Grantee Research Project Results
2004 Progress Report: Exposure to Vehicular Pollutants and Respiratory Health
EPA Grant Number: R827352C007Subproject: this is subproject number 007 , established and managed by the Center Director under grant R827352
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
Center: UC Berkeley/Stanford Children’s Environment Health Center
Center Director: Tager, Ira
Title: Exposure to Vehicular Pollutants and Respiratory Health
Investigators: McConnell, Rob Scot , Avol, Edward L. , Lurmann, Fred , Berhane, Kiros , Gauderman, William
Current Investigators: McConnell, Rob Scot , Avol, Edward L. , Lurmann, Fred , Gauderman, William
Institution: University of Southern California , Sonoma Technology, Inc.
Current Institution: University of Southern California
EPA Project Officer: Chung, Serena
Project Period: June 1, 1999 through May 31, 2005 (Extended to May 31, 2006)
Project Period Covered by this Report: June 1, 2003 through May 31, 2004
RFA: Airborne Particulate Matter (PM) Centers (1999) RFA Text | Recipients Lists
Research Category: Air Quality and Air Toxics , Particulate Matter , Air
Objective:
Project Hypotheses
This project is testing the hypothesis that exposure to vehicular pollutants is associated with respiratory health in children. We have examined this hypothesis using data from the Children’s Health Study (CHS), a longitudinal evaluation originally designed to evaluate the effect of average community pollutant exposures on average community respiratory health outcomes. Pollution has been well characterized at central site monitors in 16 southern California communities involving participants from two sets of school based cohorts. In the SCPCS we have characterized within-community variation in traffic related pollutant exposure. These new pollutant metrics have allowed us to examine the relationship of individual level exposures to the CHS outcomes.
(1) We have developed modeled metrics of exposure and tested these against measured pollutants.
(2) We have evaluated the association of these metrics and of measured indicator pollutants to: (2a) prevalence and severity of asthma at entry; (2b) incidence of asthma during follow-up; (2c) lung function at study entry and its growth during follow-up; and (2d) school absence.
(3) In addition, among asthmatics we have evaluated the association of chronic asthma exacerbation with chronic temporal variation from year-to-year in exposure to oxidant pollutants.
Project Objectives
Our approach has been to geo-code addresses of residences and schools and then to assign traffic exposures to these addresses, based on traffic estimates available from the California Department of Transportation (Caltrans). We have used ambient air quality measurements at 12 CHS community monitors to evaluate these models. Using modifications of statistical modeling strategies developed for the CHS, we have examined the effect of exposure to traffic related pollutants on asthma prevalence, and we have found relationships not previously reported in this growing literature. In addition, we have examined outcomes for which there has been little previous study of the effects of traffic modeled pollutants: asthma severity, asthma incidence, lung function and lung function growth, and school absence.
Progress Summary:
Aim 1: Characterization of Exposure to Traffic-Related Pollutants in CHS Communities
The overall objective of the exposure component of this study was to provide improved methods and databases to characterize seasonal and annual average population exposure to traffic-related pollutants with fine spatial resolution. Prior to this study, Sonoma Technology, Incorporated (Fred Lurmann) had performed some initial traffic modeling for the CHS which pointed to the need for model refinements, database improvements, and model performance evaluation. Many potential improvements were explored; improvements in the spatial accuracy of roadway and receptor data, meteorological data, and emission rates proved most important. It is important to note that most assessments of traffic-related pollutant exposure focus on worst-case 1-hr or 8-hr maximum conditions. Our focus was on estimation of long-term exposures in order to evaluate relationships with chronic health effects.
Traffic Activity and Roadway Data. Annual average daily traffic counts data for Interstate freeways, other principal arterials, minor arterials, major collectors, and minor collectors in 2000 were obtained from Caltrans. The annual traffic counts are based on continuous measurement data for freeways and intermittent measurements (usually every three years) on other arterials and some collectors. Annual traffic volumes for 1994-2004 were backcasted and forecasted from the 2000 data assuming a 2 percent per year growth rate. Diurnal traffic volume variations and day-of-week variations for light-duty and heavy-duty vehicles on freeways were determined from Caltrans weigh-in-motion (WIM) data for Southern California. Diurnal variations for collectors were determined from more limited traffic measurements (Chinkin, et al., 2003).
Additional databases were used to apportion the total traffic volume to heavy-duty diesel vehicles (HDV) and light-duty gasoline vehicles (LDV). For the seven CHS communities located within the southern California Air Basin (SoCAB), including Orange County and portions of Los Angeles, Riverside, and San Bernardino Counties, HDV fractions of total traffic volumes were derived from travel-demand transportation model simulations. The Southern California Association of Governments (SCAG) performed separate simulations of LDV and HDV traffic volumes in the SoCAB area for 1997. The simulations incorporate an accurate link-node roadway network for freeways and a surrogate link-node roadway network for nonfreeways. Despite concerns regarding the spatial inaccuracy of the nonfreeway network, it provided much better spatial resolution for the fractions of LDV and HDV than any other database. HDV fractions of total traffic volume in the five communities located outside the SoCAB were derived from the Caltrans statewide truck-traffic-volume database for freeways and state highways for 1998. It is linked to a third roadway network, the state post-mile roadway system. Both databases rely heavily on data from the WIM sensors at selected freeway locations. HDV fractions (~3%) for collector streets were determined from more limited traffic measurements (Chinkin, et al., 2003).
