Acute renal failure (ARF) is particularly frequent in the elderly. A few studies have reviewed immediate prognosis of ARF in older patients, but these do not allow any conclusion on long-term renal prognosis. Our retrospective study included 46 patients over 65 years referred to our renal unit between 1983 and 1989. Survivors were followed up 6-71 months after discharge (mean: 39 months). The evolution of renal function was evalated by measurement of serum creatinine. Data analysis employed chi-2 with Yates correction, Student t, and Mann-Whitney tests to compare survivors and decreased, and to compare patients with normal and abnormal renal function at follow-up. During hospitalisation 11 patients (24%) died. Our univariate analysis reveals that three variables independently influence mortality: consciousness disturbance (P < 0.001), high urea concentration (P < 0.01), and hypoalbuminaemia (P < 0.001). Age does not adversely affect prognosis. At follow-up 15 patients (43%) had a complete functional recovery, eight (23%) had incomplete renal recovery and two (6%) were on chronic haemodialysis. These results are similar to those observed in a younger population. In conclusion we believe that age alone should not be used to predict the immediate survival or the long-term renal outcome in ARF in the elderly.