A Two-Step Risk Stratification Strategy Integrating Caprini Score and D-Dimer for Venous Thromboembolism in Trauma Patients

A Two-Step Risk Stratification Strategy Integrating Caprini Score and D-Dimer for Venous Thromboembolism in Trauma Patients | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article A Two-Step Risk Stratification Strategy Integrating Caprini Score and D-Dimer for Venous Thromboembolism in Trauma Patients Qiang Liu, Mengqing Lin, Qian Zhang, Junfei Huang, Bingyu Xu, and 7 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9440559/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 6 You are reading this latest preprint version Abstract Background To evaluate the integration of the Caprini score with thrombosis-related biomarkers for predicting venous thromboembolism (VTE) risk in trauma patients, aiming to establish a precision re-stratification strategy for high-risk populations. Methods A retrospective cohort of 206 trauma surgical patients was analyzed. Caprini scores and baseline plasma biomarkers (DDI, TAT, PIC, TM) were assessed against VTE occurrence using ROC curves. Secondary risk re-stratification was performed within the Caprini high-risk cohort utilizing a Youden-optimized DDI threshold. Associations between trauma severity and coagulation activation were further evaluated using Injury Severity Score-based analyses. Results Overall VTE incidence was 31.1%. The Caprini score demonstrated modest discriminative capacity (AUC = 0.640). Across the entire cohort, DDI exhibited the highest predictive efficacy among individual biomarkers (AUC = 0.669, P < 0.05). However, within the Caprini high-risk subgroup (≥ 9 points, n = 118), the discriminative power of all biomarkers experienced a ubiquitous decline (DDI AUC = 0.614), reflecting trauma-induced biomarker saturation. Integrating DDI with the Caprini score elevated the combined model's AUC to 0.730. By applying an optimized threshold of DDI ≥ 5.5 mg/L, the high-risk cohort was successfully re-stratified, unmasking an "extremely high-risk" subset with a 44.8% VTE incidence. This represents a 2.78-fold risk increase compared to the general high-risk baseline (16.1%, P = 0.005). Conclusions Relying solely on clinical risk scoring provides limited VTE discrimination in trauma patients. Implementing a hierarchical, two-step "Caprini screening—DDI re-stratification" strategy effectively circumvents biomarker saturation, isolating extremely high-risk patients to provide a robust quantitative framework for precision thromboprophylaxis. venous thromboembolism trauma center Caprini score D-dimer risk stratification Figures Figure 1 Figure 2 Figure 3 1 Background Venous thromboembolism (VTE), encompassing deep venous thrombosis (DVT) and pulmonary thromboembolism (PTE), remains a formidable perioperative complication in trauma surgery[ 1 ]. The acute physiological stress of trauma—characterized by severe vascular endothelial disruption, prolonged immobilization, and a profound hypercoagulable state—predisposes these patients to an exceptionally high risk of VTE, which subsequently drives significant morbidity and mortality[ 2 – 4 ]. While the Caprini risk assessment model is universally endorsed for clinical VTE stratification[ 5 ], its reliance on the cumulative scoring of static clinical risk factors limits its diagnostic agility in trauma patients experiencing acute pathophysiological shifts[ 6 – 8 ]. Consequently, the isolated application of the Caprini score frequently funnels a disproportionate volume of trauma patients into a broad "high-risk" category[ 9 ]. This non-specific classification masks substantial inter-individual variability in actual thrombotic risk[ 10 , 11 ], ultimately complicating the formulation and execution of tailored, precision-driven thromboprophylaxis[ 12 ]. To navigate these clinical ambiguities, novel thrombosis-related biomarkers have emerged as highly sensitive instruments for detecting early hypercoagulability[ 13 ]. Specifically, the thrombin-antithrombin complex (TAT) serves as a direct barometer of thrombin generation[ 14 – 16 ], the plasmin-α2-plasmin inhibitor complex (PIC) quantifies the activation of the fibrinolytic system[ 17 – 19 ], and thrombomodulin (TM) acts as a proximal indicator of vascular endothelial disruption[ 20 – 22 ]. Although these novel metrics hold theoretical superiority over traditional D-dimer (DDI) in capturing early kinetic shifts within the coagulation cascade[ 23 – 25 ], their practical utility and optimal diagnostic thresholds for risk re-stratification in trauma cohorts remain poorly defined. To address this critical knowledge gap, we retrospectively analyzed a cohort of 206 trauma center patients to rigorously evaluate the prognostic synergy between the Caprini score and a panel of thrombosis biomarkers. Our primary objective was to investigate the hypothesized "predictive attenuation" (or biomarker saturation) within the Caprini high-risk stratum. By optimizing biomarker cutoff values specifically tailored to this vulnerable population, we aimed to engineer a pragmatic, two-step "Caprini screening—biomarker re-stratification" strategy. Ultimately, this hierarchical model seeks to deliver a robust quantitative framework that empowers clinicians to implement precise and individualized VTE prophylaxis in trauma surgery. 2 Methods 2.1 Study Population and Ethics This retrospective cohort study included 206 trauma surgical patients admitted to Zhujiang Hospital of Southern Medical University between January 2022 and December 2025. The study protocol was conducted in accordance with the Declaration of Helsinki and received formal approval from Medical Ethics Committee of Zhujiang Hospital of Southern Medical University (Approval No.: 2025-KY-329-01). Written informed consent was obtained from all patients or their legal guardians prior to enrollment. 2.2 Inclusion and Exclusion Criteria Patients were eligible for inclusion if they: (1) were aged ≥ 18 years; (2) were admitted for acute trauma and underwent subsequent surgical intervention; (3) completed both the Caprini risk assessment and initial thrombosis-related biomarker testing within 24 hours of admission; and (4) possessed complete clinical and follow-up data. Exclusion criteria were: (1) VTE diagnosed prior to or immediately upon admission based on clinical manifestations (e.g., limb swelling, positive Homans sign) or imaging confirmation; (2) coexisting severe hepatic or renal insufficiency, active malignancies, or pre-existing hematological disorders affecting coagulation; (3) administration of anticoagulant or thrombolytic therapy within 14 days prior to admission; (4) pregnancy or lactation; and (5) failure to complete mandatory lower extremity vascular imaging (color Doppler ultrasound or venography) during hospitalization. 2.3 Clinical Data Acquisition and Variable Selection This study followed a rigorous retrospective protocol for data extraction. Comprehensive clinical profiles and laboratory parameters for all eligible patients were systematically retrieved from the Hospital Information System (HIS) and Laboratory Information System (LIS). The collected dataset encompassed: (1) baseline demographics, including age and gender; (2) injury-specific metrics, primarily the Injury Severity Score (ISS) and admission Caprini score; and (3) definitive VTE diagnostic outcomes derived from postoperative imaging (e.g., color Doppler ultrasound, venography, or CTV). To ensure high data fidelity and minimize extraction bias, all variables were independently audited by two researchers, with any discrepancies resolved through consensus and cross-verification of the original electronic medical records. 2.4 Laboratory Analysis and Biomarker Measurement Peripheral venous blood samples were collected by trained nursing staff upon admission or during the immediate preoperative period. Samples were drawn into standardized vacuum tubes containing trisodium citrate anticoagulant and processed via programmed centrifugation according to established institutional protocols. The plasma concentrations of four key thrombosis-related biomarkers—plasmin-α2-plasmin inhibitor complex (PIC), thrombomodulin (TM), thrombin-antithrombin III complex (TAT), and D-dimer (DDI)—were quantified. All assays were performed by qualified laboratory personnel using unified automated analytical platforms and standardized operating procedures (SOPs) to ensure technical reproducibility. Laboratory results were integrated in real-time into the LIS, from which researchers extracted the initial test values (baseline levels) for subsequent predictive modeling. 2.5 VTE Diagnosis and Risk Grouping VTE diagnosis followed the International Consensus Statement on the Prevention and Management of Venous Thromboembolism[ 26 ], employing a hierarchical diagnostic strategy. Initial clinical probability was determined via the Caprini scale. Patients identified as moderate-to-high risk, or those with elevated D-dimer levels, underwent priority screening with color Doppler ultrasound. In cases of persistent clinical suspicion despite negative initial ultrasound, VTE was further confirmed or excluded via contrast-enhanced venography or Computed Tomography Venography (CTV). For comparative analysis, patients were stratified by admission Caprini scores into three risk tiers: low-risk (0–4 points, n = 34), intermediate-risk (5–8 points, n = 54), and high-risk (≥ 9 points, n = 118). Based on definitive postoperative imaging, the cohort was dichotomized into the VTE group (n = 64) and the non-VTE group (n = 142). 2.6 Statistical Analysis Statistical analyses were performed using SPSS version 27.0. The normality of continuous data was assessed using the Shapiro-Wilk test. Normally distributed variables are presented as mean ± standard deviation ( \(\:\stackrel{-}{\text{x}}\pm\:\text{s}\) ) and compared via independent samples t -tests. Non-normally distributed data are expressed as median (interquartile range) \(\:\left[\text{M}\left({\text{P}}_{25},{\text{P}}_{75}\right)\right]\) and analyzed using the Mann-Whitney U test. Categorical variables are reported as frequencies (%) and compared using the chi-square ( \(\:{{\chi\:}}^{2}\) ) test. The discriminative performance of the Caprini score and individual biomarkers was evaluated using Receiver Operating Characteristic (ROC) curves and the corresponding Area Under the Curve (AUC). Spearman rank correlation and linear-by-linear association tests were used to evaluate relationships between Injury Severity Score (ISS), D-dimer levels, and VTE incidence across clinically relevant ISS strata (< 16, 16–24, ≥ 25). The optimal D-dimer threshold for re-stratification within the high-risk subgroup, as well as the optimal ISS cutoff for risk separation, were determined using the Youden index. Logistic regression models were constructed to assess the incremental predictive value of combined strategies and to explore potential interactions between ISS and D-dimer. To ensure model robustness and mitigate overfitting, internal validation was performed using 1,000 bootstrap resamples. Finally, Decision Curve Analysis (DCA) was utilized to quantify the net clinical benefit of the proposed two-step strategy. All tests were two-tailed, with P < 0.05 denoting statistical significance. 3 Results 3.1 Caprini Risk Stratification and Patient Baseline Characteristics A total of 206 trauma surgical patients were included in this study, with an overall VTE incidence of 31.1%. Based on the Caprini score, the patients were divided into three groups: a low-risk group (0–4 points, n = 34), an intermediate-risk group (5–8 points, n = 54), and a high-risk group (≥ 9 points, n = 118). There were no statistically significant differences in age and gender distribution among the three groups (P > 0.05), suggesting that these variables were unlikely to confound the relationship between risk stratification and VTE occurrence. In contrast, the difference in ISS scores among the groups was significant (P = 0.034), with higher scores observed in the higher Caprini groups. This indicates that Caprini-based risk stratification is consistent with trauma severity to some extent, supporting its clinical relevance in this setting. Although the difference in VTE incidence among the three groups was marginally significant (χ² = 5.007, P = 0.082), the linear trend test confirmed that VTE incidence exhibited a progressive upward trend with increasing risk levels (χ² = 3.870, P = 0.049). Taken together, these findings suggest that while categorical comparisons alone may understate group differences, the Caprini score retains the ability to capture an overall risk gradient, highlighting its value as a baseline risk stratification tool in trauma patients (Table 1 ). Table 1 Comparison of baseline characteristics of trauma patients with different Caprini risk stratifications Characteristics Low-Risk Group (n = 34) Intermediate-Risk Group (n = 54) High-Risk Group (n = 118) Statistic P-value Age(years, \(\:\stackrel{-}{\text{x}}\pm\:\text{s}\) ) 46.1 ± 16.9 45.9 ± 17.2 48.6 ± 18.0 F = 0.572 0.566 Male, n(%) 26(76.5) 45(83.3) 88(74.6) χ²=1.625 0.444 ISS Score, \(\:\text{M}(\text{Q}1,\text{Q}3)\) 17(11,20) 20(14,25) 20(16,27) H = 6.787 0.034 VTE Occurrence, n(%) 8(23.5) 12(22.2) 44(37.3) χ²=5.007 0.082 Notes: Age follows a normal distribution and is expressed as mean ± standard deviation, analyzed using one-way ANOVA (F test); ISS score does not follow a normal distribution and is expressed as median (interquartile range), analyzed using the Kruskal–Wallis H test; categorical variables are expressed as number (percentage) and analyzed using the χ² test. 3.2 Stratification Efficacy of Caprini Score in the Trauma Population Spearman correlation analysis showed a statistically significant but weak positive correlation between the Caprini score and VTE occurrence (r = 0.148, P = 0.034), indicating that higher scores correlate with an increased risk of thrombosis, although the strength of this relationship is limited. ROC curve analysis further revealed that the AUC of the Caprini score for predicting VTE was 0.640 (95% CI: 0.553–0.726, P = 0.001) (Fig. 1 A), suggesting modest predictive performance in this population. At the optimal cutoff value of approximately 8.5 points, the sensitivity was 68.8% but the specificity was relatively low at 47.9%. This imbalance indicates that while the Caprini score is capable of identifying a substantial proportion of patients who will develop VTE, it also misclassifies a considerable number of low-risk individuals as high risk. These findings suggest that relying solely on clinical risk factor scoring involves a high false-positive rate and limited overall predictive discrimination in the trauma population. 