ArcInfo GIS software was used to process the Caltrans roadway link and traffic count data. The Caltrans roadway geometries were mostly based on TIGER files and were often inaccurate, based on comparisons with aerial photography images. Comparison to global positioning system (GPS)-accurate TeleAtlas Roadway Network data showed that the Caltrans’ TIGER roadway links occasionally had up to 250-m discrepancies from actual roadway locations. The roadway geometry errors were random and affected roadways of all sizes in most communities. Zhu, et al. (2002) reported ten-fold differences in measured concentrations of traffic-related pollutants between 30 m and 200 m downwind of Southern California freeways. Errors of 100 m to 250 m in the location of major roadways relative to residences are not acceptable for neighborhood-scale assessments of traffic effects. Since the Caltrans roadway location data did not have sufficient accuracy for our intended use, methods (software) were developed to transfer the Caltrans annual traffic volumes to the GPS-accurate TeleAtlas roadway network. The HDV fractions of total traffic volumes were also mapped from their original network to the TeleAtlas network (Wu, et al., 2005a). The TeleAtlas roadway database incorporated both more accurate and more precise location information. For example, each direction of travel on moderate and large roadways is represented as a separate link in the TeleAtlas database.
The street address of residences and schools of CHS residences were geocoded on the TeleAtlas roads. Considerable effort was made to standardize the ~8500 participant addresses, which resulted in more than 98% of the residences having the highest quality geocoding match quality. Different geocoding services were evaluated; the results indicated that it was essential to (1) geocode receptor locations of interest using the same roadway locations employed for the traffic assessment and (2) confirm geocodes of large properties (e.g., schools) with aerial photographic images.
Proximity Modeling. We wanted to examine the relationships of CHS participant health status with traffic indicators separately from the CALINE4 dispersion model estimates of concentrations from mobile source emissions. Dispersion modeling provides refinements but introduces additional uncertainties compared to analysis of traffic alone. Hence, a three-level hierarchical approach was adopted for traffic assessment that considered (1) the distance of residences to nearest roadways of various types, (2) GIS-mapped traffic density assignments at residences, and (3) the CALINE4 dispersion model estimates of traffic-related pollutant concentrations at residences.
The first sets of traffic metrics were the distances from residences to the nearest roadways of different types and the associated LDV and HDV traffic volumes on those roads. GIS tools were used to calculate the distance to the nearest (1) interstate freeway, U.S. highway, or limited access highway; (2) other highways; (3) arterial roads; (4) collector roads; and (5) local roads. An advantage of this metric is that the calculations were carried out for all roads, not just the roads for which traffic volumes were available. Traffic volumes were not available on most local roads and LDV/HDV traffic volume splits were not known with certainty on most arterial, collector, and local roads. The database of distances and volumes metrics were compiled and used to obtain an understanding of which subjects were living within the zone of influence (e.g., 50 m, 100 m, 150 m, 200 m, etc.) of busy roads.
The second approach for characterization of traffic exposures was to calculate traffic densities, which vary more smoothly in space than the distances to nearest roads. They also capture the effects of intersection and multiple roadway influences that are missed using only distance to the nearest roadways. The link-based traffic volumes are used to generate maps of traffic density using the ARCInfo Spatial Analyst software. Figure 1 shows a traffic density map created with a Gaussian decay function that has traffic densities decreasing by ~90% between the roadway and 150 m away (perpendicular) from the roadways, which is consistent with the characteristics observed by Zhu, et al. (2002). Identical mapping procedures are used in all the communities so that the results are comparable across communities. The densities reflect proximity to traffic without consideration of differential exposures caused by meteorology. The traffic densities are mapped as if the wind speeds and directions were uniformly distributed across all quadrants. The traffic density map for Riverside clearly shows high densities in narrow bands along the freeways, moderate densities along major arterials, and lower densities in the suburban neighborhoods. A database of densities at all of the CHS residences was compiled for use in the health analysis. Comparison of the traffic densities with other metrics indicates the densities are more closely related to the dispersion model estimated concentrations than distance to nearest major road.
