Hu D, Lin XY, Zhao TQ, Ji H, Chen MW, Zhao XT. Clinical significance of serum thrombospondin-4 for interventional restenosis in diabetes patients complicated by lower extremity arteriosclerosis obliterans. World J Diabetes 2026; 17(10): 124271 [DOI: 10.4239/wjd.124271]
Corresponding Author of This Article
Xiao-Tong Zhao, MD, Department of Endocrinology, The First Affiliated Hospital of Anhui Medical University, No. 218 Jixi Road, Hefei 230032, Anhui Province, China. zxt522154846@163.com
Research Domain of This Article
Endocrinology & Metabolism
Article-Type of This Article
research-article
Open-Access Policy of This Article
This article is an open-access article which was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/
Baishideng Publishing Group Inc, 7041 Koll Center Parkway, Suite 160, Pleasanton, CA 94566, USA
Share the Article
Hu D, Lin XY, Zhao TQ, Ji H, Chen MW, Zhao XT. Clinical significance of serum thrombospondin-4 for interventional restenosis in diabetes patients complicated by lower extremity arteriosclerosis obliterans. World J Diabetes 2026; 17(10): 124271 [DOI: 10.4239/wjd.124271]
World J Diabetes. Oct 15, 2026; 17(10): 124271 Published online Oct 15, 2026. doi: 10.4239/wjd.124271
Clinical significance of serum thrombospondin-4 for interventional restenosis in diabetes patients complicated by lower extremity arteriosclerosis obliterans
Die Hu, Hua Ji, Ming-Wei Chen, Xiao-Tong Zhao, Department of Endocrinology, The First Affiliated Hospital of Anhui Medical University, Hefei 230032, Anhui Province, China
Xiao-Yue Lin, Department of General Practice, The First Affiliated Hospital of Anhui Medical University, Hefei 230032, Anhui Province, China
Tian-Qi Zhao, Department of Endocrinology, The Affiliated Chuzhou Hospital of Anhui Medical University, Chuzhou 239001, Anhui Province, China
Co-corresponding authors: Ming-Wei Chen and Xiao-Tong Zhao.
Author contributions: Hu D and Lin XY contributed to the conception and design of the study, they contributed equally to this article, they are the co-first authors of this manuscript; Hu D, Lin XY, Zhao TQ, and Ji H contributed to participant recruitment and clinical sample collection; Hu D, Lin XY, and Zhao XT performed data collection and analysis; Chen MW supervised the study and acquired funding; Chen MW and Zhao XT contributed equally to this article, they are the co-corresponding authors of this manuscript; and all authors reviewed and approved the final manuscript and agreed to its publication.
AI contribution statement: AI tools, specifically DeepSeek, were used solely for language polishing and formatting assistance. No AI tool was used to generate research data, interpret results, or formulate conclusions. All AI-assisted content was critically reviewed and revised by the authors, who take full responsibility for the accuracy, originality, and integrity of the manuscript.
Supported by the Clinical Medical Research Translation Special Project of Anhui Province in China, No. 202427b10020133; and the Anhui Provincial Health Research Project, No. AHWJ2023BAc10012.
Institutional review board statement: This study was approved by the Medical Ethics Committee of the First Affiliated Hospital of Anhui Medical University, approval No. PJ 2024-09-88.
Clinical trial registration statement: This study is registered at the Chinese Clinical Trial Registry (Registration ID: ChiCTR2500096943). Details are available at https://www.chictr.org.cn/bin/project/edit?pid=258670.
Informed consent statement: All study participants provided written consent prior to study enrollment.
Conflict-of-interest statement: All the authors report no relevant conflicts of interest for this article.
CONSORT 2010 statement: The authors have read the CONSORT 2010 Statement, and the manuscript was prepared and revised according to the CONSORT 2010 Statement.
Data sharing statement: There are no additional data available.
Corresponding author: Xiao-Tong Zhao, MD, Department of Endocrinology, The First Affiliated Hospital of Anhui Medical University, No. 218 Jixi Road, Hefei 230032, Anhui Province, China. zxt522154846@163.com
Received: June 11, 2026 Revised: July 20, 2026 Accepted: September 4, 2026 Published online: October 15, 2026 Processing time: 124 Days and 0.6 Hours
Abstract
BACKGROUND
Restenosis after interventional procedures is a big challenge for type 2 diabetes mellitus (T2DM) patients who also have lower extremity arteriosclerosis obliterans (LEASO). Identifying biomarkers that predict restenosis risk and developing precise predictive models are essential for improving patient treatment outcomes. Thrombospondin-4 (TSP-4) is an extracellular matrix protein, which is associated with vascular inflammation. However, its relationship with postoperative restenosis still unclear.
