The research, published in PLOS Global Public Health, was led by Hannah M. Rickman of the Malawi-Liverpool-Wellcome Trust clinical research programme together with collaborators from the Malawi Ministry of Health, the London School of Hygiene & Tropical Medicine and the University of Liverpool. The team set out to answer a deceptively simple question: where in a large African city is Mycobacterium tuberculosis, the bacterium that causes tuberculosis, actively being transmitted right now? Their answer came not from counting sick adults, but from measuring silent infection in healthy toddlers.

The logic behind the approach rests on a fundamental feature of TB epidemiology. Young children who test positive for immune reactivity against M. tuberculosis must have been infected recently, because the window between birth and testing is short. Unlike adults, who may carry latent infection acquired decades earlier, children aged one to four years serve as sensitive biological sensors of ongoing transmission in their immediate surroundings. Every positive result in a toddler points to an undiagnosed infectious case somewhere nearby, most likely an adult in the household or neighbourhood whose illness has never reached a clinic.

To exploit this signal, the researchers conducted a cross-sectional survey across 33 high-burden neighbourhoods of Blantyre, a city of roughly one million people in southern Malawi. They enrolled 1,545 healthy children between one and four years of age using two complementary recruitment strategies. Some families were approached through health facilities, where children attend for routine immunisations and growth monitoring, while others were identified through random household sampling guided by satellite imagery of the urban landscape. This hybrid design was deliberately pragmatic, designed to be affordable and repeatable by TB programmes in resource-limited settings rather than requiring the expensive logistics of a classic disease prevalence survey.

Each child provided a small blood sample that was tested with an interferon-gamma release assay, or IGRA, a laboratory test that detects the immune response of T cells to proteins specific to M. tuberculosis. Unlike tuberculin skin tests, IGRAs are not confounded by prior vaccination with bacille Calmette-Guérin, the live vaccine given at birth across most of sub-Saharan Africa. A positive QFT-Plus result therefore indicates genuine immunological exposure to the tuberculosis bacterium, whether the child has cleared the infection, is carrying it latently, or in rare cases has progressed to active disease.

The headline finding was sobering. Of the 1,545 children tested, 102, or 6.6 percent, had a positive IGRA result. Because infection accumulates over the first years of life, the researchers converted this prevalence into an annual risk of M. tuberculosis infection of 2.7 percent, with a 95 percent credible interval of 2.2 to 3.2 percent. In practical terms, roughly one in every 37 children in these neighbourhoods acquires a new tuberculosis infection each year, a rate that speaks to a level of ongoing community transmission far higher than most residents or health workers would suspect from clinic-based figures alone.

The social gradient in the results was equally telling. Children living in the poorest third of households were substantially more likely to be immunoreactive than those in the richest third, with 8.7 percent positivity compared with 4.9 percent. After adjusting for other factors in a hierarchical Bayesian logistic regression model that accounted for clustering within neighbourhoods, the poorest children had nearly double the odds of a positive result, with an adjusted odds ratio of 1.88. Notably, the risk was not explained by the usual suspects of paediatric TB vulnerability: exposure to HIV, malnutrition, and reported contact with a household member with TB all showed no significant association with immunoreactivity in this analysis.

Perhaps the most consequential finding concerned geography. The researchers observed substantial heterogeneity between neighbourhoods, with estimated annual risks of infection ranging from 1.1 percent to 4.1 percent across the 33 communities studied. Yet when they compared these neighbourhood-level infection risks against official TB case notification rates, the correlation was absent. Areas reporting the most tuberculosis cases to health authorities were not the areas where children were most likely to be acquiring infection. The authors attribute this disconnect to disparities in healthcare access: case notifications reflect who reaches a clinic and gets diagnosed, not who is actually transmitting the disease. In a city where diagnostic capacity, transport costs and health-seeking behaviour vary sharply from one district to the next, notification data can point programmes toward the wrong neighbourhoods entirely.

The statistical machinery behind these conclusions was designed to handle sparse data and clustered sampling. The hierarchical Bayesian logistic regression framework allowed the team to estimate infection risk at both the individual and neighbourhood levels simultaneously, borrowing strength across communities while quantifying uncertainty through credible intervals rather than single point estimates. Neighbourhood-level random effects captured the unexplained variation between communities, and the model structure made it possible to test whether household wealth, HIV exposure, nutritional status, reported TB contact or notification rates predicted positivity after accounting for this spatial clustering. This approach, the authors argue, is what makes the survey a template for surveillance rather than a one-off research exercise.

The implications reach well beyond Blantyre. Tuberculosis prevalence surveys, in which teams screen entire populations for active disease, were once the gold standard for measuring epidemic trends, but they have become prohibitively expensive as countries approach elimination targets and disease becomes rare. Passive surveillance through case notifications, meanwhile, is systematically biased by unequal access to care. Measuring immunoreactivity in young children offers a third way: an indirect but biologically grounded readout of recent transmission that requires only a blood test, a laboratory and a sampling frame. The Blantyre study demonstrates that such surveys can be run at scale using convenience recruitment in clinics supplemented by satellite-guided household sampling, keeping costs within reach of national TB programmes.

For the children of Blantyre, the findings carry an immediate message. The burden of recent tuberculosis infection is high, concentrated among the poorest families, and invisible to the surveillance systems currently guiding resource allocation. The authors argue that M. tuberculosis immunoreactivity surveys could help programmes identify high-risk populations, monitor trends over time and target interventions such as active case finding and preventive therapy to the neighbourhoods where transmission actually occurs. In a global campaign that has pledged to end the tuberculosis epidemic, knowing precisely where the epidemic lives may prove to be the most valuable diagnostic of all, and it may be found in a few millilitres of blood from a healthy two-year-old.

Subject of Research: Community-based interferon-gamma release assay survey measuring Mycobacterium tuberculosis immunoreactivity and transmission risk in young children in Blantyre, Malawi

Article Title: Heterogeneity in Mycobacterium tuberculosis immunoreactivity in young children in Blantyre, Malawi: A community-based survey

Article References: Rickman, H. M., Phiri, M. D., Mbale, H., Feasey, H. R. A., Nliwasa, M., Chagaluka, G., Seddon, J. A., Mwandumba, H. C., Horton, K. C., Henrion, M. Y. R., Mwenyenkulu, T., Mbendera, K. N., Nightingale, E. S., Corbett, E. L., & MacPherson, P. (2026). Heterogeneity in Mycobacterium tuberculosis immunoreactivity in young children in Blantyre, Malawi: A community-based survey. PLOS Global Public Health, 6(10), e0007269. https://doi.org/10.1371/journal.pgph.0007269

Image Credits: AI Generated

DOI: 10.1371/journal.pgph.0007269

Keywords: tuberculosis, Mycobacterium tuberculosis, IGRA, Malawi, Blantyre, annual risk of infection, children, surveillance, health inequity, Bayesian modelling, public health, epidemiology