Location and risk factors of postoperative deep venous thrombosis in patients with tibial plateau fracture: a retrospective single-center study Authors: Zhuo-dong Tang, Ming-you Wang, Hong-ping Wang, Xun-zhou Song, Shao-jiang Liu, Yu-ping Lan, Ting-yan Li Published: 2026 ABSTRACT BackgroundDeep vein thrombosis (DVT) is a common postoperative complication in patients with tibial plateau fractures. The aim of this study is to determine the incidence and risk factors of postoperative DVT in patients with tibial plateau fractures.MethodsWe conducted a retrospective study from January 1, 2023, to December 31, 2025. The primary endpoints of this study were the prone site and risk factors of DVT. Univariate analysis and multivariate Logistic regression analysis were used to analyze the related factors, and the receiver operating characteristic (ROC) curve was drawn to evaluate its predictive value.ResultsAmong 890 patients, 207 (23.25%) had postoperative DVT, including 186 cases of distal thrombosis, 4 cases of proximal thrombosis, and 17 cases of mixed thrombosis, all detected DVT events were asymptomatic in the results section.Univariate analysis showed that there were significant differences in age, injury factors, fracture type, preoperative waiting time, operation time, incision length, intraoperative blood loss, preoperative hemoglobin (Hb), preoperative D-dimer, postoperative Hb, postoperative hematocrit (HCT), postoperative fibrinogen (FIB) and postoperative D-dimer levels between the two groups (P < 0.05). Multivariate logistic regression analysis showed that preoperative and postoperative D-dimer levels were independent predictive markers of postoperative DVT in patients with tibial plateau fractures. The area under the curve (AUC) of preoperative D-dimer was 0.928(95%CI = 0.910 ~ 0.946, cut-off value: 2397.0 ng/mL), and the AUC of postoperative D-dimer was 0.922(95%CI = 0.904 ~ 0.940, cut-off value: 2009.5 ng/mL), and the combined prediction efficiency of the two was higher (AUC = 0.960,95%CI = 0.948 ~ 0.971).ConclusionPostoperative DVT in patients with tibial plateau fracture is mostly distal thrombosis. D-dimer level in the 24 h before surgery and D-dimer level in the 24 h after surgery are independent risk factors for postoperative DVT. Tibial plateau fracture is an intra-articular injury, and patients are prone to deep vein thrombosis (DVT) due to extensive soft tissue injury, vascular injury, prolonged limb immobilization, weakened muscle pump and other factors [1–3]. However, DVT formed after lower limb fracture is mostly distal, local and asymptomatic thrombosis, which is easy to be missed, which will delay the intervention time of DVT, increase the risk of pulmonary embolism (PE) [4], and may be life-threatening in severe cases. Therefore, it is important to have the ability to identify patients at high risk of DVT early. Color Doppler ultrasound is a safe, non-invasive, low-cost and repeatable method for the diagnosis of lower extremity DVT. However, ultrasound examination has a lag effect and cannot predict the formation of DVT. “Guidelines for the prevention of venous thromboembolism in orthopedic Surgery in China” recommends the use of Caprini score to predict the risk of DVT in surgical patients [5], but Caprini score lacks the evaluation of laboratory indicators. To improve the predictive ability of postoperative DVT risk in patients with tibial plateau fractures, it is necessary to pay attention to the study of the occurrence regularity and risk factors of DVT. However, the author consulted the previous literature and found that there were few studies on the risk factors of DVT in patients with tibial plateau fractures. The current study attempts to identify and analyze the site and risk factors of postoperative deep vein thrombosis in patients with tibial plateau fracture. The key innovation of this study included the evaluation of laboratory-related indicators, and the comparison of AUC was used to verify their predictive efficacy. This study retrospectively analyzed the relevant data of patients with tibial plateau fractures, and explored the risk factors of DVT, so as to provide theoretical basis for the clinical diagnosis and treatment of these diseases. Methods The clinical data of 890 patients with tibial plateau fractures admitted to Panzhihua Central Hospital from January 1, 2023 to December 31, 2025 were retrospectively analyzed. There were 572 males and 318 females with an average age of 51.46 ± 6.50 years. Two sample size estimation methods, 10 events per variable(10EPV) and 10 events per parameter(10EPP), were used to estimate the sample size, which suggested that the sample size was sufficient. Inclusion criteria: (1) tibial plateau fracture was confirmed by clinical and imaging examination; (2) age > 18 years old; (3) unilateral closed fracture; (4) open reduction and internal fixation were performed. Exclusion criteria: (1) combined with other fractures or injuries; (2) open or pathological fractures; (3) patients with deep venous thrombosis or congenital diseases before admission; (4) coagulopathy, long-term use of anticoagulant drugs, malignant tumor, severe liver disease or renal dysfunction. The patients were divided into DVT group (207 cases) and non-DVT group (683 cases) according to the results of color Doppler ultrasound of deep veins of lower extremities. This study was reviewed by the Medical Ethics Committee of Panzhihua Central Hospital (Pankelun Review No. [2024–067]). All patients underwent color Doppler ultrasonography of the deep veins of the lower extremities 24 h before surgery, 24 h after surgery, and before discharge.The diagnostic criteria of thrombus by color Doppler ultrasound: the venous lumen at the distal end of the thrombus occlusion site was enlarged, and the venous lumen could not be closed or only partially compressed when the probe was pressed, and the proximal end of the thrombus occlusion site showed no blood flow signal or only a small amount of blood flow signal. The sonographers were unaware of the study-group assignments. The location of thrombus was defined as proximal to popliteal vein, including popliteal vein, superficial femoral vein, femoral vein and iliac vein. Vein thrombosis was distal, including: intermuscular vein, anterior tibial vein, posterior tibial vein, peroneal vein; Mixed thrombus was defined as both proximal and distal thrombus. All patients received oral rivaroxaban (10 mg, once a day) to prevent DVT after admission, and mechanical prophylaxis (plantar venous pump, 30 min each time, twice a day) was added for patients with negative results of lower extremity vascular ultrasound at admission. All patients stopped using anticoagulants 24 h before surgery, and resumed using anticoagulants 24 h after surgery until one month after discharge. If any vascular color Doppler ultrasound results were positive, the mechanical prevention measures were immediately stopped, and the drug prevention strategy was adjusted to oral rivaroxaban (20 mg, once a day). All patients in this study were treated with plaster immobilization of the knee joint before operation, and the injured limb was strictly forbidden to bear weight. Ankle pump training and isometric contraction training of quadriceps femoris muscle could be performed. All patients underwent open reduction and internal fixation of tibial fractures. All operations were performed by the same group of doctors. General anesthesia or combined spinal-epidural anesthesia was used in all operations. For the fixation of the lateral column of the tibia, the anterolateral incision was used to reveal and explore the collapse of the articular surface. The collapsed fracture fragment was lifted up with a top rod, and temporary fixation was performed with Kirschner wire. For fixation of the medial and posterior tibial columns, the fracture site was exposed through an inverted L-shaped incision from the popliteal fossa area to the posteromedial edge of the distal tibia. Under direct vision, the fracture end was cleaned, the fracture block was reduced, Kirschner wire was temporarily fixed, autologous iliac bone graft was used to support the bone defect, and the reduction was confirmed under fluoroscopy. The posteromedial and posterolateral columns were fixed with 3.0 mm reconstruction plate support, respectively. One drainage tube was placed after repeated irrigation with normal saline. All patients began to perform knee flexion and extension functional exercise 24 h after the end of the operation, ankle pump training and quadriceps femoris isometric contraction training could be performed, and the injured limb was strictly forbidden to bear weight. The general data of the two groups of patients were collected through the hospital electronic medical record management system: age, gender, body mass index (BMI), injured side, injury factors(tumble, traffic accident, and fall injury), fracture type(Schatzker classification: II ~ VI), history of cardiovascular and cerebrovascular diseases, history of diabetes, smoking history, preoperative waiting time, anesthesia method(general anesthesia and combined spinal epidural anesthesia), operation time, incision length, intraoperative blood loss, postoperative drainage volume. Hemoglobin (Hb), hematocrit (HCT), platelet (PLT), D-dimer, fibrinogen (FIB), prothrombin time (PT), and activated partial thromboplastin time (APTT) were