Air Pollution and Term Birth Weight in New York, New York
Air Pollution and Term Birth Weight in New York, New York
Our finding of a relationship between both PM2.5 and nitrogen dioxide in relation to birth weight is consistent with those of some other studies but somewhat greater in magnitude than has typically been reported. For comparison, the estimates based on absolute change (per 10 µg/m for PM2.5 or per 10 ppb for nitrogen dioxide) are most readily compared, because the exposure ranges across studies, and thus the effects of interquartile shifts, are not comparable. Our estimated birth weight effects of PM2.5 of approximately 20 g for first-trimester exposure, 30 g for third-trimester exposure, and 40 g for total pregnancy exposure per 10 µg/m are broadly in the range observed in some studies and much greater than was found in other studies.
Our findings for nitrogen dioxide suggest a birth weight reduction of approximately 18 g per 10 ppb for exposure in the first or third trimester and for the total pregnancy. These estimates are broadly comparable to some studies, much greater than those found in others, and markedly weaker than a large but imprecise estimate reported in Valencia, Spain.
Some of the differences in estimated effect sizes across studies could be due to the exposure assignment methods used and their spatial resolutions. For example, Kloog et al. estimated PM2.5 exposure on the basis of aerosol optical depth at a 10-km spatial resolution. Our study incorporated information about local emission sources to estimate exposure within 300 m of the maternal address. Among the candidate explanations besides exposure assignment accuracy are particle composition and toxicity, contributions from spatial and temporal variations in pollution, exact definition of the birth weight measure, exclusions, covariates used in adjustment, varying susceptibility to socioeconomic confounding, and analytical methods. The most distinctive features of our study, which potentially resulted in somewhat stronger effects, are the enhanced exposure assessment (i.e., high spatial resolution) from NYCCAS and restriction of the study population to those individuals and outcome measures most purely indicative of growth, excluding births in which pathology caused a reduction in size (e.g., preterm births or congenital defects). On the other hand, the more modest and irregular association of air pollution and socioeconomic status and the exclusion of smokers may well reduce apparent effect size to the extent that incomplete adjustment has affected other studies. A systematic evidence review, which is beyond the scope of this paper, would be needed to draw inferences about which of the many variables related to study setting and design may be responsible for variable findings across studies.
A major concern for studies of pregnancy is the distinctive contributions of spatial and temporal variations in exposure. A recent multicountry meta-analysis found that studies using temporal assignment found stronger inverse associations of particulate matter and birth weight. However, given that the quality of exposure assignment based on these 2 determinants likely differs, and the susceptibility to confounding clearly differs, the observed patterns are not easily interpreted. Comparable effect sizes for the 2 sources for PM2.5 might suggest a causal effect (assuming that mass concentration is the biologically relevant measure of exposure), in that it is highly unlikely that both indices would be confounded by other factors. The isolation of nitrogen dioxide associations to spatial, not temporal, variation may support either a true effect that reflects the more accurate indicators of spatial variation or confounding of 1 or both of the measures. This spatial/temporal distinction is embedded in our attempts to examine mutually adjusted results and makes the predominance of nitrogen dioxide over PM2.5 of uncertain significance.
In this study, we examined only the association with PM2.5 mass concentration, whereas the toxicity of PM2.5 and its effect on a range of health outcomes including birth weight may be modified by its chemical composition, which is related to sources. For example, Bell et al. found that the concentrations of zinc, elemental carbon, silicon, aluminum, vanadium, and nickel were associated with lower birth weight in selected Massachusetts and Connecticut communities, and numerous studies suggest effect modification of cardiovascular and other outcomes by PM2.5 composition. Nitrogen dioxide, which is less influenced by regional sources and has greater spatial variation within New York than does PM2.5, may be a surrogate for local combustion sources, such as traffic or residual oil combustion, that contribute to spatial variation in PM2.5 mass and composition within the city. In future studies, we plan to apply PM2.5 chemical speciation data from NYCCAS to develop exposure metrics and assess the association of spatial source and PM2.5 composition differences on birth outcomes. Other pollutants of concern (e.g., carbon monoxide, which was not measured in NYCCAS) were not addressed, and we did not incorporate information on particle constituents in this analysis, but plan to do so later.
