Search 110-S4 Table of Contents EHPS Archives Publications Subscribe

Environmental Health Perspectives Supplements Volume 110, Number 4, August 2002 Exposures to Multiple Air Toxics in New York City Patrick L. Kinney,1 Steven N. Chillrud,2 Sonja Ramstrom,3 James Ross,2 and John D. Spengler3 1Mailman School of Public Health and 2Lamont-Doherty Earth Observatory, Columbia University, Palisades, New York, USA; 3Harvard School of Public Health, Boston, Massachusetts, USA Introduction Methods Results Summary and Discussion

Full Article in PDF

Abstract Efforts to assess health risks associated with exposures to multiple urban air toxics have been hampered by the lack of exposure data for people living in urban areas. The TEACH (Toxic Exposure Assessment, a Columbia/Harvard) study was designed to characterize levels of and factors influencing personal exposures to urban air toxics among high school students living in inner-city neighborhoods of New York City and Los Angeles, California. This present article reports methods and data for the New York City phase of TEACH, focusing on the relationships between personal, indoor, and outdoor concentrations in winter and summer among a group of 46 high school students from the A. Philip Randolph Academy, a public high school located in the West Central Harlem section of New York City. Air pollutants monitored included a suite of 17 volatile organic compounds (VOCs) and aldehydes, particulate matter with a mass median aerodynamic diameter 2.5 µm (PM2.5), black carbon, and a suite of 28 particle-associated trace elements. Sequential 48-hr ambient samples also were collected over 8 weeks in each season at an urban fixed site and an upwind, nonurban fixed site. Personal, indoor, and outdoor concentrations of particle elements were generally similar, suggesting that ambient sources may have driven indoor and personal exposures for most elements. More varied relationships among personal, home indoor, and home outdoor concentrations were observed for VOCs and aldehydes. For formaldehyde and acetaldehyde, and several VOCs, indoor concentrations far exceeded outdoor levels and appeared to dominate personal exposures. Strong seasonal differences in indoor to outdoor concentration ratios were observed for these compounds, reflecting the influence of home air exchange rates. For other VOCs, especially those related to motor vehicle exhaust, more consistent indoor, outdoor, and personal concentrations were observed, suggesting that ambient concentrations may have been the driving force for personal exposures to some VOCs. These results demonstrate exposures to a wide range of air toxic pollutants among young people attending school in inner-city New York. Key words: adolescents, air toxics, exposure assessment, hazardous air pollutants, urban. Environ Health Perspect 110(suppl 4):539-546 (2002). http://ehpnet1.niehs.nih.gov/docs/2002/suppl-4/539-546kinney/abstract.html This article is part of the monograph Environmental Air Toxics: Role in Asthma Occurrence? Address correspondence to P.L. Kinney, Columbia School of Public Health, Dept. of Environmental Health Sciences, 60 Have Ave., B-1, New York, NY 10032 USA. Telephone: (212) 305-3663. Fax: (212) 305-4012. E-mail: plk3@columbia.edu This study was supported by contract NUATRC-96-01B from the Mickey Leland National Urban Air Toxics Research Center. Additional support was provided by Columbia's National Institute of Environmental Health Sciences (NIEHS) Center for Environmental Health in Northern Manhattan (ES09089) and the Columbia Center for Children's Environmental Health (NIEHS ES09600 and U.S. EPA R827027). Received 30 November 2001; accepted 16 April 2002.

Introduction

Considerable progress has been achieved in reducing ambient concentrations of certain criteria air pollutants in the United States over the past 30 years, including lead, sulfur dioxide, carbon monoxide, and the coarse particle fraction of particulate matter. Over the same period there have been striking upward trends in asthma morbidity, especially in urban areas. At face value these opposing trends imply little or no relationship between air pollution and asthma. On the other hand considerable evidence exists that acute exacerbations of asthma can be triggered by ambient ozone and fine particles, two pollutants for which less progress has been achieved. Furthermore, little is known about temporal or spatial trends in ambient concentrations of the noncriteria "air toxic" pollutants. Air toxics include particulate metals, diesel exhaust, and a range of volatile organic compounds (VOCs), many of which could, at elevated concentrations, have adverse impacts on the respiratory system. Little is known about the health risks posed by ambient concentrations of air toxic pollutants, and even less is known about the potential cumulative impacts of exposures to multiple air toxics.

