A cross-sectional study to determinate the relationship between body composition & neuropathy in patients with type 2 diabetes mellitus

Abstract

Background

More information is required where the relation of diabetic neuropathy and body composition through bioimpedance is addressed. Therefore the aim of this study was to investigate the association between peripheral (PN) and autonomic neuropathy (AN) with bioimpedance in patients with type 2 diabetes mellitus (T2DM) aged 18 to 64 years.

Methods

A cross-sectional study was conducted at a private hospital in Northeastern Mexico between November 2023 and March 2024. Patients with T2DM were evaluated using the Michigan Neuropathy Screening Instrument (MNSI) for PN, Sudoscan for AN, and bioimpedance analysis (InBody) for body composition. Other variables considered included diabetes duration, hemoglobin A1C (HbA1C), and lipid profile. Chi-square, T-test, and Wilcoxon tests were used for independent variables, Cohen’s kappa for concordance, and logistic regression models for association analyses.

Results

A total of 160 patients were analyzed, with a median (IQR) age of 51 (12) years, and a majority being male (n = 81, 51%). The prevalence of PN was 30% (n = 48) and AN was 35.6% (n = 57), with a Cohen’s kappa concordance of 0.282. Patients with PN had higher median (IQR) visceral fat mass [20 (8.3) vs. 17 (10.3) kg, p = 0.010], arm fat mass [3.65 (3.35) vs. 2.70 (2.30) kg, p = 0.004], torso fat percentage [485 (166) vs. 369 (173), p < 0.001], arm circumference [40.1 (7.4) vs. 36.4 (5.7) cm, p < 0.001], arm muscle circumference [32.1 (3.8) vs. 30.7 (4.7) cm, p = 0.008], and skeletal muscle index (SMI) [8.35 (1.25) vs. 7.95 (1.55), p = 0.009]. For AN, a higher torso muscle percentage was observed [105.2 (6.2) vs. 103.3 (7), p = 0.02]. No significant differences were found in HbA1C or lipid profile. Logistic regression for PN showed associations with T2DM duration (OR = 1.111, p = 0.006), age (OR = 1.080, p = 0.004), leg fat mass (OR = 187.197, p = 0.003), and SMI (OR = 0.612, p = 0.021), with a Nagelkerke R² of 0.328. No significant predictors were found for AN.

Conclusion

A high prevalence of neuropathy was observed, with a significant association between body composition and PN, highlighting greater adiposity in limbs, torso, and visceral fat, as well as age, T2DM duration, and SMI.

Background

Diabetes mellitus (DM) is a chronic disease that represents one of the greatest public health concerns worldwide. The global prevalence of diabetes has increased alarmingly, with an estimated 463 million adults affected in 2019, a figure expected to reach 700 million by 2045 [1]. In Mexico, the prevalence of diabetes among adults is equally concerning, with approximately 12.3% of the population diagnosed with this disease [2]. Type 2 diabetes (T2DM) is the most common form of diabetes, predominantly affecting the adult population, although its prevalence is also increasing among younger individuals. In the non-geriatric population, the prevalence of T2DM is significant and follows a growing trend due to factors such as sedentary lifestyle, obesity, and unhealthy diet [3]. This trend underscores the importance of adequately understanding and managing the complications associated with T2DM to improve the quality of life of these patients.

Diabetes complications, both microvascular and macrovascular, significantly contribute to the morbidity and mortality of patients. Among the most common microvascular complications is diabetic neuropathy, which affects up to 50% of diabetic patients at some point in their lives [4]. Diabetic neuropathy can be classified into autonomic neuropathy and peripheral neuropathy, each with different clinical implications and diagnostic methods. Autonomic neuropathy, which includes sudomotor dysfunction, can be evaluated using the Sudoscan^®, a device that measures electrochemical skin conductance and serves as a marker of sweat gland function [5]. Peripheral neuropathy, on the other hand, is often diagnosed using the Michigan Neuropathy Screening Instrument (MNSI), which evaluates tactile and vibratory sensitivity as well as deep tendon reflexes in the lower extremities [6].