AIM
To investigate whether TSP-4 levels predict restenosis in T2DM patients with LEASO and to develop a nomogram for risk estimation.
METHODS
This prospective cohort study included 143 T2DM patients with LEASO who underwent endovascular intervention. Clinical data and preoperative TSP-4 levels were collected. After 6 months of follow-up, they were divided into restenosis and non-restenosis groups. The cohort was randomly divided into a training and a validation set. Logistic regression was used to select independent factors and construct a nomogram model. The performance of the model was evaluated by receiver operating characteristic analysis, calibration curve, and decision curve analysis.
RESULTS
Among the 100 patients in the training cohort, 45 patients developed restenosis (45.0%). Patients in the restenosis group had significantly higher preoperative serum TSP-4 levels compared to those without restenosis (P < 0.001). Multivariate logistic regression analysis showed that TSP-4, low-density lipoprotein cholesterol, and age were independent predictors of restenosis (all P < 0.05). The nomogram demonstrated good discrimination. The area under the curve was 0.839 in the training cohort and 0.826 in the validation cohort. Calibration curves indicated a high degree of agreement between predicted probabilities and actual observations. Additionally, decision curve analysis confirmed a notable net clinical benefit across a wide range of threshold probabilities.
CONCLUSION
In T2DM patients with LEASO, serum TSP-4 is a useful predictor of restenosis after interventional treatment. The nomogram incorporating this marker provides risk stratification and guides individualized management.
Core Tip: In this prospective study on type 2 diabetes mellitus patients with lower extremity arteriosclerosis obliterans, after multivariable adjustment, serum thrombospondin-4 was found to be a new independent factor for restenosis, along with age and low-density lipoprotein cholesterol. The nomogram with these three variables, had strong discriminative power and good calibration, so it allows for earlier recognition of high risk patients and personalized interventions.
Citation: Hu D, Lin XY, Zhao TQ, Ji H, Chen MW, Zhao XT. Clinical significance of serum thrombospondin-4 for interventional restenosis in diabetes patients complicated by lower extremity arteriosclerosis obliterans. World J Diabetes 2026; 17(10): 124271
Lower extremity arteriosclerosis obliterans (LEASO) is a prevalent vascular complication of type 2 diabetes mellitus (T2DM) and a major cause of non-traumatic amputation in diabetic patients[1]. Diabetic patients with lower extremity lesions usually have multisegmental, bilateral involvement, and often affect infrapopliteal arteries, causing severe foot ischemia, which requires surgical intervention[2]. Percutaneous transluminal angioplasty (PTA) is the main treatment method for this disease[3]. However, restenosis after interventional treatment remains a challenging clinical problem, which significantly affects the quality of life of patients[4]. Restenosis after intervention involves complex pathological processes, like endothelial dysfunction, vascular smooth muscle cell phenotype switching, inflammatory reactions, and metabolic imbalance of extracellular matrix (ECM)[5,6]. Based on the above pathogenesis, beyond conventional risk factors, recent studies have discovered new predictive indicators such as the neutrophil to lymphocyte ratio[7], microRNAs[8], and high mobility group box protein B1[9]. Biomarkers that can predict restenosis are helpful for early identification of high risk patients, guiding the formulation of treatment plans, and improving prognosis. However, most previous studies were conducted on the general population and lacked systematic research specifically targeting diabetic patients. These patients have unique clinical and physiological characteristics, which increase the risk of restenosis.