The potential risks associated with exposures to multiple air toxics may be greatest for residents of central urban neighborhoods, where the pollutant mix often represents the sum of regional, citywide, and local source pollutants. Further, evidence suggests that there is disproportionately high exposure of minority and disadvantaged populations to air pollutants (1-3). In the United States, 60% of Hispanics, and 50% of African Americans, compared with 33% of Caucasians, live in areas failing to meet two or more of the national ambient air quality standards (NAAQS) (3,4). These patterns may be due in part to the concentration of minority and low socioeconomic-status populations in cities, along with the tendency for some air pollution concentrations to be higher in urban, versus suburban and rural, areas.

An additional factor enhancing potential exposure risks related to air toxics in inner cities is the strong association between socioeconomic disadvantage and compromised health status. Of special concern are the high rates of asthma morbidity and mortality among disadvantaged urban residents. To date, little information has been available to characterize the cumulative air toxic exposure burdens of inner-city residents, hampering efforts to assess and mitigate health risks.

The TEACH (Toxic Exposure Assessment, a Columbia/Harvard) study was designed to characterize levels of and factors influencing personal exposures to urban air toxics among high school students living in inner-city neighborhoods of New York City (NYC) and Los Angeles. The study included measurements of a wide range of air toxics, including VOCs, particulate matter with a mass median aerodynamic diameter 2.5 µm (PM_2.5), and associated metals and elements, and aldehydes. In this article, we report methods and data for the NYC phase of TEACH, focusing on the relationships between personal, indoor, and outdoor concentrations.

Methods

Study Design

Exposures to air toxics were assessed in a group of 46 high school students from the A. Philip Randolph Academy, a public high school located in the West Central Harlem section of NYC (Figure 1). Each of two field campaigns (winter and summer, 1999) involved 8 weeks of fixed-site ambient monitoring on the school roof and on a roof at the Lamont Doherty Earth Observatory (LDEO) in Palisades, New York. The school roof site was seven stories high and located on a ridge with one of the highest elevations in Manhattan, thus representing areawide urban concentrations. The LDEO roof site was on a three-story building near the Palisades cliffs overlooking the Hudson River, 13 miles northwest of Manhattan. The predominant winds are from the west, resulting in the LDEO rooftop usually being a representative site of the upwind air masses. However, it should be noted that, especially during the summer, sea breezes may sometimes cause winds to flow up the Hudson River Valley from NYC toward LDEO. Here we refer to these two outdoor monitoring locations as the urban fixed site and the upwind fixed site.

Figure 1. Map of sampling locations for the NYC TEACH study.

The fixed-site monitoring covered three consecutive 48-hr periods each week. Concurrent with the first of these ambient samples each week (Tuesday-Thursday), subject-based monitoring of personal, home indoor, and home outdoor samples were collected for 48 hr. Five subjects were typically monitored simultaneously each week. As shown in Table 1, pollutants monitored at every sampling event included a suite of 15 VOCs, PM_2.5, black carbon, a suite of 28 particle trace elements, and two aldehydes. Air exchange rate (AER) was monitored in each home during the air pollution measurement period.

Study Communities

The study school was located at 135^th Street and Convent Avenue in Harlem, a low-income neighborhood of mainly African-American and Hispanic (Dominican) residents. Students attending the school lived primarily in northern Manhattan and the South Bronx, with additional students coming from the boroughs of Queens and Brooklyn. From an environmental perspective, Harlem is at the center of the metropolitan New York region that in recent years has been out of compliance with the NAAQS for particulate matter with a mass median aerodynamic diameter 10 µm (PM₁₀). Ambient air toxic levels in northern Manhattan and the South Bronx result from regionwide emissions as well as from local sources such as diesel bus depots, waste incinerators, industrial operations, and the network of commuter highways, commercial truck routes, and bus routes surrounding and interlacing these communities.

