Neutrophil gelatinase-associated lipocalin is a better biomarker than cystatin C for the prediction of imminent acute kidney injury in critically ill patients

Abstract

Aims The prevention of acute kidney injury can be lifesaving for the intensive care unit patients. However, conventional methods are not sufficient for the prediction of the risk of future acute kidney injury. In this study, the promising biomarker, neutrophil gelatinase-associated lipocalin, was compared with cystatin C as an indicator for the risk of future acute kidney injury. Methods One hundred and eighty-three adult patients without chronic kidney disease or renal replacement therapy were included in this study. The plasma and urine concentrations of neutrophil gelatinase-associated lipocalin and cystatin C were assessed on the second day after intensive care unit admission and were followed for seven days to monitor the development of acute kidney injury. Acute kidney injury diagnosis was based on the risk, injury, failure, loss, end-stage renal failure criteria. Results Thirty-four per cent of the patients had acute kidney injury; 17 patients who did not fulfil criteria at the beginning, developed acute kidney injury from days 3 to 7 after admission. The mean serum creatinine on admission did not significantly differ between this and control groups (0.72 ± 0.20 and 0.83 ± 0.21; P = 0.060); however, the serum and urinary neutrophil gelatinase-associated lipocalin concentrations on the second day were significantly different (median: 75.69 [54.18–91.18] and 123.68 [90.89–166.31], P = 0.001; and median: 17.60 [8.56–34.04] and 61.37 [24.59–96.63], P = 0.001). Notably, the 48-h serum cystatin C concentration did not differ. Conclusion Neutrophil gelatinase-associated lipocalin concentrations in the urine and serum on the second day of intensive care unit admission could be used to predict the development of acute kidney injury in the following three to seven days in the intensive care unit; however, the cystatin C concentration did not have predictive value.