BACKGROUND: Studies have failed to show a survival difference between intermittent hemodialysis (IHD) and continuous renal replacement therapy (CRRT). Comparative cost analyses are limited and fail to control for differences in patient disease severity and comorbid conditions. The authors retrospectively estimated clinical and economic outcomes associated with CRRT and IHD among critically ill patients experiencing acute renal failure (ARF) in 2 tertiary care hospitals in Rochester, Minnesota, between January 1, 2000, and December 12, 2001. METHODS: 161 critically ill patients requiring dialysis for ARF were analyzed. Patient demographics, comorbid conditions, ARF etiology, mode of renal replacement therapy (RRT), renal recovery, and survival were abstracted from medical chart. APACHE II scores at dialysis initiation were calculated. Administrative data tracked length of stay (LOS) and direct medical costs from initiation of RRT to death or intensive care unit (ICU) and hospital discharge. Multivariate modeling was used to adjust outcomes for baseline differences. RESULTS: 84 (52%) of the patients received CRRT and 77 (48%) received IHD. CRRT-treated patients were younger (58 vs 65 years), less likely male (58% vs 77%), had higher APACHE II scores (32 vs 27) with a higher incidence of sepsis (46% vs 30%) and respiratory disease (56% vs 39%), and were less likely to have chronic renal insufficiency (32% vs 49%). With adjustment for differences in baseline patient characteristics, the RRT method did not affect the likelihood of renal recovery, in-hospital survival, or survival during follow-up. Mean adjusted ICU LOS was 9.5 days shorter for IHD-treated than CRRT-treated patients (P< .001), and the adjusted mean difference in hospital and total costs associated with ICU stay was $56,564 and $60 827, in favor of IHD (P< .001). Mean adjusted total costs through hospital discharge were $93 611 and $140,733 among IHD-treated and CRRT-treated patients, respectively (P< .001). CONCLUSIONS: This observational study suggests that costs may significantly differ by mode of RRT despite similar severity-adjusted patient outcomes. Future prospective comparisons of renal replacement modalities will need to include both clinical and economic outcomes.