Association of HbA1c Variability with Vibrating Perception Threshold in Middle-Aged and Elderly Patients with Type 2 Diabetes Mellitus: A Retrospective Cohort Study

Introduction

The prevalence rate of type 2 diabetes mellitus (T2DM) among adults in China is as high as 12.8%,¹ establishing it as the country with the highest diabetic population globally. The prevalence of diabetic peripheral neuropathy (DPN)² is remarkably high among individuals with diabetes mellitus, making it one of the most prevalent complications associated with this metabolic disorder. Due to the insidious onset and slow progression of DPN, clinical symptoms often manifest when irreversible pathological changes have already occurred in most peripheral nerves, thereby predisposing individuals to foot ulcers and an elevated risk of amputation. According to the Standards of Medical Care in Diabetes published by the American Diabetes Association (ADA), the prevalence of DPN is reported to be 13.1%.³ DPN significantly impacting the overall quality of life for individuals with diabetes mellitus.

Currently, the screening methods commonly employed for DPN encompass vibration sensory threshold detection, temperature threshold detection, pressure threshold detection, and ankle reflex detection and so on. The Vibrating Perception Threshold (VPT) detection⁴ is primarily employed for assessing lesions in vibrating perception, and reflecting the functionality of large fiber. Due to its non-invasive nature, high accuracy, and convenient usability, VPT has emerged as a primary modality for clinical diagnosis and evaluation of diabetic peripheral neuropathy both domestically and internationally. Ponirakis et al⁵ conducted a comprehensive assessment of the neurological function in 144 subjects, revealing that VPT exhibited the strongest correlation with sensory nerve conduction velocity and diagnostic accuracy.

In addition to glycemic control, it is imperative not to overlook the impact of blood glucose variability on the incidence and progression of chronic complications associated with diabetes mellitus. Blood glucose variability,⁶ also referred to as glycemic fluctuations, encompasses both short-term and long-term variations in blood glucose levels. Short-term fluctuations in blood sugar levels reflect daily and/or intraday variations, whereas long-term fluctuations are typically determined by multiple measurements of glycated hemoglobin A1c(HbA1c) or fasting plasma glucose(FPG) over an extended period, indicating changes in blood sugar levels over months to years. The findings of a prospective study⁷ revealed that, even among patients with type 2 diabetes mellitus who maintained well-controlled HbA1c levels (HbA1c<7%), the mean amplitude of glycemic excursions (MAGE) continued to be an independent risk factor for DPN. However, the existing literature predominantly focuses on the association between continuous dynamic glucose monitoring systems and diabetic peripheral neuropathy, with limited investigations into long-term HbA1c variability.

The latest epidemiological survey¹ reveals that the prevalence of diabetes mellitus among the elderly population in China is 30%, significantly exceeding the average incidence rate observed in the general population, thereby establishing this demographic as a prominent group affected by diabetes mellitus. Middle-aged and elderly patients with type 2 diabetes mellitus (T2DM) typically exhibit characteristics such as a prolonged disease duration, progressive renal function decline, impaired islet function, etc., which collectively contribute to the challenge of achieving optimal blood glucose control and significantly elevate the risk of complications compared to younger individuals. This retrospective study aims to investigate the association between HbA1c variability and vibration perception threshold values in middle-aged and elderly patients with type 2 diabetes mellitus, providing crucial clinical evidence for the prevention and treatment of diabetic peripheral neuropathy.

Participants and Methods

Participants

This retrospective study enrolled 367 middle-aged and elderly patients with type 2 diabetes mellitus (T2DM) who received regular visits at the Second Affiliated Hospital of Anhui Medical University from January 2021 to August 2023. Inclusion criteria: (1) Type 2 diabetes mellitus diagnosed based on the statement of the American Diabetes Association (ADA) in 2011;⁸ (2) Age exceeding 45 years; (3) A minimum of three HbA1c tests were conducted prior to the investigation. Exclusion criteria: (1) Peripheral neuropathy caused by factors other than T2DM (eg trauma, medication-induced, vitamin B deficiency); (2) Acute and chronic severe complications of diabetes mellitus (eg DKA, HHS, diabetic foot); (3) Concurrent presence of significant systemic diseases (eg, heart failure, liver dysfunction, renal insufficiency).

