Comparative modeling approaches for personal exposure to particle-associated PAH.

Several models for simulation of personal exposure (PE) to particle-associated polycyclic aromatic hydrocarbons (PAH) have been developed and tested. The modeling approaches include linear regression models (Model 1), time activity weighted models (Models 2 and 3), a hybrid model (Model 4), a univariate linear model (Model 5), and machine learning technique models (Model 6 and 7). The hybrid model (Model 4), which utilizes microenvironment data derived from time-activity diaries (TAD) with the implementation of add-on variables to account for external factors that might affect PE, proved to be the best regression model (R(2) for B(a)P = 0.346, p < 0.01; N = 68). This model was compared with results from two machine learning techniques, namely decision trees (Model 6) and neural networks (Model 7), which represent an innovative approach to PE modeling. The neural network model was promising in giving higher correlation coefficient results for all PAH (R(2) for B(a)P = 0.567, p < 0.01; N = 68) and good performance with the smaller test data set (R(2) for B(a)P = 0.640, p < 0.01; N = 23). Decision tree accuracies (Model 6) which assess how precisely the algorithm can determine the correct classification of a PE concentration range indicate good performance, but this is not comparable to the other models through R(2) values. Using neural networks (Model 7) showed significant improvements over the performance of hybrid Model 4 and the univariate general linear Model 5 for test samples (not used in developing the models). The worst performance was given by linear regression Models 1 to 3 based solely on home and workplace concentrations and time-activity data.

Measurement and modeling of exposure to selected air toxics for health effects studies and verification by biomarkers.

The overall aim of our investigation was to quantify the magnitude and range of individual personal exposures to a variety of air toxics and to develop models for exposure prediction on the basis of time-activity diaries. The specific research goals were (1) to use personal monitoring of non-smokers at a range of residential locations and exposures to non-traffic sources to assess daily exposures to a range of air toxics, especially volatile organic compounds (VOCs) including 1,3-butadiene and particulate polycyclic aromatic hydrocarbons (PAHs); (2) to determine microenvironmental concentrations of the same air toxics, taking account of spatial and temporal variations and hot spots; (3) to optimize a model of personal exposure using microenvironmental concentration data and time-activity diaries and to compare modeled exposures with exposures independently estimated from personal monitoring data; (4) to determine the relationships of urinary biomarkers with the environmental exposures to the corresponding air toxic. Personal exposure measurements were made using an actively pumped personal sampler enclosed in a briefcase. Five 24-hour integrated personal samples were collected from 100 volunteers with a range of exposure patterns for analysis of VOCs and 1,3-butadiene concentrations of ambient air. One 24-hour integrated PAH personal exposure sample was collected by each subject concurrently with 24 hours of the personal sampling for VOCs. During the period when personal exposures were being measured, workplace and home concentrations of the same air toxics were being measured simultaneously, as were seasonal levels in other microenvironments that the subjects visit during their daily activities, including street microenvironments, transport microenvironments, indoor environments, and other home environments. Information about subjects' lifestyles and daily activities were recorded by means of questionnaires and activity diaries. VOCs were collected in tubes packed with the adsorbent resins Tenax GR and Carbotrap, and separate tubes for the collection of 1,3-butadiene were packed with Carbopack B and Carbosieve S-III. After sampling, the tubes were analyzed by means of a thermal desorber interfaced with a gas chromatograph-mass spectrometer (GC-MS). Particle-phase PAHs collected onto a quartz-fiber filter were extracted with solvent, purified, and concentrated before being analyzed with a GC-MS. Urinary biomarkers were analyzed by liquid chromatography-tandem mass spectrometry (LC-MS-MS). Both the environmental concentrations and personal exposure concentrations measured in this study are lower than those in the majority of earlier published work, which is consistent with the reported application of abatement measures to the control of air toxics emissions. The environmental concentration data clearly demonstrate the influence of traffic sources and meteorologic conditions leading to higher air toxics concentrations in the winter and during peak-traffic hours. The seasonal effect was also observed in indoor environments, where indoor sources add to the effects of the previously identified outdoor sources. The variability of personal exposure concentrations of VOCs and PAHs mainly reflects the range of activities the subjects engaged in during the five-day period of sampling. A number of generic factors have been identified to influence personal exposure concentrations to VOCs, such as the presence of an integral garage (attached to the home), exposure to environmental tobacco smoke (ETS), use of solvents, and commuting. In the case of the medium- and high-molecular-weight PAHs, traffic and ETS are important contributions to personal exposure. Personal exposure concentrations generally exceed home indoor concentrations, which in turn exceed outdoor concentrations. The home microenvironment is the dominant individual contributor to personal exposure. However, for those subjects with particularly high personal exposures, activities within the home and exposure to ETS play a major role in determining exposure. Correlation analysis and principal components analysis (PCA) have been performed to identify groups of compounds that share common sources, common chemistry, or common transport or meteorologic patterns. We used these methods to identify four main factors determining the makeup of personal exposures: fossil fuel combustion, use of solvents, ETS exposure, and use of consumer products. Concurrent with sampling of the selected air toxics, a total of 500 urine samples were collected, one for each of the 100 subjects on the day after each of the five days on which the briefcases were carried for personal exposure data collection. From the 500 samples, 100 were selected to be analyzed for PAHs and ETS-related urinary biomarkers. Results showed that urinary biomarkers of ETS exposure correlated strongly with the gas-phase markers of ETS and 1,3-butadiene. The urinary ETS biomarkers also correlated strongly with high-molecular-weight PAHs in the personal exposure samples. Five different approaches have been taken to model personal exposure to VOCs and PAHs, using 75% of the measured personal exposure data set to develop the models and 25% as an independent check on the model performance. The best personal exposure model, based on measured microenvironmental concentrations and lifestyle factors, is able to account for about 50% of the variance in measured personal exposure to benzene and a higher proportion of the variance for some other compounds (e.g., 75% of the variance in 3-ethenylpyridine exposure). In the case of the PAHs, the best model for benzo[a]pyrene is able to account for about 35% of the variance among exposures, with a similar result for the rest of the PAH compounds. The models developed were validated by the independent data set for almost all the VOC compounds. The models developed for PAHs explain some of the variance in the independent data set and are good indicators of the sources affecting PAH concentrations but could not be validated statistically, with the exception of the model for pyrene. A proposal for categorizing personal exposures as low or high is also presented, according to exposure thresholds. For both VOCs and PAHs, low exposures are correctly classified for the concentrations predicted by the proposed models, but higher exposures were less successfully classified.

