Abstract: Objective To evaluation the relationship between the initiation of continuous renal replacement therapy (CRRT) guided by RIFLE grading and clinical outcomes in critically ill patients with acute kidney injury (AKI). Methods We recruited 87 AKI patients treated with CRRT in the period from 2010 to 2013. They were divided into early (RIFLE classified as risk, n=31) and late (RIFLE classified as injury or failure, n=56) initiation of CRRT by RIFLE criteria. Their demographic data and biochemistry parameters were collected. Clinical outcomes including mortality and renal function recovery were recorded. Results After the initiation of CRRT, the mortality rate at the 28th day was 64.52% and 57.14% in the early group and late group, respectively, and that at the 90th day was 67.74% and 66.07% in the early group and late group, respectively (P＞0.05). Kaplan- Meier curve revealed that the survival estimates were similar between the two groups (P=0.67). The relative death risks at the 28th and 90th days were also similar between the two groups (RR=0.724 and 0.921, respectively; P＞0.05). Renal function recovery at the 28th day was 38.70% and 30.36% in the early group and late group, respectively, and that at the 90th day was 38.70% and 32.14% in the early group and late group, respectively (P＞0.05). The relative risks for renal function recovery at the 28th and 90th days were also similar between the two groups (RR=1.449 at the 28th day, P=0.430; RR=1.333 at the 90th day, P=0.538). Cox multivariate analyses revealed that APACHE II score was the death risk factor for AKI patients. Conclusions RIFLE classification as the guide for the initiation of CRRT could not be used to predict the mortality at the 90th day and the renal function recovery in AKI patients.

Key words: continuous renal replacement therapy, acute kidney injury, timing, RIFLE