Epidemiological Survey And Development Of Mortality Risk Prediction Model For End Stage Renal Disease Patients In Zhejiang Province

Part I Epidemiological Survey of Incident End Stage Renal Disease Patients in Zhejiang Province from2008to2013Objective:Dialysis registration is an important component for standardized management of dialysis patients. Zhejiang Dialysis Quality and Management Center(ZDQM) was built in2007. It includes basic profile of dialysis patients, clinical data, medical treatment data, laboratory data and so on. Our aim was to analyze the epidemiological literature in incident end stage renal disease(ESRD) patients in Zhejiang province, which in turn to provide epidemic data to doctors, researchers, and improvement and formulation of health insurance policy.Method:We retrospectively review all incident ESRD patients registered in the ZDQM from2008-2013. The demographic information, clinical and laboratory data were collected. Their incidence of ESRD patients, primary causes changes in them, stracture of sex and gender, and cause of mortablity were investigated.Results:26310incident ESRD patients were included (19143maintenance hemodialysis patients, and5199maintenance peritoneal dialysis patients). The incidence rate of ESRD from2008to2013was46.3to119.9pmp. The average age was55.7±16.1years old, from2008to2013was53years old to58.7years old, the proportion of patients over65years old increased from25.1%to37.8%of incident ESRD.Male to female was1.3vs1.0. The first three primary cause of ESRD were chronic glomerulonephritis (CGN)(51.3%), diabetic nephropathy(DN)(17.3%) and hypertensive nephrosclerosis(HTN)(6.4%).The ratio of DN was growing year by year. In ESRD patients, the cardiovascular mortality was the leading causes of death.Conclusion:In Zhejiang province from2008-2013, the incidence of ESRD rises annually, and average age for them is gradually growing, the proporation of elderly patients is increasing. Main dialysis modality is hemodialysis, meanwhile, the proporation of peritoneal dialysis is increasing recent years. The leading cause of ESRD is CGN, DN rises with years. Cardiovascular mortality is the leading causes of mortality. Part II Hemodialysis versus Peritoneal Dialysis in Patients-a Propensity-Matched Cohort Study.Objective:It is still subject to debate about the survival differences between hemodialysis (HD) and peritoneal dialysis (PD). Here, we sought to compare the survival among incident maintenance PD to propensity score-matched maintenance HD patients in Zhejiang province.Methods:The data of incident dialysis patients were from the Zhejiang Dialysis Quality and Management Center (ZDQM) between1July2008and30June2013. All patients were followed up until30June2014, all-cause mortality was the primary outcome, patients with dialysis vintage less than90days and under18years old were excluded. Demographic, diagnosis and vascular access were collected. Propensity scores were calculated by using logistic multivariate regression analysis, we randomly selected HD patients who were matched to PD patients based on their propensity scores. Then survival analysis were performed using the Kaplan-Meier method, Log-Rank test and Cox proportional hazard model.Results:A total of19846incident dialysis patients were enrolled (15610HD patients and4236PD patients), and followed for a median of29months (3-72months).4233matched pairs were obtained, no significant difference were observed between the baseline characteristics of two groups, except liver disease. Kaplan-Meier survival curve revealed that there was no significant difference of overall survival between HD and PD patients(P=0.979), but PD had demonstrated survival advantages over HD in the first year(P<0.05). In subgroup analysis:1) There was no significant difference of survival between gender, age, cause of ESRD and comorbidities (log-rank test, P>0.05);2) No significant difference of survival was found in HD and PD patients under65years old, however, survival was higher in HD patients compared to PD patients in diabetes under 65years old (log-rank test, P<0.05), and survival was lower in HD patients compared to PD patients under65years old without diabetes.