Association of [FEV.sub.1] in asthmatic children with personal and microenvironmental exposure to airborne particulate matter.

Exposure to particulate matter (PM) air pollution has been shown to exacerbate children's asthma, but the exposure sources and temporal characteristics are still under study. Children's exposure to PM is likely to involve both combustion-related ambient PM and PM related to a child's activity in various indoor and outdoor microenvironments. Among 19 children with asthma, 9-17 years of age, we examined the relationship of temporal changes in percent predicted forced expiratory volume in 1 sec ([FEV.sub.1]) to personal continuous PM exposure and to 24-hr average gravimetric PM mass measured at home and central sites. Subjects were followed for 2 weeks during either the fall of 1999 or the spring of 2000, in a southern California region affected by transported air pollution. [FEV.sub.1] was measured by subjects in the morning, afternoon, and evening. Exposure measurements included continuous PM using a passive nephelometer carried by subjects; indoor, outdoor home, and central-site 24-hr gravimetric [PM.sub.2.5] (PM of aerodynamic diameter < 2.5 [micro]m) and [PM.sub.10]; and central-site hourly [PM.sub.10], nitrogen dioxide, and ozone. Data were analyzed with linear mixed models controlling for within-subject autocorrelation, [FEV.sub.1] maneuver time, and exposure period. We found inverse associations of [FEV.sub.1] with increasing PM exposure during the 24 hr before the [FEV.sub.1] maneuver and with increasing multiday PM averages. Deficits in percent predicted [FEV.sub.1] (95% confidence interval) for given PM interquartile ranges measured during the preceding 24-hr were as follows: 128 [micro]g/[m.sup.3] 1-hr maximum personal PM, -6.0% (-10.5 to -1.4); 30 [micro]g/[m.sup.3] 24-hr average personal PM, -5.9% (-10.8 to -1.0); 6.7 [micro]g/[m.sup.3] indoor home [PM.sub.2.5], -1.6% (-2.8 to -0.4); 16 [micro]g/[m.sup.3] indoor home [PM.sub.10], -2.1% (-3.7 to -0.4); 7.1 [micro]g/[m.sup.3] outdoor home [PM.sub.2.5], -1.1% (-2.4 to 0.1); and 7.5 [micro]g/[m.sup.3] central-site [PM.sub.2.5], -0.7% (-1.9 to 0.4). Stronger associations were found for multiday moving averages of PM for both personal and stationary-site PM. Stronger associations with personal PM were found in boys allergic to indoor allergens. [FEV.sub.1] was weakly associated with N[O.sub.2] but not with [O.sub.3]. Results suggest mixed respiratory effects of PM in asthmatic children from both ambient background exposures and personal exposures in various microenvironments. Key words: asthma, epidemiology, forced expiratory volume, longitudinal data analysis, nitrogen dioxide, ozone, panel study, particulate air pollution. Environ Health Perspect 112:932-941 (2004). doi: 10.1289/ehp.6815 available via http://dx.doi.org/[Online 4 March 2004]

**********

Most panel studies of the daily relationship between acute asthma in children and exposure to particulate matter (PM) air pollution have relied on ambient data collected at central regional sites. All subjects are usually assigned the same daily exposures in these studies. Exposure misclassification from using central regional PM data is expected to diminish the accuracy of exposure-response estimates, possibly leading to null findings. Despite this expectation, with few exceptions (Roemer et al. 1999, 2000), recent panel studies of asthmatic children are largely consistent in showing positive associations between acute increases in asthma morbidity and ambient PM (Delfino et al. 1998, 2002, 2003; Giden et al. 1997; Just et al. 2002; Koenig et al. 2003; Mortimer et al. 2000, 2002; Ostro et al. 2001; Pekkanen et al. 1997; Peters et al. 1997a, 1997b; Romieu et al. 1996; Segala et al. 1998; Slaughter et al. 2003; Thurston et al. 1997; Timonen and Pekkanen 1997; Vedal et al. 1998; Yu et al. 2000).

To improve the accuracy of the estimated associations, measurements of personal exposure to PM and adjustments to ambient exposure using time-activity data have been proposed (National Research Council 1998). In the case of children, high levels of physical activity are expected to generate higher levels of particle exposure in a variety of microenvironments. Other activities may bring the child close to an undiluted PM source such as a school bus. These phenomena have been referred to as the "personal dust doud," which was originally described by Ozkaynak et al. (1996) to account for the difference between total personal exposure as measured by a personal monitor and the estimated time-weighted exposure in indoor and outdoor microenvironments. Studies performed later further clarified the major sources and components of the personal dust cloud (e.g., Liu et al. 2003; McBride et al. 1999) and have shown that the personal dust cloud is a combined result of particles generated from personal activities (e.g., cooking or dusting) and exposures to local sources (e.g., next to traffic exhaust on the street) that are not captured by the stationary indoor and outdoor monitors (Liu et al. 2003). Children's personal cloud PM is significantly higher than adults', resulting in a low prediction power with the traditional microenvironmental model that incorporates time-place-activity data and area monitor measurements (Liu et al. 2003). In addition, short-term exposures lasting minutes to hours may be relevant to respiratory responses and may not be fully captured by time-integrated PM measurements, as is done with 24-hr gravimetric filters.

In this study we evaluated the relationship of repeated measurements of forced expiratory volume in 1 sec ([FEV.sub.1]) in asthmatic children to fine particle exposures using measurements of hourly personal PM and stationary-site 24-hr average PM of aerodynamic diameter < 2.5 [micro]m ([PM.sub.2.5]; indoor and outdoor home, and central regional). Stationary-site 24-hr average [PM.sub.10] (PM of aerodynamic diameter < 10 [micro]m) is also evaluated. To assess the relative importance of peaks and averaging times, we examined relationships of [FEV.sub.1] to hourly maxima and 24-hr mean personal [PM.sub.2.5] exposures for the period preceding the expiratory maneuver, including current and past days.

Materials and Methods

Design. This is a panel study, which involves repeated measurements of outcomes and exposures in individuals. Asthmatic children were monitored for personal PM exposures and health outcomes for 2 weeks (three subjects at a time), during September-October 1999 or April-June 2000 in Alpine, California. There were two 14-day runs in 1999 and five runs in 2000. Measurements of PM were also made at indoor and outdoor home sites of each subject in addition to central regional outdoor measurements. We conducted daily home visits for data collection and instrument preparation and calibration. Subjects also electronically recorded health outcome data daily for approximately 8 weeks. The 2-week personal exposure study was nested in the 8 weeks and is the focus of this article.

Population. The study was conducted in Alpine, a southern California community located inland from San Diego. The community receives long-range transported air pollutants from Los Angeles and San Diego, and it has experienced high ozone episodes. Wind direction is predominantly from the urban coastal regions to Alpine. People live near or above the air inversion layer base (~1,200 ft). Higibility for participation in the study included physician-diagnosed asthma with at least a 1-year history; a historical confirmation of asthma exacerbations on at least one separate occasion each week during March-October that required the use of an as-needed, oral or inhaled, bronchodilator; a home address in Alpine; no active smoking or passive exposure to tobacco smoke at home; and age 9-19 years. Subjects were recruited with the assistance of the Alpine school district nurse. The institutional review boards of the University of California, Irvine, and San Diego State University approved the study protocol. Informed written consent was obtained from all subjects and from one of their legal guardians.

Twenty-four children were recruited. Three dropped out, and another was not compliant with self-administered spirometry. Another subject did not participate in the personal exposure assessment. Nineteen subjects completed the 2-week exposure assessment and were compliant with spirometry (Table 1). One subject participated in two 14-day runs. The other 18 subjects each participated in one 14-day run, except one who started 4 days late. Only five used anti-inflammatory medications (four on inhaled corticosteroids, one on the antileukotriene zafirlukast). Over all the [FEV.sub.1] measurements in the study, 16 subjects had < 80% predicted [FEV.sub.1], three had 84-87%, and one had 92% using normal lung function equations from the Third National Health and Nutrition Examination Survey (NHANES III; Hankinson et al. 1999).

Health outcomes and daily follow-up procedures. The health outcome described in this article is [FEV.sub.1], which represents the degree of airway obstruction as represented by the volume of air that can be forcibly exhaled in the first second of an expiratory maneuver starting from a maximal inhalation (forced vital capacity). Airway obstruction is a key phenotype of asthma. [FEV.sub.1] is lower than expected when flow rate decreases due to airway obstruction. Analyses here focus on the percentage of predicted normal [FEV.sub.1] (Hankinson et al. 1999) for a given height, age, sex, and race/ ethnicity. This standardizes measurements between subjects and gives clinically meaningful overall estimates of association for the study population.

We used the Vitalograph 2110 handheld electronic expiratory flow meter (Vitalograph, Inc., Lenexa, KS) for subjects to self-administer measurements of [FEV.sub.1]. It measures flow by sensing pressure across a resistive mesh. Before beginning the 2-week exposure assessments, calibration of Vitalographs was checked with a 3-L syringe, and all subjects used the expiratory flow meter for at least 1 week to assure that maneuvers were performed properly. Subjects were trained by staff to perform expiratory maneuvers and were given a subject manual reviewing instructions. Nose clips were not used. The Vitalograph alarm prompts the subject at home to begin the maneuver (or the subject can self-initiate) and to repeat maneuvers when shown an icon of a face blowing air during each maneuver session. The unit stops prompting if two peak expiratory flow (PEF) readings are within 10%, up to a maximum of five maneuvers. The percent reliability between the highest and second highest PEF is logged in memory; the highest flow is logged as "good" for time to peak flow between 40 msec and 290 msec, and "bad" otherwise to identify maneuvers performed with either the spitting technique or insufficient initial effort. The [FEV.sub.1] stored in memory is the highest volume achieved. All maneuvers were to be performed before [beta]-agonist inhaler use. The morning maneuver was to be performed after being awake and alert soon after arising; the afternoon maneuver was to be performed around 1700-1800 hr, and the evening was to be performed around 2100 hr or near bedtime.

