Interaction of maternal smoking and preterm birth on future risk of maternal cardiovascular disease: A population-based record linkage study

Abstract

Background While associations of smoking and preterm birth (PTB) with maternal cardiovascular disease (CVD) risks have been established, it is unknown whether the coexistence of these two conditions could synergistically increase the risks. Methods We linked birth records of 902,008 mothers with singleton infants during 1994-2011 in New South Wales, Australia to the mothers' subsequent CVD hospitalisation or death. Multiplicative interaction was tested through an interaction term in a multivariate Cox-proportional hazard regression model, while additive interaction was assessed by calculating the synergy index. Results Relative to never-smokers with term babies, the CVD risk in ever-smokers with PTBs (hazard ratio (HR) 3.35, 95% confidence interval (CI) 2.96-3.80) was significantly greater than the sum of risks in ever-smokers with term babies (HR 2.10, 95% CI 1.96-2.24) and in never-smokers with PTBs (HR 1.73, 95% CI 1.55-1.93), indicating an additive interaction (synergy index = 1.29, 95% CI 1.05-1.58). In ever-smokers, the association was stronger for extremely PTB (HR 3.83, 95% CI 3.23-4.69) than moderately PTB (HR 3.18, 95% CI 2.76-3.66), and for ≥2 PTB (HR 4.47, 95% CI 3.39-5.88) than one PTB (HR 3.20, 95% CI 2.81-3.64). Conclusion Maternal smoking and PTB interact on the additive scale to synergistically increase maternal CVD risks. The interaction was dose-dependent according to both the severity and number of PTBs.