3.3 Predictive Value of Thrombosis Biomarkers in the Overall Population Single-indicator ROC curve analysis showed that among the evaluated biomarkers, D-dimer (DDI) demonstrated the highest discriminative ability for VTE, with an AUC of 0.669 (95% CI: 0.592–0.747, P < 0.001), outperforming both TAT (AUC = 0.655, P = 0.001) and PIC (AUC = 0.616, P = 0.012) (Fig. 1 A). Although the overall discriminative performance remained modest, D-dimer consistently showed superior classification capacity compared with other indicators (Supplementary Table 1). At the optimal cutoff value of 2.2 mg/L, D-dimer achieved a high sensitivity of 92.2% but a markedly low specificity of 18.3%, indicating that while it is highly effective for risk detection, its utility as a standalone screening tool is limited by a substantial false-positive rate. In contrast, thrombomodulin (TM) showed no statistically significant predictive value in the overall population (AUC = 0.577, P = 0.144), suggesting limited clinical relevance in this context. Spearman correlation analysis further revealed strong positive correlations among DDI, TAT, and PIC (r > 0.60, P < 0.001), indicating a high degree of overlap in the biological information captured by these markers (Fig. 1 D). This suggesting that TAT and PIC may reflect similar aspects of coagulation and fibrinolytic activation, whereas the relatively weaker correlation between DDI and TM (r = 0.247, P = 0.005) implies that endothelial injury-related pathways may be partially independent (Supplementary Table 2). These results indicate that although multiple biomarkers are associated with VTE risk, DDI provides the most favorable balance of predictive performance, while also highlighting the potential redundancy among certain coagulation-related indicators. 3.4 Decline in Predictive Efficacy of Thrombosis Biomarkers Within the High-Risk Group When focusing on the high-risk population with Caprini scores ≥ 9 points (n = 118), the predictive efficacy of all evaluated biomarkers declined compared with the overall cohort (Supplementary Table 3). Within this subgroup, only D-dimer (DDI) maintained a statistically significant association with VTE (P = 0.039), whereas TAT, PIC, and TM no longer demonstrated discriminatory ability (P > 0.05). Consistent with these findings, ROC curve analysis showed a uniform reduction in discriminative performance across all indicators. Specifically, the AUC of DDI decreased from 0.669 in the overall population to 0.614 in the high-risk subgroup, while similar declines were observed for TAT (0.655 to 0.612), PIC (0.616 to 0.597), and TM (0.577 to 0.542) (Fig. 1 C). Notably, despite this overall attenuation, D-dimer remained the best-performing biomarker within the high-risk group. These findings highlight the challenge of risk re-identification within populations already classified as high risk by clinical scores, and underscore the need for more refined stratification strategies beyond conventional clinical scoring and single-biomarker approaches. 3.5 Combined Model Improves VTE Predictive Efficacy Binary logistic regression was used to construct predictive models based on data from the high-risk subgroup (n = 118). Among the evaluated models, the combination of the Caprini score and D-dimer (Model 2) increased the AUC to 0.730, which outperformed the Caprini score alone (AUC = 0.701) (Fig. 1 B). This finding indicates that the incorporation of D-dimer provides incremental predictive value beyond clinical risk assessment. Model fit statistics further supported this improvement, as Model 2 demonstrated a lower − 2 log-likelihood value and higher pseudo-R2 values compared with the Caprini-only model. Although models incorporating TM (Model 3 and Model 4) exhibited numerically higher pseudo-R2 values, these analyses were based on a significantly reduced effective sample size (n = 78) due to missing data, which may limit the generalizability of these results. Meanwhile, the addition of TM to the Caprini + D-dimer model (Model 4) did not result in a meaningful improvement in discriminative performance compared with Model 2 (AUC = 0.725), despite the increased complexity. These results suggest that the combination of the Caprini score and D-dimer achieves an optimal balance between predictive performance and model simplicity, outperforming both single-variable and more complex multi-biomarker combinations (Table 2 ). Table 2 Comparison of Goodness-of-Fit for Different Logistic Regression Models Predicting VTE Model Included Variables Effective Sample Size -2LL Cox-Snell R² Nagelkerke R² Model 1 Caprini Total Score 118 147.440 0.069 0.094 Model 2 Caprini + DDI 118 144.710 0.090 0.123 Model 3 Caprini + TM 78 96.665 0.108 0.146 Model 4 Caprini + DDI+TM 78 94.239 0.136 0.183 Abbreviations: −2LL, − 2 log-likelihood. Notes: Cox–Snell R² and Nagelkerke R² are used to evaluate model explanatory power. Due to missing data for TM, the effective sample size for Model 3 and Model 4 was 78. 3.6 D-dimer Re-stratification Identifies Extremely High-Risk Patients To address the decreased biomarker discrimination within the high-risk group, we optimized the D-dimer (DDI) cutoff value using the Youden index. The optimal threshold was established at 5.53 mg/L (sensitivity 88.6%, specificity 35.1%) and pragmatically rounded to 5.5 mg/L for clinical applicability (Supplementary Table 4). Accordingly, the high-risk population was further divided into an extremely high-risk subgroup (DDI ≥ 5.5 mg/L) and a general high-risk subgroup (DDI < 5.5 mg/L). This re-stratification revealed a marked separation in clinical outcomes: VTE incidence was 44.8% in the extremely high-risk subgroup compared to 16.1% in the general high-risk subgroup, corresponding to a 2.78-fold increase in relative risk (RR = 2.78, 95% CI: 1.21–6.37, P = 0.005) (Fig. 2 B). Multivariate logistic regression analysis, adjusting for age, gender, and ISS score, further evaluated the robustness of this association (Fig. 2 A). Although the adjusted association for DDI ≥ 5.5 mg/L did not reach statistical significance (OR = 2.418, 95% CI: 0.660–8.864, P = 0.183), which was likely a consequence of constrained sample size and reduced statistical power, the direction and magnitude of the effect remained consistent with the significant univariate finding (OR = 4.225, P = 0.007). Furthermore, the Hosmer–Lemeshow test confirmed adequate model calibration (P = 0.379), and both age and ISS score remained robust independent predictors. These findings suggest that DDI–based re-stratification offers substantial clinical utility for identifying an extremely high-risk subset within the Caprini high-risk population, despite the limitations in statistical power for adjusted analyses. 3.7 Association Between Trauma Severity and Coagulation Activation To explore the biological basis for reduced biomarker discrimination in the Caprini high-risk cohort, we assessed the relationship between trauma severity and coagulation activation. Spearman correlation demonstrated a significant positive association between ISS and baseline D-dimer levels (r = 0.240, P = 0.027), indicating a trauma severity–dependent increase in hypercoagulability. When stratified by ISS categories, the proportion of patients with elevated DDI (≥ 5.5 mg/L) increased progressively: 55.6% for ISS < 16, 72.2% for ISS 16–24, and 83.9% for ISS ≥ 25 (linear-by-linear P = 0.033). VTE incidence exhibited a parallel stepwise increase from 27.8% to 51.6% across these strata, although the trend did not reach statistical significance (P = 0.077). These findings support a biologically plausible association between anatomical trauma severity and thrombotic risk escalation. 3.8 ISS-Based Re-stratification Further Refines Risk Gradients ROC analysis identified an optimal ISS threshold of 21 points (AUC = 0.618) (Supplementary Table 5). Combined stratification with DDI revealed a graded risk pattern. Among patients with moderate trauma (ISS ≤ 20), VTE incidence increased from 11.8% to 37.0% in those with elevated DDI, although this difference did not reach statistical significance (P = 0.081). In severe trauma (ISS ≥ 21), baseline VTE incidence was already elevated (33.3%) even with lower DDI, reaching 54.3% when DDI was ≥ 5.5 mg/L. Overall, DDI ≥ 5.5 mg/L remained significantly associated with VTE (χ² = 6.098, P = 0.014, OR = 4.174). Although the interaction term (ISS × DDI) was not statistically significant (P = 0.623), this combined stratification approach identified an "ultra-high-risk" subset (ISS ≥ 21 and DDI ≥ 5.5 mg/L), providing a clinically meaningful framework for risk refinement. 3.9 Internal Validation and Clinical Utility Assessment of the Two-Step Strategy The prediction model constructed for the Caprini high-risk population demonstrated strong internal validity and robustness. After 1,000 bootstrap resamples, the bias-corrected concordance index (C-index) of the two-step model reached 0.764, indicating stable and reliable discriminatory performance even after correcting for potential optimism. Calibration curve results (Fig. 3 A) showed high consistency between the predicted and observed VTE risks, characterized by a mean absolute error (MAE) of only 0.059 and a calibration slope of 0.838, suggesting no evidence of substantial overfitting. Decision curve analysis (Fig. 3 B) further confirmed the clinical utility of this strategy. Across a wide and clinically relevant threshold probability range (10% to 60%), the two-step model incorporating DDI ≥ 5.5 mg/L consistently provided a superior net benefit compared to the clinical baseline model based on the Caprini score and basic demographic variables. Overall, these findings demonstrate that the proposed two-step strategy is not only statistically robust but also offers meaningful clinical benefit, supporting its potential for guiding precision VTE risk assessment and management in trauma patients. 4 Discussion This study provides a comprehensive evaluation of integrating the Caprini score with thrombosis-related biomarkers to optimize VTE risk stratification in trauma patients. Our investigation reveals that while the Caprini score serves as a foundational assessment tool, its independent discriminative capacity remains notably modest in acute trauma settings. Among the evaluated biomarkers intended to bridge this diagnostic gap, D-dimer consistently emerged as the most robust predictor of VTE. Crucially, however, we identified a paradoxical attenuation in the discriminatory power of all biomarkers once patients were categorized into the Caprini high-risk echelon. To overcome this diagnostic ceiling, we demonstrated that applying an optimized D-dimer threshold for secondary stratification successfully unmasks a subset of patients with profoundly elevated thrombotic risk. Collectively, these insights advocate for a paradigm shift toward a hierarchical, two-step assessment strategy, one that tightly couples baseline clinical scoring with dynamic biomarker refinement to guide precision thromboprophylaxis. The Caprini score is one of the most widely adopted tools for VTE risk assessment in surgical practice, integrating multiple clinical variables including age, operative factors, and comorbidities[ 27 – 29 ]. However, its applicability in trauma populations remains controversial[ 30 – 32 ]. In the present study, the AUC of the Caprini score was 0.640, indicating statistically significant but limited discriminative performance. This finding is consistent with the notion that trauma patients differ fundamentally from elective surgical populations. In trauma settings, patients are typically admitted in the acute phase following injury, often with high ISS scores and substantial physiological derangement. The pathophysiological response to trauma involves a complex interplay of systemic inflammation, activation of the coagulation cascade, endothelial disruption, and release of tissue factor, which collectively contribute to a rapid shift toward a hypercoagulable state[ 33 – 35 ]. Importantly, these processes can occur independently of traditional risk factors captured by the Caprini score. As a result, a scoring system primarily based on static clinical variables may not adequately reflect the dynamic evolution of thrombotic risk in trauma patients. This mismatch likely contributes to the limited discriminative performance observed in this study, as well as the tendency for a large proportion of trauma patients to be classified within the high-risk category. Given these limitations, there has been increasing interest in incorporating laboratory-based indicators to enhance VTE risk assessment. Techniques such as thromboelastography (TEG) have been used to evaluate coagulation dynamics in trauma patients; however, their routine application is limited by operational complexity, cost, and limited availability in emergency settings[ 36 – 38 ]. In contrast, thrombosis-related biomarkers offer several