Discordant results of tests for tuberculosis reconsidered reply

We thank Sandra Arend and Jonathan Uzorka for their thoughtful comments on our study.1Ho CS Feng PI Narita M et al.Comparison of three tests for latent tuberculosis infection in high-risk people in the USA: an observational cohort study.Lancet Infect Dis. 2022; 22: 85-96Summary Full Text Full Text PDF PubMed Scopus (8) Google Scholar Our study population included people from countries where a BCG vaccine is administered at least once since birth, with the overall prevalence of positive results among non-US-born participants being 43% for tuberculin skin tests (TSTs) and 22–27% for IFNγ release assays. We believe that this cross-reaction with BCG accounts for the large discordance seen between TST and IFNγ release assay results among our cohort born outside of the USA; this belief is also supported by the large UK PREDICT TB study.2Abubakar I Drobniewski F Southern J et al.Prognostic value of interferon-γ assays and tuberculin skin test in predicting the development of active tuberculosis (UK PREDICT TB): a prospective cohort study.Lancet Infect Dis. 2018; 18: 1077-1087Summary Full Text Full Text PDF PubMed Scopus (91) Google Scholar Although some discordance persists, results for US-born participants (younger than 5 years, living with HIV, experiencing homelessness, or close contacts of people with infectious tuberculosis disease) are more similar for all three tests. The US testing and treatment strategy for latent tuberculosis infection includes a focus on people born outside of the USA, from countries where tuberculosis disease is more common. These people have usually received a BCG vaccine, and thus the cross-reaction of BCG with TST is a common consideration. Without a gold standard for diagnosis of latent tuberculosis infection, it is hard to know for certain whether the TST is more sensitive or the IFNγ release assays are more specific, but all tests for latent tuberculosis infection miss a substantial proportion of immunocompetent people who subsequently progress to tuberculosis disease.3Stout JE Wu Y Ho CS et al.Evaluating latent tuberculosis infection diagnostics using latent class analysis.Thorax. 2018; 73: 1062-1070Crossref PubMed Scopus (21) Google Scholar In another publication, we used Bayesian latent class analysis of the combination test results to estimate performance characteristics of the three different tests for latent tuberculosis infection.3Stout JE Wu Y Ho CS et al.Evaluating latent tuberculosis infection diagnostics using latent class analysis.Thorax. 2018; 73: 1062-1070Crossref PubMed Scopus (21) Google Scholar We used published data on contacts with known tuberculosis exposure, test results, and subsequent progression to disease as initial estimates and refined them with the Bayesian analysis. Among those born outside the USA with a history of BCG vaccination, the TST had the lowest modelled specificity of 70% for latent tuberculosis infection. The TST modelled specificity for latent tuberculosis infection was 92% among US-born individuals who did not have HIV infection. Thus, the interpretation of test results needs to account for test sensitivity and specificity as affected by an individual's previous exposure to tuberculosis cross-reacting antigens and the strength of their underlying immune response. In people who are not contacts with known tuberculosis exposure but originate from countries with high tuberculosis rates and high BCG vaccination, IFNγ release assays offer a distinct advantage over TST. Moreover, we agree with Arend and Uzorka's statement that the risk of progression to tuberculosis is very low among those with negative IFNγ release assay results if they are immunocompetent. We declare no competing interests. Discordant results of tests for tuberculosis reconsideredWe compliment Christine Ho and colleagues on their study comparing three tests for latent tuberculosis infection in more than 22 000 individuals.1 They concluded that IFNγ release assays should be preferred over the tuberculin skin test (TST) for the diagnosis of latent tuberculosis infection. Full-Text PDF Comparison of three tests for latent tuberculosis infection in high-risk people in the USA: an observational cohort studyOur findings support the preferential use of IFN-γ release assays for the diagnosis of latent tuberculosis in high-risk populations, especially in very young and older people born outside the USA. Full-Text PDF