Whether coagulation dysfunction influences the onset and progression of diabetic peripheral neuropathy: A multicenter study in middle‐aged and aged patients with type 2 diabetes

Abstract Background Nearly half of patients with diabetes experience diabetic peripheral neuropathy (DPN), resulting in a mere 53% survival rate within 3 years. Aberrations in coagulation function have been implicated in the pathogenesis of microvascular complications, prompting the need for a thorough investigation into its role as a contributing factor in the development and progression of DPN. Methods Data were gathered from 1211 type 2 diabetes patients admitted to five centers from September 2018 to October 2022 in China. DPN was evaluated by symptoms and electromyography. Motor and sensory nerve conduction velocity (NCV) was appraised and the NCV sum score was calculated for the median, ulnar, and peroneal motor or sensory nerves. Results Patients with DPN exhibited alterations in coagulation function. (i) Specifically, they exhibited prolonged thrombin time (p = 0.012), elevated fibrinogen (p < 0.001), and shortened activated partial thromboplastin time (APTT; p = 0.026) when compared to the control group. (ii) After accounting for potential confounders in linear regression, fibrinogen, and D‐dimer were negatively related to the motor NCV, motor amplitude values, and mean velocity and amplitude. Also, fibrinogen was associated with higher Michigan neuropathy screening instrument (MNSI) scores (β 0.140; p = 0.001). This result of fibrinogen can be validated in the validation cohort with 317 diabetic patients. (iii) Fibrinogen was independently associated with the risk of DPN (OR 1.172; p = 0.035). In the total age group, DPN occurred at a slower rate until the predicted fibrinogen level reached around 3.75 g/L, after which the risk sharply escalated. Conclusions Coagulation function is warranted to be concerned in patients with type 2 diabetes to predict and prevent the occurrence of DPN in clinical practice.


| BACKG ROU N D
Diabetic peripheral neuropathy (DPN), one of the most common chronic complications of diabetes mellitus, 1 is estimated to occur in 10%-15% of newly diagnosed diabetics and to reach 50% in diabetic patients. 2,3While the 3-year survival rate for diabetic patients with DPN is approximately 53%. 4 DPN is a symmetrical length-dependent sensorimotor polyneuropathy that arises from chronic hyperglycemia, relevant metabolic disorders, cardiovascular risk factors, and microvascular alterations. 5The characteristics of DPN, including neuropathic symptoms, paresthesia, and loss of sensation, lead to a rising risk of burns, injuries, and diabetic foot ulcers. 6,7Recent research indicated that neuropathy may develop not only during diabetes but also in the prediabetic state. 8 is widely recognized that several biomarkers play a role in the diagnosis and prognosis of DPN, 9 such as age, duration of diabetes, hypertension, dyslipidemia, and smoking. 8Individuals with diabetes, encompassing both type 1 and type 2, may experience a hypercoagulable state with elevated clotting factors 10 and a prethrombotic environment due to prolonged thrombolysis time. 11Furthermore, pathological and laboratory studies have demonstrated that even low levels of systemic inflammation may result in the loss of both myelinated and unmyelinated nerve fibers, leading to disruptions in the blood-nerve barrier and microvascular system, 9,12 which has been identified as a potential mechanism leading to type 2 diabetes mellitus (T2DM) and its subsequent complications. 13,14Nevertheless, the detection of inflammatory markers can be technically challenging and costly, which limits their widespread use in routine clinical practice.
The dysfunction of coagulation function can significantly impact microvascular changes, which has been widely recognized to elevate the risk of DPN in patients with T2DM. 15,16Fibrinogen (FIB), the primary protein involved in blood coagulation, was reported to have a strong association with chronic systemic inflammation compared with hypercoagulability. 17In a study by Lianlian et al., FIB was identified as an independent risk factor for predicting type 2 diabetic nephropathy, 18 a key microvascular complication that can lead to kidney failure.Additionally, Zhuang et al. found that T2DM patients with a D-dimer level of ≥0.22 mg/L had a higher risk of developing DPN. 15To date, no research has assessed the impact of coagulation function on the severity of DPN based on electromyography (EMG), nor has any such research been stratified by age groups.
In this retrospective study, we aimed to investigate the potential of coagulation biomarkers in predicting the risk of DPN (the diagnosis based on EMG and neurological assessment) in Chinese patients with T2DM.Our goal was to assess the value of these biomarkers in aiding diagnosis and estimating prognosis in patients across various age groups.

