Ambient Temperature and Risk of Preeclampsia: Biased Association?

Abstract Background Preeclampsia is associated with conception during warm months and delivery during cold months. We sought to determine whether season of conception and shorter gestation bias the associations. Methods We used hospital discharge summaries to identify 65 273 pregnancies with and 1 825 438 without preeclampsia in Quebec, Canada between 1989 and 2012. We obtained data on mean temperature for the month following conception and the month before hospital admission. We used cubic splines in log‐binomial models to estimate the association between temperature and preeclampsia (risk ratios, RR; 95% confidence interval, CI). To assess the potential for bias, we compared models progressively adjusted for baseline maternal characteristics, season of conception, and length of gestation at admission. Results With adjustment for baseline maternal characteristics only, risk of preeclampsia was higher for hot temperatures at conception and cold temperatures at end of pregnancy. Adjusting for season at conception removed the association between preeclampsia and temperature at conception. Adjustment for length of gestation removed the association between preeclampsia and temperature at end of pregnancy. Conclusions This study demonstrates that associations between ambient temperature and preeclampsia may be biased by short gestation, because preeclampsia commonly occurs earlier in pregnancy. Temperatures during gestation change with time for all women, and temperatures early in pregnancy frequently differ from temperatures later in pregnancy. Variation in temperature over gestation may lead to a coincidental association with preeclampsia.