Published online July 3, 2006
PEDIATRICS Vol. 118 No. 1 July 2006, pp. 156-164 (doi:10.1542/peds.2005-2432)
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow P3Rs: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when P3Rs are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow E-mail this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My File Cabinet
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via CrossRef
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Rogers, J. F.
Right arrow Articles by Dunlop, A. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Rogers, J. F.
Right arrow Articles by Dunlop, A. L.
Related Collections
Right arrow Office Practice

Air Pollution and Very Low Birth Weight Infants: A Target Population?

J. Felix Rogers, PhD, MPHa and Anne L. Dunlop, MD, MPHb

a National Immunization Program, Centers for Disease Control and Prevention, Atlanta, Georgia
b Department of Family and Preventive Medicine, Emory University School of Medicine, Atlanta, Georgia


   ABSTRACT
TOP
 ABSTRACT
METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
OBJECTIVE. The goal was to examine systematically the association between maternal exposure to particulate matter of <10 µm and very low birth weight (<1500 g) delivery for evidence of an effect on duration of gestation and/or intrauterine growth restriction.

METHODS. This case-control study took place between April 1, 1986, and March 30, 1988, in Georgia Health Care District 9 and included 128 mothers of very low birth weight infants, all of whom were preterm and were classified as either small for gestational age or appropriate for gestational age, and 197 mothers of term, appropriate-for-gestational-age infants weighing ≥2500 g. Maternal exposure to particulate matter of <10 µm was estimated with 2 exposure measures, namely, a county-level measure based on residence in a county with an industrial point source and an environmental transport model based on the geographic location of the birth home.

RESULTS. Considering preterm/appropriate-for-gestational-age infants as cases and term/appropriate-for-gestational-age infants as controls, adjusted odds ratios for maternal exposure to particulate matter of <10 µm were statistically significant (adjusted odds ratio for county-level model: 4.31; adjusted odds ratio for environmental transport model: 3.68). Although elevated, no statistically significant association was found between maternal exposure and preterm/appropriate-for-gestational-age delivery when compared to preterm/small-for-gestational-age delivery.

CONCLUSIONS. There are increased odds of maternal exposure to ambient particulate matter of <10 µm for very low birth weight preterm/appropriate-for-gestational-age delivery, compared with term/appropriate-for-gestational-age delivery, which suggests that the observed association between maternal exposure to air pollution and low infant birth weight (particularly <1500 g) is at least partially attributable to an effect on duration of gestation.

Key Words: air pollution • birth weight • preterm • environmental factors • fetal growth restriction

Abbreviations: AGA—appropriate for gestational age • IUGR—intrauterine growth restriction • LBW—low birth weight • PM10—particulate matter of <10 µm • SGA—small for gestational age • VLBW—very low birth weight • CI—confidence interval • GHCD9—Georgia Health Care District 9 • OR—odds ratio • aOR—adjusted odds ratio

The role of air pollution in postnatal infant death is well documented.1–3 The potential role of air pollution in adverse birth outcomes is less well understood, as described in detail in a recent literature review.4 Several studies found an association between maternal prenatal exposure to various air pollutants and low birth weight (LBW) (<2500 g).5–9 Others found no association10,11 or have failed to find an association with specific pollutants (eg, particulate matter of <10 µm [PM10]) implicated in other studies. Associations between air pollution and LBW are difficult to interpret, because birth weight is influenced by both duration of gestation and rate of intrauterine growth, and prematurity and intrauterine growth restriction (IUGR) are known to have different determinants.12,13 [IUGR indicates a situation in which a fetus has not reached its genetically predetermined size, and it can result in infants who are considered small for gestational age (SGA) or below the 10th percentile of birth weight for gestational age, according to race and gender.] The results of studies exploring the association between maternal exposure to ambient air pollution and preterm delivery and/or IUGR have yielded mixed results. Several studies reported an association between maternal exposure to air pollution and preterm delivery,9,14–17 some reported an association with preterm delivery but not IUGR or SGA status,8,18 and some reported an association with IUGR or SGA status.9,19–21

An association between maternal exposure to ambient air pollution and very low birth weight (VLBW) (<1500 g) delivery has been demonstrated.6 VLBW delivery, which confers a high risk of serious morbidity and death to the infant,22 almost always is associated with delivery before 32 completed weeks of gestation and often is accompanied by IUGR.23 To explore more thoroughly whether maternal exposure to ambient air pollution affects the length of gestation and/or intrauterine growth, particularly for infants born VLBW, we performed a case-control study in which we estimated maternal exposure to PM10 by using 2 exposure measures for women who delivered VLBW infants, all of whom were preterm (<37 completed weeks of gestation), classified as either SGA (<10th percentile of birth weight for gestational age, according to race and gender) or appropriate for gestational age (AGA) (≥10th percentile of birth weight for gestational age, according to race and gender).