Thrombospondin-4 (TSP-4) is an ECM protein related to vascular inflammation. It is produced by endothelial and vascular smooth muscle cells and exerts diverse biological effects through multiple surface receptors[10]. The protein is widely distributed across tissues, including the vascular wall[11], skin[12], myocardium[13], and nervous tissue[14]. Evidence indicated that TSP-4 plays an important role in locally regulating vascular inflammation. In healthy vascular walls, the expression level of TSP-4 is low, but it significantly increases in atherosclerotic lesions[15]. The transforming growth factor-β signaling pathway plays a core role in restenosis, and its receptors and downstream signaling molecules are promising therapeutic targets[16]. TGF-β specifically promotes the expression of TSP-4 in endothelial cells, which may mediate some of its biological functions[17]. Moreover, more and more studies have shown that TSP-4 also plays a role in tissue repair after injury. TSP-4 expression increases in microvascular endothelial cells after cerebral ischemia-reperfusion injury, possibly affecting the outcome of stroke through repair mechanisms[18]. In skin wounds, TSP-4 also aids healing by promoting fibroblast migration and keratinocyte proliferation[19]. Importantly, in endovascular treatment, the excessive reparative reaction to local mechanical trauma forms the basis of restenosis. Studies on animals have showed that TSP-4 is involved in neointimal hyperplasia after vascular injury and can reflect the inflammatory state of the vascular wall[20]. Recent studies have also found that TSP-4 is related to tumor cell proliferation and migration[21]. Based on these findings, TSP-4 may be involved in the vascular remodeling process after interventional injury. However, the link between TSP-4 and restenosis in patients with LEASO has not been studied, and whether it can act as a predictor of restenosis in the diabetic group still needs further exploration.
Therefore, this study examined the association of serum TSP-4 levels with vascular restenosis in 143 T2DM patients complicated by LEASO following endovascular intervention. Furthermore, multivariable logistic regression was performed and then a nomogram was built that combined serum TSP-4 with clinical features. The discrimination, calibration, and clinical applicability of the model were assessed in both the training and the validation cohort. The results show that serum TSP-4 level may serve as a novel biomarker for predicting restenosis in T2DM patients complicated by LEASO. The nomogram is expected to be a practical tool for classifying risk status, aiding in the decision making process, and contributing to the refinement of prognostic assessments in this patient population.
MATERIALS AND METHODS
Study design
A total of 183 patients with T2DM complicated by LEASO scheduled to undergo PTA at the First Affiliated Hospital of Anhui Medical University between September 2024 and November 2025 were enrolled. Enrollment was consecutive, with no interruptions or omissions in subject selection. The following criteria were used for inclusion: Each patient fulfilled the established criteria for LEASO diagnosis and was eligible for PTA[22], and underwent a first successful lower extremity endovascular intervention; fulfilled the definition of T2DM[23]. Fontaine stage II-IV, presenting with intermittent claudication, rest pain, ulceration, or gangrene; regular postoperative follow-up with arterial ultrasonography of the lower limbs and computed tomography angiography (CTA). Exclusion criteria included: Patients with acute complications of T2DM; patients with vascular stenosis caused by other diseases such as Takayasu arteritis or syphilis; patients with prior lower limb trauma or surgery, those requiring conventional bypass, or those with tumors or inflammatory diseases; patients with severe impairment of vital organ; patients with incomplete clinical and follow-up data; and pregnant or lactating women.
Throughout the study period, consistent inclusion and exclusion criteria were applied. Enrollment was consecutive, with no interruptions or omissions in subject selection. All treatment strategies, including the choice of surgical approach, preoperative and postoperative medications, and intraoperative procedures, were determined by experienced surgeons based on clinical guideline consensus, multidisciplinary assessment, and patient preferences. Every participant gave written informed consent before being enrolled. The entire patient cohort underwent a 6-month follow-up period after PTA, with the study endpoint being the occurrence of restenosis. Patients received lower limb arterial ultrasound or CTA at 1 month, 3 months, and 6 months during the follow-up period. At the same time points, follow-up data were collected through outpatient medical records and telephone interviews. Restenosis after intervention was defined as > 50% vascular stenosis detected by lower extremity arterial ultrasound and CTA[24-27].
After removing 16 patients who did not complete follow-up, the final statistical analysis comprised 143 patients, representing a 10.1% loss rate. To develop and validate the prediction model, a computer generated random number method was used to randomly allocate the final enrolled patients into a training set (70%, n = 100) and a validation set (30%, n = 43) at a ratio of 7:3. Independent predictors were identified and the nomogram was built using the training cohort, while the validation cohort examined the model’s robustness and applicability. We assessed whether the sample size was sufficient using the EPV criterion, a standard approach in multivariable logistic regression. Patients in the training cohort were divided into restenosis and non-restenosis groups according to the occurrence of restenosis. Figure 1 illustrates the flow chart.
Figure 1 Study design and patient flow.