As in other studies of this nature, our exposure estimates were limited to the location of the maternal residence, not considering variability due to work and other activities and the nature of housing as it affects indoor/outdoor gradients (which may be especially problematic in New York given the variable height of apartment buildings). Potential confounders that we were not able to address include noise and environmental tobacco smoke. In addition, we know from unpublished analyses of New York City Pregnancy Risk Assessment Monitoring System data that smoking during pregnancy is underreported on birth certificates in New York, as has been previously reported for multiple states. Additional limitations include the inability to account for repeat births to the same mother, slightly understating the variance of effect estimates, and the lack of residential history information other than the birth address. There are many potential influences on fetal growth, with only some pathways vulnerable to the adverse effects of air pollution, and we were not able to refine the outcome to isolate those most plausibly affected by eliminating those with known complications affecting fetal growth.
In this large and rapidly expanding research avenue, our results add support to the possible impact of common air pollutants, specifically PM2.5 and nitrogen dioxide, on fetal growth. Although the magnitude of estimated change in birth weight is clinically inconsequential for a given infant, there may be health consequences to a shift in the population birth weight distribution with regard to both near-term health outcomes (e.g., hospital stay, survival) and long-term consequences (e.g., neurodevelopment, cardiovascular risk markers). Additional analyses are needed to determine whether the predicted impacts of a small shift in birth weight on morbidity are identifiable, which are feasible for outcomes such as neonatal intensive care unit admissions and respiratory distress. Careful examination of specific indicators of fetal and infant health is needed, along with refined pollution assessment that considers temporal and spatial contributors, chemical speciation of particulates, and evaluation of pollutant sources. Although the signal relating air pollution to reproductive health is difficult to discern, there is ample encouragement to take the next steps to refine our understanding of its presence and meaning.
Discussion
Our finding of a relationship between both PM2.5 and nitrogen dioxide in relation to birth weight is consistent with those of some other studies but somewhat greater in magnitude than has typically been reported. For comparison, the estimates based on absolute change (per 10 µg/m for PM2.5 or per 10 ppb for nitrogen dioxide) are most readily compared, because the exposure ranges across studies, and thus the effects of interquartile shifts, are not comparable. Our estimated birth weight effects of PM2.5 of approximately 20 g for first-trimester exposure, 30 g for third-trimester exposure, and 40 g for total pregnancy exposure per 10 µg/m are broadly in the range observed in some studies and much greater than was found in other studies.
Our findings for nitrogen dioxide suggest a birth weight reduction of approximately 18 g per 10 ppb for exposure in the first or third trimester and for the total pregnancy. These estimates are broadly comparable to some studies, much greater than those found in others, and markedly weaker than a large but imprecise estimate reported in Valencia, Spain.
Some of the differences in estimated effect sizes across studies could be due to the exposure assignment methods used and their spatial resolutions. For example, Kloog et al. estimated PM2.5 exposure on the basis of aerosol optical depth at a 10-km spatial resolution. Our study incorporated information about local emission sources to estimate exposure within 300 m of the maternal address. Among the candidate explanations besides exposure assignment accuracy are particle composition and toxicity, contributions from spatial and temporal variations in pollution, exact definition of the birth weight measure, exclusions, covariates used in adjustment, varying susceptibility to socioeconomic confounding, and analytical methods. The most distinctive features of our study, which potentially resulted in somewhat stronger effects, are the enhanced exposure assessment (i.e., high spatial resolution) from NYCCAS and restriction of the study population to those individuals and outcome measures most purely indicative of growth, excluding births in which pathology caused a reduction in size (e.g., preterm births or congenital defects). On the other hand, the more modest and irregular association of air pollution and socioeconomic status and the exclusion of smokers may well reduce apparent effect size to the extent that incomplete adjustment has affected other studies. A systematic evidence review, which is beyond the scope of this paper, would be needed to draw inferences about which of the many variables related to study setting and design may be responsible for variable findings across studies.
