Study On The Early Influencing Factors Of Chronic Kidney Disease In Elderly Community And The Establishment Of Prediction Model

Research background and purpose:Chronic kidney disease(CKD)is currently recognized as a global public health problem.The prevalence and influencing factors of CKD in the elderly in community may differ from the general population.However,at present,there is a lack of longitudinal observation research and prediction model construction research on the occurrence of CKD in the elderly in community in Sichuan.Therefore,through medical examination database data,the purpose of this study is to observe the current status of CKD among the elderly in community,to explore the important early factors that affect CKD in the elderly in community,and to explore the construction of prediction models to provide relevant basis and reference for the further early prevention of CKD.Methods:In accordance with the inclusion and exclusion criteria,the elderly population in the community without CKD who had undergone a health checkup in our hospital from January to December 2015 was selected,and the incidence of CKD after 3 years was observed longitudinally.Consult the literature and experts to collect medical examination data that may be closely related to the occurrence of CKD.Based on the results of CKD in the study population 3 years later,baseline physical examination data were divided into CKD group and non-CKD group for comparison.Then single-factor and multi-factor analysis was performed to find the important early influencing factors of CKD in the elderly in community,and as the final predictors,a logistic regression disease prediction model was constructed.Finally,the predictive ability of the model was evaluated by receiver operating characteristic curve and Hosmer-Lemeshow test.Result:1.A total of 469 elderly community study subjects were included in this study,of which 56(11.9%)developed CKD patients after 3 years.Of the 56 CKD patients,there were 7 patients with stage 1 of CKD(12.5%),33 patients with stage 2 of CKD(58.9%),and 16 patients with stage 3 of CKD(28.6%).No patients with stage 4 and stage 5 of CKD were found.2.Results of single factor two-classification logistic regression analysis of CKD in the elderly in community:age(P=0.018,OR=2.073,95%CI:1.133-3.791),body mass index(P<0.001,OR=4.537,95%CI:2.167-9.499),systolic blood pressure(P=0.009,OR=2.341,95%CI:1.240-4.418),history of hypertension(P=0.002,OR=3.038,95%CI:1.528-6.042),history of diabetes(P<0.001,OR=4.409,95%CI:2.463-7.896),serum creatinine(P<0.001,OR=1.045,95%CI:1.023-1.067),cystatin C(P=0.003,OR=3.612,95%CI:1.757-5.468),high density lipoprotein(P=0.010,OR=0.285,95%CI:0.109-0.739),hemoglobin(P=0.006,OR=0.974,95%CI:0.955-0.992).3.Results of multivariate two-classification logistic regression analysis of CKD in the elderly in community:body mass index(P=0.005,OR=3.124,95%CI:1.413-6.906),history of hypertension(P=0.009,OR=2.825,95%CI:1.299-6.413),history of diabetes(P<0.001,OR=3.582,95%CI:1.859-6.901),blood creatinine(P<0.001,OR=1.056,95%CI:1.029-1.082),high density lipoprotein(P=0.044,OR=0.323,95%CI:0.107-0.969),hemoglobin(P<0.001,OR=0.951,95%CI:0.929-0.973).4.Logistic regression prediction model of CKD in the elderly in community:P=1/[1+exp?(0.698+1.139×body mass index+1.038×history of hypertension+1.276×history of diabetes+0.054×blood creatinine?1.131×high density lipoprotein?0.051×hemoglobin)].5.Evaluation of predictive ability of prediction model of CKD in the elderly in community:the area under the ROC curve is 0.836,the sensitivity was 0.732,and the specificity was 0.801,Hosmer-Lemeshow test results show that the model fit is feasible(X~2=13.432,df=8,P=0.100).Conclusion:1.The 3-year new incidence of CKD in the elderly study subjects included in this study is higher than that of the general population,of which CKD patients are mainly stage 2.2.Univariate and multivariate Logistic regression analysis revealed that history of diabetes,body mass index,history of hypertension,high density lipoprotein,serum creatinine,and hemoglobin were important early independent influencing factors for CKD in the elderly in community.3.This study established a logistic regression prediction model of CKD in the elderly in community using six independent influencing factors:history of diabetes,body mass index,history of hypertension,high density lipoprotein,serum creatinine,and hemoglobin.After ROC curve verification and hosmer-lemeshow test,the model has certain predictive power.