measured within 24 h before and after operation. A flowchart of the study is shown in Fig. 1.Fig. 1The flow chart of the study The flow chart of the study IBM SPSS 27.0 was used for statistical analysis. Kolmogorov–Smirnov test was used to determine whether the measurement data were in line with normal distribution. Measurement data subject to normal distribution were expressed as ± s. Independent sample t test was used for comparison between groups. Mann–Whitney U rank sum test was used for comparison between groups. Count data were expressed as rate (%) and the 2 test or Fisher’s exact test was used. Logistic regression analysis was used to determine the risk factors of postoperative DVT. ROC curve was drawn to evaluate the predictive value. Results No preoperative DVT was detected in any of the patients. A total of 207 patients developed DVT after operation. There were 186 cases(89.86%) of distal thrombosis,4 cases(1.93%) of proximal thrombosis, and 17 cases(8.21%) of mixed thrombosis. All patients had asymptomatic thrombosis, they had no leg swelling or pain (excluding postoperative pain if there is pain), superficial varicose veins, and the circumference of the side leg was less than 3 cm higher than that of the healthy side, indicating an asymptomatic lower extremity venous thrombosis. No patient had PE or bleeding complications during anticoagulation. There was no change in the results of color Doppler ultrasonography of the deep veins of the lower extremities before discharge compared with the results of color Doppler ultrasonography of the deep veins of the lower extremities after surgery. Univariate analysis showed that there were significant differences in age, injury factors, fracture type, preoperative waiting time, operation time, incision length, intraoperative blood loss, preoperative Hb, preoperative D-dimer, postoperative Hb, postoperative HCT, postoperative FIB and postoperative D-dimer between the two groups (P < 0.05) (Table 1).Table 1Results of univariate analysis of DVT formation after tibial plateau fracture surgeryVariableDVT setNon-DVT sett/\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi$$\end{document} 2/ZPAge(years)54.60 ± 10.0650.50 ± 11.454.6340.000Gender(male/female)129/78443/2400.4470.504BMI(kg/m2)24.12 ± 2.8724.57 ± 3.40−1.6980.090Side(left/right)127/80412/2710.0710.790Injury mechanism Tumble6525418.2600.000 Traffic accident94203 Fall injury48226Fracture type II ~ IV1084266.8840.009 V ~ VI99257Cardiovascular and cerebrovascular diseases(yes/no)32/175100/5830.0840.772Diabetes(yes/no)30/17779/6041.2660.261Smoking history(yes/no)81/126286/3970.4930.482Preoperative waiting time(days)7.37 ± 3.156.70 ± 2.852.8920.004Methods of anesthesia General Anesthesia1083841.0530.305 Combined Spinal Epidural Anesthesia99299Operation time (minutes)144.75 ± 88.90128.75 ± 67.452.7630.006Incision length(cm)16.41 ± 7.0914.69 ± 6.353.3170.001Intraoperative blood loss(mL)301.59 ± 271.45226.47 ± 222.994.0270.000Postoperative drainage volume(mL)118.24 ± 91.41116.55 ± 96.360.2240.823Preoperative Hb(g/L)129.89 ± 20.23133.39 ± 19.59−2.2360.026Preoperative HCT(%)39.63 ± 5.9840.42 ± 5.65−1.7400.082Preoperative PLT(109 /L)204.14 ± 49.99207.92 ± 58.33−0.8450.399Preoperative FIB(g/L)3.38 ± 0.913.38 ± 0.92−0.0460.963Preoperative PT(s)11.49 ± 1.1011.60 ± 1.10−1.2750.203Preoperative APTT(s)29.88 ± 3.0229.75 ± 3.060.5500.582Preoperative D-dimer(ng/mL)4240.49 ± 1553.871666.37 ± 1402.6022.5440.000Postoperative Hb(g/L)106.51 ± 16.89111.78 ± 20.84−3.3190.001Postoperative HCT(%)32.50 ± 4.9433.47 ± 5.81−2.1780.030Postoperative PLT(109 /L)183.79 ± 49.98188.61 ± 58.45−1.0750.283Postoperative FIB(g/L)3.93 ± 1.324.16 ± 1.34−2.1360.033Postoperative PT(s)14.10 ± 1.4113.99 ± 1.420.9230.356Postoperative APTT(s)37.16 ± 3.1436.98 ± 3.060.7130.476Postoperative D-dimer(ng/mL)3659.40 ± 2414.531408.64 ± 812.7820.8010.000Length of stay(day)13.74 ± 5.4013.33 ± 5.05−1.0070.314 Table 1Results of univariate analysis of DVT formation after tibial plateau fracture surgeryVariableDVT setNon-DVT sett/\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi$$\end{document} 2/ZPAge(years)54.60 ± 10.0650.50 ± 11.454.6340.000Gender(male/female)129/78443/2400.4470.504BMI(kg/m2)24.12 ± 2.8724.57 ± 3.40−1.6980.090Side(left/right)127/80412/2710.0710.790Injury mechanism Tumble6525418.2600.000 Traffic accident94203 Fall injury48226Fracture type II ~ IV1084266.8840.009 V ~ VI99257Cardiovascular and cerebrovascular diseases(yes/no)32/175100/5830.0840.772Diabetes(yes/no)30/17779/6041.2660.261Smoking history(yes/no)81/126286/3970.4930.482Preoperative waiting time(days)7.37 ± 3.156.70 ± 2.852.8920.004Methods of anesthesia General Anesthesia1083841.0530.305 Combined Spinal Epidural