Comments
Abstract
Rationale: Limited prior data suggest an association between traffic-related air pollution and incident asthma in adults. No published studies assess the effect of long-term exposures to particulate matter less than 2.5 μm in diameter (PM2.5) on adult incident asthma.
Objectives: To estimate the association between ambient air pollution exposures (PM2.5 and nitrogen dioxide, NO2) and development of asthma and incident respiratory symptoms.
Methods: The Sister Study is a U.S. cohort study of risk factors for breast cancer and other health outcomes (n = 50,884) in sisters of women with breast cancer (enrollment, 2003–2009). Annual average (2006) ambient PM2.5 and NO2 concentrations were estimated at participants’ addresses, using a national land-use/kriging model incorporating roadway information. Outcomes at follow-up (2008–2012) included incident self-reported wheeze, chronic cough, and doctor-diagnosed asthma in women without baseline symptoms.
Measurements and Main Results: Adjusted analyses included 254 incident cases of asthma, 1,023 of wheeze, and 1,559 of chronic cough. For an interquartile range (IQR) difference (3.6 μg/m3) in estimated PM2.5 exposure, the adjusted odds ratio (aOR) was 1.20 (95% confidence interval [CI] = 0.99–1.46, P = 0.063) for incident asthma and 1.14 (95% CI = 1.04–1.26, P = 0.008) for incident wheeze. For NO2, there was evidence for an association with incident wheeze (aOR = 1.08, 95% CI = 1.00–1.17, P = 0.048 per IQR of 5.8 ppb). Neither pollutant was significantly associated with incident cough (PM2.5: aOR = 0.95, 95% CI = 0.88–1.03, P = 0.194; NO2: aOR = 1.00, 95% CI = 0.93–1.07, P = 0.939).
Conclusions: Results suggest that PM2.5 exposure increases the risk of developing asthma and that PM2.5 and NO2 increase the risk of developing wheeze, the cardinal symptom of asthma, in adult women.
Previous research, primarily in children, has demonstrated an association between air pollution and asthma exacerbation, and some evidence suggests a role in asthma incidence in children. The few studies on incident asthma in adults have had less precise exposure measurements and modest sample sizes, and none has directly examined exposure to particulate matter less than 2.5 μm in diameter (PM2.5).
The results of this large nationwide cohort study suggest that ambient PM2.5 exposure or other related exposures may be involved in the development of respiratory symptoms, particularly wheeze, and incident asthma in women.
The causes of asthma are not well understood, and research on the etiology of adult-onset asthma, outside of specific occupational settings, is limited. There is increasing literature suggesting a role for ambient air pollution in the development of asthma in children, but there are few studies in adults (1). Two reviews have cited the need for more research on the role of air pollution in the development of asthma in adults (1, 2) and specifically for large cohorts with improved exposure assessment (1). Although nitrogen dioxide (NO2) is typically used as a marker for traffic-related air pollution, both particulate matter and NO2 are created by combustion sources such as vehicle exhaust and both are potentially harmful pollutants. Notably, although there have been studies of PM10 (particulate matter less than 10 μm in diameter) exposure and adult-onset asthma (3, 4), there are no published studies on the association between adult-onset asthma and exposure to PM2.5. Particles of various sizes may act differently with respect to potential asthmagenic characteristics such as lung penetration (5) and oxidative potential (6), and PM2.5 is often not closely correlated with PM10 (7). Therefore, analyses of PM2.5 as a possible cause of adult-onset asthma represent an important contribution to existing literature.
Adult-onset asthma is common and is an important public health problem. It has unique features that differentiate it from childhood-onset asthma. For example, the prevalence of asthma is higher in adult women (8.8%) than men (5.8%) in the United States (8), but before puberty asthma is more common in boys than girls (9). Adult-onset asthma is less likely than childhood asthma to be atopic (10). Ongoing research suggests that childhood-onset and adult-onset asthma represent at least two distinct diseases (11, 12) and thus observed associations between air pollution and development of childhood asthma may not necessarily extend to the adult phenotype.
To address the paucity of data on the potential role of air pollution in the development of adult-onset asthma, we evaluated the association between PM2.5 and NO2 exposure and incident asthma, incident chronic cough, and incident wheeze in the National Institute of Environmental Health Sciences (NIEHS, Research Triangle Park, NC) Sister Study, a nationwide study of more than 50,000 U.S. women.
Some of the results of these studies have been previously reported in the form of an abstract (13). A previous version of this article exists in the form of an unpublished master’s thesis (14).
