The long-term results of four different regimens of prednisone therapy were compared in 32 children with steroid sensitive, frequently relapsing idiopathic nephrotic syndrome with minimal glomerular lesions on renal biopsy. Prednisone was administered according to the following dosage schedules: 1) long-term daily, 2) standard intermittent, 3) standard alternate-day, and 4) short-term daily. Over a mean observation period of 7 years patients without steroid dependency received a cumulative dosage of prednisone of 10 mg/m2/day and those with steroid dependency received 19 mg/m2/day. Relapse free intervals were the longest with long-term daily prednisone therapy compared to the other three regimens. In frequently relapsing patients without dependency the relapse free intervals were similar with either intermittent or alternate-day prednisone therapy (median 75 d); however, they were significantly shorter with short-term prednisone therapy (median 33 d). In frequently relapsing patients with steroid dependency the time of remission was generally shorter than in patients without steroid dependency (median 25 d vs. 69 d) with no benefit of any of the different forms of short-term treatment.