ISONIAZID PREVENTIVE THERAPY IN HUMAN IMMUNODEFICIENCY VIRUS-INFECTEDPERSONS - LONG-TERM EFFECT ON DEVELOPMENT OF TUBERCULOSIS AND SURVIVAL

Background: Although the short-term benefit of isoniazid prophylaxis i n patients coinfected with human immunodeficiency virus (HIV) and tube rculosis has been shown, long-term benefits are unknown. Methods: Hist orical cohort study in an acquired immunodeficiency syndrome unit at a tertiary referral hospital. 4 sample of 121 HIV-infected patients wit h positive results on a purified protein derivative test were followed up for development of active tuberculosis and survival. Patients who received isoniazid prophylaxis were compared with patients who did not receive prophylaxis. Results: Of the 121 patients examined, 29 (24%) completed a 9- to 12-month course of isoniazid prophylaxis (median fol low-up, 89 months), and 92 (76%) did not receive the drug (median foll ow-up, 60 months). Active tuberculosis developed in 46 patients (38%). The incidence of tuberculosis was higher among patients with no proph ylaxis (9.4 per 100 patient-years) than among patients with isoniazid prophylaxis (1.6 per 100 patient years) (P=.006). Risk for development of tuberculosis was associated with the absence of isoniazid prophyla xis (relative risk [RR], 6.55; 95% confidence interval [CI], 2.02-21.1 9). Death during the period of study was more frequent in patients who did not receive isoniazid (50/92 or 54%) than in patients who receive d isoniazid (7/29 or 24%) (P=.008). Median survival was more than 111 months in patients who received isoniazid compared with 75 months in p atients who did not receive isoniazid (P<.001). In a proportional haza rds analysis, the development of tuberculosis (RR, 1.88; 95% CI, 1.09- 3.27), the absence of isoniazid prophylaxis (RR, 2.68; 95% CI, 1.16-6. 17), and a CD4(+) cell count lower than 0.20x10(9)/L (RR, 3.03; 95%CI, 1.39-6.61) were independently associated with death. Patients who rec eived isoniazid had a longer survival after stratifying for the CD4(+) cell count. Conclusions: Preventive therapy with isoniazid confers lo ng-term protection against tuberculosis and significantly increases su rvival in patients dually infected with HIV and Mycobacterium tubercul osis.