Published ahead of print on February 16, 2006, doi:10.1164/rccm.200509-1393OC
© 2006 American Thoracic Society doi: 10.1164/rccm.200509-1393OC
Particulate Levels Are Associated with Early Asthma Worsening in Children with Persistent DiseaseDivision of Allergy and Immunology, Department of Pediatrics; and Division of Biostatistics, National Jewish Medical and Research Center, Denver, Colorado Correspondence and requests for reprints should be addressed to Nathan Rabinovitch, M.D., National Jewish Medical and Research Center, 1400 Jackson Street, Denver, CO 80206. E-mail: rabinovitchn{at}njc.org
Rationale: Ambient particulate concentrations have been associated with variable physiologic effects in children with persistent asthma taking controller medications. Objective: To determine whether exposure to particulate matter has immediate effects on asthma control in children with persistent disease. Methods and Measurements: In a school-based cohort, 73 children, primarily with moderate and severe asthma, were followed daily over one or two winters (2001–2002, 2002–2003) in Denver, Colorado. The association among ambient fine particulate, bronchodilator use, and urinary leukotriene E4 levels was assessed. Results: Daily concentrations of fine particulate peaked in the morning hours when children were commuting to school. In a multivariable analysis that controlled for meteorology, time trends, and upper respiratory infections, an increase of one interquartile range in morning maximum fine particulate levels was related to an average increase of 3.8% in bronchodilator usage at school (95% confidence interval [CI], 0.2–7.4; p = 0.04). Children with severe asthma demonstrated significantly stronger associations (8.1% increase; 95% CI, 2.9–13.4; p = 0.003) than those with mild/moderate disease (1.6% increase; 95% CI, –2.2–5.4; p = 0.41; p = 0.03 for difference between groups). Morning maximum fine particulate levels were also associated with urinary leukotriene E4 measured during school hours (average increase of 6.2% per interquartile range increase; 95% CI, 1.9–10.5; p = 0.006). These associations were not discernable when 24-h averaged concentrations were used. Conclusions: Peak concentrations of ambient fine particulate are associated with early increases in bronchodilator use and urinary leukotriene E4 levels among children with persistent asthma, despite the use of controller medications.
Key Words: air pollution • children • LTE4 • PM2.5 • severe asthma
There is general consensus that the adverse health effects of exposure to ambient air pollution are not evenly distributed among the general population but are limited to, or magnified in, susceptible population subgroups. These subgroups largely consist of those with preexisting chronic illness, including asthma. A number of panel studies have demonstrated a relationship between levels of airborne particulate matter less than 10 or 2.5 µm in aerometric diameter (PM10 or PM2.5) and asthma admissions (1–3). Panel studies primarily assessing patients with mild, intermittent asthma requiring no medications have reported associations between increased particulate and acute disease severity indices, such as pulmonary function and symptoms (4–6). Individuals with persistent asthma requiring the chronic use of These studies are limited in a number of ways. First, the vast majority of studies use particulate concentrations and health outcomes averaged over a 24-h period. This may make it difficult to recognize health outcomes related to brief exposures to air pollution spikes if these outcomes are completely reversible with rescue medications (12). Second, none of the panel studies has observed a relationship between particulate air pollution and biological mediators related to asthma worsening in children taking controller medications. Therefore, it remains unclear whether increased particulate air pollution leads to increases in asthma mediators with masking of health end-points by medication use or whether biological or physiologic effects occur due to medication usage or some other factor related to severity. Recognition that increases in biological mediators are associated with particulate levels suggests that air pollution levels could affect the patient with poorly controlled asthma despite the use of controller medications and support a causal rather than an associative relationship between ambient particulate and health outcomes (13). In this study, we examined the relationship between particulate air pollution and asthma, taking advantage of a well-defined group of children with asthma, primarily with moderate or severe disease, who were monitored on an almost daily basis over 2 yr. In the first year of the study, we tested for signs of asthma worsening immediately after exposure to fine particulate concentrations (i.e., within 5–7 h). The second year of the study built on this initial critical information to determine if clinical associations were consistently observed from year to year and if ambient particulate concentrations were related within the same time interval to levels of an asthma-related biological mediator.