practical advantages, including rapid turnaround time, ease of measurement, and widespread accessibility[ 39 , 40 ]. Among these, D-dimer has been extensively studied as a fibrin degradation product reflecting activation of the fibrinolytic system and has been widely used in the exclusion of VTE in low-risk populations[ 41 ]. In the present study, D-dimer demonstrated the highest predictive performance among the evaluated biomarkers, outperforming TAT, PIC, and TM. This finding suggests that D-dimer may serve as a more integrative indicator of thrombotic activity, reflecting both coagulation activation and subsequent fibrinolysis. Despite this advantage, an important observation was that the predictive performance of all biomarkers declined when analysis was restricted to patients already classified as high risk by the Caprini score. This phenomenon highlights an inherent limitation of biomarker-based discrimination in trauma populations and can be conceptualized as a “floor effect”. Under conditions of trauma-induced hypercoagulability, the levels of thrombosis-related biomarkers tend to be broadly elevated across patients, resulting in a compressed distribution and reduced inter-individual variability. Consequently, the ability of these markers to further discriminate risk within an already high-risk population is diminished. This finding provides a potential explanation for the inconsistent performance of biomarkers reported in previous trauma studies and underscores the importance of considering baseline risk context when interpreting biomarker data. To address this limitation, the present study explored an integrated strategy combining clinical scoring with biomarker-based refinement. The addition of D-dimer to the Caprini score resulted in an improvement in AUC, indicating that biomarker information provides incremental predictive value beyond traditional clinical assessment. More importantly, by optimizing the D-dimer cutoff value using the Youden index, we identified a clinically meaningful threshold (5.5 mg/L) that enabled further stratification within the high-risk group. Patients exceeding this threshold exhibited a markedly higher incidence of VTE, defining an “extremely high-risk” subgroup. This approach effectively transforms the role of D-dimer from a general screening biomarker into a tool for risk re-classification within a predefined high-risk population. Further supporting this interpretation, our ISS-based analyses demonstrated a clear relationship between trauma severity and coagulation activation. We observed a significant positive association between ISS and D-dimer levels, along with progressively higher proportions of patients exceeding the D-dimer threshold across increasing ISS strata. Notably, VTE incidence also showed a stepwise increase with rising ISS, reaching over 50% among patients with ISS ≥ 25. These findings provide biological plausibility for the observed floor effect, suggesting that in patients with severe trauma, systemic coagulation activation becomes broadly elevated, thereby reducing the discriminatory capacity of individual biomarkers. Clinically, ISS-based re-stratification further refined risk gradients within the Caprini high-risk population. Patients with both severe trauma (ISS ≥ 21) and elevated D-dimer levels exhibited the highest observed VTE incidence, identifying an ultra-high-risk subgroup. Although the statistical interaction between ISS and D-dimer was not statistically significant, the combined stratification pattern highlights the value of incorporating anatomical injury severity into biomarker-guided risk assessment. From a clinical perspective, this two-step strategy offers several potential advantages. First, it aligns with real-world decision-making processes, in which initial risk stratification is often followed by more refined assessment in selected patient subsets. Second, it provides a pragmatic framework for identifying patients who may benefit from intensified thromboprophylaxis, such as earlier initiation, higher dosing, or extended duration of anticoagulation. Third, by focusing on a subset of patients with the highest absolute risk, this approach may improve the efficiency of resource allocation while minimizing unnecessary intervention in lower-risk individuals. The robustness of this strategy was supported by internal validation and decision curve analysis. Bootstrap validation demonstrated stable model performance, suggesting that the observed predictive accuracy is unlikely to be solely attributable to overfitting. Calibration analysis further indicated good agreement between predicted and observed risks, reinforcing the reliability of the model. Importantly, decision curve analysis showed that the two-step model provides a greater net clinical benefit than conventional approaches across a broad range of threshold probabilities. This finding highlights the potential of the proposed strategy to improve clinical decision-making, particularly in settings where balancing the risks and benefits of thromboprophylaxis is critical. Several limitations of this study should be acknowledged. First, the retrospective single-center design may limit the generalizability of the findings. Second, the sample size, particularly within subgroup analyses, may have reduced statistical power, as reflected in the lack of statistical significance in some multivariate analyses despite consistent effect directions. Third, missing data for certain biomarkers, especially TM, resulted in reduced effective sample sizes in some models, which may have influenced comparisons of model performance. Finally, the D-dimer threshold identified in this study was derived from a single cohort and requires external validation in larger, multicenter populations before it can be widely implemented in clinical practice. In addition, ISS-based subgroup analyses were conducted in relatively small sample subsets, which may limit the statistical power for detecting interaction effects. 5 Conclusions In conclusion, this study demonstrates that both clinical risk scoring and thrombosis-related biomarkers have inherent limitations in trauma populations, particularly within patients already classified as high risk. By introducing a D-dimer–based re-stratification approach, we propose a two-step risk assessment model that improves risk discrimination and identifies a subgroup of patients at extremely high risk of VTE. This strategy provides a practical and clinically applicable framework for more precise risk assessment and individualized thromboprophylaxis in trauma patients, and may serve as a basis for future prospective validation and guideline development. Abbreviations VTE Venous thromboembolism DVT Deep vein thrombosis PTE Pulmonary thromboembolism DDI D-dimer TAT Thrombin–antithrombin complex PIC Plasmin–α2-plasmin inhibitor complex TM Thrombomodulin ISS Injury Severity Score ROC Receiver operating characteristic AUC Area under the curve CI Confidence interval OR Odds ratio RR Relative risk CTV Computed tomography venography DCA Decision curve analysis C-index Concordance index MAE Mean absolute error HIS Hospital Information System LIS Laboratory Information System SOP Standard operating procedure Declarations 7.1 Ethics approval and consent to participate This study was conducted following the Declaration of Helsinki and approved by the Biomedical Ethics Committee of Southern Medical University Zhujiang Hospital (Approval No.2025-KY-329-01). All participants provided written informed consent. 7.2 Consent for publication Not applicable 7.3 Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. 7.4 Competing interests The authors declare that they have no competing interests. 7.5 Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. 7.6 Authors' contributions QL, ML, and QZ contributed to study design, data collection, and statistical analysis. JH and BX assisted in data acquisition and database management. JS, YK, CG, and JY contributed to patient recruitment and clinical data verification. XX contributed to methodological guidance and interpretation of results. SZ and FM supervised the study and critically revised the manuscript. QL drafted the manuscript. All authors read and approved the final manuscript. 7.7 Acknowledgements We thank all the authors for their contributions to this manuscript. The graphical abstract was created in biorender. References Streiff MB, Bockenstedt PL, Cataland SR, et al. Venous Thromboembolic Disease J Natl Compr Canc Netw. 2011;9:714–77. Toker S, Hak DJ, Morgan SJ. Deep vein thrombosis prophylaxis in trauma patients. Thrombosis. 2011; 2011: 505373. Ramli NN, Iberahim S, Mohd Noor NH, et al. Haemostatic, Inflammatory, and Haematological Biomarkers Among Orthopaedic Patients With Prolonged Immobilization and the Risk of Hypercoagulable States. Cureus. 2024;16:e51552. Whiting PS, White-Dzuro GA, Greenberg SE, et al. Risk Factors for Deep Venous Thrombosis Following Orthopaedic Trauma Surgery: An Analysis of 56,000 patients. Arch Trauma Res. 2016;5:e32915. Obi AT, Pannucci CJ, Nackashi A, et al. Validation of the Caprini Venous Thromboembolism Risk Assessment Model in Critically Ill Surgical Patients. JAMA Surg. 2015;150:941–8. Xie M, Zhou Z, Fei F, et al. Risk assessment models for venous thromboembolism in surgical inpatients: a multicenter retrospective cohort study. Eur J Med Res. 2025;30:1298. Hazeltine MD, Scott EM, Dorfman JD. An abbreviated Caprini model for VTE risk assessment in trauma. J Thromb Thrombolysis. 2022;53:878–86. Scott EM, Rios EC, Guber RD et al. VTE risk assessment in trauma patients: a comparative analysis of Caprini, Trauma Embolic Scoring System, and Greenfield Risk Assessment profile models in a single sample. J Thromb Thrombolysis 2025. Bo H, Li Y, Liu G, et al. Assessing the Risk for Development of Deep Vein Thrombosis among Chinese Patients using the 2010 Caprini Risk Assessment Model: A Prospective Multicenter Study. J Atheroscler Thromb. 2020;27:801–8. Hanh BM, Cuong LQ, Son NT, et al. Determination of Risk Factors for Venous Thromboembolism by an Adapted Caprini Scoring System in Surgical Patients. J Pers Med. 2019;9:36. Hayssen H, Cires-Drouet R, Englum B, et al. Systematic review of venous thromboembolism risk categories derived from Caprini score. J Vasc Surg Venous Lymphat Disord. 2022;10:1401–e14091407. Swanson E. The Case against Chemoprophylaxis for Venous Thromboembolism Prevention and the Rationale for SAFE Anesthesia. Plast Reconstr Surg Glob Open. 2014;2:e160. Wang L, Zhong J, Xiao D, et al. Thrombomodulin (TM), thrombin-antithrombin complex (TAT), plasmin-α2-plasmininhibitor complex (PIC), and tissue plasminogen activator-inhibitor complex (t-PAIC) assessment of fibrinolytic activity in postpartum hemorrhage: a retrospective comparative cohort study. Ann Transl Med. 2022;10:1273. Lundbech M, Krag AE, Christensen TD, et al. Thrombin generation, thrombin-antithrombin complex, and prothrombin fragment F1 + 2 as biomarkers for hypercoagulability in cancer patients. Thromb Res. 2020;186:80–5. Musiał J, Pajak A, Undas A, et al. Thrombin generation markers and coronary heart disease risk factors in a Polish population sample. Thromb Haemost. 1997;77:697–700. Hoppensteadt D, Tsuruta K, Cunanan J, et al. Thrombin generation mediators and markers in sepsis-associated coagulopathy and their modulation by recombinant thrombomodulin. Clin Appl Thromb Hemost. 2014;20:129–35. Schaller J, Gerber SS. The plasmin–antiplasmin system: structural and functional aspects. Cell Mol Life Sci. 2011;68:785–801. Liu Y, Ma J, Shi Q, et al. Quantitatively monitoring acute ischemic stroke patients post recombinant tissue plasminogen activator treatment. Health Sci Rep. 2021;4:e218. Asakura H, Ontachi Y, Mizutani T, et al. An enhanced fibrinolysis prevents the development of multiple organ failure in disseminated intravascular coagulation in spite of much activation of blood coagulation. Crit Care Med. 2001;29:1164–8. Loghmani H, Conway EM. Exploring traditional and nontraditional roles for thrombomodulin. Blood. 2018;132:148–58. Pletsch-Borba L, Grafetstätter M, Hüsing A, et al. Vascular injury biomarkers and stroke risk: A population-based study. Neurology. 2020;94:e2337–45. Conway EM. Thrombomodulin and its role in inflammation. Semin Immunopathol. 2012;34:107–25. Adam SS, Key NS, Greenberg CS. D-dimer antigen: current concepts and future prospects. Blood. 2009;113:2878–87. Ye N, Liu Z, Wang X, et al. Evaluation of analytic and clinical performance of thrombin-antithrombin complex and D-dimer assay in prognosis of acute ischemic stroke. Blood Coagul Fibrinolysis. 2020;31:303–9. Thachil J, Lippi G, Favaloro EJ. D-Dimer Testing: Laboratory Aspects and Current Issues. Methods Mol Biol. 2017;1646:91–104. Nicolaides AN, Fareed J, Spyropoulos AC, et al. Prevention and management of venous thromboembolism. International Consensus Statement. Guidelines according to scientific evidence. Int Angiol. 2024;43:1–222. Caprini JA. Risk assessment as a guide to thrombosis prophylaxis. Curr Opin Pulm Med. 2010;16:448–52. Bahl V, Hu HM, Henke PK, et al. A Validation Study of a Retrospective Venous Thromboembolism Risk Scoring Method. Ann Surg. 2010;251:344–50. Cronin M, Dengler N, Krauss ES, et al. Completion of the Updated Caprini Risk Assessment Model (2013 Version). Clin Appl Thromb Hemost. 2019;25:1076029619838052. Lobastov K, Barinov V, Schastlivtsev I, et al. Validation of the Caprini risk assessment model for venous thromboembolism in high-risk surgical patients in the background of standard prophylaxis. J Vasc Surg Venous Lymphat Disord. 