The mortality of patients with acute renal failure (ARF) remains high, and in several large studies approaches 60%. This mortality is particularly high in patients with ARF who require dialysis and has not changed substantially over several years, despite the introduction of major advances in monitoring and treatment. Increasing prevalence of comorbidities has been suggested as the major factor in this persistently high mortality. This study investigates the potential role of the dialysis membrane on patient outcome in a prospective multicenter study of 153 patients with ARF requiring dialysis. The membrane assignment was made in alternating order and was limited to membranes with low complement activation (Biocompatible [BCM]) and cellulosic, high complement activation (Bioincompatible [BICM]). Both types of membranes were low-flux membranes. Patients were dialyzed with the assigned membrane until recovery, discharge from hospital, or death. The severity of illness of each patient was assessed using the APACHE II score at the time of initiation of dialysis. A logistic regression analysis was used to adjust for the APACHE II score. The results of the study showed a statistically significant difference in survival (57% in patients on BCM, 46% in patients on BICM; P = 0.03) and in recovery of renal function (64% in patients on BICM and 43% in patients on BICM; P = 0.001). These differences were particularly marked in the patients who were nonoliguric (>400 ml/d of urine output) at initiation of the study. In the subset of patients who were nonoliguric at the start of dialysis, a larger fraction (70%) became oliguric after initiating dialysis on a BICM membrane, in contrast to 44% who were initiated on a BCM membrane (P = 0.03). It is concluded that the biocompatibility of the dialysis membrane plays a role in the outcome of patients with ARF, particularly those who are nonoliguric at the time of initiation of dialysis.