All 16 studies that met the review criteria were conducted between 1990 and 2014. Of these, 15 were retrospective cohort studies using birth registries and one was a case control study (Table 1) . There was representation of states across regions of the United States. Sample sizes ranged from 1,761 births to 1,548,904 births. Several different methods were used to estimate levels of air pollution exposure, including Bayesian measurement, Community Multi-scale Air Quality (CMAQ) model, data from United States Environmental Protection Agency (EPA), installation of ambient air monitors, proximity to major roads, and mixed methods. While most studies looked at a single pregnancy outcome, others considered multiple outcomes. Specific details of racial disparities in pregnancy outcomes across race are described below.
Air pollution and Preterm birth
Preterm birth was assessed in eight out of 16 of the studies we reviewed. Benmarhnia et al. (28) used decomposition analysis to understand the racial disparities in PTB in California. Two pollutants were included: fine-grained particulate matter (PM2.5) and nitrogen dioxide (NO2). A higher prevalence of PTB was observed for non-Hispanic Black (NHB) mothers when compared with non-Hispanic White mothers. The predicted difference in probability of PTB between Black and White infants was 0.056 (95% CI: 0.054, 0.058). All included predictors explained 37.8% of the Black–White disparity. Overall, individual variables (17.5% for PTB) such as age and level of education, and neighborhood-level variables (16.1% for PTB) such as socioeconomic environment explained a greater proportion of the Black–White difference in birth outcomes than air pollution (5.7% for PTB). Gray and colleagues reported that infants born to NHB mothers were at increased odds of PTB (adjusted OR 1.46, 95% CI: 1.42 to 1.50). (27)
Using the CMAQ model and measurements from stationary monitors, Hao et al. (29) investigated the association between 11 ambient air pollutants and the risk of PTB in the state of Georgia. They observed that all traffic-related pollutants (carbon monoxide (CO), NO2 PM2.5, elemental carbon) were associated with PTB [e.g., odds ratios for interquartile range increases in CO during the first, second, and third trimesters and total pregnancy were 1.005 (95% CI: 1.001, 1.009), 1.007 (95% CI: 1.002, 1.011), 1.010 (95% CI: 1.006, 1.014), and 1.011 (95% CI: 1.006, 1.017), respectively]. Associations were higher for mothers with low educational attainment and African American mothers. There were no clear explanations for the observed associations. Other studies had similar findings (29–32).
Woodruff et al. (31) using data from the EPA Ambient Air Monitoring observed that Hispanic, African-American, and Asian/Pacific Islander mothers experienced higher mean levels of air pollution and were more than twice as likely to live in the most polluted counties compared with White mothers. In addition, there was an increase in the odds of PTB (adjusted odds ratio (AOR) = 1.05; 95% CI, 0.99–1.12) in counties with higher air pollution.
Le et al. (32) reported a stronger association between air pollutants and PTB for Blacks than Whites. PTB was associated with SO2 (OR 1.07, 1.01 – 1.14) exposure in the last month of pregnancy, while O3 exposures exceeding 92 parts per billion (OR 1.08, 1.02 – 1.14) were associated with PTB in the first months of pregnancy.
Air pollution and Small for Gestational Age
Of all 16 studies reviewed, six assessed SGA. Benmarhnia et al. (28) explored racial disparities in SGA in California. The predicted difference in probability of SGA between Black and White infants was 0.084 (95% CI: 0.081, 0.087). Together, individual demographics, neighborhood socioeconomic environment (such as unemployment and poverty rates), and neighborhood air pollution explained 37.8% of the Black-White disparity. There was a higher prevalence of SGA among non-Hispanic Blacks (18.3%) as compared to non-Hispanic Whites (9.9%).
In North Carolina (27), infants born to NHB mothers and Hispanic mothers were at an increased odds of being small for gestational age with (AOR = 2.18, 95% CI: 2.12 to 2.24 and AOR = 1.21, CI: 1.17 to 1.26, respectively), compared to NHW mothers. After controlling for race and individual and area-level socio-economic status, this difference persisted, suggesting that air pollution is an additional contributor to the observed outcomes.
