What this study adds
Environmental neuroepidemiology classically emphasizes neurodevelopment in children or neurodegeneration in elderly populations, whereas we evaluated analogous associations among adults without known neurologic conditions. Assessing general population exposure levels in association with subclinical endpoints among adults generates information relevant to neuroepidemiology across the life course. Indeed, we present evidence that ambient levels of these pollutants are associated with neurologic impacts in adults, even below regulatory thresholds. We believe that the present work is a valuable addition to the burgeoning field of research relating to criteria air pollutants and the brain.
Over 6 million deaths worldwide were attributed to air pollution in 2015.1,2 In the United States, outdoor air pollution is associated with reduced life expectancy and 100,000 deaths annually.3–5 Under the Clean Air Act, the United States Environmental Protection Agency (USEPA) is responsible for developing criteria to set limits on outdoor levels of six common air pollutants that are harmful to human health and the environment.6 These six criteria pollutants (carbon monoxide, nitrogen dioxide, sulfur dioxide, lead, ground-level ozone, and particulate matter [PM]) are emitted from fuel combustion processes, with higher levels in urban and high-traffic areas.2 Specifically, environmental sources include emissions from internal combustion engine vehicles (nitrogen oxides and carbon monoxide), fossil fuel industries (sulfur oxides), and industrial processing and aircraft (lead). Ground-level ozone (henceforth referred to as ozone), a secondary pollutant created by chemical reactions between precursor emissions of nitrogen oxides and volatile organic compounds, is the main constituent in smog.7 Similarly, most particulate matter (PM) is generated by reactions between nitrogen oxides and sulfur oxides, although some PM is emitted directly (e.g., black carbon).8 PM is a mixture of acids, organic chemicals, metals and soil or dust particles.9 Categorized by size, coarse PM includes all particles with an aerodynamic diameter less than 10 μm (PM10), and fine PM represents particles less than 2.5 μm in diameter (PM2.5).10
Exposure to PM2.5 and ozone is causally associated with mortality, even at levels below regulatory standards,11 and accounts for over 90% of air pollution-related morbidity.1,12,13 In addition to well-known respiratory and cardiovascular effects, air pollution is also associated with neuroinflammation, leading to neurotoxicity.14 PM2.5, which can cross the blood-brain barrier, appears to be the most significant air pollutant associated with adverse neurologic outcomes.10 There is strong evidence that traffic-related air pollution, including PM2.5 and ozone, causes neurodevelopmental delay and alters neurobehavioral function in children.10,15–17 Long-term PM2.5 exposure is consistently associated with Alzheimer’s disease, Parkinson’s disease, dementia, neurodegeneration, and stroke.18–26 Evidence of relationships between PM2.5 and cognitive decline, anxiety, and depression is mixed.27–31 While there is general consensus that these pollutants have adverse effects on neurodevelopment and neurodegeneration, further synthesis of the literature and meta-analyses are often hampered because exposure assessment varies widely among studies.32
Although one study found suggestive associations between outdoor ozone levels and neurologic symptoms,33 few studies investigating the neurotoxicity of air pollution have evaluated subclinical outcomes in middle-aged populations.21,34 Given the availability of data for PM2.5 and ozone (but not precursor criteria pollutants), the present study examines cross-sectional associations between exposure to these secondary criteria pollutants and neurologic symptoms among adults in the Gulf Long-term Follow-up Study.