Anesthesia99299Operation time (minutes)144.75 ± 88.90128.75 ± 67.452.7630.006Incision length(cm)16.41 ± 7.0914.69 ± 6.353.3170.001Intraoperative blood loss(mL)301.59 ± 271.45226.47 ± 222.994.0270.000Postoperative drainage volume(mL)118.24 ± 91.41116.55 ± 96.360.2240.823Preoperative Hb(g/L)129.89 ± 20.23133.39 ± 19.59−2.2360.026Preoperative HCT(%)39.63 ± 5.9840.42 ± 5.65−1.7400.082Preoperative PLT(109 /L)204.14 ± 49.99207.92 ± 58.33−0.8450.399Preoperative FIB(g/L)3.38 ± 0.913.38 ± 0.92−0.0460.963Preoperative PT(s)11.49 ± 1.1011.60 ± 1.10−1.2750.203Preoperative APTT(s)29.88 ± 3.0229.75 ± 3.060.5500.582Preoperative D-dimer(ng/mL)4240.49 ± 1553.871666.37 ± 1402.6022.5440.000Postoperative Hb(g/L)106.51 ± 16.89111.78 ± 20.84−3.3190.001Postoperative HCT(%)32.50 ± 4.9433.47 ± 5.81−2.1780.030Postoperative PLT(109 /L)183.79 ± 49.98188.61 ± 58.45−1.0750.283Postoperative FIB(g/L)3.93 ± 1.324.16 ± 1.34−2.1360.033Postoperative PT(s)14.10 ± 1.4113.99 ± 1.420.9230.356Postoperative APTT(s)37.16 ± 3.1436.98 ± 3.060.7130.476Postoperative D-dimer(ng/mL)3659.40 ± 2414.531408.64 ± 812.7820.8010.000Length of stay(day)13.74 ± 5.4013.33 ± 5.05−1.0070.314 Results of univariate analysis of DVT formation after tibial plateau fracture surgery The dependent variable was DVT, while the independent variables were age, injury factors, fracture type, preoperative waiting time, operation time, incision length, intraoperative blood loss, preoperative Hb, preoperative D-dimer, postoperative Hb, postoperative HCT, postoperative FIB and postoperative D-dimer. For the dependent variable and every independent variable, corresponding assignment processing was carried out (Table 2). Multicollinearity analysis of the differentiating indicators in the univariate analysis revealed that postoperative Hb (VIF = 6.053) and HCT (VIF = 4.598) showed significant multicollinearity and were excluded. There was no multicollinearity relationship between age, injury factors, fracture type, preoperative waiting time, operation time, incision length, intraoperative blood loss, preoperative Hb, preoperative D-dimer, postoperative FIB and postoperative D-dimer (1 < VIF < 5, tolerance ≥ 0.1), suggesting that the above indicators could be included in the logistic regression model analysis (Table 3).Table 2Independent variable assignmentVariableFactorsAssignment methodYDVT0 = DVT; 1 = Non-DVTX1Ageactual measured valueX2Injury factors1 = Tumble; 2 = Traffic accident;3 = Fall injuryX3Fracture type1 = II ~ IV; 2 = V ~ VIX4Preoperative waiting timeactual measured valueX5Operation timeactual measured valueX6Incision lengthactual measured valueX7Intraoperative blood lossactual measured valueX8Preoperative Hbactual measured valueX9Preoperative D-dimeractual measured valueX10Postoperative Hbactual measured valueX11Postoperative HCTactual measured valueX12Postoperative FIBactual measured valueX13Postoperative D-dimeractual measured valueTable 3Collinearity analysisItemVIFToleranceAge1.1390.878Injury factors1.0360.965Fracture type1.0700.935Preoperative waiting time1.1280.887Operation time1.3580.736Incision length1.6730.598Intraoperative blood loss1.5230.657Preoperative Hb2.1100.474Preoperative D-dimer1.5260.655Postoperative Hb6.0530.165Postoperative HCT4.5980.217Postoperative FIB1.2050.830Postoperative D-dimer1.5990.625 Table 2Independent variable assignmentVariableFactorsAssignment methodYDVT0 = DVT; 1 = Non-DVTX1Ageactual measured valueX2Injury factors1 = Tumble; 2 = Traffic accident;3 = Fall injuryX3Fracture type1 = II ~ IV; 2 = V ~ VIX4Preoperative waiting timeactual measured valueX5Operation timeactual measured valueX6Incision lengthactual measured valueX7Intraoperative blood lossactual measured valueX8Preoperative Hbactual measured valueX9Preoperative D-dimeractual measured valueX10Postoperative Hbactual measured valueX11Postoperative HCTactual measured valueX12Postoperative FIBactual measured valueX13Postoperative D-dimeractual measured value Independent variable assignment Table 3Collinearity analysisItemVIFToleranceAge1.1390.878Injury factors1.0360.965Fracture type1.0700.935Preoperative waiting time1.1280.887Operation time1.3580.736Incision length1.6730.598Intraoperative blood loss1.5230.657Preoperative Hb2.1100.474Preoperative D-dimer1.5260.655Postoperative Hb6.0530.165Postoperative HCT4.5980.217Postoperative FIB1.2050.830Postoperative D-dimer1.5990.625 Collinearity analysis Multivariate Logistic regression analysis showed that preoperative D-dimer and postoperative D-dimer were risk factors for DVT formation after tibial plateau fracture surgery (Table 4). Receiver operating characteristic curves were drawn to explore the predictive power and cut-off value of preoperative and postoperative D-dimer for postoperative