Despite advances in the diagnosis and management of diabetic neuropathy, gaps remain in the understanding of its relationship with body composition. Bioelectrical impedance analysis (BIA) is a non-invasive tool that allows the evaluation of body composition, including fat and lean mass [7]. However, the association between body composition and the severity of diabetic neuropathy has not been fully elucidated [8, 9]. Recent studies have highlighted the importance of early assessment and intervention strategies for diabetic neuropathy. Strict glucose control and lifestyle interventions have been shown to positively impact the reduction of neuropathic complications [10]. Additionally, current research has explored the relationships between body composition, particularly visceral fat, and neuropathy risk, indicating that a higher amount of visceral fat may be associated with a greater risk of developing peripheral neuropathy [11, 12].

This study investigated the relationship between body composition, measured by bioimpedance, and diabetic neuropathy, both autonomic and peripheral, in non-geriatric patients with T2DM. The results of this study may provide new insights into the underlying pathogenic mechanisms and have important clinical implications for the personalized management of diabetic patients.

Methods

This observational cross-sectional study aimed to determine the association between body composition, as determined by bioimpedance and neuropathy in patients with type 2 diabetes mellitus (T2DM). The study adhered to the STROBE reporting guidelines. Ethical approval was obtained from the local institutional review board (IRB number: 25112022-CN-MI-CI). The study was conducted in compliance with the World Medical Association’s Declaration of Helsinki. Due to the nature of this research, written informed consent to participate was received from each patient. The study was conducted in Hospital Clinica Nova, a private hospital located in Northern Mexico, with recruitment occurring between November 2023 and March 2024.

Selection criteria

Patients of both sexes aged 18 to 64 years, diagnosed with T2DM with or without microvascular complications, were included. Participants be able to use the sudoscan and inbody equipment without assistance, without difficulty standing and following simple orders, have access to medical care at Hospital Clinica Nova and sign an informed consent. Exclusion criteria comprised patients with type 1 diabetes mellitus, pregnancy, hospitalization during the study, neoplastic diseases, active infections, prior motor or neurological disabilities, and current treatment with antibiotics or corticosteroids due to potential impact on lean mass. Additionally, subjects without current laboratory test results or who did not attend the testing sessions were excluded. A non-probabilistic convenience sampling method was used for selecting the patients.

Data collection

Medical data, including duration of diabetes (years), history of hypertension, dyslipidemia (either by history or by the presence of total cholesterol > 200 mg/dL, HDL < 40 mg/dL, LDL > 130 mg/dL, or triglycerides > 150 mg/dL), hypothyroidism, and chronic complications of diabetes including diabetic retinopathy and diabetic nephropathy were retrieved from the medical records of the patients. Vital signs, including heart rate (bpm), respiratory rate (rpm), blood pressure (mmHg) measured with a sphygmomanometer, temperature (°C) obtained with a digital thermometer, and partial oxygen saturation measured with pulse oximeter (Mindray^® uMEC 12 China), were recorded. Anthropometric measurements such as weight (kg), height (m), and waist circumference (cm) were obtained using a tape measure. Body composition was analyzed using bioelectrical impedance with an InBody^® 570 device (produced in Mexico). Peripheral neuropathy was assessed using the Michigan Neuropathy Screening Instrument (MNSI). In the questionnaire a value of ≥ 7 is considered indicative of neuropathy and in the physical examination, a score of ≥ 2.5 is considered abnormal. Autonomic neuropathy was measured using the Sudoscan^® electrochemical conductance test (Impeto Medical, EZS 01750010193, Paris, France), with results reported in microSiemens (µS) a score of < 60 µS in hands and < 70 µS is considered indicative of autonomic neuropathy. Laboratory results for glycated hemoglobin HbA1c (%) and lipid profile (including total cholesterol [mg/dL] HDL [mg/dL] LDL [mg/dL] and triglycerides [mg/dL]) were retrieved from the medical records of at most three months before the taking of the tests. For patients without recent laboratory tests, venipuncture was performed, and samples were processed and analyzed using the Cobas 6000 analyzer (Japan). Strength (pound) per dynamometer CAMRY (SCACAM-EH10117, China).

