[Effects of temperature extremes on daily mortality in Castile-La Mancha (Spain): trends from 1975 to 2003].

Objectives: To determine time trends and the geographical distribution of mortality trigger temperature thresholds due to extreme temperatures in Castile-La Mancha (central Spain) between 1975 and 2003.

Methods: The analysis was divided into three periods (1975-1984, 1985-1994 and 1995-2003) for each province of the region. Daily mortality due to organic causes (dependent variable) was modelled using autoregressive integrated moving average (ARIMA) procedures. The resulting residual series was related to the maximum temperature series grouped in 2 degrees C intervals to obtain a threshold temperature for cold or heat when the residuals rose significantly (p<0,05) above the mean residual mortality value of the corresponding study period.

Results: Mortality trigger temperature thresholds decreased over time in Castile- La Mancha. In Toledo, the trigger temperature diminished from 40 degrees C to 38 degrees C. In Cuenca and Guadalajara, threshold temperatures for heat events were obtained in the last few decades but not in the first. These thresholds varied from the 92nd percentile in Cuenca to the 98th percentile in Albacete in the last decade. No threshold temperatures for cold spells were observed in any province or period.

Conclusions: Castile-La Mancha registered an upward trend in the relationship between high temperatures and mortality, probably due to population aging. This trend could have been influenced by the increased frequency of extremely hot days. Prevention plans should be periodically reviewed.