Determining Mycobacterium tuberculosis infection in high prevalence groups: a comparative study among Nigerian adults

In populations where both tuberculosis and strong sensitization to nontuberculous mycobacteria are common, determining the proportion infected with M. tuberculosis is difficult. We skin tested 488 unvaccinated young Nigerian workers, typical of tropical populations having high prevalences of tuberculosis, with 4 low-dose Mantoux tests (PPD-RT-23, PPD-Tuberculin, PPD-Battey, and PPD-Gause) and with concentrated (2 mg/ml) PPD by the Heaf multiple-puncture method. Reactions (greater than or equal to 2 mm) to all 4 Mantoux sensitins (elicitins) unexpectedly were normal in distribution; however, the midpoint values (13 mm) of the RT-23 and PPD-T distributions were lower than the midpoints typically seen in groups of tuberculosis patients. Established methods for estimating the proportion tuberculous-infected (cutting point, dual-test, and curve-reconstruction) yielded a wide range, 32%-62%. Combining the 3 methods provided consistent estimates near 50%, despite interference due to strong sensitivity to PPD-B and especially to PPD-G in 80%-90%. PPD-T correlated well with RT-23 (r = .84), yet more closely resembled published PPD-S reaction distributions than did RT-23. All diagnostic-accuracy measures for the Heaf test, considering only grades 3 and 4 positive, exceeded 80%.