BACKGROUND Repeat BCG vaccination is standard practice in many countries for prevention of tuberculosis and leprosy, but its effectiveness has not been evaluated. The addition of Mycobacterium leprae antigens to BCG might improve its effectiveness against leprosy. A double-blind, randomised, controlled trial to evaluate both these procedures was carried out in Karonga District, northern Malawi, where a single BCG vaccine administered by routine health services had previously been found to afford greater than 50% protection against leprosy, but no protection against tuberculosis. METHODS Between 1986 and 1989, individuals lacking a BCG scar were randomly assigned BCG alone (27,904) or BCG plus killed M leprae (38,251). Individuals with a BCG scar were randomly allocated placebo (23,307), a second BCG (23,456), or BCG plus killed M leprae (8102). Incident cases of leprosy and tuberculosis were ascertained over the subsequent 5-9 years. FINDINGS 139 cases of leprosy were identified by May, 1995; 93 of these were diagnostically certain, definitely postvaccination cases. Among scar-positive individuals, a second BCG vaccination gave further protection against leprosy (about 50%) over a first BCG vaccination. The rate ratio for all diagnostically certain, definitely postvaccination cases, all ages, was 0.51 (95% CI 0.25-1.03, p = 0.05) for BCG versus placebo. This benefit was apparent in all subgroups, although the greatest effect was among individuals vaccinated below 15 years of age (RR = 0.40 [95% CI 0.15-1.01], p = 0.05). The addition of killed M leprae did not improve the protection afforded by a primary BCG vaccination. The rate ratio for BCG plus killed M leprae versus BCG alone among scar-negative individuals was 1.06 (0.62-1.82, p = 0.82) for all ages, though 0.37 (0.11-1.24, p = 0.09) for individuals vaccinated below 15 years of age. 376 cases of postvaccination pulmonary tuberculosis and 31 of glandular tuberculosis were ascertained by May, 1995. The rate of diagnostically certain tuberculosis was higher among scar-positive individuals who had received a second BCG (1.43 [0.88-2.35], p = 0.15) than among those who had received placebo and there was no evidence that any of the trial vaccines contributed to protection against pulmonary tuberculosis. INTERPRETATION In a population in which a single BCG vaccination affords 50% or more protection against leprosy, but none against tuberculosis, a second vaccination can add appreciably to the protection against leprosy, without providing any protection against tuberculosis.