This study was approved by the Ethics Committee of the Second Affiliated Hospital of Anhui Medical University (Ethics number: YX2022-043). Human research was registered at the Chinese Clinical Trial Registry (registered 23 December 2022), chictr.org.cn (ChiCTR2200067020). All participants voluntarily consented to take part in the study and provided written informed consent. This study adhered to the standards of the International Committee on Harmonization of Good Clinical Practice and the revised version of the Declaration of Helsinki.

Data Collection

Demographic variables such as age, gender, exercise habits, smoking and alcohol history were collected. The height, weight, systolic blood pressure (SBP), and diastolic blood pressure (DBP) of the study participants were measured, and their body mass index (BMI) was calculated. Blood tests were carried out after an overnight fasting for glucose, glycosylated hemoglobin A1c (HbA1c), alanine aminotransferase (ALT), aspartate aminotransferase (AST), total cholesterol (TC), triglycerides (TG), creatinine, 24-hour urinary protein excretion, urine albumin-to-creatinine ratio (UACR), triiodothyronine (T3), thyroxine (T4), thyroid-stimulating hormone (TSH) and other relevant parameters. All data were collected during the last HbA1c test.

HbA1c Variability

In this study, the HbA1c-CV and HbA1c-Mean are considered as indexes of HbA1c variability. HbA_1c values were also obtained from the hospital system during previous follow-up visits. The coefficient of variation of glycated hemoglobin A1c (HbA1c-CV) and the mean glycated hemoglobin A1c (HbA1c-Mean) were computed based on the recorded HbA1c values provided by the patients, where the coefficient of variation was determined as the ratio between the mean and standard deviation.

Vibration Perception Threshold (VPT) Assessment

During the last HbA1c test, the VPT using a vibratory sensory analysis device (Sensiometer A200, Beijing Blue news times technology Co. LTD) was measured: the first toe and dorsal of both left and right feet were detected respectively. The vibration intensity and amplitude were gradually increased until they reached a level that could be perceived by the individuals, and the corresponding vibration values were recorded (ranging from 0 to 50 V with an accuracy of 0.1). Each position was measured three times consecutively, and the average of these measurements was considered as the final result.

The findings of previous studies⁹ have demonstrated a significant correlation between VPT levels exceeding 25V and an increased cumulative risk of neuropathic foot ulcers. Therefore, for the purpose of this study, diabetes mellitus combined with vibration sensation deficiency was defined as the presence of VPT levels higher than 25V on at least one side of the foot.

Other Clinical Indicators Assessment

The Homeostasis Model Assessment of Beta-cell function index (HOMA-β) and the Homeostasis Model Assessment of Insulin Resistance (HOMA-IR) were evaluated using the homeostasis model assessment, which is based on paired measurements of fasting plasma glucose and fasting C-peptide (HOMA2 v2.2.3; Homacalculator/https://www.dtu.ox.ac.uk/). Calculation of body surface area (BSA) in patients was performed using the DuBois formula:¹⁰ 0.007184×weight (kg)^0.425×height(cm)^0.725. The estimation of creatinine clearance rate (Ccr) was determined using the Cockcroft-Gauh (CG) formula:¹¹ eCcr for ; eCcr for . According to gender differences, glomerular filtration rate (GFR) calculated by the Cockcroft-Gauh formula was normalized by BSA: . It is important to note that blood creatinine values in this equation are expressed in mg/dl, and 1μmol/l=0.0113mg/dl.

Statistical Analysis

All statistical analyzes using SPSS Ver. 26.0 software. Comparisons between groups were made using t-test for normally distributed variables, the Mann–Whitney U-test for asymmetrically distributed variables, and Chi-Square test for categorical variables. Variables with statistical differences were entered into a multivariate logistic regression analysis model to determine their net effects on vibration sensation deficiency. The generalized linear regression model was employed to further ascertain the independent association of HbA1c-CV with vibration value. The subjects’ working characteristic curves (ROCs) were mapped using R software (V4.3.1) and the area under the ROC curve was calculated (AUC) using MedCalc Software version 20.100 (https://www.medcalc.org/). A two-sided P-value < 0.05 was considered to indicate statistical significance.

Results

Clinical Characteristics of the Study Population

The participants’ clinical characteristics are summarized in Table 1. A total of 367 middle-aged and elderly patients diagnosed with type 2 diabetes mellitus were included in this study, and they were categorized based on the presence or absence of vibration sensation deficiency. The control group consisted of 149 male and 84 female participants, while the vibration sensation deficiency group (VSD) comprised 90 males and 44 females. The median follow-up duration for both groups was 15 months. Compared to the control group, the VSD group exhibited significant increases in age, duration of diabetes, SBP, HbA1c-Mean, HbA1c-CV, 24h urinary protein levels, and VPT values (P<0.05). Moreover, there were significant decreases in HOMA-β, ALT, eGFR and T3 levels (P<0.05).