The Sister Study involves a large cohort of women across the United States, who have at least one sister with a diagnosis of breast cancer (n = 50,884) but who had not been diagnosed themselves at enrollment (15). Participants were enrolled from August 2003 through March 2009. Each participant underwent a baseline computer-assisted telephone survey at enrollment and completed the first computer-assisted follow-up telephone interview 2–3 years after the baseline survey, depending on enrollment date (March 2008–April 2012).
Both the baseline and follow-up survey included questions about asthma history and symptoms (see the online supplement for details). Instead of individualizing follow-up questionnaires to query new diagnoses with respect to each participant’s personal baseline date, we asked whether they had been diagnosed since a date that was the earliest they could have completed the baseline questionnaire. Specifically, follow-up questions regarding new symptoms were asked with regard to a reference date, which varied by enrollment period and approximated the date of completion of the baseline survey. Because the first follow-up questionnaire was fielded over several years, the reference date changed each year. The time period since actual completion of the baseline questionnaire and follow-up questionnaires is knowable; this average follow-up time was 2.88 years (minimum, 1.54 yr).
Incident asthma was defined on the basis of self-report at follow-up of all three of the following occurring since the reference date: diagnosis of asthma by a doctor, any use of asthma medications, and asthma symptoms. Incident asthma symptoms were defined as self-report of any of the following with onset since the reference date: chronic cough, wheeze, or report of “asthma symptoms” in the 12 months preceding follow-up (additional detail on the survey questions is provided in the online supplement).
For analyses of incident asthma, we excluded prevalent asthmatics on the basis of their report at baseline of ever having had doctor-diagnosed asthma. To reduce the number of undiagnosed patients with prevalent asthma, we additionally excluded subjects who reported both wheeze and chronic cough at baseline (Figure 1).
Figure 1. Study populations and exclusions. Additional exclusions due to missing covariates are not shown.
[More] [Minimize]We identified women who reported, at follow-up, new onset of chronic cough and wheeze, chronic cough, and wheeze since the reference date. For separate analyses of each symptom category, participants were excluded if they reported an asthma diagnosis or the corresponding symptom(s) at baseline.
We estimated air pollution exposures using year 2006 PM2.5 and NO2 concentrations at the participant’s reported primary address at study enrollment. Participant home addresses were geocoded using ArcMap version 10 (ESRI, Redlands, CA), and the residential locations (Figure 2) were used for prediction of outdoor pollutant concentrations. Ambient pollutant concentrations were estimated using geographic covariates in a universal kriging regression on annual averages derived from a national network of air pollution–monitoring stations (using U.S. Environmental Protection Agency [EPA] reference methods) in a model described previously (16). The cross-validated R2 value for the PM2.5 model in year 2006 was 0.88, the same as the year 2000 R2 value reported previously (16). The cross-validated R2 value for the NO2 model in year 2006 was 0.85. We modeled exposure only for the contiguous United States; participants living in Hawaii, Alaska, and Puerto Rico were excluded.
Figure 2. Sister Study baseline home address locations (data for Alaska, Hawaii, and Puerto Rico are not shown).
[More] [Minimize]Analyses were conducted by multivariable logistic regression. Potential confounders were identified on the basis of a priori hypothesized relationships and literature review; the resulting hypothesized causal model is illustrated in Figure 3. The a priori identified confounders were age (continuous), body mass index (<25, 25–30, ≥30), race (black, Hispanic, non-Hispanic white, other), education (less than high school; high school or GED; some college; bachelor’s degree; associate, tech, or nursing degree; master’s or doctoral degree), occupational exposure to either vapor or fumes (ever/never), occupational exposure to dust (ever/never), baseline smoking status (current smokers, current social smokers, former smokers, never-smokers), age started smoking among ever-smokers (age < 20 yr, age ≥ 20 yr), packs per day at baseline (<0.5, 0.5–1, ≥1), smoked since baseline (yes/no), childhood second-hand smoke from primary caregiver (yes/no), any health care coverage (yes/no), dietary fiber consumption per day (continuous in grams).
Figure 3. Causal model of air pollution and asthma demonstrating potential confounders. BMI = body mass index.
[More] [Minimize]To assess the possibility for residual confounding, modeling was performed by a three-stage approach. Minimally adjusted models included age alone. Fully adjusted models included all a priori hypothesized confounders. We additionally adjusted for adult second-hand smoke exposure in a model presented in the online supplement; this adjustment did not appreciably affect the results. In all models, effects of NO2 and PM2.5 were estimated using separate models for the main results.
