Objective: To define the incidence of tuberculin skin test (TST) conversion and to evaluate the yield of annual testing in an era of declining tuberculosis incidence rate in the United States.
Methods: Annual TSTs were performed on initially TST-negative women (HIV infected, 995; uninfected, 260) from October 1995 through March 2002.
Results: A total of 4622 repeat TSTs were performed during 5530 person-years. The incidence of TST conversion was 0.8 case per 100 person-years for HIV-infected and 1.0 case per 100 person-years for uninfected women. Non-Hispanic blacks, women younger than 40 years of age, and HIV-infected women who had recently initiated active therapy were more likely to experience TST conversion. The incidence of conversion decreased over the course of the study from a peak of 21 cases per 937 tests in 1996 to 1 case per 179 tests in 2002 (P = 0.046 for trend). Twenty-one of 47 conversions occurred on the second TST, implying that boosting accounted for a number of conversions.
Conclusions: The yield of annual skin testing diminished from 1996 to 2002. Our data suggest that repeating testing after initiation of HIV therapy, regardless of CD4 cell change, is warranted. If serial testing is undertaken, initial 2-step testing should be performed to allow for accurate interpretation of subsequent tests and earlier identification of persons with latent Mycobacterium tuberculosis infection.