DVT formation (Fig. 2). The area under the curve (AUC) of preoperative D-dimer for predicting DVT after tibial plateau fracture surgery was 0.928 (cut-off value: 2397.0 ng/mL, specificity: 80.9%, sensitivity: 95.6%), and the area under the curve (AUC) of postoperative D-dimer for predicting DVT formation after tibial plateau fracture surgery was 0.922 (cut-off value: 2009.5 ng/mL; specificity: 78.6%; sensitivity: 95.6%) (Table 5). The Delong test showed that the combined predictive value of the preoperative D-dimer and postoperative D-dimer was higher than the single predictive value (Table 6).Table 4Results of multivariate analysis of DVT formation after tibial plateau fracture surgeryVariableβSEWaldχ2POR95%CIAge0.0070.0120.3040.5811.0070.983 ~ 1.031Injury mechanism−0.1240.1790.4770.4900.8830.622 ~ 1.256Fracture type−0.0330.0800.1640.6860.9680.827 ~ 1.133Preoperative waiting time−0.0380.0430.7760.3780.9620.884 ~ 1.048Operation time0.0000.0020.0350.8531.0000.995 ~ 1.004Incision length0.0350.0251.8680.1721.0350.985 ~ 1.088Intraoperative blood loss0.0000.0010.1590.6901.0000.999 ~ 1.001Preoperative Hb0.0130.0092.1120.1461.0130.995 ~ 1.031Preoperative D-dimer0.0010.00079.5480.0001.0011.001 ~ 1.001Postoperative FIB−0.2040.1073.6730.0550.8150.661 ~ 1.005Postoperative D-dimer0.0020.00089.7290.0001.0021.001 ~ 1.002Fig. 2ROC curve of preoperative D-dimer and postoperative D-dimer in predicting postoperative DVT formation in patients with tibial plateau fractureTable 5Predictive valueVariableAUC95% CICut off valuespecificitysensitivityCollaborative forecasting0.9600.948 ~ 0.971-0.9070.942Preoperative D-dimer0.9280.910 ~ 0.9462397.0 ng/mL0.8090.956Postoperative D-dimer0.9220.904 ~ 0.9402009.5 ng/mL0.7860.956Table 6DeLong testVariableP95% CICollaborative forecasting VS Preoperative D-dimer0.0000.023 ~ 0.052Collaborative forecasting VS Postoperative D-dimer0.0000.018 ~ 0.044Preoperative D-dimer VS Postoperative D-dimer1.832−0.017 ~ 0.030 Table 4Results of multivariate analysis of DVT formation after tibial plateau fracture surgeryVariableβSEWaldχ2POR95%CIAge0.0070.0120.3040.5811.0070.983 ~ 1.031Injury mechanism−0.1240.1790.4770.4900.8830.622 ~ 1.256Fracture type−0.0330.0800.1640.6860.9680.827 ~ 1.133Preoperative waiting time−0.0380.0430.7760.3780.9620.884 ~ 1.048Operation time0.0000.0020.0350.8531.0000.995 ~ 1.004Incision length0.0350.0251.8680.1721.0350.985 ~ 1.088Intraoperative blood loss0.0000.0010.1590.6901.0000.999 ~ 1.001Preoperative Hb0.0130.0092.1120.1461.0130.995 ~ 1.031Preoperative D-dimer0.0010.00079.5480.0001.0011.001 ~ 1.001Postoperative FIB−0.2040.1073.6730.0550.8150.661 ~ 1.005Postoperative D-dimer0.0020.00089.7290.0001.0021.001 ~ 1.002 Results of multivariate analysis of DVT formation after tibial plateau fracture surgery ROC curve of preoperative D-dimer and postoperative D-dimer in predicting postoperative DVT formation in patients with tibial plateau fracture Table 5Predictive valueVariableAUC95% CICut off valuespecificitysensitivityCollaborative forecasting0.9600.948 ~ 0.971-0.9070.942Preoperative D-dimer0.9280.910 ~ 0.9462397.0 ng/mL0.8090.956Postoperative D-dimer0.9220.904 ~ 0.9402009.5 ng/mL0.7860.956 Predictive value Table 6DeLong testVariableP95% CICollaborative forecasting VS Preoperative D-dimer0.0000.023 ~ 0.052Collaborative forecasting VS Postoperative D-dimer0.0000.018 ~ 0.044Preoperative D-dimer VS Postoperative D-dimer1.832−0.017 ~ 0.030 DeLong test Discussion In this study, 890 patients with tibial plateau fracture were retrospectively analyzed. The incidence of postoperative deep vein thrombosis (DVT) was 23.25% (207/890), and it was confirmed that the most common site of DVT was distal vein, accounting for 89.86% (186/207), and proximal thrombosis and mixed thrombosis were relatively rare. This distribution feature is consistent with the results of previous studies on tibial plateau fractures [6, 7]. In this study, through univariate and multivariate Logistic regression analysis, it was finally determined that preoperative D-dimer and postoperative D-dimer levels were independent risk factors for postoperative DVT. The area under the curve (AUC) of the combined prediction of the two was 0.960, showing a good prediction trend. This result provides a strong evidence-based basis for early identification of patients at high risk of DVT and dynamic adjustment of prevention strategies. However, in this study, some factors traditionally considered to be closely related to thrombosis, such as age, operation time, fracture type and injury factors, showed significant differences in univariate analysis, but lost statistical significance after entering the multivariate regression model. This statistical phenomenon is not an isolated