Diabetes can result in long-term health complications, with one of the most common being microvascular damage that leads to diabetic neuropathy (DN), which is an insidious, variated pathology that affects multiple body systems and increases amputation risk [
Earlier detection of DPN in at-risk individuals and in those with prediabetes (PD) or type 2 diabetes (T2D) allows for potential better management through optimal intervention and lifestyle changes [
While each measure has been shown to be reliable and valid in adults with PD and T2D, overweight, obese, and inactive (OOI) populations are also at high risk for the development of T2D and associated complications [
Volunteer subjects were recruited by flyers, email, word of mouth, and university-wide announcements. Subjects were screened by phone for exclusionary factors prior to reporting for testing. Potential subjects then reported to the Old Dominion University Wellness Institute for additional screening, informed consent, and testing measures. Assignment to groups was based on current HbA1C testing values obtained onsite during study procedures. Individuals with a history of type 1 diabetes, tobacco use, hepatitis B, hepatitis C, HIV, pregnancy, damage to the lower extremities, nerve disease (other than neuropathy), peripheral arterial disease, lower limb amputations, foot ulcers, or a serious medical condition that would compromise subject safety or the integrity of the study were excluded. Study procedures were approved by the Old Dominion University Institutional Review Board prior to recruitment, and subjects gave their informed consent before participating.
This study was designed to examine the effectiveness of the 128 Hz tuning fork, 1 g and 10 g monofilaments, and QOL-DN questionnaire for the identification of early signs of DPN in OOI, PD, and T2D study populations. All testing results were compared to the SNAP and SNCV of the NC-stat DPNCheck, which served as the criterion measure for the study.
Participants followed hydration instructions for 24 hours (8–10 cups of fluid) and 2-3 hours (2-3 cups of fluid) prior to appointment times to avoid point of care device (POCD) high total hemoglobin errors. Finger-stick testing was performed with a DCA Vantage 2000 Analyzer (Siemens, Tarrytown, NY) and DCA Vantage HbA1C test kits utilizing sterile techniques [
Nerve conduction study procedures utilized the POCD NC-stat DPNCheck (DPNCheck, NeuroMetrix Inc., Waltham, MA) and followed previously outlined methods [
Unique biosensor technology paired with the NC-stat DPNCheck unit’s two probes allowed for quick onsite evaluation in a matter of minutes. After skin preparation, the probes were coated in conductive gel and applied directly to the skin, posterior to the lateral malleolus. With the single press of a button, the unit distributed 100 mA of current, which was detected by the single patient use disposable biosensor. A built-in thermometer accounted for variances in temperature between 23°C and 30°C and notified the operator of skin temperatures too cold for testing, preventing further use until appropriate temperatures were present. Up to five individual nerve conduction study attempts per leg were utilized to collect three sets of SNCV and SNAP values, each providing individualized feedback based on the patient’s age, height, and weight data. Device errors were not recorded; however, zero readings were recorded by hand and reattempts were made up to the 5-trial limit, as individual tolerance permitted. When individuals could not tolerate the acquisition of 3 data points per leg, last observation carried forward (LOCF) methods were employed to complete the trial set [