   METHODS
TOP
 ABSTRACT
 METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Study Population
We obtained data on singleton VLBW births from the Georgia Very Low Birth Weight Study, which was conducted in 1988 by the Division of Birth Defects and Developmental Disabilities, the National Center for Environmental Health, the Centers for Disease Control and Prevention, and the Office of Epidemiology, Georgia Department of Human Resources. For this study, participants were selected from the 24 counties in Georgia Health Care District 9 (GHCD9). Except for the city of Savannah, GHCD9 is mostly rural. The case infants represented all live-born singleton infants born between April 1, 1986, and March 30, 1988, in GHCD9, after labor of spontaneous onset, who weighed <1500 g at birth. The control infants represented a selection of a 3% random sample (n = 197) of term, live-born, AGA infants who weighed ≥2500 g at birth and met the same residency and time-frame requirements as the case infants. Details of the selection procedure for control infants are specified elsewhere.6

We defined 3 birth outcomes. (1) Preterm/SGA VLBW infants were VLBW infants delivered between 20 and 37 completed weeks of gestation who fell below the 10th percentile for birth weight, given their gestational age, race, and gender (preterm/SGA; n = 69). (2) Preterm/AGA VLBW infants were VLBW infants delivered between 20 and 37 completed weeks of gestation who were at an appropriate weight for gestational age, given their race and gender (preterm/AGA; n = 59). (3) Term/AGA infants were infants delivered after 37 completed weeks of gestation who weighed ≥2500 g and who were at an appropriate weight for gestational age, given their race and gender (comparison group; n = 197). Gestational age was determined according to accepted clinical standards of preference, by using the date of the mother's last menstrual period and fetal ultrasound findings.24 Assignment of gestational age was based on the mother's last menstrual period for 38.3% of VLBW infants (49 of 128 infants) and 30.4% of term/AGA infants (60 of 197 infants), and assignment of gestational age was based on first- or second-trimester fetal ultrasound findings for 61.7% of VLBW infants (79 of 128 infants; 12.7% first trimester and 87.3% second trimester) and 69.5% of term/AGA infants (137 of 197 infants; 14.6% first trimester and 85.4% second trimester). SGA versus AGA status was determined with standard tables for birth weights according to the gestational age, race, and gender of the infant.25 Medical data on newborns and mothers were collected from labor and delivery logbooks and medical records from the birthing hospitals serving GHCD9; data related to mothers' sociodemographic status, reproductive history, health behaviors and substance abuse, stress and social support, and household exposures were collected through interviews, as described in detail previously.6

Ambient Air Pollution Exposure Measures
We considered 2 exposure measures for PM10 in this study. First, we used a county-level, ecologic exposure measure based on the county of the birth home and whether the county contained an industrial point source of PM10. Exposures for the county-level spatial model were classified as residence within a county with or without a point source of PM10, as described by Rogers et al.26 For this study, a point source was defined as an industrial facility in GHCD9 that is monitored by the Industrial Source Monitoring Program of the Georgia Department of Natural Resources for an estimate of specific air pollutants (including sulfur dioxide, nitrogen dioxide, PM10, and particulate matter of <2.5 µm) that are released into the atmosphere as a result of production processes.27 These monitored facilities account for ~95% of the total environmental releases in GHCD9.27

Second, we estimated ambient PM10 exposures at the level of the birth home with an environmental transport model that considered total annual PM10 emissions from 32 geographically identified, industrial point sources, annual meteorologic factors, and geographic location of the birth home. The environmental transport model is based on the Gaussian plume atmospheric transport model, which assumes that the plume concentration at each downwind distance has independent Gaussian distributions in both the horizontal and vertical planes.28–30 The model used in this study has been validated (and described more fully) elsewhere.26 The Gaussian plume model is used widely to predict concentrations in the atmosphere, and many of the atmospheric dispersion models recommended by the US Environmental Protection Agency are based on the Gaussian plume.31 PM10 exposure estimates from the atmospheric transport exposure model were categorized as quartile distributions; PM10 exposure estimates in the fourth quartile (>15.07 µg/m3) were considered "high exposure," those in the third quartile (3.75–15.07 µg/m3) were considered "moderate exposure," those in the second quartile (1.48–3.74 µg/m3) were considered "low exposure," and those in the first quartile (<1.48 µg/m3) were considered "unexposed" and served as the reference group for logistic analyses.

Statistical Analyses
First, to determine whether maternal exposure to PM10 was associated with VLBW delivery, we began by performing a case-control comparison with all VLBW infants as case subjects and term/AGA infants as control subjects. To determine systematically whether any observed association might be attributable to an effect on duration of gestation and/or IUGR, we performed 2 additional case-control comparisons. To address whether exposure to PM10 was associated with preterm delivery for infants born VLBW, we considered the subpopulation of VLBW infants classified preterm/AGA as case subjects and term/AGA infants as control subjects. To address whether exposure to PM10 was associated with IUGR, we compared the subpopulation of VLBW infants classified preterm/AGA with those classified preterm/SGA.