The expected number of [FEV.sub.1] measurements was 840. The observed number was 710 maneuvers, but some subjects did more than one session during one of the three daily periods (37 measurements). In these cases, the highest [FEV.sub.1] from the different sessions was chosen, leaving 673 maneuver sessions. Therefore, 167 (20% of expected) were missing because of noncompliance during the expected daily maneuver period. Although PEF was retained, an additional 67 [FEV.sub.1] measurements were missing because of inadequate expiratory maneuvers where forced expiratory time was < 1 sec. In addition, eight [FEV.sub.1] measurements were excluded because time to peak flow was not between 40 msec and 290 msec, and 50 others were excluded where the difference between the highest and second highest PEF was > 10%, which is evidence of low reliability. There were 548 [FEV.sub.1] maneuvers remaining for analysis.

Diary data on symptoms, as-needed inhaler use (short acting [[beta].sub.2]-agonists), preventive medication use, respiratory infections, and time-activity were entered by subjects into small personal digital assistants (PDAs) every 2 waking hours. Subjects were prompted with 14 one-minute alarms and program lockout at the end of 15 min to ensure short-term recall at the appropriate times. Subjects gave touch-pad responses to user-friendly interfaces: informational screens and ordinal and categorical list screens. We expect that this PDA diary enhanced data validity over paper diaries. Staff visited the subject's home daily to examine the quality of outcome data after they were downloaded from each electronic device to a laptop. Subjects were compliant with diary completion for 84.3% of data relevant to [FEV.sub.1] observations.

Analyses of time-activity diary data from the PDA and factors predicting personal PM exposures will be presented elsewhere. Although asthma symptom data and as-needed inhaler use were collected, the frequency of responses during the 2-week exposure assessment periods was low (only four subjects had any symptoms that were at least bothersome, and only nine subjects used as-needed inhalers; Table 1). Therefore, analyses of symptoms and inhaler use in relation to exposures are not presented in this article. Inhaler use in the 2-3 hr before [FEV.sub.1] measurements and respiratory infections were tested as covariates in regression models.

Environmental variables. To measure personal PM exposures, we used the Personal dataRAM (pDR; MIE Thermo Electron Corp. Inc., Franklin, MA), which is a nephelometer that measures light scatter (source wavelength, 880 nm) from PM that relates to mass concentration. The pDR is relatively small (15.2 x 9.1 x 6.4 cm) and lightweight (0.5 kg) and was easily carried by subjects at waist level using a fanny pack, shoulder harness, or vest. It was operated in passive sampling mode with the air exchange region open to air and without a size selective inlet. The pDR responds mainly to PM in the 0.1-10-[micro]m range, with the highest response in the fine PM range. The measurements approximate [PM.sub.2.5] with an [R.sup.2] between 0.77 and 0.84 when compared with collocated stationary Harvard impactor (HI) measurements and an [R.sup.2] of 0.44-0.60 when compared with collocated Harvard personal environmental monitor measurements (Howard-Reed et al. 2000; Liu et al. 2003). The pDR reads in units of [micro]g/[m.sup.3] as calibrated by the manufacturer using fine International Organization for Standardization test dust (specific gravity, 2.6 g/[cm.sup.3]; index of refraction, 1.5; mass median diameter, 2-3 [micro]m). Readings in micrograms per cubic meter can be converted to light scattering [[micro]g/[m.sup.3] = 1.027 x [b.sub.sp] (light scattering coefficient in 1 x [10.sup.-3]/m)].

The pDRs were set to report concentration averaged over 1-min intervals. Data were downloaded daily and a new 24-hr logging period was set. All pDRs had attached small relative humidity (RH), temperature, motion, and light intensity loggers recording at 1-min intervals (Onset Computer Corp, Pocasset, MA). Motion sensors (Onset's motor on/off), intended to monitor compliance with wearing the pDR, had limited memory and often did not cover the 24-hr periods of monitoring. Subjects were not informed of the purpose of the data loggers. All electronic exposure data were time synchronized and data processed to quarter-hour means to match PDA diary data. Electronic exposure data were downloaded to laptops every 24 hr during home visits by staff.

A more detailed description of our quality assurance procedure for the pDR data is given elsewhere (Wu et al. 2004). Briefly, we applied a correction equation that models the effect of RH on [b.sub.sp] to adjust for the influence of RH > 60% on the pDR data (Richards et al. 1999). Personal RH rarely exceeded 70% (1.3%). Observations with personal RH > 95%, where the equation is not reliable, were deleted, but were few in number (0.18%). In the fall of 1999, the pDRs were zeroed daily with a zeroing bag (Z-Pouch; MIE Thermo Electron Corp. Inc.) attached to a small HEPA filter as supplied by the manufacturer. In the spring of 2000, pDRs were zeroed with particle-free air through an active sampling system attached to a HEPA filter before each 2-week run. Negative and positive drift was still detected, and, where possible, correction factors were applied using early morning data from collocated indoor pDRs, which did not drift, as we had previously reported (Quintana et al. 2000). In 0.5% of the personal pDR data, we also identified instances when subjects had not worn the pDR and left it in another microenvironment using subject reports and/or data from RH, temperature, motion, and light intensity loggers attached to the pDR, which suggested prolonged stationary positioning. These data were dropped from the analysis of personal PM exposures.

Indoor and outdoor home sites were monitored for [PM.sub.2.5] and [PM.sub.10] gravimetric concentrations with His operated at 10 L/min (Air Diagnostics and Engineering, Inc., Naples, ME). The major living area of the house was the indoor sampling site. Outdoor impactor samplers were placed under a rain cover immediately outside the house. Samplers were placed at 1 m off the ground, at least 1 m from walls, and away from pollutant sources, including heavy foot traffic. The HI [PM.sub.2.5] and [PM.sub.10] samples were collected onto Teflon filters (37 mm, 2.0 [micro]m Teflon membrane; Gelman Labs, Ann Arbor, MI) for 24 hr from around 1700-2000 hr to 1700-2000 hr the next day, and weighed on a Cahn microbalance (model 30; Cahn Instruments, Madison, WI). Filters were conditioned at 23 [+ or -] 3[degrees]C and 33 [+ or -] 5% RH for at least 24 hr in an environmental chamber, and a Cahn polonium-210 source was used to eliminate interference by electrostatic charges on the filter. Forty-nine filters (3.8%) were removed from analysis because of ending pump flow rates that were < 9 or > 11 L/min, or errors including pump failure, damaged or dropped filters, and incorrect log sheet data. Mass measurements (in micrometers per cubic meter) were corrected by site of collection with one indoor alternated with one outdoor field blank per day.

The gaseous pollutants [O.sub.3] and nitrogen dioxide were measured at a stationary outdoor monitoring station located centrally in Alpine and operated by the San Diego Air Pollution Control District. Hourly data were available for [O.sub.3] measured with ultraviolet photometry, N[O.sub.2] measured with chemiluminescence, temperature, and RH. Central-site [PM.sub.10] concentrations were also measured with a tapered-element oscillating microbalance (TEOM; Rupprecht and Patashnick Co., Inc., Albany, NY). The TEOM is an inertial instrument that measures particle mass in real time on an exchangeable filter cartridge by monitoring frequency changes of a tapered element. The TEOM sampler inlet was operated at 16.7 L/min, and the inlet air stream was heated to a constant 50[degrees]C to keep water in the vapor phase. Hourly TEOM [PM.sub.10] data were used for comparison with the personal continuous PM exposure data (described above).

Statistical analysis. The pDR PM data were log normally distributed and were therefore log transformed before analysis. We wanted to compare strengths of association for different pollutant metrics regardless of the monitor type (personal vs. fixed site), units of measurement, or concentration range of normally distributed regressors. Ideally, associations for different exposure variables should be compared for their relative impact during the same period and for the same subjects. Therefore, we expressed results as the change in percent predicted [FEV.sub.1] for an interquartile range (IQR) increase in the air pollutant, an approach to standardizing associations that is discussed elsewhere (Lipfert and Wyzga 1999). This is particularly important for comparisons of personal versus stationary-site PM measurements. Sampling is done by technologically different methods that are not entirely comparable in terms of particles measured, their size distribution, or toxicity.

Data were analyzed by a general linear mixed model (with both fixed and random effects). This type of model is particularly suitable for serially correlated data in individuals (Diggle et al. 2002). We used the restricted maximum likelihood method as implemented by the MIXED procedure in SAS, version 8.2 (Littell et al. 1996). Random intercepts were estimated for each individual. The model covariance structure was determined by the Akaike's information criterion and Bayesian information criterion (Littell et al. 1996). The best covariance structure was autoregressive of order 1 [AR(1)], which allows for within-subject autocorrelated error terms.

Pollutant regression models were tested for confounding by day of week, personal temperature and RH, time of [FEV.sub.1] maneuver (morning, afternoon, or evening), season (fall 1999 or spring 2000), 2-week period of exposure assessment (seven three-subject runs), as needed medication use (inhaler puffs in the 2-3-hr period preceding the [FEV.sub.1] maneuver, or 24-hr cumulative use), and presence versus absence of upper or lower respiratory infections. Confounding was considered a [greater than or equal to] 10% change in the regression parameter estimate for the air pollutant. We also tested for interaction of air pollutants with respiratory infections, with anti-inflammatory medication use (a binary variable indicating self-reported daily use vs. no use), with sex, and with allergy to indoor allergens.