We performed a limited number of prespecified sensitivity analyses. To assess the effect of varying follow-up times on effect estimates, Cox proportional hazards modeling was performed for the primary outcome of incident asthma. Participants who developed asthma reported the month and year of doctor diagnosis of asthma, and the time between baseline and Day 15 of the diagnosis month was used as the time to failure (see the online supplement for additional details).
In addition, a limited set of potential effect modifiers was specified a priori. Interactions between air pollution exposures and smoking status (current smokers vs. both former and never-smokers) were tested.
This study was approved by the University of Washington (Seattle, WA) Institutional Review Board and the NIEHS Institutional Review Board.
Out of the 50,884 women enrolled in the Sister Study, we excluded the 7,013 reporting asthma at baseline. Of the remaining 43,352 participants eligible for analysis, 2,300 (5.3%) did not complete the follow-up questionnaire. An additional 1,027 (2.4%) participants were missing exposure data because the locations were outside of the modeling region or could not be geocoded. Additional exclusions were made for each outcome according to the corresponding case definition (Figure 1). The observed incidence rate of adult asthma was approximately 2.5 cases per 1,000 person-years.
The racial/ethnic makeup of the population differed substantially by tertiles of PM2.5 and NO2 (Table 1), with higher exposure in black subjects than in non-Hispanic white subjects. Body mass index had a slight positive association with PM2.5 but not with NO2. The proportion of women with master’s or doctoral degrees was higher in increasing quartiles of NO2; a similar but weaker association was present with PM2.5. Age, smoking status, second-hand smoke exposure, health care coverage, and occupational exposure to dust and fumes had only weak or inconsistent associations with the air pollutants. Final analytic sample sizes differed because of differences in baseline exclusion criteria for the different outcomes. After excluding individuals missing exposure and covariate data in the fully adjusted model, there were 254 cases of adult incident asthma out of 35,862 at risk; 194 cases with incident chronic cough and wheeze out of 31,545 at risk; 1,559 cases with incident chronic cough only out of 33,316 at risk; and 1,023 cases with incident wheeze only out of 33,747 at risk.
| PM2.5 | NO2 | |||||
|---|---|---|---|---|---|---|
| [1.89, 9.45] | (9.45, 11.9] | (11.9, 18] | [0.729, 7.59] | (7.59, 11.4] | (11.4, 31.5] | |
| n | 16,490 | 16,489 | 16,489 | 16,490 | 16,489 | 16,489 |
| Age, yr | 55.7 ± 8.9 | 55.1 ± 9.1 | 54.6 ± 8.9 | 55.5 ± 8.9 | 54.9 ± 9.0 | 55.0 ± 9.1 |
| BMI | 27.1 ± 5.8 | 27.5 ± 6.0 | 28.0 ± 6.3 | 27.5 ± 5.9 | 27.5 ± 6.0 | 27.6 ± 6.3 |
| Daily fiber consumption | 17.3 ± 8.4 | 17.0 ± 8.5 | 16.8 ± 8.5 | 16.9 ± 8.3 | 16.9 ± 8.4 | 17.3 ± 8.8 |
| Baseline smoking status, % | ||||||
| Current smoker | 7.6 | 8.2 | 8.9 | 8.5 | 7.6 | 8.6 |
| Never smoked | 52.8 | 53.1 | 54.6 | 54.0 | 54.6 | 51.9 |
| Past smoker | 37.4 | 36.3 | 34.3 | 35.4 | 35.6 | 37.0 |
| Social smoker | 2.2 | 2.4 | 2.3 | 2.2 | 2.2 | 2.5 |
| Education, % | ||||||
| Less than high school | 0.8 | 1.1 | 1.1 | 1.1 | 0.9 | 1.1 |
| High school or GED | 13.8 | 15.2 | 13.3 | 16.6 | 14.1 | 11.6 |
| Some college | 20.3 | 19.7 | 19.2 | 20.9 | 19.3 | 19.0 |
| Bachelor’s | 27.5 | 25.8 | 27.3 | 25.2 | 27.4 | 28.1 |
| Associate, tech, or nursing | 15.0 | 14.4 | 12.9 | 16.1 | 14.2 | 12.0 |
| Master’s/doctoral | 22.4 | 23.8 | 26.2 | 20.1 | 24.1 | 28.2 |
| Race, % | ||||||
| Black | 2.5 | 7.0 | 17.8 | 5.0 | 9.3 | 13.0 |
| Hispanic | 3.4 | 3.1 | 3.1 | 2.2 | 2.7 | 4.7 |
| Non-Hispanic white | 91.4 | 87.3 | 76.7 | 90.0 | 85.7 | 79.7 |
| Other | 2.7 | 2.6 | 2.4 | 2.8 | 2.3 | 2.7 |