Figure 1. Traffic Densities in Riverside, California
Dispersion Modeling. The CALINE4 model, developed by Caltrans and the U.S. Federal Highway Administration (Benson, 1989), is one of several Gaussian line source dispersion models that is designed to estimate local-scale pollutant concentrations from motor vehicle emissions. Often it is used to estimate worst case 1-hr or 8-hr maximum CO concentrations from a congested roadway. It has primarily been evaluated for inert traffic-related pollutants such as CO. It was selected for the CHS analyses because it has a credible scientific formulation and it is recommended by Caltrans for analysis of CO imposed from transportation projects in California (Garza, et al., 1997). The objective of the CHS application is to characterize long-term exposures (annual or warm season and cool season) from vehicle emissions on all roads in a community. The principal inputs to the model are meteorological conditions, traffic volumes, roadway geometry, and vehicle emission rates. The model was used to simulate ambient concentrations due to on-road motor vehicle emissions on all roads with traffic volume data located within a 20 km square centered in each CHS community. The approach used to specify these inputs for long-term exposure is somewhat different than is used for the worst case analyses.
A climatological approach was used to estimate long-term average concentrations. The model was applied for a wide range of meteorological cases in each community, and the seasonal or annual concentrations were calculated by weighting the results for individual cases by the frequency of occurrence of the conditions in the community. Typically, the model was run for 6 wind speeds, 16 wind directions, and 3 atmospheric stability conditions (morning, afternoon, and nighttime). The frequency of occurrence of meteorological conditions was determined from five years of surface observations (1995-1999) measured in or near the communities. The cases were run for an inert pollutant using daily average hourly traffic volumes. The results are post-processed to incorporate not only the meteorological frequency of occurrence but also the pollutant-specific emission factors, diurnal and day-of-week variations of traffic volumes, and chemical conversion (for NO to NO2). Numerous approaches were explored for using wind observations and post-processing. The approach that worked best for winds was the use of separate hourly wind frequency distribution for the warm season and cool season. For long-term NO2 modeling, we obtained more consistent results when the average measured hourly NO2/NOx ratios was applied to simulated NOx concentrations than by using the CALINE4 model’s ozone limiting method to estimate NO2.
Vehicle emission factors were obtained from the California Air Resources Board’s EMFAC2002 vehicle emissions model. County vehicle registration data and community monthly average temperatures were used in the model to estimate the fleet average NOx, CO, and PM emission rates for light-duty gasoline and heavy-duty diesel vehicles traveling at speeds ranging from 20 to 70 mph in 1994 through 2000. The emission factors for nonfreeways were based on the average of the emission factors for 20, 30, and 40 mph. Vehicle emission factors on freeways were based on the average of the emission factors for 50, 60, and 70 mph. The elemental carbon (EC) and organic carbon (OC) fractions of exhaust PM emissions were based on composite profiles from Fillies and Gertler (Gillies and Gertler, 2000). Paved road-dust emission factors for PM2.5 and PM10 were based on Southern California in-roadway measurements (Fitz and Bufalino, 2002). The EMFAC model also estimates the PM emissions from brake wear and tire debris.
Figures 2 and 3 show comparisons of CALINE4 model estimates for the air monitoring station locations to the 4-year average observed ambient concentrations at the stations. There was correlation between the model estimates and the observations for all pollutants. The dispersion model estimates local motor vehicle emissions, which contributed 24% of the observed NO2 concentrations, on average, and the coefficient of determination was 0.57. The model’s estimates for NO2 in Long Beach were higher than for other stations and were a larger percentage (40%) of the observed ambient concentration. The high predicted concentration for Long Beach was due not only to the large amount of local traffic, but also a tendency for the CALINE4 model to grossly overestimate contributions from roadway links more than 5 km from the receptor (of which there were many more in Long Beach than any other community). The model estimates for NOx were more strongly correlated with observations (r2 = 0.67) than those for NO2. This result was consistent with the dispersion model being formulated for chemically non-reactive species and, therefore, the model was more accurate for NOx than NO2. The model results were more strongly correlated when the values for Long Beach were excluded (r2 = 0.81 for NO2 and r2 = 0.97 for NOx).
The average estimated contribution of local motor vehicle emissions to the observed PM2.5 EC concentrations was 26% and the coefficient of determination was 0.57. The predicted PM2.5 EC concentrations for Long Beach were much higher than those predicted for any other community monitoring station, yet the average predicted PM2.5 EC concentration was only 40% of the observed concentration. The relationships of the model results to the observed concentrations were very similar for PM2.5 EC and NO2. This was somewhat surprising given that EC emissions from vehicles are not well characterized and most emissions experts consider PM (and EC) emissions rates to be far more uncertain than those for NOx.
The model estimates for PM2.5 OC were not well correlated with the observed concentrations (r2= 0.11). The model estimated that the highest PM2.5 OC from local mobile sources occurred in Long Beach. The model estimate for the Mira Loma station location was quite low compared to the very high OC concentration observed in the station. The correlation with the observed value was higher (r2 = 0.71) if the data from Long Beach and Mira Loma were excluded. The comparison of estimated and observed PM2.5 OC was confounded by not only the contributions of regional transport and other local sources, but also the contribution of secondary OC to the measured OC.
