Diabetologia

, Volume 51, Issue 6, pp 1046–1050

The Neuropad test: a visual indicator test for human diabetic neuropathy

  • C. Quattrini
  • M. Jeziorska
  • M. Tavakoli
  • P. Begum
  • A. J. M. Boulton
  • R. A. Malik
Article

DOI: 10.1007/s00125-008-0987-y

Cite this article as:
Quattrini, C., Jeziorska, M., Tavakoli, M. et al. Diabetologia (2008) 51: 1046. doi:10.1007/s00125-008-0987-y

Abstract

Aims/hypothesis

The commercially available Neuropad test was developed as a simple visual indicator test to evaluate diabetic neuropathy. It uses a colour change to define the integrity of skin sympathetic cholinergic innervation. We compared the results of Neuropad assessment in the foot with established measures of somatic and autonomic neuropathy.

Methods

Fifty-seven diabetic patients underwent Neuropad assessment, quantitative sensory and autonomic function testing, and evaluation of intra-epidermal nerve fibre density in foot skin biopsies.

Results

Neuropad responses correlated with the neuropathy disability score (rs = 0.450, p < 0.001), neuropathic symptom score (rs = 0.288, p = 0.03), cold detection threshold (rs = 0.394, p = 0.003), heat-as-pain perception threshold visual analogue score 0.5 (rs = 0.279, p = 0.043) and deep-breathing heart rate variability (rs = −0.525, p < 0.001). Intra-epidermal nerve fibre density (fibres/mm) compared with age- and sex-matched control subjects (11.06 ± 0.82) was non-significantly reduced (7.37 ± 0.93) in diabetic patients with a normal Neuropad response and significantly reduced in patients with a patchy (5.01 ± 0.93) or absent (5.02 ± 0.77) response (p = 0.02). The sensitivity of an abnormal Neuropad response in detecting clinical neuropathy (neuropathy disability score ≥5) was 85% (negative predictive value 71%) and the specificity was 45% (positive predictive value 69%).

Conclusions/interpretation

The Neuropad test may be a simple indicator for screening patients with diabetic neuropathy.

Keywords

Autonomic neuropathy Diabetic neuropathy Skin innervation Sudomotor dysfunction 

Abbreviations

CASE

computer-aided sensory evaluator

CDT

cold detection threshold

DB-HRV

deep-breathing heart rate variability

DNS

diabetic neuropathy symptom score

HP-VAS

heat-as-pain perception threshold visual analogue score

IENFD

intra-epidermal nerve fibre density

NDS

neuropathy disability score

Q-SART

quantitative sudomotor axon reflex testing

Introduction

The evaluation of nerve damage in human diabetic neuropathy is important to define those at risk of developing sensory loss, pain and foot ulceration. Simple clinical tests include assessment of neurological deficits using the neuropathy disability score (NDS), vibration perception threshold or the 10 g monofilament to define severe neuropathy and hence risk of ulceration [1]. Tests which evaluate earlier stages of neuropathy include neurophysiology, quantitative sensory testing [2] and nerve [3] or skin biopsy [4]. However, all of these tests are expensive, require specialist assessment and, in the case of biopsy, are invasive.

Autonomic dysfunction has been principally evaluated centrally by assessing heart rate variability [5] and peripherally by assessing sweating using complex and expensive equipment including the sympathetic skin response and quantitative sudomotor axon reflex testing (Q-SART) [6]. The commercially available Neuropad has recently been proposed as a simple test to diagnose sudomotor dysfunction and hence peripheral neuropathy. An adhesive pad containing cobalt salts is attached to the plantar aspect of the foot and changes colour from blue to pink within 10 min and defines if sudomotor and hence cholinergic sympathetic function is normal [7]. An abnormal Neuropad response is associated with sympathetic dysfunction and clinical neuropathy [7] and the reproducibility of the test has been shown to be excellent [8].

In this study we assessed the Neuropad indicator test in a cohort of diabetic patients who underwent a comprehensive assessment of somatic and autonomic neuropathy as well as assessment of intra-epidermal nerve fibre density (IENFD) in skin biopsies from the dorsum of the foot.

Methods

This study was approved by the Local Research Ethics committee and all patients gave informed consent to take part in the study.

