Background: Cooking with solid fuels such as wood and charcoal contributes to household air pollution and is associated with adverse birth outcomes. Cleaner fuel interventions may reduce HAP exposure sufficiently to prevent preterm birth (PTB, <37 weeks gestation). We assessed the effects of a randomized liquefied petroleum gas (LPG) stove and free fuel intervention on PTB and gestational duration (secondary outcomes of the trial), as well as associations with fine particulate matter (PM2.5), black carbon (BC), and carbon monoxide (CO) in the HAPIN trial conducted in Guatemala, India, Peru, and Rwanda.

Methods: Pregnant women (18-35 years of age; gestation confirmed by ultrasound at 9-20 weeks) were randomly assigned to intervention arm (n=1,593) or control arm (n=1,607). We measured 24-hour personal exposure to PM2.5, BC, and CO at baseline, 18-24 weeks, and 32-36 weeks gestation. We used intent-to-treat approach to evaluate the intervention effect. We also assessed associations between measured pollution concentrations with PTB and gestational duration using linear and log-binomial regression models.

Results: Intervention adherence was nearly exclusive; PM2.5, BC, and CO exposures were 66-80% lower post-randomization compared to control. The proportion of livebirths (n=3060) that were preterm was 5.9% (intervention arm) and 5.4% (control arm). Compared to the control arm, the relative risk of PTB among women randomized to the intervention was 1.09 (95% CI 0.82-1.46, p=0.56). The mean gestational duration was 39.3 weeks in both the intervention and control arm (adjusted mean difference 0.03 weeks, 95% CI -0.08-0.15, p=0.55). PM2.5, BC, and CO levels during pregnancy were not associated with PTB or gestational duration.

Conclusions: An LPG stove and fuel intervention did not reduce PTB risk or increase gestational duration in this study population, nor was measured personal pollution associated with either outcome.