Study Subjects Children, aged 6 to 13 yr, who attended the Kunsberg School at the National Jewish Medical and Research Center and who had physician-diagnosed asthma, were studied over two consecutive winters. In the first year of the study (Year 1), pollution and health outcomes data were collected on school days in 37 children from November 19, 2001, to March 18, 2002. In Year 2 of the study, 57 schoolchildren were monitored over a slightly longer period (October 15, 2002, to May 22, 2003). Ethical and scientific approval for each year was obtained from the National Jewish Center's Institutional Review Board.
Medication Use
Leukotriene E4
Upper Respiratory Infections
Ambient Air Monitoring and Meteorology
Analysis Natural log urinary LTE4 levels were modeled as a function of morning maximum or mean PM2.5 using a linear mixed model, with a spatial exponential covariance structure to account for within-subject repeated measures over time (16). Height and URI were significant and therefore were included in the model. Meteorologic and time-trend variables were initially included but were highly insignificant and were not included in final runs. SAS, PROC MIXED (version 9.1), was used to carry out the analysis. All reported p values are based on two-sided tests. The interquartile range (IQR) for a distribution is the 75th percentile minus the 25th percentile. This statistic was used to standardize pollutant slope estimates. For further details of statistical methods, see the online supplement.
Demographics and Asthma Severity Table 1 summarizes demographic and asthma severity characteristics based on a screening questionnaire administered to the parents before each year of the study. Just over half the panel were African American, and approximately half were admitted into an intensive care unit for asthma at least once. Almost 90% of the children had experienced an asthma exacerbation during the previous year. Based on the frequency of nighttime symptoms, most of these children were classified as moderate or severe by National Asthma Education and Prevention Program (NAEPP) guidelines (17). Average daily use of school inhalers was approximately 2 puffs/d in both years.
Pollutant Concentrations and Meteorologic Variables FRM daily mean and maximum concentrations were well within National Ambient Air Quality Standards (Table 2). The IQR for morning maximum PM2.5 was 12.0 µg/m3 in both years. For morning mean PM2.5, the IQRs were 5.6 µg/m3 in Year 1 and 6.4 µg/m3 in Year 2. TEOM concentrations tended to be lower than concentrations measured by the proximal FRM, but the two were highly correlated (r = 0.87 in both years). In previous years, high correlations were observed between the community FRM and local (outside the school) PM2.5 concentrations (r = 0.92). (A local FRM monitor was not available for use at the school during the study years.) Local concentrations were moderately correlated with daily personal exposures to ambient PM2.5 (r = 0.60), based on sulfate data, which were available in Year 2 of the study. Strand and colleagues (18) describe how pollutant health effects estimates based on a central monitor can be calibrated to reflect personal ambient PM2.5 exposures by incorporating exposures to sulfate.
Maximum PM2.5 Concentrations Occur in the Morning Figure 1 illustrates the daily pattern of PM2.5 peaks and troughs. During an average 24-h period, ambient PM2.5 concentrations were highest at approximately 7:30 A.M. to 10:00 A.M., with a smaller peak occurring in the late afternoon. Because this was primarily a wintertime study, much of the children's time was spent indoors where children were exposed to low levels of ambient PM2.5 (18). Therefore, peak exposures likely occurred during the morning (7:00 A.M. to 8:00 A.M.) commute to school when children as a group were outside of their homes.