For univariate analyses, we compared the proportion of the infants within each birth outcome category with the various measured characteristics and exposures by using the Mantel-Haenzel {chi}2 statistic (with associated P value). We calculated crude odds ratios (ORs) to estimate the risk of the various birth outcomes within each exposure category.

We performed multivariate modeling by first using a model containing all exposure variables (as main effects) and all interaction terms, with subsequent elimination of all interaction terms because none was found to be significant with a likelihood ratio test. The SUDAAN software package (Research Triangle Institute, Research Triangle Park, NC) was used for logistic analysis because of the differential sampling scheme used for selecting white and nonwhite control subjects.6 The analysis accounted for cluster-correlated responses and controlled for known risk factors for VLBW. Specifically, in our multivariate analyses, variables were included in the final logistic model because of their established relationship with infant birth weight and were defined and categorized as described below.

The maternal weight gain variable was derived by dividing the absolute weight gain experienced by the mother during pregnancy by the gestational age of the newborn. In this way, weight gain was normalized, giving an estimate of the weight gained by the mother per week of gestation. The reference level for weight gain, a mean of 0.79 lb (0.36 kg) per week for mothers of control infants, approximated the average weight gain per week for the infants born within the range of 16 to 42 weeks of gestation.32 Maternal prepregnancy weight was dichotomized as <100 lb or ≥100 lb. The cutoff points for adequate and inadequate prenatal care were based on a modified Adequacy of Prenatal Care Utilization Index.33 The Adequacy of Prenatal Care Utilization Index is based on the gestational age of the newborn, the occurrence of the first visit, the expected number of visits, and the observed number of prenatal care visits recorded from medical records. Race was dichotomized as nonwhite or white. The other independent variables were financial stress during the pregnancy (the family's ability to pay medical bills; yes versus no), income (<$10000 per year versus ≥$10000 per year), maternal education (less than high school versus high school or more), paternal education (less than high school versus high school or more), maternal prepregnancy weight (<100 lb versus ≥100 lb), toxemia (yes versus no), parity (primiparous versus multiparous), mother working during pregnancy (yes versus no), maternal age (in years), alcohol consumption during the pregnancy (yes versus no), illicit drug use during pregnancy (yes versus no), cigarette smoking during pregnancy (yes versus no), exposure to passive cigarette smoke in the household (yes versus no), and gender of the newborn (male versus female).

We considered season of birth as a potential confounder and adjusted for it in our logistic modeling procedures, because of its reported relationship with preterm birth34–37 and our observation of substantial variability in the 3-year seasonal averages for total PM10 measurements (as measured with ambient monitors in GHCD927) among the seasons. Birth months were categorized as follows, with the 3-year seasonal average for total PM10 measurements noted. Spring (March to May) had the highest seasonal average of PM10 (27.33 µg/m3), with a maximal average of 44.18 µg/m3. Summer (June to August) had a seasonal average of 25.85 µg/m3 and a maximal average of 40.88 µg/m3, fall (September to November) had a seasonal average of 25.56 µg/m3 and a maximal average of 43.08 µg/m3, and winter (December to February) had the lowest 3-year ambient PM10 pollutant average (23.99 µg/m3), with a maximal average value of 39.42 µg/m3. We did not control for gestational age in our logistic modeling procedures because of the high correlation between gestational age and the birth outcome categories under study and because birth weight given gestational age was accounted for in the algorithm for classifying infants as SGA versus AGA.

With the PM10 exposure estimates generated from the environmental transport model, we also used the logistic modeling procedure to evaluate PM10 as a continuous exposure variable (in increments of 10 µg/m3) for the various birth outcome comparisons. Finally, we used the Cochran-Armitage test for trend to determine whether the observed proportions of case infants and control infants differed in a linear manner across exposure categories (according to a dose-response relationship) for the PM10 exposure estimates generated from the environmental transport model.


   RESULTS
TOP
 ABSTRACT
 METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Table 1 shows the characteristics of the study population (women and their birthed infants) according to birth outcome category. Both the median gestational age and the mean birth weight for the preterm/AGA group were significantly less than those for the preterm/SGA group (P = .0013 and .0018, respectively). There were statistically significant differences in the prevalence of the various characteristics among the birth outcome groups, and these factors were controlled for in the multivariate analyses.


View this table:
[in this window]
[in a new window]
  		TABLE 1 Characteristics of Study Population, According to Birth Outcomes

 
Table 2 describes the exposure distribution of the study population according to birth outcome category. The exposure distribution shows a statistically significant difference in the proportions of mothers residing in a county with a point source according to birth outcomes, with the largest proportion being observed for mothers who delivered preterm/AGA infants. There was also a statistically significant difference in the median PM10 exposure estimates (according to the environmental transport model) among the 3 birth outcome categories, with the highest median value of PM10 exposure being observed for the preterm/AGA group (7.84 µg/m3), followed by the preterm/SGA group (3.38 µg/m3) and the term/AGA group (3.23 µg/m3). There was not a statistically significant difference in the observed proportions of mothers with estimated PM10 exposure in the fourth quartile among the birth outcome categories (Table 2). For mothers who delivered preterm/AGA infants, however, the test for trend indicated a statistically significant dose-response relationship across the PM10 exposure categories for the environmental transport model (P = .015).


