Sensory testing in type 2 diabetes: Insights into neuropathy and systemic complications

Abstract

Aims

Sensory tests such as vibration perception threshold (VPT), monofilament, pinprick, and temperature sensation are used to detect diabetic peripheral neuropathy (DPN). This study examined associations between these tests and chronic diabetic complications in type 2 diabetes, focusing on the clinical relevance of VPT.

Methods

In this retrospective study, 652 patients with type 2 diabetes evaluated at a tertiary center (2014–2024) underwent standardized neurological assessments. Based on VPT, patients were categorized as normal (<16 V), borderline (16–24 V), or abnormal (≥25 V). Associations between sensory tests and retinopathy, nephropathy, ischemic heart disease (IHD), cerebrovascular disease (CeVD), and diabetic foot ulcer (DFU) were examined in cross-sectional and longitudinal analyses.

Results

Abnormal VPT was present in 31.3% of patients and was independently associated with all five complications after multivariable adjustment: IHD (OR: 2.15), CeVD (OR: 3.31), DFU (OR: 15.94), retinopathy (OR: 7.66), and nephropathy (OR: 2.56). Borderline VPT was also independently associated with several outcomes. In a longitudinal subset (n=221; median follow-up ∼5 years), abnormal VPT was associated with incident IHD, DFU, retinopathy, and nephropathy.

Conclusion

VPT showed strong associations with microvascular and macrovascular complications, supporting its role as a practical indicator of systemic disease burden in type 2 diabetes.

1. AIMS

Diabetic peripheral neuropathy (DPN) is a slowly progressive and frequently under-recognized complication of diabetes mellitus, affecting approximately 50% of individuals over their lifetime. ¹ It most commonly presents as distal symmetric sensorimotor polyneuropathy and, although early stages are often asymptomatic, progression can result in neuropathic pain, loss of protective sensation, foot ulceration, and lower-limb amputation. While chronic hyperglycemia is a key driver of neural injury, growing evidence implicates additional mechanisms, including oxidative stress, ischemia, polyol pathway activation, and the accumulation of advanced glycation end products (AGEs). ² Notably, DPN shares common risk factors and pathogenic pathways with other diabetes-related complications such as retinopathy, nephropathy, and cardiovascular disease, underscoring its role as a marker of overall vascular burden. ³ In this context, early detection and integrated management of DPN are essential not only for the prevention of limb-threatening complications but also for improving long-term outcomes in people living with diabetes.

A comprehensive assessment of DPN should incorporate multiple tests, as no single measure can fully capture the complexity of sensory dysfunction.^4,5 Vibration perception threshold (VPT), which evaluates large fiber function using tools such as a biothesiometer, is a reliable and quantitative indicator of neuropathy severity. Elevated VPT values have been strongly associated with an increased risk of foot ulceration, lower-limb amputation, and mortality in individuals with diabetes. ⁶ Moreover, VPT correlates well with abnormalities on nerve conduction studies and provides objective data that can be monitored over time.^7,8 However, as VPT predominantly reflects large fiber function, it should be complemented with additional assessments such as the 10-g monofilament test and ankle reflex evaluation, both of which also assess large fiber integrity, as well as pinprick and temperature sensation testing, which target small fiber function. This multimodal approach allows for earlier and more accurate detection of DPN and supports timely clinical intervention.

Among the available clinical tools for diabetic foot assessment, VPT stands out due to its quantitative nature, ease of use, patient acceptability, and strong predictive value. Notably, beyond its established role in detecting peripheral neuropathy, previous studies have demonstrated that elevated VPT values may also be associated with systemic complications such as cardiac autonomic neuropathy (CAN), chronic kidney disease (CKD), and hepatic fibrosis due to non-alcoholic fatty liver disease.^9–11 However, direct comparisons between VPT and other bedside modalities with respect to their associations with the full spectrum of microvascular and macrovascular complications remain limited in real-world clinical cohorts. Furthermore, whether borderline VPT (16–24 V) values carry a meaningfully elevated complication burden remains poorly characterized, as most studies have applied only the conventional ≥25 volts (V) threshold. In our study, we aimed to evaluate whether these commonly used bedside foot examination methods, including both large and small fiber assessments, can reflect the burden of microvascular and macrovascular complications in a large real-world cohort of patients with type 2 diabetes. Additionally, we sought to explore whether a three-category VPT stratification offers additional clinical value compared to a conventional binary classification, and to examine the longitudinal associations between sensory test findings and incident diabetic complications in a subset of patients with available follow-up data.

2. Methods

2.1. Study design

This retrospective study included patients with a diagnosis of type 2 diabetes who attended the Endocrinology Outpatient Clinic at Ankara University Faculty of Medicine between 2014 and 2024. The study was conducted in accordance with the principles outlined in the Declaration of Helsinki, and the study protocol was approved by the Institutional Ethics Committee (Approval No İ06-542-25).

Patients were excluded if they met any of the following criteria: presence of infectious or inflammatory diseases, active diabetic foot ulcer, history of limb amputation, use of neurotoxic medications, current treatment for neuropathy, vitamin B12 deficiency, presence of neuropathic disorders other than DPN and cognitive impairment.

2.2. Clinical and laboratory measurements

The baseline demographic and clinical data of patients including age, gender, smoking history (defined as current smoking, former smoking, or both), duration of diabetes, current medications regarding antidiabetic medications and statins were documented. Participants’ weight and height were measured for the calculation of BMI by dividing weight in kilograms by height in square meters.

Blood samples were collected from each patient in the morning, after 12 h of fasting period, for biochemical analysis including creatinine (Cr), blood urea nitrogen (BUN), uric acid, albumin, lipid profiles, fasting blood glucose (FBG), glycosylated hemoglobin A1c levels (HbA1c). Estimated glomerular filtration rate (eGFR) was calculated using the Modification of Diet in Renal Disease (MDRD) formula: [eGFR = 186 ×(Cr)^-1.154 × (age) ^{- 0.203}× (0.742 if female)]. Samples for the complete blood count analysis (with differential analysis) were collected in EDTA-anticoagulated tubes. Early morning urine samples were collected to measure urinary albumin and creatinine. The urinary albumin/creatinine ratio (UACR) was calculated as the ratio of urinary albumin concentration to creatinine concentration. All laboratory parameters were reported in Système International (SI) units.

