Establishment Of Nomogram For Predicting Portal Vein Thrombosis In Cirrhotic Patients With Esophageal Varices

Background/Aims:Portal vein thrombosis(PVT)is a common complication of liver cirrhosis that can increase portal vein pressure.At present,domestic and foreign scholars generally believe that hepatic venous pressure gradient(HVPG)reflects the risk of esophageal variceal bleeding.When it is greater than 12mmhg,the risk of bleeding will be greatly increased.Once the varices rupture and bleed,the mortality rate can reach 15-20%within 6 weeks.In the entire natural history of liver cirrhosis,50%of patients will develop esophageal varices.If there is no effective intervention,the degree of varices will continue to increase.And,severe esophageal varices are often accompanied by slowing of portal blood flow,causing venous blood stasis and increasing the risk of PVT formation.As a result,esophageal varices and PVT may form a vicious circle,influencing the prognosis of patients with liver cirrhosis.In addition,in order to better eliminate esophageal varices,patients need to be hospitalized several times in a short period of time for regular endoscopic treatment.However,the formation of acute or severe PVT may directly or indirectly affect the efficacy of endoscopic varicose vein treatment.Therefore,doctors often need to conduct regular imaging examinations of the portal vein,and frequent examination not only wastes medical resources but also increases the financial burden of patients.Our research aims to identify the risk factors for Portal vein thrombosis(PVT)and develop a nomogram to predict the risk of PVT in cirrhotic patients with esophageal varices.The Nomogram was used to quantitatively predict the risk of PVT in patients with cirrhosis of esophageal varices.Necessary imaging examinations were performed for high-risk patients to detect the formation of PVT early and provide guidance for early clinical decision-making and intervention.MethodsA total of 283 patients with liver cirrhosis and esophageal varices who were treated in this hospital from December 2013 to December 2018 were admitted.Then,patients were divided into PVT group(n=119)and non-PVT group(n=164)based on imaging examinations.Continuous variables with a normal distribution were expressed as mean±SD and compared using the t-test.Distributed skewed variables are expressed as median(25th,75th percentage)and compared using the Mann-Whitney U test.Categorical variables are expressed in terms of numbers and percentages and are tested by chi-square or fisher’s exact test.Multivariable logistic regression models were established to find independent risk factors for portal vein thrombosis A nomogram for PVT was developed according to the multivariate regression model,The concordance index(C-index)and calibration curves were used to verify the performance of this nomogram.A two-tailed P value of<0.05 was considered statistically significant.ResultsThe univariate analysis indicated that the child-pugh class(χ2=9.388,P=0.009),history of splenectomy(χ2=26.805,P<0.001),WBC(Z=-2.248,P=0.025),PLT(Z=-3.323,P=0.001),D-dimer(Z=-6.236,P<0.001),and splenic thickness(Z=-2.432,P<0.001)in The PVT group were significantly higher than non-PVT group,while TG level(Z=-4.150,P<0.001)was significantly and inversely associated with PVT.Further,multivariable logistic regression showed TG(OR 0.441,95%CI 0.190-0.889),D-dimer(OR 1.158,95%CI 1.043-1.285),PT(1.160,95%Cl 1.025-1.313),and a history of splenectomy(OR 2.933,95%1.164-7.389)were independent risk factors of PVT.Among all the variables,the highest Vif is 2.23,and the risk of multicollinearity is lower.Hosmer-Lemeshow test showed that the multivariate logistic regression fitted well with the data(P=0.711).Besides,a nomogram was also constructed based on these risk factors to predict the PVT.The concordance index was 0.745.The calibration plot of PVT risk rates revealed good agreement between the observed and predicted values.ConclusionsUnivariate analysis showed that the levels of child-pugh grade,splenectomy history,WBC,PLT,D-dimer and spleen thickness in PVT group were higher than those in non-PVT group,while the level of TG in PVT group was lower than that in non-PVT group.Multivariate analysis showed that the decrease of TG level,the increase of D-dimer level,the prolongation of PT and the history of splenectomy were independent risk factors for the formation of PVT in cirrhotic patients with esophageal varices.Based on the results of multivariate regression,the C index value of the line chart is 0.745.The calibration curve shows that there is a good consistency between the observed and predicted values of PVT,which provides a quantitative and intuitive tool for clinicians to evaluate the risk of PVT.