Subject Recruitment

Initial contact with potential study subjects took place in school with the assistance of science teachers. Study staff visited classrooms to describe the goals and methods of the study and to distribute informational brochures and consent forms. At this stage, the school teachers also distributed and collected a brief student survey form on demographics, parental education, student commuting patterns, and personal and passive smoking exposures for all students. To be eligible, students needed to be nonsmokers from nonsmoking families and to be available for sampling in both winter and summer. Students interested in participating in the monitoring study were instructed to have the consent forms signed by a parent/guardian at home and then to return the signed consent forms to the teacher. Consent forms and student surveys were collected in batches and forwarded to study staff, who identified nonsmoking households and then contacted students by telephone to invite their participation. The protocol was approved by the Columbia Health Sciences Institutional Review Board and the Harvard Human Subject Committee.

Questionnaires

A detailed home characteristics questionnaire was administered to the subject or the parent/guardian in the home at the time of the initial sampling setup. This questionnaire included information on home heating and cooking methods and habits, recent renovation work or hobbies that might result in VOC emissions, and other factors. Time-activity logs were completed by subjects during the 48-hr personal sampling period. In addition, at the completion of personal sampling, activity recall questionnaires were completed by the subjects.

Air Sampling and Laboratory Analyses

Personal sampling was carried out using a portable, battery-operated pump (model 400S, BGI Inc., Waltham, MA, USA) housed in a customized daypack that the subjects carried over the shoulder. The pump flow was split three ways to collect one PM_2.5 filter at 4 L/min, one VOC thermal desorption tube at 1.8 standard cubic centimeters per minute (SCCM), and one C₁₈ aldehyde sampler at approximately 100 SCCM. Columbia black boxes (redesigned and rebuilt Harvard black boxes) containing three 7-L/min pumps (Medo, Inc., Hanover Park, IL, USA) were used to collect samples inside and outside of each subject's home and at the urban and upwind fixed sites. Two black boxes were used in each home, with one designated for indoor and the other for outdoor sampling. Duplicate indoor and outdoor PM_2.5 samples were collected onto Teflon filters, with one filter designated to be analyzed for PM_2.5, reflectance, and total metals, and the second filter being archived for future analyses (placed in a petri dish wrapped in aluminum foil and stored in a freezer). The third pump in each black box had its flow split three ways to collect one VOC thermal desorption tube at approximately 1.8 SCCM, one C₁₈ aldehyde sampler at approximately 100 SCCM, and a vent line at approximately 4 L/min. Duplicate VOC and aldehyde samples were collected for at least 10% of sampling events. Flows were checked before and after all sampling events using calibrated flow meters. No sampling was carried out for particles larger than 2.5 µm in diameter, nor did we analyze for biologic particles.

During each sampling week, a total of three 48-hr sample sets were collected at the urban and upwind fixed sites. The first of these samples was taken at the same time as the individual indoor and outdoor samples; two additional consecutive 48-hr samples were taken thereafter. The objectives for obtaining these additional rooftop samples were 2-fold: a) to place the individual personal, indoor, and outdoor sampling in context of the temporal variability, and b) to provide additional outdoor samples for source apportionment purposes.

VOC samples were collected on multisorbent "Air Toxics" tubes (Perkin-Elmer, Inc., Shelton, CT, USA) using the U.S. Environmental Protection Agency Compendium Method TO-17. The tubes were stainless steel and approximately 90 mm (3.5 inches) long and 6.35 mm (0.25 inches) in diameter containing 35 mm of Carbopack B (a medium-strength hydrophobic sorbent) and 10 mm of Carboxen 1000 (a strong sorbent, slightly hydrophilic). The mixture of sorbent strengths allowed for collection of VOCs from n-C₃ to n-C₁₂. The low sampling rate prevented breakthrough and loss of collected VOCs. A diffusion barrier was developed to prevent sampling by diffusion during the lag time before and after active sampling in the home-based and fixed-site monitors, where sampling was activated and deactivated by automatic timers. The diffusion barrier consisted of a small-inner-diameter (0.02 mm) stainless steel tube 200 mm in length. A similar diffusion barrier was tested and used in a recent European study (5). Analysis of VOC tubes was carried out using a Perkin-Elmer automatic thermal desorber (model 400) connected to a Hewlett Packard 5890II gas chromatograph /5971 mass spectrometer with quanitification software (Hewlett Packard, Palo Alto, CA, USA). Sample tubes were placed on a spiking device (tubing connected to an ultra-high-purity nitrogen tank with fitting for the tube) with carrier flow of 75 mL/min. Vapor-phase internal standard was injected into the device and the tube was kept in place 5 min. The vapor-phase internal standard is made from liquid standards in solution (usually methanol) of a known concentration that are injected as a known volume into a 2-L static dilution bottle. A volume of vapor is drawn up with a gas-tight syringe and injected into the injector-port/spiking device with flow onto the sample tube. Drawing different volumes yields different masses of analytes on tube and thus different levels of calibration.