2016;4:153–60. Hazeltine MD, Guber RD, Buettner H, et al. Venous thromboembolism risk stratification in trauma using the Caprini risk assessment model. Thromb Res. 2021;208:52–7. Lobastov K, Urbanek T, Stepanov E, et al. The Thresholds of Caprini Score Associated With Increased Risk of Venous Thromboembolism Across Different Specialties: A Systematic Review. Ann Surg. 2023;277:929–37. Gando S, Otomo Y. Local hemostasis, immunothrombosis, and systemic disseminated intravascular coagulation in trauma and traumatic shock. Crit Care. 2015;19:72. Stojkovic S, Kaun C, Basilio J, et al. Tissue factor is induced by interleukin-33 in human endothelial cells: a new link between coagulation and inflammation. Sci Rep. 2016;6:25171. Bodnar D, Bosley E, Raven S, et al. The nature and timing of coagulation dysfunction in a cohort of trauma patients in the Australian pre-hospital setting. Injury. 2024;55:111124. Baksaas-Aasen K, Gall LS, Stensballe J, et al. Viscoelastic haemostatic assay augmented protocols for major trauma haemorrhage (ITACTIC): a randomized, controlled trial. Intensive Care Med. 2021;47:49–59. Hagemo JS, Næss PA, Johansson P, et al. Evaluation of TEG® and RoTEM® inter-changeability in trauma patients. Injury. 2013;44:600–5. Burton AG, Jandrey KE. Use of Thromboelastography in Clinical Practice. Vet Clin North Am Small Anim Pract. 2020;50:1397–409. Di Nisio M, Squizzato A, Rutjes AWS, et al. Diagnostic accuracy of D-dimer test for exclusion of venous thromboembolism: a systematic review. J Thromb Haemost. 2007;5:296–304. Shaw JR, Nopp S, Stavik B, et al. Thrombosis, Translational Medicine, and Biomarker Research: Moving the Needle. J Am Heart Assoc. 2025;14:e038782. Weitz JI, Fredenburgh JC, Eikelboom JW. A Test in Context: D-Dimer. JACC. 2017;70:2411–20. Additional Declarations No competing interests reported. Supplementary Files SupplementaryMaterial.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 22 Apr, 2026 Reviewers agreed at journal 21 Apr, 2026 Reviewers invited by journal 21 Apr, 2026 Editor assigned by journal 20 Apr, 2026 Submission checks completed at journal 20 Apr, 2026 First submitted to journal 16 Apr, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9440559","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":627914822,"identity":"2ea22238-8a81-4a3b-9f86-d3f7385d6b7d","order_by":0,"name":"Qiang Liu","email":"","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Qiang","middleName":"","lastName":"Liu","suffix":""},{"id":627914823,"identity":"60bbd635-51b5-4de0-97f3-06063e69602e","order_by":1,"name":"Mengqing Lin","email":"","orcid":"","institution":"Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Mengqing","middleName":"","lastName":"Lin","suffix":""},{"id":627914824,"identity":"b4b14d1d-e878-49c4-a5f9-22bad5c7ebb3","order_by":2,"name":"Qian Zhang","email":"","orcid":"","institution":"Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Qian","middleName":"","lastName":"Zhang","suffix":""},{"id":627914825,"identity":"85a6f304-7036-4ace-a53a-0827f7a80ec3","order_by":3,"name":"Junfei Huang","email":"","orcid":"","institution":"Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Junfei","middleName":"","lastName":"Huang","suffix":""},{"id":627914826,"identity":"4a21dc02-b27d-4cbb-9c9c-52f9fae0ea18","order_by":4,"name":"Bingyu Xu","email":"","orcid":"","institution":"Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Bingyu","middleName":"","lastName":"Xu","suffix":""},{"id":627914828,"identity":"beec4668-cd6c-4af3-96e8-d6ea10e78693","order_by":5,"name":"Jinlian Shao","email":"","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jinlian","middleName":"","lastName":"Shao","suffix":""},{"id":627914830,"identity":"4a0d99fa-76da-46cc-bef5-e5f69a0fd3e4","order_by":6,"name":"Yuzhan Kang","email":"","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yuzhan","middleName":"","lastName":"Kang","suffix":""},{"id":627914832,"identity":"60aab884-8c14-405b-8f88-41267ada71d5","order_by":7,"name":"Chaohao Guan","email":"","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Chaohao","middleName":"","lastName":"Guan","suffix":""},{"id":627914834,"identity":"ef9b3dbe-a50d-4991-815b-00814b5d5ab2","order_by":8,"name":"Junhua Yang","email":"","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Junhua","middleName":"","lastName":"Yang","suffix":""},{"id":627914836,"identity":"80eb2a71-2bf1-4758-8174-fe6827402a61","order_by":9,"name":"Xun Xia","email":"","orcid":"","institution":"Maternity and Child Health Hospital of Huadu District","correspondingAuthor":false,"prefix":"","firstName":"Xun","middleName":"","lastName":"Xia","suffix":""},{"id":627914837,"identity":"5fdbc16b-4b6a-4e86-8934-e8a729f6b0c4","order_by":10,"name":"Sainan Zou","email":"","orcid":"","institution":"The Sixth Affiliated Hospital, Sun Yat-sen University","correspondingAuthor":false,"prefix":"","firstName":"Sainan","middleName":"","lastName":"Zou","suffix":""},{"id":627914838,"identity":"95a65cd7-6d69-4dbf-8a16-7cc7891eb989","order_by":11,"name":"Fanke Meng","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABCElEQVRIie3RMUsDMRTA8XcE3i2hrjmueH6EVw5aHUQ/Sopgp4LjDQ4VIbfcB3AQP8NNglvKg3Yp3CroUBFucnO5oYN3g1ubc3TIfwpJfoQkAD7ff0wAApBtB4K3qrk9xjC3fyQhXtNZsUoHcqP7zmkJtORIkspQTJ/UJTn30zqsP+XNezIRQPQqcWYUaGiy58OE5SSVVI9e7kF/PJwO5ya+s0GxeXMRjCVxUDLYVEmcm6HVIjAuEtYduSg5WMQ7FDNUmnoIjDsyLVmAUih0L4lYjqNH4quSESkqViPTPvLSdZdBta7V147Py6r67r4ySfJ8uW2yw+TE7pvdO/lbsnCt+nw+n6/rB2OtVwxr8gk0AAAAAElFTkSuQmCC","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":true,"prefix":"","firstName":"Fanke","middleName":"","lastName":"Meng","suffix":""}],"badges":[],"createdAt":"2026-04-16 16:24:16","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9440559/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9440559/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108492745,"identity":"6d7ced94-4f1b-448c-b123-ddd90ece4cfd","added_by":"auto","created_at":"2026-05-05 09:58:30","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1391926,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePredictive efficacy of the Caprini score and thrombosis-related biomarkers for VTE in trauma patients\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) ROC curves of the Caprini score and thrombosis biomarkers in the entire population.\u003c/p\u003e\n\u003cp\u003e(B) ROC curves of individual indicators and the combined model (Caprini + DDI) in the Caprini high-risk group.\u003c/p\u003e\n\u003cp\u003e(C) Trends in the AUC of various thrombosis-related biomarkers between the overall population and the Caprini high-risk group.\u003c/p\u003e\n\u003cp\u003e(D) Spearman correlation heatmap among thrombosis-related biomarkers. Values are correlation coefficients (r); *P \u0026lt; 0.05, **P \u0026lt; 0.01, ***P \u0026lt; 0.001.\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-9440559/v1/be56853d7bb658355849ef68.png"},{"id":108385708,"identity":"58013731-c3cb-4645-8ece-7ee0659f4b00","added_by":"auto","created_at":"2026-05-04 06:04:53","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1194733,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRe-stratification Strategy Based on D-dimer Threshold and Regression Analysis in the Caprini High-risk Group\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Forest plot of factors associated with VTE occurrence in the Caprini high-risk group. Results from univariate analysis and multivariate logistic regression adjusted for age, gender, and ISS score are shown. Squares represent odds ratios (ORs), and horizontal lines represent 95% confidence intervals. P values correspond to multivariate analysis.\u003c/p\u003e\n\u003cp\u003e(B) Sankey diagram of re-stratification based on the D-dimer threshold (5.5 mg/L) in the Caprini high-risk group. The incidence of VTE in the extremely high-risk subgroup (DDI ≥ 5.5 mg/L) was significantly higher than that in the general high-risk group, with a relative risk of 2.78 (P = 0.005).\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-9440559/v1/4aa8a1c011bf12000ce085d5.png"},{"id":108385709,"identity":"86afa7ed-2596-4ba8-a576-077e6b7563ef","added_by":"auto","created_at":"2026-05-04 06:04:53","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":694693,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eInternal Validation and Clinical Utility Evaluation of the VTE Risk Re-stratification Model in Trauma Patients\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Calibration curve. This figure demonstrates the consistency between the VTE risk predicted by the two-step model and the actual observed risk after 1000 Bootstrap resamples. The x-axis is the predicted risk probability, and the y-axis is the actual observed incidence. The bias-corrected line (solid line) closely follows the 45° ideal value dashed line, with a mean absolute error (MAE) of 0.059 and a calibration slope of 0.84.\u003c/p\u003e\n\u003cp\u003e(B) Decision curve analysis (DCA). This figure compares the clinical net benefit of the two-step model (Caprini score + D-dimer ≥ 5.5 mg/L) with the clinical baseline model (Caprini score + age + gender + ISS).\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-9440559/v1/93665a1c48a74444edcabfcc.png"},{"id":109081141,"identity":"9f739039-ed46-4614-acbb-dde88b440be7","added_by":"auto","created_at":"2026-05-12 12:01:02","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3299862,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9440559/v1/8507abba-53cc-4817-a29c-c0b15b1b40b7.pdf"},{"id":108385706,"identity":"fece5982-12f9-4107-a0bb-b998a019194b","added_by":"auto","created_at":"2026-05-04 06:04:53","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":30820,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-9440559/v1/8bbfbae0c14c8b42a51688e0.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"A Two-Step Risk Stratification Strategy Integrating Caprini Score and D-Dimer for Venous Thromboembolism in Trauma Patients","fulltext":[{"header":"1 Background","content":"\u003cp\u003eVenous thromboembolism (VTE), encompassing deep venous thrombosis (DVT) and pulmonary thromboembolism (PTE), remains a formidable perioperative complication in trauma surgery[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. The acute physiological stress of trauma\u0026mdash;characterized by severe vascular endothelial disruption, prolonged immobilization, and a profound hypercoagulable state\u0026mdash;predisposes these patients to an exceptionally high risk of VTE, which subsequently drives significant morbidity and mortality[\u003cspan additionalcitationids=\"CR3\" citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eWhile the Caprini risk assessment model is universally endorsed for clinical VTE stratification[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e], its reliance on the cumulative scoring of static clinical risk factors limits its diagnostic agility in trauma patients experiencing acute pathophysiological shifts[\u003cspan additionalcitationids=\"CR7\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Consequently, the isolated application of the Caprini score frequently funnels a disproportionate volume of trauma patients into a broad \"high-risk\" category[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. This non-specific classification masks substantial inter-individual variability in actual thrombotic risk[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e], ultimately complicating the formulation and execution of tailored, precision-driven thromboprophylaxis[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eTo navigate these clinical ambiguities, novel thrombosis-related biomarkers have emerged as highly sensitive instruments for detecting early hypercoagulability[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Specifically, the thrombin-antithrombin complex (TAT) serves as a direct barometer of thrombin generation[\u003cspan additionalcitationids=\"CR15\" citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e], the plasmin-α2-plasmin inhibitor complex (PIC) quantifies the activation of the fibrinolytic system[\u003cspan additionalcitationids=\"CR18\" citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e], and thrombomodulin (TM) acts as a proximal indicator of vascular endothelial disruption[\u003cspan additionalcitationids=\"CR21\" citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Although these novel metrics hold theoretical superiority over traditional D-dimer (DDI) in capturing early kinetic shifts within the coagulation cascade[\u003cspan additionalcitationids=\"CR24\" citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e], their practical utility and optimal diagnostic thresholds for risk re-stratification in trauma cohorts remain poorly defined.\u003c/p\u003e \u003cp\u003eTo address this critical knowledge gap, we retrospectively analyzed a cohort of 206 trauma center patients to rigorously evaluate the prognostic synergy between the Caprini score and a panel of thrombosis biomarkers. Our primary objective was to investigate the hypothesized \"predictive attenuation\" (or biomarker saturation) within the Caprini high-risk stratum. By optimizing biomarker cutoff values specifically tailored to this vulnerable population, we aimed to engineer a pragmatic, two-step \"Caprini screening\u0026mdash;biomarker re-stratification\" strategy. Ultimately, this hierarchical model seeks to deliver a robust quantitative framework that empowers clinicians to implement precise and individualized VTE prophylaxis in trauma surgery.\u003c/p\u003e"},{"header":"2 Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Study Population and Ethics\u003c/h2\u003e \u003cp\u003eThis retrospective cohort study included 206 trauma surgical patients admitted to Zhujiang Hospital of Southern Medical University between January 2022 and December 2025. The study protocol was conducted in accordance with the Declaration of Helsinki and received formal approval from Medical Ethics Committee of Zhujiang Hospital of Southern Medical University (Approval No.: 2025-KY-329-01). Written informed consent was obtained from all patients or their legal guardians prior to enrollment.