The influence of seasonal climate variability on mortality in pre-industrial Sweden

BACKGROUND: Recent studies have shown an association between weather and climatic factors with mortality, cardiovascular and infectious diseases. We used historical data to investigate the impact of seasonal temperature and precipitation on total mortality in Uppsala, Sweden, during the first two stages of the demographic transition, 1749–1859. DESIGN: We retrieved mortality and population numbers of the Uppsala Domkyrka parish from digitised parish records and obtained monthly temperature and precipitation measures recorded in Uppsala during that time. Statistical models were established for year-to-year variability in deaths by annual and seasonal temperature and precipitation, adjusting for longer time trends. In a second step, a model was established for three different periods to study changes in the association of climate variability and mortality over time. Relative risks (RR) with 95% confidence intervals (CI) were calculated. RESULTS: Precipitation during spring and autumn was significantly associated with annual mortality (spring RR 0.982, CI 0.965–1.000; autumn RR 1.018, CI 1.004–1.032, respectively, per centimetre increase of precipitation). Higher springtime temperature decreased annual mortality, while higher summer temperature increased the death toll; however, both were only borderline significant (p=0.07). The significant effect of springtime precipitation for mortality was present only in the first two periods (1749–1785 and 1786–1824). On the contrary, the overall effect of autumn precipitation was mainly due to its relevance during the last period, 1825–1859 (RR 1.024, CI 0.997–1.052). At that time, higher winter precipitation was found to decrease mortality. CONCLUSIONS: In urban Uppsala, during the 18th and 19th century, precipitation appeared to be a stronger predictor for mortality than temperature. Higher spring precipitation decreased and higher autumn precipitation increased the number of deaths. However, this association differed before and during the early stages of industrialisation. Further research shall take age-specific differences into account, as well as changes in socio-economic conditions during that time.