We have examined, in a prospective randomized controlled trial, the effect of 8- and 16-week initial steroid treatment on the course of idiopathic nephrotic syndrome (INS). Patients with a first episode of INS were randomized to receive standard 8-week prednisolone (2 mg/kg daily for 4 weeks, then 1.5 mg/kg on alternate days for 4 weeks) or prolonged 16-week prednisolone treatment (2 and 1.5 mg/kg daily each for 4 weeks, then 1.5 and 1 mg/kg on alternate days each for 4 weeks). Relapses were treated with prednisolone, 2 mg/kg daily for 2 weeks, then 1.5 mg/kg on alternate days for 4 weeks. Of 45 patients, 23 received standard therapy and 22 prolonged therapy. The mean duration of follow-up was 29.2 and 27.3 months in the standard and prolonged treatment groups, respectively. The time to first relapse was longer in the prolonged treatment (mean 222.2 days, median 120.0 days) than the standard group (mean 134.3 days, median 96.5 days). The percentage of patients with no relapse at 6 and 12 months after prednisolone withdrawal was 40.9% and 27.3% in the prolonged treatment and 21.7% and 8.7% in the standard groups, respectively. The inability to show statistically significant differences between the two groups was probably related to the small number of patients studied. Prolonged therapy did not affect the subsequent relapse rates and proportion of patients with frequent relapses and steroid dependence. The mean dose of prednisolone received, for the initial episode and relapses during the next year, was higher and associated with significant steroid toxicity in the prolonged treatment group. Our findings suggest that 16-week prednisolone treatment for the initial episode of INS may delay occurrence of the first relapse, but results in significant side effects. Prolongation of initial therapy may be useful in developing countries where frequent infections often induce early relapses.