Morphological and Mechanical Properties of Lower-Limb Muscles in Type 2 Diabetes: New Potential Imaging Indicators for Monitoring the Progress of DPN

In recent years, the incidence of diabetes has increased in China. Currently, the number of patients with diabetes in China is estimated at 129,800,000 (1), which means 9.1% of adults living in China had type 2 diabetes mellitus (T2DM) in comparison with 6.28% of adults in the world (2). T2DM accounts for ∼90–95% of the total population with diabetes (3). Diabetic peripheral neuropathy (DPN) is a common chronic complication of T2DM. Long-term DPN can lead to serious defects in skeletal muscles, including neurogenic muscle atrophy and muscle strength loss (4–6), which leads to motor dysfunction, gait modification, and abnormal foot biomechanics (7).

A synergistic relationship between multiple muscle components of the lower leg controls motor balance. The abductor hallux (AbH) muscle is located subcutaneously in the medial margin of the plantar (which has the largest cross-sectional area [CSA] in the foot muscle) and is associated with diabetic metatarsophalangeal joint malformation (8), hallux flexor strength (9), and flat foot (10). The largest muscle in the anterior tibial muscle group is the tibialis anterior (TA). The TA bends the ankle during a gait cycle and moves the center of gravity forward on the basis of support (11). The peroneal longus (PER) muscle is important for preventing foot varus, raising the heel, and maintaining gait balance (12,13). The AbH, TA, and PER are superficially located, which is convenient for ultrasonic imaging, with superior reliability (14).

Ultrasound shear wave imaging (USWI) is an advancing ultrasound imaging technique for estimation of the mechanical properties of tissues. The purpose of elastography is to assess tissue stiffness by quantitatively calculating the velocity of the shear waves (15). To achieve this, USWI applies a transient mechanical force to the tissue, resulting in a transient displacement, and propagating it in the form of a compressional or shear wave. The shear wave propagates perpendicular to the compressional wave’s direction, and its velocity is directly correlated with the shear modulus of tissues. The relationship between shear wave velocity (SWV) and tissue shear modulus is expressed as follows, μ = ρv², where μ is the shear modulus, ρ is the tissue density, and v is the tissue SWV. From a physical point of view, Young’s modulus (E) is the most relevant measure of stiffness of a given material. Shear modulus (μ) is directly related to Young’s modulus (E) when the material is an isotropic linear elastic body E ≈ 3μ (15). Since skeletal muscles are viscoelastic, nonlinear, and anisotropic, this formula cannot theoretically be applied. However, when the ultrasonic probe is parallel to the muscle fiber, the shear modulus of the muscle has a strong linear relationship with Young’s modulus measured by the traditional test method, and the shear modulus can still accurately characterize the hardness of the muscle (16).

Investigators of previous studies have reported that ultrasonography can quantify foot muscle–related structures and provide an opportunity to judge foot function (17–21). Ultrasound imaging can reliably evaluate the CSA of the muscle. The previous studies showed that the plantar intrinsic muscle sizes were significantly correlated with corresponding toe flexion strength (18) and abnormal plantar pressure distribution (19). Furthermore, quantifying lower-limb muscle stiffness may also contribute to improving the understanding of skeletal muscle rapid force production (20), walking time, grip strength, and isokinetic knee strength (21).

So far, ultrasonography provides additional information including nervous lesion morphology, anatomic location, relationship of lesions to surrounding soft tissues for diagnosis and surgical planning for peripheral nerve lesions (22,23). But with it one still cannot evaluate neuromuscular function. Compared with ultrasonography of peripheral nerves, quantitative evaluation of skeletal muscle morphology and mechanical properties by ultrasonography can be used as a convenient and noninvasive method to evaluate skeletal muscle function (18,20). Moreover, skeletal muscles are easier to identify than nerves. Ultrasound of properties of skeletal muscles allows clinicians to evaluate neuromuscular function with high reproducibility (24) and effectiveness (25).

