Death rates at midlatitudes tend to be highest in winter, when people are confined indoors and therefore are more exposed to infectious diseases. Human and animal mortality is high also during heat waves.
There are various definitions of heat waves, but a common one is a series of days with maxima over 35° C. Heat waves can cause great discomfort to people, and in the extreme they can cause heat stroke, which involves brain dysfunction (e.g. a stroke or convulsions), hot and dry skin, and a body-core temperature above 40° C (1). Heat stress should be measured in terms of its effect of the human energy balance, e.g. the apparent temperature or heat index. Heat-related deaths are most common amongst the elderly and sick people, and they tend to occur one day after great heat, perhaps because of the time taken to modify the blood’s composition in a way that precipitates heart or brain thrombosis.
In Australia and the USA, heat waves kill more people than any other weather-related disaster (1, 3). Moreover, it is estimated that for every death specifically attributed to high temperatures, there are ten more due to the heat aggravating a pre-existing illness.
Heat deaths and measured summer temperatures are strongly correlated in many parts of the USA but not in the southeastern states (the 'South') (4). One could hypothesise that many heat-related deaths in the South, where high dewpoints and maxima over 30° C are common from May onwards, occur earlier in the year. The 'effectiveness' of a heat wave on mortality decreases as summer progresses, because vulnerable people have already succumbed.
The daily-mean heat index threshold which deaths increase does vary between cities, and increases during April and May; values from 12 southern cities ranged between 19-32° C. However conditions in spring do not explain the anomaly, i.e. the hypothesis is wrong. It appears more likely that people living in the warmth of the south are less susceptible to heat waves because they have become acclimatised, or live in houses better adapted to hot conditions. In other words, heat deaths are most in cities subject to intense heat occurring irregularly.
In Buenos Aires a close relationship exists between high daily minimum temperatures and the rate of emergency hospitalisation (5). Skin and allergy problems in particular correlate with heat stress. Heart, lung and stomach emergencies correlated better with wind direction, either easterly (humid) or westerly (dry), than with temperature.
A direct impact of heat on health has also been observed in Tokyo in Japan (with a population of 12 million) and in Nanjing (2.6 million) and Wuhan (3.7 million) in China (6). There is a rapid increase of heat deaths amongst the elderly in summer, above threshold daily maximum temperatures of 32° C in Tokyo, and 36° C in China.
People in cities are especially prone to heat deaths because of additional urban heating, less air circulation due to congested conditions in slums, less access to air conditioning, less awareness, and the more common poor health of slum-dwellers. The threshold danger temperature, above which heat deaths become noticeable, depends on the ambient humidity and the local average summertime daytime maximum temperature. Studies in USA yield a value of around 35° C for a threshold maximum temperature. Data from five cities in Australia show that about 100 deaths (directly and indirectly heat-related) occurred annually between 1979 - 1990 when the daytime maximum was above 36° C (1). However, a graph of the death rates plotted against temperatures shows no increase above any particular value, i.e. unlike in the US and other countries, there is no particular threshold temperature in Australia (7). The reason for this might perhaps be a relative absence of slums in Australia.
Global warming is likely to dramatically increase the incidence of heat deaths, because even a slight increase in mean temperature implies a relatively large increase in extremely warm conditions.