Teflon filter samples were collected in plastic cassettes attached downstream from a cyclone with a 2.5 µm aerodynamic-diameter cut point (model KTL, BGI, Inc.) when operated at 4 L/min ± 10%. Flows were checked before and after sampling. Filters were weighed pre- and postsampling on a microbalance at the Harvard School of Public Health Laboratory after being conditioned in a temperature-humidity-controlled environment for at least 24 hr (by cracking the petri dish cover) and statically discharged via a polonium source.

After PM_2.5 analyses were complete, filters were returned to Columbia, where they were analyzed inside a class-100 flow bench for reflectance using a smoke stain reflectometer (model 43D; Diffusion Systems Ltd., London, UK). Prior work (6,7) has demonstrated that reflectance can be a good proxy for elemental carbon concentrations in outdoor filters. Measurements of five separate locations are made on each filter. However, because in this method the reflectometer head may touch the active filter area (potentially resulting in metals contamination), we designed and built a new filter holder that touches only the outer plastic ring, holding the filter in a fixed flat geometry. This modification enables reflectance measurements to be made without a significant risk of contaminating the filter for later multielement analysis. However, the reflectance measurement is sensitive to the distance between the reflectometer head and the filter (which is an additional 2.5 mm with our filter holder); consequently, to help distinguish our measurements from those of others, we will report our reflectance measurements as "modified" absorption coefficients (Abs*). Reflectance measurements are expressed as the absorption coefficient in reciprocal meters times 100,000.

After reflectance measurements, filters were prepared for multielement analysis by magnetic-sector high-resolution inductively coupled plasma mass spectrometry (HR-ICP-MS). Filters (with supporting ring removed) were microwave extracted in two steps inside 7-mL Teflon vials, which were placed inside microwave vessels containing 10 mL 65% HNO₃. For winter samples, reagents in step 1 were 60 µL water and 200 µL concentrated HNO₃; in step 2 reagents were 100 µL HNO₃ and 40 µL hydrofluoric acid. After extraction, the mass of remaining digest solution was calculated gravimetrically and brought up to 5 mL to approximately 5% acid strength. The filter was removed, transferred to a clean vial, and redigested in the same manner. Both the first and second digests were analyzed. Reagents for summer filters were 20 µL ethanol, 60 µL water, and 225 µL concentrated HNO₃ in step 1, and 10 µL ethanol, 40 µL hydrofluoric acid, and 100 µL HNO₃ in step 2. The addition of ethanol eliminated the need for redigestion of the filter, so it was not done for summer samples. This streamlining of the analytical procedure had no significant impact on the measurements. Recoveries of standards and duplicate precisions were very good in both seasons.

Diluted digests were analyzed by HR-ICP-MS. Winter New York samples were run on the Element (Finnigan-Mat, Breman, Germany) at Rutgers University; summer New York digests were run on the newly purchased Axiom (VG-Elemental, Franklin, MA, USA) at LDEO. Data were collected for all isotopes of interest at the appropriate resolving power (RP) to avoid isobaric interferences. Be, Ag, Cd, Sn, Sb, Cs, La, Pt, Tl, and Pb, for which interferences are not a problem, were run at RP 400; Na, Mg, Al, S, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, and Zn, at RP 3,000-4,300; and K, As, and Se, at RP 9,300. Indium was added to all samples, blanks, and standards as an internal drift corrector and run in all resolving powers. Quantification is done by external and internal standardization.

HR-ICP-MS data were drift corrected with indium, quantified, converted to a mass, and corrected for blanks. Samples that were below the limit of quantification (LOQ) based on daily procedural blanks were flagged. For New York winter samples, analyte mass of the first and second digests was combined. All sulfur mass concentrations are reported as sulfate concentrations, which is assumed to be the predominant sulfur species on the PM_2.5 filters.

