Acute kidney injury in intensive care units according to RIFLE classification: a population-based study.

Abstract

Recent studies of the incidence of acute kidney injury (AKI) are largely based on estimated baseline serum creatinine values. The aim of this study was to more accurately determine the incidence of AKI using the RIFLE criteria for intensive care unit (ICU) patients of a whole population. All adult patients admitted to the ICUs of Landspitali - The National University Hospital of Iceland in 2007 (n = 1026) were studied with meticulous search for baseline creatinine. The underlying risk factors and contributing causes for AKI were defined, and survival and ratio of end-stage renal failure evaluated. A measured baseline creatinine value was found for all but two patients with AKI. The incidence of AKI according to RIFLE criteria was 21.7% [95% confidence interval (CI): 19.0-24.1%], with 7.1% (95 CI: 5.6-8.9%), 6.8% (95 CI: 5.3-8.5%) and 7.8% (95 CI: 6.2-9.6%) in the risk, injury and failure subgroups. Using estimated baseline creatinine overestimated the incidence of AKI by 3.5%. The sensitivity and specificity of the RIFLE criteria using estimated baseline creatinine were 76% and 95%. Renal replacement therapy was required for 17% of the AKI patients. One year survival of AKI patients was 51%, but only 2.5% of patients surviving 90 days required chronic renal replacement therapy. The incidence of AKI in the ICU was lower than previously published, perhaps due to overestimation of AKI using estimated baseline creatinine or bias from tertiary referrals. AKI patients have high mortality, but the survivors have a low incidence of end-stage renal failure.