Diabetes & Metabolic Syndrome: Clinical Research & Reviews 6 (2012) 85–89

Original article

Quantitative vibration perception threshold in assessing diabetic neuropathy: Isthe cut-off value lower for Indian Subjects? [Q-VADIS Study]

Samit Ghosal *, Jeff Stephens, Aninda Mukherjee 1

Glamorgan University, United Kingdom

A R T I C L E I N F O

Keywords:

Type 2 diabetes mellitus

Vibration perception threshold (VPT)

Diabetic neuropathy

Neuropathy disability score

A B S T R A C T

Aims: The aim was to compute a normative data of VPT [Vibration Perception Threshold], compare

results of VPT among type 2 diabetes patients with and without neuropathy, validate VPT taking NDS

[Neuropathy Disability Scores] as gold standard and suggest a cut off value for the Indian population.

Materials and methods: A clinic based case-control study was conducted at Nightingale Hospital (NH) in

Kolkata for 2 months duration. Fifty type 2 diabetes patients (who were detected with by fasting plasma

glucose or on medication) reporting at OPD (Out Patent Department) were randomly selected and

informed consent was obtained. The age range was 20–65 years and other common causes of neuropathy

were excluded. Same number of control patients without diabetes and reporting at the same hospital

during the study period in the similar age range were selected.

Results: The normative data of VPT for mean of 4 sites (malleoli and great toe) was 11.3 � 4.9 mV. The VPT

value was significantly higher among diabetic patients with neuropathy compared to non-neuropathic and

non-diabetic patients. Considering NDS score as gold standard lowering the cutoff value of VPT from 25 mV to

20 mV increased the sensitivity from 50% to 62.5% in detecting diabetic neuropathy compared to NDS taken

as a gold standard.

Conclusions: It was found that lowering the cut off value of VPT in Indian population increased the

sensitivity of the test to detect diabetic neuropathy without hampering the specificity. There is however

no indication that a lower cut off VPT value is justified as of now.

� 2012 Diabetes India. Published by Elsevier Ltd. All rights reserved.

Contents lists available at SciVerse ScienceDirect

Diabetes & Metabolic Syndrome: Clinical Research &Reviews

jo ur n al h o mep ag e: www .e lsev ier . c om / loc ate /d s x

1. Introduction

Neuropathy is an important chronic complication of DiabetesMellitus. Distal symmetric polyneuropathy (DPN) is the prototypeand is the major cause of diabetic foot ulcers [1]. Cliniciansexamine Achilles tendon reflexes, vibration perception (withgraded 128 Hz tuning fork), pressure and touch perception (with10 g mono filament), pain sensation (Neurotip), temperaturesensation (Tiptherm rod) and biomechanical foot abnormalities(hallus valgus, callous formation) in the outdoor clinics. Abnormalfindings are added to form a neuropathy score like NeuropathyDisability Score (NDS). NDS includes findings from Achilles tendonreflexes, 128-Hz tuning fork, pinprick, and temperature perceptionat predetermined sites.

Abbreviations: VPT, vibration perception threshold; NDS, neuropathy disability

score; FBS, fasting plasma glucose; OPD, out patient department; NH, Nightingale

Hospital; ABI, ankle brachial index; DPN, distal sensory peripheral neuropathy.

* Corresponding author.

E-mail address: ramdasghosal@gmail.com (S. Ghosal).1 MD (Community Medicine), Faculty, Department of Community Medicine,

Medical College, Kolkata.

1871-4021/$ – see front matter � 2012 Diabetes India. Published by Elsevier Ltd. All r

http://dx.doi.org/10.1016/j.dsx.2012.08.002

The VPT is assessed with a special device, the Neurothesiometeror Biothesiometer. The tractor of the device is applied on the tip ofthe big toe and malleoli and the VPT is measured in volts. The VPT isabnormal when the mean voltage exceeds 25 [2].

1.1. Aims and objectives

There is dearth of data among Indian population regarding thenormative value of VPT and its validity as a screening test to detectdiabetic neuropathy. In addition most centers use a lower cut-offvalue to detect DPN [Distal sensory Peripheral Neuropathy].

We could not find any supportive data to suggest this lower cut-off. So the present pilot project was undertaken with the aim ofcomputing the normative data in non-diabetic population in anIndian sub-population as well as to compare the results of VPTamong the type 2 diabetes patients with or without neuropathy.

We also wanted to find out the sensitivity and specificity of theVPT as a screening tool to detect neuropathy among the type 2diabetes patients taking NDS as the reference gold standard. Basedupon the normative data we tried to increase the sensitivity ofVPT to detect diabetic neuropathy without compromising thespecificity.

ights reserved.

