Study On The Risk Prediction Model Of Chronic Kidney Disease In The Elderly In Community

Background and Objective:Chronic kidney disease（CKD）is characterized by a structural or functional abnormality of the kidney that persists for more than 3 months and is usually clinically diagnosed with a urinary albumin creatinine ratio≥30mg/g or an estimated glomerular filtration rate<60ml·min^-1·（1.73 m²）^-1.The global burden of chronic kidney disease is increasing and is projected to become the fifth most common cause of death by 2040.In Chinese adults,chronic kidney disease is characterized by high incidence,high mortality and low awareness rate.And morbidity and mortality rates of chronic kidney disease are higher in the elderly.The main causes of chronic kidney disease in the elderly are diabetes and hypertension,accounting for 61%,and ranked third-glomerulonephritis and other causes accounting for 40%.In community primary diagnosis and treatment,diabetes and hypertension are easy to be diagnosed,and chronic kidney disease caused by them is easy to be paid attention to,followed up and managed.However,glomerulonephritis and other causes are difficult to be detected and diagnosed in primary care,the resulting chronic kidney disease is more likely to be ignored.At present,studies on chronic kidney disease mainly focus on chronic kidney disease caused by diabetes and hypertension or the whole population of chronic kidney disease,but lack more studies on chronic kidney disease in the elderly without hypertension or diabetes.Therefore,the analysis of the risk factors of chronic kidney disease in the elderly without hypertension and diabetes,as well as the risk assessment of chronic kidney disease in the elderly with accessible clinical data in community medicine,is of great significance to the grass-roots general practitioners and community residents.Method:Clinical data of the elderly from physical examination in community health service centers were collected,including a total of 8271 subjects.They were randomly divided into modeling group and validation group in a ratio of 7:3.SPSS statistical software was used to analyze the clinical characteristics in the modeling group and the verification group,and the clinical characteristics of the elderly with chronic kidney disease and without kidney disease in the modeling group and the observation indicators with statistical differences were analyzed.The multivariate analysis was carried out based on the indicators with differences and the prediction equation model was created.Finally,in the verification group,the differentiation（ROC curve）and calibration（Hosmer-Leme）of the prediction equations were tested show goodness of fit test.Results:There was no statistical difference in each observation index between the modeling group and the test group,and the two groups were comparable.From the analysis in the modeling group,Gender,age,waist circumference,past medical history（hyperuricemia,renal calculi,renal cysts）,alcohol consumption,alpha-fetoprotein,glycosylated hemoglobin,urinary creatinine,urinary microalbumin,UACR,blood uric acid,fasting blood glucose,serum glutamic oxalacetic transaminase,serum alanine aminotransferase,serum creatinine,e GFR,blood urea nitrogen,triglyceride,and triglyceride glucose index were all associated with CKD related factors.BMI,past history（arteriosclerotic disease,dyslipidemia,fatty liver,obesity/overweight）,frequency of physical exercise,carcinoembryonic antigen,total bilirubin,total cholesterol,high-density lipoprotein,and low-density lipoprotein were not correlated.Among them,female and drinking are the protective factor of chronic kidney disease.Based on the above correlation factors,the regression equation was obtained:X=-17.351+0.072 age-1.909×sex（0=male,1=female）+0.165×urinary microalbumin+0.147×serum creatinine.Constructing the prediction equation:P=e^x/（1+ex）,where e represents the natural logarithm.The area under ROC curve of the prediction model was0.962（95%confidence interval:0.954～0.97,P<0.001）,the Yoden index was 0.783,the sensitivity was 90.3%,and the specificity was 88%.The area under ROC curve of the verification group was 0.847（95%confidence interval:0.82-0.874,P<0.001）.It is proved that the prediction model is stable and has good prediction ability.Hosmer-Lemeshow goodness of fit test was used to evaluate the calibration ability of the prediction model,X²=3.400,P=0.907（>0.05）,indicating a good goodness of fit curve.Conclusion:We established and verified a prediction model for the risk of chronic kidney disease in the elderly without hypertension or diabetes.The model has good differentiation and calibration ability,and can be used in primary medical practice as an early identification tool for the elderly without hypertension or diabetes.