case, but has been confirmed in different degrees in a number of similar studies. Aging is accompanied by venous wall degeneration, endothelial dysfunction and decreased activity, which constitute the pathophysiological basis of thrombosis. In the univariate analysis of this study, the average age of patients in the DVT group was significantly higher than that in the non-DVT group, which was consistent with Virchow theory. However, in this study, although age is an important factor affecting the occurrence of postoperative DVT, it is not an independent risk factor. The reason is that the incidence of DVT increases with the increase of age, but the correlation is low. In the multivariate model of this study, the OR value of age was 1.007 (95%CI 0.983–1.031, P = 0.581), and the effect was extremely weak and not significant. Rong [8] also found significant differences in age in univariate analysis in the study of 325 elderly patients with hip fracture over 70 years old, but age did not become an independent risk factor in multivariate regression. From the perspective of statistics and pathophysiology, the effect of age on DVT is not an independent and direct causal pathway, but indirectly realized through a series of intermediate variables. With increasing age, patients have decreased surgical tolerance and worse vascular conditions, and these intermediate pathological states will be directly reflected in variables such as higher D-dimer levels, longer operation time, and more intraoperative blood loss. When D-dimer, a sensitive laboratory indicator that directly reflects the state of hypercoagulation and fibrinolysis in vivo, was included in the regression model, the macroscopic risk effect of age was largely explained and replaced by these refined mediating variables. Of course, we also note that age has been confirmed as an independent risk factor for DVT in some large sample studies. For example, Li [9] found that age was an independent risk factor for postoperative DVT in their study of 987 patients with tibial plateau fractures, and Yang [10] also showed that age > 60 years old was an independent risk factor for postoperative DVT in their multivariate analysis of 497 patients with periknee fractures. This suggests that the independent contribution of age may fluctuate somewhat depending on fracture site, age composition of the sample, and underlying disease spectrum of the study population. In the particular injury type of tibial plateau fracture, laboratory indicators such as D-dimer seem to provide a more direct and accurate basis for risk stratification than age itself. Operative time was also significant in univariate analysis (P = 0.006), but was excluded in multivariate analysis (P = 0.853). This result is consistent with the findings of some studies. In the analysis of 186 young and middle-aged patients with hip fracture by Qiao [11], there was no significant difference in operation time in univariate analysis. Yang [10] also found no significant effect of operation time in their study of periarticular knee fractures. However, other studies have reached different conclusions. Peng [12] ‘s analysis of 615 patients with tibial plateau fracture showed that operation time > 114 min was not only a significant variable in univariate analysis, but also confirmed as an independent risk factor in multivariate regression. Li [13] also had similar results in their study. The author believes that the factors that really affect the formation of DVT are those that lead to prolonged operation time, such as the degree of fracture comminuted, the severity of soft tissue injury, and the difficulty of intraoperative reduction. When variables that can more directly reflect surgical trauma and coagulation stress such as intraoperative blood loss, fracture type, and D-dimer were included in the multivariate model, the independent effect of operation time was explained by these more biologically direct indicators, which were statistically reasonable. This also suggests that rather than focusing on the operative time itself, more attention should be paid to quantifiable indicators such as the dynamic changes of D-dimer during the perioperative period and intraoperative blood loss. In this study, fracture type (Schatzker classification) was significantly different in univariate analysis (P = 0.045), but was excluded in multivariate analysis (P = 0.686). This phenomenon is different in the study of tibial plateau fracture by Bai [6], which found that Schatzker classification was an independent risk factor for preoperative DVT (OR = 1.829, P = 0.045), but it should be noted that the P value was very close to the significance boundary and the effect size was