Dependence of home outdoor particulate mass and number concentrations on residential and traffic features in urban areas.

The associations between residential outdoor and ambient particle mass, fine particle absorbance, particle number (PN) concentrations, and residential and traffic determinants were investigated in four European urban areas (Helsinki, Athens, Amsterdam, and Birmingham). A total of 152 nonsmoking participants with respiratory diseases, not exposed to occupational pollution, were included in the study, which comprised a 7-day intensive exposure monitoring period of both indoor and home outdoor particle mass and number concentrations. The same pollutants were also continuously measured at ambient fixed sites centrally located to the studied areas (fixed ambient sites). Relationships between concentrations measured directly outside the homes (residential outdoor) and at the fixed ambient sites were pollutant-specific, with substantial variations among the urban areas. Differences were more pronounced for coarse particles due to resuspension of road dust and PN, which is strongly related to traffic emissions. Less significant outdoor-to-fixed variation for particle mass was observed for Amsterdam and Birmingham, predominantly due to regional secondary aerosol. On the contrary, a strong spatial variation was observed for Athens and to a lesser extent for Helsinki. This was attributed to the overwhelming and time-varied inputs from traffic and other local sources. The location of the residence and traffic volume and distance to street and traffic light were important determinants of residential outdoor particle concentrations. On average, particle mass levels in suburban areas were less than 30% of those measured for residences located in the city center. Residences located less than 10 m from a street experienced 133% higher PN concentrations than residences located further away. Overall, the findings of this multi-city study, indicated that (1) spatial variation was larger for PN than for fine particulate matter (PM) mass and varied between the cities, (2) vehicular emissions in the residential street and location in the center of the city were significant predictors of spatial variation, and (3) the impact of traffic and location in the city was much larger for PN than for fine particle mass.

Relationship of personal exposure to volatile organic compounds to home, work and fixed site outdoor concentrations.

