Gestational age modifies the association between exposure to fine particles and fetal death: findings from a nationwide epidemiological study in the contiguous United States

Abstract

Backgrounds: The vulnerability of fetuses differs at different developmental stages, in response to environmental stressors such as fine particulate matter (PM_2.5), a ubiquitous air pollutant. Whether gestational age (GA) modifies the association between prenatal fine particulate matter (PM_2.5) exposure and fetal death remains unclear.

Methods: We selected approximately 47.8 million eligible United States (US) livebirth and fetal death (defined as a termination at a GA of 20-43 weeks) records from 1989 to 2004. For each record, we took the level of prenatal exposure to PM_2.5 as the average concentration in the mother's residential county during the entire gestational period, or a specific trimester (i.e., GA-specific exposure), according to well-established estimates of monthly levels across the contiguous US. First, we evaluated the associations between PM_2.5 exposure and fetal death at a specific GA (i.e., GA-specific outcome) using five different logit models (unadjusted, covariate-adjusted, propensity-score, double robust, and diagnostic-score models). Double robust model was selected as the main model due to its advantages in causal inference. Then, we conducted meta-analyses to pool the estimated GA-specific associations, and explored how the pooled estimates varied with GA.

Results: According to the meta-analysis, all models suggested gestational PM_2.5 exposure was associated with fetal death. However, there was slight heterogeneity in the estimated effects, as different models revealed a range of 3.6-10.7% increase in the odds of fetal death per 5-µg/m³ increment of PM_2.5. Each 5-µg/m³ increase in PM_2.5 exposure during the entire gestation period significantly increased the odds of fetal death, by 8.1% (95% confidence interval [CI]: 5.1-11.2%). In terms of GA-specific outcomes, the odds of fetal death at a GA of 20-27, 28-36, or ≥ 37 weeks increased by 11.0% (5.9-16.4%), 5.2% (0.4-10.1%), and 8.3% (2.5-14.5%), respectively. In terms of GA-specific exposure, the odds of fetal death increased by 6.0% (3.9-8.2%), 4.1% (3.9-8.2%), and 4.3% (0.5-8.2%) with 5-µg/m³ increases in PM_2.5 exposure during the first, second, and third trimester, respectively. The association had the largest effect size (odds ratio = 1.098, 95% CI: 1.061-1.137) between PM_2.5 exposure during early gestation (i.e., first trimester) and early fetal death (i.e., 20-27 weeks).

Conclusions: Prenatal exposure to PM_2.5 was significantly associated with an increased risk of fetal death. The association was varied by gestational-age-specific exposures or outcomes, suggesting gestation age as a potential modifier on the effect of PM_2.5. The fetus was most vulnerable during the early stage of development to death associated with PM_2.5 exposure.

Keywords: Effect modification; Fetal death; Gestational age; PM2.5; Prenatal exposure.