We hypothesized that lung function would be inversely associated with particulate air pollution exposure. To test this, we examined the relationship of [FEV.sub.1] to air pollutant concentrations measured in the 24-hr preceding the maneuver (exposure lag 0) and for cumulative exposures up to 4 days before the day of the [FEV.sub.1] maneuver (2-5-day exposure averages). All three maneuvers were entered in the regression models except for models involving current-day (lag 0) 24-hr mean gravimetric measurements of [PM.sub.2.5] and [PM.sub.10], (which were completed by the afternoon or evening [FEV.sub.1] maneuver), and current-day 12-hr day-time pDR PM and TEOM [PM.sub.10]. Only the afternoon and evening maneuvers were regressed on these lag 0 PM variables because the morning maneuvers were done within or before that day's sampling period. For personal pDR PM, the real-time nature of the measurements allowed the assessment of exposures in the period immediately before each of the [FEV.sub.1] maneuvers. This approach makes it possible to assess acute exposure-response relationships, hypothetically including early phase bronchospasm.

Data for exposure variables had to include [greater than or equal to] 75% of the relevant exposure period (e.g., 18 hr out of a 24-hr averaging period). Exposure averaging times for personal PM included maximum 1-hr, 4-hr, and 8-hr during the preceding 24 hr; 2-hr average the preceding 2 hr; 24-hr average the preceding 24 hr; 12-hr average daytime (0800-2000 hr); and 12-hr average nighttime (2000-0800 hr). Lag days were set to the prior 24-hr periods (e.g., lag 1 would be the average pDR PM 24-48 hr before the [FEV.sub.1] maneuver). We tested models for multiday moving averages including lags 0, 1, 2, 3, or 4 days. The longer the PM lag, the fewer the [FEV.sub.1] observations available for analysis, and the more likely the [FEV.sub.1] data in the multiday average model differ from the full set of [FEV.sub.1] data used in models with lag 0 (current day) PM data. In other words, only the last 10 of 14 days of monitoring had [FEV.sub.1] data to regress against a 5-day moving average of PM.

Similar averaging times were used for central-site TEOM [PM.sub.10], which included 1-hr and 8-hr maximums and 24-hr averages. To correlate personal with stationary-site gravimetric measurements at the home and central sites, the personal data were also averaged for the 24-hr sampling period of the His. For central-site N[O.sub.2] and [O.sub.3], both 1-hr and 8-hr maximum concentrations in the preceding 24 hr plus multiday averages were examined.

Results

Descriptive statistics. Subject characteristics are shown in Table 1. One subject was Hispanic and the remaining subjects were white non-Hispanic, reflecting the overall composition of the community. Among asthmatic populations, lower lung function values are expected in the morning, with the maximum around noon (American Thoracic Society 1991). We saw lower values in the morning compared with afternoon for 12 of 19 subjects.

Table 2 shows univariate data for the exposures. Data for personal pDR PM show more than twice the average concentration during the daytime than during nighttime. In comparison, ambient daytime TEOM [PM.sub.10] mass was around 50% higher than nighttime TEOM [PM.sub.10]. The pDR PM variables show increased particle concentration as the averaging time decreases. Peak concentrations were much higher than seen for the ambient hourly TEOM [PM.sub.10], suggesting that personal activity led to higher exposures related to the so-called personal cloud (Liu et al. 2002). The highest 1-hr maximum TEOM [PM.sub.10] was 95 [micro]g/[m.sup.3], whereas the overall geometric mean of the 1-hr maximum pDR PM measurements (approximating [PM.sub.2.5]) was 117 [micro]g/[m.sup.3]. Gravimetric mass measurements of 24-hr average [PM.sub.2.5] at all stationary sites were around one-third the 24-hr average pDR PM mass. Home indoor:outdoor PM ratios by person-day of paired nonmissing observations were 1.2 for [PM.sub.2.5] and 1.3 for [PM.sub.10]. Neither [PM.sub.2.5] nor [PM.sub.10] 24-hr average concentrations ever came close to the U.S. National Ambient Air Quality Standards (65 [micro]g/[m.sup.3] for [PM.sub.2.5] and 150 [micro]g/[m.sup.3] for [PM.sup.10]; U.S. EPA 1990).

Table 3 shows Spearman rank correlations between selected exposures. Ambient [O.sub.3] showed only small to moderate correlations with stationary-site PM and virtually no correlation with personal PM. Ambient N[O.sub.2] was moderately correlated with outdoor home and central-site HI [PM.sub.2.5] and [PM.sub.10], and with indoor [PM.sub.2.5], but weakly correlated with personal pDR PM. Correlations between personal pDR PM and stationary-site PM were generally small. Correlations for the 1-hr maximum pDR PM (not shown) with stationary-site PM were similar to the 8-hr maximum pDR PM, but slightly weaker. Peak personal PM exposures and peak TEOM [PM.sub.10] both occurred during the daytime but were only weakly correlated. TEOM [PM.sub.10] was only moderately correlated with collocated central-site gravimetric [PM.sub.10], possibly because of loss of semivolatile components across the TEOM's 50[degrees]C heating element (Allen et al. 1997) and/or loss of other particle components on the filter with time. Correlations of indoor with outdoor [PM.sub.2.5] were moderately strong.

Regression analysis. Model fit was improved using an indicator variable for the 2-week period of exposure assessment, and this fit was better than an indicator variable for subject. In addition to being a surrogate for groups of subjects and their responses, exposure period also likely represents different outdoor sources, meteorologic conditions, home ventilation conditions, and thus infiltration efficiency (Allen et al. 2003). Although exposure period nominally predicted [FEV.sub.1] (p < 0.1), controlling for period resulted in tighter confidence intervals (CIs) and larger negative slope coefficients for PM, indicating stronger adverse effects of PM on lung function. Maneuver time (morning, afternoon, or evening [FEV.sub.1]) was alone associated with [FEV.sub.1] (p < 0.005), and when added to models with the PM variables, it improved fit through its influence on the intercept but did not confound pollutants. Including maneuver time as a heterogeneity factor for the AR(1) covariance matrix did not improve models further. Other covariates did not confound associations of air pollutants with [FEV.sub.1] and were not themselves associated with [FEV.sub.1], including mean personal temperature the preceding 2 hr (p = 0.85), 12 hr (p = 0.62), and 24 hr (10 = 0.96) before the [FEV.sub.1]; maximum 1-hr personal temperature the preceding 24 hr before the [FEV.sub.1] (p = 0.24); minimum 1-hr personal RH (p = 0.53) and maximum 1-hr personal RH the preceding 24 hr before the [FEV.sub.1] (p = 0.94); season (fall 1999, spring 2000, p = 0.29); as-needed inhaler use in the previous 2-3 hr (p = 0.37); cumulative as-needed inhaler use during that day (p = 0.36); respiratory infection (p = 0.55); and day of week (p = 0.63). The lack of association for respiratory infection and as-needed inhaler use was likely caused by the infrequency of respiratory infections (6% of observations) and lack of inhaler use in more than half the subjects and infrequency of inhaler use in most of the rest (Table 1). Therefore, all models include random subject-specific intercepts and control for 2-week period of exposure assessment, maneuver time, and autocorrelation in the covariance structure.

We found that personal pDR PM exposure was inversely associated with percent predicted [FEV.sub.1], and most pDR exposure variables showed upper 95% confidence limits that did not cross the null value of 0% (Figure 1). Values for 4-hr and 8-hr maximum pDR PM the preceding 24 hr showed similar magnitudes of association, so we present results only for the 8-hr maximum. For all averaging times (1-hr maximum, 8-hr maximum, 24-hr average, 12-hr daytime, and 12-hr nighttime), the most robust associations were for lag 0 and 2-, 4-, or 5-day averages, with associations ranging between -4 and -22% predicted [FEV.sub.1] per IQR increase in personal PM. Generally, associations with [FEV.sub.1] were stronger when more lags were added. The strongest association was for 5-day moving average 12-hr daytime personal PM (-22%; 95% CI, -34 to -11). Associations for 12-hr daytime were stronger than 12-hr nighttime average personal PM. Percent predicted [FEV.sub.1] was not associated with 2-hr average pDR PM measured over the 2 hr preceding the [FEV.sub.1] maneuver (p = 0.5; data not shown).

[FIGURE 1 OMITTED]

We also found that [PM.sub.2.5] gravimetric mass measured at indoor and outdoor home sites and at the central site was inversely associated with percent predicted [FEV.sub.1] (Figure 2), although for some exposure variables the 95% CI included the null value of no effect. There were small differences in the magnitude of association by monitoring site for lag 0 [PM.sub.2.5] gravimetric mass (indoor > outdoor > central site), but for multiple-day averages, associations were similar by site. [PM.sub.10] gravimetric mass was also inversely associated with percent predicted [FEV.sub.1] (Figure 3), although again, for some exposure variables the 95% CI included the null value of no effect. Indoor [PM.sub.10] was more strongly associated with [FEV.sub.1] than was outdoor home or central-site [PM.sub.10]. Associations shown for monitoring sites were notably smaller compared with personal PM (note graph scale difference in Figure 1).

[FIGURES 2-3 OMITTED]

Central-site TEOM [PM.sub.10] was inversely related to percent predicted [FEV.sub.1] at magnitudes similar to that of collocated central-site gravimetric [PM.sub.10], but most 95% CIs included the null value of no effect (Figure 4). There were no differences between hourly maxima and 24-hr averages.