Aldehydes were sampled by the method of Fung and Grosjean (8), with air pumped through a cartridge packed with C₁₈ coated with acidified 2,4-dinitrophenylhydrazine (DNPH). Samplers were sealed and refrigerated before and after use and during shipping. Analyses were performed at the Harvard School of Public Health. The DNPH derivatives (hydrazones) were eluted with acetonitrile and then analyzed using high-pressure liquid chromatography (model 1100, Hewlett Packard) with an ultraviolet detector (360 nm). External standards were used to determine the concentration in the samples based on the peak area. Laboratory and field blanks were used for quality control purposes. The concentration of aldehydes found in blank cartridges was subtracted from the sample concentration. The variability of the blank levels was used to determine the limit of detection.

AERs were measured using the perfluorocarbon technique, which is based on diffusional sources (continually release of tracer gas) and diffusional samplers [capillary absorption tubes (CATs)]. The sources were placed in the subject's home 24-72 hr before placement of CATs to allow equilibrium to develop between source release rate and the AER of the home. Two or three CATs were placed per home, typically in the main living area and in the subject's bedroom.

Data Analysis

To describe overall air toxic exposures in the NYC study, we computed means and standard deviations of concentrations for each pollutant measured in personal, home outdoor, and home indoor samples. Results are presented separately for winter and summer. To identify and distinguish pollutants for which exposures were dominated by outdoor concentrations as opposed to indoor concentrations, we plotted median indoor/outdoor (I/O) ratios. To examine spatial and temporal patterns in traffic-related VOC concentrations, home outdoor, urban fixed-site, and upwind fixed-site methyl tert-butyl ether (MTBE) concentrations were plotted for the winter monitoring season.

Results

The full list of 47 air pollutants we quantified is shown in Table 1. These included PM_2.5 and 29 particulate matter elemental components, 15 VOCs, and 2 aldehydes. Twenty-six of these pollutants were listed among 189 hazardous air pollutants in the 1990 Clean Air Act amendments.

A total of 46 students who self-reported not smoking and not being exposed to environmental tobacco smoke were enrolled in the NYC TEACH project, including 33 who completed both winter and summer phases. Five students in winter and eight students in summer participated in only one season. Subjects ranged from 14 to 19 years of age, with 31 (67%) female and 15 (33%) male (Table 2). The racial distribution was 43% African American, 50% Hispanic, and the remaining 7% either Asian or not reported. Most students lived in apartment buildings (76%). These and other characteristics were similar to a larger group of 611 students surveyed at the school (Table 2).

Figure 1 shows a map of the study area, with symbols indicating the locations of the school and subject homes. The majority of the students lived in the upper Manhattan (63% in the winter and 46% in the summer) and Bronx (24% in winter and 27% in summer) boroughs of NYC, with only a few in Brooklyn (8% for winter and 10% for summer) and Queens (5% for winter and 17% for summer).

Table 3 displays means and standard deviations of PM_2.5 and associated component concentrations measured on personal, home indoor, and home outdoor filter samples in winter and summer. Also listed are the limits of detection for each analyte. Most particle components, even those having the lowest concentrations, were detected at levels at least one order of magnitude above the limit of detection, reflecting the high sensitivity of HR-ICP-MS for elemental analyses of air samples. Most concentrations fell in the range from <1 to 100 ng/m³, with PM_2.5, sulfate, and iron extending into the 1-20 µg/m³ range. Many elements had very skewed distributions, with standard deviations exceeding the means. For most elements, indoor, outdoor, and personal concentrations were rather consistent, suggesting that ambient air may have been the driving force for both indoor and personal exposures for most elements. However, there were several elements, including iron and manganese in both seasons, aluminum in winter only, and antimony, chromium, and lead in summer only, for which personal exposures exceeded both indoor and outdoor concentrations by at least a factor of two. For iron, the increase was a factor of five or six. This suggests the existence of one or more personal activities or microenvironments outside the home in which high exposures to certain metals occur.