S. Ghosal et al. / Diabetes & Metabolic Syndrome: Clinical Research & Reviews 6 (2012) 85–8986

2. Materials and methods

2.1. Type of study

It is an analytical case control study.

2.2. Study design

The data were collected from a clinic-based setting atNightingale Hospital situated at Shakespeare Sarani Kolkata, India.

2.3. Study population and their selection

Diabetic patients (detected by fasting plasma glucose or onmedication) reporting at OPD clinic of NH and willing to participatein the study were consecutively selected and informed consentwas obtained. The age range was 20–65 years and other causes ofneuropathy were excluded.

During the study duration 50 type 2 diabetes patients detectedby fasting plasma glucose were recruited. For selection of controlpatients we recruited those who did not have diabetes andreported at the same OPD during study period in the similar agerange. Ruling out other common medical non-diabetic causes forperipheral neuropathy was the real challenge.

There was the issue related to a huge list of differential diagnosisas well as cost for conducting these tests. We selected the verycommon disorders prevalent in the community in question and not

Fig. 1. An example of Bio

testing for the rare and uncommon disorders like paraneoplasticsyndrome. The common disorders related to peripheral neuropathy,which were excluded include alcohol intake, indigenous drugs(Ayurvedic) intake, and chronic disorders like hypothyroidism,uremia, vitamin B12 deficiency and peripheral arterial disease.There was a need to put in place a definition of neuropathy to screenthe recruited individuals especially the controls. We took a thoroughhistory related to peripheral neuropathy and to keep things simpletwo routine tests were preformed as a screening procedure. Thesewere ankle jerk and assessment of sensation with a 10 g Semmes-Weinstein monofilament.

To keep the cost of the total investigations low we performed theankle brachial index [ABI] to rule out peripheral vascular disease. TheHospital Educational Grant, an amount reserved for research purpose,funded us. As a result there was no financial burden on the individualsrecruited for the study. For each of the cases being selectedthroughout the study period controls were selected on the same day.

The controls were selected from the non-diabetic patientsattending the OPD of the same hospital for other medical or non-medical reasons. Repeat entry was prevented both for diabetic aswell as for non-diabetic patients. So total cases is to control ratiowas maintained at 1:1.

2.4. Duration of study

Two months (February and March – 2012) including datacollection and analysis.

thesiometry Results.

S. Ghosal et al. / Diabetes & Metabolic Syndrome: Clinical Research & Reviews 6 (2012) 85–89 87

2.5. Tools and techniques

Subjects were tested in a quiet room. The procedures wereexplained and demonstrated by an experienced operator whoperformed all VPT tests using Biothesiometer according tostandard protocol. The Biothesiometer was applied lightly at thesites of sole of feet as in Fig. 1. The voltage on the Biothesiometerdisplay at that instant was recorded (Fig. 1). The procedure wasrepeated thrice on each participant and the average reading wastaken to minimize operator and patient related errors. The datawas automatically recorded in the computer attached to thebiothesiometer. The same operator performed all the tests. Thecomputerized system incorporated for VPT measurement was wellstandardized and hence the chance of intra-observer variabilitywas considered to be minimal.

Variables used for both cases and controls were age (years), sex,height (cm), weight (kg), BMI [Body Mass Index] (kg/m2), VPT value(right and left great toe), VPT value (right and left malleoli), VPTmean value right and left foot for 6 sites each, mean VPT value bothtoe and both malleoli combined and NDS score (>6 or �6).

Variables used only for cases were duration of diabetes afterfirst detection (in years) and fasting blood sugar level in mg/dl.

2.6. Methodology

The study obtained institute ethical committee clearance after aprotocol presentation. The Nightingale Hospital Ethical Committeefixed the date for the presentation of the study proposal on the25th of February 2012.

The committee reviewed the presentation and recommendedcertain changes. These included preparing the consent form inEnglish as well the local vernacular (Bengali); trying to match boththe cases and controls economically and taking a thorough historyrelated to ingestion of indigenous (Ayurvedic) drugs rich in heavymetals.

These suggestions were adhered to and a fresh form wasprepared which Nightingale Ethical Committee approved.

Normative data among controls were then obtained by VPTtesting after excluding neuropathy by NDS score of �6. Among thediabetics VPT score was similarly obtained and sensitivity,specificity and predictive efficacy was tested compared to NDSscore as gold standard.

2.7. Statistical calculation

The data collected was transferred to Microsoft Excel 2007spreadsheet and statistical tests were done by SPSS 16.0 software.Mean and standard deviations were obtained for age, height, BMI,VPT values for toe, malleoli and 6 sites toe and malleoli(combined). Unpaired t test was done for statistical significancebetween diabetic patients with and without neuropathy andbetween diabetic and non-diabetic patients for those above-mentioned variables.