Neuropathy assessment

All 57 diabetic patients underwent assessment of the NDS as previously described [1]. We set the cut-off for neuropathy/no neuropathy at NDS = 5/10. Patients underwent computer-aided sensory evaluator (CASE) IV quantitative sensory assessment [9] including: heat-as-pain perception threshold visual analogue score (HP-VAS), cold detection threshold (CDT) and deep-breathing heart rate variability (DB-HRV). Orthostatic hypotension was also assessed as a measure of sympathetic dysfunction [10] and was defined by a postural drop in BP of at least 20 mmHg [11]. Symptoms were assessed using the diabetic neuropathy symptom score (DNS) [12] and the short form of McGill’s Pain Questionnaire [13, 14].

Sudomotor dysfunction

Testing was carried out in a warm and quiet environment (24 ± 1°C). The Neuropad test (miro Verbandstoffe, Wiehl-Drabenderhöhe, Germany) was applied on the plantar aspect of the great toe and removed after 10 min to evaluate the colour change as normal (blue colour turned completely pink, score = 0), patchy (patches of blue and pink, score = 0.5), abnormal (remained blue, score = 1.0).

Skin biopsy

A 3 mm punch skin biopsy was taken from the dorsum of the foot, approximately 2 cm above the second metatarsal head, under 1% lidocaine local anaesthesia. Skin samples were immediately fixed in 4% (wt/vol.) paraformaldehyde for 18–24 h and then cryoprotected in 30% (wt/vol.) sucrose for 4 h and cut into 50 μm thick sections.

Immunohistochemistry

Melanin bleaching with 0.25% (wt/vol.) potassium permanganate and 3% (wt/vol.) oxalate solution was used prior to staining. For protein block, a mixture of Tris-buffered saline with 0.5% (wt/vol.) powdered milk, 1% Triton X-100 and 5% (vol./vol.) normal swine serum was applied for 4 h. The sections were incubated overnight at room temperature with 1:1,200 Biogenesis polyclonal rabbit anti-human protein gene product (PGP9.5) antibody (Serotec, Oxford, UK). Swine anti-rabbit secondary antibody 1:300 (45 min) was then applied; sections were quenched with 1% (vol./vol.) H2O2 in 30% (vol./vol.) MeOH-PBS (30 min) prior to a 45 min incubation with 1:500 horseradish peroxidase–streptavidin (Vector Laboratories, Peterborough, UK) and 3,3′-diaminobenzidine chromogen was used to demonstrate the reaction. Sections were transferred onto gelatin-subbed slides, coverslipped and observed by light microscopy.

Image analysis

An image analysis camera (Sony 2CCD, CCD-IRIS; Weybridge, Surrey, UK) and suitable computer program (Leica QWin Standard V2.4; Leica Microsystem Imaging, Cambridge, UK) were used to quantify IENFD (number of fibres per mm of basement membrane) as previously described [15].

Statistical analysis

Statistical analysis was performed using SPSS 15.0 for Windows. Results are presented as means ± SEM. Spearman analysis was used to test for correlation of Neuropad ranks with all other measures of neuropathy. The Mann–Whitney test was used for comparison between two groups and the Kruskal–Wallis test was used to compare more than two groups. Post hoc multi-group comparison analysis was performed by a Tukey test (in the case of equal variances as assessed by Levene’s test) or a Dunnet T3 test (in the case of unequal variances) tests. A χ2 test was used to study associations between two dichotomous variables.

Results

Fifty-seven diabetic patients (20 type 1 and 37 type 2) aged 56 ± 1.4 years were classified in accordance with the Neuropad response as: normal (n = 16), patchy (n = 16) and abnormal (n = 21). Age, BMI and HbA1c did not differ between patient groups with different Neuropad responses, whereas duration of diabetes (p < 0.01) and severity of neuropathy assessed via NDS (p < 0.05) and degree of postural hypotension did (Table 1). According to the NDS, 12 patients had no neuropathy (NDS < 3); 18 mild neuropathy (NDS 3–5), 15 moderate neuropathy (NDS 6–8) and 12 severe neuropathy (NDS 9–10). The NDS was significantly higher in patients with an abnormal Neuropad response (6.5 ± 0.7) compared with patients with a normal response (3.3 ± 0.6), p < 0.05, and the Neuropad responses correlated with the severity of neuropathy defined by the NDS (rs = 0.450, p < 0.001). Patients were further grouped into those with NDS < 5 (40%) and NDS ≥ 5 (60%). The sensitivity of an abnormal Neuropad response (either blue or patchy) in detecting neuropathy was 85% (negative predictive value 71%), while the specificity was 45% (positive predictive value 69%).
Table 1