[FIGURE 4 OMITTED]

Central-site 5-day average 8-hr maximum N[O.sub.2] was inversely associated with percent-predicted [FEV.sub.1] (per IQR increase in N[O.sub.2], -1.16%; 95% CI, -2.4 to 0.1), and associations were similar for the 3- and 4-day average and for 1-hr maximum N[O.sub.2]. Central-site [O.sub.3] was not associated with percent predicted [FEV.sub.1] for any averaging times (p [greater than or equal to] 0.4). Between-pollutant confounding was tested with two-pollutant regression models including an individual PM variable with central-site N[O.sub.2]. Interaction was tested first but was nonsignificant. We found that associations for personal PM were minimally affected by N[O.sub.2], in that the above-reported associations with [FEV.sub.1] remained unchanged (for mean 24-hr) or more inversely associated (20-27% change for mean 12-hr daytime). However, N[O.sub.2] was confounded by personal PM with parameter estimates falling near zero.

Product term models for the interaction of pDR PM mass with a binary indicator for whether a subject was taking anti-inflammatory medications (five subjects) versus not taking them (14 subjects) showed no significant interaction (p = 0.9 for 8-hr maximum pDR PM, p = 0.8 for 24-hr average pDR PM). A similar product term for respiratory infection was also nonsignificant (p = 0.5 for 8-hr maximum pDR PM, p = 0.6 for 24-hr average pDR PM). Product term models for the interaction of pDR PM mass with a binary indicator for sex were usually significant, particularly for 2-4-day moving averages. The 14 boys showed more inverse associations than did the five girls, who showed responses not different from zero. For instance, we found an increase of -16% predicted [FEV.sub.1] (95% CI, -26 to -6) per IQR in 4-day average personal PM for boys, but only -1% predicted [FEV.sub.1] (95% CI, -16 to 14) for girls. Consistent interactions were found for stationary-site PM.

To indirectly test this assumption, we performed an analysis of effect modification by the presence versus absence of allergy to common indoor allergens. Allergy was based on positive allergen reactivity as assessed using skin prick tests for cat (Fells domesticus 1) and house dust mites (HDM; Dermatophagoides farinae and D. pteronyssinus) and defined as a skin wheal 3 mm greater than the saline negative control or having a diameter [greater than or equal to] 50% of a histamine dihydrochloride-positive control. None of the girls was allergic to these allergens; therefore, the analysis was conducted only on boys because we found significant differences in associations between boys and girls as reported above. Two boys were not tested, leaving six having positive HDM/cat allergies (Table 1: ID 1, 5, 11-14) and six having negative HDM/cat allergies (Table 1: ID 2, 3, 9, 10, 15, 16). Three were allergic to HDM alone, two to cat alone, and one to both. Figure 5 shows results of product term models for personal PM and atopy to HDM/cat, and Figure 6 shows results of product term models for stationary-site gravimetric PM and atopy to HDM/cat. We found that allergic subjects had stronger responses to personal PM (product term p-value < 0.07 for 2- and 3-day averages). However, responses were generally similar for stationary-site PM. Some associations with gravimetric [PM.sub.2.5] and [PM.sub.10] were slightly and nonsignificantly stronger among nonallergic subjects, except 5-day average indoor [PM.sub.10], which was significantly stronger among allergic subjects.

[FIGURES 5-6 OMITTED]

Discussion

Findings. A key finding of this study is that percent predicted [FEV.sub.1] was inversely associated with personal exposure to fine particles (pDR PM; Figure 1). This is among the first reports of the relationships of personal PM exposure and [FEV.sub.1] in schoolchildren with asthma. The magnitudes of associations were in some cases clinically relevant at -6% predicted [FEV.sub.1] or worse. For instance, airway reversibility (one hallmark of asthma) can be identified in clinical evaluations of asthma using spirometry when there is an observed increase of [greater than or equal to] 12% in [FEV.sub.1] from pre- to post-bronchodilator administration. Given our overall average [FEV.sub.1] of 2.00 L/sec, and an overall percent predicted [FEV.sub.1] of 70%, a 12% increase from an [FEV.sub.1] of 2.00 L/sec would represent an 8.4% increase in percent predicted [FEV.sub.1]. The largest association we observed was -22% predicted [FEV.sub.1] for 5-day moving average daytime personal PM.

We also found inverse associations of [FEV.sub.1] with stationary-site indoor, outdoor and central-site gravimetric [PM.sub.2.5] and [PM.sub.10], and with hourly TEOM [PM.sub.10] (Figures 2-4). Personal PM was more strongly associated with [FEV.sub.1] than was stationary-site PM. Exposure misclassification from using stationary PM data may have diminished the accuracy of exposure-response estimates compared with personal exposures, thus potentially weakening associations with stationary PM. Personal exposures may have been not only quantitatively different (Table 2) but also qualitatively different. Recent studies have shown that personal PM includes not only infiltrated PM from vegetative burning, mobile sources, or secondary sulfate sources but also several classes of crustal materials that are related to personal activities (Hopke et al. 2003; Larson et al., in press; Yakovleva et al. 1999). It is likely that exposures to causally relevant airborne particles occurred at times not represented by the stationary-site monitors. These exposures may be encountered only in certain environments (e.g., at a bus stop) or only during certain activities (e.g., cleaning the house or classroom activities). Nevertheless, the consistency in associations with [FEV.sub.1] for personal PM compared with ambient PM suggests that some part of the association for personal PM exposure was attributable to ambient PM. The stronger associations for personal PM, especially the daytime average, imply that indoor activities, personal cloud, as well as daytime outdoor point sources contributed more to decrements in [FEV.sub.1] than did the average background PM (largely transported PM).

Associations of [FEV.sub.1] with lag 0 indoor (p < 0.01) and outdoor home [PM.sub.2.5] (p < 0.07) were more significant and stronger than for central-site [PM.sub.2.5] (p < 0.22; Figure 2). Similarly, associations for lag 0 indoor (p < 0.02) and outdoor home [PM.sub.10] (p < 0.21) were more significant and stronger than for central-site [PM.sub.10] (p < 0.55; Figure 3). We speculate that this is because the home measurements better represented personal exposures. Correlations between home and central-site PM were moderate to strong (Table 3). All of this suggests that central-site PM measurements could be sufficiently representative of ambient exposures for larger sample sizes in panel studies that are reliant on central-site data.

Results for personal PM versus stationary-site PM monitors can be compared with those of only one recent study, by Koenig et al. (2003), who reported results of another panel study in 19 Seattle schoolchildren with asthma using up to 10 consecutive daily measurements of exhaled nitric oxide (eNO), which is a marker of airway inflammation. They found a positive association between eNO and lag 0 24-hr TEOM measurements of [PM.sub.2.5] at central outdoor sites that was similar in magnitude to associations of eNO with lag 0 24-hr gravimetric indoor home, outdoor home, and personal measurements of [PM.sub.2.5]. Associations were significantly stronger for 10 subjects not using inhaled corticosteroids, consistent with our previous findings for asthma symptoms in children (Delfino et al. 1998, 2002). We may have observed no effect modification by anti-inflammatory medications because only five subjects were on them.

We found that there were significantly stronger associations of [FEV.sub.1] with PM in boys than in girls; but again inferences are somewhat limited because we compared responses in five girls with those in 14 boys. It is conceivable that boys may have been more exposed to outdoor PM and to personal cloud PM from increased physical activity. Time-activity diaries showed that boys spent more time outdoors on average (3 hr 17 min/day vs. 2 hr 32 min/day in girls), and boys were engaged in moderate to strenuous physical activity more often (37 min/day vs. 23 min/day in girls). No consistent sex differences in the relationship of PEF and ambient particulate or gaseous air pollutants were found in a large European panel study (Roemer et al. 1999), but another large U.S. panel study found stronger inverse associations between PEF and ambient [O.sub.3] in boys than in girls (Mortimer et al. 2000). Mixed sex differences in associations were found for one time series study of asthma hospitalization in relation to ambient N[O.sub.2] and sulfur dioxide in Vancouver, Canada (Lin et al. 2004). Stronger cross-sectional associations between lung function and ambient air pollution were found for girls in a general population cohort study in southern California, particularly among girls spending more time outdoors, and similar differences were seen for asthmatic subjects (Peters et al. 1999).

Evidence for PM associated with specific personal activities was shown in our previous study using time-stamped voice recorders and pDRs worn by 10 adult subjects over 1 week (Quintana et al. 2001). In that study, we found that compared with periods with no reported pollution events, pDR PM exposures were significantly higher while the subject was near pets, construction activities, cooking, barbecues, or environmental tobacco smoke or when doing yard work. These daily exposure patterns showing large excursions in PM were consistently found both within subjects and for specific daily activities such as those described above. Modest differences in PM concentrations were seen for routine daily changes in indoor and outdoor microenvironments, such as differences between home and work. Large excursions in PM driven by activities may be particularly important in active children. A recent study by Liu et al. (2003) in Seattle found that children with asthma had higher personal [PM.sub.2.5] exposures than did elderly adults, and microenvironmental models used to predict personal [PM.sub.2.5] were weakest for asthmatic children ([R.sup.2] = 0.09) compared with elderly adults ([R.sup.2] = 0.45-0.62).

We found no notable differences in associations with [FEV.sub.1] by averaging time (1-hr, 8-hr, or 24-hr average) for personal PM (Figure 1). However, the strongest association was for 5-day moving average 12-hr daytime personal PM, but other multiday personal PM averages were comparable between 12-hr nighttime and daytime periods. Contrary to previous findings (Delfino et al. 1998, 2002), there were minimal differences between averaging times for TEOM [PM.sub.10] measured at the outdoor central site.