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Inclusion and Exclusion Criteria\u003c/h2\u003e \u003cp\u003ePatients were eligible for inclusion if they: (1) were aged\u0026thinsp;\u0026ge;\u0026thinsp;18 years; (2) were admitted for acute trauma and underwent subsequent surgical intervention; (3) completed both the Caprini risk assessment and initial thrombosis-related biomarker testing within 24 hours of admission; and (4) possessed complete clinical and follow-up data.\u003c/p\u003e \u003cp\u003eExclusion criteria were: (1) VTE diagnosed prior to or immediately upon admission based on clinical manifestations (e.g., limb swelling, positive Homans sign) or imaging confirmation; (2) coexisting severe hepatic or renal insufficiency, active malignancies, or pre-existing hematological disorders affecting coagulation; (3) administration of anticoagulant or thrombolytic therapy within 14 days prior to admission; (4) pregnancy or lactation; and (5) failure to complete mandatory lower extremity vascular imaging (color Doppler ultrasound or venography) during hospitalization.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Clinical Data Acquisition and Variable Selection\u003c/h2\u003e \u003cp\u003eThis study followed a rigorous retrospective protocol for data extraction. Comprehensive clinical profiles and laboratory parameters for all eligible patients were systematically retrieved from the Hospital Information System (HIS) and Laboratory Information System (LIS). The collected dataset encompassed: (1) baseline demographics, including age and gender; (2) injury-specific metrics, primarily the Injury Severity Score (ISS) and admission Caprini score; and (3) definitive VTE diagnostic outcomes derived from postoperative imaging (e.g., color Doppler ultrasound, venography, or CTV). To ensure high data fidelity and minimize extraction bias, all variables were independently audited by two researchers, with any discrepancies resolved through consensus and cross-verification of the original electronic medical records.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Laboratory Analysis and Biomarker Measurement\u003c/h2\u003e \u003cp\u003ePeripheral venous blood samples were collected by trained nursing staff upon admission or during the immediate preoperative period. Samples were drawn into standardized vacuum tubes containing trisodium citrate anticoagulant and processed via programmed centrifugation according to established institutional protocols.\u003c/p\u003e \u003cp\u003eThe plasma concentrations of four key thrombosis-related biomarkers\u0026mdash;plasmin-α2-plasmin inhibitor complex (PIC), thrombomodulin (TM), thrombin-antithrombin III complex (TAT), and D-dimer (DDI)\u0026mdash;were quantified. All assays were performed by qualified laboratory personnel using unified automated analytical platforms and standardized operating procedures (SOPs) to ensure technical reproducibility. Laboratory results were integrated in real-time into the LIS, from which researchers extracted the initial test values (baseline levels) for subsequent predictive modeling.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 VTE Diagnosis and Risk Grouping\u003c/h2\u003e \u003cp\u003eVTE diagnosis followed the International Consensus Statement on the Prevention and Management of Venous Thromboembolism[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e], employing a hierarchical diagnostic strategy. Initial clinical probability was determined via the Caprini scale. Patients identified as moderate-to-high risk, or those with elevated D-dimer levels, underwent priority screening with color Doppler ultrasound. In cases of persistent clinical suspicion despite negative initial ultrasound, VTE was further confirmed or excluded via contrast-enhanced venography or Computed Tomography Venography (CTV).\u003c/p\u003e \u003cp\u003eFor comparative analysis, patients were stratified by admission Caprini scores into three risk tiers: low-risk (0\u0026ndash;4 points, n\u0026thinsp;=\u0026thinsp;34), intermediate-risk (5\u0026ndash;8 points, n\u0026thinsp;=\u0026thinsp;54), and high-risk (\u0026ge;\u0026thinsp;9 points, n\u0026thinsp;=\u0026thinsp;118). Based on definitive postoperative imaging, the cohort was dichotomized into the VTE group (n\u0026thinsp;=\u0026thinsp;64) and the non-VTE group (n\u0026thinsp;=\u0026thinsp;142).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.6 Statistical Analysis\u003c/h2\u003e \u003cp\u003eStatistical analyses were performed using SPSS version 27.0. The normality of continuous data was assessed using the Shapiro-Wilk test. Normally distributed variables are presented as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\stackrel{-}{\\text{x}}\\pm\\:\\text{s}\\)\u003c/span\u003e\u003c/span\u003e) and compared via independent samples \u003cem\u003et\u003c/em\u003e-tests. Non-normally distributed data are expressed as median (interquartile range) \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\left[\\text{M}\\left({\\text{P}}_{25},{\\text{P}}_{75}\\right)\\right]\\)\u003c/span\u003e\u003c/span\u003e and analyzed using the Mann-Whitney \u003cem\u003eU\u003c/em\u003e test. Categorical variables are reported as frequencies (%) and compared using the chi-square (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{{\\chi\\:}}^{2}\\)\u003c/span\u003e\u003c/span\u003e) test.\u003c/p\u003e \u003cp\u003eThe discriminative performance of the Caprini score and individual biomarkers was evaluated using Receiver Operating Characteristic (ROC) curves and the corresponding Area Under the Curve (AUC). Spearman rank correlation and linear-by-linear association tests were used to evaluate relationships between Injury Severity Score (ISS), D-dimer levels, and VTE incidence across clinically relevant ISS strata (\u0026lt;\u0026thinsp;16, 16\u0026ndash;24, \u0026ge;\u0026thinsp;25). The optimal D-dimer threshold for re-stratification within the high-risk subgroup, as well as the optimal ISS cutoff for risk separation, were determined using the Youden index. Logistic regression models were constructed to assess the incremental predictive value of combined strategies and to explore potential interactions between ISS and D-dimer. To ensure model robustness and mitigate overfitting, internal validation was performed using 1,000 bootstrap resamples. Finally, Decision Curve Analysis (DCA) was utilized to quantify the net clinical benefit of the proposed two-step strategy. All tests were two-tailed, with P\u0026thinsp;\u0026lt;\u0026thinsp;0.05 denoting statistical significance.\u003c/p\u003e \u003c/div\u003e"},{"header":"3 Results","content":"\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Caprini Risk Stratification and Patient Baseline Characteristics\u003c/h2\u003e \u003cp\u003eA total of 206 trauma surgical patients were included in this study, with an overall VTE incidence of 31.1%. Based on the Caprini score, the patients were divided into three groups: a low-risk group (0\u0026ndash;4 points, n\u0026thinsp;=\u0026thinsp;34), an intermediate-risk group (5\u0026ndash;8 points, n\u0026thinsp;=\u0026thinsp;54), and a high-risk group (\u0026ge;\u0026thinsp;9 points, n\u0026thinsp;=\u0026thinsp;118).\u003c/p\u003e \u003cp\u003eThere were no statistically significant differences in age and gender distribution among the three groups (P\u0026thinsp;\u0026gt;\u0026thinsp;0.05), suggesting that these variables were unlikely to confound the relationship between risk stratification and VTE occurrence. In contrast, the difference in ISS scores among the groups was significant (P\u0026thinsp;=\u0026thinsp;0.034), with higher scores observed in the higher Caprini groups. This indicates that Caprini-based risk stratification is consistent with trauma severity to some extent, supporting its clinical relevance in this setting.\u003c/p\u003e \u003cp\u003eAlthough the difference in VTE incidence among the three groups was marginally significant (χ\u0026sup2; = 5.007, P\u0026thinsp;=\u0026thinsp;0.082), the linear trend test confirmed that VTE incidence exhibited a progressive upward trend with increasing risk levels (χ\u0026sup2; = 3.870, P\u0026thinsp;=\u0026thinsp;0.049). Taken together, these findings suggest that while categorical comparisons alone may understate group differences, the Caprini score retains the ability to capture an overall risk gradient, highlighting its value as a baseline risk stratification tool in trauma patients (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eComparison of baseline characteristics of trauma patients with different Caprini risk stratifications\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCharacteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLow-Risk Group (n\u0026thinsp;=\u0026thinsp;34)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIntermediate-Risk Group (n\u0026thinsp;=\u0026thinsp;54)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHigh-Risk Group (n\u0026thinsp;=\u0026thinsp;118)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eStatistic\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eP-value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge(years, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\stackrel{-}{\\text{x}}\\pm\\:\\text{s}\\)\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e46.1\u0026thinsp;\u0026plusmn;\u0026thinsp;16.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e45.9\u0026thinsp;\u0026plusmn;\u0026thinsp;17.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e48.6\u0026thinsp;\u0026plusmn;\u0026thinsp;18.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eF\u0026thinsp;=\u0026thinsp;0.572\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.566\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale, n(%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e26(76.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e45(83.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e88(74.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eχ\u0026sup2;=1.625\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.444\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISS Score, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\text{M}(\\text{Q}1,\\text{Q}3)\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e17(11,20)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e20(14,25)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e20(16,27)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eH\u0026thinsp;=\u0026thinsp;6.787\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.034\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVTE Occurrence, n(%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8(23.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e12(22.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e44(37.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eχ\u0026sup2;=5.007\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.082\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003eNotes: Age follows a normal distribution and is expressed as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation, analyzed using one-way ANOVA (F test); ISS score does not follow a normal distribution and is expressed as median (interquartile range), analyzed using the Kruskal\u0026ndash;Wallis H test; categorical variables are expressed as number (percentage) and analyzed using the χ\u0026sup2; test.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Stratification Efficacy of Caprini Score in the Trauma Population\u003c/h2\u003e \u003cp\u003eSpearman correlation analysis showed a statistically significant but weak positive correlation between the Caprini score and VTE occurrence (r\u0026thinsp;=\u0026thinsp;0.148, P\u0026thinsp;=\u0026thinsp;0.034), indicating that higher scores correlate with an increased risk of thrombosis, although the strength of this relationship is limited. ROC curve analysis further revealed that the AUC of the Caprini score for predicting VTE was 0.640 (95% CI: 0.553\u0026ndash;0.726, P\u0026thinsp;=\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA), suggesting modest predictive performance in this population. At the optimal cutoff value of approximately 8.5 points, the sensitivity was 68.8% but the specificity was relatively low at 47.9%. This imbalance indicates that while the Caprini score is capable of identifying a substantial proportion of patients who will develop VTE, it also misclassifies a considerable number of low-risk individuals as high risk. These findings suggest that relying solely on clinical risk factor scoring involves a high false-positive rate and limited overall predictive discrimination in the trauma population.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Predictive Value of Thrombosis Biomarkers in the Overall Population\u003c/h2\u003e \u003cp\u003eSingle-indicator ROC curve analysis showed that among the evaluated biomarkers, D-dimer (DDI) demonstrated the highest discriminative ability for VTE, with an AUC of 0.669 (95% CI: 0.592\u0026ndash;0.747, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001), outperforming both TAT (AUC\u0026thinsp;=\u0026thinsp;0.655, P\u0026thinsp;=\u0026thinsp;0.001) and PIC (AUC\u0026thinsp;=\u0026thinsp;0.616, P\u0026thinsp;=\u0026thinsp;0.012) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). Although the overall discriminative performance remained modest, D-dimer consistently showed superior classification capacity compared with other indicators (Supplementary Table\u0026nbsp;1). At the optimal cutoff value of 2.2 mg/L, D-dimer achieved a high sensitivity of 92.2% but a markedly low specificity of 18.3%, indicating that while it is highly effective for risk detection, its utility as a standalone screening tool is limited by a substantial false-positive rate. In contrast, thrombomodulin (TM) showed no statistically significant predictive value in the overall population (AUC\u0026thinsp;=\u0026thinsp;0.577, P\u0026thinsp;=\u0026thinsp;0.144), suggesting limited clinical relevance in this context.