Interferon-γ Release Assays for the Diagnosis of Tuberculosis

To the Editors: We read with interest the recent report by Bianchi et al entitled “Interferon-γ release assay improves the diagnosis of tuberculosis in children.”1 We are concerned that this title may be misleading to readers and suggest that a more cautious interpretation of the study findings is warranted. In the study by Bianchi et al, the sensitivity of the QuantiFERON-TB Gold in Tube (QFT) was not different from that of the tuberculin skin test (TST) for the diagnosis of active tuberculosis (TB) (TST positive 14/16 vs. QFT positive 15/16, P = 1.0). However, of the 44 children with TST-defined latent TB infection (LTBI), QFT was positive in only half. Of these 44 children, 22 had a TST induration ≥15 mm and QFT was negative in approximately one-third of this subgroup. We have previously published similar findings showing significant discordance between TST and QuantiFERON-TB Gold.2 We have also reported a further study that showed 43% and 37% of children at high risk of LTBI, with a TST ≥15 mm were QFT and T.SPOT.TB negative, respectively.3 A critical question, with implications for future use of interferon-γ release assays (IGRA), is whether this type of discordance is attributable to a false-positive TST result or a false-negative IGRA result. Some researchers have advocated the former, based on the fact that IGRA incorporate Mycobacterium tuberculosis-specific antigens, and are therefore likely to be more specific than the TST. Previous BCG immunization is frequently postulated to be the primary epidemiologic factor underlying reduced specificity of the TST, though definitive proof to support this assumption is lacking. Bianchi et al contend that prior BCG immunization likely affected TST results in their cohort, as children with TST-defined LTBI and a negative QFT were more likely to be BCG immunized compared with those with a positive QFT. However, in the whole cohort this was not the case. In fact, a significantly higher proportion of uninfected (TST negative) children in whom the BCG vaccination status was known were BCG immunized compared with those deemed infected (TST positive) (152/175; 87% vs. 20/28; 71%, P = 0.047/odds ratio: 2.64, 95% CI: 1.04–6.7). Multivariate linear regression analysis assessing the influence of various factors on the results of TST and IGRA, including BCG vaccination, would be more informative to confirm potential associations. In the absence of a standard for the diagnosis of LTBI, the debate regarding which test is more sensitive or specific will likely continue.4 Currently available data suggest that, compared with adults, the performance of IGRA may be inferior in children.2,3,5–7 At present, the correct interpretation of a negative IGRA in a child with a significant TST induration remains a management dilemma in routine clinical practice. At this stage, it is uncertain which of the tests more accurately identifies children with LTBI and the influence of prior BCG immunization on the magnitude of TST induration remains contentious. We believe that in the context of a child at risk for LTBI it is not safe to attribute a TST ≥15 mm to a “false positive” due to prior BCG immunization. We note that Bianchi et al chose to treat children at risk for LTBI in whom TST and IGRA results were discordant. We also take this approach in the absence of sufficient data to suggest these children are not at risk for progression to active TB. Important strengths of the study by Bianchi et al include the young age of the cohort and novel data provided on internationally adopted children. The study also adds to the small number of studies investigating IGRA free from commercial involvement.8 Bianchi et al provide a reasonably balanced discussion but the data provided do not support the statement that IGRA “improves the diagnosis of tuberculosis in children.” We agree with the authors that large prospective studies with a sufficiently long follow-up are required to determine the true value of IGRA in children. Thomas Connell, MBBAO, Bch BMedSc, MRCPI Marc Tebruegge, MRCPCH, MD Nicole Ritz, MD Nigel Curtis, FRCPCH, PhD Department of Paediatrics, The University of Melbourne Infectious Disease Unit, Department of General Medicine; and Murdock Children’s Research Institute Royal Children’s Hospital Melbourne, Parkville, VIC 3052, Australia