Abstract: Objective   By optimizing and improving the traditional RAPT score,the screening accuracy of thrombosis tendency is improved,and the relevant basis for thrombosis prevention in patients with early traumatic fracture is provided,so as to select an efficient individualized anticoagulation therapy.  Methods   The clinical data of 300 patients with traumatic fracture admitted to Shiyan Taihe Hospital from January 2021 to December 2021 were retrospectively analyzed.According to whether venous thromboembolism (VTE) occurred during hospitalization,they were divided into thrombosis group(150 cases) and non-thrombosis group(150 cases).In addition to the four categories included in the traditional RAPT score,the history of smoking and drinking,diseases of vascular endothelial injury,time from injury to admission,preoperative prophylactic anticoagulation measures,upper limb fractures or simple lower limb fractures,type of anesthesia,emergency treatment,blood biochemical index,and dynamic data of D-dimer were also collected.The clinical data of the two groups were compared,and the independent risk factors were screened by univariate and multivariate logistic regression analysis to formulate the optimized RAPT score.The traditional and optimized RAPT scores during hospitalization were dynamically evaluated.The ROC curve was used to evaluate the efficacy of traditional and optimized RAPT scores in the evaluation of traumatic fracture thrombosis.  Results    Single factor analysis showed that there were statistically significant differences between the two groups in smoking history, time from injury to admission, anesthesia method, VTE history, vascular endothelial injury diseases (diabetes, coronary heart disease, hyperlipidemia, cerebrovascular diseases, varicose veins of lower limbs or arteriosclerosis oblast), no use of preventive antithrombotic measures before surgery, and central venous catheterization (P<0.05). Logistic regression analysis after adding age and gender factors showed that with the increase of age, people with smoking history, diabetes and cerebrovascular disease increased the incidence of thrombosis. Men with a history of smoking, diabetes, hyperlipidemia, and women with pre-operative anticoagulation and simple upper or lower limb fractures are increased risk of thrombosis.The AUC of traditional RAPT score at admission,before operation and after operation were 0.636 1,0.639 5 and 0.649 2,respectively.The optimized RAPT scores were 0.806 4,0.806 7 and 0.803 1.  Conclusion   In this study,the predictive value of the optimized RAPT score for thrombosis at admission,before operation and after operation was 15%-17% higher than that of the traditional RAPT score.The optimized RAPT score has a higher predictive value and is more conducive to early detection of thrombosis risk and early treatment than the traditional score.

Key words: traumatic, fracture, venous thromboembolism, analysis of regression, risk assessment tool