How urban characteristics affect vulnerability to heat and cold: A multi-country analysis

Francesco Sera, Ben Armstrong, Aurelio Tobias, Ana Maria Vicedo-Cabrera, Christofer Åström, Michelle L. Bell, Bing Yu Chen, Micheline De Sousa Zanotti Stagliorio Coelho, Patricia Matus Correa, Julio Cesar Cruz, Tran Ngoc Dang, Magali Hurtado-Diaz, Dung Do Van, Bertil Forsberg, Yue Leon Guo, Yuming Guo, Masahiro Hashizume, Yasushi Honda, Carmen Iñiguez, Jouni J.K. JaakkolaHaidong Kan, Ho Kim, Eric Lavigne, Paola Michelozzi, Nicolas Valdes Ortega, Samuel Osorio, Mathilde Pascal, Martina S. Ragettli, Niilo R.I. Ryti, Paulo Hilario Nascimento Saldiva, Joel Schwartz, Matteo Scortichini, Xerxes Seposo, Shilu Tong, Antonella Zanobetti, Antonio Gasparrini

Research output: Contribution to journalArticleResearchpeer-review

103 Citations (Scopus)


Background: The health burden associated with temperature is expected to increase due to a warming climate. Populations living in cities are likely to be particularly at risk, but the role of urban characteristics in modifying the direct effects of temperature on health is still unclear. In this contribution, we used a multi-country dataset to study effect modification of temperature-mortality relationships by a range of city-specific indicators. Methods: We collected ambient temperature and mortality daily time-series data for 340 cities in 22 countries, in periods between 1985 and 2014. Standardized measures of demographic, socio-economic, infrastructural and environmental indicators were derived from the Organisation for Economic Co-operation and Development (OECD) Regional and Metropolitan Database. We used distributed lag non-linear and multivariate meta-regression models to estimate fractions of mortality attributable to heat and cold (AF%) in each city, and to evaluate the effect modification of each indicator across cities. Results: Heat- and cold-related deaths amounted to 0.54% (95% confidence interval: 0.49 to 0.58%) and 6.05% (5.59 to 6.36%) of total deaths, respectively. Several city indicators modify the effect of heat, with a higher mortality impact associated with increases in population density, fine particles (PM2.5), gross domestic product (GDP) and Gini index (a measure of income inequality), whereas higher levels of green spaces were linked with a decreased effect of heat. Conclusions: This represents the largest study to date assessing the effect modification of temperature-mortality relationships. Evidence from this study can inform public-health interventions and urban planning under various climate-change and urban-development scenarios.

Original languageEnglish
Pages (from-to)1101-1112
Number of pages12
JournalInternational Journal of Epidemiology
Issue number4
Publication statusPublished - Aug 2019


  • cities
  • climate
  • epidemiology
  • heat
  • mortality
  • Temperature

Cite this