It is with much interest to read ‘Under the weather with stroke; more data emerge’ by Dawson et al. (1). Following their observation and summary, more discussions with regard to epidemiological methods are needed because the relationship between stroke and meteorological variables is complex, and a number of potentially relevant meteorological variables, differences in study design and climate between geographical areas and potential for confounding are seen (2). In line with the request of well-conducted prospective studies to apply and clarify confusing and conflicting data worldwide, it is argued that using a regression model with controlling for possible confounders such as humidity and wind speed in previous epidemiological studies is inappropriate. Confounding occurs when effect measures confound the study outcome, which results in before–after adjustment difference not being equal (3). Therefore, examining air temperature and stroke events adjusting for humidity and/or wind speed is based on the assumption that the two factors, humidity and wind speed, are associated with the risk of stroke. However, it is yet to be clearly addressed. On the other hand, taking only air temperature in the regression model is not accurate because other climatic factors to some extent impact on thermal environment close to human body experience. Like Xun et al. (4) recently addressed, climate change studies require epidemiology to endeavour to identify research needs and gaps in empirical knowledge and prioritise according to need, to develop methodologies that address the ‘real-world’ complexities. Quantification of climate and risk of stroke is, in fact, more complicated than anticipated.


  1. Top of page
  2. References