A major concern for studies of pregnancy is the distinctive contributions of spatial and temporal variations in exposure. A recent multicountry meta-analysis found that studies using temporal assignment found stronger inverse associations of particulate matter and birth weight. However, given that the quality of exposure assignment based on these 2 determinants likely differs, and the susceptibility to confounding clearly differs, the observed patterns are not easily interpreted. Comparable effect sizes for the 2 sources for PM2.5 might suggest a causal effect (assuming that mass concentration is the biologically relevant measure of exposure), in that it is highly unlikely that both indices would be confounded by other factors. The isolation of nitrogen dioxide associations to spatial, not temporal, variation may support either a true effect that reflects the more accurate indicators of spatial variation or confounding of 1 or both of the measures. This spatial/temporal distinction is embedded in our attempts to examine mutually adjusted results and makes the predominance of nitrogen dioxide over PM2.5 of uncertain significance.
In this study, we examined only the association with PM2.5 mass concentration, whereas the toxicity of PM2.5 and its effect on a range of health outcomes including birth weight may be modified by its chemical composition, which is related to sources. For example, Bell et al. found that the concentrations of zinc, elemental carbon, silicon, aluminum, vanadium, and nickel were associated with lower birth weight in selected Massachusetts and Connecticut communities, and numerous studies suggest effect modification of cardiovascular and other outcomes by PM2.5 composition. Nitrogen dioxide, which is less influenced by regional sources and has greater spatial variation within New York than does PM2.5, may be a surrogate for local combustion sources, such as traffic or residual oil combustion, that contribute to spatial variation in PM2.5 mass and composition within the city. In future studies, we plan to apply PM2.5 chemical speciation data from NYCCAS to develop exposure metrics and assess the association of spatial source and PM2.5 composition differences on birth outcomes. Other pollutants of concern (e.g., carbon monoxide, which was not measured in NYCCAS) were not addressed, and we did not incorporate information on particle constituents in this analysis, but plan to do so later.
As in other studies of this nature, our exposure estimates were limited to the location of the maternal residence, not considering variability due to work and other activities and the nature of housing as it affects indoor/outdoor gradients (which may be especially problematic in New York given the variable height of apartment buildings). Potential confounders that we were not able to address include noise and environmental tobacco smoke. In addition, we know from unpublished analyses of New York City Pregnancy Risk Assessment Monitoring System data that smoking during pregnancy is underreported on birth certificates in New York, as has been previously reported for multiple states. Additional limitations include the inability to account for repeat births to the same mother, slightly understating the variance of effect estimates, and the lack of residential history information other than the birth address. There are many potential influences on fetal growth, with only some pathways vulnerable to the adverse effects of air pollution, and we were not able to refine the outcome to isolate those most plausibly affected by eliminating those with known complications affecting fetal growth.
In this large and rapidly expanding research avenue, our results add support to the possible impact of common air pollutants, specifically PM2.5 and nitrogen dioxide, on fetal growth. Although the magnitude of estimated change in birth weight is clinically inconsequential for a given infant, there may be health consequences to a shift in the population birth weight distribution with regard to both near-term health outcomes (e.g., hospital stay, survival) and long-term consequences (e.g., neurodevelopment, cardiovascular risk markers). Additional analyses are needed to determine whether the predicted impacts of a small shift in birth weight on morbidity are identifiable, which are feasible for outcomes such as neonatal intensive care unit admissions and respiratory distress. Careful examination of specific indicators of fetal and infant health is needed, along with refined pollution assessment that considers temporal and spatial contributors, chemical speciation of particulates, and evaluation of pollutant sources. Although the signal relating air pollution to reproductive health is difficult to discern, there is ample encouragement to take the next steps to refine our understanding of its presence and meaning.
Source...