View this table:
[in this window]
[in a new window]
  		TABLE 2 Estimated Exposures of Study Population, According to Birth Outcomes

 
Results of bivariate and multivariate logistic regression analyses for the 2 exposure measures for PM10 with all VLBW infants as case subjects and term/AGA infants as control subjects are shown in Table 3. The county-level spatial logistic model resulted in a significant adjusted OR (aOR) of 2.54 (95% confidence interval [CI]: 1.46–4.22) for residence within a county with an industrial point source. The PM10 environmental transport logistic model resulted in an aOR of 1.94 (95% CI: 0.98–3.83) for mothers with PM10 exposures in the fourth quartile. Evaluation of estimated PM10 (from the environmental transport model) as a continuous exposure variable (in increments of 10 µg/m3) in the logistic model resulted in an aOR of 1.21 (95% CI: 1.03–1.42).


View this table:
[in this window]
[in a new window]
  		TABLE 3 Maternal Exposure to PM10 and Delivery of VLBW Infants, Compared With Term/AGA Infants

 
Results of bivariate and multivariate logistic regression analyses for the 2 exposure measures for PM10 with preterm/AGA infants as case subjects and term/AGA infants as control subjects are shown in Table 4. The county-level spatial logistic model resulted in a statistically significant aOR of 4.31 (95% CI: 1.88–9.87) for residence within a county with an industrial point source. The PM10 environmental transport logistic model also resulted in a statistically significant aOR of 3.68 (95% CI: 1.44–9.44) for mothers with PM10 exposures in the fourth quartile. Evaluation of estimated PM10 (from the environmental transport model) as a continuous exposure variable (in increments of 10 µg/m3) in the logistic model resulted in an aOR of 1.24 (95% CI: 1.04–1.47).


View this table:
[in this window]
[in a new window]
  		TABLE 4 Maternal Exposure to PM10 and Delivery of Preterm/AGA Infants, Compared With Term/AGA Infants

 
Results of bivariate and multivariate logistic regression analyses for the 2 exposure measures for PM10 for preterm/AGA and preterm/SGA infants are shown in Table 5. The county-level spatial logistic model resulted in an aOR of 2.07 (95% CI: 0.83–5.16) for residence within a county with an industrial point source for preterm/AGA infants, compared with preterm/SGA infants. The PM10 environmental transport logistic model resulted in an aOR of 2.58 (95% CI: 0.78–8.51) for mothers with PM10 exposures in the fourth quartile for preterm/AGA delivery, compared with preterm/SGA delivery. The measure of association was >1 for each exposure measure, but statistical significance was not achieved in either case.


View this table:
[in this window]
[in a new window]
  		TABLE 5 Maternal Exposure to PM10 and Delivery of Preterm/AGA Infants, Compared With Preterm/SGA Infants

 

   DISCUSSION
TOP
 ABSTRACT
 METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
This study demonstrated a significant increase in the odds of maternal exposure to PM10, with 2 different exposure measures (ie, a county-level spatial measure based on residence within a county with an industrial point source of PM10 and an environmental transport model based on annual PM10 emissions from geographically identified point sources and the geographic location of the birth homes, yielding an ambient PM10 residence estimate), for VLBW infants delivered preterm/AGA, compared with term/AGA infants. Similarly, this study demonstrated an increase in the odds of maternal exposure to PM10, with both exposure measures, for the subpopulation of VLBW infants classified as preterm/AGA, compared with infants classified as preterm/SGA, although statistical significance was not achieved. In this study, the VLBW infants classified as preterm/AGA were of statistically significantly lesser gestational age (and lower birth weight) than were those classified as preterm/SGA (Table 1). Results from this study are consistent with those of previous studies that used a county-based exposure index of air pollution8,18 and an exposure index based on birth home location9 and found a statistically significant relationship with preterm delivery. Taken together, findings from this study and others suggest that the observed association between maternal exposure to air pollution and low infant birth weight can be attributed at least in part to the effect of air pollution on the duration of gestation, particularly when infants born in the VLBW category are considered.

Power calculations reveal that, for the sample of mothers who delivered preterm/AGA (n = 59) and preterm/SGA (n = 69) infants in GHCD9 during the study period, our analysis had 69% and 76% power to detect a significant difference in odds of maternal exposure to PM10 with the county-level and environmental transport exposure measures, respectively. Although the analytical results for the comparison between VLBW preterm/AGA and preterm/SGA infants did not achieve statistical significance, scrutiny of other aspects of the data (Table 2) supports our conclusion that maternal exposure to PM10 seems to influence the duration of gestation for infants delivered VLBW. There is a significant difference in the median PM10 exposure estimates between mothers who delivered preterm/AGA infants and preterm/SGA infants, with the median exposure for preterm/AGA deliveries being almost twice that for preterm/SGA deliveries. In addition, a significantly higher percentage of mothers with preterm/AGA deliveries lived in counties with a point source, compared with mothers with preterm/SGA or term/AGA deliveries.