| Patients
In the main cohort, patients who received nerve conduction examinations were recruited from the First Affiliated Hospital of Wenzhou Medical University between September 2018 and October 2022.Those with type 2 diabetic peripheral neuropathy were categorized into the DPN group (n = 614), while those with type 2 diabetes mellitus but without DPN were assigned to the T2DM group (n = 280).The remaining individuals without T2DM were selected as the control group (n = 133).In the external validation cohort, 196 patients with DPN and 121 patients with T2DM were also enrolled.A total of 317 individuals were included at Zhejiang Provincial People's Hospital, Ruian People's (p < 0.001), and shortened activated partial thromboplastin time (APTT; p = 0.026) when compared to the control group.(ii) After accounting for potential confounders in linear regression, fibrinogen, and D-dimer were negatively related to the motor NCV, motor amplitude values, and mean velocity and amplitude.Also, fibrinogen was associated with higher Michigan neuropathy screening instrument (MNSI) scores (β 0.140; p = 0.001).This result of fibrinogen can be validated in the validation cohort with 317 diabetic patients.(iii) Fibrinogen was independently associated with the risk of DPN (OR 1.172; p = 0.035).In the total age group, DPN occurred at a slower rate until the predicted fibrinogen level reached around 3.75 g/L, after which the risk sharply escalated.

| Inclusion and exclusion criteria
The inclusion criteria of the main cohort and external validation were patients with diabetes aged 18 years or older who had undergone nerve conduction examination and Michigan Neuropathy Screening Instrument (MNSI).The exclusion criteria of the main cohort and the external validation gathered: Patients with (1) type

| Clinical assessment
Clinical data were obtained from the patient's electronic history.
Diabetes complications were also recorded.Results were compared with normal reference ranges.The study protocol was approved by the Ethics Committee of the First Affiliated Hospital of Wenzhou Medical University, Zhejiang Provincial People's Hospital, Ruian People's Hospital, Shaoxing People's Hospital, and Anyang District Hospital, and all participants provided written informed consent to participate in this study.

| Neurological symptoms and examinations
The effects of confounding variables were minimized before (e.g., 24 h before) and during neuropathy detection: Stop or avoid drugs (such as beta-blockers), foods (such as spicy foods), or beverages (such as coffee) that affect heart rate or neurological function; avoid or reduce previous physical activity and emotional stress.The patients with symptomatic hyperglycemia, hypoglycemia, or ketonuria were excluded, or these symptoms were corrected beforehand.All the tests were conducted in a quiet laboratory. 20urological symptoms of DPN include burning, numbness, tingling, fatigue, cramps, and pain in the leg and/or foot.
Neurological signs were defined as reduced or absence of ankle reflexes (using an appropriate tendon reflex hammer) along with reduced or absence of distal sensation, including vibratory sensation (using a standard 128 Hz tuning fork), tactile perception (using a 10 g monofilament on 5 sits per foot), temperature sensation (using cold and warm objects), tingling sensation (using pins), and proprioception. 21