General clinical characteristics of the patients distributed according to the three categories of Neuropad responses

 

Age (years)

Type 1/type 2

Duration (years)

HbA1c (%)

BMI (kg/m2)

PH (mmHg)

NDS

p value

NS

NS

0.01

NS

NS

NS

0.004

Neuropad normal

54 ± 3

6/10

14 ± 3

7.5 ± 0.4

28.5 ± 1.2

−13 ± 3

3.3 ± 0.6

Neuropad patchy

56 ± 3

6/11

24 ± 3*

8.4 ± 0.2

29.7 ± 1.3

−19 ± 2

5.5 ± 0.8

Neuropad abnormal

59 ± 2

8/16

25 ± 2*

8.0 ± 0.3

30.4 ± 1.0

−20 ± 3

6.5 ± 0.7*

PH, Postural hypotension

*p < 0.05 compared with the Neuropad-normal group. Patients with a patchy and abnormal Neuropad associated significantly with PH (>20 mmHg systolic reduction) (p = 0.001, χ2 test)

Symptoms

Seventeen patients had no symptoms of neuropathy; 70% had a DNS > 0, 55% DNS > 1, 39% DNS > 2 and 25% DNS > 3. Thirty-two patients had neuropathic pain (HP-VAS score ≥1). The Neuropad response correlated with the neuropathic symptoms as assessed by the DNS Questionnaire (rs = 0.288, p = 0.03) and pain subtypes ‘unsteady’ (rs = 0.379, p = 0.004), ‘numb’ (rs = 0.363, p = 0.006) and ‘throbbing’ (rs = 0.336, p = 0.012) as characterised by the McGill Pain Questionnaire.

Small fibre function

The results of CASE IV assessment are summarised in Table 2. The CDT response was significantly greater in diabetic patients with a patchy (91 ± 3) and an abnormal (92 ± 4) response compared with patients with a normal response (69 ± 7), p = 0.007, and the Neuropad ranking correlated with CDT (rs = 0.394, p = 0.003; Table 3). While there was no significant difference in HP-VAS 0.5 between a normal and abnormal Neuropad response there was a correlation between it and Neuropad ranking overall (rs = 0.279, p = 0.043; Table 3). The DB-HRV response was significantly lower in diabetic patients with a patchy (29 ± 9), p < 0.05, and abnormal (19 ± 7), p < 0.05, response compared with diabetic patients with a normal response (64 ± 8) and correlated with the overall Neuropad ranking (rs = −0.525, p < 0.001; Table 3). No relationship was found with HP-VAS 5.0, HP-VAS 0.5–5.0 or postural hypotension (drop in BP > 20 mmHg), which occurred in 36% of patients. However, the χ2 test showed a significant association between postural hypotension and an abnormal Neuropad test (p = 0.001).
Table 2

CASE IV assessment in the patients distributed according to the three categories of Neuropad response

 

CDT (pc)

HP-VAS 0.5 (pc)

HP-VAS 5.0 (pc)

HP-VAS 0.5–5.0 (pc)

DB-HRV (pc)

p value

0.007

NS

0.020

0.015

0.002

Neuropad normal

69 ± 7

31 ± 7

26 ± 8

33 ± 8

64 ± 8

Neuropad patchy

91 ± 3*

54 ± 11

62 ± 9*

69 ± 9*

29 ± 9*

Neuropad abnormal

92 ± 4*

58 ± 7

51 ± 7

36 ± 7

19 ± 7*

pc, Percentile

*p < 0.05 compared with the Neuropad-normal group

Table 3

Correlation of the Neuropad ranking (normal, abnormal and patchy) with the three quantitative sensory function tests of small fibre denervation

Test

Correlation

CDT

 

 rs

0.394

 p value

0.003

HP-VAS 0.5

 

 rs

0.279

 p value

0.043

DB-HRV

 

 rs

−0.525

 p value

0.0003

IENFD

The IENFD was significantly reduced in diabetic patients (5.69 ± 0.51) compared with 15 age and sex matched non-diabetic control individuals (11.06 ± 0.82, p < 0.001, Fig. 1). Diabetic patients with a normal Neuropad result had a non-significant reduction in IENFD (7.37 ± 0.93). This was significantly reduced in patients with either a patchy (5.01 ± 0.93) or absent (5.02 ± 0.77) result (p = 0.02). IENFD correlated with the Neuropad response (rs = −0.271, p = 0.04).
Fig. 1

IENFD in diabetic patients with normal and abnormal Neuropad responses. There were significantly lower IENFDs in patients with a pathological Neuropad test (abnormal) compared with those with a normal Neuropad test (normal), p = 0.02. IENFD in non-diabetic control individuals is included for reference (C). Horizontal bars, mean; error bars, SEM

Discussion

Sweat function has an important protective role and the presence of peripheral autonomic neuropathy characterised by anhidrosis is integral to the pathogenesis of diabetic foot ulceration [16]. When sweat gland function is lost because of peripheral autonomic denervation, foot ulceration is facilitated [17].