Morning PM2.5 Is Associated with Increased Morning Medication Use There were 2,385 (subject-day) records available for analysis in Year 1 and 5,928 in Year 2. All estimates reported below are standardized per 25th through 75th IQR increase of the pollutant. For morning maximum PM2.5, medication use increased in Year 1 by 4.0% (95% confidence interval [CI], 0.5–7.6; p = 0.02). In Year 2, the increase was 2.3% (95% CI, 0.7–4.0; p = 0.009). The combined (Years 1 and 2) estimate was 2.6% (95% CI, 0.9–4.2; p = 0.002; Table 3). Using 24-h averages from the TEOM or FRM monitors, no significant effects were observed on the same day or up to 3 d later (Table 4 and Figure 2). We also explored asthma severity, age, and height as potential pollutant effect modifiers (see online supplement for methods). Over the two study years, children with severe asthma had higher Doser medication increases (3.9%; 95% CI, 1.1–6.8; p = 0.006) than those with mild/moderate asthma (1.9%; 95% CI, –0.2–4.1; p = 0.07). Estimates for patients with severe asthma increased further when days with URI symptoms were controlled in the analysis (8.1%; 95% CI, 2.9–13.4; p = 0.003), but no significant change was seen in patients with mild/moderate asthma (p = 0.03 for difference between groups) (Table 5 and Figure 3). Age and height were not significant effect modifiers. Estimates based on morning mean PM2.5 were similar to those observed for morning maximum PM2.5 (Tables 3 and 5 and Figure 2).
Morning PM2.5 Is Associated with Increased Urine LTE4 Levels The geometric mean of the subject average LTE4 values within the Year 2 study period was 83.6 pg/mg creatinine (geometric SD, 1.5 pg/mg). After controlling for days with URI symptoms, urinary LTE4 levels were increased by 6.2% for each IQR increment in morning maximum PM2.5 (95% CI, 1.9–10.5; p = 0.006), based on 388 records. Estimates based on morning mean PM2.5 were similar. No significant effects were observed on the same day or up to 3 d later, based on 24-h averaged concentrations from the TEOM or FRM monitors (Table 6 and Figure 4). We also explored the same potential pollutant effect modifiers as in the medication models (asthma severity, age, and height). Increases in LTE4 tended to be greater in smaller children (p = 0.06 and 0.11 for height x pollutant interaction for morning mean and maximum PM2.5, respectively; Table 7). LTE4 and fine particulate relationships were not dependent on severity or age (see online supplement for methods.)
This study examined the timing of particulate associations with disease control in children with primarily moderate or severe asthma taking controller medications. The first year of this study determined that the interval between exposure and the initiation of health effects occurred within the first few minutes or hours after exposure. In Year 2, this observation of an early health outcome was similarly observed with almost identical estimates of effect. These effects were strongest in children with more severe asthma. Within this same time interval, urinary LTE4 levels increased in association with increases in fine particulate, suggesting that, in children with persistent asthma, particulate exposure may lead to early release of mediators related to asthma worsening. If one assumes that exposure to higher levels of particulate occurs or begins to occur during the daily particulate peak when children are commuting to school, then the duration of time between morning exposure and onset of health effects was no more than 5 h for LTE4 (data collected between 11:00 A.M. and 1:00 P.M.) and no more than 7 h for Doser use (school Doser used until 2:30 P.M.). It is possible that health effects began minutes after exposure or closer to the time that measurements were taken (i.e., hours after exposure). The interval between exposure and changes in health, therefore, seems to be within a fraction of a day, but further studies are needed to define a more precise time interval. In a similar cohort of children, health effects were not discernable when data were analyzed using conventional models that assess particulate concentrations and health outcomes averaged over a 24-h period (7). In the previous study (which included data from Year 1 of the present study), no significant associations were observed between particulate concentrations and daily lung function, asthma symptoms, medication use, and asthma exacerbations. Twenty-four–hour particulate averages were used in this earlier study, as they have been in the majority of previously published studies (1–10), and medication usage was averaged over a 24-h period. In the present analysis, the association was examined using shorter observations periods for the pollutant and health outcomes. This allowed for a determination of associations that were not previously observed when using more diluted or averaged measures. The timing of particulate association with acute asthma worsening has been explored in previous studies. A number of studies have reported maximal effects when averaged over 3 or 5 d (19). This has sometimes been interpreted as evidence for a lag effect possibly indicative of inflammation. However, most of these studies (3–10, 19–22) used current-day