The QOL-DN, a validated instrument and method for assessing neuropathy and differentiating between autonomic, large, and small fiber impairments [
A 128 Hz tuning fork was used to assess vibration perception [
Commercially produced 1 g and 10 g monofilaments (North Coast Medical, San Jose, CA) were used with a standard lab testing table to evaluate sensation perception. Monofilament storage and testing took place in a temperature-controlled environment, within the parameters established by previous research [
Data analyses were performed using SPSS version 22.0 for Windows (SPSS, Chicago, IL). Participants, group characteristics, SNAP, and SNCV are presented as raw data. Criterion and dependent variable data were logarithmically transformed to best achieve normality for statistical analysis. Partial correlations were analyzed using Spearman’s coefficients for the tuning fork, 1 g and 10 g monofilaments, QOL-DN and NC-stat DPNCheck results, to determine the relationship between the variables. Interpretation of nerve conduction studies and diabetic neuropathy across the study for means was based on a cutoff definition of <6
In total, 34 adults (10 males, 24 females) of varying ethnicities [Caucasian (64.7%); African-American (12%)] participated in the study and were assigned to one of the three groups based on their HbA1C values: 10 (29.4%) normoglycemic OOI adults (6 females, 4 males; 59.6 years, ±13.0 years), 13 (38.2%) with PD (11 females, 2 males; 56.4 years, ±12.2 years), and 11 (32.4%) with T2D (7 females, 4 males; 59.6 years, ±12.1 years). Group characteristics included mean weight (OOI: 87.93 kg, SD ± 10.94 kg; PD: 98.03 kg, SD ± 23.26 kg; and T2D: 101.30 kg, SD ± 17.68 kg), mean BMI (OOI: 30.90 kg/m2, SD ± 3.17 kg/m2; PD: 34.20 kg/m2, SD ± 6.71 kg/m2; and T2D: 35.10 kg/m2, SD ± 5.03 kg/m2), and mean HbA1C (OOI: 5.3%, SD ± .36%; PD: 5.9%, SD ± .22%; and T2D: 7.8%, SD ± 2.12%). The ten males (29.4%) in the study had a mean age of 61 years (SD ± 13.53 years), a mean height of 1.745 m (SD ± .08 m), a mean weight of 105.90 kg (SD ± 20.62 kg), a mean BMI of 34.85 kg/m2 (SD ± 4.97 kg/m2), and a mean HbA1C of 6.0% (SD ± .96%). Twenty-four females (70.6%) who participated in the study had mean group characteristics of the following: mean age of 57.2 years (SD ± 11.58 years), mean height of 1.66 m (SD ± .06 m), mean weight of 89.4 kg (SD ± 15.10 kg), mean BMI of 32.99 kg/m2 (SD ± 5.67 kg/m2), and mean HbA1C of 6.5% (SD ± 1.79%).
Fifteen participants reported no prior diagnosis or knowledge of hyperglycemia. Five had PD (based on HbA1C). Without specific recruitment for OOI, 33 of the 34 subjects in all groups were overweight (9) or obese (24). Six individuals reported prior neuropathy diagnosis, while 28 individuals reported having no prior neuropathy diagnosis or knowledge. Ten reported T2D-specific medication usage as part of their personal medical plan, and two T2D subjects reported taking a combination of T2D and neuropathy medications.
Group means for SNAP and SNCV did not significantly vary by HbA1C level (data not shown). No significant differences were evident among OOI, PD, and T2D groups on SNAP and SNCV values (SNAP: R
NCS results.
Sural nerve amplitude potentials and conduction velocities
|
Min | Max | Mean | Std. error | Std. dev. | |
---|---|---|---|---|---|---|
SNAP-R ( |
||||||
OOI | 10 | 2 | 14.3 | 6.631 | 1.444 | 4.567 |
PD | 13 | 2 | 24.7 | 7.691 | 1.674 | 6.037 |
T2D | 11 | 2 | 25 | 9.875 | 2.133 | 7.076 |
SNAP-L ( |
||||||
OOI | 10 | 2.3 | 21.7 | 7.129 | 1.834 | 5.798 |
PD | 13 | 3 | 21.7 | 7.277 | 1.186 | 4.277 |
T2D | 11 | 3 | 21.7 | 10.572 | 2.064 | 6.847 |
SNCV-R ( |
||||||