2.3. Definitions

A diagnosis of type 2 diabetes was based on documented medical records or laboratory findings according to the following criteria: a fasting plasma glucose level ≥126 mg/dL (≥7.0 mmol/L), a 2-hour plasma glucose ≥200 mg/dL (≥11.1 mmol/L), during a 75 g oral glucose tolerance test, HbA1c ≥6.5% (≥48 mmol/mol), or a random plasma glucose ≥200 mg/dL (≥11.1 mmol/L), in the presence of classic hyperglycemic symptoms. ¹²

Hypertension was defined as a documented diagnosis or the presence of persistently elevated office blood pressure ≥140/90 mmHg on at least two separate occasions, or current use of antihypertensive medication.^13,14

Ischemic heart disease (IHD) was defined as a documented history of myocardial infarction, angina pectoris, or coronary revascularization (i.e., coronary angiography with stent placement or coronary artery bypass grafting). When available, supporting evidence from electrocardiography, echocardiography, or non-invasive ischemia testing was reviewed.

Cerebrovascular disease (CeVD) was defined as a documented history of stroke or transient ischemic attack, confirmed by neuroimaging when available.

Patients with a previously diagnosed diabetic foot ulcer (DFU) that had completely healed at the time of evaluation were also classified as having a positive history of DFU. DFU history was systematically verified using electronic medical records, including documentation from wound care clinics. Available vascular assessments (e.g., ankle–brachial index, Doppler ultrasonography, and angiography) were also reviewed when present.

Diabetic nephropathy (DN) was defined as the presence of either persistent albuminuria or reduced eGFR (<60 mL/min/1.73 m² for >3 months). Albuminuria was assessed using UACR; microalbuminuria was defined as two or more readings between 30–299 mg/g (3.4–33.8 mg/mmol), while values ≥300 mg/g (≥33.9 mg/mmol) were considered macroalbuminuria. Patients on renal replacement therapy due to diabetes-related kidney disease were considered to have macroalbuminuria, regardless of current UACR levels.^15,16 Available clinical records were reviewed to identify potential alternative causes of chronic kidney disease (e.g., obstructive uropathy, polycystic kidney disease, or other structural abnormalities), and patients with a documented non-diabetic etiology of renal impairment were excluded from the DN group.

Diabetic retinopathy (DRP) was diagnosed based on clinical findings during annual comprehensive dilated eye examinations conducted by an ophthalmologist. ⁵

2.4. DPN assessment

DPN diagnosis was based on clinical evaluation, including the presence of typical symptoms (e.g., numbness, tingling, burning, or pain in the extremities) and/or signs (e.g., loss of vibration, pressure, or temperature sensation; diminished ankle reflexes). All patients underwent a standardized neurological examination, including evaluation of deep tendon reflexes, monofilament testing, pinprick and temperature sensation assessments, and VPT measurements. ⁵

Cutaneous pressure perception was evaluated using the 10 g Semmes–Weinstein monofilament test applied to the seven sites of each lower extremity: the distal part of the hallux, 3rd and 5th toe, 1st, 3rd and 5th metatarsal heads, and heel. Responses were categorized as normal or abnormal. ¹⁷

Pinprick sensation was evaluated on both the dorsum and plantar aspects of the feet using a disposable safety pin applied with light pressure. Patients were asked to report whether the stimulus felt sharp or dull, and the response was recorded as normal or abnormal.

Temperature sensation was assessed using the Tip-Therm® device (Auchela GmbH, Germany), which has a metal (cold) and plastic (neutral) end; patients, with eyes closed, were asked to distinguish between the two sensations applied to the plantar surface of the foot. Responses were categorized as normal or abnormal.

VPT was assessed on the plantar surface of the hallux using a biothesiometer (Biomedical Instruments, OH, USA). Although the conventional threshold of 25 V has been widely used to define clinically significant large fiber impairment, ⁶ recent studies suggest that lower cut-offs may improve sensitivity for detecting early neuropathy. ¹⁸ Therefore, we stratified patients into three groups based on VPT values: <16 V (normal), 16–24 V (borderline), and ≥25 V (abnormal), allowing for a more nuanced evaluation of neuropathy severity.

Neuropathic symptoms were systematically recorded at each visit and included pain, burning, tingling, numbness, cold sensation, and other complaints. For the purpose of this analysis, the presence of any neuropathic symptom was defined as the report of at least one of these symptoms. Patients reporting none of the above symptoms were classified as asymptomatic.

2.5. Statistical analysis

All statistical analyses were conducted using IBM SPSS Statistics version 30 (IBM Corp., Armonk, NY, USA). The normality of continuous variables was assessed both visually (via histograms and Q–Q plots) and analytically using the Kolmogorov–Smirnov and Shapiro–Wilk tests. Descriptive statistics were expressed as mean ± standard deviation (SD) for normally distributed variables, median (interquartile range [IQR]) for non-normally distributed variables, and frequency (percentage) for categorical variables.

The Chi-square or Fisher’s exact test was used for categorical variables, depending on the expected cell counts. For comparisons between two groups, the Student’s t-test was used for normally distributed continuous variables, while the Mann–Whitney U test was applied for non-normally distributed ones. For the comparison of continuous variables across three or more independent groups, one-way analysis of variance (ANOVA) was performed if the assumption of normality was met. If the variable was not normally distributed, the Kruskal–Wallis test was used as a non-parametric alternative. For categorical variables, the Chi-square test was used, and in cases where the expected cell counts were low, Monte Carlo simulation was applied to obtain more accurate p-values. For ordinal variables, the Kruskal–Wallis test was conducted to evaluate differences between groups.

To assess the association between sensory test abnormalities and prevalent diabetic complications, univariate and multivariable logistic regression analyses were performed. Odds ratios (ORs) with 95% confidence intervals (CIs) were reported. Multivariable models were adjusted for age, gender, body mass index, diabetes duration, hypertension, smoking status, and HbA1c.

