A 2-year prospective study was done in 45 patients to see whether the FENa test, done at first nephrology consultation visit for acute renal failure, was useful in prognosticating whether patients would recover renal function spontaneously or need dialysis. Of 26 patients who recovered renal function spontaneously, 8 had an FENa of less than 1; of the 19 who needed dialysis 8 also had an FeNa of less than 1. The mean FENa was 4.95% in those who recovered, and 6.28% in those who needed dialysis. There was thus no difference in false-positive or negative tests or numerical results of the FENa. Dividing the patients up into those who received and did not receive furosemide before made no difference. There were no differences in plasma or urine, creatinine or sodium, nor in blood pressure nor rate of rehydration measured as change in body weight between the groups. The FENa in our study was of no clinical importance, and of less prognostic use than simple hourly urine volume at time of consultation. We feel that the clinical usefulness of the FENa has been overemphasized in the literature.