We used data from the Gulf Long-term Follow-up Study (GuLF Study), a prospective cohort of 32,608 adults from across the United States (ages 21 years and older) who participated in oil spill response activities and others who received safety training, but did not participate in the response, following the 2010 Deepwater Horizon (DWH) disaster. Participants enrolled between March 2011 and March 2013. A detailed description of this study is available elsewhere.35,36 Of the 26,828 English- or Spanish-speaking GuLF Study participants living in the Gulf region (Alabama, Florida, Louisiana, Mississippi, and Texas) at enrollment, 25,834 (96%) reported addresses that were successfully geocoded to a 2010 US Census tract. From this sample of participants with known residential locations, we excluded participants with any missing neurologic symptom reporting (n = 1,343) or covariate information (n = 3,024), leaving 21,467 eligible participants. Of the 3,024 participants excluded based on incomplete covariate information, 88% of exclusions were due to missing income. Covariate and health information, including neurologic symptoms, was collected via. Computer-Assisted Telephone Interview during the enrollment interview. Participants provided verbal informed consent for the telephone interview and the Institutional Review Board of the National Institute of Environmental Health Sciences approved this study.
At enrollment, participants were asked to report how often they experienced each of nine neurologic symptoms—dizziness, nausea, headaches, sweating, palpitations, tingling, numbness, blurred vision, and stumbling while walking—during the preceding 30 days. Frequency of symptoms was reported as: all of the time, most of the time, sometimes, rarely, or never. For analyses, symptoms were classified as a binary indicator of the “presence” (all or most of the time) or “absence” (sometimes, rarely, or never) of occurrence. Based on results of a principal components analysis of all reported symptoms that yielded five distinct symptom clusters (data not shown), we identified two neurologic clusters: a central nervous system (CNS) and a peripheral nervous system (PNS) cluster. The CNS cluster included dizziness, headache, nausea, sweating, and palpitations (i.e., the presence of CNS symptoms was defined as experiencing at least one of these symptoms all or most of the time). The PNS cluster included tingling and numbness in the extremities, blurred vision, and stumbling while walking.
Air pollution exposures
Daily census tract estimates of PM2.5 (μg/m3) and ozone [parts per billion (ppb)] were obtained from the United States Environmental Protection Agency (USEPA) fused Downscaler Model for the time period corresponding with participant enrollment (2011–2013).37,38 These daily predictions of ambient PM2.5 and ozone levels are generated by using Bayesian spatially-varying weighted regression to combine modeled output (from the Community Multiscale Air Quality [CMAQ] model) with daily air pollution measurements (from USEPA monitors), thereby improving fine-scale estimation of pollution levels for US census tract centroid locations.39
We matched each participant’s residential location at enrollment to their corresponding census tract and applied the USEPA-estimated daily PM2.5 and ozone levels. We then calculated each participant’s 7-day and 30-day moving average exposure, capturing the average levels over the week and month leading up to the date they reported neurologic symptoms (i.e., their enrollment date). Because symptoms were ascertained with respect to the previous 30 days, the 30-day metrics match the entire symptom reporting window, whereas the 7-day metrics reflect air pollution levels closest to the date of symptom reporting, which addresses possible recency bias. Conceptually, the 30-day exposure metrics correspond to typical exposure, whereas the 7-day metrics represent acute or recent exposure. In addition to these average exposure metrics, we also extracted the maximum daily level of each pollutant over the same 7- and 30-day intervals. Finally, we counted the number of days during the 30-day reporting period when concentrations exceeded the maximum daily levels set by the USEPA in 2020, known as National Ambient Air Quality Standards (NAAQS), for PM2.5 (35 μg/m3) or ozone (70 ppb).40,41 These primary regulatory standards set the maximum allowed daily concentration levels in outdoor air to protect human health.
We used multivariable log-binomial regression to estimate associations of PM2.5 and ozone levels with neurologic symptoms, reported as prevalence ratios (PRs) and corresponding 95% confidence intervals (CIs). We modeled outcomes as symptom clusters, separately examining associations between air pollution levels and the presence of any neurologic symptom, any CNS symptom, more than one CNS symptom, any PNS symptom, and more than one PNS symptom.
