BACKGROUND Diagnosis of acute kidney injury (AKI) has been a major concern due to its association with increased morbidity and mortality. However, the clinical implication of the urine output criterion (UOCr) in diagnosing AKI has not been fully established.

METHODS We assessed the incidence of AKI among 1625 critically ill patients and analysed the overall survival rates based on the serum creatinine criterion (CrCr) and UOCr, both of which have been defined by the AKI Network (AKIN).

RESULTS Within 7 days of admission, the risk rate of AKI was 57.0% and the rate determined by UOCr alone was 25.7%. AKI determined by the UOCr alone increased hazard ratios (HRs) for mortality; 1.81 (Stage 1), 2.96 (Stage 2) and 4.17 (Stage 3) compared to non-AKI. However, the difference in mortality between Stages 2 and 3 using the CrCr alone was not significant (P = 0.881). In patients with Stages 2 and 3 by the CrCr, the UOCr further separated the survival rates (P = 0.001 among the four UOCr stages). The diuretic dose did not alter the discriminative function of the UOCr for survival rates. However, 42.1% of non-AKI cases, as determined by the UOCr, were identified as AKI cases by the CrCr.

CONCLUSION Although some AKI cases were not identified by the UOCr alone, the UOCr has an additional role in AKI staging, regardless of diuretic use.