The effect of prolonged treatment with levamisole was examined in 43 patients (30 boys) with steroid-dependent nephrotic syndrome (SDNS). The mean age at institution of treatment was 4.0 +/- 2.0 years. Fourteen patients had previously received cyclophosphamide with an ensuing remission of 8.5 +/- 10 months. Following induction of remission with prednisolone, levamisole was administered at a dose of 2.5 mg/kg body weight on alternate days. Prednisolone was tapered by 2.5-5 mg every 4 weeks to 0.5 mg/kg on alternate days. The duration of levamisole therapy ranged from 6 to 31 months (mean 17.4 +/- 8.4 months); 15 patients received levamisole for more than 18 months and 10 for more than 24 months. Prednisolone was discontinued in 18 patients after a mean duration of 11.7 +/- 7.1 months, whereas in 21 patients its dose was reduced to 0.2-0.4 mg/kg on alternate days. The mean relapse rate prior to levamisole therapy was 3.0 +/- 1.5 relapses/year, which reduced to 0.9 +/- 0.7 relapses/year during levamisole treatment (P < 0.001). A comparison of the response in 14 patients who had previously received cyclophosphamide with the other 29 patients did not show any significant difference. There were no side effects of levamisole therapy. Our findings suggest that prolonged treatment with levamisole is beneficial and safe in SDNS, with a marked steroid-sparing effect. A significant proportion of these patients can be kept in remission on levamisole alone.