Grantee Research Project Results
Final Report: Ultrafine Particles in Urban and Respiratory Health Among Children with Respiratory Symptoms
EPA Grant Number: R825265Title: Ultrafine Particles in Urban and Respiratory Health Among Children with Respiratory Symptoms
Investigators:
Institution: Harvard University
EPA Project Officer: Chung, Serena
Project Period: December 2, 1996 through December 1, 1999
Project Amount: $196,185
RFA: Air Quality (1996) RFA Text | Recipients Lists
Research Category: Air Quality and Air Toxics , Air
Objective:
The effects of particulate air pollution on the respiratory health of symptomatic children were studied in Kuopio, Finland, during the winter and spring of 1994, and during the spring of 1995. In 1994, primary school children were followed with daily diaries and peak expiratory flow (PEF) measurements for 3 months. In addition, 33 of the children participated in exercise challenge tests every second school week. In the next spring, a part of the children continued the followup for 6 weeks. During the panel studies, air pollution data were collected in a fixed monitoring site close to the center of the town.The primary hypothesis was that airborne particles were associated with pulmonary function. A secondary hypothesis was that the association differed by type of particle.
Summary/Accomplishments (Outputs/Outcomes):
An aerosol monitor computed the number concentration of particles in 12 size ranges from ultrafine to 10 µm. In addition, PM10 and black smoke (BS) were measured. Black smoke is predominantly from local sources and usually dominated by traffic. In addition, the elemental composition of PM filters was analyzed for part of the winter and all of the spring diary period, using inductively coupled plasma mass spectrometry. This was used to aid in the source apportionment, and to identify the toxicity of specific elements.An initial analysis of the 1994 diary identified a weak, but significant, association between airborne particles (PM10) and PEF (Timonen and Pekkanen, 1997). During the study period, the mean daily concentration of particulate air pollution (PM10) was 18 µg/m3 in the urban area, and 13 µg/m3 in the suburban area. Lagged concentrations of PM10, BS, and NO2 were significantly associated with declines in morning PEF among asthmatic children. The regression coefficient (x10) for a 2-day lag of PM10 was -0.911 (SE, 0.386) in the urban area, and -1.05 (0.596) in the suburban area. Among children with cough only, PM10, BS, and NO2 were not significantly associated with PEF.
To test the hypothesis that the effects were predominantly with the ultrafine particles, we compared the effects of daily variations in particles of different sizes on PEF during a 57-day followup of the 39 asthmatic children aged 7-12 years. The main source of particulate air pollution in the area was traffic. In addition to the measurements of PM10 and BS concentrations, an electric aerosol spectrometer was used to measure particle number concentrations in six size classes ranging from 0.01 to 10.0 microns. Daily variations in BS and particle number concentrations in size ranges between 0.032 and 0.32 micron and between 1.0 and 10.0 microns were highly intercorrelated (correlation coefficients about 0.9). Correlations with PM10 were somewhat lower (below 0.7). All these pollutants also tended to be associated with declines in morning PEF. However, the only statistically significant associations were observed with PM10 and BS. Different time lags of PM10 also were most consistently associated with declines in PEF. These results have been published (Pekkanen, et al., 1997). This failure to distinguish particles of any particular size range may have been because of the high correlation among all of the particle measures, the limited number of days for analysis (57 days), and the particle counts that were only available for the urban and not suburban children.
We were able to examine the effect of indoor particles, however. We found that environmental tobacco smoke was associated with increased bronchodilator use, increased cough, and decreased peak flow. These results have been published (Schwartz, et al., 2000).
An initial attempt was used to identify source contributions using factor analyses and to examine their association with peak flow. The major limitation of this analysis was that only the children in the Kuopio area, and not in the suburban area, could be included. Because of the strong association of ultrafine particle counts with traffic density, we did not feel it would be appropriate to assign the particle count measurements to the children in the suburban community. This reduced the sample size, and therefore limited our power.
The first step in our analysis was to perform a factor analysis to identify individual sources. In addition to particle count data in the 12 different size ranges, we included PM10 and BS in the factor analysis. We then examined the effect of adding individual elements that might help distinguish sources. Aluminum is predominantly from soil and was chosen to help identify crustal particles. Sulfur is predominantly from longer range transport particles and was chosen to help separate out that source. In addition, several metals were chosen because they might help identify industrial sources.