The present findings for some stationary-site PM variables are far from statistically significant, but most are in the direction expected (inverse). In general, all PM metrics were more robustly associated with [FEV.sub.1] when multiday moving averages were used. However, it is difficult to make a direct comparison between current day and cumulative lag exposures because the first days of an individual's [FEV.sub.1] data from the 14-day session could not be used in models involving the multiday moving averages. Other asthma panel studies have also found more robust associations of asthma morbidity with multiday moving averages of PM compared with single lag days (Delfino et al. 1998, 2002; Gielen et al. 1997; von Klot et al. 2002). This suggests either delayed or cumulative respiratory effects, and/or a smoothing of exposure misclassification for any single day lag.

Associations of [FEV.sub.1] with ambient N[O.sub.2] was isolated to 3-5-day average exposures but was confounded by personal PM, which was itself largely independent of N[O.sub.2]. We previously reported associations between asthma symptoms and lag 0 N[O.sub.2] in another asthma panel study in Alpine, California (Delfino et al. 2002). Despite occasional high levels of central-site [O.sub.3] (8-hr maximum [O.sub.3] 90th percentile, 83.9 ppb), [O.sub.3] was not associated with percent predicted [FEV.sub.1] in the present study. Time series studies investigating the relationship between gaseous ambient air pollutants and asthma hospital admissions or emergency department visits have been inconsistent (for a discussion, see Lin et al. 2004).

Associations for [O.sub.3] in the present study could be lacking because on hot days in this inland semi-arid desert region, when [O.sub.3] concentrations are highest, most subjects were indoors in air-conditioned buildings (r = 0.50 for 8-hr [O.sub.3] and 1-hr maximum outdoor temperature, vs. r = 0.10 for 8-hr [O.sub.3] and 1-hr maximum personal temperature). [O.sub.3] concentrations in southern California homes are considerably lower in air-conditioned buildings (Lee et al. 1999). We previously reported results of an [O.sub.3] exposure assessment study in Alpine that showed only small correlations between 12-hr daytime personal passive measurements of [O.sub.3] (median: spring 1994, 15.5 ppb; fall 1994, 12.7 ppb) and central-site [O.sub.3] (median: spring 1994, 54 ppb; fall 1994, 60 ppb). Although significant at p < 0.001, using stationary-site [O.sub.3] as the sole predictor of personal [O.sub.3], the [R.sup.2] ranged from only 0.04 in the spring to 0.07 in the fall of 1994 (Liu et al. 1997). In comparison, for the present study, using central-site gravimetric [PM.sub.2.5] as the sole predictor of 24-hr average personal pDR PM (approximates [PM.sub.2.5]), the [R.sup.2] is 0.18. More detailed exposure assessment analyses will be presented in a separate report.

Limitations. The relatively wider CIs for stationary monitor data compared with personal PM may occur because the personal sampling work was limited to three subjects at a time each followed over 2 weeks. This practical constraint is not an issue in other asthma panel studies that often rely exclusively on central-site PM data and are able to incorporate more person-days of outcome observation per subject. Given the limited duration of the individual time series, the likelihood is greater in personal exposure studies that exposure misclassification inherent in central-site data will inflate the error term for exposure-response relationships.

The pDR, like other nephelometers, is biased under high humidity conditions. We found this to be a minor and correctable problem in personal sampling data. The light scatter and mass relationship also depend on the diameter, refractive index, and density of particles (Thomas and Gebhart 1994), and particle size distribution in ambient air varies over space (Hering et al. 1997) and time (Morawska et al. 1999). Furthermore, although the pDR data may accurately record the timing of PM excursions, it does not inform us of the various sources and types of PM. Because of the differential light scattering responses to different PM components and lack of composition data, [FEV.sub.1] associations with maximum personal PM exposure should be interpreted with caution. Despite the drawbacks, we believe the pDR is the best available technology for real-time personal PM monitoring.

We made no measurements of aeroallergens. Our previous studies showed no confounding of air pollutant associations with asthma outcomes by outdoor fungal spores or pollen (Delfino et al. 1998, 2002). Nevertheless, it is conceivable that some of the relevant fine PM in the personal samples that led to [FEV.sub.1] deficits was allergenic. We know of no asthma panel study that has examined acute effects of aerometric measurements of indoor allergens such as dust mites and air pollutants simultaneously. Even though most high-molecular-weight allergens are likely carried by particles > 2.5 [micro]m in diameter, a sizable fraction, including that of dust mite allergen, is carried by fine PM (Custovic et al. 1999). It is of interest in this regard that stronger and more robust associations were found for indoor [PM.sub.10] than for outdoor [PM.sub.10], although CIs overlapped considerably (Figure 3), and we did not measure coarse particles directly ([PM.sub.10] can be in large part [PM.sub.2.5]). We speculate that this difference may have been a result of indoor activities that increased coarse particle fractions containing indoor allergens. This is supported by our finding of stronger associations between personal PM and [FEV.sub.1] among boys allergic to indoor allergens (Figure 5). For gravimetric stationary-site PM, some associations were nonsignificantly stronger among nonallergic subjects. However, 5-day average indoor [PM.sub.10] was significantly stronger among allergic subjects (Figure 6).

Another asthma panel study found that asthmatic children who were both exposed (by dust assay or self-report) and sensitized to cat showed [O.sub.3] more strongly associated with morning PEF decreases and with asthma symptom increases, but this was not found for HDM (Mortimer et al. 2000). In the present study, only one of the subjects with cat allergies reported having cats in the home.

Conclusions

The inverse associations of [FEV.sub.1] with stationary-site and personal PM found in the present study may reflect mixed respiratory effects of personal cloud, microenvironmental, and ambient background exposures to airborne PM. Although ambient PM clearly affected [FEV.sub.1], additional high personal cloud exposures were likely captured by the pDR and may have involved exposures or sources not well represented by the stationary-site monitors at either home microenvironments or central outdoor sites. This is supported by several findings as follows: a) Personal exposures showed stronger associations than stationary measurements; b) over all subjects, indoor PM exposures showed stronger associations than outdoor and central-site measurements; and c) among boys, subjects allergic to indoor allergens showed stronger associations for personal PM but generally similar associations for most stationary PM measurements compared with nonallergic subjects. This suggests two possible sources of PM may be driving associations: combustion sources (ambient infiltrated into homes or indoor sources) and personal and indoor activities that dominate during the daytime.

The present study, as with previous epidemiologic studies of particulate air pollution and asthma, is limited in the scope of inferences that can be made because particle exposures were represented only by total mass measurements. Further advancement in methods for assessing personal exposure to ambient PM is needed. This includes assessment of particle components that may influence airway inflammation or induce broncho-constriction, and better assessment of particle sources, including allergenic or combustion sources that may be controlled or avoided. These exposure assessment methods should be developed for use in epidemiologic research.

We thank the staff at the San Diego Air Pollution Control District for providing stationary-site data for air pollutant gases and weather and for their assistance in placement of our air monitors.

The project described was supported by grant ES06214 from the National Institute of Environmental Health Sciences (NIEHS), and its contents are solely the responsibility of the author and do not necessarily represent the official views of the NIEHS.

The authors declare they have no competing financial interests.

Received 24 October 2003; accepted 4 March 2004.

REFERENCES

Allen G, Sioutas C, Koutrakis P, Reiss R, Lurmann FW, Roberts PT. 1997. Evaluation of the TEOM method for measurement of ambient particulate mass in urban areas. J Air Waste Manage Assoc 47:682-689.

Allen R, Larson T, Wallace L, Sheppard L, Liu L-J S. 2003. Investigation of indoor and outdoor contributions to total indoor particulate matter exposure. Environ Sci Technol 37:3484-3492.

American Thoracic Society. 1991. Lung function testing: selection of reference values and interpretative strategies. Am Rev Respir Dis 144:1202-1218.

Custovic A, Woodcock H, Craven M, Hassall R, Hadley E, Simpson A, et al. 1999. Dust mite allergens are carried on not only large particles. Pediatr Allergy Immunol 10:258-260.

Delfino RJ, Gong H Jr, Linn WS, Hu Y, Pellizzari ED. 2003. Asthma symptoms in Hispanic children and daily ambient exposures to toxic and criteria air pollutants. Environ Health Perspect 111:647-656.

Delfino RJ, Zeiger RS, Seltzer JM, Street DH. 1998. Symptoms in pediatric asthmatics and air pollution: differences in effects by symptom severity, anti-inflammatory medication use, and particulate averaging time. Environ Health Perspect 106:751-761.

Delfino RJ, Zeiger RS, Seltzer JM, Street DH, McLaren C. 2002. Effects of hourly particulate air pollution on asthma symptoms: interaction with use of anti-inflammatory medications. Environ Health Perspect 110:A607-A617.

Diggle PJ, Heagerty P, Liang K-Y, Zeger SL. 2002. Analysis of Longitudinal Data. New York:Oxford University Press.

Gielen MH, van der Zee SC, Wijnen JH, van Steen CJ, Brunekreef B. 1997. Acute effects of summer air pollution on respiratory health of asthmatic children. Am J Respir Crit Care Med 155:2105-2108.

Hankinson JL, Odencrantz JR, Fedan KB. 1999. Spirometric reference values from a sample of the general U.S. population. Am J Respir Crit Care Med 159:179--187.

