Objective: The Acute Dialysis Quality Initiative (ADQI) Group published a consensus definition (the RIFLE criteria) for acute renal failure. We sought to assess the ability of the RIFLE criteria to predict mortality in hospital patients.
Design: Retrospective single-center study.
Setting: University-affiliated hospital.
Patients: All patients admitted to the study hospital between January 2000 and December 2002. Patients were excluded if they were younger than 15 yrs old, were on chronic dialysis, or had kidney transplant or if their length of hospital stay was <24 hrs.
Measurements and main results: We included 20,126 patients. Mean age was 64 yrs, 14.7% of patients required intensive care unit admission, and hospital mortality was 8.0%. According to the RIFLE criteria, 9.1% of all patients were in the Risk category for acute renal failure, 5.2% were in the Injury category, and 3.7% were in the Failure category. There was an almost linear increase in hospital mortality from Normal to Failure (Normal, 4.4%; Risk, 15.1%; Injury, 29.2%; and Failure, 41.1%). Multivariate logistic regression analysis showed that all RIFLE criteria were significantly predictive factors for hospital mortality, with an almost linear increase in odds ratios from Risk to Failure (odds ratios, Risk 2.5, Injury 5.4, Failure 10.1).
Conclusions: The RIFLE criteria for acute renal failure classified close to 20% of our study patients as having some degrees of acute impairment in renal function and were useful in predicting their hospital mortality.