The factor analysis used the Procrustes method, which has two major advantages. First, it allows us to set targets for individual factors. Hence, we can use our existing knowledge of which components are strongly associated with a given source to improve the separation of the source components. Equally important, this method uses an oblique rather than an orthogonal factor rotation. Because the individual sources are not, in fact, orthogonal, this approach makes much better sense. Finally, the Procrustes method minimizes the appearance of negative loading, which makes little sense for physical data.
We chose to estimate four factors, and then regressed PM10 on the standardized factor scores for each factor for each day, to see how much of the variation in total particle mass was explained by the four factors. Table 1 shows the results of our initial factor analysis. Using 15 components, we estimated four factors. Factor 1 in this analysis has high standardized loading for ultrafine particles, and essentially no loading for PM10 or for aluminum. This would seem to represent the fresh ultrafine source. Factor 2 has loadings that peak in the .1 to 1 m range, and a high loading of black carbon. These clearly represent fine combustion particles, with the high black carbon loading indicating a predominance of traffic as the source. The third factor also is loaded on the fine particles, but shifted somewhat to the larger end of the size range. It has little black carbon, more PM10 loading, and some loading of aluminum. This probably represents fine particles of non-traffic origin, possibly including some crustal intrusion. Finally, the last factor has high loadings of aluminum and PM10, and substantial weighting in the size range of 3 to 7 µm, indicating particles of crustal origin.
Table 1. Rotated factor pattern (standardized regression coefficients)
Factor 1 | Factor 2 | Factor 3 | Factor 4 | |
EAS1 N of particles in size .010-.018 um | 1.15435057 | -0.1006731 | 0.03481171 | -0.1003874 |
EAS2 N of particles in size .018-.032 um | 0.68983638 | 0.22815169 | -0.0072277 | 0.15093211 |
EAS3 N of particles in size .032-.056 um | 0.19683269 | 0.51661981 | -0.0307371 | 0.35826999 |
EAS4 N of particles in size .056-.100 um | 0.00833829 | 0.74211039 | 0.0559908 | 0.31842451 |
EAS5 N of particles in size .100-.178 um | -0.0685846 | 0.94360785 | 0.25149599 | 0.16481786 |
EAS6 N of particles in size .178-.316 um | -0.0481868 | 1.06674876 | 0.63221843 | -0.1252674 |
EAS7 N of particles in size .316-.562 um | 0.04118721 | 1.02087218 | 0.76112432 | -0.275065 |
EAS8 N of particles in size .562-1.00 um | 0.01055691 | 0.95567776 | 0.56707554 | 0.04815543 |
EAS9 N of particles in size 1.00-1.778 um | 0.04637262 | 0.54882289 | 0.09970801 | 0.50711545 |
EAS10 N of particles in size 1.778-3.162 um | 0.12513353 | 0.42994898 | 0.08359134 | 0.54919105 |
EAS11 N of particles in size 3.162-5.624 um | 0.05834429 | 0.36479168 | 0.05327662 | 0.65833594 |
EAS12 N of particles in size 5.624-10.0 um | -0.1245305 | 0.32521452 | -0.0089659 | 0.77469976 |
PM10 PM10 | 0.00386146 | 0.18049273 | 0.63249378 | 0.85477117 |
BC Black Carbon | -0.0285995 | 0.88012202 | 0.16014258 | 0.18026249 |
ALX Aluminum | -0.0185818 | -0.5537574 | 0.43485643 | 1.38488908 |
When PM10 was regressed on these four factor scores, a high R2 (.96) was obtained. Adding lead, nickel, titanium, vanadium, or potassium did not improve the model fit further. However, adding sulfur did improve the fit of the model. This is shown in Table 2.
Once again, the individual factors can be associated with different sources. Factor 1 again represents ultrafines with little mass. They have some sulfur, reflecting the sulfur content of transportation (mostly diesel) fuel. Factor 2 again represents fine particles, with black carbon, again predominantly from traffic. Factor 3 represents slightly larger fine particles with no carbon, but a high sulfur loading. These represent long-range transport sulfates. Factor 4 represents the crustal particles.
Having developed these source models, we then examined how they were associated with children's peak flow. In contrast to the previous analyses where individual size ranges were examined, the factor analysis did provide some evidence on the effects of specific sources. Table 3 shows the results of that analysis. The first factor, representing the very small particles from traffic sources, was a significant predictor of morning PEF in children. No association was seen for the other sources.
