Temporal trends of the association between ambient temperature and hospitalisations for cardiovascular diseases in Queensland, Australia from 1995 to 2016: A time-stratified case-crossover study

Background In the context of global warming, studies have turned to assess the temporal trend of the association between temperature and health outcomes, which can be used to reflect whether human beings have adapted to the local temperature. However, most studies have only focused on hot temperature and mortality. We aim to investigate the temporal variations in the association between ambient temperature and hospitalisations for cardiovascular diseases in Queensland, Australia from 1995 to 2016. Methods and findings We obtained data on 1,855,717 cardiovascular hospitalisations (mean age: 65.9 years, 42.7% female) from all 443 postal areas in Queensland, Australia between January 1, 1995 and December 31, 2016. Grid-level meteorological data were downloaded from scientific information for landowners. We used a time-stratified case-crossover design fitted with a conditional quasi-Poisson regression model and time-varying distributed lag nonlinear model (DLNM) to evaluate the association between temperature and cardiovascular hospitalisations and the temporal trends of the associations. Stratified analyses were performed in different age, sex, and climate zones. In all groups, relative risks (RRs) of cardiovascular hospitalisations associated with high temperatures (heat effects) increased, but cold effects showed a decreasing trend from 1995 to 2016. The increasing magnitude of heat effects was larger (p= 0.002) in men than in women and larger (p< 0.001) in people aged <= 69 years than in those aged >= 70 years. There was no apparent difference amongst different climate zones. The study was limited by the switch from ICD-9 to ICD-10 coding systems, by being unable to separate first-time hospitalisation from repeated hospitalisations, and possibly by confounding by air pollution or by influenza infections. Conclusion The impacts of cold temperatures on cardiovascular hospitalisations have decreased, but the impacts of high temperatures have increased in Queensland, Australia. The findings highlight that Queensland people have adapted to the impacts of cold temperatures, but not high temperatures. The burden of cardiovascular hospitalisations due to high temperatures is likely to increase in the context of global warming. Author summaryWhy was this study done? Studies have reported temporal decline in the association between hot temperature and cardiovascular mortality, which indicates that people may have adapted to the warming climate. However, little is known about the change in the association between ambient temperature and cardiovascular hospitalisation in recent decades. What did the researchers do and find? Using 1.9 million cardiovascular hospitalisations during 1995-2016 in Queensland, Australia, we developed a model to evaluate the association between temperature and cardiovascular hospitalisations and their trend over time. We found that the impact of cold temperature on cardiovascular hospitalisation decreased, but the impact of high temperature on cardiovascular hospitalisations increased from 1995 to 2016. The results were robust amongst different subgroups. The increasing heat effect was more serious for men than women and for people <= 69 years than people >= 70 years. What do these findings mean? Given the increased associations between hot temperature and cardiovascular hospitalisation, the acceleration of global warming, and an aging population, we highlight a worrying trend that the burden of cardiovascular healthcare facilities is to increase over time on high-temperature days. Heat warning systems should focus more on men and people <= 69 years old, since heat effects on these groups increased more noticeably than for women and people >= 70 years old, respectively.