In addition, a longitudinal analysis was performed to evaluate the association between sensory test abnormalities and incident diabetic complications. Only patients who were free of all evaluated complications at baseline and with at least 1 year of follow-up were included in the analysis (n=221). Time-to-event analyses were conducted using Cox proportional hazards regression models, with follow-up duration defined as the time from baseline assessment to the first occurrence of the outcome or last follow-up. Univariate Cox regression models were constructed for each sensory test (VPT, monofilament, pinprick, and temperature perception), and hazard ratios (HRs) with 95% CIs were reported. Multivariable models were not performed due to the limited number of outcome events, in order to avoid model overfitting and unstable estimates.

Both Pearson and Spearman correlation tests were used to evaluate the relationship between continuous variables, depending on the distribution of the data.

A two-tailed p-value <0.05 was considered statistically significant.

3. Results

3.1. Baseline characteristics of the study population

A total of 652 patients with type 2 diabetes mellitus were included in the study. The mean age was 58.52 ± 8.92 years, and 53.8% were female. The mean BMI was 28.24 ± 4.52 kg/m². The median duration of diabetes was 7.00 years [IQR: 4.00–15.00]. The median HbA1c was 62.0 mmol/mol [IQR:50.0–79.0]. The mean total cholesterol level was 4.82 ± 1.26 mmol/L, LDL was 2.75 ± 0.97 mmol/L, HDL was 1.17 ± 0.33 mmol/L, and triglycerides were 1.69 [IQR: 1.22–2.36] mmol/L. Hypertension was present in 54.9% of patients, and 36.2% had a history of smoking. Regarding antidiabetic treatment, 55.2% were on oral antidiabetic drugs (OAD) alone, 11.0% on insulin alone, and 33.7% on combined therapy. The most frequently used medications were metformin (79.6%), Dipeptidyl peptidase-4 (DPP-4) inhibitors (43.4%), and sodium-glucose co-transporter 2 (SGLT-2) inhibitors (32.2%). ACE inhibitors or angiotensin II receptor blockers (ARBs) were used by 46.3% of the cohort, and statins by 52.1%. The prevalence of complications was as follows: IHD 23.6%, CeVD 9.4%, DFU 5.8%, DN 37.1%, and DRP 22.5%.

Additional clinical, laboratory, and treatment-related variables are also presented in Table 1.

Table 1.

Comparison of clinical, laboratory, and treatment variables across the VPT groups.

Variable	Total n=652	Normal VPT n=249 (38.2%)	Borderline VPT n=199 (30.5%)	Abnormal VPT n=204 (31.3%)	P Value
Gender, female, n (%)	351 (53.8)	139 (55.8)	117 (58.8)	95 (46.6)	0.035
Age, years (mean ± SD)	58.52 ± 8.92	54.57 ± 9.39	59.90 ± 7.18	61.99 ± 7.98	<0.001
BMI, kg/m 2 (mean ± SD)	28.24 ± 4.52	28.09 ± 4.55	28.15 ± 4.82	28.50 ± 4.18	0.595
Diabetes duration, years (median [IQR])	7.00 [4.00–15.00]	4.83 [2.42–10.00]	10.00 [4.00–15.00]	11.00 [5.00–20.00]	<0.001
HbA1c, %, (median [IQR])	7.80 [6.70–9.40]	7.30 [6.50–9.05]	7.90 [6.80–9.40]	8.30 [6.90–10.10]	<0.001
HbA1c, mmol/mol, (median [IQR])	62.0 [50.0–79.0]	56.0 [48.0–75.0]	63.0 [51.0–79.0]	67.0 [52.0–87.0]	<0.001​
FBG, mmol/L (median [IQR])	7.78 [6.22–10.89]	7.36 [6.04–9.44]	7.56 [6.06–10.86]	8.39 [6.56–12.39]	<0.001
eGFR, mL/min/1.73 m 2 (median [IQR])	87.78 [71.01–105.03]	95.16 [78.62–115.19]	86.08 [72.15–103.20]	75.84 [60.40–95.33]	<0.001
BUN, mmol/L (median [IQR])	5.0 [3.93–6.43]	4.3 [3.6–5.4]	5.4 [4.3–6.8]	6.1 [5.0–8.6]	<0.001
Cr, µmol/L (median [IQR])	69.0 [57.5–83.1]	64.5 [55.7–73.4]	69.0 [56.6–81.3]	78.7 [62.8–97.2]	<0.001
UACR, mg/mmol (median [IQR])	1.36 [0.57–4.52]	0.9 [0.45–2.06]	1.24 [0.55–4.13]	3.22 [1.13–21.72]	<0.001
Uric acid, µmol/L (median [IQR])	291.5 [237.9–356.9]	285.5 [237.9–336.1]	285.5 [234.9–362.8]	309.3 [249.8–383.6]	0.001
Albumin, g/L (median [IQR])	44.00 [42.00–46.00]	44.00 [42.00–46.00]	45.00 [43.00–47.00]	43.00 [40.00–45.00]	<0.001
Total cholesterol, mmol/L, (mean ± SD)	4.82 ± 1.26	4.92 ± 1.11	4.89 ± 1.54	4.62 ± 1.12	0.036
Triglycerides, mmol/L, (median [IQR])	1.69 [1.22–2.36]	1.59 [1.19–2.35]	1.79 [1.29–2.42]	1.66 [1.21–2.34]	0.495
HDL cholesterol, mmol/L, (mean ± SD)	1.17 ± 0.33	1.16 ± 0.31	1.2 ± 0.33	1.14 ± 0.36	0.187
LDL cholesterol, mmol/L, (mean ± SD)	2.75 ± 0.97	2.88 ± 0.94	2.7 ± 0.98	2.65 ± 0.99	0.029
WBC, ×10 9 /L (mean ± SD)	8.30 ± 2.31	8.10 ± 2.16	8.23 ± 2.25	8.60 ± 2.50	0.058
PLT, ×10 9 /L (mean ± SD)	266.52 ± 72.85	271.84 ± 69.73	266.40 ± 74.26	260.15 ± 75.00	0.236
Hb, g/dL (mean ± SD)	13.88 ± 1.68	14.09 ± 1.62	14.04 ± 1.65	13.48 ± 1.71	<0.001
TSH, µIU/mL (median [IQR])	1.76 [1.15–2.68]	1.73 [1.19–2.55]	1.73 [1.12–2.77]	1.81 [1.06–2.72]	0.973
Any neuropathic symptom, n (%)	​	​	​	​	<0.001
No	410 (62.9)	179 (71.9)	153 (76.9)	78 (38.2)	​
Yes	242 (37.1)	70 (28.1)	46 (23.1)	126 (61.8)	​
VPT (right), Volts (mean ± SD)	21.65 ± 11.19	12.02 ± 2.01	19.21 ± 2.56	35.79 ± 8.36	<0.001
VPT (left), Volts (mean ± SD)	21.80 ± 11.37	12.42 ± 2.96	19.54 ± 2.42	35.46 ± 9.19	<0.001
Monofilament test, n (%)	​	​	​	​	<0.001
Normal	409 (62.7)	177 (71.1)	155 (77.9)	77 (37.7)	​
Abnormal	243 (37.3)	72 (28.9)	44 (22.1)	127 (62.3)	​
Pinprick sensation, n (%)	​	​	​	​	<0.001
Normal	441 (67.6)	181 (72.7)	165 (82.9)	95 (46.6)	​
Abnormal	211 (32.4)	68 (27.3)	34 (17.1)	109 (53.4)	​
Temperature sensation, n (%)	​	​	​	​	<0.001
Normal	435 (66.7)	183 (73.5)	155 (77.9)	97 (47.5)	​
Abnormal	217 (33.3)	66 (26.5)	44 (22.1)	107 (52.5)	​
Hypertension, n (%)	358 (54.9)	113 (45.4)	114 (57.3)	131 (64.2)	<0.001
Smoking, n (%)	236 (36.2)	96 (38.6)	65 (32.7)	75 (36.8)	0.427
Diabetes treatment, n (%)	​	​	​	​	<0.001
OAD	360 (55.2)	162 (65.1)	114 (57.3)	84 (41.2)	​
Insulin	72 (11.0)	26 (10.4)	12 (6.0)	34 (16.7)	​
OAD + Insulin	220 (33.7)	61 (24.5)	73 (36.7)	86 (42.2)	​
Metformin, n (%)	519 (79.6)	199 (79.9)	171 (85.9)	149 (73.0)	0.006
Sulfonylureas, n (%)	78 (12.0)	28 (11.2)	25 (12.6)	25 (12.3)	0.902
Thiazolidinediones, n (%)	53 (8.1)	14 (5.6)	29 (14.6)	10 (4.9)	<0.001
DPP-4 inhibitors, n (%)	283 (43.4)	85 (34.1)	112 (56.3)	86 (42.2)	<0.001
SGLT-2 inhibitors, n (%)	210 (32.2)	63 (25.3)	84 (42.2)	63 (30.9)	<0.001
GLP-1 receptor analogs, n(%)	21 (3.2)	11 (4.4)	3 (1.5)	7 (3.4)	0.218
ACE inhibitors/ARBs, n (%)	302 (46.3)	85 (34.1)	115 (57.8)	102 (50.0)	<0.001
Statin, n (%)	340 (52.1)	124 (49.8)	117 (58.8)	99 (48.5)	0.076
Ischemic heart disease, n (%)	154 (23.6)	30 (12.0)	49 (24.6)	75 (36.8)	<0.001
Cerebrovascular disease, n (%)	61 (9.4)	9 (3.6)	13 (6.5)	39 (19.1)	<0.001
Diabetic foot ulcer, n (%)	38 (5.8)	2 (0.8)	8 (4.0)	28 (13.7)	<0.001
Diabetic nephropathy, n (%)	242 (37.1)	58 (23.3)	75 (37.7)	109 (53.4)	<0.001
Diabetic retinopathy, n (%)	147 (22.5)	15 (6.0)	41 (20.6)	91 (44.6)	<0.001

