People with immunoreactivity to tuberculosis are thought to have lifelong asymptomatic infection and remain at risk for active tuberculosis. Marcel A Behr and colleagues argue that most of these people are no longer infected.
Two billion people worldwide are thought to be asymptomatically (latently) infected with Mycobacterium tuberculosis and at risk of developing active tuberculosis (TB)
The prevalence of latent TB infection is inferred from tests that detect immunoreactivity to mycobacterial antigens rather than live bacteria and from mathematical modelling
Longitudinal studies and clinical trials show that this TB immunoreactivity can persist after curative treatment
Most people with TB immunoreactivity do not develop active TB upon immunosuppression, suggesting that they have cleared their infection while retaining immunological memory to it
TB immunoreactivity cannot distinguish cleared from persistent infection, emphasising the urgent need for tests that can identify people with asymptomatic infections
BCG and false positives
BCG vaccination can result in false positive results on TSTs in those vaccinated when older than 1 year of age.24 This could have falsely inflated the number of people estimated to be TB immunoreactive in the studies that we analysed. If so, the proportion of TB immunoreactive people who developed TB associated with immunosuppression would have been higher than our estimates. To ensure that this was not the case, for each study we determined the use of BCG vaccine in the appropriate country and whether the more specific IGRA tests25 were used to determine TB immunoreactivity. We concluded that our estimates of TB immunoreactivity were not falsely high because of BCG vaccination (see supplementary table 2).
Summary and implications
Whether infected people can clear M tuberculosis has consumed TB researchers for over a century, yet multiple approaches have failed to yield a definitive answer. Our analysis of epidemiological experiments suggests the answer: between 1% and 11% of people with TB immunoreactivity continue to harbour viable bacteria capable of causing disease. In the remainder, the organisms are either dead or have lost their pathogenic potential. The latter scenario would be extraordinary. It would differ from a multitude of other quiescent infections where immunosuppression is recognised to cause a reactivation or exacerbation of infection.26 Cytomegalovirus, herpes simplex virus, herpes zoster virus, the protozoan Toxoplasma gondii, and the fungus Cryptococcus neoformans are known to produce full blown devastating infections from a long term latent infection.
This analysis highlights that the currently available tests for latent TB infection detect only TB reactivity—that is, immunological memory and recall responses. So they would not be expected to distinguish between long lived memory that persists after elimination of M tuberculosis antigens and T cell responses maintained by repeated antigen stimulation from chronic infection. Many infectious diseases (including hepatitis A, B, and C) as well as immunisation with attenuated or subunit vaccines generate memory responses (antibody, T cell, or both) that do not reflect pathogen persistence.
We hope that our analysis will stimulate further discussion about the research and biological paradigms of TB. If infection is forever, researchers will aim to study host tolerance. If infection can be cleared, researchers will try to understand the mechanisms of pathogen elimination.
Our findings indicate that the tests used in biomarker and immunological studies of persistent infection (TST and IGRA)2728 might have two major confounders: the control participants who do not develop TB might be a heterogeneous mix of currently and previously infected people; and efforts to find a profile that predicts increased risk of disease might be confounded by the control group (TST positive) being at reduced risk of TB, as has been shown in multiple observational studies.29
Notwithstanding the challenges of evaluating new biomarkers, the importance of identifying the proportion of people with persistent infection and distinguishing them from those with immunological memory of past infection cannot be overstated. A test that could identify the 10% who are infected should reduce the cost and morbidity of