Know your tuberculosis epidemic-Is it time to add Mycobacterium tuberculosis immunoreactivity back into global surveillance?

Abstract

Tuberculosis (TB) still causes 1.5 million deaths globally each year. Over recent decades, slow and uneven declines in TB incidence have resulted in a falling prevalence of TB disease, which increasingly concentrates in vulnerable populations. Falling prevalence, while welcome, poses new challenges for TB surveillance. Cross-sectional disease surveys require very large sample sizes to accurately estimate disease burden, and even more participants to detect trends over time or identify high-risk areas or populations, making them prohibitively resource-intensive. In the past, tuberculin skin surveys measuring Mycobacterium tuberculosis (Mtb) immunoreactivity were widely used to monitor TB epidemiology in high-incidence settings, but were limited by challenges with both delivering and interpreting the test. Here we argue that the shifting epidemiology of tuberculosis, and the development of new tests for Mtb infection, make it timely and important to revisit the strategy of TB surveillance based on infection or immunoreactivity. Mtb infection surveys carry their own operational challenges and fundamental questions, for example: around survey design and frequency; which groups should be included; how the prevalence of immunoreactivity in a population should be used to estimate force of infection; how individual results should be interpreted and managed; and how surveillance can be delivered efficiently and ethically. However, if these knowledge gaps are addressed, the relative feasibility and lower costs of Mtb infection surveillance offer a powerful and affordable opportunity to better "know your TB epidemic", understand trends, identify high-risk and underserved communities, and tailor public health responses to dynamic epidemiology.

Fig 1. Simplified diagram of the causal stages between TB transmission, case notification and mortality.

Data sources (bottom) may capture more proximal or distal stages of the process. As methods capture later stages, it becomes increasingly challenging to draw inferences about the original causative transmission event.

Fig 2

a. Age-dependent cross-sectional population prevalence of Mycobacterium tuberculosis (Mtb) immunoreactivity under differing epidemiological assumptions: i) Constant annual risk of tuberculosis infection (ARTI) of 1%, with no reversion, secular trends or age-related trends; ii) Childhood ARTI of 1% which increases from age 12 to a maximum of 4% in adulthood [58]; iii) Current ARTI of 1% in all age groups on a background of a secular decline in ARTI of 2% per year[4]; iv) Constant ARTI of 1%, but with 5% annual reversion [59]; v) Constant ARTI of 1%, assuming 10% of the population are resistant to infection and remain immune-nonreactive [60]. b. The degree to which age-specific ARTIs would be under- or over-estimated in each scenario, by age group sampled, if a constant rate of infection with no reversion were assumed (i.e. using ARTI = 1−(1−Prevalence) ^{1/Mean age}).