Table 1 Characteristics of the Study Cohort

Multivariate Logistic Regression Analysis of T2DM Combined with VSD

The multivariate logistic regression analysis revealed that age, SBP, and HbA1c-CV were identified as risk factors for vibration sensation deficiency in middle-aged and elderly patients with type 2 diabetes mellitus. Conversely, eGFR, ALT, and T3 were found to be protective factors (Table 2).

Table 2 Multivariate Logistic Regression Analysis of T2DM Patients with VSD-Related Influencing Factors

Correlation Between HbA1c-CV and VPT Values

In order to further elucidate the relationship between HbA1c-CV and VPT values on both sides, the generalized linear regression analysis was conducted with VPT values as dependent variables. In the unadjusted analyses, a significant association was observed between HbA1c-CV and both left and right VPT values. After adjusting for age, duration of diabetes, SBP, HOMA-β, ALT, eGFR, T3, 24-hour urinary protein excretion levels, and HbA1c-Mean, HbA1c-CV remained significantly correlated with VPT values on both the left and right sides. (left side, B=2.560, 95% CI 1.298~3.823; P<0.001; right side, B=2.608, 95% CI 1.498~3.718, P<0.001)(Table 3).

Table 3 Generalized Linear Regression Analysis of HbA1c-CV and the Left and Right VPT Values in Middle-Aged and Elderly Patients with Type 2 Diabetes Mellitus

ROC Curve Analysis for HbA1c-CV and VSD Prevalence

The area under the curve (AUC) for HbA1c-CV and VSD prevalence was 0.723(95% CI [0.674~0.768], P<0.001), the sensitivity was 79.85%, the specificity was 56.22%, and the best diagnostic value was 0.08 (Figure 1).

Figure 1 ROC curve analysis for HbA1c-CV and VSD prevalence. Abbreviation: VSD, vibration sensation deficiency.

Discussion

This study aimed to investigate the association between HbA1c variability and other influencing factors with vibration sensation deficiency in middle-aged and elderly patients diagnosed with type 2 diabetes mellitus. The assessment of HbA1c variability was conducted using HbA1c-CV and HbA1c-Mean. The findings revealed a positive correlation between increased HbA1c-CV variability and elevated VPT levels, thereby indicating a significantly heightened risk of peripheral neuropathy. However, no statistically significant correlation was observed between HbA1c-Mean and peripheral neuropathy. Furthermore, age and SBP are recognized as risk factors for vibration sensation deficiency in middle-aged and elderly individuals with type 2 diabetes mellitus, while ALT, eGFR, and T3 may exert a protective effect.

Diabetic peripheral neuropathy (DPN), being the most prevalent complication of diabetes mellitus, stands as the primary cause of neuropathy on a global scale.¹² It can be divided into large fiber neuropathy, small fiber neuropathy or mixed neuropathy. The predominant manifestation of diabetic peripheral neuropathy is characterized by progressive paresthesia, which gradually extends from the peripheral extremities to the central regions, exhibiting a gloves or socks distribution pattern. In severe cases, patients may experience intense pain, muscle weakness and atrophy, along with diminished tendon reflexes. The vibration sensory threshold detection (VPT), as a standardized and quantified method for assessing nerve function, reflecting the functionality of large fiber. Dominguez-Munoz et al¹³ conducted repeated assessments of the vibration sensory threshold in individuals with type 2 diabetes mellitus, considering variations in gender and BMI. Their findings demonstrate robust reliability and stability in the VPT. The studies^14–16 have shown that using VPT to diagnose DPN has high sensitivity and specificity. McIllhatton et al¹⁷ demonstrated that VPT can be recommended as a valuable assessment tool for diabetic peripheral neuropathy.