Personal exposures of 100 adult non-smokers living in the UK, as well as home and workplace microenvironment concentrations of 15 volatile organic compounds were investigated. The strength of the association between personal exposure and indoor home and workplace concentrations as well as with central site ambient air concentrations in medium to low pollution areas was assessed. Home microenvironment concentrations were strongly associated with personal exposures indicating that the home is the driving factor determining personal exposures to VOCs, explaining between 11 and 75% of the total variability. Workplace and central site ambient concentrations were less correlated with the corresponding personal concentrations, explaining up to 11-22% of the variability only at the low exposure end of the concentration range (e.g. benzene concentrations <2.5 µg m(-3)). One of the reasons for the discrepancies between personal exposures and central site data was that the latter does not account for exposure due to personal activities (e.g. commuting, painting). A moderate effect of season on the strength of the association between personal exposure and ambient concentrations was found. This needs to be taken into account when using fixed site measurements to infer exposures.

Temporal variations of atmospheric aerosol in four European urban areas.

PURPOSE: The concentrations of PM(10) mass, PM(2.5) mass and particle number were continuously measured for 18 months in urban background locations across Europe to determine the spatial and temporal variability of particulate matter. METHODS: Daily PM(10) and PM(2.5) samples were continuously collected from October 2002 to April 2004 in background areas in Helsinki, Athens, Amsterdam and Birmingham. Particle mass was determined using analytical microbalances with precision of 1 µg. Pre- and post-reflectance measurements were taken using smoke-stain reflectometers. One-minute measurements of particle number were obtained using condensation particle counters. RESULTS: The 18-month mean PM(10) and PM(2.5) mass concentrations ranged from 15.4 µg/m(3) in Helsinki to 56.7 µg/m(3) in Athens and from 9.0 µg/m(3) in Helsinki to 25.0 µg/m(3) in Athens, respectively. Particle number concentrations ranged from 10,091 part/cm(3) in Helsinki to 24,180 part/cm(3) in Athens with highest levels being measured in winter. Fine particles accounted for more than 60% of PM(10) with the exception of Athens where PM(2.5) comprised 43% of PM(10). Higher PM mass and number concentrations were measured in winter as compared to summer in all urban areas at a significance level p < 0.05. CONCLUSIONS: Significant quantitative and qualitative differences for particle mass across the four urban areas in Europe were observed. These were due to strong local and regional characteristics of particulate pollution sources which contribute to the heterogeneity of health responses. In addition, these findings also bear on the ability of different countries to comply with existing directives and the effectiveness of mitigation policies.

Environmental and biological monitoring of exposures to PAHs and ETS in the general population.

The objective of this study was to analyse environmental tobacco smoke (ETS) and PAH metabolites in urine samples of non-occupationally exposed non-smoker adult subjects and to establish relationships between airborne exposures and urinary concentrations in order to (a) assess the suitability of the studied metabolites as biomarkers of PAH and ETS, (b) study the use of 3-ethenypyridine as ETS tracer and (c) link ETS scenarios with exposures to carcinogenic PAH and VOC. Urine samples from 100 subjects were collected and concentrations of monophenolic metabolites of naphthalene, fluorene, phenanthrene, and pyrene and the nicotine metabolites cotinine and trans-3'-hydroxycotinine were measured using liquid chromatography-tandem mass spectrometry (LC-MS/MS) to assess PAH and ETS exposures. Airborne exposures were measured using personal exposure samplers and analysed using GC-MS. These included 1,3-butadiene (BUT), 3-ethenylpyridine (3-EP) (a tobacco-specific tracer derived from nicotine pyrolysis) and PAHs. ETS was reported by the subjects in 30-min time-activity questionnaires and specific comments were collected in an ETS questionnaire each time ETS exposure occurred. The values of 3-EP (>0.25 microg/m(3) for ETS) were used to confirm the ETS exposure status of the subject. Concentrations as geometric mean, GM, and standard deviation (GSD) of personal exposures were 0.16 (5.50)microg/m(3) for 3-EP, 0.22 (4.28)microg/m(3) for BUT and 0.09 (3.03)ng/m(3) for benzo(a)pyrene. Concentrations of urinary metabolites were 0.44 (1.70)ng/mL for 1-hydroxypyrene and 0.88 (5.28)ng/mL for cotinine. Concentrations of urinary metabolites of nicotine were lower than in most previous studies, suggesting very low exposures in the ETS-exposed group. Nonetheless, concentrations were higher in the ETS population for cotinine, trans-3'hydroxycotinine, 3-EP, BUT and most high molecular weight PAH, whilst 2-hydroxyphenanthrene, 3+4-hydroxyphenanthrene and 1-hydroxyphenanthrene were only higher in the high-ETS subpopulation. There were not many significant correlations between either personal exposures to PAH and their urinary metabolites, or of the latter with ETS markers. However, it was found that the urinary log cotinine concentration showed significant correlation with log concentrations of 3-EP (R=0.75), BUT (R=0.47), and high molecular weight PAHs (MW>200), especially chrysene (R=0.55) at the p=0.01 level. On the other hand, low correlation was observed between the PAH metabolite 2-naphthol and the parent PAH, gas-phase naphthalene. These results suggest that (1) ETS is a significant source of inhalation exposure to the carcinogen 1,3-butadiene and high molecular weight PAHs, many of which are carcinogenic, and (2) that for lower molecular weight PAHs such as naphthalene, exposure by routes other than inhalation predominate, since metabolite levels correlated poorly with personal exposure air sampling.

