Improving the histologic classification of advanced diabetic nephropathy based on the Transformative Research in Diabetic Nephropathy (TRIDENT) findings.

INTRODUCTION

Diabetic nephropathy (DN) remains the leading cause of kidney failure worldwide. Histopathologic assessment is the diagnostic standard, and the Renal Pathology Society (RPS) classification system is widely used in clinical practice and research. However, its prognostic performance is limited, particularly in advanced disease.

METHODS

To examine this, we used TRIDENT, a multicenter, prospective observational cohort of patients with diabetes undergoing clinically indicated kidney biopsy (176 individuals). Light microscopy features were scored and analyzed by unsupervised k-means clustering to derive the RPS-TRIDENT modification (RPS-TM) classification. Associations between RPS-TM classes and kidney outcomes, death, dialysis initiation, or 40% or more decline in estimated glomerular filtration rate, were evaluated with Kaplan-Meier curves and Cox proportional hazards models. Clinical utility was assessed by decision-curve analysis. An external cohort of 101 individuals was used for validation.

RESULTS

The RPS classification showed only modest prognostic discrimination, with no significant survival difference between classes 3 and 4. The RPS-TM classification introduced an additional high-risk category (Class 5) defined by visceral epithelial hyperplasia, which more clearly separated risk groups, modestly improved prediction at one year (area under the curve 0.68 vs. 0.65) and two years (0.75 vs. 0.71), and demonstrated higher net benefit in decision curve analysis. RPS-TM Class 5 had the most rapid progression to kidney failure.

CONCLUSIONS

The RPS-TM classification enhances prognostication in DN by identifying a morphologic subgroup characterized by visceral epithelial hyperplasia with aggressive clinical course. This framework offers more precise risk stratification and could guide targeted management and trial design in DN.