Comparison of validity and reliability of the risk equationwith KDIGO staging, in the prediction of the end stage renal failure

Introduction and purpose: Chronic kidney disease (CKD) is a chronic disease defined as impairment of kidney structure and function for more than 3 months. Glomerular filtration rate is used in disease staging and predicting progression, but it alone is not sufficient to predict progression. Various models have been emphasized in predicting KBH progression. One of these models is the renal failure risk equation created by Tangri and his colleagues (BYRD). In our study, it is aimed to evaluate the applicability of the renal failure risk equation model in our country and to compare it with KDIGO staging using data from hemodialysis patients who are being monitored in nephrology outpatient clinics. Equipment and Methods: In this study, 129 patients were included in their evaluations 2 and 4 years before the onset of hemodialysis, consisting of those who were looked at their rate of urine albumin creatinine among the hemodialysis patients who were followed up at nephrology outpatient clinic between January 2015 and December 2020. The patients whose KBH etiology were glomerulonephritis, vasculitis, connective tissue disease and patients with malignancies were not included in the study. All data in the study were analysed using the SPSS 21.0 program. Sensitivity is calculated using the ROC (Receiver Operating Characteristic) curve. Findings: Demographic and laboratory data were obtained retrospectively from the file records of the patients. The mean age of the patients included in the study was 64.09 ±11.67% and 51.2% of the patients were female and 48.8% were male. Risk calculation was made with laboratory data of patients 2 and 4 years ago using 4 variable BYRD and risk situations were determined according to KDIGO with this data. Patients were divided into two groups as low-medium (<15%) and high (>15%) risk. In the artificial neural network approach made with BYRD results and KDIGO estimates 4 years ago, the correct classification rate for the high-risk group vii was 97.1%, the correct classification rate for the low-risk group was found to be 80.8%, and the area under the ROC curve was found to be as high as 0.983 in both the low and high risk groups. Likewise, BYRD and KDIGO estimates calculated with data 2 years before the onset of hemodialysis and the area below the ROC curve were found to be 0.943. Result: The c-statistical value in comparison with KDIGO and prediction of the progress of the renal failure risk equation to end-stage renal failure in the Turkish population was found to be significant and studies involving a large number of patients are needed.