Two-stage tuberculin skin testing in an HIV-infected population: a preliminary report

To identify differences in purified protein derivative (PPD) tuberculin positivity rates and to assess the utility of sequential PPD testing among persons with HIV infection, we performed two sequential administrations of 5 tuberculin units of PPD in a group of persons with HIV infection. Eligible patients were skin-tested with the tuberculin using the Mantoux method. Patients who tested < 5 mm of induration were skin-tested seven days later along with Candida and mumps antigens. Of 37 tested patients, only 18 (49%) returned to have the first test read. Using a > or = 5 mm cutoff, 5 (28%) tested positive. Positivity varied markedly between patients with CD4+ cell counts under and at or over 400 cells/mm3 (0% vs. 56%). Among patients who had a reaction to the first test, the reaction was smaller in patients with CD4+ cell counts under 400 than in those with counts at or over 400 cells/mm3 (mean induration: 2.8 vs. 30.4 mm). Positivity was also less frequent in intravenous drug users than in nonusers (9% vs. 57%); these two groups did not differ with respect to CD4+ cell counts. Of the 13 patients who tested < 5 mm, only 8 (62%) kept their appointments to have the second test placed and read. Only 1 of these, a drug user with a CD4+ count of 6 cells/mm3, had a positive reaction with boosting from 4 mm on the first test to 10 mm on the second. These data indicate that PPD testing may be unreliable in screening for tuberculosis infection in persons with CD4+ counts < 400 cell/mm3.(ABSTRACT TRUNCATED AT 250 WORDS)