Detection of latent tuberculosis among HIV-infected patients after initiation of highly active antiretroviral therapy

This cross-sectional study of 110 individuals examined skin testing for latent tuberculosis infection (LTBI) after the initiation of highly active antiretroviral therapy. Skin test reactivity to one or more of four antigens was found in 98 out of 110 subjects (89%), and was maximal in those whose CD4 cell counts recovered to ≥ 100 cells/mm3. Skin testing is reliable for the identification or exclusion of LTBI once the CD4 cell count recovers to ≥ 100 cells/mm3. Patients co-infected with HIV and tuberculosis have high rates of reactivation of latent tuberculosis infection (LTBI) [1,2] and tuberculosis disease increases mortality in HIV-infected patients [3–5]. Therefore co-infected patients should receive treatment for LTBI to prevent tuberculosis disease. [6–8]. The diagnosis of LTBI relies on delayed-type hypersensitivity skin testing with Mycobacterium tuberculosis purified protein derivative (PPD). Many patients with HIV infection and low CD4 cell counts are anergic, but those who receive treatment with highly active antiretroviral therapy (HAART) often experience a recovery of cellular immune function [9–11]. This study sought to determine when, in patients with a history of a low CD4 cell count who have initiated HAART, skin testing with PPD can reliably identify or exclude LTBI. A cross-sectional study was performed at the Grady Memorial Hospital Infectious Diseases Program in Atlanta, Georgia, USA, enrolling HIV-infected patients who had previous CD4 cell counts of less than 100 cells/mm3 and were subsequently treated with HAART for at least 2 months. The study was approved by the human investigations committee at Emory University, and participants gave written informed consent. Four skin test antigens were studied: PPD (5 TU; Connaught, Swiftwater, PA, USA), Candida antigen (Candin; ALK, Wallingford, CT, USA), tetanus toxoid (Connaught) diluted 1 : 10 in albumin-saline diluent, and Mycobacterium avium sensitin (MAS 10/2, filling lot 68, 1 μg/ml; Statens Seruminstitut, Copenhagen, Denmark). For each antigen, 0.1 ml was injected intradermally on the forearm. Skin tests were read 48–72 h after placement by the same skin test reader. Anergy was defined as intradermal induration of less than 3 mm for all four antigens. A 5 mm or greater PPD skin test reaction was considered positive and indicative of LTBI. A total of 110 patients were enrolled. The median age was 40 years (range 19–62). Eighty-six patients (78%) were men and 71 (65%) were African-American. The patients had been diagnosed with HIV infection a median of 6.3 years earlier (range 4 months to 20 years) and previously had very low CD4 cell counts, with a median nadir of 12 cells/mm3 (range 0–98). Patients had received HAART for a median of 22 months (range 2–52 months). At the time of skin testing, the patients’ median CD4 cell count was 185 cells/mm3 (range 2–950 cells/mm3). The HIV load at the time of skin testing determined by RNA polymerase chain reaction ranged from below the limit of detection (400 copies/ml) in 45% of patients, to greater than 750 000 copies/ml, with a median of 690 copies/ml. Twelve of the 110 patients (11%) were anergic. Skin test reactivity was most common for Candida (83/110), with less frequent reactivity for tetanus (35/110), M. avium sensitin (75/110) and PPD (13/110). Previous anergy testing information was available for 21 patients. Seventeen of these (81%) were known to have been anergic before starting HAART, but only one (6%) remained anergic after the initiation of HAART. Skin test reactivity did not differ by age or sex. African-Americans were less likely than Caucasians to manifest skin test reactivity (60/71 versus 38/39, P = 0.053), but they also had lower CD4 cell counts (median 161 versus 240 cells/mm3, P = 0.005). Patients with a CD4 cell count of 100 cells/mm3 or more at the time of testing were significantly more likely to manifest skin test reactivity than those with less than 100 cells (75/77 versus 23/33, P < 0.001). A CD4 cell count nadir of less than 50 (80/92 versus 18/18, P = 0.21), viral load less than 400 copies (46/49 versus 52/61, P = 0.15) and a duration of HAART of less than 6 months (11/14 versus 87/96, P = 0.18) were not significantly associated with skin test reactivity. In multivariate analysis, only the CD4 cell count at time of testing was significantly associated with skin test reactivity (P = 0.002). Thresholds for CD4 cell counts at the time of skin testing were examined to determine at what point in CD4 cell recovery skin testing becomes sufficiently reliable for clinical use. As seen in Fig. 1, skin test reactivity increased with each interval of CD4 cell count (chi square for trend 1.97, P < 0.001).Fig. 1.: Skin test reactivity of patients on highly active antiretroviral therapy at various stages of recovery of CD4 cell count. Shaded area indicates the expected percentage of individuals with skin test reactivity in the absence of HIV infection.Of the 13 patients with positive tuberculin skin tests, 11 had a history of a positive PPD skin test, whereas two had previously been PPD skin test negative. One of these two had been incarcerated since his most recent negative PPD skin test, but the other had no known potential exposure to M. tuberculosis. This study demonstrated skin test reactivity in 89% of patients with advanced AIDS on HAART and restoration of skin test reactivity in 16 out of 17 (94%) patients with previous known anergy. Despite very low CD4 cell nadirs in these patients, the frequency of skin test reactivity seen after HAART was similar to that seen among immunocompetent patients [12–14]. Two patients with previously unrecognized positive PPD skin tests were discovered in the course of the investigation. It is unknown whether these two positive skin tests reflected a recovery of previous skin test reactivity or recent infection with M. tuberculosis. Notwithstanding, the skin testing of patients after they began HAART was beneficial, as these two patients were previously unrecognized candidates for LTBI treatment [6–8]. As this study was cross-sectional, it was not possible to follow the development of skin test reactivity over time to define prospectively the earliest threshold for retesting after the initiation of HAART. However, 75 out of 77 patients (97%) including 16 out of 17 (94%) with known previous anergy manifested skin test reactivity after their CD4 cell counts reached 100 cells/mm3. Patients with CD4 cell counts of 50 cells/mm3 or greater were significantly more likely to have skin test reactivity than those with less than 50 cells/mm3, but only 77% of patients with CD4 cell counts between 50 and 100 cells/mm3 had skin test reactivity; thus LTBI could not be excluded in this group. Therefore, 100 CD4 cells/mm3 appears to be the optimal threshold for performing PPD skin testing to identify or exclude latent tuberculosis infection after the initiation of HAART.