Table 2. Rotated factor pattern (standardized regression coefficients)
Factor 1 | Factor 2 | Factor 3 | Factor 4 | |
EAS1 N of particles in size .010-.018 um | 1.00419305 | -0.2531288 | 0.17642425 | 0.23195838 |
EAS2 N of particles in size .018-.032 um | 0.71961363 | 0.13396093 | 0.07128672 | 0.2559892 |
EAS3 N of particles in size .032-.056 um | 0.35611365 | 0.49612076 | -0.0428987 | 0.25897418 |
EAS4 N of particles in size .056-.100 um | 0.09620836 | 0.76193609 | -0.0654013 | 0.21162053 |
EAS5 N of particles in size .100-.178 um | -0.1570643 | 0.97754002 | 0.00264416 | 0.14985295 |
EAS6 N of particles in size .178-.316 um | -0.2976367 | 0.98919568 | 0.35964988 | 0.02917345 |
EAS7 N of particles in size .316-.562 um | -0.1968449 | 0.84311673 | 0.57588465 | -0.054627 |
EAS8 N of particles in size .562-1.00 um | -0.1103559 | 0.84966554 | 0.37444521 | 0.13969215 |
EAS9 N of particles in size 1.00-1.778 um | 0.15268065 | 0.55846579 | -0.0099839 | 0.39354968 |
EAS10 N of particles in size 1.778-3.162 um | 0.20070268 | 0.45053501 | -0.0290961 | 0.4606332 |
EAS11 N of particles in size 3.162-5.624 um | 0.2177604 | 0.37966911 | -0.0255509 | 0.50810108 |
EAS12 N of particles in size 5.624-10.0 um | 0.25643124 | 0.29586726 | 0.0232555 | 0.48174276 |
PM10 PM10 | -0.0005953 | 0.08746818 | 0.47589323 | 0.84291969 |
BC Black Carbon | 0.07539124 | 0.83846082 | 0.07382722 | 0.09376431 |
ALX Aluminum | -0.1409572 | -0.4449261 | 0.1383638 | 1.33106237 |
SX | 0.36951178 | 0.01539951 | 1.04894307 | -0.2315735 |
Table 3. Results from mixed model analysis, controlling for child,
temperature, humidity, holidays,
weekend, ETS exposure, pollen, and time
trend
Estimate | Error | DF | t Value | Pr > |t| | |
Factor1 | -1.5319 | 0.7579 | 2013 | -2.02 | 0.0434 |
Factor2 | 0.4047 | 0.6750 | 2013 | 0.60 | 0.5489 |
Factor3 | -0.2129 | 0.4493 | 2013 | -0.47 | 0.6357 |
Factor4 | 0.6102 | 0.7052 | 2013 | 0.87 | 0.3870 |
An alternative approach to addressing the size issue resulted from the time course of particle concentrations by size. Spring in Finland is marked by a sudden increase in resuspended road dust. Forty-nine children with chronic respiratory symptoms, aged 8-13 years, were followed daily for 6 weeks in spring 1995, in Kuopio. Daily concentrations of particulate material with a 50 percent cutoff aerodynamic diameter < or = 10 microns and < or = 2.5 microns (PM10 and PM2.5, respectively), black carbon, and the number concentrations of particles from 0.01-10 microm diameter were measured. During the study period, PM10 was mainly resuspended soil and street dust, and the concentration was estimated using aluminum content of PM10 samples. No consistent effect of particles was found as the associations varied by lag. Of the lags examined, only 1-day lagged PM2.5 was statistically significantly associated with morning PEF (beta=-1.06, SE=0.52 [per interquartile increase in pollutant]). Evening PEF was significantly associated with the 1-day lagged number of particles in the size range 0.1-1.0 micron (beta=-1.56, SE=0.72). One-day lagged PM10, PM2.5-10, PM2.5, and resuspended PM10, and 4-day average of PM2.5 were significantly associated with increased risk of cough. Given the short duration of the study, separating the effects of different types of particles was difficult. Nevertheless, the lack of association with coarse mass particles suggests that only the smaller particles are responsible (Tiittanen, et al., 1999).