Measurement of personal exposure to volatile organic compounds and particle associated PAH in three UK regions.

Personal exposures to 15 volatile organic compounds (VOC) and 16 polycyclic aromatic hydrocarbons (PAH) of 100 adult nonsmokers living in three UK areas, namely London, West Midlands, and rural South Wales, were measured using an actively pumped sampler carried around by the volunteers for 5/1 (VOC/PAH) consecutive 24-h periods, following their normal lifestyle. Results from personal exposure measurements categorized by geographical location, type of dwelling, and exposure to environmental tobacco smoke (ETS) are presented. The average personal exposure concentration to benzene, 1,3-butadiene, and benzo(a)pyrene representing the main carcinogenic components of the VOC and PAH mixture were 2.2 +/- 2.5 microg/m3, 0.4 +/- 0.7 microg/m3, and 0.3 +/- 0.7 ng/m3 respectively. The association of a number of generic factors with personal exposure concentrations was investigated, including first-line property, traffic, the presence of an integral garage, and ETS. Only living in houses with integral garages and being exposed to ETS were identified as unequivocal contributors to VOC personal exposure, while only ETS had a clear effect upon PAH personal exposures. The measurements of personal exposures were compared with health-based European and UK air quality guidelines, with some exceedences occurring. Activities contributing to high personal exposures included the use of a fireplace in the home, ETS exposure, DIY (i.e., construction and craftwork activities), and photocopying, among others.

Model development and validation of personal exposure to volatile organic compound concentrations.

BACKGROUND: Direct measurement of exposure to volatile organic compounds (VOCs) via personal monitoring is the most accurate exposure assessment method available. However, its wide-scale application to evaluating exposures at the population level is prohibitive in terms of both cost and time. Consequently, indirect measurements via a combination of microenvironment concentrations and personal activity diaries represent a potentially useful alternative. OBJECTIVE: The aim of this study was to optimize a model of personal exposures (PEs) based on microenvironment concentrations and time/activity diaries and to compare modeled with measured exposures in an independent data set. MATERIALS: VOC PEs and a range of microenvironment concentrations were collected with active samplers and sorbent tubes. Data were supplemented with information collected through questionnaires. Seven models were tested to predict PE to VOCs in 75% (n = 370) of the measured PE data set, whereas the other 25% (n = 120) was used for validation purposes. RESULTS: The best model able to predict PE with independence of measurements was based upon stratified microenvironment concentrations, lifestyle factors, and individual-level activities. The proposed model accounts for 40-85% of the variance for individual VOCs and was validated for almost all VOCs, showing normalized mean bias and mean fractional bias below 25% and predicting 60% of the values within a factor of 2. CONCLUSIONS: The models proposed identify the most important non-weather-related variables for VOC exposures; highlight the effect of personal activities, use of solvents, and exposure to environmental tobacco smoke on PE levels; and may assist in the development of specific models for other locations.