Effect of Previous-Winter Mortality on the Association between Summer Temperature and Mortality in South Korea

BACKGROUND: It has recently been postulated that low mortality levels in the previous winter may increase the proportion of vulnerable individuals in the pool of people at risk of heat-related death during the summer months. OBJECTIVES: We explored the sensitivity of heat-related mortality in summer (June-August) to mortality in the previous winter (December-February) in Seoul, Daegu, and Incheon in South Korea, from 1992 through 2007, excluding the summer of 1994. METHODS: Poisson regression models adapted for time-series data were used to estimate associations between a 1 degrees C increase in average summer temperature (on the same day and the previous day) above thresholds specific for city, age, and cause of death, and daily mortality counts. Effects were estimated separately for summers preceded by winters with low and high mortality, with adjustment for secular trends. RESULTS: Temperatures above city-specific thresholds were associated with increased mortality in all three cities. Associations were stronger in summers preceded by winters with low versus high mortality levels for all nonaccidental deaths and, to a lesser extent, among persons >= 65 years of age. Effect modification by previous-winter mortality was not evident when we restricted deaths to cardiovascular disease outcomes in Seoul. CONCLUSIONS: Our results suggest that low winter all-cause mortality leads to higher mortality during the next summer. Evidence of a relation between increased summer heat-related mortality and previous wintertime deaths has the potential to inform public health efforts to mitigate effects of hot weather.