Of 183 initially enrolled patients, 143 were finally analyzed after exclusions and loss to follow-up and randomly allocated to a training cohort and a validation cohort. The training cohort was divided into the restenosis and non-restenosis groups. Risk factors were analyzed, and a nomogram was developed and validated. T2DM: Type 2 diabetes mellitus; LEASO: Lower extremity arteriosclerosis obliterans; PTA: Percutaneous transluminal angioplasty; TSP-4: Thrombospondin-4; ROC: Receiver operating characteristic.
General data collection and laboratory index testing
Baseline information was collected upon admission, including age, sex, diabetes duration, medical history (coronary heart disease, cerebral infarction), personal history (smoking history, drinking history), Fontaine stage, and body mass index (BMI). Laboratory parameters recorded were hemoglobin (Hb), albumin (ALB), alanine aminotransferase (ALT), aspartate aminotransferase (AST), total bilirubin (TBIL), serum creatinine (Scr), fasting plasma glucose (FPG), glycated hemoglobin A1c (HbA1c), total cholesterol (TC), triglycerides (TG), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), fibrinogen (FIB), D-dimer (D-D), and high-sensitivity C-reactive protein (hs-CRP). Serum TSP-4 levels were measured using an enzyme-linked immunosorbent assay kit (Wuhan Huamei Bioengineering Co., China) in line with the manufacturer’s instructions.
Statistical analysis
Statistical analyses were performed using SPSS version 25.0 and R software version 4.2.2. Quantitative data following a normal distribution were expressed as mean ± SD, and comparisons between two groups were performed using the independent samples t test. Quantitative data not following a normal distribution were expressed as median (interquartile range) [M (P25, P75)], and comparisons between two groups were performed using the Mann-Whitney U test. Categorical data were expressed as number (percentage) [n (%)]. The χ2 test was used for comparisons of categorical data, and the Mann-Whitney U test was applied for ordinal data. Spearman correlation analysis was applied to examine the associations of TSP-4 with other clinical parameters. Univariate analysis and multivariable logistic regression were performed in the training cohort to identify independent risk factors. The variance inflation factor (VIF) was calculated to detect multicollinearity, and variables with a VIF ≥ 10 were excluded. Using the rms package in R software, we built a nomogram prediction model from the logistic regression outputs. Model discrimination was quantified by the receiver operating characteristic curve and the area under the curve (AUC). Assessment of calibration relied on the HosmerLemeshow test and calibration plots. Bootstrap internal validation estimated the overfitting risk, and decision curve analysis (DCA) assessed the clinical applicability of the model. A two-tailed P < 0.05 was considered statistically significant.
RESULTS
Baseline characteristics
A total of 143 patients were finally enrolled and randomized into the training cohort (n = 100) and the validation cohort (n = 43). The preoperative serum TSP-4 levels and clinical parameters were compared between the two cohorts, and the results are presented in Table 1. The training and validation cohorts exhibited highly comparable baseline features, with no major differences or apparent bias. The two cohorts were comparable in terms of demographic characteristics, unhealthy habits, and medical history (P > 0.05). Furthermore, the Fontaine stage, which was used to assess the severity of limb ischemia, also showed no significant difference between the cohorts. Intergroup comparisons of routine laboratory parameters showed no significant differences (P > 0.05). The training cohort had a mean serum TSP-4 level of 61.26 ± 29.46 ng/mL, compared to 64.47 ± 28.34 ng/mL in the validation cohort, and the difference was not statistically significant (P > 0.05).
Table 1 Comparison of baseline features of the training and validation cohorts, n (%)/mean ± SD.
Univariate analysis of postoperative restenosis in the training cohort
In the training cohort (n = 100), 45 patients experienced restenosis within 6 months of endovascular treatment, corresponding to a 45% incidence rate. As shown in Table 2, the restenosis group was older and had a higher proportion of patients with Fontaine stage IV than the non-restenosis group (P < 0.05). Compared with the non-restenosis group, the restenosis group exhibited markedly elevated HbA1c, LDL-C, D-D, and hs-CRP levels (P < 0.05). In addition, patients with restenosis had markedly higher serum TSP-4 levels than those without (Z = -5.020, P < 0.001). No marked intergroup disparities were found for sex, smoking, drinking, coronary heart disease, cerebral infarction, BMI, Hb, ALB, ALT, AST, TBIL, Scr, FPG, TG, TC, HDL-C, or FIB (P > 0.05). These results suggest that restenosis significantly correlates with older age, severity of symptoms, dyslipidemia, hyperglycemia, inflammation, and coagulation dysfunction.