Simulations of concentrations from local roads in CHS communities confirmed the importance of regional-scale and urban-scale pollutant sources and meteorological transport in southern California. At the types of locations selected for community monitoring, the ambient pollutant concentrations due to transport from upwind and local non-mobile sources were always larger than the simulated concentrations from local on-road motor vehicle sources. This finding was consistent with regional modeling results completed for another SCPCS project le d by Dr. Arthur Winer.
The CALINE4 dispersion model results were also used in the modeling of personal exposure of CHS participants. They were used as input to the Individual Exposure Model (IEM) developed under the SCPCS project le d by Dr. Arthur Winer and described in Wu, et al. (2005b).
Figure 2. Comparison of Annual Average NO2 and NOx Concentrations Estimated by the CALINE4 Model for the Central Air Monitoring Station Locations and the Four-Year Average Observed Ambient Concentrations at the Stations. Note, the observations for San Dimas are based on only two years of data.
- Figure 3. Comparison of Annual Average PM2.5 EC and OC Concentrations Estimated by the CALINE4 Model for the Central Air Monitoring Station Locations and the Four-Year Average Observed Ambient Concentrations at the Stations. Note, the observations for San Dimas and Glendora are based on only two years of data.
At the homes of children in the CHS, the traffic-modeled pollutant exposures should be considered to be indicators of annual average incremental increases due to local traffic on top of background ambient levels. In addition, the estimates for the various pollutants from the dispersion models at homes in the CHS were highly correlated (generally R>0.90). Therefore, health associations with any specific pollutant modeled at homes may be due to that pollutant or to some other highly correlated pollutant in fresh traffic exhaust, including pollutants for which we did not model exposures. Although it is likely from the extensive toxicological results from the SCPCS and elsewhere that components of fine or ultrafine particulate pollutants are responsible for observed associations with health outcomes, other modeled indicator pollutants were sometimes used for the health analyses in the CHS, described below. NO2 or NOx modeled at homes were sometimes used as an indicator of exposure to traffic related pollutants, because we have measured NO2 at a sample of CHS homes. NO2 can be measured inexpensively and is commonly used in epidemiologic studies as an indicator pollutant for within-community variation in traffic related pollutant exposure (although the correlation with actual measurements of particulate exposures of interest has rarely been evaluated). In other analyses, we used traffic-modeled PM2.5 as an indicator of variability in exposure within communities, because we evaluated the correlation of this exposure with measured pollutants. We also observed correlations between traffic-modeled NOx at homes and variability within communities at those homes, as described below.
Aim 2a:Effect of Traffic on Asthma Prevalence and Severity
We evaluated the risk of lifetime asthma at study entry and traffic related pollutant exposure. We observed large risks associated with measured NO2 at homes in a cohort of children recruited into the original CHS study in 1993 and 1996. There were also high risks associated with residential distance to a freeway and with freeway modeled exposure to traffic related pollutants, using modeled NOx exposures developed in Aim 1 (Gauderman, et al., 2005). The strength of this study is that we measured a traffic related pollutant at the homes, which has not been the case for most studies of traffic and asthma, and the consistency of results from measured and modeled exposures makes a causal relationship more plausible. Using modeled exposures, which were available for the entire cohort, we observed a similar association in the entire cohort.
We then examined risks associated with early life asthma in younger children (5-6 years old) in a new cohort of CHS children recruited in 2003, using modeled exposures. The pattern of results largely replicated those observed in the earlier cohorts, and we found a strong relationship between prevalent and lifetime asthma and modeled exposure. In addition, there was a strong relationship with a relatively simple traffic metric, distance to a major road (McConnell, et al., 2005a, 2005b). (See Table 1.)
Table 1. Association of Asthma and Wheeze with Traffic Related Pollution Among Long Term Residents
-
*N is total exposed in each category of exposure; O.R. (95% C.I.) odds ratio (95% confidence interval), adjusted for age, sex, language of questionnaire, community, and race.
†p<.05; §p<.01
The effect varied by parental history of asthma and age of exposure, with much stronger associations observed in those with exposure before age 3 and with no family history. We believe this is an important finding, as some of the inconsistency in the literature may be explainable based on different proportions of susceptible children in different studies. It is also relevant to evaluation of the effect of traffic on incident asthma (discussed below).
Among children with a lifetime history of asthma in the 2003 cohort, we have also examined the relationship of traffic-related pollutants to asthma severity (nocturnal wheeze, shortness of breath with wheezing, or more than 3 attacks of wheezing in the previous year). Traffic related pollution was associated only with asthma severity (OR 1.7 across the IQR for traffic modeled PM2.5 (95% C.I. 1.0, 3.0). A manuscript is in preparation.