OOI | 10 | 35.3 | 55.7 | 46.2 | 1.902 | 6.016 |
PD | 13 | 30 | 57 | 48.2 | 1.871 | 6.747 |
T2D | 11 | 35.3 | 57 | 45.5 | 1.816 | 6.022 |
SNCV-L ( |
||||||
OOI | 10 | 41.3 | 55 | 47.265 | 1.519 | 4.803 |
PD | 13 | 43 | 55 | 49.637 | 1.072 | 3.865 |
T2D | 11 | 37.3 | 57 | 46.876 | 1.946 | 6.455 |
Sensitivity and specificity of screening tests and subcomponents
Prevalence | Sensitivity | Specificity | PPV |
NPV |
|
---|---|---|---|---|---|
128 Hz tuning fork | 52.90% | 50.00% | 75.00% | 69.20% | 57.10% |
1 g monofilament | 26.50% | 66.70% | 72.00% | 46.20% | 85.70% |
10 g monofilament | 55.90% | 47.40% | 73.30% | 69.20% | 52.40% |
QOL-DN total | 29.40% | 60.00% | 70.80% | 46.20% | 81.00% |
QOL-DN symptoms | 32.40% | 36.40% | 60.90% | 30.80% | 66.70% |
QOL-DN large fiber | 35.30% | 58.30% | 72.70% | 53.80% | 76.20% |
QOL-DN small fiber | 97.10% | 39.40% | 100.00% | 100.00% | 4.80% |
QOL-DN ADLS | 76.50% | 42.30% | 75.00% | 84.60% | 28.60% |
QOL-DN autonomic | 61.80% | 42.90% | 69.20% | 69.20% | 42.90% |
Sural NCS, signs, and symptoms
Variable | Total | Group | |||
---|---|---|---|---|---|
OOI | PD | T2D | |||
Sural NCS |
Normal | 7 | 1 | 3 | 3 |
Abnormal |
27 | 10 | 9 | 8 | |
|
|||||
Signs |
Tuning fork | ||||
Normal | 14 | 3 | 6 | 5 | |
Abnormal |
20 | 7 | 7 | 6 | |
Monofilaments | |||||
1 g | |||||
Normal | 3 | 1 | 0 | 2 | |
Abnormal |
31 | 9 | 13 | 9 | |
10 g | |||||
Normal | 3 | 1 | 0 | 2 | |
Abnormal |
31 | 9 | 13 | 9 | |
|
|||||
Symptoms |
None reported | 11 | 6 | 1 | 4 |
Reported |
23 | 4 | 12 | 7 | |
|
|||||
Autonomic |
None reported | 21 | 7 | 8 | 6 |
Reported |
13 | 3 | 5 | 5 | |
|
|||||
ADLS |
None reported | 26 | 8 | 10 | 8 |
Reported |
8 | 2 | 3 | 3 | |
|
|||||
NCS, sign & symptom combinations | AbNCS, signs & symptoms | 17 | 3 | 9 | 5 |
AbNCS, signs, or symptoms | 9 | 5 | 1 | 3 | |
AbNCS, no signs, or symptoms | 1 | 1 | 0 | 0 | |
NNCS, signs & symptoms | 7 | 1 | 3 | 3 |
SNAP values
Min | Max | Mean (SD) | Std. error | ||
---|---|---|---|---|---|
SNAP-R ( |
|||||
NNCS |
|
7.300 | 25.000 | 14.022 ± 5.64 | 1.564 |
AbNCS |
|
3.618 | 5.202 | 4.41 ± 1.740 | 0.379 |
SNAP-L ( |
|||||
NNCS |
|
11.050 | 17.138 | 14.096 ± 5.033 | 1.396 |
AbNCS |
|
3.898 | 5.524 | 4.712 ± 1.786 | 0.389 |
The tuning fork on/off test did not correlate with criterion variables used in this study (see Table
NC-stat DPNCheck Spearman’s partial correlations (log transformed).
SNAP-R | SNAP-L | SNCV-R | SNCV-L | |
---|---|---|---|---|
|
|
|
|
|
|
||||
On/off | 0.221 | 0.137 | 0.235 | −0.089 |
|
0.121 | 0.235 | 0.106 | 0.319 |
Timed R | −0.066 | −0.019 | −0.019 | −0.099 |
|
0.365 | 0.461 | 0.459 | 0.302 |
Timed L | −0.063 | −0.052 | −0.018 | −0.081 |
|
0.371 | 0.392 | 0.463 | 0.355 |
|
||||
Total 1 g | 0.364 |
0.312 |
−0.06 | −0.141 |
|
0.024 | 0.047 | 0.377 | 0.229 |
1 g R | 0.229 | 0.206 | 0.024 | 0.077 |
|
0.112 | 0.138 | 0.451 | 0.342 |
1 g L | 0.393 |
0.301 |
−0.191 | −0.313 |
|
0.016 | 0.053 | 0.155 | 0.046 |
Total 10 g | 0.098 | 0.088 | 0.032 | 0.03 |
|
0.304 | 0.321 | 0.432 | 0.438 |
10 g R | 0.096 | 0.16 | 0.005 | −0.066 |
|
0.306 | 0.2 | 0.489 | 0.364 |
10 g L | 0.137 | 0.07 | 0.031 | 0.054 |
|
0.235 | 0.356 | 0.436 | 0.388 |
|
||||
Total | −0.317 |
−0.311 |
0.162 | −0.117 |
|
0.044 | 0.047 | 0.197 | 0.269 |
Symptoms | −0.332 |
−0.375 |
0.213 | −0.003 |
|
0.036 | 0.021 | 0.129 | 0.493 |
Large fiber | −0.297 | −0.284 | 0.107 | −0.163 |
|
0.056 | 0.064 | 0.286 | 0.195 |
Small fiber | −0.241 | −0.187 | −0.311 |
−0.366 |
|
0.099 | 0.161 | 0.047 | 0.023 |
ADLS | −0.354 |
−0.263 | 0.104 | −0.065 |
|
0.028 | 0.08 | 0.293 | 0.366 |
Autonomic | −0.236 | −0.245 | 0.149 | −0.044 |
|
0.105 | 0.096 | 0.216 | 0.408 |
Accounts for HbA1C, age, and waist (cm).