Hering S, Eldering A, Seinfeld JH. 1997. Bimodal character of accumulation mode aerosol mass distributions in southern California. Atmos Environ 3:1-11.

Hopke P, Ramadan Z, Paatero P, Norris G, Landis MS. Williams R, et al. 2003. Receptor modeling of ambient and personal exposure samples: 1998 Baltimore Particulate Matter Epidemiology-Exposure Study. Atmos Environ 37:3289-3302.

Howard-Reed C, Rea AW, Zufall MJ, Burke JM, Williams RW, Suggs JC, et al. 2000. Use of a continuous nephelometer to measure personal exposure to particles during the US Environmental Protection Agency Baltimore and Fresno panel studies. J Air Waste Manage Assoc 50:1125-1132.

Just J, Segala C, Sahraoui F, Priol G, Grimfeld A, Neukirch F. 2002. Short-term health effects of particulate and photochemical air pollution in asthmatic children. Eur Respir J 20:899-906.

Koenig JQ, Jansen K, Mar TF, Lumley T, Kaufman J, Trenga CA, et al. 2003. Measurement of offline exhaled nitric oxide in a study of community exposure to air pollution. Environ Health Perspect 111:1625-1629.

Larson T, Gould T, Simpson C, Claiborn C, Lewtas J, Liu L-JS. In press. Source apportionment of indoor, outdoor and personal [PM.sub.2.5] in Seattle, WA using positive matrix factorization. J Air Waste Manage Assoc.

Lee K, Vallarino J, Dumyahn T, Ozkaynak H, Spengler JD. 1999. Ozone decay rates in residences. J Air Waste Manag Assoc 49:1238-1244.

Lin M, Chen Y, Villeneuve PJ, Burnett RT, Lemyre L, Hertzman C, et al. 2004. Gaseous air pollutants and asthma hospitalization of children with low household income in Vancouver, British Columbia, Canada. Am J Epidemiol 159:294-303.

Lipfert FW, Wyzga RE. 1999. Statistical considerations in determining the health significance of constituents of airborne particulate matter. J Air Waste Manag Assoc 49:182-191.

Littell RC, Milliken GA, Stroup WW, Wolfinger RD. 1996. SAS System for Mixed Models. Cary, NC:SAS Institute Inc.

Liu L-JS, Box M, Kalman D, Kaufman J, Koenig J, Larson T, et al. 2003. Exposure assessment of particulate matter for susceptible populations in Seattle. Environ Health Perspect 111:909-918.

Liu L-JS, Delfino RJ, Koutrakis P. 1997. Ozone exposure assessment in a southern California community. Environ Health Perspect 105:58-65.

Liu L-JS, Slaughter C, Larson T. 2002. Comparison of light scattering devices and impactors for particulate measurements in indoor, outdoor, and personal environments. Environ Science Technol 36:2977-2986.

McBride SJ, Ferro AR, Ott WR, Switzer P, Hildemann LM. 1999. Investigations of the proximity effect for pollutants in the indoor environment. J Expo Anal Environ Epidemiol 9:602-621.

Morawska L, Thomas S, Jamriska M, Johnson G. 1999. The modality of particle size distribution of environmental aerosols. Atmos Environ 33:4401-4411.

Mortimer KM, Neas LM, Dockery DW, Redline S, Tager lB. 2002. The effect of air pollution on inner-city children with asthma. Eur Respir J 19:699-705.

Mortimer KM, Tager IB, Dockery DW, Neas LM, Redline S. 2000. The effect of ozone on inner-city children with asthma. Am J Respir Crit Care Med 162:1838-1845.

National Research Council. 1998. Research Priorities for Airborne Particulate Matter. I. Immediate Priorities and Long-Range Research Portfolio. Washington, DC:National Academy Press.

Ostro B, Lipsett M, Mann J, Braxton-0wens H, White M. 2001. Air pollution and exacerbation of asthma in African-American children in Los Angeles. Epidemiology 12:200-208.

Ozkaynak H, Xue J, Spengler J, Wallace L, Pellizzari E, Jenkins P. 1996. Personal exposure to airborne particles and metals: results from the particle team study in Riverside, California. J Expo Anal Environ Epidemiol 6:57-78.

Pekkanen J, Timonen KL, Ruuskanen J, Reponen A, Mirme A. 1997. Effects of ultrafine and fine particles in urban air on peak expiratory flow among children with asthmatic symptoms. Environ Res 74:24-33.

Peters A, Dockery DW, Heinrich J, Wichmann HE. 1997a. Short-term effects of particulate air pollution on respiratory morbidity in asthmatic children. Eur Respir J 10:872-879.

Peters A, Dockery DW, Heinrich J, Wichmann HE, 1997b. Medication use modifies the health effects of particulate sulfate air pollution in children with asthma. Environ Health Perspect 105:430-435.

Peters JM, Avol E, Gauderman WJ, Linn WS, Navidi W, London SJ, et al. 1999. A study of twelve southern California communities with differing levels and types of air pollution. II. Effects on pulmonary function. Am J Respir Crit Care Med 159:768-775.

Quintana PJE, Samimi BS, Kleinman MT, Liu L-JS, Soto K, Buffalino C, et al. 2000. Evaluation of a real-time passive personal particle monitor in fixed site residential indoor and ambient measurements. J Expo Analysis Environ Epidemiol 10:437-445.

Quintana PJE, Valenzia JR, Delfino, RJ, Liu L-JS. 2001. Monitoring of 1-minute personal particulate matter exposures in relation to voice-recorded time-activity data. Environ Res 87:199-213.

Richards LW, Alcorn SH, McDade C, Couture T, Lowenthal D, Chow JC, et al. 1999. Optical properties of the San Joaquin Valley aerosol collected during the 1995 Integrated Monitoring Study. Atmos Environ 33:4787-4795.

Roemer W, Clench-Aas J, Englert N, Hoek G, Katsouyanni K, Pekkanen J, et al. 1999. Inhomogeneity in response to air pollution in European children (PEACE project). Occup Environ Med 56:86-92.

Roemer W, Hoek G, Brunekreef B. 2000. Pollution effects on asthmatic children in Europe, the PEACE study. Clin Exp Allergy 30:1067-1075.

Romieu I, Meneses F, Ruiz S, Sienra JJ, Huerta J, White MC, et al. 1996. Effects of air pollution on the respiratory health of asthmatic children living in Mexico City. Am J Respir Crit Care Med 154:306-307.

Segala C, Fauroux B, Just J, Pascual L, Grimfeld A, Neukirch F. 1998. Short-term effect of winter air pollution on respiratory health of asthmatic children in Paris. Eur Respir J 11:677-685.

Slaughter JC, Lumley T, Sheppard L, Koenig JQ, Shapiro GG. 2003. Effects of ambient air pollution on symptom severity and medication use in children with asthma. Ann Allergy Asthma Immunol 91:346-363.

Thomas A, Gebhart J. 1994. Correlations between gravimetry and light scattering photometry for atmospheric aerosols. Atmos Environ 28:935-938.

Thurston GD, Lippmann M, Scott MB, Fine JM. 1997. Summertime haze air pollution and children with asthma. Am J Respir Crit Care Med 155:654-660.

Timonen KL, Pekkanen J. 1997. Air pollution and respiratory health among children with asthmatic or cough symptoms. Am J Respir Crit Care Med 156:546-552.

U.S. EPA. 1990. National Air Quality and Emissions Trends Report. Research Triangle Park, NC:U.S. Environmental Protection Agency Office of Air Quality Planning and Standards.

Vedal S, Petkau J, White R, Blair J. 1998. Acute effects of ambient inhalable particles in asthmatic and nonasthmatic children. Am J Respir Crit Care Med 157:1034-1043.

von Klot S, Wolke G, Tuch T, Heinrich J, Dockery DW, Schwartz J, et al. 2002. Increased asthma medication use in association with ambient fine and ultrafine particles. Eur Respir J 20:691-702.

Wu CF, Delfino RJ, Floro JN, Samimi BS, Quintana PJE, Kleinman MT, Liu L-JS. In press. Field evaluation and quality control of personal nephelometers in indoor, outdoor and personal environments. J Expo Analysis Environ Epidemiol doi:10.1038/sj.jea/'/500351 [Online 24 March 2004].

Yakovleva E, Hopke P, Wallace L. 1999. Positive matrix factorization in determining sources of particles measured in EPA's Particle TEAM Study. Environ Sci Technol 33:3645-3652.

Yu O, Sheppard L, Lurnley T, Koenig JQ, Shapiro GG. 2000. Effects of ambient air pollution on symptoms of asthma in Seattle-area children enrolled in the CAMP study. Environ Health Perspect 108:1209-1214.

Ralph J. Delfino, (1) Penelope J.E. Quintana, (2) Josh Floro, (1) Victor M. Gastanaga, (1) Behzad S. Samimi, (2) Michael T. Kleinman, (3) L.-J. Sally Liu, (4) Charles Bufalino, (1) Chang-Fu Wu, (4) and Christine E. McLaren (1)

(1) Epidemiology Division, Department of Medicine, College of Medicine, University of California at Irvine, Irvine, California, USA; (2) Division of Occupational and Environmental Health, Graduate School of Public Health, San Diego State University, San Diego, California, USA; (3) Department of Community and Environmental Medicine, College of Medicine, University of California at Irvine, Irvine, California, USA; (4) Department of Environmental and Occupational Health Sciences, University of Washington, Seattle, Washington, USA

Address correspondence to R.J. Delfino, Epidemiology Division, Department of Medicine, College of Medicine, 224 Irvine Hall, University of California Irvine, Irvine, CA 92697-7550 USA. Telephone: [phone]. Fax: [phone]. E-mail: [e-mail]

Table 1. Subject characteristics and health outcomes during
2-week exposure assessment periods, August-October 1999
and April-June 2000: Alpine, California, Asthma Panel Study.