During the past year we have undertaken a pilot study to evaluate the spatial relationship of traffic modeled PM2.5 to measured particulate pollution in Long Beach, a CHS community with heavy primary particulate pollution. We used nephelometers as indicators of PM2.5 pollution that could be measured at reasonable cost simultaneously in a relatively large number of sites. Residences of cases of lifetime asthma and controls from the 2003 cohort formed the sampling frame. The goal was to see if short term sampling at many locations would accurately indicate long term intra-community variation in exposure and predict health effects, if the short term samples were co-located with a central site monitor for which continuous historical data are available. Nephelometers were co-located with Ogawa samplers to measure ozone and NOx during seven 2-week cycles at a total of 69 locations. There were 8-12 different residences per cycle. In addition, measurements were made at a central site and 4 other locations during all sampling periods. Traffic related exposure at all locations also was modeled. At selected sites, filters were also collected for PM10, PM2.5, and PM0.25 EC and OC, and these are used in analyses by other SCPCS principal investigators. Preliminary results show that measurements made with nephelometers were only weakly associated with NO, NO2 and ozone. This indicates that NO2, which is often measured as an inexpensive surrogate for traffic related pollutants, may not be a good indicator of intra-community variation in PM2.5, at least in Long Beach. However, traffic modeled residential PM2.5 exposure was associated with lifetime asthma, but the measured pollutants were not. Likely explanations for this discrepancy include the short sampling period, which was unlikely to reflect lifetime exposure. The measurements made at the community central site monitor varied by sampling period (due to varying meteorology), as might be expected. It was disappointing that neither the residual of measurements at the central site and co-located locations in every cycle, nor the ratio of co-located measurements to the central site, were constant across cycles. A possible lesson for future epidemiologic studies is that short term sampling may not be adequate to characterize long term exposures, even if co-located with central site monitoring for which longer term exposure is well characterized. The generalizability of these results from Long Beach, which has a dense network of freeways and other heavy traffic corridors, has not been evaluated in other communities.
Aim 2b: Effect of Traffic on Incident Asthma
Although there is a growing literature of studies of prevalent asthma and traffic, few previous studies have examined the relationship of traffic related pollutants with incident asthma. We have evaluated the effect of traffic modeled exposures on incident asthma in both sets of CHS cohorts. As for prevalent asthma, absence of parental history of asthma identified a group with stronger associations of traffic with health. Among children recruited at age 5-7, preliminary results indicate that there was little effect of traffic related exposure on subsequent new onset of asthma in high ozone communities. However, an increased risk of incident asthma was associated with traffic-modeled PM2.5 in low ozone communities. We hypothesize that this was due to competing risks between the effect of particulate and the effect of ozone. In low ozone communities, there was an association of modeled exposure to fresh traffic exhaust with increasing asthma incidence. This relationship is consistent with biological effects of fine and ultrafine particulate matter that have been observed by other SPCS investigators. In high ozone communities, ozone concentrations vary inversely with fresh traffic exhaust, because NO in fresh exhaust scavenges ozone. Therefore, near busy roadways, there might be decreased asthma risk due to ozone in high ozone communities that obscures the deleterious effect of other pollutants in fresh exhaust.
We also observed relationships between traffic modeled exposure, ozone, and incident asthma in the cohorts of CHS children enrolled in 1994 and 1996. In contrast to the observed effect of traffic on lifetime asthma at study entry, in these children traffic modeled exposure (to NOx) was associated with a decreased risk of incident asthma. One possible explanation is that exposure to traffic related pollutants, which was associated with increased risk of lifetime asthma at study entry, had advanced the onset of asthma to an earlier age among children who would have developed during follow-up if not for early life traffic exposures. This would result in an apparent protective effect of traffic at later ages. However, the strength of the association of traffic modeled pollutants with incident asthma varied by background community ambient ozone exposure. Therefore, these results may also be consistent with a within-community ozone effect. To explore this hypothesis, we first examined how ozone varied within the CHS communities based on traffic patterns, using an existing data set of homes for which ozone was measured simultaneously at the home and the central site monitor (Avol, et al., 1998). There was a strong inverse relationship observed between measured ozone and traffic modeled exposure to NOx that varied spatially within communities (McConnell, et al., 2005). Although it is well known that ozone concentrations are low near major roadways, there has been little previous systematic study of the pattern of spatial variation within communities due to scavenging. Using the proposed prediction model derived from the observed relationship of ozone and traffic modeled NOx from these analyses, we have identified large risks of new onset asthma associated with the within-community variability in ozone exposure (Table 2). We also examined residential distance to a freeway, because this metric indicative of particulate exposure had been associated with lifetime asthma at study entry. There was no significant association with freeway distance by itself, although the risk of asthma was higher close to the freeway and decreased with increasing distance. However, if we adjusted for the effect of ozone, the association of incident asthma with freeway distance became significant. It is possible that an effect of the particulate exposure from freeways was observed only after accounting for the competing risk of asthma associated with ozone, which varied inversely with traffic modeled pollutants.
