Construction of a prognostic assessment model for diabetic nephropathy based on serum fibrinogen and renal tissue IFTA score

OBJECTIVE: To explore the association between serum fibrinogen (FIB) and clinicopathological features and renal prognosis in type 2 diabetic nephropathy (T2DN), and develop a web-based dynamic model to predict renal progression. METHODS: This paper retrospectively enrolled 173 biopsy-proven T2DN patients and stratified them by the optimal FIB cutoff. Renal progression was defined as a >50% decline in eGFR, doubling of creatinine, or onset of end-stage renal disease (ESRD). Cox regression analysis was used to screen out the independent predictors of T2DN progression; a web-based dynamic prediction model was established and evaluated using time-dependent receiver operating characteristic (Time-ROC) curves, calibration curves, and decision curve analysis (DCA). RESULTS: Of the 173 patients, 81 (46.82%) experienced the renal endpoint event. Multifactorial Cox regression analysis showed that eGFR, hemoglobin, FIB, parathyroid hormone, and interstitial fibrosis and tubular atrophy (IFTA) scores were independent risk factors for T2DN progression (P < 0.05). Among the three models constructed based on these factors, model 3 had the highest areas under the curve (AUC) (90.42; 95% CI, 85.80-95.04). The AUC of the risk prediction model constructed by the nomogram was 0.846, 0.752, and 0.754 at 1, 3, and 5 years, showing good discrimination, and the Web-based dynamic nomogram demonstrated good calibration. CONCLUSIONS: FIB is an independent predictor of T2DN progression. A web-based dynamic model was developed to predict renal progression risk in T2DN.