Table 2 Univariate analysis of postoperative restenosis in the training cohort, n (%)/mean ± SD.
Correlation analysis of TSP-4 with clinical parameters in the training cohort
To further investigate the association between TSP-4 and other clinical parameters, a correlation analysis was performed. As most variables exhibited skewed distributions, Spearman correlation analysis was used. The correlation heatmap is shown in Figure 2. The analysis showed that serum TSP-4 level was positively correlated with age (rs = 0.300, P = 0.001), smoking (rs = 0.202, P = 0.044), Fontaine stage (rs = 0.326, P = 0.001), FPG (rs = 0.214, P = 0.033), HbA1c (rs = 0.465, P < 0.001), LDL-C (rs = 0.662, P < 0.001), D-D (rs = 0.249, P = 0.012), and hs-CRP (rs = 0.937, P < 0.001).
Figure 2 Spearman correlation heatmap of thrombospondin-4 and other clinical parameters in type 2 diabetes mellitus patients with lower extremity arteriosclerosis obliterans.
Positive correlations appear in yellow, negative correlations in blue. Color intensity scales with the magnitude of the relationship. BMI: Body mass index; Hb: Hemoglobin; ALB: Albumin; TBIL: Total bilirubin; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; Scr: Serum creatinine; FPG: Fasting plasma glucose; HbA1c: Glycated hemoglobin A1c; TG: Triglycerides; TC: Total cholesterol; LDL-C: Low-density lipoprotein cholesterol; HDL-C: High-density lipoprotein cholesterol; FIB: Fibrinogen; D-D: D-dimer; hs-CRP: High-sensitivity C-reactive protein; TSP-4: Thrombospondin-4.
Feature selection by multivariable logistic regression analysis
A multivariable logistic regression model was built by performing backward stepwise elimination, which involved identifying and including variables that were found to be significant in univariate analysis (Table 3). The coding of each variable is shown in Table 4. After adjusting for other confounding factors, serum TSP-4 remained a strong independent predictor of restenosis [odds ratio (OR) = 1.033, 95% confidence interval (CI): 1.007-1.059; P < 0.05], indicating that each 1 ng/mL increase in preoperative TSP-4 was associated with an approximately 3.3% higher risk of restenosis. Age (OR = 1.058, 95%CI: 1.000-1.120; P < 0.05) and LDL-C (OR = 2.437, 95%CI: 1.051-5.653; P < 0.05) were also independent risk factors. All VIF values were below 5, indicating no significant multicollinearity (Table 5).
Table 3 Multivariable logistic regression analysis for restenosis.
Construction of the nomogram model for interventional vascular restenosis
To evaluate restenosis risk following intervention in T2DM patients with LEASO, we developed a nomogram model by multivariable logistic regression. The nomogram included age, LDL-C, and TSP-4 (Figure 3). To obtain the predicted restenosis risk, clinicians located the value on each axis, read its point, summed the three points, and projected the total onto the probability scale. For example, for every 20 ng/mL increase in preoperative serum TSP-4 level, the corresponding point increased by 12.5 points. When the total score reached 161, the predicted risk of restenosis was 0.9.
Figure 3 Nomogram of interventional restenosis risk in type 2 diabetes mellitus patients complicated by lower extremity arteriosclerosis obliterans.
The nomogram incorporates age, low-density lipoprotein cholesterol, and thrombospondin-4. The total points of the three predictors correspond to the predicted probability of restenosis. LDL-C: Low-density lipoprotein cholesterol; TSP-4: Thrombospondin-4.
Interpretation of the nomogram model
We used receiver operating characteristic analysis to evaluate the model in the training and validation cohorts (Table 6 and Figure 4). In the training cohort, the AUC was 0.839 (95%CI: 0.762-0.916). Sensitivity was 0.689, and specificity was 0.855 (Figure 4A). The validation cohort produced an AUC of 0.826 (95%CI: 0.703-0.950), with a sensitivity of 0.800 and a specificity of 0.739 (Figure 4B). These results indicate that the model possesses good discriminatory ability. The Hosmer-Lemeshow test showed χ2 = 9.606, P = 0.294 in the training cohort and χ2 = 9.140, P = 0.331 in the validation cohort, indicating good calibration. Model performance was further assessed using bootstrap resampling (500 iterations). The calibration curve results showed that the bias-corrected curve closely matched the ideal line, and the deviation of the observed curve from the corrected curve was extremely small. These results supported that the calibration effect of the model was good (Figure 5).