ARBs: Angiotensin II Receptor Blockers, BUN: Blood Urea Nitrogen, Cr: Creatinine, DPP-4 inhibitors: Dipeptidyl Peptidase-4 Inhibitors, eGFR: Estimated Glomerular Filtration Rate, FBG: Fasting Blood Glucose, GLP-1 receptor analogs: Glucagon-Like Peptide-1 Analogs, HbA1c: Hemoglobin A1c (Glycated Hemoglobin), IQR: Interquartile Range, OAD: Oral Antidiabetic Drug, PLT: Platelet Count, SD: Standard Deviation, SGLT-2 inhibitors: Sodium–Glucose Cotransporter-2 Inhibitors, TSH: Thyroid-Stimulating Hormone, UACR: Urine Albumin-to-Creatinine Ratio, VPT: Vibration Perception Threshold, WBC: White Blood Cell Count.

3.2. Comparison across VPT categories

Participants were categorized into normal (n = 249), borderline (n = 199), and abnormal (n = 204) VPT groups. Age, diabetes duration, HbA1c, and fasting blood glucose levels increased significantly across VPT categories (p < 0.001). Renal function worsened with higher VPT: median eGFR decreased from 95.16 to 75.84 mL/min/1.73 m², while BUN, creatinine, UACR, and uric acid increased (p ≤ 0.001 for all). Serum albumin was lowest in the abnormal group (p < 0.001). Hemoglobin levels were lower in the abnormal VPT group (13.48 ± 1.71 g/dL) compared to the normal and borderline groups (14.09 ± 1.62 and 14.04 ± 1.65 g/dL, respectively) (p < 0.001). Any neuropathic symptom was present in 37.1% of the study population, with the highest prevalence observed in the abnormal VPT group (61.8%) compared to the normal (28.1%) and borderline (23.1%) groups (p<0.001). The frequencies of abnormal findings in the monofilament (62.3%), pinprick (53.4%), and temperature sensation (52.5%) tests were highest in the abnormal VPT group (p < 0.001). The prevalence of hypertension increased across the groups (from 45.4% to 64.2%, p < 0.001), while smoking rates did not differ significantly. Use of insulin therapy (alone or combined), DPP-4 inhibitors, SGLT-2 inhibitors, thiazolidinediones, and ACE inhibitors/ARBs increased with higher VPT (p < 0.001 for all). Diabetic complications were more prevalent in the abnormal VPT group: IHD (36.8%), CeVD (19.1%), DFU (13.7%), DN (53.4%), and DRP (44.6%) (p < 0.001 for all).