| Childhood secondhand smoke, % | 47.0 | 47.9 | 47.6 | 46.5 | 47.3 | 48.6 |
| Health care coverage, % | 96.3 | 96.2 | 95.9 | 96.1 | 96.3 | 96.1 |
| Occupational dust exposure, % | 21.2 | 22.1 | 23.4 | 22.7 | 21.6 | 22.4 |
| Occupational fumes exposure, % | 23.9 | 23.3 | 23.7 | 24.7 | 23.2 | 23.1 |
The medians (and interquartile ranges) for estimated exposures at participant locations were 10.8 μg/m3 (3.6 μg/m3) for PM2.5 and 9.3 ppb (5.8 ppb) for NO2. The fully adjusted odds ratio (OR) for incident asthma was 1.20 (95% confidence interval [CI] = 0.99–1.46, P = 0.063) for an interquartile range (IQR) increase in PM2.5 and 1.12 (95% CI = 0.96–1.30, P = 0.141) for an IQR increase in NO2 (Table 2). PM2.5 was significantly associated with incident wheeze (fully adjusted OR = 1.14, 95% CI = 1.04–1.26, P = 0.008). The association of NO2 with incident wheeze was slightly weaker (fully adjusted OR = 1.08, 95% CI = 1.00–1.17, P = 0.048). Fully adjusted estimates were not appreciably different from the age-adjusted estimates with the exception of the NO2 and wheeze result, which was nonsignificant in the age-adjusted model.
| Exposure (IQR) | Outcome | Age-adjusted* | Fully Adjusted† | ||
|---|---|---|---|---|---|
| OR (95% CI) | P Value | OR (95% CI) | P Value | ||
| PM2.5 (3.6 μg/m3) | Incident asthma | 1.19 (0.99–1.42) | 0.059 | 1.20 (0.99–1.46) | 0.063 |
| Cough and wheeze | 0.98 (0.81–1.19) | 0.855 | 0.94 (0.75–1.16) | 0.555 | |
| Cough | 0.98 (0.91–1.06) | 0.621 | 0.95 (0.88–1.03) | 0.194 | |
| Wheeze | 1.17 (1.07–1.28) | 0.001 | 1.14 (1.04–1.26) | 0.008 | |
| NO2 (5.8 ppb) | Incident asthma | 1.12 (0.97–1.28) | 0.123 | 1.12 (0.96–1.30) | 0.141 |
| Cough and wheeze | 1.00 (0.85–1.18) | 0.977 | 1.01 (0.84–1.21) | 0.913 | |
| Cough | 0.99 (0.93–1.05) | 0.680 | 1.00 (0.93–1.07) | 0.939 | |
| Wheeze | 1.05 (0.98–1.13) | 0.151 | 1.08 (1.00–1.17) | 0.048 | |
The Cox proportional hazard model, assessing the sensitivity of the analysis to differences in follow-up time, gave comparable results. Individuals missing date of diagnosis were excluded, leaving 231 cases of incident asthma after excluding individuals missing exposure or covariate data. The fully adjusted hazard ratio (HR) for an IQR increase in PM2.5 was 1.20 (95% CI = 0.98–1.47, P = 0.079). The fully adjusted HR for an IQR increase in NO2 was 1.06 (95% CI = 0.90–1.25, P = 0.486).
An association between NO2 and incident wheeze was evident among never-smokers or former smokers but not in the smaller group of current smokers (interaction P = 0.012). The OR for wheeze for an IQR increase in NO2 was 1.14 (95% CI = 1.04–1.24, P = 0.003) in former/never-smokers (792 cases) and 0.89 (95% CI = 0.74–1.06, P = 0.179) in current smokers (232 cases). For wheeze plus cough, the OR for an IQR increase in NO2 was 1.12 (95% CI = 0.92–1.37, P = 0.272) in never-smokers or former smokers (144 cases) and 0.71 (95% CI = 0.48–1.07, P = 0.102) in current smokers (50 cases) (interaction P = 0.049).
This is the first study of PM2.5 and adult-onset asthma. Higher PM2.5 concentrations were associated with incident wheeze and asthma in a large national cohort of women, with a prospective study design, strict definition of incident asthma, and the use of hybrid partial least squares regression universal kriging models. Our results were also consistent with an association between higher NO2 and current wheeze, especially in women who did not currently smoke. Collectively, these results support the hypothesis that air pollution exposure may be a risk factor for new-onset wheeze and diagnosis of asthma in previously asymptomatic adults. Neither incident cough alone nor cough and wheeze combined was significantly associated with either pollutant. Although cough is not specific to asthma, cough may represent other pathologies potentially relevant to air pollution and thus we included this analysis.