Our data show that the Neuropad test is able to detect the presence of clinically relevant neuropathy with good sensitivity, confirming the data from Papanas et al. [7]. A lower specificity in comparison with sensitivity is also confirmed [18] but, notably, with limited reduction in positive predictive value. The Neuropad correlations with CASE IV measurements, as well as with neuropathic symptoms, substantiate its validity as an effective screening instrument for small fibre neuropathy. Therefore the test appears to have the potential to be used as a first-line screening tool, and once a patient is found to have a positive Neuropad test, additional assessment is recommended.

Interestingly, tests of autonomic dysfunction, including altered heart rate variability and postural hypotension were also associated with an abnormal Neuropad response. Although somatic and cardiac autonomic neuropathies are considered to be separate clinical entities [19], a common link is that of involvement of small fibres [20, 21]. Although the association with somatic autonomic neuropathy demands a more specific validation with quantitative sudomotor testing using the gold standard for sudomotor testing, Q-SART [22, 23], this was not available in our laboratory.

With regard to the site of testing using Neuropad we chose the pulp of the great toe, which has a dense cholinergic sympathetic innervation and avoided the metatarsal head area as the latter is a common site of ulceration. Furthermore, the biopsy was not taken from the pulp of the toe as this would have imposed a considerable risk in terms of non-healing. Nevertheless the biopsied area was from a nearby area, the dorsum of the foot, which was clearly more distal than the normal site for skin biopsy, which is the thigh or ankle [24]. Studying the nerve fibres entering the sweat glands in the dermis would have been more anatomically relevant in demonstrating a sympathetic deficit [25]; however, the skin biopsy was neither large enough nor deep enough to allow quantification of sweat gland innervation adequately. The correlation between the Neuropad response and IENFD was poorer than with the other measures of neuropathy. We can only speculate that assessing IENFD, which provides a structural measure of the number of fibres, may not reflect nerve fibre and specifically sudomotor nerve function [25], as recent studies certainly suggest that IENFD relates well to other measures of somatic neuropathy [26, 27]. Nevertheless, as IENFD is widely accepted as a gold standard measure of skin denervation and neuropathy [26, 27, 28], we believe these findings add strength to the assertion that Neuropad reflects the severity of distal somatic neuropathy.

A potential clinical limitation of Neuropad may be the 10 min needed for screening in a busy diabetic clinic. A recent study has suggested that the maximum threshold time to define an abnormal Neuropad test may be lowered to 530 s [29]. However, the distinct practical clinical advantage of the Neuropad test is that patients can self-administer the test at home. Furthermore, the visual reinforcement of an abnormal test will hopefully make the patient more aware of their risk of neuropathy and ulceration [30].

In conclusion, this study has shown that an abnormal Neuropad response indicates both functional and structural denervation in the feet of diabetic patients. This has considerable clinical relevance in screening for diabetic neuropathy.

Acknowledgements

The study was supported by Diabetes UK (RD03/0002624) and National Institutes of Health (R01 NS46259-01) and Neuropad tests were supplied by miro Verbandstoffe.

Duality of interest

The authors declare that there is no duality of interest associated with this manuscript.

Copyright information

© Springer-Verlag 2008

Authors and Affiliations

  • C. Quattrini
    • 1
    • 4
  • M. Jeziorska
    • 2
  • M. Tavakoli
    • 1
    • 4
  • P. Begum
    • 3
  • A. J. M. Boulton
    • 1
    • 4
  • R. A. Malik
    • 1
    • 4
  1. 1.Division of Cardiovascular Medicine, Core Technology FacilityUniversity of ManchesterManchesterUK
  2. 2.Division of Regenerative MedicineUniversity of ManchesterManchesterUK
  3. 3.Department of Gastrointestinal Sciences, Clinical Sciences BuildingSalford Royal HospitalsSalfordUK
  4. 4.Manchester Diabetes CentreManchester Royal InfirmaryManchesterUK

Personalised recommendations