RESUMEN
PURPOSE: We evaluated the renal arterial resistive index (RRI), urine monocyte chemotactic protein 1 (uMCP-1), and urine neutrophil gelatinase-associated lipocalin (uNGAL) to predict acute kidney injury (AKI) in critically ill cancer patients. METHODS: In this prospective study, we included patients without AKI. We compared the area under the curve (AUC) of RRI, uMCP-1, and uNGAL to predict any stage of AKI and stage-3 AKI with the DeLong method, and we established cutoff points with the Youden index. RESULTS: We included 64 patients, and 43 (67.2%) developed AKI. The AUC to predict AKI were: 0.714 (95% CI 0.587-0.820) for the RRI, 0.656 (95% CI 0.526-0.770) for uMCP-1, and 0.677 (95% CI 0.549-0.789) for uNGAL. The AUC to predict stage-3 AKI were: 0.740 (95% CI 0.615-0.842) for the RRI, 0.757 (95% CI 0.633-0.855) for uMCP-1, and 0.817 (95% CI 0.701-0.903) for uNGAL, without statistical differences among them. For stage 3 AKI prediction, the sensitivity and specificity were: 56.3% and 87.5% for a RRI > 0.705; 70% and 79.2% for an uMCP-1 > 2169 ng/mL; and 87.5% and 70.8% for a uNGAL > 200 ng/mL. The RRI was significantly correlated to age (r = 0.280), estimated glomerular filtration rate (r = - 0.259), mean arterial pressure (r = - 0.357), and serum lactate (r = 0.276). CONCLUSION: The RRI, uMCP-1, and uNGAL have a similar ability to predict AKI. The RRI is more specific, while urine biomarkers are more sensitive to predict stage 3 AKI. The RRI correlates with hemodynamic variables. The novel uMCP-1 could be a useful biomarker that needs to be extensively studied.