| Objectives: To explore the distribution characteristics of multilevel visceral vascular calcification in hemodialysis patients,establish a model to predict their mortality and evaluate its value.Methods: A retrospective cohort study was established,which included patients on regular hemodialysis in the Blood Purification Center of Zhongda Hospital Affiliated to Southeast University from January 2015 to September 2017.The calcification volume of abdominal aorta and its branches in abdominal computed tomography images were quantitatively evaluated.Then the distribution characteristics of arterial calcification in the abdomen were described.Patients were followed up to death or until September 2020.IBM SPSS Statistics 26.0(IBM Corporation,Armonk,NY,USA)and R version 4.1.2(R package car,rms,p ROC,time ROC,gg DCA,glmnet,VIM,pec,dplyr,survival)were used for data analysis.Kaplan-Meier methods were applied to analyze the survival of patients.All of the patients were randomly divided into the training set and testing set in a ratio of 7:3.Univariate Cox regression analysis and Least Absolute Shrinkage and Selection Operator(Lasso)regression analysis were used to screen the prognostic factors in the training set.Multivariate Cox proportional risk regression and nomograms were used to construct a predictive model.The model was verified and evaluated in the training set and testing set respectively.Results: According to inclusion and exclusion criteria,237 patients(125 males and 112 females)were included with a median age of 62(52,72)years and a median dialysis vintage of 42.00(15.50,93.00)months.They were divided into mild calcification group and severe calcification group according to the median of total volume of abdominal vascular calcification.Compared to the mild calcification group,patients with severe calcification are older,have longer dialysis vintage and complicated with heart failure,cardio-cerebrovascular disease,cardiac valve disease and atrial fibrillation(P<0.05).In hemodialysis patients,the prevalence of abdominal vascular calcification is95.36%,and the prevalence of abdominal aortic calcification is the highest(88.61%).However,the overall prevalence of calcification in branches(94.51%)is higher than that in abdominal aorta,and the prevalence of internal iliac artery calcification is the highest(85.65%)among branches.The median survival time of the patients was 37.00 months(95% Confidence Interval(CI)29.05-44.95),and 137 patients(57.81%)died,most of which died of cardio-cerebrovascular diseases(54.01%).In the mild calcification group,the cumulative survival rate of 1-5 years is higher than that of the severe calcification group.The area under the time-dependent Receiver Operator Characteristic(ROC)curve for the calcification of the internal iliac artery and mesenteric artery is larger than that of other arteries,and the area under the time-dependent ROC curve for the calcification of the renal artery is the lowest.A total of 237 patients were divided into training set(165 cases)and testing set(72 cases)by the ratio of 7:3.Univariate Cox regression shows that age,body mass index,body surface area,diastolic blood pressure,the history of diabetes,heart failure,cardio-cerebrovascular disease and cardiac valve disease,the use of angiotensin converting enzyme inhibitor/angiotensin receptor blocker and nitrates,serum albumin level,alkaline phosphatase,fasting glucose,serum creatinine,abdominal aortic calcification,mesenteric artery calcification,renal artery calcification,common iliac artery calcification and internal iliac artery calcification are associated with prognosis of hemodialysis patients in the training set(P<0.05).Lasso regression shows that age,the history of cardiocerebrovascular disease,serum albumin level and internal iliac artery calcification are the best prognostic factors for hemodialysis patients.Multivariate Cox regression model shows internal iliac artery calcification(Hazard Ratio(HR)=1.503,95%CI: 1.214-1.862,P<0.001),serum albumin level(HR=0.911,95%CI: 0.875-0.950,P<0.001)and the history of cardio-cerebrovascular diseases(HR=1.735,95%CI: 1.102-2.734,P=0.017)are independently correlated with the death of hemodialysis patients.A prediction nomogram model is constructed to predict the prognosis of hemodialysis patients based on age,serum albumin level,the history of cardio-cerebrovascular disease and calcification volume of internal iliac artery.The time-dependent ROC curve shows that the model had a high discriminability in predicting 3-year and 5-year survival.The calibration curve shows that the model has a high calibration degree in predicting 1 year and 3 years survival.The clinical decision curve analysis shows that the model has a high value of clinical application.Conclusion: The prevalence of abdominal vascular calcification in hemodialysis patients is extremely high and the prevalence of branches of abdominal aorta is higher than abdominal aorta itself.Internal iliac artery calcification,low albumin levels and the history of cardio-cerebrovascular disease are independent risk factors of death in hemodialysis patients.The nomogram constructed according to the factors to predict the prognosis of hemodialysis patients is of high value. |