| Nerve conduction examination
The nerve conduction examination was conducted at a room temperature of 24°C, with the legs warmed using an electric heating pad for at least 10 min to achieve a skin temperature of 32-35°C.
A skilled technician performed all tests.The nerve conduction velocities (NCV) and nerve conduction amplitudes were measured for both sides of the upper and lower limbs, including motor and sensory branches of the median nerve; motor and sensory branches of the ulnar nerve; motor branch of peroneal nerve; motor branch of the tibial nerve; sensory branch of the superficial peroneal nerve.In addition, both sides of the tibial nerve Fwave were recorded, and the lower one was regarded as the final F-wave.Slowed/blocked nerve conduction was defined as more than 2.5 SD below the control nerve conduction threshold.Nerve conduction is defined as abnormal when two or more nerve abnormalities are detected. 22DPN was defined as the presence of abnormal nerve conduction together with neurological symptoms or signs. 20Vibration-sensing thresholds of the first metatarsal and tibia (10 cm below the knee) were determined by additional probes (vibrometer).
The velocities and amplitudes of all peripheral nerves examined in the study were standardized, and the standardized values for each patient were aggregated into the corresponding Z score.Slow/low nerve conduction (defined as velocity/amplitude Z score <−1, and approximately equal to the lowest tertile of the Z score) is diagnosed as abnormal nerve conduction by trained electrophysiologists. 23 The amplitude and velocity Z scores for males and females were calculated by the following formulas 24 : Summed amplitude:

| Michigan neuropathy screening instrument
The MNSI score is a comprehensive tool that includes both a 15item self-administered questionnaire and a structured examination of the feet (MNSIE).The MNSIE is scored for abnormalities of appearance, presence of ulcers, vibration perception at the distal great toe, and ankle reflexes.The score for each parameter is as follows: appearance of feet (normal = 0, abnormal = 1), ulceration (absent = 0, present = 1), ankle reflexes (absent = 1, present with reinforcement = 0.5, present = 0), and vibration perception (absent = 1, reduced = 0.5, present = 0).A trained and certified health professional performs the MNSIE to reduce interobserver variability.Evaluation of each parameter is made on both sides, and the maximum score is 8 points. 25A higher score indicates more severe neuropathy in the patient.

| Definition of disease
T2DM was diagnosed based on blood glucose measurements, medication data (linked to pharmacy dispensing data), and records from general practitioners.Patients with fasting glucose ≥7.0 mmol/L or nonfasting glucose levels ≥11.1 mmol/L (if there is no fasting glucose sample available), and those receiving glucose-lowering therapy were considered to have T2DM. 26normal nerve conduction is defined as the abnormality in two or more measured nerves.The presence of nerve conduction abnormality along with one or more neurological symptoms or signs is confirmed DPN. 20When any of the three elements of the screening process-symptoms, signs, or abnormal nerve conduction parame-  APTT (p = 0.026), while exhibiting higher levels of FIB (p < 0.001).
Notably, the DPN group demonstrated significantly higher levels of FIB (p < 0.001) than the T2DM group.Conversely, there were no significant differences in PT, INR, and D-dimer levels among the three groups.

| Coagulation function and EMG parameters
In the main cohort, the potential correlation between coagulation function and EMG parameters was assessed using multiple linear regression (Table 1).Notably, FIB levels exhibited a negative association with the NCV and amplitude of both upper and lower limbs,  S2).

| Coagulation function and MNSI exam
In the main cohort, through multivariate linear regression analysis, we discovered that FIB levels were associated with higher MNSI scores after adjusting for various DPN-related indicators, indicating severe neuropathy (β = 0.140, 95% CI: 0.056, 0.224, p = 0.001), whereas no significant association was observed with other coagulation biomarkers (Table 2).In the external validation, the relationship between FIB and MNSI exam in multivariate linear regression analysis was evident, after adjusting for various DPN-related indicators (β = 0.223, 95% CI: 0.027-0.419,p = 0.026, Table 2).TT levels were negatively associated with the severity of neuropathy (β = −0.165,95% CI: −0.283, −0.048, p = 0.006), whereas no significant association was observed between other coagulation biomarkers and MNSI exam.Importantly, the results of multiple linear regression validated the finding that FIB levels were positively correlated with the severity of neuropathy in both the main cohort and the validation cohort, consistent with the outcomes presented in Table 1.