For both NO2 and PM2.5, the effect of each pollutant on incident asthma was comparable in size to the effect of NO2 on incident wheeze (1.12 vs. 1.08 for NO2 and 1.16 vs. 1.12 for wheeze). Although analyses of incident asthma were not statistically significant, they included fewer cases than the wheeze analyses (254 vs. 1,023) and thus are not as well powered.
Notably, the current EPA National Ambient Airy Quality Standard (NAAQS) for annual NO2 is 53 ppb. By comparison, the maximum estimated participant exposure in this analysis was 31.5 ppb. Our results indicate respiratory effects with long-term exposure may be occurring at levels well below the current national standards.
Associations between acute exposure to air pollution and exacerbation of respiratory symptoms in subjects with asthma are fairly well established, especially in children (2). Symptom exacerbations in subjects with asthma have been shown to be related to PM2.5 (17), PM10 (18), SO2 (18), and NO2 (19). There are fewer studies on long-term exposure and incident asthma and most are in children (20, 21). A meta-analysis, including four studies of childhood asthma and one study of adult asthma symptom scores, found an association between asthma and PM2.5 (22).
The only previous study on PM2.5 and adult incident asthma used visibility data to estimate PM2.5 exposure and symptom scores to define incident asthma (23). We cannot directly compare our effect size with that of Abbey and colleagues (23) because their reported exposure contrast of 45 μg/m3 is considerably larger than our observed variability; nevertheless, despite substantially reduced exposure variability in the Sister Study population, we find effects at lower levels of exposure. The results are also not directly comparable because of the different outcome classification approaches, greater exposure misclassification due to the indirect estimation method used in the earlier study, or different exposure characteristics given the different periods and locations of observation.
The present study adds to the small, but growing, body of literature indicating that air pollution may be a risk factor for the development of asthma in adults (1). An analysis of the Danish Diet, Cancer and Health cohort of adults aged 50–65 years found an association between NO2 and new hospitalization for asthma (977 cases) (19). A prospective cohort in Swedish cities found an elevated risk of incident asthma for NO2 with 55 cases (24). An earlier case–control study in Sweden by the same authors of the Swedish cohort study found an elevated but nonsignificant association with NO2 (203 cases) (25). Two cohort studies have reported associations between incident asthma and PM10 (3, 4).
We found some evidence that smoking may modify the effect of NO2 on the development of wheezing. This is consistent with the finding of Künzli and colleagues (3) that the association between NO2 and wheeze was present only in nonsmokers. Air pollution may be plausibly more asthmagenic in nonsmokers if smoking saturates biologic pathways between air pollution exposure and development of asthma. Alternatively, the observed effect of NO2 may be masked in smokers who have a higher risk of incident airway obstruction in adulthood.
Whereas the mechanisms by which air pollution might cause asthma are uncertain, the hypothesis that air pollution induces oxidative stress leading to airway inflammation has been proposed by various investigators (1). NO2 and specific components of particulate matter, possibly transition metals, increase oxidative stress in lung tissue (26). Lung lining fluid, containing antioxidants such as reduced glutathione, ascorbic acid, uric acid, and α-tocopherol, can act as defensive barrier against reactive oxygen species, but air pollutant exposure may deplete antioxidative capacity, leading to secondary oxidant species that can cause inflammatory responses (6). In response to secondary oxidants that reach lung tissue, cells may release proinflammatory molecules such as cytokines and chemokines that promote migration of additional inflammatory cells to the lung, and this could initiate an inflammatory cascade resulting in the initiation of asthma or asthma-like conditions (6).
Like any epidemiologic study of air pollution, ours has limitations for drawing causal inference. Although adjusting for well-recognized risk factors resulted in little or no attenuation of our effect estimates, the potential for residual confounding by unmeasured factors remains. We were not able to adjust for mold, dampness, gas cooking, or pets. This could confound our results to the extent that these factors are related to the spatial pattern of air pollution exposure.
