OBJECTIVES: To assess the feasibility of establishing a pneumococcal vaccine trial among HIV-1-infected adults in Uganda and to characterize their responses to 23-valent pneumococcal polysaccharide vaccine. DESIGN: An open-label pilot trial to assess recruitment and compliance of HIV-1-infected adults in Uganda to vaccination and to determine the immunogenicity of the vaccine. SETTING: A community clinic for HIV-1-infected adults in Entebbe, Uganda. METHODS: Levels of capsule-specific IgG to four common vaccine capsular serotypes were measured before vaccination and 1 month after vaccination. Subsequent rates of disease episodes and deaths, and immunologic responses in two vaccine failures, were followed. RESULTS: One month after-vaccination, both HIV-1-infected (n = 77) and seronegative control subjects (n = 10) demonstrated a significant rise in capsule-specific immunoglobulin G (IgG) for three of four serotypes tested, but levels were significantly lower among HIV-1-infected patients. In 149 patient-years of follow-up, two (2.6%) developed pneumococcal pneumonia, one bacteremic with serotype 1 and one non-bacteremic with serotype 13, a non-vaccine serotype; both patients showed inadequate killing of the organism in vitro. In this same follow-up period, 29 (38%) patients died. CONCLUSION: HIV-1-infected adults in Uganda are at high risk of pneumococcal disease and show a significant but suboptimal response to pneumococcal vaccine. Although reliable recruitment and follow-up of vaccinees is feasible, evaluation of vaccine efficacy may be compromised by limited responses to common vaccine serotypes, an unknown incidence of disease with non-vaccine serotypes, and a high rate of mortality unrelated to Streptococcus pneumoniae infection.