Acute renal failure with concomitant sepsis in the intensive care unit is associated with significant mortality. The purpose of this study was to determine if the timing of initiation of renal replacement therapy (RRT) in septic patients had an effect on the 28-day mortality. Retrospective data on medical intensive care unit patients with sepsis and acute renal failure requiring RRT were included. Renal replacement therapy started with a blood urea nitrogen (BUN) of <100 mg/dL was defined as "early" initiation, and initiation with a BUN >or=100 mg/dL was defined as "late." Multivariate logistic regression analysis with the primary outcome of death at 14, 28, and 365 days following the initiation of RRT was performed. One hundred thirty patients were studied. The early dialysis (mean BUN 66 mg/dL) group had 85 patients; the late group (mean BUN 137 mg/dL) had 62 patients. The mean acute physiology and chronic health evaluation II score was 24.5 in both groups. The overall 14, 28, and 365-day survival rates were 58.1%, 41.9%, and 23.6%. Survival rates for the early group were 67%, 47.7%, and 30.7% at 14, 28, and 365 days. Survival rates for the late group were 46.7%, 31.7%, and 13.3% at 14, 28, and 365 days. Upon logistic regression analysis, initiating dialysis with a BUN >100 mg/dL predicted death at 14 days (odds ratio [OR] 3.6, 95% confidence interval [CI] 1.7-7.6, P=0.001), 28 days (OR 2.6, 95% CI 1.2-5.7, P=0.01), and 365 days (OR 3.5, 95% CI 1.2-10, P=0.02). Septic patients who started dialysis with a BUN <100 mg/dL had improved mortality rates up to 1 year after initiation of dialysis in this single-center, retrospective analysis.