Although the estimates of effect for the statistically significant relationships between maternal exposure to air pollution and adverse birth outcomes found in this study might seem high, compared with those found in other studies, it is difficult to make direct comparisons, for several reasons. Other studies evaluated principally LBW deliveries, rather than VLBW deliveries. In many cases, other studies evaluated continuous increments of exposure, rather than categories of exposure. In addition, the subjects for this study were drawn from GHCD9, which has a moderately high overall VLBW delivery rate of 1.8% (ranging from 1.5% in some counties to 5.2% in others),38 and were of comparatively low socioeconomic status, with high prevalence of a variety of other conditions linked to LBW delivery (Table 1). Therefore, the results of this study may indicate that those already at high risk of VLBW delivery are especially sensitive to the potential effects of air pollution. More research is necessary to clarify this issue.

Results from this study seem inconsistent with those of the Czech-US Environmental Protection Study,19 which found an association between exposure to PM10 in the first month of pregnancy and IUGR, and with a study that found an association between high levels of maternal exposure to carbon monoxide and sulfur dioxide, but not PM10, and IUGR.7 Both of those studies, however, excluded preterm infants, whereas our study included only VLBW infants as case subjects, all of whom were born preterm (most at <32 weeks of gestation). It is possible that, if we had evaluated preterm and SGA outcomes for infants not classified as VLBW, then we might have observed an association between IUGR and PM10 exposure. Two other studies8,18 did include SGA infants not classified as VLBW and still failed to find an association between maternal exposure to ambient air pollution and IUGR.

Discrepant results among the various studies might also be attributable to differences in specific air pollutants studied, exposure levels for air pollutants (attributable to differences in geographic locations and sources of pollutants), methods used for estimating exposure, and variables controlled for in the analysis. Particulate air pollution, in particular, is difficult to study because it is not an exactly defined toxicant and its composition depends on the particular industrial source. Furthermore, particulate matter may serve as a surrogate for an associated component of air pollution, because it has been correlated significantly with sulfur dioxide (PM10, r = 0.75), nitrogen oxide species (PM10, r = 0.37), and polycyclic aromatic hydrocarbons (PM10, r = 0.79) (all P < .001), or may serve as a surrogate for some unmeasured toxicant.19 Seemingly variable results from studies of maternal exposure to air pollution and birth outcomes may be related to differences in measurements related to the timing of maternal exposure.23 In particular, exposure to PM10, carbon monoxide, sulfur dioxide, and nitrogen dioxide during early to middle pregnancy seems to contribute to the risk of LBW.39

The results of our study are likely not attributable to bias. Although there was likely some misclassification of true maternal exposure to PM10 with the exposure measures, this misclassification was most likely random, which would tend to underestimate the effects of PM10. Although it is true that mothers who are clustered geographically are likely to be more similar to each other, with respect to a number of factors, than are mothers who live farther from one another, we were able to account for this clustering by using multilevel modeling in our logistic regression procedures. Furthermore, confounding by socioeconomic status or race is unlikely, because adjustment for socioeconomic and race risk did not attenuate the effect estimates.

This study has limitations because of its inability to measure some potential sources of air pollution and other factors linked to infant birth weight. In this study, we were not able to assess residential proximity to heavy traffic, which has been found to be correlated with preterm delivery9,40 and might be correlated with our measures of PM10 exposure, potentially confounding the relationships observed in our study. Of note, however, GHCD9 is mostly rural, except for the city of Savannah. Other potential sources of exposure, such as living near an unmonitored industrial point source or landfill, unmeasured environmental stressors such as noise, access to health care (although appropriate prenatal care did not have a significant association with preterm birth in GHCD96), and nutrition may confound the association between VLBW delivery and ambient air pollution.41,42 Nonpoint sources of particulate matter and sources outside the Georgia Environmental Protection Division requirements for monitoring were not included in our model and might contribute to exposures, particularly in urban areas. However, the environmental transport model was validated with data from ambient monitors in Chatham County and was found to predict with reasonable accuracy the recorded values at the ambient monitoring stations, whose recordings include all sources of particulate matter.6

Data from this study were drawn from the Georgia Very Low Birth Weight Study; therefore, another limitation is that we were able to consider only the birth outcomes of preterm/SGA and preterm/AGA for infants born VLBW in GHCD9. It is possible that, if we had been able to consider preterm/SGA and preterm/AGA infants in the LBW category (1500-2499 g), then our results might have been tempered. For example, if the effect of PM10 on intrauterine growth did not manifest until later in pregnancy, then any observed association with SGA might not have been realized until later gestational ages. To evaluate more fully the potential relationship between maternal exposure to air pollution and SGA delivery, it would be useful to compare preterm/SGA infants and term/SGA infants, who are typically born in the LBW category.

