Evaluation of the prognostic value of the risk, injury, failure, loss and end-stage renal failure (RIFLE) criteria for acute kidney injury.

AIM The experts have argued about the use of the risk, injury, failure, loss and end-stage renal failure (RIFLE) criteria as a prognosis scoring system. We examined the association between in-hospital mortality and the RIFLE criteria, and discussed its accuracy as a prognosis factor. METHODS In this prospective study, we analysed the data gathered from a cohort of 956 patients admitted in a Spanish tertiary hospital between January 1998 and April 2006. Hazard ratios for mortality, and survival curves within 60 days were calculated. Discrimination and calibration of the model were also assessed. RESULTS Excluding 53 patients, 903 patients were finally analysed. We classified them into groups according to the maximum RIFLE class reached during their admission. The RIFLE class was assessed by the glomerular filtration rate criterion. We found an increase in the in-hospital mortality risk. Cox proportional hazard models showed that RIFLE classes risk, injury, and failure were significant predictive factors (hazard ratios were 2.77, 3.23 and 3.52, respectively; P for trend was 0.005). The multivariate analyses from the cross-classification of the participants according to Liano score values (severity of illness) and RIFLE classes showed additive effects of the exposures on in-hospital mortality. CONCLUSION In this population, the risk of in-hospital mortality during the acute kidney injury (AKI) episode was positively associated with RIFLE classes. We showed that the RIFLE classification system had discriminative power in predicting hospital mortality within 60 days in AKI patients, but not better than a specific AKI predictive model. However, a combined use of both may give a more robust prognosis system.