Abstract
INTRODUCTION: Diabetic nephropathy (DN) is a major cause of chronic kidney disease and end-stage renal failure worldwide. The current diagnostic marker, albuminuria, lacks specificity and often detects renal damage only at advanced stages. OBJECTIVES: This study aimed to characterize urinary metabolic alterations associated with DN and explore metabolite panels with diagnostic potential. METHODS: A targeted urinary metabolomics analysis was performed using the validated TMIC Urine MEGA Assay, quantifying 268 metabolites in 60 participants (20 controls, 20 type 2 diabetes mellitus [DM-2], and 20 DN patients). Data were analyzed by Partial Least Squares Discriminant Analysis (PLS-DA) for visualization, and penalized regression algorithms [Least Absolute Shrinkage and Selection Operator (LASSO) and Elastic Net (EN) with a Genetic Algorithm (GA)] followed by logistic regression (LR) modeling to identify potential discriminative variables. RESULTS: DN patients showed marked alterations in metabolites related to oxidative stress, mitochondrial dysfunction, and inflammation. Twenty-four of 86 quantified uremic toxins differed significantly between DN and comparison groups. The LASSO-derived model identified β-alanine, kynurenine, glucose and argininic acid as key discriminants (AUC = 0.905, 10-fold CV), while inclusion of GFR and additional metabolites (2-hydroxybutyric acid, shikimic acid) improved performance (AUC = 0.96). CONCLUSIONS: Quantitative urinary metabolomics revealed metabolic perturbations reflective of DN pathophysiology and identified candidate metabolite panels with potential for non-invasive disease characterization. These findings, though preliminary, provide a foundation for validation in larger, longitudinal cohorts and for integrating urinary metabolomics into precision diagnostics for diabetic kidney disease.