relatively small. There may be two reasons why the effect of fracture type did not reach independent significance in this study: Firstly, the Schatzker classification of tibial plateau fractures may be naturally weaker in predicting DVT than the classification of hip fractures. In hip fractures, the vascular anatomy difference between intra-articular and external capsular fractures is extremely significant, while the impact of the injury range and surgical approach on the popliteal vascular region is relatively limited regardless of the classification of tibial plateau fractures. Second, this study showed that patients with high Schatzker classification (types IV to VI) had longer operation time, more intraoperative blood loss, and longer preoperative waiting time, and the risk effect of fracture type was largely transmitted and expressed through these “downstream” variables. The same was true for injury factors—road traffic injuries and falls from height (high-energy injuries) were associated with a higher incidence of DVT in univariate analyses, but multivariate analyses did not show independent effects. Li [14] also found that there was no significant difference in the incidence of preoperative and postoperative DVT between the high-energy group and the low-energy group, but the high-energy group had longer hospital stay, more intraoperative blood loss, and higher preoperative D-dimer levels, which were similar to the results of this study. The results of this study showed that preoperative and postoperative D-dimer levels were independent risk factors for DVT after tibial plateau fracture surgery, and DeLong test confirmed that the combined predictive efficacy of the two indicators was significantly better than either single indicator. D-dimer, as the specific end product of cross-linked fibrin decomposition under the action of plasmin, the increase of its plasma concentration directly reflects the dynamic process of coagulation system activation, fibrin formation and secondary hyperfibrinolysis in the body, which has been widely confirmed as a sensitive screening index for thrombotic diseases [15, 16]. An important innovation of this study is the combined assessment of preoperative and postoperative D-dimer. Preoperative D-dimer reflects the “baseline thrombosis susceptibility” from trauma to surgery, which is affected by factors such as individual genetic background, comorbidities, and the time from injury to admission. On this basis, postoperative D-dimer added new procoagulant factors such as surgical trauma, anesthetic effect, and postoperative immobilization. The combination of the two dynamic monitoring essentially constructs a “trend surface of perioperative coagulation status” in the time dimension, and its leap in predictive ability is logically self-consistent. In clinical practice, if a patient has a high level of D-dimer before surgery (e.g., > 2397 ng/mL), it suggests that there may be an excessive hypercoagulable state after trauma in the body before surgery. If the D-dimer continues to rise instead of leveling off after operation, it is a strong warning that thrombosis is forming or spreading.However, the limitations of D-dimer as a single marker cannot be ignored. Its specificity is affected by a variety of non-thrombotic factors, such as blood transfusion, tumor, infection, trauma, blood transfusion, and the use of anticoagulant drugs [17]. Through strict implementation of inclusion and exclusion criteria, this study excluded patients with known interfering factors such as malignant tumors, severe liver and kidney insufficiency, and coagulation dysfunction, so as to minimize confounding. However, clinicians should not make decisions based on a single D-dimer test value in isolation, but should integrate it into a comprehensive evaluation framework including clinical symptoms and signs, ultrasound results, and make dynamic and continuous individualized judgments. This study provides a reference for risk prediction and clinical management of DVT after tibial plateau fracture surgery. Firstly, we identified the incidence of postoperative DVT in this population (23.25%) and the distribution of common sites (89.86% of distal thrombosis). This data can provide orthopedic surgeons with more accurate risk quantification information and provide a baseline basis for the development of standardized prevention procedures within the department. However, even with standard drug + mechanical prophylaxis, the 23.25% asymptomatic DVT incidence still indicates high thrombotic risk of tibial plateau fracture patients, and suggest targeted enhanced prevention for patients exceeding D-dimer cut-off value. Secondly, the preoperative D-dimer