An additional limitation involves the uncertainties in the transport model used in this study. Unpublished results from models estimating the uncertainty in exposure (accounting for uncertainties in the release quantities, transport model, ambient temperature, wind speed and direction, and stack gas temperature and velocities) suggest that ambient PM10 exposures may range from <1 µg/m3 to >48 µg/m3. Such a range of exposures results in aORs that range from a 5th percentile of 2.02 (95% CI: 0.36-8.58) to a 95th percentile of 10.64 (95% CI: 2.19-76.30), with a median of 4.30 (95% CI: 1.64-10.91), for the comparison between VLBW deliveries classified as preterm/AGA and term/AGA deliveries. Finally, the sample size for this study was relatively small and drawn from a geographically distinct area (GHCD9); therefore, we must be cautious in drawing conclusions and must look at this study as hypothesis generating.

The study also has several strengths. The modeling procedure was able to consider independently the outcomes of preterm/SGA and preterm/AGA, as determined with data obtained from a reliable medical source, specifically among infants born VLBW, and included a host of other factors known to influence birth weight. Data on air pollution also were obtained from a reliable source (Georgia Department of Natural Resources), and the monitored facilities are known to account for ~95% of the total environmental releases in GHCD9.27 In addition, the study results were quite consistent across the 2 logistic models for the evaluated PM10 exposure measures and for other evaluated variables known to be important for VLBW delivery. For example, for both logistic modeling procedures for preterm/AGA delivery versus term/AGA delivery, the variables that we found to have a significant association with preterm/AGA delivery were maternal age, passive smoking, maternal weight gain, birth month, and maternal anemia (data not shown). For both logistic modeling procedures for preterm/SGA delivery versus preterm/AGA delivery, the variables that we found to have a significant association with preterm/SGA delivery were maternal race, maternal prepregnancy weight, active or passive smoking, and maternal asthma (data not shown). Finally, the environmental transport model suggested a dose-response effect for maternal exposure to PM10 and preterm/AGA delivery, lending support to the existence of a true association.


   CONCLUSIONS
TOP
 ABSTRACT
 METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
Numerous studies performed in various geographic locales have linked maternal exposure to ambient air pollution to adverse birth outcomes (including LBW, SGA, and preterm delivery) that are themselves associated with increased infant mortality and morbidity rates.4 With this study, we found increased odds of maternal exposure to ambient PM10 for VLBW infants delivered preterm/AGA, compared with term/AGA infants, which suggests that the observed association between maternal exposure to PM10 and infant LBW (particularly <1500 g) may be at least partially attributable to an effect on duration of gestation. With this study, we are unable to evaluate fully the potential relationship between maternal exposure to PM10 and IUGR, because the lack of an observed statistically significant association could be attributable to low statistical power for evaluation of this relationship, particularly if the true effect size is small. Although this study assessed specifically the association between various measures of maternal exposure to PM10, it should be noted that PM10 is one of many pollutants in the atmosphere, some of which are monitored and some not. Therefore, PM10 may actually be a surrogate marker for the causative agent or agents responsible for the observed associations, particularly because PM10 measurements are correlated strongly with sulfur dioxide, nitrogen oxide species, and polycyclic aromatic hydrocarbon measurements.6,19

Across the United States, the prevalence of exposure to ambient air pollution among pregnant women, particularly in urban settings or settings near a point source of pollution, is appreciable. Therefore, even a small increase in the risk of adverse birth outcomes resulting from maternal exposure to ambient air pollution could translate into substantial adverse health events. Because there are still unanswered questions regarding the role of specific components of air pollution and their effects on birth outcomes, future studies should assess birth outcomes defined carefully according to gestational age and size for gestational age across the birth weight spectrum and should include examination of the physical and chemical properties of the air pollutants under study. Future studies also should consider other adverse birth outcomes (eg, spontaneous abortion and stillborn delivery) in assessments of the impact of maternal exposure to air pollution. Such data should increase our understanding of the potential role of the environment on adverse birth outcomes.


   ACKNOWLEDGMENTS
 
This study was supported in part by funds from the Comprehensive Environmental Response, Compensation, and Liability Act Trust Fund, through an interagency agreement between the Agency for Toxic Substances and Disease Registry and the National Center on Birth Defects and Developmental Disabilities, Centers for Disease Control and Prevention.

We thank Dr Anne Sowell, Office of the Director, National Center for Environmental Health; Dr Adolfo Correa, National Center on Birth Defects and Developmental Disabilities; Dr Dana Flanders, Dr Carol Gotway, and Vicky Booth, National Center for Environmental Health/Environmental Hazards and Health Effects; and Dr Jane Ellis, Department of Obstetrics and Gynecology, Maternal and Fetal Medicine, Grady Health Care System (Atlanta, GA); for review of the manuscript. We also thank Amy Guinn for editorial review.