Table 2. Relative Risk (Hazard Ratios) for Incidence Asthma
The effect of ozone was robust to adjustment for likely confounders. In addition, the effect was modified by parental history of asthma (bigger risk of asthma among those children with no parental history), and the effect of within community variability in ozone was larger in communities with higher average background concentrations of ozone. Although there are other possible interpretations of these results, they are compatible with previous associations of incident asthma with ozone we have observed in the CHS, effects which were also larger in children without a history of atopic characteristics (McConnell, et al., 2002). They are also compatible with recent toxicologic evidence that ozone pollutant exposure causes asthma (Larson, et al., 2004; Miller, et al., 2003; Schelegle, et al., 2003). A manuscript is in preparation.
This confounding of health effects of traffic related pollutants by background ozone merits further study. A new study is underway to explore further this hypothesis using measured ozone and NOx at homes of children with and without incident asthma.
Aim 2c:Effect of Traffic on Lung Function and Lung Function Growth
We have shown deficits in community average lung function in the CHS and particulate measurements made at central site monitors. Deficits in flow rates were observed at study entry and in replicated assessments of 4-year lung function growth rates in different cohorts of children (Gauderman, et al., 2002; Gauderman, et al., 2000; Peters, et al., 1999). Recently, we reported deficits in 8-year growth rates (Gauderman, et al., 2004). These findings are part of an emerging literature indicating that traffic related pollutants compromise lung function. However, there has been little study of within-community variability in traffic-related pollution and lung function and no previous study of its association with childhood lung function growth. We have examined the association of lung function deficits at study entry using a novel estimator of NO2 exposure at homes, using both measured and traffic modeled indicators of exposure in a Bayesian modeling framework (Molitor, et al., 2005). We found deficits in both lung flows and lung volumes associated with exposure to NO2, suggesting a detrimental effect of traffic related pollutants. (As noted above, local variation in NO2 within communities may be considered a surrogate of particulate exposure associated with fresh traffic emissions.)
We have now incorporated traffic modeled exposure into the evaluation of 8-year lung function growth in the CHS. Preliminary analyses indicate that there are effects of traffic modeled exposure, which are independent of the effect of regional background particulate pollution measured at the central site monitors. Thus, both local traffic and ambient background PM2.5 and PM10 are associated with impaired lung function. These results are important because impaired lung function in childhood is associated with lung deficits in adulthood. In adults, impaired lung function is a strong predictor of respiratory morbidity and death. A manuscript is in preparation.
Aim 2d: School Absence and Traffic
School absences are an outcome that has been strongly associated with temporal variability in central site air pollution in the CHS. We have now observed strong associations with CALINE4 derived NOx and with distance to a freeway. The effects were observed exclusively among children with asthma. A manuscript is in preparation.
Aim 3: Chronic Asthma Exacerbation is Associated with Yearly Variation in Central Site Particulate Measurements
Most of our analyses have examined the effect of spatial variation within communities in pollutants modeled from traffic or measured at a sample of homes. Many other studies have examined acute effects of temporal variation in pollution over days to weeks, but little previous work has examined the effect of temporal variation over longer periods on chronic respiratory outcomes.
We developed novel statistical methods to evaluate the effect of year-to-year temporal variation in the average ambient pollutants measured at the central monitoring sites in each community. The relationship of bronchitic symptoms to ambient particulate matter and to particulate EC and OC , nitrogen dioxide (NO2) and other gaseous pollutants was examined in asthmatic children in the CHS (McConnell, et al., 2003a). Symptoms, assessed yearly by questionnaire from 1996-1999, were associated with the yearly variability of PM2.5, organic carbon, NO2 and ozone. The odds ratios associated with yearly within-community variability in air pollution were larger than the effect of the between-community four-year average concentrations. In two pollutant models, the effects of yearly variation in OC and NO2 were only modestly reduced by adjusting for other pollutants, except in a model containing both OC and NO2; the effects of all other pollutants were reduced after adjusting for OC or NO2. We concluded that previous cross-sectional studies of bronchitic symptoms may have underestimated the risks associated with air pollution.
There is considerable experimental literature suggesting that particulate pollution promotes the sensitizing effect of allergen exposure, and chamber studies have demonstrated that co-exposure of asthmatics to relevant allergens and oxidant pollutants results in acute exacerbation of symptoms. Recent literature suggests that there may be similar synergistic respiratory effects of oxidants and endotoxin. However, there has been little supporting evidence from population-based studies that these effects are relevant for public health. Therefore, using our framework for examining the relationship of bronchitic symptoms to yearly variability in air pollution, we hypothesized that stronger effects of air pollution among children with a dog in the home would indicate interaction between air pollution and endotoxin, as existing literature indicates that dogs are a strong predictor of indoor endotoxin exposure, and few asthmatic children in southern California are allergic to dogs. Stronger effects of air pollution among children with a cat would suggest an interaction with allergen, as cat allergy is common in asthmatic children, and cats are weak predictors of endotoxin exposure. For this analysis the effect of each pollutant was scaled to the range of yearly variation in a representative community (with the median range in yearly variability for that pollutant). Among children with a dog, there were strong associations between bronchitic symptoms and all pollutants examined. Odds ratios ranged from 1.30 per 4.2 μg/m3 for PM10 − PM2.5 (95% confidence interval 0.91, 1.87) to 1.91 per 1.2 μg/m3 for organic carbon (1.34, 2.71). Effects were somewhat larger among children who owned both a cat and dog. There were no effects or small effects with wide confidence intervals among children without a dog, and among children who only owned a cat. Our results suggest that dog ownership, a source of residential exposure to endotoxin, worsens the relationship between air pollution and respiratory symptoms in asthmatic children (McConnell, et al., 2003b ; McConnell, et al., 2005).