Figure 4 Receiver operating characteristic curve analysis of the prediction model.
A: Training cohort; B: Validation cohort. In the training cohort, the area under the curve was 0.839 (95% confidence interval: 0.762-0.916). In the validation cohort, the area under the curve was 0.826 (95% confidence interval: 0.703-0.950).
Figure 5 Calibration plots of the predictive model.
A: Training cohort; B: Validation cohort. The black diagonal line represents the ideal curve. The yellow curve is the observed curve, and the blue curve is the bias-corrected calibration curve after 500 bootstrap iterations.
Table 6 Comparison of predictive efficacy of the nomogram model for interventional restenosis between the training and validation cohorts.
DCA was subsequently performed to evaluate the clinical utility of the nomogram. Within a wide range of threshold probability intervals, the model showed a net benefit that was higher than the treat-all or treat-none strategies in both the training (Figure 6A) and the validation (Figure 6B) cohorts. This indicates that the model has clinical practical value and can be used for individualized risk assessment and to provide a basis for clinical treatment plan formulation.
Figure 6 Decision curve analysis of the prediction model.
A: Training cohort; B: Validation cohort. Yellow curve: Treat all strategy; Black horizontal line: Treat none strategy; Blue curve: Nomogram.
DISCUSSION
LEASO is a common vascular complication of T2DM, characterized pathologically by progressive luminal stenosis or occlusion due to atherosclerosis, leading to chronic limb ischemia[28]. Without timely and effective intervention, it can result in limb necrosis and even amputation[29]. As a minimally invasive endovascular procedure, PTA is now the first line therapy for LEASO, in contrast to conventional open surgery[30]. However, the high incidence of restenosis after endovascular therapy significantly compromises its long term efficacy[31-33]. Therefore, identifying potential risk factors to predict postoperative restenosis and implementing targeted interventions for high risk populations have become crucial for improving long term outcomes in patients with LEASO. In recent years, although some progress has been made in identifying risk factors for restenosis, discovering novel biomarkers that reflect the underlying pathological mechanisms and constructing and validating effective predictive models remain major challenges. Moreover, systematic studies specifically focusing on the diabetic population are still lacking. This study systematically analyzed the relationship of serum TSP-4 levels with postoperative restenosis in T2DM patients complicated by LEASO, and constructed and validated a nomogram model incorporating TSP-4 to predict the risk of restenosis.
In this study, univariate analysis was initially performed to screen a range of clinical indicators. The results showed that patients in the restenosis group were older, had higher Fontaine stages, and presented with higher preoperative levels of HbA1c, LDL-C, hs-CRP, D-D, and TSP-4 compared with those without restenosis. These findings suggest that restenosis after endovascular therapy is influenced by multiple factors. Given that univariate analysis is subject to confounding, we further performed multivariate logistic regression analysis, which revealed that only age, LDL-C, and TSP-4 were retained as independent risk factors (OR = 1.058, 2.437, and 1.033, respectively; all P < 0.05). This suggests that T2DM patients with older age, higher preoperative LDL-C, and higher TSP-4 levels are at increased risk of restenosis.
Of the three independent predictors, TSP-4 is a novel serum biomarker that we focused on in this study. TSP-4 is an ECM protein composed of five 140 kDa polypeptide chains and is involved in regulating vascular inflammation and tissue remodeling[34-36]. Notably, inflammatory response is a core pathological feature driving restenosis after endovascular intervention, operating throughout all key processes including inflammatory cell recruitment, vascular smooth muscle cell proliferation, and neointimal hyperplasia[37]. Based on its inflammatory regulatory function, we hypothesized that TSP-4 may be involved in restenosis, yet clinical evidence was previously lacking. In this study, TSP-4 was identified as an independent risk factor for restenosis among numerous inflammatory and metabolic indicators, highlighting that in diabetic patients it may reflect the risk characteristics associated with restenosis more effectively than traditional indicators. Spearman correlation analysis further revealed that TSP-4 was significantly correlated with FPG, HbA1c, LDL-C, hs-CRP, D-D, smoking, age, and Fontaine stage. Previous studies also show that TSP-4 levels correlate positively with Fontaine stage and rise significantly in diabetic metabolic disorders[38]. These findings show that TSP-4 may serve as a key molecule linking metabolic disorders, inflammation and vascular remodeling. Compared with circulating inflammatory markers such as hs-CRP, TSP-4, as an ECM protein, is closer to the pathological process of local inflammation at the vascular wall. This might explain why TSP-4 still maintains an independent correlation even after multiple factor adjustments. In addition, age and LDL-C were also identified as independent predictors, consistent with previous studies[31,39], suggesting that elderly patients and those with dyslipidemia require more intensive postoperative management.