   FOOTNOTES
 
Accepted Feb 10, 2006.

Address correspondence to Anne L. Dunlop, MD, MPH, Department of Family and Preventive Medicine, Emory University School of Medicine, 735 Gatewood Rd NE, Atlanta, GA 30322. E-mail: amlang{at}emory.edu

The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.

The authors have indicated they have no financial relationships relevant to this article to disclose.


   REFERENCES
TOP
 ABSTRACT
 METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 

  1. Woodruff TJ, Grillo J, Schoendrof KC. The relationship between selected causes of postnatal infant mortality and particulate air pollution in the United States. Environ Health Perspect. 1997;105 :608 –612[ISI][Medline]
  2. Loomis D, Castilljos M, Gold DR, McDowell W, Borja-Aburto VH. Air pollution and infant mortality in Mexico City. Epidemiology. 1999;10 :118 –123[CrossRef][ISI][Medline]
  3. Samet JM, Dominici F, Curriero FC, Coursac I, Zeger SL. Fine particulate air pollutants and mortality in 20 US cities, 1987–1994. N Engl J Med. 2000;343 :1742 –1749[Abstract/Free Full Text]
  4. Srám RJ, Binková B, Dejmek J, Bobak M. Ambient air pollution and pregnancy outcomes: a review of the literature. Environ Health Perspect. 2005;113 :375 –382[ISI][Medline]
  5. Wang X, Ding H, Ryan L, Xu X. Association between air pollution and low birth weight: a community-based study. Environ Health Perspect. 1997;105 :514 –520[ISI][Medline]
  6. Rogers JF, Thompson SJ, Addy CL, McKeown RE, Cowen DJ, DeCoulfe P. The association of very low birth weight with exposures to environmental sulfur dioxide and total suspended particulates. Am J Epidemiol. 2000;151 :602 –613[Abstract/Free Full Text]
  7. Maisonet M, Bush TJ, Correa A, Jaakkola JJK. Relation between ambient air pollution and low birth weight in the northeastern United States. Environ Health Perspect. 2001;109 (suppl 3):351–356
  8. Bobak M. Outdoor air pollution, low birth weight, and prematurity. Environ Health Perspect. 2000;108 :173 –176[ISI][Medline]
  9. Wilhelm M, Ritz B. Residential proximity to traffic and adverse birth outcomes in Los Angeles County, California. Environ Health Perspect. 2003;111 :207 –216[ISI][Medline]
  10. Dolk H, Pattenden S, Vrijheid M, Thakrar B, Armstrong B. Perinatal and infant mortality and low birth weight among residents near cokeworks in Great Britain. Arch Environ Health. 2000;55 :26 –30[ISI][Medline]
  11. Bhopal RS, Phillimore P, Moffatt S, Foy C. Is living near a coking works harmful to health? A study of industrial air pollution. J Epidemiol Community Health. 1994;48 :237 –247[Abstract]
  12. Villar J, Khoury MJ, Finucane FF, Delgado HL. Differences in the epidemiology of prematurity and intrauterine growth retardation. Early Hum Dev. 1986;14 :307 –322[CrossRef][ISI][Medline]
  13. Lang JM, Lieberman E, Cohen A. A comparison of risk factors for preterm labor and term small-for-gestational age birth. Epidemiology. 1996;7 :369 –376[ISI][Medline]
  14. Xu X, Ding H, Wang X. Acute effects of total suspended particulates and sulfur dioxide on preterm delivery: a community-based cohort study. Arch Environ Health. 1995;50 :407 –415[ISI][Medline]
  15. Ritz B, Yu F, Chapa G, Fruin S. Effect of air pollution on preterm birth among children born in Southern California between 1989 and 1993. Epidemiology. 2000;1 :502 –511
  16. Yang CY, Tseng YT, Chang CC. Effects of air pollution on birth weight among children born between 1995 and 1997 in Kaohsiung, Taiwan. J Toxicol Environ Health A. 2003;66 :807 –816[CrossRef][ISI][Medline]
  17. Tsai SS, Yu HS, Liu CC, Yang CT. Increased incidence of preterm delivery in mothers residing in an industrialized area in Taiwan. J Toxicol Environ Health A. 2003;66 :987 –994[CrossRef][ISI][Medline]
  18. Woodruff TJ, Parker JD, Kyle AD, Schoendrof KC. Disparities in exposure to air pollution during pregnancy. Environ Health Perspect. 2003;111 :942 –946[ISI][Medline]
  19. Dejmek J, Selevan SG, Benes I, Solansky I, Sram RJ. Fetal growth and maternal exposure to particulate matter during pregnancy. Environ Health Perspect. 1999;107 :475 –480[ISI][Medline]
  20. Berkowitz GS, Janevic TM, Holzman IR, Yehuda R, Landrigan PJ. The World Trade Center disaster and intrauterine growth restriction. Res Lett. 2003;290 :595a
  21. Parker JD, Woodruff TJ, Basu R, Schoendorf KC. Air pollution and birth weight among term infants in California. Pediatrics. 2005;115 :121 –128[Abstract/Free Full Text]
