| Objective:To explore the risk factors related to post thrombotic syndrome(PTS)in patients with deep venous thrombosis(DVT),develop a prediction model and risk score for PTS,and provide individual prognostic information for DVT patients.Methods:A retrospective cohort study method was used.The hospitalization information of patients with lower extremity DVT who were treated in vascular surgery in Shanxi Bethune Hospital from June 2016 to December 2017 was collected,and the occurrence of PTS was followed up.The subjects were randomly divided into training set and validation set according to the ratio of 2: 1.The training set is used for the establishment and evaluation of the prediction model and risk score,and the validation set is used for internal verification of the prediction model and risk score.The calibration of the prediction model and risk score was evaluated by the calibration plot and the P value of Hosmer-Lemeshow goodness of fit test.The discrimination was evaluated by the ROC curve and the area under the ROC curve(AUC).Wilcoxon’s rank sum test was used to compare the prediction model and risk score differentiation in the same data set.Comparison of prediction models/risk score differentiation in different data sets was conducted by Mann-whitney U test.Results:1.808 patients were included in the analysis.The training set and the validation setonly have statistical differences in the distribution of rheumatoid immune disease,active cancer,and TT,there was no statistically significant difference in the distribution of other clinical variables between the two groups.2.Logistic regression results in training set showed BMI> 27.43 Kg / ㎡(OR =1.983;95% CI: 1.080 ~ 3.640;P = 0.027),iliofemoral deep venous thrombosis(OR =3.621;95% CI: 1.988 ~ 6.595;P <0.001),active cancer(OR = 4.250;95% CI: 2.003 ~9.021;P <0.001),history of DVT(OR = 3.509;95% CI: 1.597 ~ 7.711;P <0.001),no short-term risk factors(OR = 2.216;95% CI: 1.229 ~ 3.996;P = 0.008),varicose veins(OR = 3.617;95% CI: 1.860 ~ 7.032;P <0.001),no elastic support(OR = 2.324;95% CI:1.335 ~ 4.043;P = 0.003)is an independent risk factor for patients with DVT to develop into PTS.The Hosmer-Lemeshow goodness-of-fit test of the prediction model shows that P= 0.535,the model fits well;the area under the ROC curve is 0.785(95% CI: 0.728-0.842),P <0.001,and the model has a good discriminative ability.3.According to the regression coefficient β value of each variable in the model,assign a value and establish a risk scoring system with a total score of 0-11.Among them,BMI > 27.43 Kg / ㎡ is assigned 1 point;iliofemoral deep venous thrombosis is assigned 2 points;active cancer is assigned 2 points;DVT history is assigned 2 points;no transient risk factors are assigned 1 point;varicose veins are assigned 1 point;no elastic support is assigned.1 point.For the risk score,a total score of 4 was used as the cut-off value,and ≥4 was divided into a high-risk group of PTS.The area under the risk score ROC curve was 0.789(95% CI: 0.682 to 0.775),P <0.001,and the ability to discriminate the risk score was good.The sensitivity of the risk score in the training set was 69.74%,the specificity was 75.32%,and the crude consistency rate was 53.64%.The Wilcoxon Signed Rank Test results of the area under the ROC curve of the risk score and the prediction model in the training set showed that P=0.4497>0.05,and the difference between the prediction model and the risk score was not statistically significant.4.In the validation set the Hosmer-Lemeshow goodness-of-fit test of the prediction model shows that P = 0.591,the model fits well;the area under the ROC curve is 0.789(95% CI: 0.716 ~ 0.862),P <0.001,the discriminative ability is good;In the validationset the area under the ROC curve of the risk score is 0.804(95% CI: 0.734 to 0.874),P<0.001,and the discriminative ability of the risk score is good.The Wilcoxon Signed Rank Test results of the area under the ROC curve of the risk score and the prediction model in the validation set showed that P=0.0631>0.05,and the difference between the prediction model and the risk score was not statistically significant.5.The Mann-Whitney U test of the area under the ROC curve of the prediction model in the training set and the validation set showed P=0.217>0.05,and the difference of the prediction model in the two data sets was not statistically significant.The Mann-Whitney U test of the area under the ROC curve of risk score in the training set and the validation set showed that P=0.237>0.05,and the difference of risk score in the two data sets was not statistically significant.Conclusion:The prediction model and risk score can help clinicians identify high-risk groups who may develop PTS in DVT patients,and then monitor them closely or adopt active treatment strategies.However,external validation is still required before applying this risk score to the clinic. |
PDF Full Text Request