Development And Validation Of Differential Diagnosis And Prognosis Models For Diabetic Nephropathy And Non-diabetic Renal Disease

BackgroundThe incidence of type 2 diabetes mellitus(T2DM)has risen rapidly in the past few decades.DN is a major complication of T2DM.It has surpassed primary glomerulonephritis and become the leading cause of Chronic kidney disease(CKD)and end-stage renal disease(ESRD).With the development of renal biopsy technology,it is increasingly recognized that non-diabetic renal disease(NDRD)is also common in patients with type 2 diabetes.NDRD has better efficacy and renal prognosis than DN.Therefore,it is very important to establish a differential diagnosis prediction model of DN and NDRD,identify the risk factors affecting the prognosis of DN,and establish a prognostic risk prediction model,which can provide an important basis for the diagnosis,treatment and prognosis assessment of DN patients in clinical work.Objective1.To develop and validate a predictive model for the differential diagnosis of diabetic nephropathy(DN)and non-diabetic renal disease(NDRD)in patients with type 2 diabetes mellitus.2.To develop and validate a predictive model for the risk of end-stage renal disease(ESRD)in patients with diabetic nephropathy(DN)confirmed by renal biopsy.Methods1.Establishment and validation of differential diagnosis modelWe conducted a retrospective study with 938 patients with type 2 diabetes who underwent renal biopsy in the first affiliated hospital of Zhengzhou university from February 2012 to January 2015.The dataset was randomly split into development(70%)and validation(30%)cohorts.Use univariable and multivariable logistic regression to identify baseline predictors for model development.The model's performance in the two cohorts,including discrimination and calibration,was evaluated by the C-statistic,calibration curve and the P value of the Hosmer-Lemeshow test.Establish the final prediction model and draw the corresponding nomogram.2.Establishment and validation of prognostic risk modelWe conducted a retrospective study with 478 patients with T2DM who underwent renal biopsy for the pathological confirmation of DN at the First Affiliated Hospital of Zhengzhou University from February 2012 to January 2015.The patients were followed until December 2018.The outcome was defined as a fatal or nonfatal ESRD event(peritoneal dialysis or hemodialysis for ESRD,renal transplantation,or death due to chronic renal failure or ESRD).The dataset was randomly split into development(75%)and validation(25%)cohorts.We used stepwise multivariable logistic regression to identify baseline predictors for model development.The model's performance in the two cohorts,including discrimination and calibration,was evaluated by the C-statistic and the P value of the Hosmer-Lemeshow test.Finally,the optimal model is selected and the nomogram of the model is drawn.Results1.Establishment and validation of differential diagnosis model1.1 Pathological features of baseline dataAmong the 938 patients with type 2 diabetes,478 cases(51%)diagnosed as DN alone,217 cases(23.1%)as NDRD alone and 243 cases(25.9%)as DN plus superimposed NDRD(MIX).Among NDRD and MIX patients,membranous nephropathy was the most common pathological type,followed by IgA nephropathy.1.2 Model derivationThe variables selected in the final predictive model were Age,history of diabetes,diabetic retinopathy,systolic blood pressure,hemoglobin,fasting blood glucose,glycosylated hemoglobin,cystatin C.1.3 Model validationInternal verification showed that the model performed well with good discrimination and calibration.The C-statistics were 0.903(95%CI(0.901-0.905))in the derivation cohort and 0.899(95%CI(0.896-0.902))in the validation cohort.The model had the best P value of 0.907 of the Hosmer-Lemeshow test.1.4 Draw a nomogramAll the predictors in the model were assigned values and the corresponding nomogram was drawn.It can be used as a probabilistic prediction tool for differential diagnosis of DN and NDRD.2.Establishment and validation of prognostic risk model2.1 Clinical features of baseline dataDuring the 3-year follow-up period,there were 225 outcome events(47.1%)during follow-up.Outcomes occurred in 187 patients(52.2%)in the derivation cohort and 38 patients(31.7%)in the validation cohort.2.2 Model derivationThe variables selected in the final multivariable logistic regression after backward selection were pathological grade,24hUTP(g),cystatin C,eGFR(ml/min/1.73 m2)and BNP.2.3 Model selection and validationAmong the four models,the clinical-pathological model that included pathological grade,cystatin C,BNP and Log ACR performed well,with good discrimination and calibration in the derivation cohort and the validation cohort.The areas under the receiver operating characteristic curves for the derivation and validation cohorts were 0.865(95%CI0.863-0.867)and 0.876(95%CI 0.874-0.878).It had the lowest AIC of 332.53 and the best P value of 0.909 of the Hosmer-Lemeshow test.2.4 Draw a nomogramThe nomogram can be used in clinical practice to predict the risk ratio of progression to ESRD for patients with DN within 3 years.Conclusion1.We constructed a simple,robust predictive model to predicting the presence of NDRD and MIX for type 2 diabetic patients with high accuracy.The model showed that age,duration of diabetes,diabetic retinopathy,fasting blood glucose,glycosylated hemoglobin,systolic blood pressure,cystatin C and hemoglobin were important predictors of NDRD and MIX.The nomogram can be used as a decision support tool to provide a non-invasive method for differential diagnosis of DN and NDRD,which may help clinicians assess the risk-benefit ratio of kidney biopsy for type 2 diabetic patients with renal impairment.2.We have developed and validated a model to predict the risk of progression to ESRD in patients diagnosed with biopsy-confirmed DN.The model demonstrated that cystatin C,eGFR and pathological grade influenced the risk of DN progression to ESRD.The model performed well among individuals who share characteristics similar to those of our source population and can be a practical and convenient tool for the early identification of and prognostic prediction in high-risk patients with DN in routine clinical practice.