Skeletal muscle atrophy in patients with DPN usually occurs first in the distal limbs (18,26). However, morphological and mechanical changes in ABH, TA, and PER muscles in patients with T2DM have not been reported. A systematic study on the morphological and mechanical properties of AbH, PER, and TA muscles is helpful for understanding gait balance disorders and foot biomechanical changes caused by DPN. This would be conducive to the development of muscle strength intervention plans, effectively reducing the risk of falls (5) and the incidence of foot ulcers caused by biomechanical abnormalities in this population (7).

In the current study we aimed to estimate the synergistic changes in morphology and mechanics of lower-limb skeletal muscles during the progression of diabetes. Additionally, we sought to explore the correlation of the properties of lower-extremity muscles with clinical indicators associated with DPN progression and find a potential image indicator for monitoring the progress of DPN in patients with T2DM.

A total of 240 patients with T2DM were continuously enrolled in the outpatient and inpatient departments of the Second Xiangya Hospital of Central South University between December 2019 and May 2021 (in line with the 1999 World Health Organization diagnostic and classification criteria) (27). Details of the study design and the inclusion and exclusion criteria can be found in Fig. 1. Diagnosis of clinical DPN was based on the diagnostic standard proposed by the Toronto Diabetic Neuropathy Expert Group: the presence of an abnormality of nerve conduction testing and a symptom or symptoms or a sign or signs of neuropathy confirm DPN (28).

Individuals were excluded from the study if they met any of the following exclusion criteria: current or previous plantar ulcer or foot and ankle bone deformity, recent history of major surgery or any lower-extremity orthopedic surgery, history of lower-extremity vascular disease and vascular reconstruction from other causes, history of cerebral hemorrhage and infarction, toxic peripheral neuritis, infectious polyneuritis, pregnancy, malignant tumor, acute and chronic skeletal muscle system diseases, and neuroendocrine system diseases from other causes.

Among the 240 patients recruited according to the inclusion criteria, 37 patients were excluded according to the exclusion criteria, and the residual 203 patients were divided into groups: T2DM with DPN (102 subjects) and T2DM without DPN (101 subjects). Details of the study design and inclusion and exclusion criteria can be found in Fig. 1.

All procedures were reviewed and approved by the ethics committee of Second Xiangya Hospital (ethics review no. LY2021-019). The participants signed an informed consent form approved by the ethics committee before participating in the study.

Clinical data including age, sex, systolic blood pressure (SBP), diastolic blood pressure, ankle-brachial index (ABI), BMI, and course of diabetes were collected. In the laboratory, venous blood was collected and examined the morning after a 10-h fast to determine the levels of triglyceride (TG), total cholesterol (CHO), LDL cholesterol (LDL-C), HDL cholesterol (HDL-C), and glycosylated hemoglobin (HbA_1c).

Based on the previous study (29), the patients’ Toronto Clinical Scoring System (CSS) scores were obtained with inclusion of neurological symptom scores (foot pain, numbness, tingling, ataxia, and upper-limb symptoms), neuro-reflex scores (knee reflexes and ankle reflexes), and sensory function test scores (pinprick, temperature, light touch and vibration, position). The outcome score is a continuous variable ranging from a minimum of 0 points (no neuropathy) to a maximum of 19 points. Six points are derived from symptoms, 8 from lower-limb reflexes, and 5 from sensory examination distally at the toes.

The motor conduction velocity of the bilateral median nerve (elbow to wrist), bilateral tibial nerve (medial malleolus to popliteal fossa), and bilateral common peroneal nerve (below the fibular head to the anterior ankle) and sensory conduction velocities of the bilateral median nerve and bilateral superficial peroneal nerve (median to lateral malleolus at the posterior margin of the leg) were recorded using Dantec Keypoint electromyography surface electrodes.