\u003c/p\u003e \u003cp\u003eSpearman correlation analysis further revealed strong positive correlations among DDI, TAT, and PIC (r \u0026gt; 0.60, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001), indicating a high degree of overlap in the biological information captured by these markers (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD). This suggesting that TAT and PIC may reflect similar aspects of coagulation and fibrinolytic activation, whereas the relatively weaker correlation between DDI and TM (r\u0026thinsp;=\u0026thinsp;0.247, P\u0026thinsp;=\u0026thinsp;0.005) implies that endothelial injury-related pathways may be partially independent (Supplementary Table\u0026nbsp;2). These results indicate that although multiple biomarkers are associated with VTE risk, DDI provides the most favorable balance of predictive performance, while also highlighting the potential redundancy among certain coagulation-related indicators.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Decline in Predictive Efficacy of Thrombosis Biomarkers Within the High-Risk Group\u003c/h2\u003e \u003cp\u003eWhen focusing on the high-risk population with Caprini scores\u0026thinsp;\u0026ge;\u0026thinsp;9 points (n\u0026thinsp;=\u0026thinsp;118), the predictive efficacy of all evaluated biomarkers declined compared with the overall cohort (Supplementary Table\u0026nbsp;3). Within this subgroup, only D-dimer (DDI) maintained a statistically significant association with VTE (P\u0026thinsp;=\u0026thinsp;0.039), whereas TAT, PIC, and TM no longer demonstrated discriminatory ability (P\u0026thinsp;\u0026gt;\u0026thinsp;0.05). Consistent with these findings, ROC curve analysis showed a uniform reduction in discriminative performance across all indicators. Specifically, the AUC of DDI decreased from 0.669 in the overall population to 0.614 in the high-risk subgroup, while similar declines were observed for TAT (0.655 to 0.612), PIC (0.616 to 0.597), and TM (0.577 to 0.542) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC). Notably, despite this overall attenuation, D-dimer remained the best-performing biomarker within the high-risk group. These findings highlight the challenge of risk re-identification within populations already classified as high risk by clinical scores, and underscore the need for more refined stratification strategies beyond conventional clinical scoring and single-biomarker approaches.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e3.5 Combined Model Improves VTE Predictive Efficacy\u003c/h2\u003e \u003cp\u003eBinary logistic regression was used to construct predictive models based on data from the high-risk subgroup (n\u0026thinsp;=\u0026thinsp;118). Among the evaluated models, the combination of the Caprini score and D-dimer (Model 2) increased the AUC to 0.730, which outperformed the Caprini score alone (AUC\u0026thinsp;=\u0026thinsp;0.701) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB). This finding indicates that the incorporation of D-dimer provides incremental predictive value beyond clinical risk assessment. Model fit statistics further supported this improvement, as Model 2 demonstrated a lower\u0026thinsp;\u0026minus;\u0026thinsp;2 log-likelihood value and higher pseudo-R2 values compared with the Caprini-only model. Although models incorporating TM (Model 3 and Model 4) exhibited numerically higher pseudo-R2 values, these analyses were based on a significantly reduced effective sample size (n\u0026thinsp;=\u0026thinsp;78) due to missing data, which may limit the generalizability of these results. Meanwhile, the addition of TM to the Caprini\u0026thinsp;+\u0026thinsp;D-dimer model (Model 4) did not result in a meaningful improvement in discriminative performance compared with Model 2 (AUC\u0026thinsp;=\u0026thinsp;0.725), despite the increased complexity. These results suggest that the combination of the Caprini score and D-dimer achieves an optimal balance between predictive performance and model simplicity, outperforming both single-variable and more complex multi-biomarker combinations (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eComparison of Goodness-of-Fit for Different Logistic Regression Models Predicting VTE\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModel\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIncluded Variables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEffective Sample Size\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-2LL\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eCox-Snell R\u0026sup2;\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNagelkerke R\u0026sup2;\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModel 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCaprini Total Score\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e118\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e147.440\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.069\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.094\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModel 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCaprini\u0026thinsp;+\u0026thinsp;DDI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e118\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e144.710\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.090\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.123\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModel 3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCaprini\u0026thinsp;+\u0026thinsp;TM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e96.665\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.108\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.146\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModel 4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCaprini\u0026thinsp;+\u0026thinsp;DDI+TM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e94.239\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.136\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.183\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003eAbbreviations: \u0026minus;2LL, \u0026minus;\u0026thinsp;2 log-likelihood.\u003c/td\u003e\u003c/tr\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003eNotes: Cox\u0026ndash;Snell R\u0026sup2; and Nagelkerke R\u0026sup2; are used to evaluate model explanatory power. Due to missing data for TM, the effective sample size for Model 3 and Model 4 was 78.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e3.6 D-dimer Re-stratification Identifies Extremely High-Risk Patients\u003c/h2\u003e \u003cp\u003eTo address the decreased biomarker discrimination within the high-risk group, we optimized the D-dimer (DDI) cutoff value using the Youden index. The optimal threshold was established at 5.53 mg/L (sensitivity 88.6%, specificity 35.1%) and pragmatically rounded to 5.5 mg/L for clinical applicability (Supplementary Table\u0026nbsp;4). Accordingly, the high-risk population was further divided into an extremely high-risk subgroup (DDI\u0026thinsp;\u0026ge;\u0026thinsp;5.5 mg/L) and a general high-risk subgroup (DDI\u0026thinsp;\u0026lt;\u0026thinsp;5.5 mg/L). This re-stratification revealed a marked separation in clinical outcomes: VTE incidence was 44.8% in the extremely high-risk subgroup compared to 16.1% in the general high-risk subgroup, corresponding to a 2.78-fold increase in relative risk (RR\u0026thinsp;=\u0026thinsp;2.78, 95% CI: 1.21\u0026ndash;6.37, P\u0026thinsp;=\u0026thinsp;0.005) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eMultivariate logistic regression analysis, adjusting for age, gender, and ISS score, further evaluated the robustness of this association (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). Although the adjusted association for DDI\u0026thinsp;\u0026ge;\u0026thinsp;5.5 mg/L did not reach statistical significance (OR\u0026thinsp;=\u0026thinsp;2.418, 95% CI: 0.660\u0026ndash;8.864, P\u0026thinsp;=\u0026thinsp;0.183), which was likely a consequence of constrained sample size and reduced statistical power, the direction and magnitude of the effect remained consistent with the significant univariate finding (OR\u0026thinsp;=\u0026thinsp;4.225, P\u0026thinsp;=\u0026thinsp;0.007). Furthermore, the Hosmer\u0026ndash;Lemeshow test confirmed adequate model calibration (P\u0026thinsp;=\u0026thinsp;0.379), and both age and ISS score remained robust independent predictors. These findings suggest that DDI\u0026ndash;based re-stratification offers substantial clinical utility for identifying an extremely high-risk subset within the Caprini high-risk population, despite the limitations in statistical power for adjusted analyses.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e3.7 Association Between Trauma Severity and Coagulation Activation\u003c/h2\u003e \u003cp\u003eTo explore the biological basis for reduced biomarker discrimination in the Caprini high-risk cohort, we assessed the relationship between trauma severity and coagulation activation. Spearman correlation demonstrated a significant positive association between ISS and baseline D-dimer levels (r\u0026thinsp;=\u0026thinsp;0.240, P\u0026thinsp;=\u0026thinsp;0.027), indicating a trauma severity\u0026ndash;dependent increase in hypercoagulability.\u003c/p\u003e \u003cp\u003eWhen stratified by ISS categories, the proportion of patients with elevated DDI (\u0026ge;\u0026thinsp;5.5 mg/L) increased progressively: 55.6% for ISS\u0026thinsp;\u0026lt;\u0026thinsp;16, 72.2% for ISS 16\u0026ndash;24, and 83.9% for ISS\u0026thinsp;\u0026ge;\u0026thinsp;25 (linear-by-linear P\u0026thinsp;=\u0026thinsp;0.033). VTE incidence exhibited a parallel stepwise increase from 27.8% to 51.6% across these strata, although the trend did not reach statistical significance (P\u0026thinsp;=\u0026thinsp;0.077). These findings support a biologically plausible association between anatomical trauma severity and thrombotic risk escalation.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e3.8 ISS-Based Re-stratification Further Refines Risk Gradients\u003c/h2\u003e \u003cp\u003eROC analysis identified an optimal ISS threshold of 21 points (AUC\u0026thinsp;=\u0026thinsp;0.618) (Supplementary Table\u0026nbsp;5). Combined stratification with DDI revealed a graded risk pattern. Among patients with moderate trauma (ISS\u0026thinsp;\u0026le;\u0026thinsp;20), VTE incidence increased from 11.8% to 37.0% in those with elevated DDI, although this difference did not reach statistical significance (P\u0026thinsp;=\u0026thinsp;0.081). In severe trauma (ISS\u0026thinsp;\u0026ge;\u0026thinsp;21), baseline VTE incidence was already elevated (33.3%) even with lower DDI, reaching 54.3% when DDI was \u0026ge;\u0026thinsp;5.5 mg/L.\u003c/p\u003e \u003cp\u003eOverall, DDI\u0026thinsp;\u0026ge;\u0026thinsp;5.5 mg/L remained significantly associated with VTE (χ\u0026sup2; = 6.098, P\u0026thinsp;=\u0026thinsp;0.014, OR\u0026thinsp;=\u0026thinsp;4.174). Although the interaction term (ISS \u0026times; DDI) was not statistically significant (P\u0026thinsp;=\u0026thinsp;0.623), this combined stratification approach identified an \"ultra-high-risk\" subset (ISS\u0026thinsp;\u0026ge;\u0026thinsp;21 and DDI\u0026thinsp;\u0026ge;\u0026thinsp;5.5 mg/L), providing a clinically meaningful framework for risk refinement.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e3.9 Internal Validation and Clinical Utility Assessment of the Two-Step Strategy\u003c/h2\u003e \u003cp\u003eThe prediction model constructed for the Caprini high-risk population demonstrated strong internal validity and robustness. After 1,000 bootstrap resamples, the bias-corrected concordance index (C-index) of the two-step model reached 0.764, indicating stable and reliable discriminatory performance even after correcting for potential optimism. Calibration curve results (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA) showed high consistency between the predicted and observed VTE risks, characterized by a mean absolute error (MAE) of only 0.059 and a calibration slope of 0.838, suggesting no evidence of substantial overfitting.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eDecision curve analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB) further confirmed the clinical utility of this strategy. Across a wide and clinically relevant threshold probability range (10% to 60%), the two-step model incorporating DDI\u0026thinsp;\u0026ge;\u0026thinsp;5.5 mg/L consistently provided a superior net benefit compared to the clinical baseline model based on the Caprini score and basic demographic variables. Overall, these findings demonstrate that the proposed two-step strategy is not only statistically robust but also offers meaningful clinical benefit, supporting its potential for guiding precision VTE risk assessment and management in trauma patients.\u003c/p\u003e \u003c/div\u003e"},{"header":"4 Discussion","content":"\u003cp\u003eThis study provides a comprehensive evaluation of integrating the Caprini score with thrombosis-related biomarkers to optimize VTE risk stratification in trauma patients. Our investigation reveals that while the Caprini score serves as a foundational assessment tool, its independent discriminative capacity remains notably modest in acute trauma settings. Among the evaluated biomarkers intended to bridge this diagnostic gap, D-dimer consistently emerged as the most robust predictor of VTE. Crucially, however, we identified a paradoxical attenuation in the discriminatory power of all biomarkers once patients were categorized into the Caprini high-risk echelon. To overcome this diagnostic ceiling, we demonstrated that applying an optimized D-dimer threshold for secondary stratification successfully unmasks a subset of patients with profoundly elevated thrombotic risk. Collectively, these insights advocate for a paradigm shift toward a hierarchical, two-step assessment strategy, one that tightly couples baseline clinical scoring with dynamic biomarker refinement to guide precision thromboprophylaxis.