Biomarkers for acute kidney injury: combining the new silver with the old gold

Abstract

Over the last decade, the pursuit for drugs to prevent ortreat acute kidney injury (AKI) has been replaced by asearch for novel biomarkers of kidney damage. Promptedby expert opinion that lack of sensitive and specific bio-markers has thwarted progress in the field, several newbiomarkers have emerged and are jockeying to becomethe ‘golden’ test for AKI. As clinical experience withthese biomarkers accumulates, it is increasingly evidentthat no single biomarker will likely take the crown. Thereis a slow realization that the consistent emphasis on theweakness and limitations of existing markers may be mis-guided. In fact, there may be several opportunities to re-evaluate existing techniques in combination with newerbiomarkers for managing patients with AKI.Urine microscopy has long been considered to be awindow to the kidney; however, its clinical value wasquestioned after publications suggested that urinalysis wasnot helpful in discriminating between functional andintrinsic renal disorders, especially in sepsis [1–3]. Never-theless, there has been a recent resurgence in interest inurine microscopy as a tool to characterize AKI spurred bydata from two groups. Chawla et al. [4] developed an AKIcast scoring index to standardize urine sediment analysisand were able to show good precision of the index todetect acute tubular necrosis (ATN). In that study, urinesediment also correlated with outcomes in patients withATN. Renal recovery was worse in patients with a highercast scoring index (2.55±0.9 versus 1.7±0.79; P=0.04),and the area under the receiver operating characteristic(ROC) curve of the cast scoring index for the prediction ofnon-renal recovery was 0.79. Perazella et al. [5] proposeda different scoring system for differentiating ATN from de-creased kidney perfusion in AKI (pre-renal AKI). Usingfinal AKI diagnosis at discharge as the gold standard,urinary microscopy on the day of nephrology consultationwas highly predictive of ATN. The odds ratio for ATN in-crementally increased with a higher score. In patients withan initial diagnosis of ATN, any granular casts or renal epi-thelial tubular cells (corresponding to a score of 2) resultedin a positive predictive value of 100% and a negative pre-dictive value of 44%. Lack of renal epithelial tubular cellsor granular casts in patients with an initial diagnosis ofdecreased kidney perfusion (functional AKI) had a sensi-tivity of 0.73 and a specificity of 0.75 for the final diagno-sis of ATN. The same group developed a scoring pointsystem of urinary microscopy findings to predict adverseoutcomes [6]. They evaluated the correlation of urinarysediment score and the Acute Kidney Injury Network(AKIN) stage at nephrology consultation. The score ofurinary microscopy was associated with higher risk ofworsening AKI in a dose-dependent manner.In this issue of the journal, Schinstock et al. reinforcethe value of urinary microscopy as a predictor of AKI andAKI severity. They performed urine microscopy andmeasured Neutrophil Gelatinase-associated Lipocalin(NGAL) levels in urine samples collected in the emer-gency room (ER) from 363 patients who were sub-sequently admitted to the hospital. Patient charts werereviewed to determine whether they developed AKI usingthe AKIN creatinine criteria of a rise in creatinine of >0.3mg/dL within the 48 h following admission. Normalranges and thresholds for the NGAL ELISA were estab-lished in a cohort of 125 healthy volunteers. Seventy-sixpatients (21%) developed AKI with 65% in AKIN Stage1. Finding a renal epithelial cell, renal epithelial cellcast or granular cast had separately a high specificity(93.0–98.6%) to discriminate AKI from non-AKI patients.The presence of these elements as a group showed ahigher sensitivity (from 6 to 22%) and good specificity(91%), determining a low negative but high positive pre-dictive value (81.6). The discriminative performance waseven better when comparing more severe AKIN Stage 2or 3 from AKIN Stage 0 or 1: sensitivity 30%, specificity90% and positive predictive value 94%. In contrast, urineNGAL at ER admission had a sensitivity of 64.5% and aspecificity of 64.5% to predict the development of AKI,which was significantly lower than the admission serumcreatinine (sensitivity of 84.2% and specificity of 77.7%).Urine NGAL levels also predicted AKIN stage; however,urine microscopy additionally predicted death.Nickolas et al. [7], in a single-center study of ERpatients, showed sensitivities and specificities of urinary