Seasonal variation in cardiovascular mortality and hospitalisations and its implications for public health planning: a prospective cohort study

Background We characterised seasonal patterns in cardiovascular (CV) mortality and hospitalisations to understand how CV health might be affected by climate change.Methods We analysed the UK Biobank cohort recruited from 22 centres between 2006 and 2010. Seasonal mortality rates and hospitalisation rates were determined for death (all cause, CV and non-CV death) and first recorded of CV hospitalisations (heart failure (HF), myocardial infarction (MI), stroke and peripheral vascular disease (PVD)). Seasonal trends were evaluated using a cosinor model with a binomial family and by sex, age, ethnicity and socioeconomic deprivation. Multivariable models were adjusted for demographic, socioeconomic, lifestyle and clinical covariates, and assessment centre was subsequently included as a random intercept to account for clustering within centres.Results Data from 453 597 participants were included, 56% were women, with a mean age of 56 years (SD 8). There were 33 857 deaths (10 807 (31.9%) were CV and 23 050 (68.1%) non-CV deaths). All-cause mortality showed evidence of seasonal variation, and CV and non-CV deaths had significant seasonal variation (p<0.001), with the highest rates in winter. The maximum and minimum of all-cause mortality rates were 62.2 and 51.5 per 10 000 person-years in winter and summer respectively. There were 12 273, 9520, 7137, 14 492 hospitalisations for HF, MI, stroke and PVD. MI and stroke showed clear winter peaks (p<0.001), HF and PVD also displayed seasonal variation (p<0.001) with the highest hospitalisation rate in spring.Conclusions Rates of deaths, MI and stroke hospitalisations were highest during winter months. The findings imply that colder temperatures rather than warmer temperatures are a more significant driver of adverse outcomes.