OBJECTIVES: To 1) study a variant delayed reaction to tuberculin testing as a way to enhance screening for tuberculosis among high-risk persons and 2) correlate the delayed reaction with lymphocyte blastogenesis. DESIGN: Cross-sectional study. SETTING: 2 public health department clinics in North Carolina. Participants: 121 adults who had recently emigrated from Vietnam to North Carolina and who were ethnic Vietnamese and ethnic Dega, a minority population group from the central highlands region of Vietnam. MEASUREMENTS: Medical history, physical examination, laboratory evaluation, and standard purified protein derivative (PPD) testing (Mantoux method). Skin test results were read at 72 hours and again at 6 days. Variant reactivity was defined as induration of less than 10 mm at 72 hours that, when reassessed at 6 days, had increased in size to 10 mm or greater. Persons with negative (n=54) and variant (n=32) PPD results also had booster testing at 10 to 12 weeks. Serum samples were obtained from 57 participants for lymphocyte blastogenesis studies. RESULTS: 26% of participants had variant tuberculin reactivity. Variant reactivity was strongly associated with booster positivity: Sixty-five percent of persons with variant PPD results had booster positivity compared with 16% of persons with negative PPD results (P<0.001). The lymphocyte blastogenesis response of persons with variant PPD results was between the response of persons with negative PPD results and that of persons with positive PPD results. CONCLUSION: Variant reactivity in this high-risk group was a predictor of booster positivity. Together with the blastogenic response pattern, this association strongly suggests that variant reactivity has a high positive predictive value for tuberculous infection. Clinicians should incorporate these findings into their approach for choosing candidates for preventive therapy.