Statistical analysis

Sample size was calculated using a formula for studies that require a logistic regression model [13], with an effect size of 5%, an alpha of 5%, and a power of 80%, and 16 predictors, resulting in a total sample of 160 patients. An exploration of the variables was conducted to determine their respective normality, using histograms and the Shapiro-Wilk test. Subsequently, logarithmic transformations were performed in order to normalize non-normal variables. Descriptive statistics were then carried out through frequencies, percentages, mean, median, standard deviation, and interquartile ranges. The Chi-square test was used for the comparison of categorical variables and the groups of peripheral neuropathy and autonomic neuropathy; while the independent t-test and Wilcoxon independent tests were used for comparison of quantitative variables and the groups of peripheral neuropathy and autonomic neuropathy. Two binary logistic regression models were utilized to examine the relationships between the dependent variables (peripheral and autonomic neuropathy) and the independent variables, which included data from the patient’s medical history, InBody results, and laboratory results. Independent variables were selected based on the results of the univariate analysis. Variables were simultaneously entered into the model, and those statistically insignificant or with minimal impact on the R² value were subsequently removed. The use of multiple imputations for missing values, which were less than 20%, was considered. A p-value of less than 0.05 was considered significant. The analysis was performed using IBM SPSS Statistics V.26.

Results

Study population

A total of 162 patients were recruited for the study. Two subjects were excluded, one did not show up for laboratory sampling, and the other did not attend the Sudoscan test. Consequently, 160 subjects were analyzed. A total of 81 (51%) patients were male and the overall median (IQR) age was of 51 [12].

The median (IQR) diabetes evolution was 6 (8.96) years. Hypertension was present in 64 (40%) patients, obesity in 104 (65%) patients, and dyslipidemia in 106 (66%) patients. The median (IQR) waist circumference was 100 [14] cm, and the mean (SD) BMI was 33.1 (1.2).

In the studied population, neuropathy of any kind (autonomic and peripheral) was found in 105 (65.6%) patients. Peripheral neuropathy was observed in 48 (30%) patients, and autonomic neuropathy in 57 (35.6%). The concordance between cases of peripheral and autonomic neuropathy was assessed by a Kappa coefficient of 0.280, p = < 0.001. For the peripheral neuropathy there was a statistically significant difference regarding the age of patients with patients affected by the disease having an older age (50 [12] vs. 55 [8]) p = 0.004, longer duration of diabetes with the presence of disease (6 [8.06] vs. 9.5 [9.56]) p = 0.004, BMI higher in the presence of disease (34.6 [1.17] vs. 31.6 [1.58]) p < 0.001, higher waist circumference (104 [18] vs. 98 [23]) p = 0.025, a greater prevalence of hypertension (27 [56%] vs. 37 [33%]) p = 0.006, obesity (38 [79%] vs. 66 [59%]) p = 0.014 and diabetic retinopathy (6 [13%] vs. 2 [1.8%]) p = 0.010. Medical history details divided by type of neuropathy are provided in Table 1.

Table 1 Medical history in subjects divided by type o neuropathy

Neuropathy assessment and body composition analysis

Among individuals screening positive for peripheral neuropathy, lower sudoscan values were observed in the feet, with a median (IQR) of 74 [15] compared to 78 [9], p = 0.001. There were no statistical difference between subjects with neuropathy and those that tested normally in lipid profile, glycosylated hemoglobin, and dynamometer strength. Clinical and laboratory findings stratified by neuropathy type are presented in Table 2.

Table 2 Clinical and laboratory tests

According to body composition significant differences were observed in patients with peripheral neuropathy with higher levels of visceral fat, with a median (IQR) of 20 (8.3) vs. 17 (10.3), p = 0.010, fat per segment in arms 3. 65 (3.35) vs. 2.70 (2.30), p = 0.004, fat percentage in torso 485 (166) vs. 369 (173) p < 0.001, total arm circumference 40.1 (7.4) vs. 36.4 (5.7), p < 0.001, arm muscle circumference 32.1 (3.8) vs. 30.7 (4.7), p = 0.008 and skeletal muscle mass index 8.35 (1.25) vs. 7.95 (1.55), p = 0.009. As for the patients with autonomic neuropathy, it was observed that they presented a higher percentage torso muscle with a median (IQR) of 105.2 (6.2) vs. 103.3 [7], p = 0.02. Detailed results of body composition variables divided by type of neuropathy are provided in Table 3.