                                      No. of 2-hr
                                      periods with
                                       bothersome       Daily
                                       or worse       prescribed
                                        asthma          asthma
          Age               Height      symptoms      medications
ID      (years)    Sex     (inches)    (% total)         taken

1         10       Male       57           0             None
1 (c)     11                  59           0             None
2         12       Male       61           0             None
3         13       Male       70           0             None
4         11      Female      57           0             None
5         12       Male       62        1 (1.2)      Triamcinolone
6         10      Female      60        4 (14.3)         None
7         11      Female      58           0             None
8         11      Female      62        4 (4.9)          None
9         14       Male       64           0             None
10         9       Male       59           0             None
11         9       Male       57           0          Budesonide,
                                                      Salmeterol
12        12       Male       62       24 (25.0)         None
13         9       Male       52           0          Budesonide,
                                                      Salmeterol
14        13       Male       62           0             None
15        11       Male       54           0             None
16        15       Male       64           0          Fluticasone
17        14       Male       65           0          Zafirlukast
18        17      Female      67           0             None
19        11       Male       61           0             None

                               Mean daily              Mean
                               as-needed              morning
          Age                 inhaler use           [FEV.sub.1]
ID      (years)    Sex        [+ or -] SD       (L/sec [+ or -] SD)

1         10       Male    0.29 [+ or -] 0.73   1.89 [+ or -] 0.07
1 (c)     11                      0.0           2.16 [+ or -] 0.10
2         12       Male           0.0           1.78 [+ or -] 0.10
3         13       Male           0.0           3.40 [+ or -] 0.12
4         11      Female   2.17 [+ or -] 2.33   1.51 [+ or -] 0.07
5         12       Male    3.00 [+ or -] 1.30   2.63 [+ or -] 0.11
6         10      Female   0.86 [+ or -] 2.27   1.78 [+ or -] 0.29
7         11      Female   0.29 [+ or -] 0.73   1.26 [+ or -] 0.11
8         11      Female   0.43 [+ or -] 0.85   2.04 [+ or -] 0.20
9         14       Male           0.0           1.99 [+ or -] 0.11
10         9       Male    1.00 [+ or -] 1.73   1.80 [+ or -] 0.08
11         9       Male           0.0           1.51 [+ or -] 0.19

12        12       Male    2.07 [+ or -] 4.05   1.46 [+ or -] 0.30
13         9       Male           0.0           1.45 [+ or -] 0.10

14        13       Male           0.0           1.72 [+ or -] 0.12
15        11       Male           0.0           1.44 [+ or -] 0.13
16        15       Male           0.0           2.39 [+ or -] 0.25
17        14       Male    0.46 [+ or -] 0.88   1.80 [+ or -] 0.15
18        17      Female          0.0           2.81 [+ or -] 0.14
19        11       Male           0.0           1.70 [+ or -] 0.19

                                  Mean
                               afternoon               Mean
                              [FEV.sub.1]             evening
          Age               (L/sec [+ or -]         [FEV.sub.1]
ID      (years)    Sex          SD) (a)         (L/sec [+ or -] SD)

1         10       Male    1.93 [+ or -] 0.13   1.86 [+ or -] 0.16
1 (c)     11               2.11 [+ or -] 0.14   2.11 [+ or -] 0.07
2         12       Male    1.56 [+ or -] 0.27   1.64 [+ or -] 0.13
3         13       Male    3.45 [+ or -] 0.10   3.40 [+ or -] 0.17
4         11      Female   1.37 [+ or -] 0.11   1.43 [+ or -] 0.07
5         12       Male    2.75 [+ or -] 0.15   2.66 [+ or -] 0.14
6         10      Female   1.89 [+ or -] 0.18   1.69 [+ or -] 0.24
7         11      Female   1.34 [+ or -] 0.06   1.31 [+ or -] 0.15
8         11      Female   2.04 [+ or -] 0.18   2.09 [+ or -] 0.15
9         14       Male    2.03 [+ or -] 0.08   2.02 [+ or -] 0.11
10         9       Male    1.84 [+ or -] 0.17   1.80 [+ or -] 0.11
11         9       Male    1.51 [+ or -] 0.27   1.55 [+ or -] 0.15

12        12       Male    1.63 [+ or -] 0.29   1.63 [+ or -] 0.32
13         9       Male            --           1.19 [+ or -] 0.27

14        13       Male    1.90 [+ or -] 0.21   1.66 [+ or -] 0.14
15        11       Male    1.62 [+ or -] 0.04   1.38 [+ or -] 0.16
16        15       Male    2.35 [+ or -] 0.24   2.29 [+ or -] 0.27
17        14       Male    1.89 [+ or -] 0.14   1.82 [+ or -] 0.17
18        17      Female   2.71 [+ or -] 0.18   2.70 [+ or -] 0.16
19        11       Male    1.98 [+ or -] 0.19   1.88 [+ or -] 0.08

                               Overall
                               percent
          Age                 predicted
ID      (years)    Sex     [FEV.sub.1] (b)

1         10       Male         84.1
1 (c)     11                    84.7
2         12       Male         56.1
3         13       Male         86.9
4         11      Female        64.5
5         12       Male         92.3
6         10      Female        73.8
7         11      Female        55.4
8         11      Female        76.0
9         14       Male         61.4
10         9       Male         75.1
11         9       Male         69.3

12        12       Male         54.3
13         9       Male         77.7

14        13       Male         59.4
15        11       Male         71.5
16        15       Male         71.5
17        14       Male         54.1
18        17      Female        76.9
19        11       Male         69.8

ID, identification number.

(a) Subjects 11 and 13 were noncompliant with
afternoon lung function maneuvers (ID11 performed
four afternoon maneuvers, and ID13 performed none).
(b) NHANES III (Hankinson et al., 1999). (c) Subject
1 participated in two 2-week exposure assessments
in fall 1999 and spring 2000.

Table 2. Daily air pollution and weather measurements during a
personal particle exposure assessment study, September-October
1999, and April-June 2000: Alpine, California, Asthma Panel Study.

                                                       Arithmetic
Exposure and averaging time             No. (a)     mean [+ or -] SD

pDR PM ([micro]g/[m.sup.3])
  Mean personal PM last 2 hr              452      34.4 [+ or -] 33.7
  12-hr Daytime mean personal PM          178      55.7 [+ or -] 31.6
  12-hr Nighttime mean personal PM        200      22.3 [+ or -] 13.6
  1-hr Maximum personal PM last 24 hr     419     151.0 [+ or -] 120.3
  4-hr Maximum personal PM last 24 hr     419      87.5 [+ or -] 55.3
  8-hr Maximum personal PM last 24 hr     419      67.6 [+ or -] 39.0
  Mean personal PM last 24 hr             419      37.9 [+ or -] 19.9
Home stationary-site PM ([micro]g/
    [m.sup.3])
  24-hr Mean indoor gravimetric           216      30.3 [+ or -] 11.9
    [PM.sub.10]
  24-hr Mean indoor gravimetric           219      12.1 [+ or -] 5.4
    [PM.sub.2.5]
  24-hr Mean outdoor gravimetric          226      25.9 [+ or -] 10.4
    [PM.sub.10]
  24-hr Mean outdoor gravimetric          226      11.0 [+ or -] 5.4
    [PM.sub.2.5]
Central outdoor stationary-site PM
    ([micro]g/[m.sup.3])
  12-hr Daytime TEOM [PM.sub.10]          232      35.1 [+ or -] 11.3
  12-hr Nighttime TEOM [PM.sub.10]        232      23.3 [+ or -] 8.4
  1-hr Maximum TEOM [PM.sub.10] last      535      54.4 [+ or -] 13.8
    24 hr
  4-hr Maximum TEOM [PM.sub.10] last      535      44.8 [+ or -] 12.4
    24 hr
  8-hr Maximum TEOM [PM.sub.10] last      535      39.8 [+ or -] 11.2
    24 hr
  Mean TEOM [PM.sub.10] last 24 hr        535      29.7 [+ or -] 8.6
  24-hr Mean gravimetric [PM.sub.10]      236      23.6 [+ or -] 9.1
  24-hr Mean gravimetric [PM.sub.2.5]     232      10.3 [+ or -] 5.6
Central outdoor stationary-site
    pollutant gases and weather
  8-hr Maximum daily [O.sub.3] last       535      62.9 [+ or -] 15.1
    24 hr (ppb)
  8-hr Maximum daily N[O.sub.2] last      535      19.6 [+ or -] 7.0
    24 hr (ppb)
  1-hr Maximum temperature last 24 hr     543      79.1 [+ or -] 9.1
    ([degrees]F)
  1-hr Minimum RH last 24 hr (%)          526      36.3 [+ or -] 19.7
Personal weather
  1-hr Maximum temperature last 24 hr     530      86.4 [+ or -] 5.6
    ([degrees]F)
  1-hr Minimum RH last 24 hr (%)          530      30.8 [+ or -] 7.5