The CSA and stiffness of the AbH, TA, and PER muscles were measured with a Canon ultrasonic diagnostic instrument (Aplio 500; Canon Medical Systems, Canon, Tokyo, Japan) coupled with one linear transducer (6–15 MHz; Canon) used in shear wave elastography mode (musculoskeletal preset) on nondominant feet. The patient was asked to remain supine on the examination bed by extending their knees and keeping the ankle in a neutral position, with position maintained by a fixator (Supplementary Fig. 1). The lower leg was 90° perpendicular to the outer edge of the foot. Ankle angle and hallux positioning were kept constant with elastic bands throughout the probe scans. During the scanning, the probe exerted only minimal pressure on the muscle for measurement of the passive stiffness of the muscle.

Spatial variation of SWV in lower-limb muscles has been reported (30,31). Thus, it is important that the measurements are performed in invariable muscle regions such that the results can be compared between participants and studies. In the current study, the CSAs of AbH, TA, and PER were measured according to the method described in a previous study (32). The CSA of AbH muscle was acquired on a scanning line perpendicular to the long axis of the foot at the anterior aspect of the medial malleolus. The CSA of PER muscle was captured with transverse scanning at a distance of 50% between the fibular head and the inferior border of the lateral malleolus. The CSA of TA muscle was measured with longitudinal scanning at 20% of the distance between the fibular head and the inferior edge of the lateral malleolus. In the ultrasonographic images, the measured muscle boundary was based on the inner edge of muscle fascia around the muscle (19). The patients were instructed to do corresponding muscle movements to define muscle boundaries. Representative images of skeletal muscle CSA measurements are shown in Fig. 2.

After the CSA of each muscle was obtained, the probe was then cautiously rotated to the longitudinal section of the muscle fiber and used to measure the SWV of each muscle. The AbH muscle was longitudinally located on a scanning line between the muscle’s origin on the medial calcaneal tuberosity and the navicular tuberosity. The PER muscle was longitudinally identified at the lateral side of the fibula, and the TA muscle was longitudinally obtained laterally to the tibial crest. The muscle-fiber pennation angle (PA) is the angle of insertions of muscle fascicles into the deep aponeurosis. It was measured on the longitudinal view by the probe in parallel to muscle fibers on the position where we measured the SWV of the lower-limb muscle.

During the measurement of SWV, a color-coded SWV map was displayed overlapping with standard B-mode images in real time, while the temporal resolutions of USWI measurements were 0.7 fps (frame per second). Tissue density of 1 g/cm³ was adopted in measuring muscle properties. The image was frozen until a stable map was obtained after ∼3–5 s (33). The color map area was 1.5–2.0 cm² in measuring the SWV for all the muscles, and the depth of the color map area ranged from 0.5 to 4.0 cm. For avoidance of artifacts on the edge of the USWI map, several circular regions of interest (ROIs) with a diameter of 2 mm were placed in the central area of the USWI map to measure the SWV. Notably, the thicker aponeurosis in muscle should be excluded from the ROIs to avoid the effects of muscle connective tissues on muscle stiffness (19). Representative images of skeletal muscle SWV measurements were shown in Fig. 3.

The SPSS 20.0 statistical software was used for statistical analysis. Where the measurement data between the T2DM with DPN group and the T2DM without DPN group were normally distributed, they are expressed as mean ± SD. The Student t test was used for comparison between the two groups. Nonnormally distributed data are expressed by the median (interquartile spacing), and the comparison between the two groups adopts a nonparametric test. Count data are expressed as n (%) using χ² inspections. All significance tests were two tailed, with P < 0.05 as the difference of statistical significance. The correlation of clinical and biochemical metabolic parameters with SWV or CSA of the measured muscles was expressed with Pearson or Spearman bivariate correlation analysis. Parameters with statistical significance in the univariate analysis were included in multiple stepwise linear regression analysis. Logistic regression analysis was used to analyze the risk factors for DNP complications and to adjust for potential miscellaneous covariates. The correlation is expressed as the odds ratio and corresponding 95% CI, and odds ratio indicates the possibility of more DNP complications. Forty patients were randomly selected to be scanned twice by one observer for testing the repeatability of SWV and CSA measurements. The repeatability was evaluated using intraclass correlation coefficient (ICC) (two-way mixed, single measures).