Participants were asked to report month and year of asthma diagnosis, which may not have corresponded to actual disease onset. Therefore, the Cox regression analysis, which modeled time to diagnosis, was considered a sensitivity analysis rather than the primary analysis because of concerns regarding uncertainty of time of onset and nonetheless produced similar results. Our study relied on self-report of doctor-diagnosed asthma rather than methacholine challenge or other objective tests of nonspecific bronchial hyperresponsiveness. However, this type of testing is not feasible in national population studies of this size, and a past study of air pollution and asthma with a methacholine challenge component was unable to incorporate the results into the case definition (27). Although relying on self-reported doctor diagnosis to identify incident asthma may lead to outcome misclassification, it remains a frequently used approach for large epidemiological studies. Self-report of symptoms and doctor diagnosis provide useful, if imperfect, information about incident asthma. Consistency between the direction of the effect between these complementary analyses, despite differences in case counts, strengthens the conclusions of our study. In addition, highly replicable genetic loci have been identified in genome-wide association studies using self-reported doctor diagnosis of incident asthma (28).
The observed incidence of asthma, 2.5 cases per 1,000 person-years, was somewhat lower in our population than reported in other studies. A meta-analysis found an average of 4.6 cases per 1,000 person-years in women (29). However, using a less restrictive definition of asthma, and relying only on doctor diagnosis to exclude baseline prevalent cases and to define incident cases at follow-up, the incidence in the Sister Study cohort was 4.7 cases per 1,000 person-years, indicating that doctor diagnosis alone provides incidence rates consistent with other studies of adult asthma.
The possibility exists that the observed association between air pollution and asthma diagnosis may result from an association between air pollution and subclinical asthma exacerbations (leading to diagnosis during follow-up). Although we cannot completely eliminate this possibility, we addressed it by excluding from the incident asthma analysis women reporting both cough and wheeze in the 12 months before baseline and making similar exclusions in our analysis of incident symptoms. To help reduce outcome misclassification at follow-up, we used a strict definition of asthma requiring cases to have symptoms, medication use, and a doctor diagnosis. This definition therefore excluded individuals whose incident asthma was well controlled by medication. One study has suggested that more severe adult-onset asthma may differ phenotypically from milder adult-onset asthma in that it is less frequently atopic and associated with greater airway inflammation (30), characteristics that are consistent with the proposed action of air pollution on the lung.
We classified subjects as incident asthmatics only if they reported no prior asthma, even in childhood, at baseline. Some proportion of incident asthma in adults may represent recrudescence of childhood-onset asthma that was either never diagnosed or forgotten by the subject. Among all subjects who reported asthma before baseline, there were only 3.6% who reported a diagnosis of childhood asthma before the age of 16 with subsequent disappearance of symptoms for 10 years or greater. Although we cannot eliminate the possibility that air pollution is simply exacerbating occult childhood-onset asthma, a substantial proportion of incident cases is likely to have true adult-onset asthma.
Exposure misclassification is a concern for all air pollution studies that rely on predicted concentrations. Exposure models have several potential sources of misclassification, including ambient outdoor concentrations not fully characterizing indoor exposures. Furthermore, estimation of ambient concentrations at home addresses does not fully capture participant exposure because it cannot account for infiltration into the home or potentially important exposures from commuting or other activities. Nonetheless, our exposure model represents a substantial improvement over the previous generation of nearest monitor (4) and may be more comparable to high-quality dispersion models (3, 24) used in previous publications on incident asthma because our model better captures within-region and between-region contrasts in air pollution, reduces exposure misclassification through spatial smoothing, and has overall excellent model performance. We anticipate that misclassification in exposure, which would not be expected to differ by case status, would tend to attenuate the observed associations rather than lead to spurious positive ones. We also used annual average exposures, although the true biologically relevant exposure period could be shorter or longer. Misspecification of the exposure period should, at worst, lead to attenuation. Furthermore, although it is possible that seasonal bias in asthma symptom recall exists, exposure was an annual average. Therefore, seasonality in the outcome would not be related to seasonality in the exposure and would, at worse, attenuate the results.
In conclusion, our study, conducted in a nationwide prospective cohort of women and using advanced exposure modeling, provides evidence for an association between air pollution and an increased risk of incident asthma symptoms in nonasthmatic adults. We also found some evidence to support an association with asthma development. PM2.5 and NO2 were the only air pollutants analyzed in this study. Air pollution is a complex mixture; therefore, despite suggestive associations, PM2.5 and NO2 may not be the underlying etiologic agents but instead may be acting as surrogates for unmeasured pollutants. Further research is needed investigating complex multipollutant mixtures in the context of adult-onset asthma.
