Stroke Triggered by Respiratory Viruses: Ecological Study of Temporal and Spatiotemporal Associations

BACKGROUND: Infections have been linked as short-term triggers for stroke, but the extent of their contribution is unclear. We examined population-level temporal and spatiotemporal associations between common respiratory viruses and stroke in an ecological study in Victoria, Australia. METHODS: The statewide Victorian Admitted Episodes data set, covering all hospital admissions for stroke between June 2010 and December 2022, was analyzed against statewide representative data from SnotWatch respiratory polymerase chain reaction tests. Population size, climate, and air pollution data sets for the same period were treated as population-level confounders. Time-series analysis was used to model weekly stroke events and respiratory virus levels (% of polymerase chain reaction tests positive for each pathogen), and attributable fractions were computed using distributed lag nonlinear modeling. To examine spatiotemporal variation, the weekly association by Statistical Area Level 2 was modeled using a Bayesian Integrated Nested Laplace Approximation. RESULTS: We analyzed 153 730 stroke events (61% ischemic, 25% hemorrhagic, and 14% unspecified) and 6 180 896 respiratory specimens. In the time-series analysis, higher weekly levels of specific respiratory viruses were associated with increased rates of stroke after controlling for confounders. Our model estimated that the fractions of strokes attributed to enterovirus and human metapneumovirus 4.6% (95% CI, 2.5-6.6) and 1.3% (95% CI, 0.3-2.3), respectively, which rose during peak viral weeks. The attributable fractions differed between ischemic stroke (human metapneumovirus 3.7%) and hemorrhagic stroke (enterovirus, 5.8%; parainfluenza, 3.2%). Spatiotemporal analysis revealed greater effect sizes in areas potentially associated with agricultural/industrial activity, as well as in outer-urban areas, based on the regionality of the Statistical Area Level 2 of usual patient residence. CONCLUSIONS: At the population level, there were clear, observable time- and space-time-related associations between respiratory virus activity and stroke. People living in higher-risk areas during peak respiratory virus periods may be more susceptible to acute stroke. This warrants further investigation.