| Coagulation function and diabetic peripheral neuropathy
In the main cohort, the relationship between coagulation function  other coagulation function biomarkers, like PT, TT, APTT, INR, and D-dimer, the results were insignificant.Given that FIB emerged as an independent predictor of the occurrence and severity of DPN, a more detailed analysis using EMG was conducted in the study.

| Unstandardized parameters and standardized Z score of EMG by tertiles of fibrinogen
Further analysis of EMG parameters after tripartite grouping according to FIB levels in the main cohort with 894 patients is performed in  S3.Similar results were also observed in the groups aged more than 55 years.

| Serum fibrinogen and diabetic peripheral neuropathy stratified by age
In the main cohort, correlation analysis with Spearman's test revealed a negative association between FIB levels and NCV and amplitude of upper and lower limbs, with all variables showing a p value of <0.001, except for superficial peroneal amplitude and velocity (p < 0.01) (Supplementary Table S4).Still, after stratified by age, FIB was correlated with a significant reduction in motor NCV, F-wave minimum latency, MNAmp, MNCV, SNAmp, and SNCV in each age group (p < 0.03).
Also, the nonlinear relationship between elevated FIB and the higher risk of DPN was further manifested in the main cohort with restricted cubic spline regressions with three knots (Figure 3).In the total age group, the development of DPN decreased until the predicted FIB levels reached 3.75 g/L, and then increased thereafter (p for nonlinearity = 0.019).A similar trend was observed in the group aged between 55 and 65 years, where the risk of DPN was relatively stable until around 3.82 g/L of the predicted FIB levels and then increased thereafter (p for nonlinearity = 0.028).However, no nonlinear relationship was found between FIB levels and DPN in groups with ages <55 years or >65 years.In the validation cohort, the risk of developing DPN decreased until the predicted FIB level reached 3.78 g/L, after which it increased sharply (Supplementary Figure S2; p nonlinearity = 0.027).

| DISCUSS ION
In light of our findings, we systemically and comprehensively examined the relationship among coagulation function, the prevalence TA B L E 2 Multiple linear regression analysis of the correlation between coagulation function and MNSI exam in both the main cohort and external validation.

F I G U R E 2
The odds ratio (95% confidence interval) of the presence of DPN in unadjusted analysis (A) and adjusted analysis (B).APTT, activated partial thromboplastin time; FIB, fibrinogen; INR, international normalized ratio.complications.Diabetes-induced thrombotic propensity and hypercoagulability, 11 along with chronic systemic inflammation 9,12 have all been identified as potential mechanisms underlying neuropathy (Figure 4).PT and APTT reflect the exogenous and endogenous functions of the thrombin system, respectively.TT reflects

Age > 65 years
to mediate platelet aggregation and ultimately lead to widespread microthrombosis, resulting in severe microvascular disease. 30n et al. screened several clinical frequently-used coagulation indicators and then concluded that FIB was an independent risk factor for predicting type 2 diabetic nephropathy, and the main microvascular complication leading to kidney failure. 18In a single-center study with a total of 561 subjects, the k value and angle α, reflecting the plasma FIB function, were discovered to be associated with the early diagnosis of DPN.They revealed that compared with diabetic patients without DPN, the level of FIB and the angle α were higher, and the k value was lower in patients with DPN. 16In contrast to the above study, our study paid attention to the nerve conduction obtained from EMG, applied a large sample of nerve conduction studies, and stratified patients based on different ages, to seek a further correlation between FIB and DPN.Moreover, the nerve conduction test, which is the gold standard for diagnosis, 1 is also employed for diagnosing and measuring the severity of DPN in our study, rather than merely applying questionnaires and examinations.
The course of diabetes plays an important role in the development of DPN, as a longer disease course is a potential risk factor for microvascular disease. 28Compared to patients with diabetes for <5 years, those with a of >5 years were four times more likely to develop microvascular complications. 31A cross-sectional study from Saudi Arabia involving 430 diabetic patients found that those with type 2 diabetes for >10 years were more likely to develop chronic diabetic complications. 32Consistent with the previous study, in our study, the course of diabetes was strongly associated with the risk of DPN (adjusted OR, 1.110(1.077-1.146),p < 0.001).Prolonged hyperglycemia can damage blood vessels, leading to cellular damage in neurons and Schwann cells of peripheral nerves, which may result in DPN. 33e middle-aged group (aged 55-65 years) is a unique middlemore time window, during which age-related chronic diseases may be in the preclinical stage, and symptoms and signs are too mild to be detected by clinicians. 34In our study, a significant correlation between high levels of FIB and abnormal nerve conduction was observed at all stages, whereas, there exists a big gap in each age group.
For instance, the absolute value of the Spearman correlation coefficient in the middle-aged group (55 ≤ age ≤ 65) and elder (age > 65) group was much greater than in the younger group (age < 55) for most of the EMG parameters.Additionally, the probability of developing DPN was found to be higher in the middle-aged group (65.8%) and elder group (67.0%) than in the younger group (52.3%).Hence, in clinical practice, more frequent DPN screening tests or earlier EMG