Table 1 presents the full comparison of clinical, laboratory, and treatment variables across the VPT groups.

3.3. Correlation between VPT and renal function parameters

There was a moderate negative correlation between VPT and eGFR (ρ = –0.309, p < 0.001) (Figure 1), while VPT showed moderate positive correlations with BUN (ρ = 0.433, p < 0.001), serum Cr (ρ = 0.296, p < 0.001), and UACR (ρ = 0.340, p < 0.001).

Figure 1.

Negative correlation of eGFR and VPT values.

3.4. Association between DPN assessment tools and diabetic complications

In univariate analysis, having any neuropathic symptoms was significantly associated with IHD (OR 1.878, 95% CI 1.303–2.706; p = 0.001), CeVD (OR 1.970, 95% CI 1.161–3.354; p = 0.012), DFU (OR 2.542, 95% CI 1.300–4.971; p = 0.006), DRP (OR 2.553, 95% CI 1.742–3.684; p = 0.001), and DN (OR 1.626, 95% CI 1.176–2.249; p = 0.003). Similarly, sensory tests were significantly associated with various diabetic complications, though the magnitude and consistency of associations were strongest for VPT. Patients with borderline VPT had significantly increased odds of IHD (OR: 2.39, 95% CI: 1.45–3.33, p<0.001), DFU (OR: 5.17, 95% CI: 1.09–24.64, p=0.039), DRP (OR: 4.05, 95% CI: 2.17–7.56, p<0.001) and DN (OR: 1.99, 95% CI: 1.32–3.00, p=0.001); the association with CeVD was not statistically significant (OR: 1.86, 95% CI: 0.78–4.45, p=0.161). Abnormal VPT showed strong and highly significant associations with all five complications, including IHD (OR: 4.24, 95% CI: 2.64–6.83, p<0.001), CeVD (OR: 6.30, 95% CI: 2.97–13.36, p<0.001), DFU (OR: 19.65, 95% CI: 4.62–83.55, p<0.001), DRP (OR: 12.56, 95% CI: 6.96–22.68, p<0.001), and DN (OR: 3.78, 95% CI: 2.53–5.65, p<0.001). Among the other modalities, monofilament testing was significantly associated with each complication (ORs ranging from 1.69 to 2.75, all p≤0.005), while pinprick and temperature sensation tests were also associated with most outcomes, especially DRP, DN, IHD, and CeVD, though the strength of these associations was generally lower, and pinprick testing failed to reach significance for DFU (p=0.097) (Table 2).

Table 2.

Univariate analysis of sensory tests associated with diabetic complications.

Sensory test	IHD OR (95% CI)	CeVD OR (95% CI)	DFU OR (95% CI)	Diabetic retinopathy OR (95% CI)	Diabetic Nephropathy OR (95% CI)
VPT (Borderline)	2.385 (1.447 – 3.930)	1.864 (0.780 – 4.454)	5.173 (1.086 – 24.640)	4.048 (2.167 – 7.562)	1.992 (1.321 – 3.003)
p value	<0.001	0.161	0.039	<0.001	0.001
VPT (Abnormal)	4.244 (2.637 – 6.832)	6.303 (2.973 – 13.362)	19.648 (4.620 -83.553)	12.563 (6.960 – 22.675)	3.778 (2.526 – 5.651)
p value	<0.001	<0.001	<0.001	<0.001	<0.001
Monofilament (Abnormal)	1.931 (1.338 – 2.786)	2.138 (1.257 – 3.636)	2.746 (1.404 – 5.372)	2.598 (1.786 – 3.780)	1.688 (1.218 – 2.340)
p value	<0.001	0.005	0.003	<0.001	0.002
Pinprick (Abnormal)	1.963 (1.352 – 2.849)	2.540 (1.492 – 4.325)	1.753 (0.904 – 3.397)	2.179 (1.494 – 3.179)	1.591 (1.137 – 2.225)
p value	<0.001	<0.001	0.097	<0.001	0.007
Temperature (Abnormal)	1.603 (1.105 – 2.328)	2.091 (1.229 – 3.556)	2.634 (1.359 – 5.105)	1.841 (1.263 – 2.684)	1.764 (1.263 – 2.462)
p value	0.013	0.006	0.004	0.002	<0.001

CeVD: Cerebrovascular disease, DFU: Diabetic foot ulcer, IHD: Ischemic heart disease, OR: Odds ratio, VPT: Vibration perception threshold.

In multivariable logistic regression models adjusted for age, gender, BMI, diabetes duration, hypertension, smoking, and HbA1c, having any neuropathic symptoms was not significantly associated with IHD (OR 1.420, 95% CI 0.924–2.198; p = 0.123), CeVD (OR 0.843, 95% CI 0.350–1.948; p = 0.698), DFU (OR 1.534, 95% CI 0.735–3.094; p = 0.232), DRP (OR 1.853, 95% CI 0.967–3.550; p = 0.063), or DN (OR 1.142, 95% CI 0.909–1.435; p = 0.254). Among sensory tests, VPT remained the most robust and independent factor associated with diabetic complications. Borderline VPT was significantly associated with increased risk of IHD (OR: 1.92, 95% CI: 1.12–3.29, p=0.017), DFU (OR: 5.20, 95% CI: 1.05–25.81, p=0.044), DRP (OR: 3.68, 95% CI: 1.90–7.11, p<0.001), and DN (OR: 1.71, 95% CI: 1.10–2.65, p=0.017), but not with CeVD (p=0.427). Abnormal VPT showed strong and consistent associations with all five complications, including IHD (OR: 2.15, 95% CI: 1.22–3.79, p=0.008), CeVD (OR: 3.31, 95% CI: 1.40–7.83, p=0.007), DFU (OR: 15.94, 95% CI: 3.38–75.06, p<0.001), retinopathy (OR: 7.66, 95% CI: 3.95–14.89, p<0.001), and DN (OR: 2.56, 95% CI: 1.59–4.12, p<0.001). Among the other sensory tests, only monofilament testing retained a significant association with DRP (OR: 1.75, 95% CI: 1.03–2.97, p=0.039); all other associations with monofilament, pinprick, and temperature sensation were nonsignificant after adjustment for confounders (Table 3).

