The relationship between heat waves and mortality has been widely described, but there are few studies using long daily data on specific-cause mortality. This study is undertaken in central Spain and analysing natural causes, circulatory and respiratory causes of mortality from 1975 to 2008. Time-series analysis was performed using ARIMA models, including data on specific-cause mortality and maximum and mean daily temperature and mean daily air pressure. The length of heat waves and their chronological number were analysed. Data were stratified in three decadal stages: 1975-1985, 1986-1996 and 1997-2008. Heat-related mortality was triggered by a threshold temperature of 37 °C. For each degree that the daily maximum temperature exceeded 37 °C, the percentage increase in mortality due to circulatory causes was 19.3 % (17.3-21.3) in 1975-1985, 30.3 % (28.3-32.3) in 1986-1996 and 7.3 % (6.2-8.4) in 1997-2008. The increase in respiratory cause ranged from 12.4 % (7.8-17.0) in the first period, to 16.3 % (14.1-18.4) in the second and 13.7 % (11.5-15.9) in the last. Each day of heat-wave duration explained 5.3 % (2.6-8.0) increase in respiratory mortality in the first period and 2.3 % (1.6-3.0) in the last. Decadal scale differences exist for specific-causes mortality induced by extreme heat. The impact on heat-related mortality by natural and circulatory causes increases between the first and the second period and falls significantly in the last. For respiratory causes, the increase is no reduced in the last period. These results are of particular importance for the estimation of future impacts of climate change on health.