effects into this average, suggesting that effects could be occurring partially concurrently with exposure. Unlike some other studies where the use of lagged moving averages increased effect estimates, associations between particulate and bronchodilator use were not apparent with longer averaging times. This may be due to early interventions by caretakers or by the children themselves, including decreases in outdoor activity or the addition of other medications. Alternatively, the differences between the present study and others may be related to the nature of particulate spikes in the Denver area, which generally are not sustained for longer than 24–48 h. Health outcomes reported in other studies may be persistent due to airway inflammation or may reflect transient repetitive responses to sustained pollution levels. This temporal pattern of brief exposure and early asthma-related outcomes is consistent with reports that decreases in pulmonary function and increases in exhaled nitric oxide occur within 1 to 4 h after exposure (23, 24). Similarly, cardiovascular morbidity (25) is associated with relatively brief exposures to traffic emissions. If so, reliance on 24-h averages by investigators and regulators may obscure important relationships between air pollution and cardiopulmonary morbidity if outcomes are reversible or if peak exposures are not well correlated with 24-h concentrations. Stimulation of macrophages, mast cells, epithelial cells, eosinophils, and endothelial cells may lead to the secretion of cysteinyl leukotrienes C4, D4, and the stable end-product LTE4 (26). This response does not seem to be affected by the use of inhaled corticosteroids (27). The significance of cysteinyl leukotrienes as key mediators and modulators in the pathogenesis of asthma is widely recognized. Inhaled LTD4 is a bronchoconstrictor 10,000 times more potent than histamine or methacholine (28). In addition to their potent bronchoconstrictive properties, cysteinyl leukotrienes induce other pathophysiologic responses characteristic of asthma, including tissue edema, mucous secretion, and increased airway responsiveness to histamine (26). They also play an important role in chronic changes due to airway remodeling (26). The inflammatory response to allergen challenge has been divided into an early response and a late response. Elevations in urinary LTE4 levels occur during early allergen-induced bronchoconstriction (29). Studies also report increases in LTE4 levels in association with other stimuli, such as exercise, that are primarily associated with an immediate bronchospastic response (30, 31). LTE4 levels are increased in the sputum 30 min after exercising, and urinary LTE4 levels are elevated within 4 h, returning to baseline within 24 h after exercise (30). A number of studies have reported that blocking leukotriene receptors with a leukotriene receptor antagonist such as montelukast attenuates but does not abolish exercise-induced lung-function declines (31, 32). A recent study reported that montelukast provided greater protection against bronchoconstriction after exercise during high PM1 than low PM1 exposure (approximately 90% vs. approximately 35%) (33). Similarly, Gong and colleagues reported that in a challenge setting, montelukast blocked the effects of sulfur dioxide on lung function declines after exercise in volunteers with asthma (34). These studies suggest that air pollution exposure can augment exercise-induced leukotriene release and effects on lung function. In the present study, LTE4 levels were increased in association with morning levels of ambient particulate but not with 24-h or lagged measurements, suggesting release of leukotrienes relatively early after exposure. This association was stronger in smaller children, which is consistent with studies reporting stronger pollutant associations with asthma admissions in younger children (3), perhaps related in part to higher ventilation rates and relative concentrations of particulate deposition into smaller lung volumes. Asthma severity was not related to changes in LTE4 associated with PM2.5, although health effects were. This suggests that interventions that increase asthma control may blunt the effects of air pollution by decreasing airway hyperresponsiveness to mediators such as LTE4. A randomized study will soon be performed to determine whether LTE4 levels are directly related to pollution-associated health outcomes in this panel. Because leukotrienes need to be released into the general circulation to be secreted into the urine, these findings may be relevant to nonrespiratory disease processes. A number of epidemiologic studies, for example, have demonstrated relationships between ambient levels of particulate matter and cardiopulmonary mortality (35, 36). Previous studies have reported that cysteinyl leukotrienes are markedly elevated in urine of patients experiencing a myocardial infarction or acute chest pain due to unstable angina (37). Diseased arteries from heart transplant patients have high levels of leukotriene receptors and enzymes involved in leukotriene synthesis, such as 5-lipoxygenase activating protein. Polymorphisms in the 5-lipoxygenase activating protein gene, which