| Objective:Venous thromboembolism(VTE)is a common complication of cancer.The mortality of patients with cancer is increased by VTE.Preventive anticoagulation can reduce the incidence of VTE,but first,it is necessary to screen high-risk patients of VTE through the risk assessment model(RAM).The Padua Prediction Score(PPS)and the Khorana score(KRS)are commonly used RAMs for VTE in department of medical oncology.However,the predictive ability of the previous RAMs in medical oncology patients is still questioned.In this paper,we attempt to propose a RAM of VTE for in-patients at all stages of treatment in department of medical oncology,and compare it with the PPS and KRS in this population.To evaluate the ability of each model to predict the risk of VTE and its clinical feasibility.Methods:In this retrospective case-control study,63 patients with cancer who had VTE were enrolled in the Department of Oncology of the Cancer Hospital Of China Medical University from July 2017 to March 2019 as the case group(VTE group).We randomly selected 189 patients without VTE as the control group(non-VTE group),using a 1:3 ratio.We collected patients’basic information,medical history,laboratory results and other data,and performed univariate and multivariate Logistic regression analysis on the risk factors of cancer-related VTE,including coagulation-related biomarkers.According to the actual clinical significance of risk factors and previous literature,the significant risk factors in Logistic regression analysis were screened and supplemented to determine the risk factor items of the new RAM.According to the regression coefficient in the multivariate logistic regression analysis and taking into account the clinical feasibility,the scores of each risk factor were determined,and a new RAM was proposed.Receiver operating characteristic(ROC)curve was used to determine the cutoff value for the new RAM.By calculating the area under receiver operating characteristic curve(AUC),sensitivity,specificity,positive likelihood ratio and other indicators,and analyzing the distribution of patients in the RAM,To compare the predictive ability and clinical feasibility of the new RAM with PPS and KRS.Results:The following five risk factors were included in the new RAM by Logistic regression analysis:reduced mobility,previous VTE,stage IV cancer,platelet count≥350x10~9/L,and D-dimer>0.55mg/L.The cutoff value for high-risk of the model was set to≥3 points.In this study,the AUC of the new RAM,PPS and KRS were 0.824,0.612 and 0.649,respectively.The sensitivity of the new RAM was 92.1%,the specificity was 64.0%,the positive likelihood ratio was 2.558,the negative likelihood ratio was 0.123,the positive predictive value was 71.9%,and the negative predictive value was 89.0%.The sensitivity and specificity of PPS were 31.7%and79.9%,respectively.The sensitivity and specificity of KRS were 9.5%and 92.2%,respectively.The odds ratios of high risk to VTE in the new RAM,PPS,and KRS were 20.641,1.848,and 2.105,respectively.In PPS,“Already known thrombophilic condition”could not be detected in our hospital.The patients with“Acute myocardial infarction or ischemic stroke,Ongoing hormonal treatment and Body mass index(BMI)≥30kg/m~2”,etc were less distributed.In KRS,there was no distribution of patients with BMI≥35kg/m~2.Each risk factor group of the new RAM was distributed among patients,and the number of patients in the case group was higher than that in the control group.Conclusions:In this study,we proposed a new RAM of VTE based on a population of medical oncology inpatients with various cancers throughout their entire treatment.As risk factors in the model,the new RAM includes a effective coagulation-related biomarker:D-dimer which is easily measurable in clinical practice.In this population,the new RAM is superior to PPS and KRS in predicting VTE,and is simpler and more feasible,but needs validation. |
PDF Full Text Request