                                        Geometric
Exposure and averaging time               mean       IQR

pDR PM ([micro]g/[m.sup.3])
  Mean personal PM last 2 hr               23.4      36.0
  12-hr Daytime mean personal PM           47.9      40.0
  12-hr Nighttime mean personal PM         19.2      18.4
  1-hr Maximum personal PM last 24 hr     117.3     128.0
  4-hr Maximum personal PM last 24 hr      72.8      63.8
  8-hr Maximum personal PM last 24 hr      57.5      53.7
  Mean personal PM last 24 hr              33.4      30.3
Home stationary-site PM ([micro]g/
    [m.sup.3])
  24-hr Mean indoor gravimetric            29.1      15.5
    [PM.sub.10]
  24-hr Mean indoor gravimetric            12.0       6.7
    [PM.sub.2.5]
  24-hr Mean outdoor gravimetric           24.8      14.9
    [PM.sub.10]
  24-hr Mean outdoor gravimetric           10.7       7.1
    [PM.sub.2.5]
Central outdoor stationary-site PM
    ([micro]g/[m.sup.3])
  12-hr Daytime TEOM [PM.sub.10]           34.2      19.3
  12-hr Nighttime TEOM [PM.sub.10]         22.8      13.5
  1-hr Maximum TEOM [PM.sub.10] last       53.7      17.7
    24 hr
  4-hr Maximum TEOM [PM.sub.10] last       44.1      16.1
    24 hr
  8-hr Maximum TEOM [PM.sub.10] last       39.2      16.6
    24 hr
  Mean TEOM [PM.sub.10] last 24 hr         29.4      13.1
  24-hr Mean gravimetric [PM.sub.10]       22.7      14.6
  24-hr Mean gravimetric [PM.sub.2.5]       9.9       7.5
Central outdoor stationary-site
    pollutant gases and weather
  8-hr Maximum daily [O.sub.3] last        62.1      22.0
    24 hr (ppb)
  8-hr Maximum daily N[O.sub.2] last       19.3      10.5
    24 hr (ppb)
  1-hr Maximum temperature last 24 hr      79.6      13.0
    ([degrees]F)
  1-hr Minimum RH last 24 hr (%)           31.3      34.0
Personal weather
  1-hr Maximum temperature last 24 hr      87.2       6.9
    ([degrees]F)
  1-hr Minimum RH last 24 hr (%)           31.0      12.4

                                         Minimum/       90th
Exposure and averaging time              Maximum     percentile

pDR PM ([micro]g/[m.sup.3])
  Mean personal PM last 2 hr             0.0/305.3       67.6
  12-hr Daytime mean personal PM         5.7/181.9      100.8
  12-hr Nighttime mean personal PM       2.7/65.6        40.2
  1-hr Maximum personal PM last 24 hr    9.1/996.8      292.4
  4-hr Maximum personal PM last 24 hr    7.1/344.1      168.2
  8-hr Maximum personal PM last 24 hr    5.3/225.9      121.9
  Mean personal PM last 24 hr            3.9/113.8       65.1
Home stationary-site PM ([micro]g/
    [m.sup.3])
  24-hr Mean indoor gravimetric          8.7/74.8        45.6
    [PM.sub.10]
  24-hr Mean indoor gravimetric          2.8/35.3        20.2
    [PM.sub.2.5]
  24-hr Mean outdoor gravimetric         6.6/68.4        38.6
    [PM.sub.10]
  24-hr Mean outdoor gravimetric         1.8/31.0        18.4
    [PM.sub.2.5]
Central outdoor stationary-site PM
    ([micro]g/[m.sup.3])
  12-hr Daytime TEOM [PM.sub.10]        16.4/60.4        50.4
  12-hr Nighttime TEOM [PM.sub.10]       9.0/45.1        33.5
  1-hr Maximum TEOM [PM.sub.10] last    24.4/95.4        71.0
    24 hr
  4-hr Maximum TEOM [PM.sub.10] last    20.8/77.6        62.4
    24 hr
  8-hr Maximum TEOM [PM.sub.10] last    16.8/70.7        53.9
    24 hr
  Mean TEOM [PM.sub.10] last 24 hr      12.9/50.7        40.9
  24-hr Mean gravimetric [PM.sub.10]     3.2/48.0        34.6
  24-hr Mean gravimetric [PM.sub.2.5]    1.7/29.1        18.4
Central outdoor stationary-site
    pollutant gases and weather
  8-hr Maximum daily [O.sub.3] last     25.0/105.9       83.9
    24 hr (ppb)
  8-hr Maximum daily N[O.sub.2] last     5.3/38.4        28.9
    24 hr (ppb)
  1-hr Maximum temperature last 24 hr   57.0/97.0        89.0
    ([degrees]F)
  1-hr Minimum RH last 24 hr (%)         5.0/92.0        60.0
Personal weather
  1-hr Maximum temperature last 24 hr   73.1/103.5       93.8
    ([degrees]F)
  1-hr Minimum RH last 24 hr (%)        22.1/59.2        42.3

(a) The sample size refers to all unique data used in [FEV.sub.1]
regression models. Where the data are continuously measured over
the period before an [FEV.sub.1] maneuver (exposure "last 2" or
"last 24 hr"), more than one measurement per person-day is possible
(e.g., 1-hr maximum the preceding 24 hr before the morning, afternoon,
and evening [FEV.sub.1]). Data thus include up to 535
person-observation times with both [FEV.sub.1] and exposure. For
measurements made during fixed time intervals, namely, gravimetric
PM or 12-hr daytime or nighttime PM, the no. represents one unique
observation per day, per site (up to 236 person-observation times).

Table 3. Exposure correlation matrix, (a) August-October 1999 and
April-June 2000: Alpine, California, Asthma Panel Study.

                                             8-hr Max   24-hr Mean
                                 8-hr Max    personal    personal
Exposure                        N[O.sub.2]      PM          PM

8-hr Max [O.sub.3]                -0.05        0.03        0.01
8-hr Max N[O.sub.2]                1.00        0.26        0.27
8-hr Max personal PM                           1.00        0.94
24-hr Mean personal PM                                     1.00
8-hr Max TEOM [PM.sub.10]
24-hr Mean TEOM [PM.sub.10]
24-hr Central HI [PM.sub.10]
24-hr Central HI [PM.sub.2.5]
24-hr Outdoor HI [PM.sub.10]
24-hr Outdoor HI [PM.sub.2.5]
24-hr Indoor HI [PM.sub.10]

                                 8-hr Max     24-hr Mean       24-hr
                                   TEOM          TEOM       Central HI
Exposure                        [PM.sub.10]   [PM.sub.10]   [PM.sub.10]

8-hr Max [O.sub.3]                 0.40          0.41          0.23
8-hr Max N[O.sub.2]                0.57          0.58          0.73
8-hr Max personal PM               0.38          0.40          0.37
24-hr Mean personal PM             0.36          0.39          0.36
8-hr Max TEOM [PM.sub.10]          1.00          0.92          0.78
24-hr Mean TEOM [PM.sub.10]                      1.00          0.86
24-hr Central HI [PM.sub.10]                                   1.00
24-hr Central HI [PM.sub.2.5]
24-hr Outdoor HI [PM.sub.10]
24-hr Outdoor HI [PM.sub.2.5]
24-hr Indoor HI [PM.sub.10]

                                   24-hr          24-hr
                                 Central HI    Outdoor HI
Exposure                        [PM.sub.2.5]   [PM.sub.10]

8-hr Max [O.sub.3]                  0.24          0.19
8-hr Max N[O.sub.2]                 0.73          0.56
8-hr Max personal PM                0.38          0.32
24-hr Mean personal PM              0.43          0.34
8-hr Max TEOM [PM.sub.10]           0.71          0.72
24-hr Mean TEOM [PM.sub.10]         0.78          0.79
24-hr Central HI [PM.sub.10]        0.90          0.80
24-hr Central HI [PM.sub.2.5]       1.00          0.72
24-hr Outdoor HI [PM.sub.10]                      1.00
24-hr Outdoor HI [PM.sub.2.5]
24-hr Indoor HI [PM.sub.10]

                                   24-hr          24-hr
                                 Outdoor HI     Indoor HI
Exposure                        [PM.sub.2.5]   [PM.sub.10]

8-hr Max [O.sub.3]                  0.27          0.04
8-hr Max N[O.sub.2]                 0.66          0.24
8-hr Max personal PM                0.39          0.23
24-hr Mean personal PM              0.44          0.29
8-hr Max TEOM [PM.sub.10]           0.75          0.36
24-hr Mean TEOM [PM.sub.10]         0.78          0.36
24-hr Central HI [PM.sub.10]        0.83          0.37
24-hr Central HI [PM.sub.2.5]       0.89          0.40
24-hr Outdoor HI [PM.sub.10]        0.86          0.44
24-hr Outdoor HI [PM.sub.2.5]       1.00          0.45
24-hr Indoor HI [PM.sub.10]                       1.00

                                   24-hr
                                 Indoor HI
Exposure                        [PM.sub.2.5]

8-hr Max [O.sub.3]                  0.15
8-hr Max N[O.sub.2]                 0.50
8-hr Max personal PM                0.37
24-hr Mean personal PM              0.46
8-hr Max TEOM [PM.sub.10]           0.58
24-hr Mean TEOM [PM.sub.10]         0.59
24-hr Central HI [PM.sub.10]        0.67
24-hr Central HI [PM.sub.2.5]       0.73
24-hr Outdoor HI [PM.sub.10]        0.66
24-hr Outdoor HI [PM.sub.2.5]       0.79
24-hr Indoor HI [PM.sub.10]         0.74

Definitions: Central, central regional U.S. EPA site; HI, Harvard
impactor gravimetric mass on Teflon filters; Indoor, indoor home
sites; Max, maximum; Outdoor, outdoor home sites.

(a) Spearman's rank correlations. Exposure measurements are
observations from the period around 1800 hr to 1800 hr to match
24-hr sampling periods of the gravimetric samplers (HIs).