High-mortality days during the winter season: comparing meteorological conditions across 5 US cities

Despite the high burden from exposure to both hot and cold weather each year in England and Wales, there has been relatively little investigation on who is most at risk, resulting in uncertainties in informing government interventions. To determine the subgroups of the population that are most vulnerable to heat-related and cold-related mortality. Ecological time-series study of daily mortality in all regions of England and Wales between 1993 and 2003, with postcode linkage of individual deaths to a UK database of all care and nursing homes, and 2001 UK census small-area indicators. A risk of mortality was observed for both heat and cold exposure in all regions, with the strongest heat effects in London and strongest cold effects in the Eastern region. For all regions, a mean relative risk of 1.03 (95% confidence interval (CI) 1.02 to 1.03) was estimated per degree increase above the heat threshold, defined as the 95th centile of the temperature distribution in each region, and 1.06 (95% CI 1.05 to 1.06) per degree decrease below the cold threshold (set at the 5th centile). Elderly people, particularly those in nursing and care homes, were most vulnerable. The greatest risk of heat mortality was observed for respiratory and external causes, and in women, which remained after control for age. Vulnerability to either heat or cold was not modified by deprivation, except in rural populations where cold effects were slightly stronger in more deprived areas. Interventions to reduce vulnerability to both hot and cold weather should target all elderly people. Specific interventions should also be developed for people in nursing and care homes as heat illness is easily preventable.

Weather-related health effects have attracted renewed interest because of the observed and predicted climate change. The authors studied the short-term effects of cold weather on mortality in 15 European cities. The effects of minimum apparent temperature on cause- and age-specific daily mortality were assessed for the cold season (October-March) by using data from 1990-2000. For city-specific analysis, the authors used Poisson regression and distributed lag models, controlling for potential confounders. Meta-regression models summarized the results and explored heterogeneity. A 1 degrees C decrease in temperature was associated with a 1.35% (95% confidence interval (CI): 1.16, 1.53) increase in the daily number of total natural deaths and a 1.72% (95% CI: 1.44, 2.01), 3.30% (95% CI: 2.61, 3.99), and 1.25% (95% CI: 0.77, 1.73) increase in cardiovascular, respiratory, and cerebrovascular deaths, respectively. The increase was greater for the older age groups. The cold effect was found to be greater in warmer (southern) cities and persisted up to 23 days, with no evidence of mortality displacement. Cold-related mortality is an important public health problem across Europe. It should not be underestimated by public health authorities because of the recent focus on heat-wave episodes.

Ambient temperature is an important determinant of daily mortality that is of interest both in its own right and as a confounder of other determinants investigated using time-series regressions, in particular, air pollution. The temperature-mortality relationship is often found to be substantially nonlinear and to persist (but change shape) with increasing lag. We review and extend models for such nonlinear multilag forms. Popular models for mortality by temperature at given lag include polynomial and natural cubic spline curves, and the simple but more easily interpreted linear thresholds model, comprising linear relationships for temperatures below and above thresholds and a flat middle section. Most published analyses that have allowed the relationship to persist over multiple lags have done so by assuming that spline or threshold models apply to mean temperature in several lag strata (e.g., lags 0-1, 2-6, and 7-13). However, more flexible models are possible, and a modeling framework using products of basis functions ("cross-basis" functions) suggests a wide range, some used previously and some new. These allow for stepped or smooth changes in the model coefficients as lags increase. Applying a range of models to data from London suggest evidence for relationships up to at least 2 weeks' lag, with smooth models fitting best but lag-stratified threshold models allowing the most direct interpretation. A wide range of multilag nonlinear temperature-mortality relationships can be modeled. More awareness of options should improve investigation of these relationships and help control for confounding by them.