Objectives. To estimate the magnitude of heat-related mortality in Finland in the 2000s.

Study Design. Daily numbers of deaths during the period 2000–2005 in Finland were classified according to the mean daily temperature.

Methods. The temperature at which mortality was lowest was first determined from smoothed data based on the loess regression. Heat-and cold-related mortalities were estimated by subtracting deaths at this optimal temperature from actual deaths on both sides of this temperature.

Results. In 2000–2005, the fewest deaths (126 per day) occurred at a mean daily temperature of 12 degrees C, and they increased to 138/day (by 10%) on the warmest days (+24 degrees C) and to 151/day (20%) on the coldest days (-31 degrees C). An estimated 160 deaths per year (0.3% of all deaths) were due to higher than optimal temperatures and 2,400 /year (5%) to low temperatures. In individual years, the fraction of deaths attributable to heat varied from 0-0.5%, with little consistency with mean summer temperatures. While the relative risk of an individual dying from heat increased consistently with rising temperatures, most heat-related deaths occurred at temperatures less than +20 degrees C. During the warm spell in summer 2000, deaths increased by an estimated 360 cases (0.7% of annual deaths), but this decreased to 250 (0.5%) once the Midsummer Festival was excluded.

Conclusions. As heat spells may occur in the future in an unpredictable way and because heat is not recognized as a health hazard in the North, more research should be devoted to clarifying the causal mechanisms underlying heat mortality, and pre-emptive measures should be planned.