Healthy individuals exhibit normal islet secretion function and maintain a stable blood glucose fluctuation curve throughout the day. Conversely, in patients with type 2 diabetes mellitus resulting from either islet failure or insulin resistance, alterations in the internal glycemic environment not only lead to a significant increase in average blood glucose levels but also amplify the amplitude of blood glucose fluctuations. Notably, even among diabetics with similar HbA1c levels, there can be substantial variations in the magnitude of blood sugar fluctuations. Continuous glucose monitoring (CGM), as an emerging method for blood glucose monitoring, offers uninterrupted, comprehensive, and reliable information on daily glycemic levels. Consequently, it has gained widespread adoption in the clinical diagnosis and treatment of individuals with diabetes mellitus. Hu et al¹⁸ discovered that glycaemic variability assessed by MAGE is a independent risk factor for diabetic peripheral neuropathy, with an optimal cut-off value of 4.60 mmol/l, exhibiting a sensitivity of 64.47% and specificity of 75.54%. The study¹⁹ showed that Time in range (TIR) was significantly correlated with the prevalence of painful diabetic neuropathy. However, continuous glucose monitoring systems solely capture hyperglycemia or hypoglycemia episodes over a span of several days, whereas HbA1c variability serves as an indicator of glycemic trends and may offer a more comprehensive assessment of diabetes-related risk. Therefore, in the process of diagnosis and treatment, it is imperative to consider both short-term and glycemic fluctuations management. In this study, HbA1c-CV and HbA1c-Mean were employed as measures to assess the HbA1c variability in middle-aged and elderly patients with type 2 diabetes mellitus, thereby providing a more accurate reflection of blood glucose fluctuations. It was found that the HbA1c-CV and HbA1c-Mean of patients in VSD group were significantly higher than those in the control group, and multivariate logistic regression analysis showed that HbA1c-CV was a risk factor for the development of vibration sensation deficiency. The results of further generalized linear regression analysis revealed a significant association between HbA1c-CV and both left and right VPT values, even without adjusting for confounding factors. Moreover, this correlation remained statistically significant after controlling for relevant risk factors. (left side, B=2.560, 95% CI 1.298~3.823; P<0.001; right side, B=2.608, 95% CI 1.498~3.718, P<0.001). There was no statistically significant difference observed in the HbA1c-Mean values as determined by multivariate logistic regression analysis. Therefore, HbA1c-CV may serve as a significant risk factor for diabetes mellitus in conjunction with vibration sensation deficiency, whereas HbA1c-Mean does not exhibit the same association. The underlying mechanism through which blood glucose fluctuations contribute to the development of peripheral neuropathy in diabetes patients remains elusive. Animal studies²⁰ have shown that blood glucose fluctuation can exacerbate oxidative stress and inflammatory response by up-regulating the expression of inflammatory regulatory factors and intercellular adhesion molecule-1 gene, which are considered to be important pathways of neurological dysfunction and apoptosis.²¹ Quagliaro et al²² found that exposure of endothelial cells to fluctuating hyperglycemia results in an upregulation of apoptosis through the activation of PKC-dependent NAD(P)H oxidase.

Through multivariate regression analysis, we identified age as a potential risk factor for diabetes individuals with vibration sensation deficiency. (OR=1.066, 95% CI 1.024~1.108, P=0.002). The prevalence of peripheral neuropathy in patients with type 1 diabetes mellitus has been observed to increase with age according to epidemiological studies,²³ ranging from 18% among individuals aged 18–29 years to 58% among those aged 30 and above. In the ACCORD trial,²⁴ the incidence of peripheral neuropathy in adults with type 2 diabetes mellitus was reported to be 42%. The incidence of peripheral neuropathy may gradually increase as diabetes individuals age. Moreover, the prevalence of diabetic peripheral neuropathy varies across different countries and regions. Pop-Busui et al² reported a 28% incidence of diabetic peripheral neuropathy in the United States, while the corresponding San Luis Valley cohort study²⁵ documented a slightly lower incidence rate of 25.8% for peripheral neuropathy. A cross-sectional study conducted in Beijing, China²⁶ revealed that peripheral neuropathy was present in 23.5% of individuals diagnosed with type 2 diabetes mellitus. The prevalence of type 2 diabetic peripheral neuropathy in the middle-aged and elderly population was found to be 35.6% in this study, which is significantly higher compared to previous research findings. This discrepancy may be attributed to variations in the age range or assessment methodologies employed within the studied population.

The patients enrolled in this study exhibited glomerular filtration rates at stage 3 or higher (eGFR>30 mL/min/1.73m²). The multivariate logistic regression analysis revealed that maintaining a high estimated glomerular filtration rate (eGFR) could potentially serve as a protective factor for vibration sensation in middle-aged and elderly patients with type 2 diabetes mellitus. (OR=0.983, 95% CI 0.969~0.998, P=0.022). Similar to previous findings,²⁷ it was also shown that the prevalence and severity of peripheral neuropathy would increase with the deterioration of renal function. Peripheral neuropathy is more likely to occur in patients with severe renal insufficiency or uremia due to metabolic disorders, acid-base imbalance, toxin accumulation, and other factors.²⁸ The prospective studies²⁹ demonstrated that elderly patients with chronic kidney disease had a 2.3-fold higher risk of motor nerve conduction velocity impairment compared to non-CKD subjects. In recent years, Moorthi et al³⁰ have discovered a significant association between renal function and peripheral neuropathy in middle-aged and elderly diabetic patients, even during the early stages of renal decline. Hence, it is suggested that there exists a substantial correlation between renal function status and diabetic peripheral neuropathy.