\u003c/p\u003e \u003cp\u003eThe Caprini score is one of the most widely adopted tools for VTE risk assessment in surgical practice, integrating multiple clinical variables including age, operative factors, and comorbidities[\u003cspan additionalcitationids=\"CR28\" citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. However, its applicability in trauma populations remains controversial[\u003cspan additionalcitationids=\"CR31\" citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. In the present study, the AUC of the Caprini score was 0.640, indicating statistically significant but limited discriminative performance. This finding is consistent with the notion that trauma patients differ fundamentally from elective surgical populations. In trauma settings, patients are typically admitted in the acute phase following injury, often with high ISS scores and substantial physiological derangement. The pathophysiological response to trauma involves a complex interplay of systemic inflammation, activation of the coagulation cascade, endothelial disruption, and release of tissue factor, which collectively contribute to a rapid shift toward a hypercoagulable state[\u003cspan additionalcitationids=\"CR34\" citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. Importantly, these processes can occur independently of traditional risk factors captured by the Caprini score. As a result, a scoring system primarily based on static clinical variables may not adequately reflect the dynamic evolution of thrombotic risk in trauma patients. This mismatch likely contributes to the limited discriminative performance observed in this study, as well as the tendency for a large proportion of trauma patients to be classified within the high-risk category.\u003c/p\u003e \u003cp\u003eGiven these limitations, there has been increasing interest in incorporating laboratory-based indicators to enhance VTE risk assessment. Techniques such as thromboelastography (TEG) have been used to evaluate coagulation dynamics in trauma patients; however, their routine application is limited by operational complexity, cost, and limited availability in emergency settings[\u003cspan additionalcitationids=\"CR37\" citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. In contrast, thrombosis-related biomarkers offer several practical advantages, including rapid turnaround time, ease of measurement, and widespread accessibility[\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]. Among these, D-dimer has been extensively studied as a fibrin degradation product reflecting activation of the fibrinolytic system and has been widely used in the exclusion of VTE in low-risk populations[\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e]. In the present study, D-dimer demonstrated the highest predictive performance among the evaluated biomarkers, outperforming TAT, PIC, and TM. This finding suggests that D-dimer may serve as a more integrative indicator of thrombotic activity, reflecting both coagulation activation and subsequent fibrinolysis.\u003c/p\u003e \u003cp\u003eDespite this advantage, an important observation was that the predictive performance of all biomarkers declined when analysis was restricted to patients already classified as high risk by the Caprini score. This phenomenon highlights an inherent limitation of biomarker-based discrimination in trauma populations and can be conceptualized as a \u0026ldquo;floor effect\u0026rdquo;. Under conditions of trauma-induced hypercoagulability, the levels of thrombosis-related biomarkers tend to be broadly elevated across patients, resulting in a compressed distribution and reduced inter-individual variability. Consequently, the ability of these markers to further discriminate risk within an already high-risk population is diminished. This finding provides a potential explanation for the inconsistent performance of biomarkers reported in previous trauma studies and underscores the importance of considering baseline risk context when interpreting biomarker data.\u003c/p\u003e \u003cp\u003eTo address this limitation, the present study explored an integrated strategy combining clinical scoring with biomarker-based refinement. The addition of D-dimer to the Caprini score resulted in an improvement in AUC, indicating that biomarker information provides incremental predictive value beyond traditional clinical assessment. More importantly, by optimizing the D-dimer cutoff value using the Youden index, we identified a clinically meaningful threshold (5.5 mg/L) that enabled further stratification within the high-risk group. Patients exceeding this threshold exhibited a markedly higher incidence of VTE, defining an \u0026ldquo;extremely high-risk\u0026rdquo; subgroup. This approach effectively transforms the role of D-dimer from a general screening biomarker into a tool for risk re-classification within a predefined high-risk population.\u003c/p\u003e \u003cp\u003eFurther supporting this interpretation, our ISS-based analyses demonstrated a clear relationship between trauma severity and coagulation activation. We observed a significant positive association between ISS and D-dimer levels, along with progressively higher proportions of patients exceeding the D-dimer threshold across increasing ISS strata. Notably, VTE incidence also showed a stepwise increase with rising ISS, reaching over 50% among patients with ISS\u0026thinsp;\u0026ge;\u0026thinsp;25. These findings provide biological plausibility for the observed floor effect, suggesting that in patients with severe trauma, systemic coagulation activation becomes broadly elevated, thereby reducing the discriminatory capacity of individual biomarkers. Clinically, ISS-based re-stratification further refined risk gradients within the Caprini high-risk population. Patients with both severe trauma (ISS\u0026thinsp;\u0026ge;\u0026thinsp;21) and elevated D-dimer levels exhibited the highest observed VTE incidence, identifying an ultra-high-risk subgroup. Although the statistical interaction between ISS and D-dimer was not statistically significant, the combined stratification pattern highlights the value of incorporating anatomical injury severity into biomarker-guided risk assessment.\u003c/p\u003e \u003cp\u003eFrom a clinical perspective, this two-step strategy offers several potential advantages. First, it aligns with real-world decision-making processes, in which initial risk stratification is often followed by more refined assessment in selected patient subsets. Second, it provides a pragmatic framework for identifying patients who may benefit from intensified thromboprophylaxis, such as earlier initiation, higher dosing, or extended duration of anticoagulation. Third, by focusing on a subset of patients with the highest absolute risk, this approach may improve the efficiency of resource allocation while minimizing unnecessary intervention in lower-risk individuals.\u003c/p\u003e \u003cp\u003eThe robustness of this strategy was supported by internal validation and decision curve analysis. Bootstrap validation demonstrated stable model performance, suggesting that the observed predictive accuracy is unlikely to be solely attributable to overfitting. Calibration analysis further indicated good agreement between predicted and observed risks, reinforcing the reliability of the model. Importantly, decision curve analysis showed that the two-step model provides a greater net clinical benefit than conventional approaches across a broad range of threshold probabilities. This finding highlights the potential of the proposed strategy to improve clinical decision-making, particularly in settings where balancing the risks and benefits of thromboprophylaxis is critical.\u003c/p\u003e \u003cp\u003eSeveral limitations of this study should be acknowledged. First, the retrospective single-center design may limit the generalizability of the findings. Second, the sample size, particularly within subgroup analyses, may have reduced statistical power, as reflected in the lack of statistical significance in some multivariate analyses despite consistent effect directions. Third, missing data for certain biomarkers, especially TM, resulted in reduced effective sample sizes in some models, which may have influenced comparisons of model performance. Finally, the D-dimer threshold identified in this study was derived from a single cohort and requires external validation in larger, multicenter populations before it can be widely implemented in clinical practice. In addition, ISS-based subgroup analyses were conducted in relatively small sample subsets, which may limit the statistical power for detecting interaction effects.\u003c/p\u003e"},{"header":"5 Conclusions","content":"\u003cp\u003eIn conclusion, this study demonstrates that both clinical risk scoring and thrombosis-related biomarkers have inherent limitations in trauma populations, particularly within patients already classified as high risk. By introducing a D-dimer\u0026ndash;based re-stratification approach, we propose a two-step risk assessment model that improves risk discrimination and identifies a subgroup of patients at extremely high risk of VTE. This strategy provides a practical and clinically applicable framework for more precise risk assessment and individualized thromboprophylaxis in trauma patients, and may serve as a basis for future prospective validation and guideline development.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eVTE\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eVenous thromboembolism\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDVT\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eDeep vein thrombosis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePTE\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ePulmonary thromboembolism\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDDI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eD-dimer\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTAT\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eThrombin\u0026ndash;antithrombin complex\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePIC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ePlasmin\u0026ndash;α2-plasmin inhibitor complex\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTM\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eThrombomodulin\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eISS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eInjury Severity Score\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eROC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eReceiver operating characteristic\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eAUC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eArea under the curve\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eConfidence interval\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eOR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eOdds ratio\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eRR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eRelative risk\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCTV\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eComputed tomography venography\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDCA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eDecision curve analysis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eC-index\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eConcordance index\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMAE\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eMean absolute error\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eHIS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eHospital Information System\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eLIS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eLaboratory Information System\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eSOP\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eStandard operating procedure\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003e7.1 Ethics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was conducted following the Declaration of Helsinki and approved by the Biomedical Ethics Committee of Southern Medical University Zhujiang Hospital (Approval No.2025-KY-329-01). All participants provided written informed consent.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e7.2 Consent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e7.3 Availability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e7.4 Competing interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e7.5 Funding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e7.6 Authors' contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eQL, ML, and QZ contributed to study design, data collection, and statistical analysis. JH and BX assisted in data acquisition and database management. JS, YK, CG, and JY contributed to patient recruitment and clinical data verification. XX contributed to methodological guidance and interpretation of results. SZ and FM supervised the study and critically revised the manuscript. QL drafted the manuscript. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e7.7 Acknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank all the authors for their contributions to this manuscript. The graphical abstract was created in biorender.