Summary
- Residential proximity to high traffic corridors and exposure to intra-community variability in traffic-related pollutants in early life were associated with lifetime asthma at study entry in the CHS.
- In later childhood, it is possible that there are competing exposures to ozone and traffic related particulate pollutants that are involved in asthma pathogenesis.
- Lung function and lung function growth were associated with traffic related air pollution estimated both at the home and at the central site monitor.
- School absence was associated with residential traffic modeled NOx at the homes of children with asthma.
- Yearly variation in organic carbon and other particulate pollutants were strongly associated with bronchitis among children with asthma; these effects were modified by dog ownership, an indicator of home endotoxin exposure.
References:
Avol EL, Navidi WC, Colome SD. Modeling ozone levels in and around Southern California homes. Environmental Science & Technology 1998;32:463-468.
Benson P. CALINE4—A dispersion model for predicting air pollutant concentrations near roadways. 1989.
Chinkin LR, Coe DL, Funk TH, Hafner HR, Roberts PT, Ryan PA, et al. Weekday versus weekend activity patterns for ozone precursor emissions in California’s South Coast Air Basin. Journal of the Air & Waste Management Association 2003;53(STI-999679-2225):829-843.
Fitz DR, Bufalino C. Measurement of PM10 emission factors from paved roads using on-board particle sensors. Presented at the U.S. Environmental Protection Agency’s 11th Annual Emission Inventory Conference, Emission Inventories—Partnering for the Future, 2002.
Garza V, Graney P, Sperling D, Niemeier D, Eisinger D, Kear T, et al. Transportation project-level carbon monoxide protocol revised. 1997.
Gauderman WJ, Avol E, Gilliland F, Vora H, Thomas D, Berhane K, et al. 2004. The effect of air pollution on lung development from 10 to 18 years of age. New England Journal of Medicine 2004;351(11):1057-1067.
Gauderman WJ, Avol E, Lurmann F, Kunzli N, Gilliland F, Peters J, et al. Associations between childhood asthma, nitrogen dioxide, and exposure to traffic. Epidemiology (in press, 2005).
Gauderman WJ, Gilliland GF, Vora H, Avol E, Stram D, McConnell R, et al. Association between air pollution and lung function growth in southern California children: results from a second cohort. American Journal of Respiratory and Critical Care Medicine 2002;166(1):76-84.
Gauderman WJ, McConnell R, Gilliland F, London S, Thomas D, Avol E, et al. Association between air pollution and lung function growth in southern California children. American Journal of Respiratory and Critical Care Medicine 2000;162(4 Pt 1):1383-1390.
Gillies JA, Gertler AW. Comparison and evaluation of chemically speciated mobile source PM2.5 particulate matter profiles. Journal of the Air & Waste Management Association 2000;50:1459-1480.
Larson SD, Schelegle ES, Walby WF, Gershwin LJ, Fanuccihi MV, Evans MJ, et al. Postnatal remodeling of the neural components of the epithelial-mesenchymal trophic unit in the proximal airways of infant rhesus monkeys exposed to ozone and allergen. Toxicology and Applied Pharmacology 2004;194(3):211-220.
McConnell R, Berhane K, Gilliland F, London SJ, Islam T, Gauderman WJ, et al. Asthma in exercising children exposed to ozone: a cohort study. Lancet 2002;359(9304):386-391.
McConnell R, Berhane K, Gilliland F, Molitor J, Thomas D, Lurmann F, et al. Prospective study of air pollution and bronchitic symptoms in children with asthma. American Journal of Respiratory and Critical Care Medicine 2003a;168(7):790-797.
McConnell R, Berhane K, Molitor J, Gilliland F, Kuenzli N, Thorne P, et al. Dog ownership enhances symptomatic responses to air pollution in children with asthma, (submitted, 2005).
McConnell R, Berhane K, Molitor J, Gilliland FD. Air pollution and bronchitic symptoms in children: effect modification by dog ownership. American Journal of Respiratory and Critical Care Medicine 2003b;167(7):A36.
McConnell R, Berhane K, Yao L, Lurmann F, Avol E. Predicting residential ozone deficits from nearby traffic. Science of the Total Environment (in press, 2005).
McConnell R, Yao L, Berhane K, Lurmann F, Kuenzli N, Jerrett M. Association of childhood asthma with residence near a major road. American Journal of Respiratory and Critical Care Medicine 2005a;2A522.
