BACKGROUND: Mortality rates of critically ill patients with acute renal failure (ARF) requiring renal replacement therapy (RRT) are high. Intermittent and continuous RRT are available for these patients on the intensive care units (ICUs). It is unknown which technique is superior with respect to patient outcome. METHODS: We randomized 125 patients to treatment with either continuous venovenous haemodiafiltration (CVVHDF) or intermittent haemodialysis (IHD) from a total of 191 patients with ARF in a tertiary-care university hospital ICU. The primary end-point was ICU and in-hospital mortality, while recovery of renal function and hospital length of stay were secondary end-points. RESULTS: During 30 months, no patient escaped randomization for medical reasons. Sixty-six patients were not randomized for non-medical reasons. Of the 125 randomized patients, 70 were treated with CVVHDF and 55 with IHD. The two groups were comparable at the start of RRT with respect to age (62+/-15 vs 62+/-15 years, CVVHDF vs IHD), gender (66 vs 73% male sex), number of failed organ systems (2.4+/-1.5 vs 2.5+/-1.6), Simplified Acute Physiology Scores (57+/-17 vs 58+/-23), septicaemia (43 vs 51%), shock (59 vs 58%) or previous surgery (53 vs 45%). Mortality rates in the hospital (47 vs 51%, CVVHDF vs IHD, P = 0.72) or in the ICU (34 vs 38%, P = 0.71) were independent of the technique of RRT applied. Hospital length of stay in the survivors was comparable in patients on CVVHDF [median (range) 20 (6-71) days, n = 36] and in those on IHD [30 (2-89) days, n = 27, P = 0.25]. The duration of RRT required was the same in both groups. CONCLUSION: The present investigation provides no evidence for a survival benefit of continuous vs intermittent RRT in ICU patients with ARF.