Analysis Of Factors Associated With Deep Vein Thrombosis In Stroke Patients

ObjectiveTo investigate the factors associated with deep vein thrombosis(DVT)after stroke,to establish a predictive model for DVT in stroke patients,and to help clinicians identify people at high risk of developing DVT after stroke at an early stage,and to provide timely intervention and treatment.MethodsTwo hundred and twenty stroke patients who were hospitalized in the Department of Rehabilitation Medicine of our hospital since January 2019 to August2022 and met the inclusion criteria were gathered and grouped into the DVT group(n=109)and the no-thrombosis group(n=111)according to whether the patients were complicated by deep vein thrombosis(DVT)or not.All variables were then entered into the LASSO regression model for screening of the characteristic variables;the intersection of the two models and the variables that were also consistent with the one-way analysis of variance were selected.The variables that were considered with statistical significance(p<0.1)were included in the multi-factorial logistic regression logistic regression analysis to construct a predictive model for DVT in stroke patients.Data were statistically analysed using the SPSS 23 software package and Extreme Wisdom Analysis software.Results(1)There were 109 in the group of deep vein thrombosis and 111 in the group of no thrombosis.Comparison of the baseline data between the two groups showed statistically significant differences in Age,Barthel index,gender,hypertension,surgical history,days in hospital,Hct,lymphocyte ratio,TBi L,MON,Apo A-I,D-dimer,Hcy,CRP,creatinine,LMR and NLR(p< 0.1).(2)In this study,Lasso regressor was used for variable importance analysis and the 10 variables with the highest importance(from highest to lowest)were selected as follows: D-dimer,homocysteine,age,Apo A-I,SD,venous blood glycated haemoglobin,Barthel index,Hct,TBIL,and serum indirect bilirubin.(3)In this study,all variables were entered into the LASSO regression model and the standard error of the minimum distance was taken to be λ of 0.057,corresponding to the choice of variables for the model: D-dimer + Hcy + Apo A-I +Barthel index + age.(4)The final variables included in the DVT prediction model for stroke patients in this study were: D-dimer,Hcy,Apo A-I,Barthel Index > 60 and age ≥ 65 years.Conclusions(1)D-dimer,Hcy,and age ≥65 years are risk factors for DVT in stroke patients.(2)Barthel index >60 and Apo A-I are protective factors for the development of DVT in stroke patients.