Five hundred forty-three patients with completely resected malignant melanoma who were considered to have a significant risk of developing recurrent disease were randomized to one of four study groups. One group received levamisole 2.5 mg/kg on 2 consecutive days weekly for 3 years, a second group received bacillus Calmette-Guérin (BCG) for 3 years. A third group alternated 8-week courses of BCG and levamisole for 3 years and a fourth group underwent clinical assessment at the same frequency as the three treatment groups. The median duration of follow-up is 8.5 years. The percentage of reduction in the death rate and the recurrence rate in the treatment groups compared with the control group was calculated using the Cox proportional hazards model and adjusted for age, sex, and stage as covariants. The patients treated with levamisole were estimated to have a 29% reduction in both the death rate (P = .08) and the recurrence rate (P = .09) compared with patients receiving no further treatment. Fifty-five patients discontinued levamisole early because of gastrointestinal intolerance or arthralgia, myalgia, fever, and immune leukopenia. The patients treated with BCG alternating with levamisole experienced a 10% reduction in the death rate and a 6% reduction in the recurrence rate, and the patients treated with BCG alone experienced a 4% reduction in the death rate and a 3% increase in the recurrence rate compared with the control group. The degree of improvement experienced by the patients that were treated by levamisole is of sufficient magnitude to warrant further investigation of this dose of levamisole as adjuvant treatment in patients with melanoma.