VOC and aldehyde concentrations are similarly displayed in Table 4 for winter and summer. For VOCs and aldehydes, mean concentrations ranged about two orders of magnitude, with many means falling in the range from less than 1 to about 20 µg/m³. For a few elements, detection limits were less than a factor of two below the lowest mean concentration (i.e., outdoors). This was the case for benzene in both seasons, chloroform in winter, and methylene chloride in summer. Indoor and personal concentrations were frequently higher than those outdoors (i.e., both indoor and personal two times higher than outdoors), especially in winter. This indicates that for many VOCs, indoor concentrations may dominate personal exposures in this population. On the other hand, somewhat more consistent indoor, outdoor, and personal concentrations were observed for the group of VOCs often associated with motor vehicle emissions, including benzene, toluene, ethylbenzene, xylenes, and especially MTBE. For these compounds, ambient air may be an important driver of personal concentrations in this community.

To further explore the relationships between indoor and outdoor concentrations for the full suite of air toxics measured, Figures 2 and 3 display box plots of I/O ratios for particle-associated elements and VOCs/aldehydes, respectively. From bottom to top, the box plots present the 5^th, 25^th, 50^th (i.e., median), 75^th, and 95^th percentiles of each distribution. Winter and summer I/O ratios are plotted together in each figure. Median I/O ratios for most particle-associated elements were close to or less than 1.0 in both winter and summer, consistent with the notion

that ambient levels were the driving force for indoor concentrations for most elements. However, there was a subset of elements that exhibited median I/O ratios greater than 1. Particle components with elevated I/O ratios included zinc, titanium, cadmium, silver, tin, PM_2.5, and potassium in winter and arsenic, chromium, silver, and tin in summer (note that arsenic and chromium were analyzed only in summer). For elements with higher outdoor than indoor concentrations, I/O ratios were slightly lower in winter than in summer, consistent with decreased winter AERs and resulting diminished penetration of outdoor particles. There was a weak tendency for this relationship to reverse for those elements with evidence of indoor sources (i.e., I/O ratios >1), where I/O ratios were often slightly higher in winter than in summer. This is consistent with increased trapping of indoor-generated pollutants in winter when home AER was lower.

I/O ratios for VOCs and aldehydes were generally much higher than those seen for the elements, with nearly half the compounds exhibiting median I/O ratios of 2 or greater (Figure 3). There was a tendency for I/O ratios to be lower (i.e., closer to 1) in summer than in winter, reflecting enhanced AER and clearance of indoor-generated pollutants. Indoor concentrations far exceeded outdoor levels (I/O ratio > 4) for chloroform in both seasons and for the two aldehydes and 1,4-dichlorobenzene in winter. These findings are consistent with more significant and widespread indoor sources for these compounds. As suggested by the results in Table 4, there also were several VOCs for which I/O ratios clustered close to 1, including the traffic-related compounds such as MTBE.

As an example of how the data set can be used to explore the relative magnitudes of temporal and spatial variability for traffic-related VOCs, Figure 4 displays a plot of MTBE concentrations measured at all of the outdoor sites throughout the winter monitoring season, including home outdoors (three to five homes simultaneously each week), the urban fixed site (on the school roof), and the upwind fixed site. Temporal cycles with roughly weekly periods were evident in both the upwind and urban concentrations, reflecting synoptic weather patterns influencing the larger region. A substantial "urban influence" was observed, with urban fixed-site concentrations ranging from 2 to 13 µg/m³ compared with a range of 0-5 µg/m³ at the upwind site. This difference was statistically significant at the 0.05 level using a two-sided paired t-test. A further increase in concentrations above those observed at the urban fixed site was seen at the home outdoor locations, perhaps reflecting greater influences of local traffic at sites closer to the ground (recall that the urban fixed site was on the roof of the seven-story school). Considerable interhome spatial variability also was evident, suggestive of differential impacts of local sources. This inference was supported by the monotonic relationship observed between home outdoor MTBE concentrations and self-reported truck and bus traffic near the home (data not shown).

Figure 4. Time series plot of winter outdoor MTBE concentrations measured at individual homes and at the urban and upwind fixed sites.