:| 3 of 12 XIE
Coagulation function is warranted to be concerned in patients with type 2 diabetes to predict and prevent the occurrence of DPN in clinical practice.K E Y W O R D S coagulation function, diabetic peripheral neuropathy, electromyography, type 2 diabetes et al.Hospital, Shaoxing People's Hospital in Zhejiang Province, and Anyang District Hospital in Henan Province, from October 2021 to October 2022 (Supplementary Figure S1).

1 |
logistic regression analysis was performed to determine the relationship between coagulation function and the risk of DPN, manifesting that serum FIB served as an independent factor for the occurrence of DPN.Restricted cubic spline curves with three knots (defined at the 10th, 50th, and 90th percentiles) for the development of DPN were further plotted to provide more precise estimates and examine the nonlinear relationships.All tests were bilateral, and p < 0.05 was considered statistically significant.3| RE SULTS3.The distribution of coagulation function indicators in three groupsIn the main cohort, after applying the inclusion and exclusion criteria, a total of 614 patients diagnosed with type 2 diabetic peripheral neuropathy (DPN group), 280 patients with type 2 diabetes without DPN (T2DM group), and 133 individuals without diabetes (control group) were recruited from the First Affiliated Hospital of Wenzhou Medical University and underwent EMG examination.The distribution of coagulation function indicators among the DPN, T2DM, and control groups is illustrated in Figure 1.Compared to the control group, the DPN group showed longer TT (p = 0.012) and shorter except superficial peroneal sensory amplitude, after adjusting for possible confounding factors (all p < 0.03).Furthermore, FIB levels were found to be significantly correlated with reduced mean motor nerve amplitude (MNAmp; β = −0.464,p < 0.001), mean motor nerve conduction velocity (MNCV; β = −0.644,p < 0.001), mean sensory nerve amplitude (SNAmp; β = −1.595,p < 0.001), and mean sensory nerve conduction velocity (SNCV; β = −0.589,p < 0.001).Moreover, D-dimer levels were observed to be associated with reduced motor NCV and motor amplitude (p < 0.01) and were significantly correlated with depressed MNAmp (β = −1.224,p < 0.001), MNCV (β = −1.251,p < 0.001), whereas no association with EMG parameters was identified for other coagulation function indicators, such as PT, TT, APTT, and INR.In the external validation cohort, the baseline characteristics of 196 patients with DPN and 121 patients with type 2 diabetes without DPN are detailed in Supplementary Table S1.The DPN group exhibited higher MNSI scores (p < 0.001) and FIB levels (p = 0.021) compared to the T2DM group.In multiple linear regression, we found that FIB levels were negatively associated with motor amplitude values (p < 0.01) and significantly correlated with the reduced MNAmp (β = −0.793,p < 0.001), MNCV (β = −0.830,p < 0.001), SNAmp (β = −2.664,p < 0.001), and SNCV (β = −0.783,p < 0.001, Supplementary Table and the risk of DPN in type 2 diabetic patients in logistics regression is shown in Figure2.After adjusting for potential confounders, serum FIB was found to be independently correlated with the incremental odds of DPN in the multivariate logistic analysis (adjusted OR, 1.172 (1.011-1.360),p = 0.034).However, concerning F I G U R E 1 Distribution of coagulation function parameters in each disease group.The distribution of coagulation function parameters (Prothrombin time, thrombin time, APTT, INR, D-dimer, and FIB) in DPN, T2DM, and control groups.DPN group patients with type 2 diabetic peripheral neuropathy, T2DM group patients with type 2 diabetes without DPN, control group individuals without diabetes, APTT, activated partial thromboplastin time; FIB, fibrinogen; INR, international normalized ratio.***p < 0.001, **p < 0.01, *p < 0.05.TA B L E 1 Multiple linear regression analysis of the correlation between coagulation function and electromyography parameters (N = 894).

