Tuberculin Skin Testing Is Useful in the Screening for Nontuberculous

(See the editorial commentary by Hill on pages 1552–4) Background. We evaluated the diagnostic usefulness of tuberculin skin testing in the screening for nontuber-culous mycobacterial (NTM) infection in children. Methods. We enrolled 180 children who had chronic cervicofacial lymphadenitis in our study. Skin testing was done using antigens of Mycobacterium tuberculosis, Mycobacterium avium, Mycobacterium kansasii, and My-cobacterium scrophulaceum. The reference standard for NTM infection was a positive culture result, identification by PCR, or both. Receiver operating characteristic analysis was used to identify the optimal cutoff point in skin induration for the detection of NTM infection. Accuracy of the mycobacterial skin tests was quantified using sensitivity and specificity rates and positive and negative predictive values at the optimal skin induration cutoff. Results. A total of 112 NTM infections were identified, of which 83 were caused by M. avium, 21 by Myco-bacterium haemophilum, and 8 by other NTM species. At the optimal cutoff for a positive test (5 mm), tuberculin skin testing had a sensitivity and specificity of 70 % and 98%, respectively, and a positive predictive value and a negative predictive value of 98 % and 64%, respectively, compared with a sensitivity and a specificity of 93 % and 97%, respectively; M. avium sensitin, the best-performing skin test, had positive and negative predictive values of 98 % and 90%, respectively