Kobashi et al1recently reported on the sensitivity of the QuantiFERON TB-2G (QFT-2G) [Cellestis Ltd; Carnegie, VIC, Australia] interferon-γ release assay (IGRA) in 30 subjects aged ≥80 years with active tuberculosis (TB) [77%], with similar results than a previous study by Mori et al2(n = 10; sensitivity, 80%), thus confirming the higher sensitivity of QFT-2G vs tuberculin skin testing (TST) in this population. We disagree, however, with the conclusion that the IGRA can be used as a diagnostic method for TB in elderly patients. The IGRA and TST are markers of infection by Mycobacterium tuberculosis complex, but are not reliable markers of activity of the disease. Although interferon-γ levels do reflect to some extent disease activity, IGRAs should not be used in a clinical setting to distinguish TB from latent tuberculosis infection (LTBI) and are not recommended in this indication.3 The infection rate by M tuberculosis in individuals aged ≥80 years (LTBI) in most developed countries is approximately 80%; thus, a positive test result only emphasizes the epidemiologic fact that most elderly patients were infected earlier in life.4 Conversely, a negative test result cannot rule out TB in elderly subjects.5–Figure 1
shows the relationship between pretest and posttest probability of disease with a sensitivity of 77% and a specificity of 97% (ie, that of QFT-2G), according to the Bayes theorem.6 Because the IGRA yield positive results in TB and LTBI, pretest probability is based on estimated prevalence of LTBI in the population studied (ie, approximately 80%). As illustrated, with a negative QFT-2G result, the posttest probability of either LTBI or TB is still 50%. Assuming a hypothetical lower LTBI infection rate in Japan in the very old (40 to 60%), the posttest probability of disease is still 10 to 30% with a negative IGRA result. The IGRA should not be used in the very old as a diagnostic test for TB; the diagnosis must rely on clinical and radiologic evaluation and adequate bacteriologic sampling.