McConnell R, Yao L, Berhane K, Lurmann F, Kuenzli N, Jerrett M. Traffic, susceptibility and asthma (s ubmitted, 2005b).
Miller LA, Hyde DM, Gershwin LJ, Schelegle ES, Fanucchi MV, Evans MJ, et al. The effect of house dust mite aeroallergen and air pollutant exposures during infancy. Chest 2003;123(3 Suppl): 434S.
Molitor J, Molitor J, Jerrett M, McConnell R, Gauderman J, Thomas D, et al. Bayesian modeling of air pollution health effects with missing exposure data (submitted, 2005).
Peters JM, Avol E, Gauderman WJ, Linn WS, Navidi W, London SJ, et al. A study of twelve Southern California communities with differing levels and types of air pollution. II. Effects on pulmonary function. American Journal of Respiratory and Critical Care Medicine 1999;159(3):768-775.
Schelegle ES, Miller LA, Gershwin LJ, Fanucchi MV, Van Winkle LS, Gerriets JE, et al. Repeated episodes of ozone inhalation amplifies the effects of allergen sensitization and inhalation on airway immune and structural development in Rhesus monkeys. Toxicology and Applied Pharmacology 2003;191(1): 74-85.
Wu J, Funk T, Lurmann F, Winer A. 2005a. Improving spatial accuracy of roadway networks and geocoded addresses. Transactions in GIS 2005a;9(4):585-601.
Wu J, Lurmann F, Winer A, Lu R, Turco R, Funk T. Development of an individual exposure model for application to the Southern California children ’s health study. Atmospheric Environment 2005b;39:259-273.
Zhu Y, Hinds WC, Kim S, Sioutas C. Concentration and size distribution of ultrafine particles near a major highway. Journal of Air and Waste Management Association 2002;52(9):10321042.
Journal Articles:
No journal articles submitted with this report: View all 7 publications for this subprojectSupplemental Keywords:
RFA, Scientific Discipline, Air, Geographic Area, HUMAN HEALTH, particulate matter, Environmental Chemistry, Health Risk Assessment, Air Pollutants, State, mobile sources, Health Effects, Environmental Monitoring, engine exhaust, ambient aerosol, asthma, motor vehicle emissions, epidemiology, human health effects, quinones, automotive emissions, particulate emissions, automobiles, automotive exhaust, PAH, air pollution, children, human exposure, PM characteristics, California (CA), allergens, indoor air quality, aerosols, atmospheric chemistryRelevant Websites:
http://www.scpcs.ucla.edu Exit
Progress and Final Reports:
Original AbstractMain Center Abstract and Reports:
R827352 UC Berkeley/Stanford Children’s Environment Health Center Subprojects under this Center: (EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
R827352C001 The Chemical Toxicology of Particulate Matter
R827352C002 Pro-inflammatory and the Pro-oxidative Effects of Diesel Exhaust Particulate in Vivo and in Vitro
R827352C003 Measurement of the “Effective” Surface Area of Ultrafine and Accumulation Mode PM (Pilot Project)
R827352C004 Effect of Exposure to Freeways with Heavy Diesel Traffic and Gasoline Traffic on Asthma Mouse Model
R827352C005 Effects of Exposure to Fine and Ultrafine Concentrated Ambient Particles near a Heavily Trafficked Freeway in Geriatric Rats (Pilot Project)
R827352C006 Relationship Between Ultrafine Particle Size Distribution and Distance From Highways
R827352C007 Exposure to Vehicular Pollutants and Respiratory Health
R827352C008 Traffic Density and Human Reproductive Health
R827352C009 The Role of Quinones, Aldehydes, Polycyclic Aromatic Hydrocarbons, and other Atmospheric Transformation Products on Chronic Health Effects in Children
R827352C010 Novel Method for Measurement of Acrolein in Aerosols
R827352C011 Off-Line Sampling of Exhaled Nitric Oxide in Respiratory Health Surveys
R827352C012 Controlled Human Exposure Studies with Concentrated PM
R827352C013 Particle Size Distributions of Polycyclic Aromatic Hydrocarbons in the LAB
R827352C014 Physical and Chemical Characteristics of PM in the LAB (Source Receptor Study)
R827352C015 Exposure Assessment and Airshed Modeling Applications in Support of SCPC and CHS Projects
R827352C016 Particle Dosimetry
R827352C017 Conduct Research and Monitoring That Contributes to a Better Understanding of the Measurement, Sources, Size Distribution, Chemical Composition, Physical State, Spatial and Temporal Variability, and Health Effects of Suspended PM in the Los Angeles Basin (LAB)
The perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.
Project Research Results
- Final Report
- 2003 Progress Report
- 2002 Progress Report
- 2001 Progress Report
- 2000
- 1999
- Original Abstract
7 journal articles for this subproject
Main Center: R827352
150 publications for this center
149 journal articles for this center