Table 3.

Multivariable analysis of sensory tests associated with diabetic complications^*.

Sensory test	IHD OR (95% CI)	CeVD OR (95% CI)	DFU OR (95% CI)	Diabetic retinopathy OR (95% CI)	Diabetic nephropathy OR (95% CI)
VPT (Borderline)	1.921 (1.123 – 3.285)	1.445 (0.583 – 3.585)	5.197 (1.046 – 25.813)	3.679 (1.904 – 7.109)	1.707 (1.099 – 2.651)
p value	0.017	0.427	0.044	<0.001	0.017
VPT (Abnormal)	2.149 (1.218 – 3.790)	3.305 (1.395 – 7.826)	15.938 (3.384 – 75.055)	7.663 (3.949 – 14.889)	2.562 (1.592 – 4.124)
p value	0.008	0.007	<0.001	<0.001	<0.001
Monofilament (Abnormal)	1.226 (0.719 – 2.088)	0.924 (0.449 – 1.902)	1.722 (0.726 – 4.081)	1.748 (1.028 – 2.972)	1.133 (0.698 – 1.838)
p value	0.625	0.830	0.217	0.039	0.614
Pinprick (Abnormal)	1.596 (0.904 – 2.817)	1.740 (0.816 – 3.711)	0.567 (0.227 – 1.412)	1.327 (0.752 – 2.342)	0.998 (0.589 – 1.690)
p value	0.107	0.152	0.223	0.329	0.994
Temperature (Abnormal)	0.876 (0.515 – 1.490)	1.048 (0.514 – 2.137)	1.711 (0.741 – 3.950)	0.922 (0.541 – 1.572)	1.329 (0.824 – 2.144)
p value	0.625	0.898	0.208	0.766	0.244

*Adjusted for age, gender, BMI, diabetes duration, HT, smoking, and HbA1c. Each sensory test was evaluated in a separate multivariable model. CeVD: Cerebrovascular disease, DFU: Diabetic foot ulcer, IHD: Ischemic heart disease, OR: Odds ratio, VPT: Vibration perception threshold.

3.5. Longitudinal analysis of DPN assessment tools and diabetic complications

A total of 221 patients without baseline complications and with at least 1 year of follow-up data were included in the longitudinal analysis. The median follow-up duration was 5.2 years (IQR: 2.45–6.80). The mean age of the cohort was 55.97 ± 9.59 years. The median HbA1c level was 58 mmol/mol (IQR: 48–76).

During follow-up, diabetic complications developed in 18 patients (8.1%) for IHD, 9 patients (4.1%) for CeVD, 6 patients (2.7%) for DFU, 30 patients (13.6%) for DRP, and 27 patients (12.2%) for DN. In the longitudinal analysis, having any neuropathic symptoms was not significantly associated with IHD (HR 1.987, 95% CI 0.678–5.827; p = 0.211), CeVD (HR 2.423, 95% CI 0.581–9.312; p = 0.279), DFU (HR 1.265, 95% CI 0.564–3.392; p = 0.612), DRP (HR 1.911, 95% CI 0.682–5.355; p = 0.218), or DN (HR 2.119, 95% CI 0.767–4.212; p = 0.114). Nevertheless, VPT was significantly associated with an increased risk of several complications, including IHD (HR: 3.44, 95% CI: 1.03–11.54, p=0.045), DFU (HR: 15.25, 95% CI: 2.45–95.24, p=0.004), DRP (HR: 3.19, 95% CI: 1.24–8.23, p=0.016), and DN (HR: 3.55, 95% CI: 1.31–9.55, p=0.012), whereas its association with CeVD was not statistically significant (p=0.172).Borderline VPT was associated with DN (HR 2.98, 95% CI 1.24–7.15, p=0.015), while other associations did not reach statistical significance. No consistent significant associations were observed for monofilament, pinprick, or temperature tests across the evaluated outcomes (Table 4).

Table 4.

Univariate Cox analysis of sensory tests associated with diabetic complications^*.

Sensory test	IHD HR (95% CI)	CeVD HR (95% CI)	DFU HR (95% CI)	Diabetic retinopathy HR (95% CI)	Diabetic nephropathy HR (95% CI)
VPT (Borderline)	2.770 (0.951 – 8.052)	2.583 (0.574 –11.635)	2.010 (0.181 –22.322)	2.073 (0.881 – 4.864)	2.981 (1.242– 7.154)
p value	0.061	0.217	0.571	0.097	0.015
VPT (Abnormal)	3.443 (1.032 – 11.541)	3.302 (0.602 – 18.273)	15.247 (2.445 –95.241)	3.191 (1.242 – 8.234)	3.546 (1.311 –9.548)
p value	0.045	0.172	0.004	0.016	0.012
Monofilament (Abnormal)	2.287 (0.463 –11.443)	2.612 (0.671 – 10.212)	1.182 (0.444 – 3.186)	1.812 (0.873 – 3.784)	2.091 (0.979 – 4.529)
p value	0.316	0.168	0.741	0.113	0.060
Pinprick (Abnormal)	2.364 (0.472 –11.922)	2.687 (0.671 –10.722)	1.132 (0.414 – 3.116)	1.520 (0.721 – 3.232)	2.160 (0.991 – 4.703)
p value	0.299	0.163	0.815	0.275	0.054
Temperature (Abnormal)	2.53 (0.502– 12.781)	1.815 (0.456 – 7.236)	1.694 (0.632 – 4.501)	2.027 (0.971 – 4.221)	0.417 (0.163 – 1.072)
p value	0.261	0.402	0.295	0.061	0.069

*Analysis was performed in the subset of 221 patients with available follow-up data.

CeVD: Cerebrovascular disease, DFU: Diabetic foot ulcer, IHD: Ischemic heart disease, HR: Hazard ratio, VPT: Vibration perception threshold.

