Underestimation of Mycobacterium tuberculosis infection in HIV-infected subjects using reactivity to tuberculin and anergy panel

BACKGROUND: This study aimed to evaluate purified protein derivative (PPD) reactivity and its interrelationship with anergy panel and CD4+ lymphocytes in HIV-infected subjects as compared to PPD reactivity in HIV-uninfected individuals in a tuberculosis endemic and high Bacillus Calmette-Guérin (BCG) coverage environment. METHODS: Clients of four Mexico City HIV detection centres were screened for HIV-1 antibodies (ELISA or haemagglutination, Western Blot); reactivity to PPD (Mantoux PPD, 5TU RT-23), Candida (1:1000, 0.1 ml), and tetanus toxoid (10Lf, 0.1 ml); and CD4+ T cells. Active tuberculosis was excluded. Informed consent was obtained. RESULTS: From 5130 clients 1168 subjects were enrolled; of these 801 (68.6%) were HIV positive. Reactivity to PPD among HIV-positive subjects was found in 174 (22%), 261 (32.6%), and 296 (37%), at PPD cutoff levels of > or =10 mm, > or =5 mm, and > or =2 mm as compared to 224 (61%) of 367 HIV-negative individuals' reactors to PPD (> or =10 mm) (P < 0.001). After exclusion of anergic individuals using two cutoff levels for cutaneous allergens (< or =2 mm and < or =5 mm), PPD reactivity between HIV-infected and uninfected individuals continued to be significantly different. Only HIV-infected individuals with CD4+ T cells > or =500 cells/mm3 had similar reactivity to PPD as HIV-uninfected individuals. Variables associated with PPD reactivity were CD4+ T cell counts, BCG scar, HIV infection and age. CONCLUSIONS: PPD reactivity was useful to diagnose tuberculosis infection only among HIV-infected individuals with CD4+ counts > or =500 cells/mm3. Among individuals with lower counts, lowering cutoff levels or using anergy panel did not permit comparable reactivity as that observed among HIV-uninfected individuals.