A placebo-controlled, double-blind study on the efficacy of a hepatitis A vaccine (SmithKline Beecham Biologicals) was started in a region of Chile in September 1990, using hepatitis B vaccine as control. A total of 260 healthy children, 6-15 years of age, negative for antibody to hepatitis A virus (anti-HAV), antibody to HAV immunoglobulin M (IgM), hepatitis B surface antigen, and antibody to hepatitis B surface and core antigens by ELISA tests within 7 days before vaccination, were randomly assigned to two study groups: 128 children received the vaccine with a yellow label (group 1), and 132 children the vaccine with an orange label (group 2) at months 0, 1 and 6. Blood for serology and transaminase determination was drawn at months 1, 2, 6, 7 and 12. Both vaccines were tolerated equally well and no serious side effects were seen. In group 1 (presumed hepatitis A vaccine group), anti-HAV was detected (20% inhibition was used as the cut-off level) in 122 of 128 children (95.5%) tested at month 1, in 126 of 127 (99.2%) at month 2, in 126 of 127 (99.2%) at month 6 and in 126 of 126 (100%) at month 7. One anti-HAV seroconversion seen at month 1 was associated with presence of anti-HAV IgM and therefore probably represents HAV infection. Geometric mean anti-HAV concentration of the other children was 128, 342, 214 and 2301 mIU/ml at months 1, 2, 6 and 7, respectively. In group 2 (presumed hepatitis B vaccine group), three anti-HAV seroconversions, all associated with transient appearance of anti-HAV IgM, were detected at months 1 and 2. The difference in HAV infection rate seen in the two vaccine groups is not significant at month 12 of the study. © 1992.