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eStreiff MB, Bockenstedt PL, Cataland SR, et al. Venous Thromboembolic Disease J Natl Compr Canc Netw. 2011;9:714\u0026ndash;77.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eToker S, Hak DJ, Morgan SJ. Deep vein thrombosis prophylaxis in trauma patients. Thrombosis. 2011; 2011: 505373.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRamli NN, Iberahim S, Mohd Noor NH, et al. Haemostatic, Inflammatory, and Haematological Biomarkers Among Orthopaedic Patients With Prolonged Immobilization and the Risk of Hypercoagulable States. Cureus. 2024;16:e51552.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWhiting PS, White-Dzuro GA, Greenberg SE, et al. Risk Factors for Deep Venous Thrombosis Following Orthopaedic Trauma Surgery: An Analysis of 56,000 patients. Arch Trauma Res. 2016;5:e32915.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eObi AT, Pannucci CJ, Nackashi A, et al. Validation of the Caprini Venous Thromboembolism Risk Assessment Model in Critically Ill Surgical Patients. JAMA Surg. 2015;150:941\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXie M, Zhou Z, Fei F, et al. Risk assessment models for venous thromboembolism in surgical inpatients: a multicenter retrospective cohort study. Eur J Med Res. 2025;30:1298.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHazeltine MD, Scott EM, Dorfman JD. An abbreviated Caprini model for VTE risk assessment in trauma. J Thromb Thrombolysis. 2022;53:878\u0026ndash;86.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eScott EM, Rios EC, Guber RD et al. VTE risk assessment in trauma patients: a comparative analysis of Caprini, Trauma Embolic Scoring System, and Greenfield Risk Assessment profile models in a single sample. J Thromb Thrombolysis 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBo H, Li Y, Liu G, et al. Assessing the Risk for Development of Deep Vein Thrombosis among Chinese Patients using the 2010 Caprini Risk Assessment Model: A Prospective Multicenter Study. J Atheroscler Thromb. 2020;27:801\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHanh BM, Cuong LQ, Son NT, et al. Determination of Risk Factors for Venous Thromboembolism by an Adapted Caprini Scoring System in Surgical Patients. J Pers Med. 2019;9:36.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHayssen H, Cires-Drouet R, Englum B, et al. Systematic review of venous thromboembolism risk categories derived from Caprini score. J Vasc Surg Venous Lymphat Disord. 2022;10:1401\u0026ndash;e14091407.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSwanson E. The Case against Chemoprophylaxis for Venous Thromboembolism Prevention and the Rationale for SAFE Anesthesia. Plast Reconstr Surg Glob Open. 2014;2:e160.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang L, Zhong J, Xiao D, et al. Thrombomodulin (TM), thrombin-antithrombin complex (TAT), plasmin-α2-plasmininhibitor complex (PIC), and tissue plasminogen activator-inhibitor complex (t-PAIC) assessment of fibrinolytic activity in postpartum hemorrhage: a retrospective comparative cohort study. Ann Transl Med. 2022;10:1273.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLundbech M, Krag AE, Christensen TD, et al. Thrombin generation, thrombin-antithrombin complex, and prothrombin fragment F1\u0026thinsp;+\u0026thinsp;2 as biomarkers for hypercoagulability in cancer patients. Thromb Res. 2020;186:80\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMusiał J, Pajak A, Undas A, et al. Thrombin generation markers and coronary heart disease risk factors in a Polish population sample. Thromb Haemost. 1997;77:697\u0026ndash;700.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHoppensteadt D, Tsuruta K, Cunanan J, et al. Thrombin generation mediators and markers in sepsis-associated coagulopathy and their modulation by recombinant thrombomodulin. Clin Appl Thromb Hemost. 2014;20:129\u0026ndash;35.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchaller J, Gerber SS. The plasmin\u0026ndash;antiplasmin system: structural and functional aspects. Cell Mol Life Sci. 2011;68:785\u0026ndash;801.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu Y, Ma J, Shi Q, et al. Quantitatively monitoring acute ischemic stroke patients post recombinant tissue plasminogen activator treatment. Health Sci Rep. 2021;4:e218.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAsakura H, Ontachi Y, Mizutani T, et al. An enhanced fibrinolysis prevents the development of multiple organ failure in disseminated intravascular coagulation in spite of much activation of blood coagulation. Crit Care Med. 2001;29:1164\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLoghmani H, Conway EM. Exploring traditional and nontraditional roles for thrombomodulin. Blood. 2018;132:148\u0026ndash;58.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePletsch-Borba L, Grafetst\u0026auml;tter M, H\u0026uuml;sing A, et al. Vascular injury biomarkers and stroke risk: A population-based study. Neurology. 2020;94:e2337\u0026ndash;45.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eConway EM. Thrombomodulin and its role in inflammation. Semin Immunopathol. 2012;34:107\u0026ndash;25.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAdam SS, Key NS, Greenberg CS. D-dimer antigen: current concepts and future prospects. Blood. 2009;113:2878\u0026ndash;87.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYe N, Liu Z, Wang X, et al. Evaluation of analytic and clinical performance of thrombin-antithrombin complex and D-dimer assay in prognosis of acute ischemic stroke. Blood Coagul Fibrinolysis. 2020;31:303\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThachil J, Lippi G, Favaloro EJ. D-Dimer Testing: Laboratory Aspects and Current Issues. Methods Mol Biol. 2017;1646:91\u0026ndash;104.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNicolaides AN, Fareed J, Spyropoulos AC, et al. Prevention and management of venous thromboembolism. International Consensus Statement. Guidelines according to scientific evidence. Int Angiol. 2024;43:1\u0026ndash;222.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCaprini JA. Risk assessment as a guide to thrombosis prophylaxis. Curr Opin Pulm Med. 2010;16:448\u0026ndash;52.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBahl V, Hu HM, Henke PK, et al. A Validation Study of a Retrospective Venous Thromboembolism Risk Scoring Method. Ann Surg. 2010;251:344\u0026ndash;50.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCronin M, Dengler N, Krauss ES, et al. Completion of the Updated Caprini Risk Assessment Model (2013 Version). Clin Appl Thromb Hemost. 2019;25:1076029619838052.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLobastov K, Barinov V, Schastlivtsev I, et al. Validation of the Caprini risk assessment model for venous thromboembolism in high-risk surgical patients in the background of standard prophylaxis. J Vasc Surg Venous Lymphat Disord. 2016;4:153\u0026ndash;60.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHazeltine MD, Guber RD, Buettner H, et al. Venous thromboembolism risk stratification in trauma using the Caprini risk assessment model. Thromb Res. 2021;208:52\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLobastov K, Urbanek T, Stepanov E, et al. The Thresholds of Caprini Score Associated With Increased Risk of Venous Thromboembolism Across Different Specialties: A Systematic Review. Ann Surg. 2023;277:929\u0026ndash;37.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGando S, Otomo Y. Local hemostasis, immunothrombosis, and systemic disseminated intravascular coagulation in trauma and traumatic shock. Crit Care. 2015;19:72.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStojkovic S, Kaun C, Basilio J, et al. Tissue factor is induced by interleukin-33 in human endothelial cells: a new link between coagulation and inflammation. Sci Rep. 2016;6:25171.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBodnar D, Bosley E, Raven S, et al. The nature and timing of coagulation dysfunction in a cohort of trauma patients in the Australian pre-hospital setting. Injury. 2024;55:111124.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaksaas-Aasen K, Gall LS, Stensballe J, et al. Viscoelastic haemostatic assay augmented protocols for major trauma haemorrhage (ITACTIC): a randomized, controlled trial. Intensive Care Med. 2021;47:49\u0026ndash;59.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHagemo JS, N\u0026aelig;ss PA, Johansson P, et al. Evaluation of TEG\u0026reg; and RoTEM\u0026reg; inter-changeability in trauma patients. Injury. 2013;44:600\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBurton AG, Jandrey KE. Use of Thromboelastography in Clinical Practice. Vet Clin North Am Small Anim Pract. 2020;50:1397\u0026ndash;409.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDi Nisio M, Squizzato A, Rutjes AWS, et al. Diagnostic accuracy of D-dimer test for exclusion of venous thromboembolism: a systematic review. J Thromb Haemost. 2007;5:296\u0026ndash;304.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShaw JR, Nopp S, Stavik B, et al. Thrombosis, Translational Medicine, and Biomarker Research: Moving the Needle. J Am Heart Assoc. 2025;14:e038782.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWeitz JI, Fredenburgh JC, Eikelboom JW. A Test in Context: D-Dimer. JACC. 2017;70:2411\u0026ndash;20.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"thrombosis-journal","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"thrj","sideBox":"Learn more about [Thrombosis Journal](http://thrombosisjournal.biomedcentral.com/)","snPcode":"12959","submissionUrl":"https://submission.nature.com/new-submission/12959/3","title":"Thrombosis Journal","twitterHandle":"@Thrombosis_J","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"venous thromboembolism, trauma center, Caprini score, D-dimer, risk stratification","lastPublishedDoi":"10.21203/rs.3.rs-9440559/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9440559/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eTo evaluate the integration of the Caprini score with thrombosis-related biomarkers for predicting venous thromboembolism (VTE) risk in trauma patients, aiming to establish a precision re-stratification strategy for high-risk populations.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eA retrospective cohort of 206 trauma surgical patients was analyzed. Caprini scores and baseline plasma biomarkers (DDI, TAT, PIC, TM) were assessed against VTE occurrence using ROC curves. Secondary risk re-stratification was performed within the Caprini high-risk cohort utilizing a Youden-optimized DDI threshold. Associations between trauma severity and coagulation activation were further evaluated using Injury Severity Score-based analyses.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eOverall VTE incidence was 31.1%. The Caprini score demonstrated modest discriminative capacity (AUC\u0026thinsp;=\u0026thinsp;0.640). Across the entire cohort, DDI exhibited the highest predictive efficacy among individual biomarkers (AUC\u0026thinsp;=\u0026thinsp;0.669, P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). However, within the Caprini high-risk subgroup (\u0026ge;\u0026thinsp;9 points, n\u0026thinsp;=\u0026thinsp;118), the discriminative power of all biomarkers experienced a ubiquitous decline (DDI AUC\u0026thinsp;=\u0026thinsp;0.614), reflecting trauma-induced biomarker saturation. Integrating DDI with the Caprini score elevated the combined model's AUC to 0.730. By applying an optimized threshold of DDI\u0026thinsp;\u0026ge;\u0026thinsp;5.5 mg/L, the high-risk cohort was successfully re-stratified, unmasking an \"extremely high-risk\" subset with a 44.8% VTE incidence. This represents a 2.78-fold risk increase compared to the general high-risk baseline (16.1%, P\u0026thinsp;=\u0026thinsp;0.005).\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eRelying solely on clinical risk scoring provides limited VTE discrimination in trauma patients. Implementing a hierarchical, two-step \"Caprini screening\u0026mdash;DDI re-stratification\" strategy effectively circumvents biomarker saturation, isolating extremely high-risk patients to provide a robust quantitative framework for precision thromboprophylaxis.\u003c/p\u003e","manuscriptTitle":"A Two-Step Risk Stratification Strategy Integrating Caprini Score and D-Dimer for Venous Thromboembolism in Trauma Patients","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-04 06:04:48","doi":"10.21203/rs.3.rs-9440559/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-04-22T16:50:08+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"44506117041327034599579777219621343775","date":"2026-04-21T10:16:19+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-21T10:00:09+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-21T00:19:02+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-21T00:18:21+00:00","index":"","fulltext":""},{"type":"submitted","content":"Thrombosis Journal","date":"2026-04-16T16:14:07+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"thrombosis-journal","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"thrj","sideBox":"Learn more about [Thrombosis Journal](http://thrombosisjournal.biomedcentral.com/)","snPcode":"12959","submissionUrl":"https://submission.nature.com/new-submission/12959/3","title":"Thrombosis Journal","twitterHandle":"@Thrombosis_J","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"1c934de2-f3c8-4fbf-9c1e-af8c421fef94","owner":[],"postedDate":"May 4th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-04T06:04:49+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-04 06:04:48","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9440559","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9440559","identity":"rs-9440559","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}