Major heat waves of 2003 and 2006 and health outcomes in Prague

We have investigated the association between heat waves and mortality and hospital admissions for Prague inhabitants for the summer heat waves of August 2003 and July 2006. The effect of heat waves was investigated using negative binomial regression in a generalized additive model. We used a linear model on a logarithmic scale, having 1-day lagged temperature differences from the long-term average, 1-day lagged ambient O-3 and PM10 concentration, relative humidity, simple heat wave indicator, and smooth seasonal effect as explanatory variables. We found a small increase in daily mortality for the examined period. This increase can be attributed to PM10 concentrations in most cases, and in fewer instances, to air temperature and O-3 concentrations. The heat wave indicator did not significantly increase the relative risk; the same held for the relative humidity. For the general unstratified population, the highest increase in relative risk of 1.072 (95% CI 1.001-1.147) was observed for cardiovascular mortality and was associated with an increase in temperature of 10 A degrees C, followed by an increase in relative risk of 1.056 (95% CI 1.025-1.087) for respiratory mortality associated with an increase in O-3 concentrations by 10 mu g.m(-3). A higher risk in most cases was found for women. A significant increase of relative risk of 1.013 (95 % CI 1.002-1.024) due to PM10 was found for hospital admissions for cardiovascular causes. This issue should be studied further in view of the anticipated increase in meteorological extremes, including heat waves, in the future, to prepare prevention plans for eliminating their negative effects as far as possible.