4. Discussion

In clinical practice, nerve conduction studies (NCS) are considered the gold standard for diagnosing neuropathy, however, their use is limited by cost and accessibility. ¹⁹ Therefore, current guidelines recommend simpler bedside tests, including pinprick, temperature sensation, VPT, 10-g monofilament, and ankle reflexes, to assess small- and large-fiber function. ⁵ Our study evaluated the associations between these routinely used sensory tests and chronic diabetic complications in a large real-world cohort of patients with type 2 diabetes. The mean age of the patients was 58.5 years, and the median diabetes duration was 7 years. Approximately one-third of the patients exhibited abnormal findings in at least one sensory test, with 31.3% showing abnormal VPT.

Our findings highlight that abnormal VPT is closely linked to broader vascular and metabolic disturbances commonly observed in individuals with type 2 diabetes. Patients with abnormal VPT were more likely to be male, older, have a longer duration of diabetes, a history of hypertension, poorer glycemic control, significantly impaired renal function, and a higher prevalence of neuropathic symptoms compared to those with normal or borderline VPT. In addition, serum uric acid levels were significantly higher, while serum albumin levels were lower in patients with abnormal VPT. These results align with existing literature indicating that DPN is strongly associated with prolonged diabetes duration, inadequate glycemic control, and worse renal parameters.^20,21 The seemingly favorable association between LDL-cholesterol and VPT in our study may reflect the high prevalence of statin use, particularly among individuals with more advanced disease receiving intensive lipid-lowering therapy, and is consistent with prior studies reporting no clear relationship between serum lipid levels and large-fiber nerve function. ²² Similarly, the lack of a significant association between smoking and VPT may be attributable to limitations of self-reported smoking data, which did not account for intensity or duration and may have obscured a potential dose–response effect.

Prior studies examining the relationship between diabetic neuropathy and systemic complications have employed a range of assessment methods, including NCS, vibration-based measures, questionnaire-based screening instruments, and monofilament testing, which limits direct comparability across cohorts. Nevertheless, converging evidence from these methodologically diverse studies supports an association between diabetic neuropathies and both macrovascular and microvascular outcomes. Among diabetic neuropathies, CAN is well established as an independent predictor of adverse cardiovascular outcomes, including silent myocardial ischemia, arrhythmias, and sudden cardiac death.^23–25 However, the association between DPN and cardiovascular disease outcomes in patients with type 2 diabetes remains less well defined and has received comparatively less attention. DPN is a recognized risk factor for foot ulceration and amputation, but emerging evidence also implicates it in macrovascular complications and increased cardiovascular mortality. DPN has been shown to be associated with coronary and medial arterial calcification; both predictors of IHD and all-cause mortality in type 2 diabetes. ²⁶ Coppini et al. reported that elevated VPT values were more strongly linked to cardiovascular mortality than other microvascular complications. ²⁷ Similarly, a prospective study of 1,345 patients with diabetes demonstrated that impaired VPT was independently associated with premature cardiovascular death after multivariable adjustment (OR: 7.8), outperforming other clinical markers including retinopathy, and remained significantly associated with peripheral artery disease and cerebrovascular events in subgroup analyses. ²⁸ Supporting this, a large study by Chung et al. (n = 1,041) using NCS found that DPN was significantly associated with multiple macrovascular risk factors and with higher cardiovascular disease prevalence. ²⁹ Brownrigg et al. demonstrated in a large primary care cohort of 13,043 individuals with type 2 diabetes free of cardiovascular disease at baseline that peripheral neuropathy, assessed via monofilament testing, was independently associated with incident cardiovascular events after adjustment for standard risk factors (HR: 1.33). ³⁰ Similarly, Bjerg et al., using data from two Danish cohorts (n = 6,473), and employing the Michigan Neuropathy Screening Instrument questionnaire (MNSIq; score ≥4), reported that DPN early in type 2 diabetes was associated with a markedly higher incidence rate of cardiovascular diseases beyond established risk factors, with a combined incidence rate ratio of 1.65. ³¹ Furthermore, a large prospective cohort study by Kim et al. utilizing the UK Biobank (n = 459,127), in which polyneuropathy was identified through nurse-led interviews and ICD diagnostic codes, found that polyneuropathy was significantly associated with all-cause and cardiovascular mortality across all diabetes statuses, with the highest cardiovascular mortality risk observed in those with type 2 diabetes and polyneuropathy (HR: 3.77; 95% CI: 2.62–5.43); HbA1c, abdominal obesity, and C-reactive protein were identified as significant mediators of this association. ³² In agreement with current literature, we found that abnormal VPT was independently associated with IHD (OR: 2.15), CeVD (OR: 3.31), and DFU (OR: 15.94) after multivariable adjustment, with the strongest association observed for DFU. While other sensory tests contributed information regarding cardiovascular complications in univariate analyses, only VPT retained consistent and significant associations across all five outcomes after adjustment for major confounders, and these findings were further supported in the longitudinal setting, where abnormal VPT was associated with incident IHD and DFU.

