The diagnosis and prognosis of acute kidney injury (AKI) by current clinical means is inadequate. Biomarkers of kidney injury that are easily measured and unaffected by physiological variables could revolutionize the management of AKI. Our objective was to systematically review the diagnostic and prognostic utility of urine and serum biomarkers of AKI in humans. We searched MEDLINE, PubMed and EMBASE databases (January 2000-August 2009) for biomarker studies that could be classified into the following categories: (a) confirmation of the diagnosis of established AKI, (b) early prediction of AKI, and (c) prognostication of AKI. We identified 54 manuscripts published since 2000 that met our inclusion and exclusion criteria. Urinary interleukin-18 (IL-18), kidney injury molecule-1 (KIM-1), neutrophil gelatinase-associated lipocalin (NGAL) and N-acetyl-beta-D-glucosaminidase (NAG) are potentially useful biomarkers for the diagnosis of established AKI. Urinary NGAL, IL-18, and liver-type fatty acid binding protein, and serum NGAL and cystatin C represent the most promising biomarkers for early prediction of AKI. Urinary cystatin C, alpha1-microglobulin, NAG and retinol-binding protein may be useful to predict severity and outcomes of AKI. In conclusion, we identified several studies of promising biomarkers for the diagnosis, prediction and prognostication of AKI. However, we note several limitations, including small sample sizes, inadequate gold standard, exclusion of patients with chronic kidney disease, incomplete statistical analyses, utilization of research-based assays and a paucity of studies examining prediction for clinical outcomes. Future studies will need to address these limitations in order for further progress to be made.