No consensus exists regarding the definition and interpretation of a significant boosting reaction after sequential tuberculin testing. The booster phenomenon is thought to represent remote tuberculous infection where tuberculin reactivity has waned, but it has also been described among persons with previous exposure to other mycobacteria or Bacille Calmette-Guérin (BCG) vaccine. We studied the factors affecting the booster phenomenon among Canadian-born young adults to determine the definition that would maximize sensitivity and specificity of a positive booster reaction in these persons. Point-prevalence survey of initial tuberculin reactions and response to repeated tuberculin testing after 1 to 4 weeks. Community-based study of all students entering health professional training programs at six post-secondary institutions. In 1989, 1990, and 1991, students completed self-administered questionnaires, underwent two-step tuberculin testing with purified protein derivative-tuberculin (PPD-T), and had their childhood BCG vaccination status verified. In 1991, students were also tested with purified protein derivative-Battey (PPD-B) (for Mycobacterium intracellulare). Overall, 74 students (5.2%) had positive booster reactions, which were significantly associated with older age (P < 0.001), larger initial tuberculin reactions (P < 0.001), previous BCG vaccination (P < 0.001), older age when vaccinated (P < 0.02), longer interval from vaccination to testing (P < 0.01), and sensitivity to PPD-B (P < 0.001). Boosting was not associated with the number of BCG vaccinations, sex, or risk factors for tuberculous infection. The pattern, mean, and mode of the frequency distributions of booster reactions among those with BCG vaccination and sensitivity to PPD-B were similar to those with assumed tuberculous infection. In young adults, booster reactions due to previous tuberculous infection are uncommon and cannot be distinguished from false-positive reactions due to past exposure to other mycobacteria.