increases leukotriene production, have been associated with increased risk of heart attack and stroke (38). This study used concentrations from the TEOM for analyses because hourly data, and hence morning statistics, were available. (For the FRM, only 24-h concentrations were available.) Mass measurements by the TEOM have been shown to be precise and well correlated with FRM manual filter measurements (39). However, TEOM measurements typically underestimate concentrations in the colder months due to loss of semivolatile compounds from the heated sample filter (40, 41). These findings were verified in our study. Specifically, a 1-µg/m3 increase in 24-h PM2.5 concentration based on the TEOM was associated with average increases of 1.87 and 1.29 µg/m3 for the FRM in Years 1 and 2, respectively, based on simple linear regression. Differences between years were probably related to Year 2 extending into the warmer spring months (see mean temperatures in Table 2). Although morning FRM data were not available, regression calibration methods (18) could be applied to approximate morning mean or maximum PM2.5 health effect estimates based on this monitor if the slope relationships between TEOM and FRM monitors for the morning period were assumed to be similar to those observed for the entire day. In particular, we would expect health effect estimates per µg/m3 increase to be 1.87 and 1.29 times smaller for the FRM monitor in Years 1 and 2, respectively, but because the IQRs for FRM concentrations were 1.8 and 1.3 times greater than the TEOM concentrations in Years 1 and 2, respectively, effects per IQR increase would be expected to be similar between monitors. In earlier panel studies, it has been difficult to separate the issue of asthma severity from medication use because children taking controller medications are presumed to have worse disease than those not on these medications. The present study may help to dissociate these factors by examining the issue of severity in a panel of children already taking controller medication. The stratification into mild, moderate, and severe was based on the study by Colice and colleagues (17), which assessed the sensitivity of each of the severity parameters in the National Asthma Education and Prevention Program expert panel II criteria. These authors reported that the most sensitive parameter for severity was the frequency of nighttime symptoms. We therefore used this parameter (answered by parents before study initiation) to stratify the panel into severity categories. Children with "moderate" disease were those who had nighttime symptoms at least twice per week, and those classified as "severe" had symptoms at least four nights per week. Using this classification, associations between medication use and morning particulate concentrations were strongest among children with more severe asthma. These observations suggest that in children with poorly controlled asthma, air pollution effects may be observed despite the use of daily controller medications and that panel studies that do not include such children may underestimate potential health effects. Given these observations, we hypothesize that particulate-related asthma morbidity would be greatest in smaller children with poorly controlled disease. In summary, consistent associations between peak ambient fine particulate concentrations in Denver and early increases in medication use were observed in urban children with persistent asthma over a 2-yr period. Peak particulate concentrations were related to early increases in urinary LTE4, an important mediator involved in asthma pathogenesis. These associations were not observed using FRM measurements. Although we cannot rule out confounding from unmeasured covariables, it seems that fine particulate concentrations well below National Ambient Air Quality Standards can lead to rapid-onset increases in mediator release and early asthma worsening in urban children with moderate to severe disease despite the use of daily controller medications.
The authors thank Michele Freas, Karen Musselman, Elaine Cline, and the teachers, students, and staff of the Kunsberg School at National Jewish Medical and Research Center; Pat McGraw, Bradley Rink, and the staff at the Air Pollution Control Division of the Colorado Department of Public Health and Environment for their expertise and technical assistance; Nahid Nazminia, Rachel Wood, and Greg Whitman for asthma severity data collection; Steve Dutton for exposure data collection and expertise; Jay Westcott for performing the LTE4 assays; and Diana Nabighian and Gretchen Hugen for help in preparation of the manuscript.
Supported by EPA grant R825702, Thrasher Research Fund grant 02816, and Colorado Tobacco Research Program grant 2R-020. This article contains an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200509-1393OC on February 16, 2006 Conflict of Interest Statement: None of the authors have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form September 7, 2005; accepted in final form February 13, 2006
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
2-agonists or daily inhaled corticosteroids do not seem to be susceptible to the effects of air pollution exposure (