We modeled exposures continuously and categorically. For continuous linear exposure, we presented results associated with an interquartile range (IQR) difference in the 30-day average concentration (10 ppb for ozone and 2 μg/m3 for PM2.5). To facilitate comparison with other studies, we also reported associations with a 10 μg/m3 difference in PM2.5, although this exposure contrast is disproportionately large for the observed range of PM2.5 levels in our study. In categorical analyses, we modeled 7- and 30-day average pollutant concentrations in quartiles, designating the lowest quartile as the referent group. We evaluated exposure-response relationships using Wald tests of linear trend by modeling exposure as an ordinal variable with the median concentration for each quartile of exposure and reported corresponding P values (P trend). Based on the distribution of days exceeding USEPA NAAQS for each pollutant, we modeled these exposures as zero, 1, or more than 1 day of elevated pollution levels.
We identified a common covariate adjustment set using directed acyclic graph analysis of the relationship between PM and ozone and neurologic symptoms (Supplemental Figure 1; http://links.lww.com/EE/A148).42 All models were adjusted for age (<30 years, 30–45 years, >45 years), sex (male, female), season (winter, spring, summer, fall), self-reported race (White, Black, other), current employment status (currently working, not currently working), current alcohol consumption (any self-reported alcoholic drinks in the last year, no drinks in the last year), smoking (self-reported current smokers, nonsmokers), educational attainment (less than high school diploma or general equivalency degree [GED], completed high school/GED, completed some college, college graduate), annual household income (≤$20,000, $20,001–50,000, >$50,000), and area deprivation index (ADI) 2013.43–45 The ADI is a measure of neighborhood socioeconomic status derived from American Community Survey (ACS) indicators of income, education, employment, and housing quality. ADI raw scores are ranked and reported as national percentile rankings for each census block group, which we matched to participant addresses and modeled as a continuous measure. Due to correlation between PM2.5 and ozone, we mutually adjusted for PM2.5 and ozone in primary analyses. For comparison, we provide effect estimates adjusted for covariates only (not co-pollutants) in Supplemental Tables (http://links.lww.com/EE/A148).
To evaluate whether ongoing exposure to cigarette smoke was either potentiating or obscuring air pollution exposure-related associations among smokers, we conducted subanalyses stratified by smoking (current smokers, current nonsmokers). Given that neurologic symptom reporting may vary according to demographic factors, we assessed potential heterogeneity by sex (women, men) and age (<30 years, 30–45, >45). Based on evidence that people of color are exposed to disproportionately higher levels of air pollution than are White people,46 we also stratified by race (collapsed to non-White and White because of sample size). We conducted analyses stratified by season (spring/summer, fall/winter) to explore the impacts of seasonal variability in air pollution and symptom reporting. Because higher ADI tends to correlate with higher air pollution exposures, less green space, and higher rates of underlying health conditions, we also evaluated modification by ADI (above the analytic cohort’s median deprivation index [>63], below analytic cohort’s median deprivation index). In stratified analyses, interaction P values (P interaction) were derived from multivariate analysis of variance using F tests of the overall interaction between quartiles of exposure and the modifier of interest. We also modeled peak exposure levels, applying the maximum PM2.5 and ozone concentration recorded on a single day during the 7- and 30-day windows as the exposure of interest. Peak exposures were modeled using identical methods to those of the primary (average exposure) analyses. Finally, we assessed the impact of using other parameters for ADI (various categorical terms based on the distribution), elevated PM days (alternative cutpoints), and self-reported smoking (lifetime ever, never smokers). All statistical analyses were conducted in SAS 9.4 (Cary, NC). All tests were two-sided with α = 0.05.