Table 1Distribution of study population according to demographic and bio-physical character

Study variables Diabetic patients

Neuropathy Non-neuropathy

N 32 18

Sex (M/F) 24/8 9/9

Mean age (years) � SD 55.2 � 6.8 50.8 � 8.6

Height (cm) � SD 164.0 � 7.5 158.9 � 7.4

Weight (kg) � SD 62.2 � 9.6 61.7 � 6.7

BMI � SD 23.0 � 2.6 24.5 � 2.9

Correlation was drawn both for diabetic and non diabeticsubjects regarding VPT values for right and left great toe, right andleft malleoli, mean VPT toe and malleoli and mean VPT right andleft foot (4 sites were 2 toe and 2 malleoli). Sensitivity, specificityand predictive efficacy of VPT as screening test compared to NDSscore taken as gold standard was calculated along with adjustmentfor reduced cut off.

3. Results

3.1. Neuropathy statistics

Out of total 50 diabetic patients 32(64%) had neuropathy atthe time of presentation as detected by NDS score. Non-diabeticpatients 50 in number were all devoid of neuropathy (as per NDSscore). Analysis showed mean age was significantly higheramong diabetic patients as compared to their non-diabeticcounterparts (p < 0.05). Height, weight, BMI and gender showedno statistical significant difference among diabetic patients withor without neuropathy and among diabetic and non-diabeticpatients (Table 1).

3.2. VPT Values in diabetics with neuropathy versus those without

neuropathy

Result of VPT in millivolt showed significantly higher valuesamong type 2 diabetes patients who had neuropathy at the timeof presentation compared to the diabetic patients withoutneuropathy in all the following areas: right great toe, left greattoe, right malleolus, left malleolus, right foot (6 sites), left foot (6sites) and mean value of 4 sites (2 great toe and 2 malleoli)(p < 0.05).

3.3. VPT values in diabetic versus non-diabetic individuals

VPT in millivolt was significantly higher among diabeticpatients compared to the non-diabetic patients in all the above-discussed parameters (p < 0.05). The normative data obtained formean value of 4 sites (2 great toe and 2 malleoli) was 11.3 � 4.9 mV(Table 2).

The correlation coefficient between VPT values of right and leftgreat toe among diabetic patients was 0.745 (p < 0.0001)compared to 0.650 (p < 0.0001) in the non-diabetic patients.

Correlation coefficient for right and left malleoli in the diabeticpatients was 0.812 (p < 0.001) and 0.882 (p < 0.001) among non-diabetic patients.

The correlation coefficient between mean VPT values of greattoe and malleoli in the diabetic patients was 0.828 (p < 0.001), and0.886 (p < 0.0001) among non-diabetic patients.

Correlation coefficient for mean VPT value for right foot and leftfoot among diabetic patients was 0.922 (p < 0.001) compared to0.915 (p < 0.001) among non-diabetic patients.

s.

Non-diabetics Statistical tests

Overall

50 50

33/17 33/17 p = 0.07/1.0

53.6 � 7.7 39.8 � 7.9 p = 0.06/<0.001

162.2 � 7.8 162.3 � 7.8 p = 0.02/0.94

62.0 � 8.6 64.1 � 12.1 p = 0.84/0.32

23.6 � 2.8 24.3 � 4.1 p = 0.07/0.29

Table 2Distribution of study population according to VPT mean and SD values at various sites.

VPT in mV mean � SD Diabetic patients Non-diabetics Statistical tests

Neuropathy Non-neuropathy Overall

Rt Great toe 24.1 � 11.9 14.4 � 7.9 20.6 � 11.6 10.6 � 4.2 p = 0.004/<0.001

Lt Great toe 28.2 � 11.6 15.2 � 9.1 23.5 � 12.4 11.7 � 5.2 p < 0.001/0.001

Rt malleolus 28.2 � 13.1 19.5 � 11.6 25.1 � 13.2 11.2 � 5.6 p = 0.023/<0.001

Lt malleolus 29.7 � 13.5 19.3 � 11.9 25.9 � 13.8 11.7 � 6.4 p = 0.009/<0.001

Rt foot 25.7 � 11.5 15.1 � 6.8 21.9 � 11.2 10.8 � 4.8 p < 0.001/<0.001

Lt foot 27.7 � 11.9 15.1 � 7.5 23.1 � 12.2 11.9 � 5.2 p < 0.001/0.001

Mean 4 site 27.6 � 11.4 17.1 � 8.4 23.8 � 11.5 11.3 � 4.9 p < 0.001/0.001 Quantitative

S. Ghosal et al. / Diabetes & Metabolic Syndrome: Clinical Research & Reviews 6 (2012) 85–8988

3.4. The issue of VPT cut off

Taking VPT value of 25 V as cut off for neuropathy detection andconsidering NDS score as the gold standard to diagnose diabeticneuropathy we got a sensitivity of 50%, specificity of 92.6%, positivepredictive value of 76.2% and negative predictive value of 79.7% forthe above-mentioned test.