F I G U R E 3
the conversion time of fibrinogen to fibrin, and D-dimer is a degradation product of fibrinogen.15Elevated levels of hyperlipemia and hyperglycemia induce chronic systemic inflammation, which can lead to platelet activation and disorder of the coagulation and anticoagulation systems, ultimately resulting in hypercoagulability and thrombosis.27This exacerbates the course of diabetes and increases the risk of complications, such as the occurrence of DPN.28 FIB, a major component of the temporary extracellular matrix formed after tissue damage, participates in blood clotting and is associated with insulin resistance syndrome and thrombosis, thereby promoting a state of hypercoagulation and hyperviscosity in the blood.29Moreover, FIB can directly participate in coagulation and combine with platelet membrane protein IIb/IIIa Multivariable adjusted odd ratios for DPN according to levels of FIB on (A) all groups, (B) group aged <55 years, (C) group aged between 55 and 65 years, and (D) group aged >65 years.Multiple spline regression analysis with three knots is utilized in DPN.The solid line indicates the odds ratio, while the shadow indicates 95% CI.Reference lines for no association are indicated by the dashed lines at an odd ratio of 1.0.Data are adjusted for age, course of diabetes, hypertension, hyperlipidemia, BMI, smoking, FIB, FPG, HbA1c, TG, TC, and CRP.DPN, diabetic peripheral neuropathy; FIB, plasma fibrinogen; FPG, fasting plasma glucose; TC, total cholesterol; TG, triglycerides.

F I G U R E 4
Schematic diagram of multiple influences affecting DPN.The left panel represents the biomarkers associated with DPN and the right panel represents the coagulation pathways involved in the coagulation index.
Spearman's test was used to explore the correlation between FIB and EMG parameters.After adjusting for age, course of diabetes, hypertension, hyperlipidemia, BMI, smoking, FPG, HbA1c, TG, TC, and CRP, multiple linear regression analysis was carried out to further analyze the relationship between the levels of coagulation function, abnormal nerve conduction, and MNSI exam.Multivariate ters-was abnormal, medical records would be scrutinized to see if these participants had accepted a previous DPN diagnosis by a specialist.Participants were allowed to self-report any previous diagnosis of DPN, which was later verified and recorded.All data obtained from each participant's screening process and medical records underwent evaluation by a panel of experts with considerable experience in diagnosing neuromuscular diseases.This panel included a seasoned neuromuscular specialist, a neurophysiologist, and a physician with expertise in epidemiology.2.8 | Statistical analysisnormal distribution, and the Chi-square test or Fisher's exact test was used for the dichotomous variables.Normally distributed continuous variables were expressed as the mean ± standard error of the mean (mean ± SEM), while nonnormally distributed continuous variables were expressed as median and quartile ranges (median, IQR).The dichotomous variables were presented as percentages (%).

Table 3 .
The age and proportion of patients with DPN increased with among the low, medium, and high FIB groups were still apparent in the group aged under 55 years, the detailed results are listed in Supplementary Table