The coexistence of chronic microvascular complications of diabetes is frequently attributable to shared risk factors, including advanced age, prolonged disease duration, hypertension, dyslipidemia, smoking, and poor glycemic control. ³³ These complications are the result of common pathophysiological mechanisms involving oxidative stress, mitochondrial dysfunction, and metabolic dysregulation.^34,35 Large-scale studies in both type 1 and type 2 diabetes have consistently demonstrated significant overlap among these complications. For instance, the EURODIAB Prospective Complications Study and a Swedish population-based study revealed strong associations among neuropathy, nephropathy, and retinopathy.^36,37 Supporting the notion of complication clustering, Nabrdalik et al., using the MNSI (score >2), found that nearly half of patients with DPN had concurrent cardiovascular disease compared with approximately one-fifth of those without DPN, and that DPN was independently associated with diabetic kidney disease. ³⁸ Interestingly, a recent long-term study in individuals with type 1 diabetes and end-stage kidney disease undergoing pancreas transplantation demonstrated that persistently abnormal VPT values were independently associated with both pancreas graft failure and incident cardiovascular events over a median follow-up of 11.3 years. Importantly, VPT values significantly improved post-transplant, suggesting their responsiveness to metabolic correction. ³⁹ In our study, both borderline and abnormal VPT values were independently associated with diabetic DRP and DN. Compared to normal VPT, borderline VPT was associated with DRP (OR: 3.68) and DN (OR: 1.71), while abnormal VPT was associated with even higher odds DRP OR: 7.66; DN OR: 2.56). In addition, VPT showed a moderate negative correlation with eGFR (ρ = –0.309), and positive correlations with BUN (ρ = 0.433), serum creatinine (ρ = 0.296), and UACR (ρ = 0.340), supporting its relevance to renal function impairment. These findings align with recent literature. Lindholm et al. demonstrated that large fiber neuropathy, assessed by multifrequency vibrometer, developed after retinopathy but before nephropathy in type 1 diabetes mellitus, and was significantly associated with both. ⁴⁰ Similarly, Zhang et al. reported that elevated VPT in type 2 diabetes was linked to both glomerular (UACR, urinary IgG) and tubular (α1-microglobulin) injury markers, with stronger associations observed for UACR than for eGFR. ¹¹ Extending these cross-sectional observations, our longitudinal analysis demonstrated that abnormal VPT was associated with incident DRP (HR: 3.19) and DN (HR: 3.55). Notably, even borderline VPT was associated with incident DN (HR: 2.98), suggesting that the risk of renal deterioration may begin to accumulate at intermediate levels of large fiber dysfunction. Altogether, these data support the utility of VPT as a surrogate marker for detecting early microvascular disease.

Among the other sensory tests, only monofilament testing retained a significant association with DRP (OR: 1.75, p=0.039); all other associations with monofilament, pinprick, and temperature sensation were nonsignificant after adjustment for confounders. This may partly be explained by methodological and physiological differences between the tests. Although both monofilament and VPT assess large fiber function, the monofilament test provides only a binary outcome and is primarily designed to detect advanced neuropathy characterized by loss of protective sensation. ¹⁷ In contrast, VPT allows for quantitative assessment across a spectrum of nerve dysfunction and has shown stronger correlations with early neuropathy and nerve conduction abnormalities.^7,8,41 Moreover, the sensitivity of the monofilament test is limited in detecting mild neuropathy and its reproducibility is moderate, with results susceptible to variation due to skin condition, temperature, and operator technique.^42,43 These limitations likely contributed to its reduced explanatory value in multivariable models compared to VPT.

The absence of significant associations between diabetic complications and pinprick or temperature perception may be explained by the fact that these tests primarily assess small fiber function, which may not align temporally or mechanistically with the development of vascular complications. Small fiber neuropathy typically emerges earlier in the disease course, often presenting with pain, dysesthesia, or thermal sensitivity changes. ⁴⁴ However, it may occur independently of the onset of microvascular complications and large fiber damage, potentially limiting its relevance as an indicator of systemic diabetic outcomes. Moreover, these tests are semi-quantitative, subjective, and less standardized compared to large-fiber assessments. Gold-standard methods such as intraepidermal nerve fiber density (IENFD) or quantitative sensory testing (QST) are more sensitive to subtle small fiber changes but were not employed in our study. ⁴⁵ Therefore, the limited associations observed likely reflect both the potential independence of small fiber involvement from other diabetic complications and the suboptimal sensitivity of the assessment tools used—rather than a true lack of clinical relevance.

This study has several notable strengths. It includes a large and well-characterized cohort of individuals with type 2 diabetes, enabling robust statistical analyses. The comprehensive assessment of diabetic peripheral neuropathy using both large fiber and small fiber modalities enhances diagnostic accuracy and reflects real-world clinical practice in line with current guidelines. A particularly valuable methodological aspect is the stratification of VPT values into three clinically meaningful categories—normal (<16 V), borderline (16–24 V), and abnormal (≥25 V)—based on both traditional thresholds and emerging evidence from recent studies. This allows for a more granular evaluation of neuropathy severity and its associations with systemic complications. The consistent and independent relationships observed between abnormal VPT and outcomes underscore the potential of VPT as a surrogate marker for vascular dysfunction. Furthermore, the inclusion of a longitudinal analysis in a subset of patients with available follow-up data represents an additional strength, as it extends the cross-sectional findings and provides preliminary evidence for the longitudinal association between VPT and incident diabetic complications.

Nonetheless, several limitations should be acknowledged. The primary analysis is retrospective and cross-sectional in design, which precludes causal inference; although a longitudinal analysis was performed in a subset of patients, the limited number of outcome events precluded multivariable adjustment in this component of the study. The use of a single tertiary care center may introduce selection bias toward more advanced disease. Although tests for small fiber function were included, more sensitive and quantitative techniques were not employed. NCS, considered the gold standard for large fiber assessment, were also not available. Additionally, data on alcohol consumption were not systematically collected, which may represent a potential confounding factor given the known association between alcohol intake and peripheral neuropathy. Despite these limitations, the findings provide clinically relevant insights and lay the groundwork for prospective studies incorporating longitudinal outcomes and advanced neurophysiological testing.

In conclusion, our findings demonstrate that VPT is a valuable and practical tool not only for detecting DPN but also for identifying individuals at higher risk of chronic diabetic complications. The stratification of VPT into clinically meaningful categories improved its capacity to distinguish patients with a higher vascular burden in type 2 diabetes. While small fiber assessments yielded limited associations, the strong and consistent links observed with VPT in both cross-sectional and longitudinal analyses highlight the importance of incorporating large fiber evaluations into routine diabetic foot screening. These findings suggest that VPT may serve not only as a diagnostic indicator but also as a clinically useful marker of future complication risk in type 2 diabetes. Future prospective studies with larger cohorts and standardized neurophysiological assessments are warranted to validate and extend these findings.

ORCID iDs

Sinem Başak Tan Öksüz https://orcid.org/0000-0002-6969-6128

Emine Selin Yıldırım https://orcid.org/0009-0002-6340-9201

Ethical considerations

The study was conducted in accordance with the principles outlined in the Declaration of Helsinki, and the study protocol was approved by the Institutional Ethics Committee (Approval No İ06-542-25).

Consent to participate

This retrospective study used anonymized patient data and was approved by the institutional ethics committee; informed consent was not required.

Data Availability Statement

Data supporting the findings of this study are available from the corresponding author upon reasonable request.*