Taking VPT value of 20 V as cut off (considering 2 standarddeviation above the normative data obtained in our study) weagain similarly obtained a sensitivity of 62.5%, specificity of 89.7%,positive predictive value of 74.1% and negative predictive value of83.6%.

It implied that sensitivity to detect neuropathy was increasedconsiderably without hampering specificity (p > 0.05) (Fig. 2).

3.5. Other important findings in our study were as follows

The mean duration of diabetes was 11.7 years with a SD of 4.6years. The duration of diabetes predicting neuropathy develop-ment showed poor correlation (Spearman rho = 0.10), (p = 0.50).Mean FBS level recorded among diabetics was 124.8 mg/dl with aSD of 26.1 mg/dl. FBS level showed poor correlation withneuropathy development with Spearman rho = 0.17 (p = 0.25).

Fig. 2. Doughnut (double) shows changes in sensitivity after altering cut off value of

VPT Sens = sensitivity. Nonsense = 100 � sensitivity (in percentage). Sensitivity

was increased from 50% to 62% after lowering of cut off.

4. Discussion

The present pilot study was conducted with 50 diabetic and 50non-diabetic patients with 66% males in both the groups and meanage was 53.6 � 7.7 years in diabetic group compared to 39.8 � 7.9years in the non-diabetic group. Thereby only sex matching was done.

In a study conducted on 30 diabetic patients with foot ulcer and85 control diabetic patients without ulcer the mean age was52.2 � 8.9 years in those with foot ulcers and 51.4 � 11.6 years inthose without foot ulcers [3].

In another study where 15 diabetic patients with neuropathyand 15 non-diabetic patients were recruited (9 male and 6 female)both the groups with diabetic neuropathy had a mean age of62.1 � 8.4 years compared to 62.3 � 8.3 years in their non-diabeticcounterparts [4].

In a data based on 110 diabetic and 64 non-diabetic patientswith 51 male and 59 female distribution in diabetic group and 32male and 32 female in non-diabetic group the mean age was48.3 � 12.3 years in those with diabetes [5].

Mean duration of diabetes was found to be 11.7 � 4.6 years inpresent study compared to 14.7 � 8.8 years in the ulcer patients and10.2 � 9 years among the non-ulcer patients as recorded by aprevious study [3]. In a study looking at the errors associated withbiothesiometer measurements the recorded mean duration ofdiabetes was 9.3 � 9.1 years [5].

The present pilot study found out 50% sensitivity and 92.6%specificity of VPT test in comparison to previous data that recorded100% sensitivity and 76.2% specificity [3]. VPT was proved to be asensitive measure of confirmed clinical neuropathy (87%) and ofdefinite clinical neuropathy (80%) and a specific measure ofabnormal nerve conduction (62%) in a study on type 1 diabeticpatients [6]. Higher VPT cut off values in their study improved testsensitivity and lower cut off values improved specificity unlike thepresent study although higher mean values of VPT in older andconfirmed neuropathy patients corroborated with our findings [6].

Assessment of neuropathy among diabetic patients withbiothesiometer to record VPT was also done in an Indian datawhere mean �2 SD was used for determining upper cut off for non-diabetic study subjects aged 20–45 years and the results came out tobe 19.7 corroborating well with our findings [7].

Mean height was 164 � 7.5 cm among the cases and 162 � 7.8 cmamong the controls in the present study along with mean weight of62 � 8.6 kg among the diabetic patients. On the contrary a previousstudy recorded mean height of 174 � 9 cm among neuropathy casesand 173 � 9 cm in non-diabetes patients thereby showing that meanheight of Europeans are more than that of Indians which could be animportant determinant for VPT values and cut off [4].

Mean weight was 89 � 13.4 kg among neuropathy patients,which is higher than their Indian counterparts [4].

