This study examined the prevalence and determinants of tuberculin reactors and the booster response in 272 (97%) residents and 218 (69%) staff of an urban nursing home where a case of miliary TB was being investigated. Many subjects were foreign born (56%), and 15% had a history of BCG vaccination (34% of staff). Testing was done with 5 TU PPD-S (Connaught). Those with < 10 mm induration at 48 h were retested at 1 wk. Test sizes > 9 mm were classified positive (reactors). A questionnaire was given to collect information on risk factors for a positive test. Results showed that 28% (36% staff and 22% of residents) of subjects were reactors at the initial test, 6% at the booster test (staff and residents), and 32% at either test (40% staff and 26% residents). BCG (OR 4.8) and foreign birth (OR 1.7) were significantly associated with total reactors. The association with foreign birth was inversely related to the time since immigration. Only BCG was associated with a positive booster test (OR 6.7). Of positive tuberculin reactions in individuals with a history of BCG, 72% were attributed to the vaccine. We concluded that (1) staff as well as residents should be two-step tested when retesting is anticipated; (2) the prevalence of tuberculin reactors is highly related to the prevalence of BCG vaccinees and foreign birth; and (3) chemoprophylaxis is probably not indicated for isolated tuberculin reactors with a history of BCG.