The data generated during the current study are not publicly available because they contain information that could compromise research participant privacy but are available from the corresponding author on reasonable request.

Average age and duration of diabetes in our cohort of 203 patients were 67.8 years and 14 years, respectively. Mean HbA_1c value was 8.87%, indicating that the subjects had poor blood glucose control. We compared the baseline characteristics of T2DM patients with and without DPN (Table 1). Patients with DPN were older, had longer diabetes duration, and had higher HbA_1c and Toronto CSS scores (P < 0.05). In contrast, there was no significant difference in blood lipids (TG, CHO, HDL-C, and LDL-C) and renal function (BUN, creatinine) between the two groups. In addition, all nerve conduction velocities (NCV) in the T2DM without DPN group were significantly higher than those in the T2DM with DPN group.

The ultrasonic parameters of the lower-limb muscles are compared in Table 2. Both SWV and CSA of the AbH in patients with DPN decreased significantly (P < 0.05) compared with the T2DM without DPN group. The results suggest that AbH in patients with DPN became smaller and softer, as shown in Fig. 4. However, for TA and PER, no significant differences were noted in the SWV and CSA between the two groups. In addition, there were no significant differences in PAs of all the muscles between the two groups (P > 0.05).

To further analyze the indication of AbH differences, we performed linear regression analysis to investigate the correlation between the morphological and mechanical properties of AbH and various diabetes variables. Results are presented in Table 3. The results showed that AbH_CSA was positively correlated with AbH_SWV, and AbH_CSA was positively correlated with AbH_PA. By contrast, AbH_SWV was not significantly associated with AbH_PA. AbH_CSA was positively correlated with NCV (P < 0.05), except for the sensory conduction velocities of the median and superficial peroneal nerves. In contrast, AbH_SWV was positively correlated with NCV in all the parts (P < 0.05). AbH_CSA and AbH_SWV were negatively correlated with the course of diabetes, Toronto CSS score, and HbA_1c level (P < 0.05), as shown in Fig. 5. AbH_SWV was negatively correlated with age, whereas AbH_CSA showed little correlation. In contrast, AbH_CSA and AbH_SWV were not related to blood lipids, creatinine, urea, ABI, or SBP.

In the subsequent multiple stepwise regression analysis, parameters with statistical significance in univariate analysis and age as a covariate were included. We found that AbH_SWV, HbA_1c, and Toronto CSS scores and AbH_PA were independent predictors of AbH_CSA (Supplementary Table 1). In contrast, HbA_1c, AbH_CSA, and Toronto CSS were independent predictors of AbH_SWV (Supplementary Table 2). The results showed that the AbH became smaller with a higher Toronto CSS score, increase in HbA_1c, smaller AbH_PA, and softer AbH. On the other hand, with increase in HbA_1c, higher Toronto CSS score, and smaller AbH, AbH became softer.

Taking as the dependent variable whether DPN was merged and as independent variables the common DPN risk factors including age, course of disease, SBP, blood lipids, HbA_1c, Toronto CSS scores, and muscle ultrasound parameters (such as SWV, CSA, and PA), we performed a multivariate binary logistic regression analysis for patients with diabetes. The results showed that patients with higher HbA_1c and Toronto CSS scores, with smaller and softer AbH, were more likely to have DPN (Supplementary Table 3), whereas a contribution of PA to the prediction of DPN occurrence was not found.

In the repeatability test, the ICCs for intraobserver variation in measuring the CSA and SWV of AbH with ultrasonography were 0.64 (95% CI 0.42–0.79) and 0.73 (95% CI 0.54–0.85), respectively. The ICCs for intraobserver variation in measuring the CSA and SWV of TA with ultrasonography were 0.68 (95% CI 0.45–0.81) and 0.69 (95% CI 0.49–0.82). The ICCs for intraobserver variation in measuring the CSA and SWV of PER with ultrasonography were 0.62 (95% CI 0.39–0.78) and 0.67 (95% CI 0.46–0.81).

