Diagnostic markers in AKI

A major obstacle to improved treatment of AKI is the inability to make an early diagnosis. Timely recognition of the causes of AKI will lead to improved understanding of the pathophysiology and improved outcomes. Unfortunately no useful diagnostic markers for the causes of AKI are currently available. Several proteins and small molecules have been proposed as markers for specific renal diseases but they have significant numbers of false negatives and false positives, have not been tested in independent sets and have only been used to identify a single disease, not to identify one disease from a panel of diseases. In this study, we will attempt to differentiate between prerenal azotemia and intrinsic renal diseases of glomerular, tubular, interstitial or vascular etiology and contrast nephropathy. Urine and clinical data will be collected from five participating centers: The Medical University of South Carolina; The Ralph H Johnson VA Medical Center; George Washington University; Duke University Medical Center and a University of Tennessee-affiliated hospital in Chattanooga, Tennessee. We will measure a panel of candidate biomarkers in the urine. These studies will be done using multiplexed, immunologic assays (Luminex bead arrays), ELISA’s, immunonephelometry, western blotting of One- and Two-D gels, and enzyme activity assays. We will perform additional discovery studies to identify new candidate markers to predict the cause of renal failure. In these studies we will use 2DE/DIGE, MALDI peptide analysis and liquid chromatography/mass spectrometry with isotopic labeling. Studies will be performed on desalted concentrated whole urine, on samples which have had high abundance proteins subtracted out and on urinary exosomes. Predictive patterns of biomarkers will be identified using multiple regression, random forest analysis, artificial neural networks, nearest related neighbor analysis and other bioinformatic analysis techniques. We hypothesize that combinations of these candidate markers that come from renal diseases can be used to predict the cause of kidney failure. As we identify new markers from the discovery studies, we will add them to our panel of measured markers. The result of the studies using the multiplexed assays will be an algorithm which can be tested for its diagnostic accuracy. This algorithm will be tested in a large set of patients from five different sites. The urine samples will be tested blindly and the diagnostic prediction compared to “true” disease as determined by a panel of experts in AKI based on the complete clinical history, response to therapy and course of the disease.

Diagnostic markers in acute kidney injury

Southern Acute Kidney Injury Network

SAKInet