Multicentre prospective validation of a urinary peptidome-based classifier for the diagnosis of type 2 diabetic nephropathy

Abstract

We provide for the first time validation of this urinary proteome-based classifier in a multicentre prospective setting and show the suitability of the CKD273 classifier to be used in the PRIORITY trial.

Urinary peptidomics provides a noninvasive humanized readout of diabetic nephropathy in mice

“…By this 10 design, macroalbuminuria and eGFR were strongly associated to the presence of DN (AUC of 0.94 and 0.88 for albuminuria and eGFR, respectively). Patient treatment information was not available in this study 27 . The humanized classifier clearly separated the 165 T2D-patients with and without DN, although with lower sensitivity and specificity than the original CKD273 classifier (AUC of 0.73 and 0.95 for the humanized-and CKD273-classifier, respectively) ( Fig.…”

“…Human CE-MS data and general patient characteristics (Table 5) were extracted from the human urinary peptide database 35 and from 27 . The study was designed and conducted fulfilling all of the requisites of the laws on the protection of individuals collaborating in medical research and was in accordance with the principles of the Declaration of Helsinki.…”

“…Capillary electrophoresis coupled to mass spectrometry (CE-MS) was used to study the urinary peptidome of 230 patients compared to 379 healthy control subjects and led to the identification of a panel of 273 peptides associated with advanced CKD 24 . These 273 peptides combined into a classifier, termed CKD273, was subsequently validated for its capacity to detect CKD in independent cross-sectional studies with over 2000 patients including a significant number of patients (~70%) with diabetes [25][26][27][28] .…”

Urinary Protein Biomarkers in Chronic Kidney Disease

“…In the first blinded validation, CKD 273 classifier significantly outperformed albuminuria, showing sensitivity of 86% and a specificity of 100% [28]. It was also validated in another cohort of CKD patients with different disease etiologies and healthy controls [29], and in diabetic patients with or without overt diabetic nephropathy [27,30]. Besides proving its capability to identify patients with established CKD in independent studies, CKD273 classifier was also able to predict progression of CKD.…”

“…Combination of multiple biomarkers in high-dimensional classifiers, substantially outperform linear combination of biomarkers [26]. Electrophoresis coupled to mass spectrometry (CE-MS) appears to be an applicable method for urinary proteome analysis and has been extensively used in discovering and validating biomarkers for CKD [17,27].…”