Development And Validation Of Predictive Model For Lower Extremities Deep Vein Thrombosis Of Patients With Acute Ischemic Stroke

Objectives To examine the independent risk factors for acute ischemic stroke combined with lower extremity deep vein thrombosis(DVT)and to construct a nomogram prediction model to predict the risk of lower extremity deep vein thrombosis in acute ischemic stroke.The nomogram was used to predict the risk of DVT associated with acute ischemic stroke.Methods Data related to DVT examinations of patients with acute ischemic stroke treated at the Department of Neurology of the Hebei General Hospital between January2019 and August 2020 were collected,and patients collected between January 2019 and December 2019 were used as a model group for predictive modeling.Patients were collected between January 2020 and August 2020 as a validation group for validating the clinical predictive value of the model.Patients who met the inclusion and exclusion criteria were screened in strict accordance.Relevant therapeutic measures,ancillary test results,and general data of the enrolled patients were collected in conjunction with current studies on risk factors in patients with acute ischemic stroke.The data collected were statistically analyzed and multivariate logistic regression was used to screen for independent risk factors for lower extremity deep vein thrombosis in patients with acute ischemic stroke.The R language software analyzes the incorporated risk factors and can assign different scores to different factors.The nomogram were drawn,allowing calculation of the specific probability of developing lower extremity deep vein thrombosis in acute ischemic stroke.The prediction models to validate by differentiation and calibration degree,for the modeling group and the validation group,the differentiation degree was evaluated by plotting the receiver operating characteristic(ROC)curve of the prediction models separately,and the area under the curve(AUC)was calculated comprehensively.The AUC and the discrimination of the prediction models were quantified with specific values to determine the difference and the degree of variation among the prediction models.The calibration of the modeling group and the validation group were tested by the HosmerLemeshow goodness-of-fit test,and the calibration of the prediction model judged by the magnitude of the P-value to validate the clinical predictive efficacy of the nomogram in predicting lower extremity deep vein thrombosis in patients with acute ischemic stroke.Results 1 A total of 412 patients with acute ischemic stroke are included in the study.291 patients are included in the modeling group for prediction modeling,and 121 patients are in the validation group to verify the clinical suitability of the newly constructed prediction model.2 Univariate analysis results indicate that patients’ age over 70 years old,basal ganglia infarction,parietal infarction,D-dimer,fibrinogen,prothrombin time,platelet count,muscle strength<2,and impaired consciousness were statistically significant(P<0.05).3 The results of the multifactorial logistic analysis shows that age over 70 years old,basal ganglia infarction,D-dimer,fibrinogen,muscle strength below grade 2,and platelet count are statistically different,and the above six factors are independent risk factors for the formation of lower extremity DVT in patients with acute ischemic stroke.4 A nomogram are drawn,allowing calculation of the specific probability of developing acute ischemic stroke complicated by lower extremity deep vein thrombosis.5 The results of the differentiation evaluation shows AUC values of 0.769 and 0.803 for the modeling and validation groups,respectively,with higher AUC values for both groups,indicating that the new prediction model has better differentiation ability to predict lower extremity deep vein thrombosis.The results of calibration evaluation for the modeled and validation populations showed that the P values were greater than 0.05,and the P values were 0.735 and 0.786 for the modeled and validation groups,respectively,indicating that the new prediction model has higher predictive accuracy in predicting lower extremity deep vein thrombosis.Conclusions 1 An individualized predictive model for acute ischemic stroke patients allows for screening and early identification of patients with DVT.2 The quantified risk parameters for DVT in patients with acute ischemic stroke can be used as a reference for the early administration of DVT prophylaxis in acute ischemic stroke.Figure 5;Table 4;Reference 118...