We treated patients with idiopathic membranous nephropathy (iMGN) and renal insufficiency, using: (i) (n = 15) monthly cycles of steroids (1 g methyl-prednisolone i.v. on three consecutive days, followed by oral prednisone 0.5 mg/kg/day months 1, 3 and 5) and chlorambucil (0.15 mg/kg/day months 2, 4 and 6); or (ii) (n = 17) oral cyclophosphamide (1.5-2.0 mg/kg/day for 1 year) and steroids in a comparable dose. The groups were comparable in age, renal function and levels of proteinuria. During the 6 months preceding treatment, serum creatinine levels increased from 148 +/- 50 to 219 +/- 73 mumol/l in the chlorambucil group and from 164 +/- 86 to 274 +/- 126 mumol/l in the cyclophosphamide group. Median (range) follow-ups were: chlorambucil 38 months (8-71); cyclophosphamide 26 months (5-68) (NS). Renal function improved in both groups, but the improvement was short-lived in the chlorambucil group; 12 months after starting treatment, mean serum creatinine was 6.3 mumol/l lower in the chlorambucil group and 121 mumol/l lower in the cyclophosphamide group (p < 0.01). Four chlorambucil-treated patients developed ESRD, and five needed a second course of therapy, whereas only one cyclophosphamide-treated patient developed ESRD (p < 0.05). Remissions of proteinuria occurred more frequently after cyclophosphamide treatment (15/17 vs. 5/15; p < 0.01). Side-effects necessitated interruption of treatment in six patients on cyclophosphamide and in 11 on chlorambucil (p < 0.05). In our patients, oral cyclophosphamide was better tolerated than oral chlorambucil. The suggested greater efficacy of the oral cyclophosphamide regimen needs to be ascertained by longer follow-up.