Lifetime Exposure to Traffic-related Pollution and Lung Function in Early Adolescence

To the Editor:

Long-term ambient air pollution exposure has been associated with reduced child lung function (1). Although some studies have demonstrated that early-life exposure has a persistent effect on child lung function, others have shown that more recent exposures have a greater impact, and improvements in air quality may reverse or attenuate these effects (2–6). We previously found that living close to a major roadway and past-year and lifetime exposure to particulate matter ⩽2.5 μm in aerodynamic diameter (PM_2.5) were associated with lower lung function among elementary school–aged children in the Boston area (2). In this follow-up study, we evaluate pollution exposures at different time points and associations with adolescent lung function. As PM_2.5 concentrations have been reduced, we also examined if change in PM_2.5 concentration during follow-up was associated with lung function growth.

Project Viva is a prospective prebirth cohort study of mother–child pairs followed from pregnancy to adolescence. We recruited 2,128 women in early pregnancy between 1999 and 2003 from Atrius Harvard Vanguard Medical Associates, a group practice in eastern Massachusetts. Children completed spirometry at a midchildhood visit (median age, 7.7 years; n = 510) and an early adolescent visit completed between 2013 and 2016 (median age, 12.8 years; n = 844). Details of the exposures, spirometry measurements, and study design have been described elsewhere (2, 7, 8). Participants’ home addresses were collected at each visit, and any changes in address were estimated to have occurred halfway between the two visits. On the basis of published findings that traffic-related pollutants decay to background concentrations exponentially with distance from a freeway (9), we calculated the natural logarithm of the distance from each participant’s home to the nearest major roadway. Annual and lifetime PM_2.5 exposures were estimated using a satellite model, as previously described (2). We analyzed associations of residential proximity to a major roadway (at birth, midchildhood, and adolescence) and PM_2.5 exposure (for the first year of life, the year before spirometry, average from birth to midchildhood, average since midchildhood, and lifetime at adolescence) with adolescent lung function (forced expiratory volume in 1 second [FEV₁], forced vital capacity [FVC], and FEV₁/FVC ratio) and lung function growth between visits. Finally, we examined if change in PM_2.5 between visits was associated with change in FEV₁ and FVC between visits.

All associations were analyzed using linear regression. We adjusted for child age, race or ethnicity, sex, height, household income, household smoking status, date of visit, season (as the sine and cosine of visit date), temperature and humidity the day before spirometry, and census tract medium income and education. As in our previous analysis in this cohort (2), participants with missing covariate data (n = 41 [4.9%]) were excluded. We tested for effect modification by sex, race (White vs. non-White), household income (⩽$70,000 vs. >$70,000), and current asthma.

Among participants with exposure and outcome data, 50.7% were female, 64.9% identified as White and 16.1% as Black, and 15.3% had asthma. Seventy-seven percent of households had annual incomes greater than $70,000, and 6.5% included at least one smoker. Mean FEV₁ was 2.7 L (97.9% predicted; standard deviation [SD], 12.9%), and mean FVC was 3.2 L (100.9% predicted; SD, 11.9%). Mean annual increases in FEV₁ and FVC between visits were 246 ml/yr (SD, 76.8 ml/yr) and 287 ml/yr (SD, 84.3 ml/yr), respectively. Median distance to a major roadway in adolescence was 1,350 m (interquartile range [IQR], 2,459 m). Median annual PM_2.5 exposure was 11.2 μg/m³ (IQR, 1.5 μg/m³) for the first year of life, 9.9 μg/m³ (IQR, 1.3 μg/m³) for the year before the midchildhood visit, 7.8 μg/m³ (IQR, 0.8 μg/m³) for the year before the adolescent visit, 8.0 μg/m³ (IQR, 1.2 μg/m³) for the years between visits, and 9.3 μg/m³ (IQR, 1.1 μg/m³) for lifetime exposure up to the adolescent visit. The logarithms of distance to a major roadway and annual average PM_2.5 preceding the adolescent visit were not correlated (Pearson r = −0.18).