Table 3 Body composition divided by type of neuropathy

A binary logistic regression was performed to predict autonomic neuropathy as the dependent variable; however no significant independent variables fitted the model. We computed another separate binary logistic regression, where peripheral neuropathy was the dependent variable, and the independent variables were years of T2DM evolution, age, leg fat mass, and musculoskeletal mass index (MSMI). The results showed a positive statistically significant relationship with T2DM time of evolution (OR = 1.111, p = 0.006), age (OR = 1.080, p = 0.004), leg fat mass (OR = 187.197, p = 0.003), and skeletal muscle mass index (OR = 0.612, p = 0.021), Nagelkerke’s R² = 0.328. Binary logistic regression for prediction of peripheral neuropathy are show in Table 4. The model can be represented in the following equation:

$$\:\text{log}\left(p\right)=\text{log}\left(\frac{p}{1-p}\right)={\:\beta\:}_{0}+\:{\beta\:}_{1}{X}_{1}+\:{\beta\:}_{2}{X}_{2}+{\beta\:}_{3}{X}_{3}+{\beta\:}_{4}{X}_{4}$$

where Y is the binary outcome variable (presence of peripheral neuropathy); the probability of the presence of disease is denoted as P(Y = 1) = p. The logit function is the natural logarithm of the odds $\:\frac{p}{1-p}.$ Where β₀ (intercept)= -23.076, β₁ = 0.105, X₁ is Time of DMT2 evolution (years), β₂ = 5.232, X₂ is leg fat mass, β₃ = 0.77, X₃ is age, β₄ = 0.612, X₄ is MSMI.

Table 4 Binary logistic regression for prediction of peripheral neuropathy

Discussion

In this cross-sectional observational study, our aim was to examine the association between body composition and neuropathy in patients with type 2 diabetes mellitus (T2DM). Our study found that neuropathy, whether peripheral or autonomic, is highly prevalent among patients with T2DM, peripheral neuropathy was observed in 30% and autonomic neuropathy in 35.6% of patients which aligns to results from previous studies [4, 15,16,17]. The concordance test, indicates a mild to moderate agreement between cases of peripheral and autonomic neuropathy. This could be explained in a clinical context where not all patients with peripheral neuropathy have autonomic neuropathy and vice versa. It was observed that body composition plays an important role in this population in relation to peripheral neuropathy, specifically with adiposity.

Patients with peripheral neuropathy had significantly higher percentages of visceral fat, arm fat, and torso fat. Additionally, these patients showed greater arm and muscle circumference, a higher skeletal muscle mass index (SMMI) and leg fat mass. These findings suggest that adiposity, especially central and segmental fat, may play a role in the development or exacerbation of neuropathy in patients with T2DM. Recent studies have also explored these associations. A study by Kim et al. (2021) found that higher visceral fat and lower muscle mass were associated with a higher prevalence of peripheral neuropathy in patients with T2DM [18], which aligns to our findings. Similarly, a study by García de la Torre et al. (2023) reported that reducing visceral fat through dietary intervention and exercise improved neuropathic symptoms in patients with T2DM [19]. Previous studies have shown that visceral fat is metabolically active and contributes to insulin resistance and systemic inflammation, which are key factors in the pathogenesis of diabetic neuropathy [20, 21], but the pathophysiological mechanisms underlying the association between body composition and diabetic neuropathy are complex and multifactorial. Chronic hyperglycemia in patients with T2DM leads to the formation of advanced glycation end products (AGEs), which accumulate in nerve tissues and contribute to neuronal damage [22]. Furthermore, insulin resistance, common in patients with obesity and T2DM, is associated with a chronic inflammatory state that can exacerbate neuronal damage. Adipose cells, especially in visceral fat, secrete pro-inflammatory cytokines such as tumor necrosis factor-alpha (TNF-α) and interleukin-6 (IL-6), which can damage peripheral nerves [18].