Study participants who experienced neurologic symptoms at enrollment (n = 6,479, 30%; Supplemental Table 1; http://links.lww.com/EE/A148) were more likely to be non-White and to report characteristics of lower socioeconomic status (e.g., they were less likely to be employed or have a college degree and more likely to earn under $20,000 and live in areas with higher ADI) (Table 1). The rate of smoking was higher among those reporting symptoms (44% vs. 30%), but current alcohol consumption was higher among those who did not report symptoms (66% vs. 77%). Exposure distributions did not differ according to symptom reporting (Table 1). The overall mean 30-day exposures for PM2.5 and ozone were 8.6 μg/m3 and 40.1 ppb, respectively (Supplemental Table 2; http://links.lww.com/EE/A148). Measures of central tendency were similar between the 7- and 30-day exposure metrics, although higher SDs indicated more variability among 7-day exposure metrics. The rank correlation between average PM2.5 and ozone ranged from 0.41 (for 30-day metrics) to 0.46 (for 7-day metrics) (Supplemental Table 3; http://links.lww.com/EE/A148). When comparing 30- and 7-day averages within a pollutant, the correlation was 0.59 for PM2.5 and 0.74 for ozone. Distributions of PM2.5 and ozone were identical between those excluded and included in the analytic population, while the prevalence of neurologic symptoms was 3–5% higher in the analytic population. We did not observe differential covariate patterns between those who were included and excluded.
Table 1. -
Bivariate distribution of demographic characteristics among Gulf Long-term Follow-up Study participants by neurologic symptom-reporting status, Gulf States, United States, 2011–2013 (n = 21,467)
||Any neurologic symptoms (n = 6,479), N (%)
||No neurologic symptoms (n = 14,988), N (%)
| 30–45 years
| >45 years
|Completed < high school diploma/GED
| High school diploma/GED
| Some college
| ≥College degree
|Annual household income ≤$20,000
|ADI, median (IQR)
|Currently consumes alcohol
|Currently smokes cigarettes
|Experienced ≥1 day exceeding NAAQSa
| PM2.5 >35 μg/m3
| Ozone >70 ppb
|Average exposure, mean (SD)
| 30-day PM2.5 (μg/m3)
| 7-day PM2.5 (μg/m3)
| 30-day ozone (ppb)
| 7-day ozone (ppb)
|Average exposure, median (IQR)
| 30-day PM2.5 (μg/m3)
| 7-day PM2.5 (μg/m3)
| 30-day ozone (ppb)
|7-day ozone (ppb)
aThe 2020 United States Environmental Protection Agency (USEPA) National Ambient Air Quality Standards (NAAQS) are 35 μg/m3 PM2.5 and 70 ppb ozone. The cells report the number (and percent) of people who lived in census tracts where concentrations of either pollutant exceeded the standard on at least one day during the 30-day symptom reporting period.
ppb indicates parts per billion.
A co-exposure-adjusted IQR difference in 30-day average PM2.5 concentration was associated with slightly increased prevalence of any neurologic symptoms (PR = 1.06; 95% CI = 1.03, 1.09), whereas there was no association with ozone (PR = 1.02; 95% CI = 0.98, 1.06) (Table 2). PM2.5 exposure was associated with prevalence of each symptom cluster, with the strongest association for multiple CNS symptoms (PR = 1.13; 95% CI = 1.08, 1.19). Unadjusted estimates appeared to be biased away from the null, with the addition of covariates and co-exposures slightly attenuating associations.