Living closer to a major roadway at the adolescent visit was associated with lower adolescent FEV₁ (−31.2 ml; 95% confidence interval [CI], −62.5 to −0.01 ml) and FVC (−38.2 ml; 95% CI, −71.3 to −5.1 ml) per IQR difference in distance (log scale) (Figure 1). Lifetime PM_2.5 exposure was associated with lower FEV₁ (−34.2 ml; 95% CI, −67.9 to −0.5 ml), and average PM_2.5 since midchildhood was associated with lower FEV₁ (−42.5 ml; 95% CI, −76.4 to −8.5 ml) and FVC (−47.6 ml; 95% CI, −83.4 to −11.8 ml) per 1 μg/m³ (Figure 2). There were no associations of proximity to a roadway at birth or midchildhood, PM_2.5 exposure during the first year of life, or PM_2.5 exposure in the prior year with adolescent lung function. There were no associations of these exposures with subsequent change in lung function between the two visits. Reduction in average annual PM_2.5 exposure between the midchildhood and early adolescent visits was associated with faster lung function growth: the mean difference in annual PM_2.5 between the visits of 0.4 μg/m³ was associated with a 17.5 ml/yr (95% CI, 6.5 to 28.4 ml/yr) faster growth in FEV₁ and a 14.5 ml/yr (95% CI, 2.7 to 26.3 ml/yr) faster growth in FVC.

Figure 1.

Associations of proximity to a major roadway and lung function in early adolescence. Results are scaled from the 75th to the 25th percentile of the log-transformed distance to a major roadway. All models are adjusted for child age, sex, race/ethnicity, and height; household income; household smoking; season (as sine and cosine terms); temperature and humidity the day before the spirometry examination; and census tract median household income and education. FEV₁ = forced expiratory volume in 1 second; FVC = forced vital capacity.

Figure 2.

PM_2.5 exposure and lung function in adolescence. Results are scaled per 1 μg/m³. All models are adjusted for child age, sex, race/ethnicity, height; household income; household smoking; season (as sine and cosine terms); temperature and humidity the day before the spirometry examination; and census tract median household income and education. FEV₁ = forced expiratory volume in 1 second; FVC = forced vital capacity; PM_2.5 = particulate matter ⩽2.5 μm in aerodynamic diameter.

Associations between pollution and adolescent lung function differed by sex. Living closer to a roadway and PM_2.5 exposure in the prior year, since midchildhood, and lifetime were negatively associated with lung function for boys but were null among girls (P_interaction < 0.05). Among non-White participants, living closer to a road at birth was associated with lower FEV₁ in adolescence (P_interaction < 0.05). Associations did not differ by household income or asthma.

In this follow-up study, we found that lifetime PM_2.5 exposure and traffic-related pollution in adolescence were associated with lower adolescent lung function. Recent reduction in PM_2.5 exposure was also associated with more rapid lung function growth during adolescence.

Consistent with our prior findings regarding midchildhood lung function, exposures occurring closer to spirometry completion were more consistently associated with lower lung function compared with exposures in early life around the time of birth. Exposures in midchildhood had similar effect estimates as adolescent exposures, indicating a possible issue with power rather than a lack of association. In our study and the Children’s Health Study in southern California, improvements in air quality were associated with greater lung function growth, suggesting at least a partial reversibility of the harmful effects of pollution and the importance of more recent exposures on adolescent lung function (3, 10). A study in the Netherlands and the Children’s Health Study both revealed decrements in FEV₁ in adolescence and slower FEV₁ growth in association with greater long-term PM_2.5 exposure (5, 11, 12). Notably, these studies were conducted in areas with higher PM_2.5 concentrations than our study. Our findings suggest that respiratory health benefits of improved air quality can be realized with small changes in PM_2.5 at even lower pollution concentrations within current U.S. Environmental Protection Agency standards.

Several of the exposures in our study were associated with reduced lung function in boys, with no effect in girls. Similar effect modification for boys has been reported in other studies (3, 6), though results have been mixed and there is yet to be conclusive evidence of differing susceptibility to the detrimental effects of pollution on the basis of sex (1). There was no consistent difference in associations on the basis of race in our study.

There are several limitations to our study. Proximity to a roadway may not reflect actual pollution exposures and may incorporate nonpollution exposures such as noise and social stress. In addition, our analyses evaluating changes in lung function used only two time points for each child, and we were unable to assess trajectories in early childhood.

In conclusion, we found that more recent air pollution exposures may have a greater effect on adolescent lung function than early-life exposure, and improving air quality, even at already low air pollution concentrations, appears to be beneficial to adolescent respiratory health.