Impact of combination antiretroviral therapy on the risk of tuberculosis among persons with HIV infection Enrico Girardi a , Giorgio Antonucci a , Paola Vanacore a , Marco Libanore b , Isabella Errante c , Alberto Matteelli d , Giuseppe Ippolito a and the Gruppo Italiano di Studio Tubercolosi e AIDS (GISTA)  Objective: To assess the association between use of different antiretroviral regimens and incidence of tuberculosis among HIV-infected individuals. Design: Observational, multicenter, prospective cohort study. Setting and patients: Twenty-eight infectious diseases hospital units in Italy. A total of 2160 HIV-infected persons were considered for enrolment in a study on the imple- mentation of tuberculosis preventive therapy between 1 May 1995 and 30 April 1996. The 1360 subjects who completed tuberculin screening at base-line were included in this analysis. Information on the use of antiretroviral therapies over time was collected. The median duration of follow-up was 104 weeks and 997 subjects (73.3%) com- pleted the study. Main outcome measure: Incidence of active tuberculosis according to different types of antiretroviral therapy. Results: Eighteen cases of tuberculosis were observed with an overall incidence rate of 0.79 per 100 person±years of observation [95% con®dence interval (CI), 0.51± 1.31]. Tuberculin positivity and low CD4 lymphocyte count were the only base-line variables independently associated with the risk of tuberculosis. During follow-up, 637 patients took double combination antiretroviral therapy and 387 took triple combination therapy. After adjusting for base-line characteristics of enrolled indivi- duals, the relative hazard of tuberculosis was 0.16 (95% CI, 0.03±0.74) for double combination therapy and 0.08 (95% CI, 0.01±0.88) for triple combination therapy compared with no therapy or monotherapy. Conclusions: Combination antiretroviral therapy signi®cantly reduced the risk of tuberculosis in HIV-infected persons. In industrialized countries, the widespread use of this treatment may determine a decrease in the incidence of HIV-associated tuberculosis, possibly contributing to a reduction in the overall incidence of tubercu- losis. & 2000 Lippincott Williams & Wilkins From the a Centro di Riferimento AIDS - Servizio di Epidemiologia delle Malattie Infettive, IRCCS L. Spallanzani, Rome, the b Divisione di Malattie Infettive, Arcispedale S. Anna, Ferrara, the c Divisione di Malattie Infettive, Ospedale Niguarda, Milan and the d Clinica di Malattie Infettive e Tropicali , Spedali Civili, Brescia, Italy.  See Appendix. Sponsorship: This work was ®nancially supported by Ministero della Sanita Á- Progetto AIDS, grants No. 9401.09 and 50.A.0.20 and Fondi per la Ricerca Corrente degli IRCCS. Correspondence to Dr. Enrico Girardi, Centro di Riferimento AIDS- Servizio di Epidemiologia delle Malattie Infettive, IRCCS L. Spallanzani, Via Portuense, 292 - 00149 Rome, Italy. Tel: 39 065594223; fax: 39 065594224; e-mail: craids@tiscalinet.it Received: 28 January 2000; revised: 19 April 2000; accepted: 3 May 2000. ISSN 0269-9370 & 2000 Lippincott Williams & Wilkins 1985