In a retrospective cohort study of 4,098,750 births in California, Woodruff et al. (31) observed that Hispanic, African-American, and Asian/Pacific Islander mothers experienced higher mean levels of air pollution, determined from strategically-placed EPA monitors. They were also more than twice as likely to live in the most polluted counties compared with White mothers, after controlling for maternal risk factors, region, and educational status [Hispanic mothers: AOR = 4.66; 95% CI: 1.92–11.32; African-American mothers: AOR = 2.58; 95% CI: 1.00–6.62; Asian/Pacific Islander mothers: AOR = 2.82; 95% CI, 1.07–7.39]. However, there was no significant increase in the odds of SGA (AOR = 0.96; 95% CI, 0.86–1.07) in counties with higher air pollution.
In a cohort of 164,905 births in Michigan, Le et al. (32) showed that there was an association between term SGA with exposure to CO and NO2 during the first trimester of pregnancy. They also reported an association between term SGA and exposure to O3 and PM10 during the later stages of pregnancy. There was evidence of stronger associations between CO and term-SGA, NO2 and term-SGA and SO2 and term-SGA for infants of Black mothers as compared to White mothers.
Rich et al. (33) found significantly increased risk of SGA associated with first and third trimester exposures to PM2.5 and increased risk of very small for gestational age (VSGA) associated with first, second and third trimester exposures to high NO2 concentrations. According to this study, mothers of SGA and VSGA infants were more likely to be less than 25 years old and less likely to have completed high school, compared to mothers of appropriate-size births. They were also more likely to be single, African American, and to have smoked during pregnancy.
Air Pollution and Low Birth Weight
Low birth weight was assessed in five of the 14 studies reviewed. Gray et al. (27) used multivariate analysis of factors, including PM2.5 and O3 to assess their association with weight differences in grams and 95% CI for all births. PM2.5 exposures were associated with LBW among infants born to NHB and Hispanic mothers more than those born to NHW mothers (-187.5g, 95% CI: -183.6 to -191.4 and -46.8g, 95% CI: -41.8 to -51.7, respectively).
Miranda et al. (34) characterized maternal exposure to traffic-related air pollution during pregnancy by using residential proximity to major road ways as a proxy. Women residing within 250 meters of a major roadway were at a 3-5% increased odds of having a LBW baby than women residing more than 250 meters away (p<0.05). The mean birth weight was 3376g for NHW, 2114g for NHB, and 3330g for Hispanics.
In a retrospective cohort study in California (35), African Americans, despite being a significantly lower percentage of the study population, had 3.19% of the LBW infants, compared to 1.23% for whites and 1.59% for Hispanics. Bell et al. (36) also reported that the association between air pollutants (especially PM2.5) and LBW for infants of Black mothers was stronger than for White mothers.
In a large cohort study of 1,548,904 births in Texas, Geer et al. (37) reported that interquartile increases in ambient air pollutant concentrations of SO2 and O3 were associated with a 4.99g (95% CI, 1.87–8.11) and 2.72g (95% CI, 1.11–4.33) decreases in birth weight, respectively. Lower birth weight was associated with exposure to O3 in the first and second trimester, whereas results for other pollutants did not differ significantly by trimester.
In a case-control study, Ghosh et al. (38), observed that women who were exposed to secondhand smoke at home had increased odds of term LBW (AOR = 1.36; 95% CI: 0.85, 2.18) compared to unexposed women. Blacks and Hispanic had higher odds of having LBW babies compared to White mothers.
Air pollution and stillbirths
We found only one cohort study that assessed stillbirths in relation to air pollution. According to DeFranco et al. (11) high average PM2.5 exposure through pregnancy was not associated with a significant increase in stillbirth risk (AOR 1.21; 95% CI: 0.96,1.53). There was also no higher risk of stillbirth associated with exposure in either the first or second trimester. However, exposure to high levels of PM2.5 in the third trimester of pregnancy was associated with 42% increased risk of stillbirth, (AOR = 1.42; 1.06, 1.91). Stillbirth rates were higher among mothers older than 40 years (11.6 per 1000), NHB mothers (8.6 per 1000), and mothers with lower education level and tobacco use.