DPN is an irreversible disease associated with weakness or dysfunction of the plantar muscles (18,26). Early identification of DPN-related neuromuscular diseases and intervention programs to strengthen muscles are proven to effectively reduce the risk of falls in this population (7) and the incidence rate of foot ulcers caused by biomechanical abnormalities (5). The purpose of this study was to evaluate the morphological and mechanical changes of the lower-limb skeletal muscle during the progression of diabetes and explore the correlation of lower-extremity muscles’ properties with clinical indicators associated with DPN progression. Additionally, we sought to find a potential image indicator for monitoring the progress of DPN in patients with T2DM.

It is well-known that the loss of sensory feedback function and foot motor dysfunction damage the gait balance of patients with DPN. Currently, distal limb muscle strength has not been accurately evaluated. The evaluation indices of clinical paper-grip tests tend to be subjective (34,35). Toe flexor strength measurement is used to evaluate the combined strength of plantar muscles and external foot muscles and cannot distinguish between plantar intrinsic muscles and external foot muscles (36–38). MRI can accurately distinguish the intrinsic and external muscles of the foot in terms of morphology, and muscle morphology measurement is considered an alternative method to evaluate muscle mechanical function (5), whereas MRI is not conducive to large-scale population screening. Currently, high-resolution ultrasonography can dynamically and noninvasively measure small plantar muscles with limited accessibility. The convenience and speed of ultrasound examination are conducive to large-scale population screening; therefore, increasing amount of attention is being paid to clinical practice. However, there are few reports on the use of ultrasound technology to evaluate synergistic changes in intrafoot and extrafoot muscle morphology and mechanics in patients with DPN.

Our results showed that the AbH in DPN patients became smaller and softer, but there was no significant difference in the CSA and stiffness of TA and PER between the two groups, suggesting that the morphological and mechanical properties of muscle changes in the DPN group first occurred in the distal limb. Andersen et al. (39) held that atrophy of the plantar intrinsic muscles in DPN patients would damage the structure and function of the foot but the larger external muscle of the foot reflected compensatory activities supporting the function of the foot and ankle. However, in our measurements, the CSA and stiffness of the TA and PER in DPN patients did not increase compared with the control group. One possible reason was that the selected patients were older and had a longer disease course, resulting in muscle degeneration. Muscle size and stiffness are important indicators of muscle activity and functional status (18–21); additionally, muscle stiffness is helpful for regulating human motion and controlling joint stability (40). Further studies are needed to evaluate the effects of changes in the morphological and mechanical properties of intrinsic foot muscles on gait and plantar pressure of patients with DPN, in order to clarify the cause of gait imbalance and reduce the risk of falls.

The most used neuroelectrophysiological examination to diagnose DPN is measurement of the NCV, which is a useful quantitative method to evaluate DPN (41). As the largest intrinsic muscle of the internal foot, the AbH is closely related to foot function (8–10). Its position is shallow, and its muscle boundary is also clear, which means that it can be considered as a research object with good repeatability. Researchers have confirmed the correlation between neuroelectrophysiological examination and the thickness of the extensor digitorum brevis and second metatarsal muscle, determined with ultrasonography (42,43). To the best of our knowledge, no study has included investigation of the correlation between the morphological and mechanical properties of AbH and NCV. It was reported that lower NCV is related to the loss of the conducting axons (44). Length-dependent loss of motor axons may result in motor unit loss and DPN-related muscle atrophy. There is a consensus that intrinsic resting muscle tension remains in the form of spontaneous contractions due to “active” interaction between the filaments even when subjects are relaxed (45). Thus, resting muscle stiffness involves active spontaneous contractions in addition to indicating the muscle itself as elastic and/or viscoelastic. The muscle mechanical properties of AbH and NCV in the DPN group were worse than those in the control group, and a positive correlation was found between the mechanical properties of AbH and almost all the NCV results in this study. Therefore, the reduced stiffness of the AbH muscle may be partly due to the loss of axons in DPN leading to a reduction in muscle active components.