Summary and Discussion

We presented methods and preliminary results of urban air toxic exposures of Harlem high school students measured in the NYC TEACH study. Median concentrations for personal, home outdoor, and home indoor samples were displayed to examine general ranges of concentrations, relationships among concentrations measured in the three locations, and analytical detection limits. Median I/O ratios were examined for all pollutants to determine the relative importance of indoor and outdoor sources. Finally, to explore spatial patterns in traffic-related VOCs, ambient concentrations of MTBE measured at the subject's homes, on the school roof, and at the upwind site were examined.

In general, personal, indoor, and outdoor median concentrations of most particle-associated elements were similar, suggesting that ambient sources may have driven indoor and personal exposures for most elements measured. Little evidence was seen for major indoor sources of particle-associated elements in this subject population, which excluded homes with smokers. Evidence for indoor and personal sources of particle-associated elements did exist for a small number of elements, however. Seasonal differences in home AERs (lower in winter than in summer) appeared to influence I/O ratios.

In contrast, a more varied pattern of relationships among personal, home indoor, and home outdoor concentrations was observed for VOCs and aldehydes. For formaldehyde and acetaldehyde, and several VOCs, indoor concentrations far exceeded outdoor levels and appeared to dominate personal exposures. Strong seasonal differences in I/O ratios were observed for these compounds, reflecting the influence of home AERs. For other VOCs, more consistent indoor, outdoor, and personal concentrations were observed, suggesting that ambient concentrations may have been the driving force for personal exposures to VOCs in some cases. This was especially evident for the group of VOCs associated with motor vehicle emissions, including benzene, toluene, ethylbenzene, xylenes, and MTBE. Thus, there were two classes of VOCs apparent in this study, those related mainly to indoor sources and those related mainly to outdoor sources. Spatial and temporal patterns of winter MTBE concentrations demonstrated a strong urban influence on exposure levels, with spatial variability across homes that was consistent with differential traffic impacts.

Active or passive smoking can have a large impact on indoor and personal concentrations. For this reason, we excluded smokers and smoking families from the study based on self-report. We chose not to burden the subjects with urine sampling and thus were unable to use cotinine to verify nonexposure status. If smoking had occurred by subjects during the 48-hr sampling period, it is likely that personal PM_2.5 concentrations would have exceeded the indoor values. This was not observed (Table 2), which provides some measure of confidence that the self-reporting was valid.

These results demonstrate exposures to a wide range of air toxic pollutants among young people attending school in inner-city New York. Although exposures to some air toxics are clearly driven by indoor sources, exposures to many other pollutants appear to relate more to ambient sources. For several VOCs, urban motor vehicle emissions appear to play an important role in exposures.

The implications of multiple air toxic exposures to the health of young people in NYC are not known. However, given the multiple social, economic, and health stresses faced by many urban residents, the possible health impacts of these cumulative exposures should be a high priority for future research.

References and Notes

1. Olden K, Poje J. Environmental justice and environmental health. Bull Soc Occup Environ Health 4:3-4 (1995).

2. Heritage J. Environmental protection--has it been fair? EPA J 18 (1992).

3. Wernette DR, Nieves LA. Breathing polluted air: minority disproportionately exposed. EPA J 18:16-17 (1995).

4. Metzger R, Delgado JL, Herrell R. Environmental health and Hispanic children. Environ Health Perspect 103(6):25-32 (1995).

5. Jantunen MJ, Hänninen O, Katsouyanni K, Knöppel H, Kuenzil N, Lebret E, Maroni M, Saarela K, Sram R, Zmirou D. Air pollution exposure in European cities: The "EXPOLIS" study. J Expos Assess Environ Epidemiol 8(4):495-518 (1998).

6. Edwards JD, Ogren JA, Weiss RE, Charlson RJ. Particulate air pollutants: a comparison of British "smoke" with optical absorption coefficient and elemental carbon concentration. Atmos Environ 17:2337-2341 (1983).

7. Kinney PL, Aggarwal M, Northridge ME, Janssen NAH, Shepard P. Sidewalk exposures to PM_2.5 and diesel exhaust particles in Harlem. Environ Health Perspect 108:213-218 (2000).

8. Fung K, Grosjean D. Determination of nanogram amounts of carbonyls as 2,4-dinitrophenylhydrazine by high performance liquid chromatography. Aerosol Sci Technol 53:168-171 (1981).

Last Updated: August 5, 2002