Objective: In vitro interferon-gamma (IFN-gamma) assays have emerged as novel alternatives to the tuberculin skin test (TST) for the diagnosis of latent tuberculosis (TB) infection. These assays have been evaluated in low incidence countries, mainly in adults, and have been shown to be more specific than TST. Because few studies have been done in high incidence countries, and because paediatric data are limited, we compared a whole-blood IFN-gamma assay with TST among hospitalized Indian children.
Methods: Between July 2004 and June 2005, a total of 105 consecutively admitted children (median age 6 years; 82% had BCG scars) in whom TB was suspected or had history of contact with an index case were recruited at a rural hospital in India. All children underwent TST, and the QuantiFERON-TB-Gold In Tube (QFT) assay.
Results: The overall prevalence of TB infection was similar with both tests. With a TST cut-off point of > or =10mm, 10 of 105 (9.5%; 95% CI 3.8, 15.2) children were TST positive. With a cut-off point of IFN-gamma> or =0.35IU/ml, 11 of 105 (10.5%; 95% CI 4.5, 16.4) were QFT positive. The concordance between TST and QFT was substantial (agreement 95.2%; kappa [kappa] 0.73; 95% CI for kappa 0.53, 0.92). Agreement between TST and QFT results was 100% (kappa 1.0) in BCG scar-negative children as compared to 94% (kappa 0.63) in scar-positive children. BCG was not associated with the results of either TST or QFT (P>0.05 for both tests). The number of children with bacteriologically confirmed active TB was too small to permit the estimation of sensitivity of the tests.
Conclusions: In a rural, predominantly BCG-vaccinated paediatric population in India, the TST and QFT assay produced comparable results. BCG vaccination did not significantly affect either TST or QFT results. Larger studies are needed to compare the sensitivity of the IFN-gamma assay with that of the TST in children with bacteriologically and/or clinically confirmed TB.