Background: Previous studies concerning the relative infectiousness of HIV-1-positive individuals with pulmonary tuberculosis have produced conflicting results. Thus, we assessed the effect of HIV-1 on the infectiousness of Mycobacterium tuberculosis in a prospective study.
Methods: We organised in Santo Domingo, Dominican Republic, a cohort study of household contacts of HIV-1-positive and HIV-1-negative individuals with newly diagnosed pulmonary tuberculosis. Household contacts were assessed at their houses at baseline and followed up for 14 months for evidence of M tuberculosis infection and tuberculosis with a multi-step tuberculin skin test, anergy skin test, physical examinations, chest radiographs, and sputum smears.
Findings: Tuberculin induration of 5 mm or greater was seen in 153 (61%) of 252 household contacts of HIV-1-positive index cases and in 418 (76%) of 551 household contacts of HIV-1-negative index cases (odds ratio 0.49 [95% CI 0.35-0.67], p=0.00001). In multivariate logistic-regression analysis after allowance for between-household variation in tuberculin response, HIV-1 infection of the index case remained inversely associated with the tuberculin response of the household contacts (0.52 [0.29-0.93], p=0.02). When the analysis was restricted to household contacts aged between 2 years and 15 years the adjusted association remained significant (0.37 [0.14-0.98], p=0.04). Among household contacts who had a negative tuberculin skin test at baseline, conversion to tuberculin skin test positivity was less frequent among household contacts of HIV-1-positive index cases (cut-off > or =5 mm: 32/131 [24%] vs 71/204 [35%], p=0.05; cut-off > or =10 mm: 23/153 [15%] vs 55/245 [22%], p=0.07).
Interpretation: These data suggest that HIV-1-positive individuals with tuberculosis are less likely than HIV-1-negative individuals with tuberculosis to transmit M tuberculosis to their close contacts. No changes in the current policy regarding tuberculosis contact tracing are needed in the presence of HIV-1.