Recent studies (26,46) have included investigation of the morphological changes of AbH in patients with DPN, and the decrease in foot muscle volume is already considered to be related to the severity of neuropathy in clinical evaluation (47). However, no reports have discussed the correlation between the mechanical properties of AbH in patients with DPN and the severity of neuropathy. The Toronto CSS score, as an evaluation tool for the severity of DPN complications (48,49), is used for clinical trial efficacy observations and epidemiological investigations (50,51). Our multivariate stepwise regression analysis showed that Toronto CSS scores were independently influencing factors for AbH_CSA and AbH_SWV. Therefore, we believe that the mechanical properties of AbH are associated with the severity of DPN complications.

Aside from the Toronto CSS score, blood glucose control levels were also independent influencing factors of AbH_CSA and AbH_SWV. Hyperglycemia causes microvascular damage and ultrastructural changes in nerves, leading to slow NCV and deterioration of neurological function (41). Denervation causes muscle weakness, atrophy, and increased intramuscular fat (6). In addition, increased glycation end products and receptors caused by hyperglycemia can also lead to skeletal muscle atrophy and dysfunction (52). Therefore, the increase of blood glucose levels may lead to changes in the morphology and mechanical properties of AbH in patients with DPN.

There is still no complete consensus in the literature on whether muscle stiffness will decrease (53), increase (54), or not change (55) with age. The muscle aging process seems to involve a more significant decline for those >75 years old (56). The average age in our cohort of 203 patients was 67.8 years. Although potential age effects could exist, the influence of HbA_1c and Toronto CSS scores on the morphology and mechanical properties of muscle should be prior to age effects in our results.

It was reported that the muscle-fiber PA is positively related with the muscle CSA measured by ultrasonography in triceps brachii and quadriceps femoris (57,58). Similarly, we also found that the AbH presents a positive correlation between the CSA and PA. In contrast, no significant correlation was noted between the SWV and PA of the AbH. The propagation of the shear wave does not occur in the longitudinal direction of fibers because of the angle of probe spindle relative to muscle bundle (59), so that particular attention should be paid to interpreting shear wave elastography data obtained from pennate muscles (60). However, when the probe was positioned in parallel or 20° obliquely to the fascicle across the B-mode images, the reproducibility of shear modulus measurements was high for both parallel and oblique conditions (25). In the present results, the mean value of PA for AbH is only 9.18°, far less than 20°. No significant correlation was found between the SWV and PA of AbH. Thus, the effect of PA on SWV can be excluded.

Skeletal muscle is a complex active and passive tissue. Compared with the decreased AbH muscle stiffness in DPN patients found in this study, the increased muscle stiffness can be observed in other neurological diseases, such as stroke, spastic cerebral palsy, or Parkinson (61–64). The primary cause of the increased muscle stiffness in these diseases is abnormal excitability of spinal cord reflex and the loss of control on the spinal cord, which leads to hyperactive stretch reflex and tendon twitching (65). In addition to these, the mechanical properties of the muscle are also affected by its composition. Different from increased intramuscular fat in DPN patients, it was indicated that changes in the amount and arrangement of collagen in the extracellular matrix increase passive muscle stiffness in Parkinson disease and cerebral palsy (66–68).

After analyzing the clinical factors affecting the morphology and mechanics of AbH, we analyzed the independent factors affecting DPN complications. Age, course of diabetes (69), and blood glucose control (70) were previously considered to be related to DPN complications. However, our multivariate logistic regression analysis revealed that Toronto CSS score and HbA_1c level were independent influencing factors for DPN complications—not age or course. The Toronto CSS score includes neurological symptoms, nerve reflex, and sensory function test scores, so it is suspected that the Toronto CSS score weaken the influence of age and course on DPN complications. In addition, multivariate logistic regression analysis showed that the likelihood of diabetes with DPN increased with atrophy and softening of AbH. This indicated that the morphological and mechanical properties of AbH can be a potential imaging indicator for monitoring the progress of DPN in T2DM patients.

