Population Attributable Mortality Associated with Respiratory Viruses in Ontario

BackgroundRespiratory viruses are major contributors to population mortality, but cause-of-death coding undercounts their impact. Ecological regression models linking viral circulation to mortality fluctuations can address this limitation. AimTo estimate the population attributable fraction (PAF) of mortality associated with influenza A and B, respiratory syncytial virus (RSV), and SARS-CoV-2 in Ontario, Canada (1993-2025), and to characterize temporal changes in virus-attributable mortality across the pandemic transition. MethodsWe analysed monthly all-cause mortality data with laboratory surveillance indicators for influenza A, B, RSV, and SARS-CoV-2. Negative binomial models with secular trends, Fourier seasonal terms, and population offsets were fit for pre-pandemic (January 1993-February 2020) and combined pandemic (March 2020-February 2025) periods. PAFs were derived from counterfactual predictions setting viral coefficients to zero. Sensitivity analyses examined temporal stratification of the pandemic period (Public Health Emergency of International Concern [PHEIC] period: March 2020-April 2023; post-PHEIC: May 2023-February 2025), exclusion of the early pandemic period (March-June 2020), and models without Fourier seasonal adjustment. Wald tests compared coefficients across specifications. ResultsPre-pandemic, influenza A accounted for 1.8% (95% CI 1.4-2.3%) of mortality; influenza B showed no detectable impact. RSV demonstrated inverse associations in seasonally adjusted models but positive associations (PAF 1.9%, 95% CI 1.3-2.4%) without seasonal adjustment. Over the combined pandemic period (March 2020-February 2025; n=60 months), amid elevated baseline mortality (IRR 1.050, P=0.027), SARS-CoV-2 accounted for 6.1% (95% CI 4.2-8.0%) of deaths, approximately 4-fold the pre-pandemic influenza A burden, despite widespread vaccination and antiviral availability. Model-estimated SARS-CoV-2-attributable deaths closely matched reported COVID-19 deaths from Public Health Ontario over the same period. Temporal stratification identified a significant increase in SARS-CoV-2-attributable mortality in the post-PHEIC period (PAF 9.8%, 95% CI 1.1-17.7%; p=0.027), while post-PHEIC influenza A and B attributable fractions did not differ significantly from pre-pandemic baselines. Excluding March-June 2020 yielded a conservative SARS-CoV-2 PAF of 5.7% (95% CI 3.3-8.1%), confirming robustness of primary estimates. Meta-analyses showed substantial heterogeneity for influenza A (I{superscript 2}=92.8%) and RSV (I{superscript 2}=89.1%) across modeling approaches, but minimal heterogeneity for SARS-CoV-2 (I{superscript 2}=5.5%). ConclusionSARS-CoV-2 was associated with a 3-4-fold higher population mortality burden than seasonal influenza A despite available countermeasures. Post-PHEIC data suggest that the burden of respiratory virus mortality, including for influenza, has not returned to pre-pandemic levels, highlighting the continued importance of respiratory virus prevention strategies. Estimates for influenza A and RSV were sensitive to seasonal adjustment, highlighting the importance of modelling choices when quantifying virus-attributable mortality.