  22. Lemons JA, Bauer CR, Oh W, et al. Very low birth weight outcomes of the National Institute of Child Health and Human Development Neonatal Research Network, January 1995 through December 1996. Pediatrics. 2001;107 (1). Available at: www.pediatrics.org/cgi/content/full/107/1/e1
  23. Wilson JG. Embryological considerations. In: Wilson JG, Warkney J, eds. Teratology: Principles and Techniques Chicago, IL: University of Chicago Press; 1965:256 –277
  24. Ott WJ. Accurate gestational dating: revisited. Am J Perinatol. 1994;11 :404 –408[ISI][Medline]
  25. Zhang I, Bowes WA Jr. Birth weight-for-gestational-age patterns by race, sex, and parity in the United States population. Obstet Gynecol. 1995;86 :200 –208[Abstract]
  26. Rogers JF, Killough GG, Thompson SJ, Addy CL, McKeown RE, Cowen DJ. Estimating environmental exposures to sulfur dioxide from multiple industrial sources for a case-control study. J Expo Anal Environ Epidemiol. 1999;9 :535 –545[CrossRef][ISI][Medline]
  27. Georgia Department of Natural Resources. Georgia Environmental Protection Division Emissions Data, 1986–1988 Atlanta, GA: Georgia Department of Natural Resources; 1995
  28. Holland JZ. A Meterological Survey of the Oak Ridge Area: Final Report Covering the Period 1948–1952 Oak Ridge, TN: Weather Bureau; 1953. USAEC report ORO-99.
  29. Meade PJ, Pasquill F. A study of the average distribution of pollution around Staythorpe. Int J Air Pollut. 1951;1 :60 –70
  30. Hamby DM. The Gaussian atmospheric transport model and its sensitivity to the joint frequency distribution and parametric variability. Health Phys. 2002;82 :64 –73[Medline]
  31. Hanna SR, Briggs GA, eds. National Technical Information Service Handbook on Atmospheric Diffusion Springfield, VA: US Department of Commerce; 1982
  32. Gabbe SG, Niebyl JR, Simpson JL, eds. Obstetrics New York, NY: Churchill Livingston; 1986
  33. Kotelchuck M. An evaluation of the Kessner Adequacy of Prenatal Care Index and a proposed Adequacy of Prenatal Care Utilization Index. Am J Public Health. 1994;84 :1414 –1420[Abstract/Free Full Text]
  34. Cooperstock M, Wolfe RA. Seasonality of preterm birth in the Collaborative Perinatal Project: demographic factors. Am J Epidemiol. 1986;124 :234 –241[Abstract/Free Full Text]
  35. Wendl Richter HU. Birthweight distribution in rural north-west Burkina Faso. Trop Med Int Health. 1997;2 :404 –408[ISI][Medline]
  36. Rousham EK, Gracey M. Seasonality of low birthweight in indigenous Australians: an increase in pre-term birth or intrauterine growth retardation? Aust NZ J Public Health. 1998;22 :669 –672[Medline]
  37. Rayco-Solon P, Fulford AJ, Prentice AM. Differential effects of seasonality on preterm birth and intrauterine growth restriction in rural Africans. Am J Clin Nutr. 2005;81 :134 –139[Abstract/Free Full Text]
  38. Georgia Department of Human Resources, Division of Public Health. Online Analytical Statistical Information System. Available at: http://oasis.state.ga.us/. Accessed February 1, 2006
  39. Lee BE, Ha EH, Park HS, et al. Exposure to air pollution during different gestational phases contributes to risks of low birth weight. Hum Reprod. 2003;18 :638 –643[Abstract/Free Full Text]
  40. Yang CY, Chang CC, Chuang HY, Ho CK, Wu TN, Tsai SS. Evidence for increased risks of preterm delivery in a population residing near a freeway in Taiwan. Arch Environ Health. 2003;58 :649 –654[Medline]
  41. Clay R. Still moving toward environmental justice. Environ Health Perspect. 1999;107 :A308 –A310[Medline]
  42. Fox MA, Groopman JD, Burke TA. Evaluating cumulative risk assessment for environmental justice: a community case study. Environ Health Perspect. 2002;110 :203 –209
PEDIATRICS (ISSN 1098-4275). ©2006 by the American Academy of Pediatrics




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow P3Rs: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when P3Rs are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow E-mail this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My File Cabinet
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via CrossRef
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Rogers, J. F.
Right arrow Articles by Dunlop, A. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Rogers, J. F.
Right arrow Articles by Dunlop, A. L.
Related Collections
Right arrow Office Practice