoster (shingles). J Pain 9, S10S18.
  • Pare, M., Elde, R., Mazurkiewicz, J.E., Smith, A.M., Rice, F.L. (2001). The Meissner corpuscle revised: A multiafferented mechanoreceptor with nociceptor immunochemical properties. J Neurosci 21, 72367246.
  • Petersen, K.L., Rice, F.L., Farhadi, M., Reda, H., Rowbotham, M.C. (2010). Natural history of cutaneous innervation following herpes zoster. Pain 150, 7582.
  • Polydefkis, M., Yiannoutsos, C.T., Cohen, B.A., Hollander, H., Schifitto, G., Clifford, D.B., Simpson, D.M., Katzenstein, D., Shriver, S., Hauer, P., Brown, A., Haidich, A.B., Moo, L., McArthur, J.C. (2002). Reduced intraepidermal nerve fiber density in HIV-associated sensory neuropathy. Neurology 58, 115119.
  • Rice, F.L., Vickers, J., Albrecht, P.J. (2010). Cutaneous vascular sensory innervation may contribute to conscious tactile perception in a patient with congenital absence of pain and hyperhydrosis: Response to a letter to the editor by Nolano et al. Pain 151, 553554.
  • Rolke, R., Baron, R., Maier, C., Tolle, T.R., Treede, R.D., Beyer, A., Binder, A., Birbaumer, N., Birklein, F., Botefur, I.C., Braune, S., Flor, H., Huge, V., Klug, R., Landwehrmeyer, G.B., Magerl, W., Maihofner, C., Rolko, C., Schaub, C., Scherens, A., Sprenger, T., Valet, M., Wasserka, B. (2006). Quantitative sensory testing in the German Research Network on Neuropathic Pain (DFNS): Standardized protocol and reference values. Pain 123, 231243.
  • Rukwied, R., Mayer, A., Kluschina, O., Obreja, O., Schley, M., Schmelz, M. (2010). NGF induces non-inflammatory localized and lasting mechanical and thermal hypersensitivity in human skin. Pain 148, 407413.
  • Scholz, J., Mannion, R.J., Hord, D.E., Griffin, R.S., Rawal, B., Zheng, H., Scoffings, D., Phillips, A., Guo, J., Laing, R.J., Abdi, S., Decosterd, I., Woolf, C.J. (2009). A novel tool for the assessment of pain: Validation in low back pain. PLoS Med 6, e1000047.
  • Selim, M.M., Wendelschafer-Crabb, G., Hodges, J.S., Simone, D.A., Foster, S.X., Vanhove, G.F., Kennedy, W.R. (2010). Variation in quantitative sensory testing and epidermal nerve fiber density in repeated measurements. Pain 151, 575581.
  • Shun, C.T., Chang, Y.C., Wu, H.P., Hsieh, S.C., Lin, W.M., Lin, Y.H., Tai, T.Y., Hsieh, S.T. (2004). Skin denervation in type 2 diabetes: Correlations with diabetic duration and functional impairments. Brain 127, 15931605.
  • Sorensen, L., Molyneaux, L., Yue, D.K. (2006). The relationship among pain, sensory loss, and small nerve fibers in diabetes. Diabetes Care 29, 883887.
  • Vlckova-Moravcova, E., Bednarik, J., Belobradkova, J., Sommer, C. (2008). Small-fibre involvement in diabetic patients with neuropathic foot pain. Diabet Med 25, 692699.
  • Wildgaard, K., Ringsted, T.K., Hansen, H.J., Petersen, R.H., Werner, M.U., Kehlet, H. (2012). Quantitative sensory testing of persistent pain after video-assisted thoracic surgery lobectomy. Br J Anaesth 108, 126133.