Tuberculosis detection and transmission : societal disruption, diagnostic innovation, and aerosol characterisation

<p dir="ltr"><b>Background</b></p><p dir="ltr">Tuberculosis (TB) remains a major global public health challenge. In 2024, an estimated 10.6 million people developed TB worldwide, yet approximately 22% of individuals with incident TB were not diagnosed or notified, contributing to ongoing morbidity, mortality, and community transmission. Slow but consistent progress towards United Nations and World Health Organization targets was badly affected by the COVID-19 pandemic and continues to be at risk from dramatic reductions in international aid commitments. Improved understanding of Mtb transmission and individual infectiousness is essential to better target interventions and reduce preventable transmission.</p><p dir="ltr"><b>Aim</b></p><p dir="ltr">The overall aim of this thesis is to improve understanding of tuberculosis infectiousness by investigating its determinants and evaluating practical proxies using both population-level analyses and individual-level measures, including aerosolised Mycobacterium tuberculosis (Mtb) detection and radiographic severity.</p><p dir="ltr"><b>Methods</b></p><p dir="ltr">This thesis comprises four quantitative studies applying population- and individual-level analyses. Study I used global TB notification data (2014-2020) and time-series modelling to estimate changes in childhood TB notifications during the COVID-19 pandemic and examine associations with government-imposed social restrictions. Study II and III collected prospective clinical data in Cape Town, South Africa, to evaluate detection of aerosolised Mtb using a novel electrostatic sampler and to identify host, clinical, and behavioural factors associated with aerosol detection. Study IV analysed multi-country data to assess associations between computer-aided detection (CAD)-interpreted chest x-ray scores and sputum bacillary load using regression and predictive modelling approaches.</p><p dir="ltr"><b>Results</b></p><p dir="ltr">Global TB notifications increased between 2014 and 2019, particularly among children, but declined sharply in 2020, with larger relative shortfalls in younger age groups compared with adults. Observed notifications in 2020 were consistently lower than counterfactual predictions, and no association was identified between the magnitude of notification shortfalls and government-imposed COVID-19 social restrictions. At the individual level, viable aerosolised Mtb was detected in a subset of people with untreated pulmonary TB, predominantly during forced coughing and within respirable particle sizes, and was associated with higher sputum bacillary load, cough frequency, and particle generation. Sputum bacillary burden was positively associated with detection of aerosolised Mtb DNA in primary care settings with evidence of effect modification by sex. As a diagnostic test, aerosolised Mtb DNA detection had limited sensitivity when compared with sputum Xpert MTB/RIF Ultra testing. Across a multi-country cohort, CAD-interpreted chest x-ray scores are independently associated with sputum bacillary burden in people with pulmonary TB. Multivariable models incorporating CAD scores with clinical factors show promise for triage applications, particularly ruling out high sputum bacterial loads.</p><p dir="ltr"><b>Conclusions</b></p><p dir="ltr">This thesis highlights substantial heterogeneity in TB infectiousness and demonstrates the strengths and limitations of commonly used proxies. Population-level analyses show that health-system disruption substantially influences TB notifications, limiting their usefulness as real-time indicators of transmission under such conditions. At the individual level, aerosolised Mtb detection and radiographic severity were associated with sputum bacillary burden but exhibited marked variability. Together, these findings emphasise the complexity of measuring infectiousness and the need for complementary approaches that better capture underlying transmission processes.</p><h3 dir="ltr">List of scientific papers</h3><p dir="ltr">I. Ranasinghe L*, <b>Achar J*</b>, Gröschel MI, Whittaker E, Dodd PJ, Seddon JA. Global impact of COVID-19 on childhood tuberculosis: an analysis of notification data. Lancet Glob Health. 2022;10(12):e1774-e1781. * Joint first authors <a href="https://doi.org/10.1016/S2214-109X(22)00414-4">https://doi.org/10.1016/S2214-109X(22)00414-4</a></p><p dir="ltr">II. <b>Achar J,</b> van Schalkwyk J, Mahlobo Z, et al. Determinants of Viable Aerosolised Mycobacterium tuberculosis Detection and Evaluation of Electrostatic DNA Detection in Untreated Pulmonary Tuberculosis [Manuscript]</p><p dir="ltr">III. <b>Achar J,</b> Venter R, van Schalkwyk J, et al. Detection of Aerosolized Mycobacterium tuberculosis DNA From Adults Being Investigated for Pulmonary Tuberculosis via an Electrostatic Sampler in a South African Primary Care Setting. Open Forum Infect Dis. 2025;12(10):ofaf593. Published 2025 Sep 24. <a href="https://doi.org/10.1093/ofid/ofaf593" rel="noreferrer" target="_blank">https://doi.org/10.1093/ofid/ofaf593</a></p><p dir="ltr">IV. <b>Achar J,</b> Seddon JA, Crowder R et al. Digital chest X-ray scores with computer-aided detection for tuberculosis and their relationship with sputum bacterial load: A multi-country cross- sectional study. [Manuscript]</p>