A comparison of the time from sepsis to inception of continuous renal replacement therapy versus RIFLE criteria in patients with septic acute kidney injury.

We hypothesized that the time from sepsis to inception of continuous renal replacement therapy (CRRT) can be used to predict survival rates in patients with septic acute kidney injury (AKI). The survival predictability of CRRT inception time was compared with that of RIFLE criteria, which were previously used in clinical practice. We retrospectively analyzed outcomes in 55 patients with septic AKI admitted to the medical intensive care unit at Asan Medical Center (Seoul, Korea) between April 2009 and October 2010. These patients were stratified by the time of inception of CRRT from sepsis (early: ≤ 24 h and late: >24 h) and also by the RIFLE criteria (RIFLE-I and RIFLE-F). The primary outcome was 28-day mortality. Of the 55 patients, 38 (69.1%) were male. Patients' mean age was 62.6 years, the most common infection site was the lung (32, 58.2%), and 47 patients (85.5%) were on mechanical ventilation. Thirty patients (54.5%) were in the RIFLE-I, and the others were in the RIFLE-F. Twenty-eight-day mortality rates were lower in the early group than in the late group (19.4% vs. 47.4%; P = 0.03), but did not differ between RIFLE-I and RIFLE-F. Ventilator-free day at day 28 was longer in the early group than that in the late group (7.5 vs. 0 d; P = 0.033). After adjustment for covariates, we found that the late group (hazard ratio, 3.106; 95% confidence interval, 1.066-9.047) and Sequential Organ Failure Assessment at sepsis (hazard ratio, 1.410; 95% confidence interval, 1.108-1.796) were independent factors associated with 28-day mortality. This study suggests that the time interval from sepsis to CRRT inception may be a more useful predictor of 28-day mortality than RIFLE criteria in patients with septic AKI.