The mechanism by which TSP-4 predicts restenosis may involve its biological functions and related molecular pathways. In the diabetic hyperglycemic state, the TGF-β/SMAD family member 3 signaling pathway is activated, upregulating TSP-4 expression[40-42]. Elevated TSP-4 may then drive restenosis through multiple routes. In inflammation, TSP-4 activates the p38 mitogen-activated protein kinase signaling pathway in endothelial cells via integrin α2 and α2δ-1 receptors, upregulating adhesion molecules such as vascular cell adhesion molecule-1 and intercellular adhesion molecule-1, which recruit inflammatory cells to the site of vascular injury[43,44]. TSP-4 also binds to integrins αMβ2 and αvβ3 on macrophages, inducing p38 mitogen-activated protein kinase phosphorylation, which drives the transcription and synthesis of proinflammatory cytokines, enhances macrophage adhesion and migration, and contributes to neointimal formation[15]. During vascular remodeling, TSP-4 may regulate the proliferation of vascular smooth muscle cells through the phosphatidylinositol 3-kinase/protein kinase B pathway, providing new cells to form the neointima[45]. These pathways collectively suggest that TSP-4 might be a predictor of restenosis and provide a potential target for clinical treatment.
Logistic regression analysis was used to build a nomogram model including age, LDL-C, and TSP-4, which has a high predictive accuracy. The AUC values of the training set and validation set was 0.839 and 0.826, respectively. The calibration curves showed a high consistency between predicted and actual restenosis risks, and DCA showed considerable clinical net benefit within a wide range of threshold probability intervals, supporting the clinical applicability of the model. Based on these findings, the model can help physicians assess individual restenosis risks before treatment, and help formulate more individualized treatment plans and follow-up strategies for T2DM patients with LEASO. Among the predictive factors of this model, age and LDL-C, as traditional risk indicators, have been repeatedly verified in previous models[31,39]. The constructed nomogram in this study has several unique advantages. Previous models mostly use traditional indicators like chronic comorbidities, blood glucose, and lipids[46,47], our model includes TSP-4, which deepens mechanistic understanding and provides a basis for targeted intervention. In addition, this nomogram only included three indicators, which was more concise than most existing models while maintaining stable predictive efficacy and clinical net benefit[31]. This convenient characteristic has important practical value in daily clinical diagnosis and treatment. Clinicians tend to use easily accessible predictive indicators to quickly carry out risk stratification. Moreover, most existing models are constructed based on the general population[48], while our study focused on T2DM patients with LEASO, providing more targeted clinical references for this specific group.
But there a few limitations to consider. First, the model was developed and validated in a single center study with a relatively small cohort, with no external validation to assess its generalizability. Second, our loss rate was 10.1%, which is within the range of similar studies[49,50]. The reasons for dropout mainly included patients refusing to answer calls, difficulties in outpatient follow-up, and actively withdrawing from the study. Although the impact of this loss rate on the research conclusion is extremely limited, it still cannot completely rule out the possibility of bias. Future studies could develop more flexible follow-up plans to improve the completeness of outcome data. Third, this study only measured the preoperative serum TSP-4 levels and did not monitor its dynamic changes after the operation. Future studies should adopt a multicenter, prospective research design with an expanded patient sample size, extended follow-up duration, and continuous serum sample collection at multiple time points. Such studies would more clearly elucidate the dynamic function of TSP-4 and validate the universality of the conclusions of this study in different populations.
CONCLUSION
Preoperative serum TSP-4 levels have high predictive value for restenosis after endovascular treatment in T2DM patients with LEASO. This research result not only provides a new perspective on the pathological mechanism of restenosis in diabetic patients, but also identifies a potential new therapeutic target for clinical practice. Moreover, the nomogram constructed based on age, LDL-C, and TSP-4 demonstrated good discriminative power, precise calibration, and high clinical practical value in the internal validation set. Clinicians can use this model to identify patients at high risk of restenosis early, enabling individualized risk assessment and the development of precise treatment strategies.
