Establishment Of Predictive Model Of DVT Formation And Analysis Of Its Value In General Surgical Patients

Objective: Explore the value of thromboelastogram and conventional coagulation indexes in predicting the formation of deep venous thrombosis（DVT）of lower limbs in patients undergoing general surgery,then establish a prediction model.Methods: 272 patients undergoing general surgery and meeting the inclusion and exclusion criteria in Taixing people’s Hospital Affiliated to Bengbu Medical College from January 2018 to January 2021 were selected as the research objects.Subsequently,they were divided into DVT group and non DVT group based on the results of postoperative venous color Doppler ultrasound.Conventional coagulation indexes include prothrombin time（PT）,activated partial thromboplastin time（APTT）,thrombin time（TT）,fibrinogen（FIB）,D-dimer.Thromboelastogram parameters: reaction time（R）,clotting time（K）and coagulation angle（α Angle）,maximum amplitude（MA）.Patients were divided into three groups according to operation time: with 140 cases in group A（operation time < 1 hour）,90 cases in group B（operation time 1 ～ 2 hours）and 46 cases in group C（operation time >2 hours）.The incidence of postoperative DVT,as well as the surgery duration was compared among these three groups（A,B,C）.Subsequently,the clinical data,conventional coagulation indexes,thrombo-elastogram results and imaging data of patients were analyzed retrospectively.SPSS 24.0 was adopted to carry out t-test for those consistent with the normal variance in the measurement data,approximate t-test was adopted for those with uneven variance,and Chi square test was adopted for the comparison of counting data.Furthermore,R version 4.1.0 was adopted to draw the ROC curve,and the Med Calc20.0 was adopted to calculate the area under the curve（AUC）,cut-off value,sensitivity,specificity and Youden index.Results: DVT occurred in a total of 23 patients after the surgery,and thrombosis generally occurred within 10 days after the surgery,with a total incidence rate of 8.4%,of which 20 cases（86.9%）had no obvious discomfort.All patients were recovered and discharged from hospital after clinical treatment.The preoperative R value,PT value,TT value and APTT value of patients in DVT group were lower than those in non DVT group（P < 0.05）,and the K value was higher than those in non DVT group（P < 0.05）,furthermore,the α angle,MA value,age and surgery time of DVT group were higher than those of non DVT group（P < 0.05）,and the level of D-dimer was higher than that of non DVT group（P < 0.05）.There was no significant difference in terms of gender,body mass index,hypertension,diabetes,smoking,drinking and FIB between these two groups（P >0.05）.In terms of surgery duration,DVT occurred in 2 cases among 140 cases in group A,with an incidence of 1.4%;DVT occurred in 3 cases among 90 cases in group B,with an incidence of 3.3%;DVT occurred in 8 cases of 46 cases in group C,with an incidence of 17.4%;The incidence of DVT in group C was significantly higher than that in group A and group B（P < 0.05）.The receiver operating characteristic（ROC）curve was drawn according to whether the patient had DVT after surgery.According to the result,the αAngle,MA,K,D-dimer and PT before the surgery are of medium predictive significance for the occurrence of DVT.The area under the curve（AUC）is 0.7 ～ 0.9.The predictive value of R,APTT,TT and FIB is comparatively low（AUC < 0.7）.Incorporating indexes with statistical significance in univariate analysis into logistic regression,the prediction model of thrombosis will be established.ROC curve: AUC 0.964（95% CI 0.934.-0.983）,P < 0.01,cut-off value = 0.174,with sensitivity of 91.3%,and specificity of 95.18%.Conclusion: 1.In terms of patients undergoing general surgeries,the older they are,the higher the incidence rate of postoperative thrombosis;2.The incidence of postoperative DVT is significantly higher in the patients with longer operation time;3.Nine preoperative variables including age,surgery time,PT,APTT,TT,D-dimer,and R,α,Ma index,in the established prediction model were incorporated,which can effectively sort out the population with high-risk of developing DVT.