Short-term fluctuations in death by cause, temperature, and income in the United States, 1930-1985.

Disclosures that this decade has had the five hottest years ever recorded globally raise concern that extreme temperatures might be associated with higher mortality. An analysis of fluctuations in annual case-specific deaths, seasonal temperatures, and annual income per capita in Massachusetts, Michigan, Washington, Utah, North Carolina, and Mississippi, 1930 to 1985, suggests that, on the contrary, a temperature increase throughout the year was associated with fewer deaths from all causes combined, including deaths from infectious diseases, heart diseases, cerebrovascular diseases, pneumonia, and influenza. An average temperature increase of one degree Fahrenheit was associated with a more than 2 per cent decline in deaths from pneumonia and influenza. The only category of deaths showing no significant association was death from malignant neoplasms. Compared to spring, summer, and fall temperature fluctuations, unusually cold winter temperatures had the strongest fatal effects, but only in North Carolina and Mississippi. The greatest cumulative temperature effects on mortality were found in the same two states. Controlling for annual fluctuations in income per capita did not influence the relationship between temperature and mortality. There was evidence suggesting that the level of wealth ameliorated the fatal effects of extreme temperatures. In conclusion, unusually warm weather was followed by fewer deaths; unusually cold weather, by more deaths.