Mortality in the elderly and ambient ozone concentration during the hot summer, 1994, in Belgium

Extensive investigations were carried out to study the relationship between daily mortality in the elderly, outdoor air temperature, and ozone concentration observed in Belgium during the hot summer, 1994. The two environmental variables were assessed through mean daily temperature and 24-hr ozone concentration, both measured the day before and averaged over the country. Data were stratified by terciles of mean daily temperature in order to reduce the degree of collinearity between the investigated environmental variables. In the first stratum, which ranged from 9.9 to 15.4 degrees C (41 days), mean daily temperature and 24-hr ozone concentration were not correlated while the mean number of daily deaths was higher when 24-hr ozone concentration increased from 45 to 55 micrograms/m3 (P < 0.05). In the second stratum, which ranged from 15.6 to 20.3 degrees C (42 days), mean daily temperature and 24-hr ozone concentration were strongly correlated (r = 0.54, P < 0.0001). In this stratum, the number of daily deaths did not depend on the mean daily temperature but increased linearly with 24-hr ozone concentration within the range 25 to 85.5 micrograms/m3 (P < 0.001). After having examined the possible confounding effect of sulfur dioxide, nitrogen dioxide, fine particulates, and humidity, ozone was found to be the only investigated variable contributing to the increased daily mortality. In the third stratum, which ranged from 20.4 to 27.6 degrees C (40 days), mean daily temperature and 24-hr ozone concentration were also strongly correlated (r = 0.71, P < 0.0001). Daily mortality, in this stratum, was correlated more with mean daily temperature (r = 0.68, P < 0.001) than with 24-hr ozone concentration (r = 0.55, P < 0.001). Nonparametric regression analyses were performed to model the number of daily deaths in the whole range of temperatures. These analyses confirmed the effect of 24-hr ozone concentration on daily mortality already uncovered by the least-squares regression analysis in the second stratum of mean daily temperature. In addition, at levels exceeding 20 degrees C, the effect of ozone concentration on daily mortality was enhanced by temperature owing to a positive interaction between these two variables. The present study thus demonstrated a statistical association between daily mortality, observed in the elderly during the hot summer, 1994, in Belgium, and ambient ozone concentration. This relationship was dependent on the range of temperatures.