Receiver operator characteristics (ROC) curve.
The 10 g monofilament did not significantly correlate to criterion variables. Sensitivity for the 10 g monofilament was 47.4%, with specificity at 73.3% and a PPV of 69.2%. Total QOL-DN scores negatively correlated with both SNAPs [R:
The integration of these testing methods provided foundational work necessary to develop a better understanding of the onset of dysfunctional physiological processes within OOI, PD, and T2D populations during the beginning of disease onset, shedding light on associations between symptoms and diseases. Moderate positive correlations were found between the 1 g monofilament and total and left leg scores with the recorded SNAP values. Additionally, the total QOL-DN, ADLS, and symptom scores negatively and moderately correlated to SNAPs, while small fiber scores negatively moderately correlated to SNCV. These correlations suggest that these tools may be useful for incorporation into low-cost screenings.
Detecting diabetes complications is an unfolding evolution that involves multiple dynamics to consider. DPN may present in a completely silent manner, without pain, burning, or symptoms of annoyance, and the utilization of the QOL-DN provides a unique, previously validated means of evaluating symptomology in at-risk populations [
This study utilized SNAP and SNCV values to evaluate nerve function in participants. Sural nerve conduction and amplitude values are validated quantitative physiological markers that assist in the assessment and confirmation of DPN status with or without the presence of signs or symptoms [
Other research has reported alternate findings, which may be at least partially explained by differences in the populations studied. Perkins et al. [
Assessment in this study differed from previous research by evaluating each individual participant according to age, height, and weight to determine appropriate cutoffs for normal and abnormal findings. This method individualized results to each participant with the built-in NC-stat software and included the potential impact of being overweight or obese. Having noted discrepancies between values of traditional NCS and the NC-stat DPNCheck, Lee et al. [
To detect early DPN in normoglycemic OOI individuals, it had been hypothesized that the 128 Hz tuning fork and QOL-DN would provide the best mechanisms for detection; however, current results indicated only partial support for this. The tuning fork on/off test did not correlate well with the NC-stat DPNCheck SNAP criterion variables, although the QOL-DN did yield correlational results on several measures. This finding is different than some prior research, as the QOL-DN has not always been found to correlate with electrophysiological measures [
The QOL-DN ranged in sensitivity (36.4–60.0%) and specificity (60.9–100.0%), differing from previous research that resulted in high specificity and sensitivity. The small fiber component weighed in with high specificity, yet overall, specificity averaged 60.9–75.0% with the other components. These results may have been affected by this study’s population and its small number of subjects across three groups in attempting to uncover DPN at the earliest point possible. Previous research expressed concern relating to the QOL-DN: Hogg et al. [
The 1 g monofilament proved to be a useful tool, with 31 individuals in this study experiencing abnormal findings. This measure indicated relatively high sensitivity (66.7%) and specificity (72.0%). However, validation of 1 g physical findings was achieved with moderate correlations back to the criterion SNAP variables, corroborating previous research that reported mixed sensitivity and specificity, such as Feng et al. and Taksande et al. [
The QOL-DN (S: 60.0%; SP: 70.8%) was hypothesized to be the most sensitive measure to detect undisclosed DPN in the study population; instead, the 1 g monofilament (S: 66.7%; SP: 72.0%) performed slightly better within this limited cohort. The 128 Hz tuning fork on/off test (S: 50.0%; SP: 75.0%) fell not too far behind. The QOL-DN proved to be lower in sensitivity in this study and ranged in specificity, yet many components (total QOL, symptoms, ADLS, and small fiber) correlated to the criterion measure used. Sample size was quite small, and with the significant previous validation history of the QOL-DN, it is likely that this screening instrument in its entirety may be useful in determining both small and large fiber deficits in a larger study in overweight, obese, and inactive individuals [
There is a strong indication of early-onset and subclinical neuropathy in the populations in this study, suggesting that careful screening of individuals at earlier stages may be quite beneficial in the early detection of DPN, even prior to hyperglycemia onset in OOI and PD. Smith and Singleton [