The authors thank Cynthia Curl, Amanda Gassett, and Marnie Hazlehurst for data management and support related to this analysis, and the staff at Social and Scientific Systems, Inc. and Westat, Inc. for overall study support and management of data acquisition and analysis in the Sister Study. The authors thank the University of Washington Center for Clean Air Research (UW CCAR) and the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air) for providing groundwork and infrastructure in developing the exposure models. The authors also thank the many thousands of women who joined the Sister Study and continue to contribute data on exposures and health.
| 1. | Jacquemin B, Schikowski T, Carsin AE, Hansell A, Krämer U, Sunyer J, Probst-Hensch N, Kauffmann F, Künzli N. The role of air pollution in adult-onset asthma: a review of the current evidence. Semin Respir Crit Care Med 2012;33:606–619. |
| 2. | Health Effects Institute (HEI) Panel on the Health Effects of Traffic-Related Air Pollution. Traffic-related air pollution: a critical review of the literature on emissions, exposure, and health effects. Boston: Health Effects Institute; 2010. |
| 3. | Künzli N, Bridevaux P-O, Liu L-JS, Garcia-Esteban R, Schindler C, Gerbase MW, Sunyer J, Keidel D, Rochat T; Swiss Cohort Study on Air Pollution and Lung Diseases in Adults. Traffic-related air pollution correlates with adult-onset asthma among never-smokers. Thorax 2009;64:664–670. |
| 4. | McDonnell WF, Abbey DE, Nishino N, Lebowitz MD. Long-term ambient ozone concentration and the incidence of asthma in nonsmoking adults: the AHSMOG Study. Environ Res 1999;80:110–121. |
| 5. | Brunekreef B, Holgate ST. Air pollution and health. Lancet 2002;360:1233–1242. |
| 6. | Kelly FJ. Oxidative stress: its role in air pollution and adverse health effects. Occup Environ Med 2003;60:612–616. |
| 7. | Brook JR, Dann TF, Burnett RT. The relationship among TSP, PM10, PM2.5, and inorganic constituents of atmospheric particulate matter at multiple Canadian locations. J Air Waste Manage Assoc 1997;47:2–19. |
| 8. | McHugh MK, Symanski E, Pompeii LA, Delclos GL. Prevalence of asthma among adult females and males in the United States: results from the National Health and Nutrition Examination Survey (NHANES), 2001–2004. J Asthma 2009;46:759–766. |
| 9. | de Marco R, Locatelli F, Sunyer J, Burney P. Differences in incidence of reported asthma related to age in men and women: a retrospective analysis of the data of the European Respiratory Health Survey. Am J Respir Crit Care Med 2000;162:68–74. |
| 10. | de Nijs SB, Venekamp LN, Bel EH. Adult-onset asthma: is it really different? Eur Respir Rev 2013;22:44–52. |
| 11. | Bel EH. Clinical phenotypes of asthma. Curr Opin Pulm Med 2004;10:44–50. |
| 12. | Siroux V, Basagaña X, Boudier A, Pin I, Garcia-Aymerich J, Vesin A, Slama R, Jarvis D, Anto JM, Kauffmann F, et al. Identifying adult asthma phenotypes using a clustering approach. Eur Respir J 2011;38:310–317. |
| 13. | Young MT, Sandler DP, DeRoo LA, Vedal S, Kaufman JD, London S. Traffic-related air pollution exposure and adult incident asthma in a cohort of US women [abstract]. Am J Respir Crit Care Med 2014;189:A2298. |
| 14. | Young MT. Traffic-related air pollution exposure and adult asthma in the Sister Study. 2013. Available from: https://digital.lib.washington.edu/researchworks/handle/1773/23732 |
| 15. | D’Aloisio AA, Baird DD, DeRoo LA, Sandler DP. Association of intrauterine and early-life exposures with diagnosis of uterine leiomyomata by 35 years of age in the Sister Study. Environ Health Perspect 2010;118:375–381. |
| 16. | Sampson PD, Richards M, Szpiro AA, Bergen S, Sheppard L, Larson TV, Kaufman JD. A regionalized national universal kriging model using partial least squares regression for estimating annual PM2.5 concentrations in epidemiology. Atmos Environ (1994) 2013;75:383–392. |
| 17. | Ostro B, Lipsett M, Mann J, Braxton-Owens H, White M. Air pollution and exacerbation of asthma in African-American children in Los Angeles. Epidemiology 2001;12:200–208. |
| 18. | Samoli E, Nastos PT, Paliatsos AG, Katsouyanni K, Priftis KN. Acute effects of air pollution on pediatric asthma exacerbation: evidence of association and effect modification. Environ Res 2011;111:418–424. |
