TB prevention in HIV clinics in New York City.

SETTING: Ten hospital-based human immunodeficiency virus (HIV) clinics in New York City. OBJECTIVE: To evaluate tuberculosis (TB) prevention in HIV clinics based on the prevalence and incidence of TB and the efficacy of preventive therapy with isoniazid (INH). DESIGN: The medical records of 2393 HIV-infected patients with a first clinic visit in 1995 were reviewed retrospectively. Deaths and TB cases through December 1997 were ascertained through a match with the TB and AIDS registries. RESULTS: At first visit, 92 patients (4%) had a history of TB, 98 (4%) were being treated for TB, and six (<1%) were diagnosed with TB. During follow-up, 23 cases were diagnosed, an incidence of 0.53 per 100 person-years (py) (95%CI 0.34-0.77). Among 439 tuberculin skin test (TST) positive patients, the incidence of TB/100 py was 1.63 (95%CI 0.27-5.02) in patients with no INH, 1.28 (95%CI 0.40-2.98) in patients with <12 months of INH, and 1.06 (95%CI 0.38-2.28) in patients with 12 months of INH. The incidence/100 py was 0.0 (95%CI 0.0-0.78) in TST-negative patients and 0.37 (95%CI 0.09-0.95) in anergic patients. The relative risk of TB was 0.65 (95%CI 0.14-4.56) in TST-positive patients with 12 months of INH (vs. none). CONCLUSIONS: The benefits of TB prevention efforts in these HIV clinics from 1995 to 1997 were limited because most TB occurred before the first clinic visit. Methods for reaching HIV-infected patients earlier should be identified.