In the present study VPT mean values recorded were 24 � 12 inright hallux and 28 � 13 in right malleoli in patients with neuropathyand 10 � 4 in right hallux, 11 � 6 in right malleoli in non-diabeticpatients. A study looking into the ceiling effects of VPT found values of

S. Ghosal et al. / Diabetes & Metabolic Syndrome: Clinical Research & Reviews 6 (2012) 85–89 89

47 � 5 in hallux and 43 � 11 in malleoli of neuropathy patients and15 � 8 in hallux and 10 � 7 in malleoli of non-neuropathy patientsthat were higher than our findings [4].

Correlation of VPT values in right and left toe was found to be0.75 in diabetics compared to 0.65 in non-diabetes patients in thepresent study whereas it was 0.89 and 0.85 in diabetes and non-diabetes patients respectively according to a previous study [5].

VPT testing is often criticized for being not sufficiently specificto large fiber with results being influenced by subject attentive-ness, motivation, fatigue and reproducibility varying in non-diabetic and diabetics along with device variation [8,9].

However it is a simple, quick to perform, painless, welltolerated, not significantly affected by the presence of foot callusor by limb temperature and standardized testing algorithms areavailable. So VPT is an attractive option for diabetic peripheralneuropathy assessment in research settings [8,9].

4.1. Study limitations

� Single center.� Age was not matched between the diabetic and non-diabetic

recruits.� Small number of subjects. The sensitivity of the lower cut off

value could have increased further if a larger population wasrecruited.

5. Conclusion

The normative data of VPT for mean of 4 sites (two malleoli andtwo great toes) obtained was 11.3 � 4.9 mV. VPT value wassignificantly higher among diabetic patients with neuropathy.Considering NDS score as gold standard lowering the cut off of VPTfrom 25 mV to 20 mV increased the sensitivity from 50% to 62.5% fordetecting diabetic neuropathy.

Lowering the cut off of VPT in Indian population may increasethe sensitivity of the test to detect diabetic neuropathy althoughfurther studies are required with larger sample size. However, thesensitivity is still low to make a definitive recommendation.

We conclude it is not advisable to use a lower VPT cut-off valueuntil further investigation is done on larger population in a multi-centric fashion. It is prudent we hold on with the presentinternational recommended VPT cut-off of 25 V for the time being.

6. Conflict of interest statement

None.

Acknowledgements

1. Patients who participated in the study.2. Hospital staff in the neurology testing section.3. Nightingale Hospital Management for providing the infrastruc-

ture as well as providing the educational grant for the pilotproject.

4. Dr. Jeff Stephens for his guidance and directing me into fine-tuning the project from the preliminary draft stage to the finalform.

5. Dr. Anindya Muhkerjee MD. Department of community medi-cine for his valuable help in getting the statistical calculationsdone.

6. University of Glamorgan for giving me this platform to conductthe pilot project.

References

[1] Boulton AJM, Kirsner RS, Vileikyte L. Neuropathic diabetic foot ulcers. NewEngland Journal of Medicine 2004;351:48–55.

[2] Feldman EL, Stevens MJ, Thomas PK, Brown MB, Canal N, Greene DA. Apractical two-step quantitative clinical and neurophysiological assessmentfor the diagnosis and staging of diabetic neuropathy. Diabetes Care1994;17:1281–9.

[3] Armstrong DG, Lavery LA, Vela SA, Quebedeaux TL, Fleischli JG. Choosing apractical screening instrument to identify patients at risk for diabetic footulceration. Archives of Internal Medicine 1998;158:289–92.

[4] Van Deursen RWM, Sanchez MM, Derr JA, Becker AB, Ulbrecht JS, Cavanagh PR.Vibration perception threshold testing in patients with diabetic neuropathy:ceiling effects and reliability. Diabetic Medicine 2001;18:469–75.

[5] Williams G, Gill JS, Aber V, Mather HM. Variability in vibration perceptionthreshold among sites: a potential source of error in biothesiometry. BMJ1998;296:233–5.

[6] Martin CL, Waberski BH, Pop-Busui R, Cleary PA, Catton S, Albers JW. Vibrationperception threshold as a measure of distal symmetrical peripheral neuropathyin type 1 diabetes. Diabetes Care 2010;33(12):2635–41.

[7] Mohan V, Vassy JL, Pradeepa R, Deepa M, Subashini S. The Indian type 2 diabetesrisk score also helps identify those at risk of macrovasvular disease andneuropathy (CURES-77). JAPI 2010;58:430–3.

[8] Elliott J, Tesfaye S, Chaturvedi N, Gandhi RA, Stevens LK, Emery C, et al. Large-fiber dysfunction in diabetic peripheral neuropathy is predicted by cardiovas-cular risk factors. Diabetes Care 2009;32:1896–900.

[9] Chong PS, Cros DP. Technology literature review: quantitative sensory testing.Muscle Nerve 2004;29:734–47.