Risk Factor Analysis And Risk Prediction Model Of Venous Thromboembolism In Patients With Lung Cancer

Objectives:This study intends to identify the risk factors of VTE in patients with lung cancer,and compare the predictive value of three commonly used VTE risk assessment models for the formation of VTE in patients with lung cancer,and establish an individualized VTE prediction model for patients with lung cancer.Methods:(1)The literature review method and the research group discussion were used to preliminarily summarize the risk factors of lung cancer patients complicated with VTE to form a questionnaire.(2)Through a retrospective case-control study method,the inpatients with stage III and IV lung cancer who met the inclusion and exclusion criteria from January 2017 to July 2021 were retrospectively reviewed.Among them,102 patients in the VTE group were used as the case group,and 258 patients in the non-VTE group were randomly selected.Humans were used as the control group,and univariate and multivariate Logistic regression was used to analyze the risk factors of VTE in lung cancer patients.(3)The risk assessment models of Caprini,Padua and Khorana were used to score lung cancer patients,the area under the ROC curve,sensitivity and specificity of the three thrombosis risk assessment models were compared,and the high-risk cut-off values of the three thrombosis risk assessment models were determined.(4)According to the logistic regression analysis results,a risk prediction model for lung cancer patients complicated with VTE was established,and its predictive value was evaluated and compared with the risk assessment models of Caprini,Padua and Khorana.Results:(1)After literature review and discussion by the research group,the questionnaire for patients with lung cancer complicated by VTE consists of 5 parts and 22 risk factors.(2)Univariate analysis was performed on 22 variables,and the results showed that there were 15 variables with statistical significance at P<0.05,namely: bed rest time ≥ 3 days(P<0.001),clinical stage IV(P=0.009),pathological The type was adenocarcinoma(P=0.014),surgical history(P=0.019),central venous catheterization(P<0.001),D-dimer ≥ 1.5mg/L(P<0.001),serum albumin ≥ 35g/L(P < 0.001),fibrinogen≥4 g/L(P=0.038),history of VTE(P<0.001),pulmonary infection(P<0.001),number of cardiovascular complications≥2(P<0.001),respiratory failure or heart failure(P<0.001),lower extremity varicose veins(P=0.021),history of blood transfusion(P<0.001),Khorana risk assessment model score≥2 points(P<0.001).(3)The 15 variables with P<0.05 in the univariate analysis were included in the multivariate Logistic regression analysis,and the results showed: bed rest time≥3days(P=0.036),clinical stage IV(P=0.046),central venous placement tube(P=0.004),D-dimer≥1.5mg/L(P<0.001),serum albumin≥35 g/L(P=0.011),history of VTE(P<0.001),number of cardiovascular complications≥2(P=0.020),Khorana risk assessment model score≥2(P<0.001)8 independent risk factors.(4)The thrombosis risk scores of Caprini,Padua and Khorana risk assessment models in lung cancer patients were(5.553±2.677),(4.591±2.006)and(1.611±0.850)points,respectively.Area under the ROC curve of Caprini,Padua and Khorana risk assessment models They were(0.756±0.028),(0.693±0.031),and(0.746 ±0.027)respectively.The high-risk cut-off values of Caprini,Padua and Khorana were6 points,6 points,and 2 points,respectively,the sensitivity was 65.69%,51.96%,73.53%,and the specificity was 76.36%,83.72%,72.09%,respectively.The Youden indices were 0.420,0.357,and 0.456,respectively.(5)According to the results of Logistic regression analysis,the risk prediction model of VTE in lung cancer patients was established as Y=4×VTE history+2 ×D-dimer≥1.5mg/L+2×Central venous catheterization+2×Whether or not Khorana risk assessment model score≥2 points+1×whether bed rest≥3 days+1×whether the number of cardiovascular complications ≥ 2+1 × whether serum albumin ≤ 35g/L+1×whether clinical stage is IV.Scores < 4 were classified as low thrombosis risk patients,and≥4 were considered as high thrombosis risk patients.The area under the ROC curve was(0.880±0.019).The sensitivity,specificity and Youden index were90.2%,70.7% and 0.609,respectively.Compared with the three risk assessment models of Caprini,Padua and Khorana,the area under the ROC curve,the Youden index,and the sensitivity of the predictive model were higher than those of the three risk assessment models,and the specificity was lower than that of the three risk assessment models.Conclusions:(1)bed rest time≥3 days,clinical stage IV,central venous catheterization,D-dimer≥1.5mg/L,serum albumin≥35g/L,history of VTE,number of cardiovascu-lar complications≥2,Khorana Risk assessment model score≥2 was an independ-ent risk factor for venous thromboembolism in patients with lung cancer.(2)The prediction accuracy of Caprini and Khorana risk assessment model for the occurrence of VTE in lung cancer patients was at a moderate level,and the prediction accuracy of Padua risk assessment model for the occurrence of VTE in lung cancer patients was poor.(3)The prediction accuracy of the prediction model established in this study for the occurrence of VTE in lung cancer patients is higher than that of the risk assessment models of Caprini,Padua and Khorana.The established prediction model is simple and practical,and can be applied to clinical work,enabling medical staff to detect patients with high risk of VTE in a timely manner.Take precautionary measures.