| 19. | Andersen ZJ, Bønnelykke K, Hvidberg M, Jensen SS, Ketzel M, Loft S, Sørensen M, Tjønneland A, Overvad K, Raaschou-Nielsen O. Long-term exposure to air pollution and asthma hospitalisations in older adults: a cohort study. Thorax 2012;67:6–11. |
| 20. | Nishimura KK, Galanter JM, Roth LA, Oh SS, Thakur N, Nguyen EA, Thyne S, Farber HJ, Serebrisky D, Kumar R, et al. Early-life air pollution and asthma risk in minority children: the GALA II and SAGE II studies. Am J Respir Crit Care Med 2013;188:309–318. |
| 21. | Brauer M, Hoek G, Van Vliet P, Meliefste K, Fischer PH, Wijga A, Koopman LP, Neijens HJ, Gerritsen J, Kerkhof M, et al. Air pollution from traffic and the development of respiratory infections and asthmatic and allergic symptoms in children. Am J Respir Crit Care Med 2002;166:1092–1098. |
| 22. | Anderson HR, Favarato G, Atkinson RW. Long-term exposure to air pollution and the incidence of asthma: meta-analysis of cohort studies. Air Qual Atmos Health 2013;6:47–56. [Published erratum appears in Air Qual Atmos Health 6:541–542.] |
| 23. | Abbey DE, Lebowitz MD, Mills PK, Petersen FF, Beeson WL, Burchette RJ. Long-term ambient concentrations of particulates and oxidants and development of chronic disease in a cohort of nonsmoking California residents. Inhal Toxicol 1995;7:19–34. |
| 24. | Modig L, Torén K, Janson C, Jarvholm B, Forsberg B. Vehicle exhaust outside the home and onset of asthma among adults. Eur Respir J 2009;33:1261–1267. |
| 25. | Modig L, Järvholm B, Rönnmark E, Nyström L, Lundbäck B, Andersson C, Forsberg B. Vehicle exhaust exposure in an incident case–control study of adult asthma. Eur Respir J 2006;28:75–81. |
| 26. | Nel A. Atmosphere. Air pollution-related illness: effects of particles. Science 2005;308:804–806. [Published erratum appears in Science 309:1326.] |
| 27. | Jacquemin B, Sunyer J, Forsberg B, Aguilera I, Briggs D, García-Esteban R, Götschi T, Heinrich J, Järvholm B, Jarvis D, et al. Home outdoor NO2 and new onset of self-reported asthma in adults. Epidemiology 2009;20:119–126. |
| 28. | Moffatt MF, Gut IG, Demenais F, Strachan DP, Bouzigon E, Heath S, von Mutius E, Farrall M, Lathrop M, Cookson WO; GABRIEL Consortium. A large-scale, consortium-based genomewide association study of asthma. N Engl J Med 2010;363:1211–1221. |
| 29. | Eagan TML, Brøgger JC, Eide GE, Bakke PS. The incidence of adult asthma: a review. Int J Tuberc Lung Dis 2005;9:603–612. |
| 30. | Amelink M, de Groot JC, de Nijs SB, Lutter R, Zwinderman AH, Sterk PJ, ten Brinke A, Bel EH. Severe adult-onset asthma: a distinct phenotype. J Allergy Clin Immunol 2013;132:336–341. |
* J.D.K. and S.J.L. contributed equally to this work.
Supported by a Biostatistics, Epidemiologic and Bioinformatic Training in Environmental Health Training Grant (ES015459) and by the U.S. Department of Health and Human Services, Health Resources and Services Administration’s Maternal and Child Health Bureau (Title V, Social Security Act) (grant T76MC00011-21-00) (M.T.Y.). Supported in part by the Intramural Program of the NIH, National Institute of Environmental Health Sciences (Z01 ES044005 and ES49019). J.D.K. was also supported by K24 ES013195. Exposure modeling was supported by the U.S. Environmental Protection Agency (RD831697 and R834796). This article does not necessarily reflect the views of the Agency, and no official endorsement should be inferred.
Author Contributions: All authors had a role in the conception and design of the analysis. M.T.Y. drafted the manuscript. S.J.L., J.D.K., and S.V. jointly supervised M.T.Y. in the analysis and drafting and revisions of the manuscript. D.P.S. is the principal investigator of the NIEHS Sister Study. L.A.D. played a major role in the conception and management of the project.
This article has an online supplement, which is accessible from this issue’s table of contents online at www.atsjournals.org
Originally Published in Press as DOI: 10.1164/rccm.201403-0525OC on August 29, 2014
Author disclosures are available with the text of this article at www.atsjournals.org.