In middle-aged and elderly patients with type 2 diabetes mellitus, the majority exhibit concomitant blood lipid deposition and arterial wall sclerosis, leading to an elevation in systolic blood pressure and pulse pressure difference. The results of animal experiments³¹ have demonstrated that hypertension can expedite the pathogenesis of peripheral neuropathy in ZDF rats, primarily characterized by a reduction in nerve myelin thickness. In order to ascertain the specific blood pressure index that contributes to neuropathy, Jarmuzewska et al³² conducted a cross-sectional survey on a limited sample size, revealing significant correlations between systolic blood pressure (SBP), pulse pressure difference (PP), and nerve function. Similar to the above results, after adjusting for potential confounders such as age and medical history, a positive correlation was observed between systolic blood pressure (SBP) and the occurrence of VSD (OR=1.018, 95% CI 1.000–1.037, P=0.048). These results suggest that SBP may serve as one of the risk factors for peripheral neuropathy in middle-aged and elderly patients with type 2 diabetes mellitus.

The thyroid hormone, synthesized and secreted by the thyroid follicular epithelial cells, exerts regulatory effects on human body metabolism. Several studies^33,34 have demonstrated that thyroid hormone secretion disorder is associated with an increased likelihood of developing metabolic syndrome and significantly elevates the risk of complications related to diabetes mellitus. Lin et al³⁵ demonstrated a negative correlation between FT3 levels and the prevalence of diabetic peripheral neuropathy, as well as an increased risk of DPN in patients with elevated FT4 levels. He et al³⁶ found that the presence of low T3 syndrome in diabetic patients can lead to severe glucose toxicity and nerve fiber damage, thereby significantly elevating the risk of diabetic peripheral neuropathy (DPN). The findings of this study suggest that T3 may serve as a protective factor for vibration perception in middle-aged and elderly patients with T2DM (OR=0.285, 95% CI 0.111~0.738, P=0.010). Consequently, there exists an association between T3 levels and diabetic neuropathy.

This study suggests that the presence of alanine aminotransferase (ALT) may confer a protective effect on vibration perception (OR=0.968, 95% CI 0.943~0.995, P=0.020). Reiterating the findings of Mohsen et al,³⁷ their study indicated a negative correlation between markers of liver injury (such as elevated ALT, fatty liver or fatty liver combined with elevated ALT) and diabetic peripheral neuropathy. However, inconsistent research findings have also been reported. For instance, Huang et al³⁸ observed a significant increase in the prevalence of diabetic peripheral neuropathy among individuals with type 2 diabetes mellitus (T2DM) and non-alcoholic liver disease (NAFLD). Mantovani et al³⁹ also reported a significantly higher incidence of distal symmetric polyneuropathy in patients with non-alcoholic liver disease (NAFLD) compared to those without NAFLD. The disparities observed in the aforementioned study findings may be attributed to variations in the racial composition, age distribution, and duration of diabetes among the enrolled patients.

Limitations

However, it is important to acknowledge the limitations of this study. Firstly, it employed a retrospective design with a small sample size at a single center, thereby restricting the statistical power of the analysis. Moreover, the age range of participants was limited to individuals aged over 45 years, leaving the relationship between glycemic variability and vibration sensory threshold (VPT) in other age groups uncertain. We intend to gradually address these inquiries in future investigations.

Conclusions

This study shows an association between HbA1c variability and vibration sensation in middle-aged and elderly patients with type 2 diabetes mellitus. Specifically, a higher risk of vibration sensation deficiency is observed in individuals with greater HbA1c-CV values within this patient population. Furthermore, age and SBP are identified as risk factors for vibration sensation deficiency in T2DM patients, while ALT, eGFR, and T3 serve as protective factors. Therefore, effective management of blood glucose levels in middle-aged and elderly patients with T2DM should aim to minimize fluctuations, which can contribute to the mitigation of diabetes peripheral neuropathy occurrence.