Investigators examined first-coded acute-care presentations for tuberculosis (TB) and nontuberculous mycobacteria (NTM) in the Greater Toronto/Hamilton Area Plus (GTHA+), Ontario, over April 2011–March 2024. The dataset combined hospital, emergency-department and ambulatory clinic administrative records; the authors identified 3,985 first TB-coded presentations and 5,214 first NTM-coded presentations across the study period. NTM served as an epidemiologic comparator because it traverses similar diagnostic pathways but has different transmission and risk profiles than TB.
The study used a population-based ecological time-series approach. The team applied negative binomial projection and full-series models to estimate pandemic-era changes under three alternate assumptions about underlying trends; in total six model specifications were reported. To test temporal associations between respiratory-virus circulation and mycobacterial presentations, the authors fitted distributed lag non-linear models to estimate cumulative rate ratios for standardized community viral activity over lags of 0–12 weeks. Sensitivity analyses extended the maximum lag to 26 and 52 weeks. SARS-CoV-2 was the primary viral exposure of interest; influenza and respiratory syncytial virus (RSV) were prespecified comparators. Data for five additional viral exposures were obtained after the primary analysis and analyzed as exploratory comparisons.
Across the six model specifications, estimated pandemic-era increases in TB-coded acute-care presentations ranged from 3% to 22%. The size of the estimated increase depended on the underlying trend assumptions embedded in the projection models. By contrast, none of the models suggested an increase in NTM-coded presentations during the pandemic era. The authors therefore reported a modest population-level rise in TB presentations that was not mirrored by NTM when using the same administrative data streams and analytic framework.
Time-series modelling did not find a positive association between community SARS-CoV-2 activity and subsequent TB presentations. The reported cumulative rate ratio per 1-standard-deviation increase in SARS-CoV-2 activity over 0–12 weeks was 0.909 (95% CI 0.822 to 1.006) for TB and 0.953 (0.868 to 1.047) for NTM; the study team observed no positive associations in subgroup analyses. Influenza activity was not associated with TB or NTM in prespecified comparisons. RSV activity was associated with both TB and NTM in the distributed lag analyses; among the prespecified comparisons only the RSV–NTM association met the Bonferroni threshold. The five additional viral exposures examined after the prespecified analysis did not meet experiment-wide significance thresholds in exploratory analyses.
Sensitivity analyses extended lag windows to 26 and 52 weeks; the authors reported that no positive associations emerged for SARS-CoV-2 even with these longer maximum lags. Subgroup analyses similarly failed to identify positive temporal associations between community SARS-CoV-2 activity and later TB or NTM presentations. The team emphasized that results were robust to the alternative lag specifications reported in the source.
This work used de-identified administrative health data accessed within the Canadian Institute for Health Information Secure Access Environment. The study received Research Ethics Board approval from the University of Toronto (Protocol #41690). Analysis code is publicly archived (DOI: 10.5281/zenodo.22216420) and an interactive scenario tool is available online and archived at the same DOI. Respiratory-virus surveillance data used in the analyses are publicly available through the Public Health Agency of Canada, according to the source.
The authors caution that the ecological, population-level design cannot exclude individual-level effects; in other words, a lack of positive association at the community scale does not prove absence of SARS-CoV-2–related risk for individual patients. The estimated magnitude of the pandemic-era TB increase depended on model assumptions about preexisting trends, which the authors explicitly report as a source of uncertainty. The preprint status of the report indicates the findings have not been peer reviewed. The study leaves unresolved whether individual SARS-CoV-2 infections alter TB risk at the person level, and the source did not report individual-level clinical or treatment data that could address that question. The authors noted competing interest declarations and provided funder information; detailed patient-level data remain controlled within the CIHI Secure Access Environment and require separate access requests.