This pilot study has some limitations. Besides being small, the cohort was limited to one site of investigation and only two ethnic groups; therefore, its findings must be placed within the context of the larger body of literature available. Lack of random assignment and use of volunteers for subjects created potential selection bias, with the use of clinical populations and low available funding heavily influencing these methods. HbA1C testing was performed with a validated machine, yet oral glucose tolerance testing is preferred by some researchers, particularly for individuals with cardiac autonomic neuropathy (CAN) [
Although we applied rigorous testing preparation and methods, it is possible that there was an error we remain unaware of that affects the validity of these findings [
Unique features of testing done in this study include accounting for age, height, and weight individually, securing a minimum of 3 readings per leg, therefore allowing a more individualized and accurate assessment of the large myelinated nerve fibers. Small fiber-associated deficits were not directly assessed in this study. The testing done does offer a nonclinical analysis based on the criteria required by Tesfaye et al. aiming to achieve minimal definition requirements for confirmed and subclinical DSPN classification, with the intent of developing early screening measures for DPN-prone populations [
In summary, early DPN signs and symptomology can be detected in OOI, PD, and T2D populations using low-cost, established tools. The 1 g monofilament proved to be more useful for the early detection of DPN than the 10 g monofilament within this population by correlating to the study standard and providing the highest sensitivity and specificity combination. The total QOL measure also proved to be useful, correlating to the standard, yielding the second highest combination of sensitivity and specificity to SNAP and SNCV values. Several QOL-DN subscales (ADLS, Small Fiber, Symptoms) provided valuable, standardized information that can be incorporated into low-cost community screening models for early DPN detection in populations with or without PD and T2D, while providing varied sensitivity and specificity across multiple categories. The 128 Hz tuning fork did not prove to be quite as accurate in this population as the QOL-DN and should be used as a tandem measure when screening. Future research should consider a larger study with the same populations, aiming to continue to refine and develop screening methods targeted towards disclosing both symptomatic and asymptomatic DPN.
The participant was first shown the monofilament, by allowing them to see it and then gently touching their forehead with the instrument. They were instructed to close their eyes and say “yes” each time that the stimulus was felt. The monofilament was applied to the dorsum of the great toe, just proximal to the nail bed for approximately one second, 4 times per foot, in an irregular pattern in such a way that the stimulus could not be anticipated. Scoring: 0 was assigned for no perception, 0.5 substantially less perception and 1 for normal perception. Totals: 0–3, the presence of neuropathy is likely; 3.5–5 indicates that the onset of neuropathy within four years is high; 5.5 and above indicates low risk.
The tuning fork was struck against the palm of the testers hand so that it would vibrate for approximately 40 seconds and then applied to the base of the forehead so that participants could understand the concept of the vibration sensation. After asking the patient to close their eyes, the tuning fork was placed on the bony prominence at the dorsum of the first toe, which is just proximal to the nail bed. The participant was asked to report when the vibration stopped, and the tester dampened the tuning fork with the other hand. Scoring: 1 point was assigned for each correct perception of (vibration “on” or “off”). The procedure was performed twice on each foot in such a way that the participants could not anticipate the testers actions. This is a “rule out” test for the presence of neuropathy and does not indicate risks related to future onset.
This is adapted from the Canadian Diabetes Association, Rapid Screening for Diabetic Neuropathy [
Activities of daily living scores
Cardiac autonomic neuropathy
Diabetic neuropathy
Diabetic peripheral neuropathy
Hemoglobin A1C
Last observation carried forward
Negative predictive value
Norfolk Quality of Life Diabetic Neuropathy
Overweight, obese, inactive
Point of care device
Prediabetes
Positive predictive value
Sural nerve amplitude potential
Sural nerve conduction velocity
Type 1 diabetes
Type 2 diabetes
Quality of life.
An alternate version of this article was published as part of the doctoral dissertation of Jennifer J. Brown [
The authors declare that they have nothing to disclose and no conflicts of interest.
Jennifer J. Brown contributed to conception, design, analysis, interpretation, writing first draft, editing, and final approval. Sheri R. Colberg contributed to reviewing drafts, editing, and final approval. All remaining authors contributed to the study design, review of drafts, and final editing.
The authors thank the researchers at Eastern Virginia Medical School for providing the NC-stat DPNCheck and associated consumable supplies used for the project.