The effects of daily variation of ambient air pollution during cold wintertime days on lung function and exercise-induced bronchial responsiveness were studied among 33 primary school children with chronic respiratory symptoms. Every second school week from February to April, 1994, the children took part in an exercise challenge test conducted outdoors at the schoolyard. Each child participated at most five times in the test. Respiratory resistance and spirometric lung function were measured indoors before the exercise, and 3 and 10 minutes thereafter. Daily mean levels of PM10, BS, NO2, CO, SO2, and particle number concentration were monitored at a fixed monitoring site. Daily variation in ambient air pollution was not associated with enhanced bronchial responsiveness. However, impairment of pre-exercise lung function was related to increased levels of ambient air pollutants mostly originating from combustion processes. The reductions in FVC and FEV1 were 0.5 percent and 0.6 percent, respectively, for each 10 µg/m3 increase in BS concentration (lag 2).
A receptor modeling study was conducted on the air pollution data collected during the winter and spring of 1994. Daily mean concentrations of PM10, sulfur dioxide, CO, and BS were measured. Elemental concentrations of PM10 samples for 38 days were analyzed by ICP-MS. The main sources and their contributions to measured concentrations of PM10 particles were solved by receptor modeling using a factor analyses-multiple linear regression (FA-MLR) model. Because a dust episode was very strong during two sampling days, the factor analysis was strongly influenced by this episode and did not give main factors. The factor analysis, when the two episode days were omitted, gave credible factors related to the sources in the study area. The four major sources and the estimated contributions to the average PM10 concentration of 27.2 µg/m3 were: soil and street dust 46-48 percent, heavy fuel oil burning 12-18 percent, wood burning ca. 11 percent, and unidentified sources 15-25 percent. However, during spring dust episode days, with maximum PM10 concentration of 150 µg/m3, the main source of PM10 was soil (Hosiokangas, et al., 1999).
Journal Articles on this Report : 6 Displayed | Download in RIS Format
Other project views: | All 6 publications | 6 publications in selected types | All 6 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Hosiokangas J, Ruuskanen J, Pekkanen J. Effects of soil dust episodes and mixed fuel sources on source apportionment of PM10 particles in Kuopio, Finland. Atmospheric Environment 1999;33(23):3821-3829. |
R825265 (Final) |
Exit Exit |
|
Pekkanen J, Timonen KL, Ruuskanen J, Reponen A, Mirme A. Effects of ultrafine and fine particles in urban air on peak expiratory flow among children with asthmatic symptoms. Environmental Research 1997;74(1):24-33. |
R825265 (Final) |
Exit Exit |
|
Schwartz J, Timonen KL, Pekkanen J. Respiratory effects of environmental tobacco smoke in a panel study of asthmatic and symptomatic children. American Journal of Respiratory and Critical Care Medicine 2000;161(3 Part 1):802-806. |
R825265 (Final) |
Exit Exit Exit |
|
Tiittanen P, Timonen KL, Ruuskanen J, Mirme A, Pekkanen J. Fine particulate air pollution, resuspended road dust and respiratory health among symptomatic children. European Respiratory Journal 1999;13(2):266-273. |
R825265 (Final) |
Exit Exit |
|
Timonen KL, Pekkanen J. Air pollution and respiratory health among children with asthmatic or cough symptoms. American Journal of Respiratory and Critical Care Medicine 1997;156(2 Part 1):546-552. |
R825265 (Final) |
Exit Exit Exit |
|
Timonen KL, Pekkanen J, Tiittanen P, Salonen RO. Effects of air pollution on changes in lung function induced by exercise in children with chronic respiratory symptoms. Occupational and Environmental Medicine 2002;59(2):129-134. |
R825265 (Final) |
Exit Exit |
Supplemental Keywords:
ultrafine particles, PM10, PM2.5, coarse particles, dust, traffic pollution, health., RFA, Health, Scientific Discipline, Air, particulate matter, Epidemiology, Risk Assessments, Susceptibility/Sensitive Population/Genetic Susceptibility, Allergens/Asthma, Children's Health, Atmospheric Sciences, genetic susceptability, ambient aerosol, ambient air quality, asthma, particle size, particulates, urban air, environmental monitoring, lungs, health effects, peak flow, sensitive populations, cardiopulmonary responses, chemical characteristics, fine particles, human health effects, respiration, infants, exposure, age-related differences, airway disease, respiratory problems, soluble transition metals, air pollution, children, airway inflammation, human exposure, atmospheric transport, chronic health effects, lung inflammation, airborne pollutants, inhalation, lung dysfunction, Acute health effects, inhaled, harmful environmental agents, environmentally caused disease, inhaled particles, atmospheric chemistry, respiratory, ultrafine particles, exposure assessment, air quality, environmental hazard exposuresProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.