(log-rank test, P＜O.05);3) No significant difference of survival was found in HD and PD patients over65years old (log-rank test, P>0.05);4) In the multivariate Cox proportional risks model, age, cause of ESRD, cardiovascular disease, cancer and chronic obstructive pulmonary disease (COPD) showed statistical significance in explaining survival of incident ESRD patients. Patients with chronic glomerulonephritis(CGN) and hypertensive nephropathy(HTN) showed higher survival rate than diabetic nephropathy(DN) patients (P<0.05);5) Incident HD with arteriovenous fistula (AVF) experienced better survival than HD patients with central venous catheter and PD patients(log-rank test, P<0.001).Conclusions:In incident ESRD patients, the overall survival rate of PD patients was comparable to that of HD patients, mortality of HD patients was significantly higher than PD patients in the first year after the initiation of dialysis. Better survival of HD patients as compared to PD ones in patients with diabetes under65years old, survival rate was higher in PD patients compared to HD ones in patients without diabetes under65years old. The variables that most influenced survival were age, cause of ESRD, cardiovascular disease, cancer and COPD. Patients with CGN and HTN had a better survival than DN ones. Incident HD patients with AVF had a better survival prognosis than PD patients and HD patients with central venous catheter. Part III Development and Validation of a Predictive Mortality Risk Score in Incident Maintenance Hemodialysis PatientsObjective:The risk of death in hemodialysis(HD) patients is higher than general population. In order to alert doctors to HD patients with increase risk of the mortality, we aimed to predict2-Year all-cause mortality by using readily available clinical and laboratory data of incident maintenance hemodialysis(MHD) patients, and to establish a risk scoring model.Methods:Data of incident hemodialysis patients were from Zhejiang Dialysis Quality and Management Center(ZDQM) in the period of January2008to June2012. All patients were≥18years old and dialysis vintage>90days, and were followed up until30June2014. The demographic information, clinical and laboratory data were collected, mortality was the primary outcome. Patients were randomly divided into training dataset to establish a predictive model (60%) and validation dataset to validate the model(40%), the predictive model was developed by using a logistic regression model according to the clinical data in training dataset. The risk model discrimination and calibration was tested by using the area under the receiver operating characteristic(ROC) curve, described in terms of sensitivity and specificity, and the risk scoring model for2-Year mortality was set up according to the coefficient and rank of variables in the risk models, respectively.Results:Of4295patients from ZDQM,691patients(16.1%) died within2years. In our predictive model, predictors were age, causes of end stage kidney disease(ESRD), use of a vascular access catheter, history of cancer, serum albumin and serum total calcium, serum albumin was a protective factor. The predictive model=0.614*age (=1,2,3,4, or5)+0.864*vascular access (=0, or1)+(0.784*(if diabetic nephropathy(DN)=l), or0.217(if hypertensive nephrosclerosis(HTN)=l), or0.796 (if cause of ESRD was other/unknown=l))+0.709*cancer (=0, or1)-0.554*serum albumin (=0, or1)+0.270*serum total calcium (=0,1, or2)-4.943; The area under the ROC curve of predictive model in training dataset was0.767,95%CI(0.744-0.790), Hosmer-Lemeshow Chi-test,X2=3.144, P=0.925, which was highly discriminatory when applied to validation dataset(the area under the ROC curve0.732,95%CI (0.700-0.765)), and the sensitivity and specificity were71.5%and64.2%, respectively. Risk-score model:age<30years old:3points, points for every15years add3,>75years old:15points; causes of end stage kidney disease:chronic glomerulonephritis:0point, diabetes:4points, hypertension:1point, other/unknown:4points; vascular access catheter:4points; comorbidity:cancer:3points; laboratory data: serum albumin:>35g/l:-3points, serum total calcium:<2.1mmol/l:1point,2.1-<2.6mmol/1:0point,>2.6mmol/l:3points. The scores of low risk group, medium risk group, high risk group and very high risk group were<9points,10-13points,14-17points and≥18points.Conclusion:We retrospectively reviewed the incident MHD patients in ZDQM from2008～2012. The final risk predictive and risk-score model for2-Year all-cause mortality in MHD patients are developed based on readily available clinical and laboratory data with sufficient accuracy. They may be used to alert medical staff and researchers to hemodialysis patients with increased risk of death, and to help them make clinical decisions.