Interestingly, autonomic neuropathy was associated with a significantly greater difference in the percentage of torso muscle, but not with other body composition variables. Previous studies have not mentioned this association with muscle percentage, but with other parameters such as weight and fat distribution by segments [16]. Autonomic neuropathy may be influenced by autonomic nervous system dysfunction due to hyperglycemia and AGEs accumulation in autonomic nerves [12].

Obesity induces a chronic inflammatory state, characterized by elevated levels of proinflammatory cytokines such as TNF-α and IL-6 that can induce oxidative stress, endothelial dysfunction, and nerve ischemia, exacerbating neuronal damage [23, 14].

The underlying pathophysiological mechanisms in diabetic neuropathy are complex and multifactorial, involving both metabolic alterations and autonomic nervous system dysfunction. Sympathetic nervous system (SNS) hyperactivity plays a key role in this interaction, contributing to neuronal damage through catecholamine release and oxidative stress induction [24].

The involvement in peripheral neuropathy is mainly in the sensory and motor nerves, due to demyelination, axonal damage, inflammation and metabolic stress. In autonomic neuropathy the main damage is in type C and B fibers which are small unmyelinated fibers, there is involvement and dysfunction of the autonomic ganglia due to the inflammatory process leading to chronic ischemia [25,26,27,28].

The presence of peripheral neuropathy was significantly associated with various clinical and demographic variables, such as older age, longer duration of diabetes, higher BMI, larger waist circumference, and the presence of hypertension, obesity, and diabetic retinopathy, although the logistic regression showed just as significant factors age and time of evolution of T2DM. These results are consistent with existing literature, which has systematically identified similar risk factors for diabetic neuropathy. The United Kingdom Prospective Diabetes Study (UKPDS) highlighted the role of prolonged hyperglycemia and hypertension in the development of microvascular complications, including neuropathy [17, 29]. In our study, patients with neuropathy had a mean age of 62 years compared to 56 years (p < 0.05) and a mean duration of diabetes of 15 years compared to 10 years in patients without neuropathy.

We found no significant differences in lipid profile or glycated hemoglobin levels between patients with and without neuropathy. This contrasts with some studies that have reported associations between poor glycemic control and lipid abnormalities with the presence of neuropathy [29]. However, the lack of significant findings in our study could be due to relatively well-controlled glycemic and lipid parameters in our study population, as indicated by average HbA1c levels around 6.6–6.7%, and because of the cross-sectional approach to our study.

This research emphasizes that fat distribution significantly influences not only the development of diabetes as previous studies have mentioned [30,31,32], but also the exacerbation of microvascular complications including neuropathy, with segmental fat distribution, particularly in the extremities, being a notable predictor. The study presented a cohort of young adults with a median age of 51 years with considerable number of these individuals exhibiting microvascular complications despite having a relatively short duration of the disease and maintaining good glycemic control. The detection of autonomic neuropathy in a significant proportion of these patients, even in the absence of peripheral neuropathy, highlights a potentially overlooked aspect of cardiovascular risk in this population. This observation suggests that autonomic neuropathy may often go undiagnosed, presenting a substantial threat to cardiovascular health.

There are some limitations to our study, firstly, the cross-sectional design limits our ability to establish causality between body composition and neuropathy. Secondly, the relatively small sample size may restrict the generalizability of the findings. Additionally, the reliance on clinical tests and questionnaires for neuropathy assessment may introduce measurement biases. Future research should consider employing longitudinal designs and larger sample sizes to confirm and further explore these associations.

Conclusion

The research identified a high prevalence of diabetic neuropathy among the participants. A significant association was observed between body composition and peripheral neuropathy, particularly with increased adiposity in the lower and upper extremities, torso and visceral fat. In addition, factors such as age and time of evolution were also strongly associated with peripheral neuropathy. Skeletal muscle mass index emerged as a possible predictor of peripheral neuropathy, suggesting that muscle function may be causally related to the development of neuropathy. Torso muscle was the solely variable associated with autonomic neuropathy. Although, further research is needed to fully understand the contribution of muscle lean mass and function to the progression of neuropathy.