ACKNOWLEDGEMENTS
We are grateful to all participants, especially the doctors and nurses in the Department of Vascular Surgery at our institution, who provided support for the recruitment of subjects and sample collection.
Fitridge R, Chuter V, Mills J, Hinchliffe R, Azuma N, Behrendt CA, Boyko EJ, Conte MS, Humphries M, Kirksey L, McGinigle KC, Nikol S, Nordanstig J, Rowe V, Russell D, van den Berg JC, Venermo M, Schaper N. The intersocietal IWGDF, ESVS, SVS guidelines on peripheral artery disease in people with diabetes mellitus and a foot ulcer.J Vasc Surg. 2023;78:1101-1131.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 59][Cited by in RCA: 51][Article Influence: 17.0][Reference Citation Analysis (3)]
Rodriguez-carrio J, Bangueses R, Rodriguez I, Pevida M, Llames S, Meana A, Lozano I, Suarez A. Distinct profiles of immune cell populations underlie in-stent restenosis: a cluster analysis approach.Eur Heart J. 2021;42:ehab724.3398.
[PubMed] [DOI] [Full Text]
Hu L, Lin Y, Zheng J, Wan L, Zhao R, Ma Y, Li J. Transcriptome sequencing revealed that lymph node metastasis of papillary thyroid microcarcinoma is associated with high THBS4 expression and PDGFRA+ cancer-associated fibroblasts.Front Oncol. 2025;15:1536063.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 2][Reference Citation Analysis (0)]
Zhang Q, Zhang Z, Xiu Y, Zou T, Quan Y. Sodium ferulate attenuates ischaemic stroke by mediating the upregulation of thrombospondin-4 expression and combined treatment with bone marrow mesenchymal stem cells.Exp Neurol. 2025;385:115124.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 4][Reference Citation Analysis (0)]
Mazzolai L, Teixido-Tura G, Lanzi S, Boc V, Bossone E, Brodmann M, Bura-Rivière A, De Backer J, Deglise S, Della Corte A, Heiss C, Kałużna-Oleksy M, Kurpas D, McEniery CM, Mirault T, Pasquet AA, Pitcher A, Schaubroeck HAI, Schlager O, Sirnes PA, Sprynger MG, Stabile E, Steinbach F, Thielmann M, van Kimmenade RRJ, Venermo M, Rodriguez-Palomares JF; ESC Scientific Document Group. 2024 ESC Guidelines for the management of peripheral arterial and aortic diseases.Eur Heart J. 2024;45:3538-3700.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 532][Cited by in RCA: 742][Article Influence: 371.0][Reference Citation Analysis (3)]
Genitsaridi I, Salpea P, Salim A, Sajjadi SF, Tomic D, James S, Thirunavukkarasu S, Issaka A, Chen L, Basit A, Luk AOY, Ma RCW, Mbanya JC, Ramachandran A, Wild SH, Duncan BB, Boyko EJ, Magliano DJ. 11th edition of the IDF Diabetes Atlas: global, regional, and national diabetes prevalence estimates for 2024 and projections for 2050.Lancet Diabetes Endocrinol. 2026;14:149-156.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 467][Cited by in RCA: 268][Article Influence: 268.0][Reference Citation Analysis (5)]
Writing Committee Members; Gornik HL, Aronow HD, Goodney PP, Arya S, Brewster LP, Byrd L, Chandra V, Drachman DE, Eaves JM, Ehrman JK, Evans JN, Getchius TSD, Gutiérrez JA, Hawkins BM, Hess CN, Ho KJ, Jones WS, Kim ESH, Kinlay S, Kirksey L, Kohlman-Trigoboff D, Long CA, Pollak AW, Sabri SS, Sadwin LB, Secemsky EA, Serhal M, Shishehbor MH, Treat-Jacobson D, Wilkins LR. 2024 ACC/AHA/AACVPR/APMA/ABC/SCAI/SVM/SVN/SVS/SIR/VESS Guideline for the Management of Lower Extremity Peripheral Artery Disease: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines.J Am Coll Cardiol. 2024;83:2497-2604.
[RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)][Cited by in Crossref: 151][Cited by in RCA: 212][Article Influence: 106.0][Reference Citation Analysis (0)]
