Background: While rapid response capacity has been instituted by many cities worldwide following recent catastrophic heat-wave events, the recognition that theoretically preventable heat-related deaths may occur throughout the summer has provoked much less notice and response.
Methods: We reviewed published estimates of the general summertime temperature-mortality relationship characterized in different settings around the world. Then, in order to explain heterogeneity of effects, we applied a random-effects meta-regression to the estimates in relation to a number of standardized city-level characteristics of demography, economy and climate.
Results: Estimates of heat thresholds and slopes were identified for 64 locations across 6 continents, with most evidence to date coming from North American and European populations. Heat thresholds were generally higher in communities closer to the equator, suggesting some degree of population adaptation. In almost half of the locations, the risk of mortality increased by between 1% - 3% per 1°C change in high temperature, including in cities where annual heat exposure may not previously have been perceived to be a problem. Heat impacts were noticeably small in many US cities. In random-effects meta-regression models, increasing population density, decreasing city GDP and increasing percentage of people aged 65 or more were all independently associated with an increase in the heat slope.
Conclusion: Improved care of the elderly, residential architecture and urban planning measures which help to reduce high temperatures in densely populated areas are likely to play a key role alongside more targeted heat-health warning systems in reducing future heat burdens.