Although the mean differences in both SWV and CSA of the AbH muscle are small, they are significant. To our knowledge, there are not agreed minimum clinically important differences in SWV and CSA of the AbH muscle in the design, analysis, and interpretation of clinical research of muscle properties in patients with diabetes. This work provides a foundation for future exploration of minimum clinically important differences in clinical research of muscle properties in patients with diabetes.

In interpreting USWI measurements, it is important to understand the factors that potentially affect the measurements. In addition to general technical factors such as fiber orientation, ROI size, probe load, and measurement depth (71,72), muscle tissue poses a specific challenge for USWI; e.g., the nonzero tensile stress resulting from passive stretch and/or muscle contraction may confound measurements (71). Although the ankle and hallux were placed in neutral (and static) position for all ultrasound shear wave elastography measurements, the three muscles may still have small amounts of stretch, so SWV may be evaluated in the three muscles at different fiber length ranges (i.e., within a slack length range or absence of tensile load vs. tensile loaded). Moreover, long-term DPN can cause muscle atrophy, which is typically associated with muscle contractures or muscle shortening. As a result, the two testing groups may have different muscle lengths for the same ankle angle, which may increase data variability. Therefore, measuring SWV and CSA of the AbH muscle at rest and under stress, respectively, may reduce data variability and improve their correlation with clinical indicators.

In short, this study can aid in development of programs to strengthen muscles and monitor the progress of DPN complications. However, there are several limitations to be recognized. First, the small sample size may not represent a wider population of people with diabetes, and future research is needed with recruitment of more people of different occupations and backgrounds. Additionally, for reduction of data variability, all measurements of ultrasound data were performed by an operator and thus lacked interobserver assessment. Thirdly, lacking functional data on muscle strength is also a limitation of this research. In future, it could be necessary to further validate the morphological and mechanical characteristics of AbH in DPN patients measured by ultrasonography with functional data on muscle strength. Fourthly, with diabetes progression, muscle anisotropy might increase, which would mean that muscle SWV would be biased between T2DM with and without DPN patients. Finally, we did not compare the morphological and mechanical properties of the muscles with histology, as it is difficult to perform muscle biopsies in the current clinical setting.

Based on a series of experimental tests in this study, it was noted that AbH of the T2DM patients with DPN became smaller and softer, while the morphological and mechanical properties of TA and PER present nonsignificant change. This suggests that the changes in morphological and mechanical properties of muscle in the T2DM patients with DPN first occurred in the distal limb. With an increase in the Toronto CSS score and HbA_1c, the morphological and mechanical properties of AbH deteriorate. Multivariate logistic regression analysis indicated a decrease in AbH_SWV and AbH_CSA, an increase in the Toronto CSS score and HbA_1c, and a significant increase in incidence of DPN. All of these indicated that the morphological and mechanical properties of AbH could be a potential imaging indicator for monitoring the progress of DPN in T2DM patients.

Acknowledgments. The authors thank all the participants in the research of morphological and mechanical properties of lower-limb muscles in T2DM for their substantial contributions.

Funding. This work is supported by the National Natural Science Foundation of China (grant 51875187), Hunan Provincial Natural Science Fund (grant 2020JJ8048), Hunan Youth Talent Program (grant 2020RC3016), Key Research and Development Program of Hunan Province of China (grant 2022SK2105), and Major Projects of Changsha (grant kh2202003).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. Y.Z., L.Y., and T.X. collected data. Y.Z. and L.Y. wrote the draft of the manuscript. Y.Z., L.Y., and Z.J. analyzed data. M.F., L.T., and Z.J. reviewed the manuscript, made critical revisions, and approved the manuscript before submission. M.F. conceived and designed the studies. All authors reviewed the draft and approved the final version. Y.Z. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.