cut-off value (2397.0 ng/mL) and postoperative D-dimer cut-off value (2009.5 ng/mL) and the corresponding sensitivity and specificity presented in this study have direct bedside application value. For patients with D-dimer more than 2397.0 ng/mL on admission, clinicians can consider taking enhanced measures on the basis of standard prevention, such as shortening the preoperative waiting time, standardized use of tourniquet during surgery, and earlier initiation of ultrasound screening after surgery. If the postoperative D-dimer level is persistently higher than 2009.5 ng/mL or there is no significant decrease compared with that before operation, the possibility of thrombosis should be highly suspected and ultrasound examination should be arranged in time. It is worth emphasizing that this study reaffirms the concept that dynamic surveillance is superior to a single test. The AUC of preoperative and postoperative D-dimer combined prediction (0.960) was not only higher than that of single index, but also the specificity increased from 78.6%−80.9% of single index to 90.7% of combined prediction. In addition, we confirmed that all patients with confirmed DVT had asymptomatic thrombi under standard medical and physical prophylaxis regimens (rivaroxaban 10 mg once daily plus a plantar venous pump), and no PE or bleeding complications occurred during anticoagulation therapy. This suggests that the current prevention strategy is effective in preventing the progression of thrombosis and fatal complications. Although it can not completely prevent the formation of DVT, its clinical significance mainly lies in avoiding the occurrence of adverse clinical events and reducing the risk of complications. Although the sample size of this study is large and the data collection is detailed, there are several methodological limitations that cannot be avoided, which need to be carefully treated when interpreting and applying the conclusions. First, this study was a single-center retrospective study, and all cases were from Panzhihua Central Hospital, which is located in a specific geographical area and may have unique population composition, disease spectrum and perioperative management process. Single-center retrospective study inevitably has selection bias and lacks an external validation cohort. Future studies can be verified by large sample and multi-center studies. Data from a single center cannot fully represent the actual situation of medical institutions in other regions and levels, and the universality and extrapolation of conclusions are limited. Second, this study did not include an independent external validation cohort. The AUC value, cut-off value, sensitivity, specificity and other indicators obtained by ROC curve analysis are all from the same data set constructed by the model. Even if the standard regression analysis process is adopted, the predictive efficiency under such internal verification may be overestimated to a certain extent. Third, the diagnosis of DVT was based on color Doppler ultrasonography of the lower limbs. Although this method is commonly used and reliable in clinical practice, its sensitivity in the diagnosis of isolated distal thrombosis is still insufficient compared with venography, especially in the conditions of postoperative swelling and dressing coverage. Due to the limitations of different departments, we were unable to provide the ultrasound machine model, probe frequency, and standard scanning segment, which may have an impact on the reproducibility of the study results. Fourth, this study excluded patients with open fractures, multiple injuries, and severe medical diseases, which ensured the homogeneity of the study population, but also made the conclusions not applicable to these more high-risk subgroups. Guo [18] also pointed out that the risk of DVT was significantly increased in patients with a history of malignant tumors (P = 0.047), but these patients were excluded from this study. It is necessary for future research to specifically evaluate and model these high-risk subgroups. Fifth, this study only analyzed the occurrence of DVT during hospitalization, and did not follow up the outcome of thrombosis, the incidence of post-thrombotic syndrome and the long-term functional recovery of patients after discharge. Future studies can extend the follow-up window to 3 to 6 months after surgery to comprehensively evaluate the long-term clinical outcome of DVT. Conclusion The study highlighted the incidence and location of postoperative DVT in patients with tibial plateau fracture, and for patients with tibial plateau fracture, preoperative and postoperative D-dimer are independent risk factors for postoperative DVT formation.