Ambient temperature and preterm birth: Interaction by maternal medical factors and medications

Associations between high temperatures and preterm birth (PTB) were reported. However, characteristics of vulnerable pregnancies are unknown. We aimed to examine pre-pregnancy comorbidities, PTB history, and prescribed medications during pregnancy as potential modifiers.

We conducted a secondary analysis of a historical cohort of 131 599 births in Southern Israel in 2005-2019. Maternal address and a spatiotemporal model were used to assess weekly average temperatures. We employed a Cox model with time-dependent covariates using a distributed lag non-linear model, adjusted for ethnicity, socioeconomic status, and time-varying seasonality. Associations were stratified by PTB history and restricted by comorbidities and medications. P-values for interaction were from likelihood ratio tests of models with multiplicative interaction term. Corrections for multiple comparisons were done using the Holm-Bonferroni method.

Associations differed by PTB history (p for interaction <0.001). In women with previous PTB, high temperatures were associated with PTB from week 9, peaking at the last follow-up week (HR = 1.09, 95% CI: 1.01-1.17, 31 °C vs 16 °C); in contrast to the null associations in primiparous women and multiparous women without PTB history. We also found stronger associations for heat in women with autoimmune diseases and for cold in those with an antihistamine prescription, but interactions were not significant. Other comorbidities and medications we examined yielded inconsistent associations with high uncertainty.

Our study highlights pregnancies with a PTB history as a high-risk group for focused interventions. Further studies with larger population sizes are needed to examine interactions by comorbidities and medications.