Table 2. -
Associations between continuous linear 30-day average exposure metrics and neurologic symptoms (n = 21,467)
|PR (95% CI)
||PR (95% CI)
||PR (95% CI)
|Ozone, IQR difference (10 ppb)
||1.12 (1.08, 1.15)
||1.06 (1.03, 1.10)
||1.02 (0.98, 1.06)
||1.14 (1.10, 1.18)
||1.06 (1.02, 1.10)
||1.01 (0.96, 1.05)
||1.13 (1.07, 1.20)
||1.03 (0.98, 1.09)
||0.95 (0.89, 1.01)
||1.13 (1.09, 1.18)
||1.06 (1.02, 1.11)
||1.01 (0.96, 1.06)
||1.17 (1.11, 1.23)
||1.11 (1.05, 1.17)
||1.05 (0.98, 1.12)
|PM2.5, IQR difference (2 μg/m3)
||1.11 (1.09, 1.14)
||1.07 (1.04, 1.09)
||1.06 (1.03, 1.09)
||1.15 (1.12, 1.18)
||1.08 (1.05, 1.11)
||1.08 (1.05, 1.11)
||1.18 (1.13, 1.23)
||1.11 (1.07, 1.16)
||1.13 (1.08, 1.19)
||1.11 (1.07, 1.14)
||1.08 (1.05, 1.11)
||1.08 (1.04, 1.11)
||1.12 (1.07, 1.16)
||1.10 (1.06, 1.14)
||1.08 (1.03, 1.13)
|PM2.5,10 μg/m3 difference
||1.69 (1.51, 1.89)
||1.38 (1.23, 1.54)
||1.33 (1.17, 1.52)
||2.02 (1.77, 2.30)
||1.49 (1.31, 1.71)
||1.47 (1.26, 1.72)
||2.26 (1.83, 2.79)
||1.69 (1.39, 2.07)
||1.87 (1.49, 2.35)
||1.66 (1.43, 1.93)
||1.46 (1.26, 1.70)
||1.45 (1.22, 1.72)
||1.74 (1.43, 2.12)
||1.60 (1.31, 1.94)
||1.47 (1.17, 1.85)
aAdjusted for age, sex, season, race, employment status, alcohol status, smoking status, education, income, and ADI.
bAdjusted as in (a) and additionally adjusted for either continuous average 30-day PM2.5 or ozone concentration.ppb indicates parts per billion.
When modeled categorically, 30-day average PM2.5 was associated with increased prevalence of each neurologic symptom cluster (Figure 1 and Supplemental Table 4; http://links.lww.com/EE/A148). We observed statistically significant increasing monotonic relationships with each outcome (P-trend ≤ 0.01), and the third and fourth quartiles of exposure were significantly associated with each cluster. The highest quartile of 30-day average PM2.5 was associated with a 16% increase in any neurologic symptom (PR = 1.16; 95% CI = 1.09, 1.23; P-trend = 0.002). We did not, however, observe similar relationships with 7-day average PM2.5. While 30-day average ozone was not associated with neurologic symptoms (Figure 2 and Supplemental Table 5; http://links.lww.com/EE/A148), the highest quartile of 7-day ozone was associated with increased PNS symptoms [any PNS symptom: PR = 1.09; 95% CI = 1.00, 1.19; P-trend = 0.03; multiple PNS symptoms: PR = 1.17; 95% CI = 1.04, 1.30; P-trend = 0.006]).
Multiple elevated ozone days (exceeding the NAAQS limit of 70 ppb) during the 30-day reporting period was associated with increased prevalence of any neurologic symptoms in covariate-adjusted models (PR = 1.12; 95% CI = 1.05, 1.20), but these associations were attenuated when PM2.5 was included in the model (PR = 1.06; 95% CI = 0.99, 1.14) (Table 3). Because the frequency of exposure to PM2.5 levels exceeding the NAAQS (35 μg/m3) was rare in this study (n = 28; Table 1), analyses of elevated PM2.5 days were insufficiently powered to detect potential associations with neurologic symptoms.
Table 3. -
Associations between the frequency of days exceeding the USEPA NAAQS for ozone (70 ppb) during the 30-day reporting period and neurologic symptoms (n = 21,467)
||High ozone days (>70 ppb)
||Covariate- and PM2.5-adjustedb
|PR (95% CI)
||PR (95% CI)
||PR (95% CI)
||1.08 (1.01, 1.16)
||1.00 (0.94, 1.06)
||0.96 (0.90, 1.03)
||1.08 (1.00, 1.17)
||1.12 (1.05, 1.20)
||1.06 (0.99, 1.14)
||1.05 (0.97, 1.15)
||0.97 (0.90, 1.04)
||0.95 (0.86, 0.99)
||1.11 (1.01, 1.23)
||1.12 (1.03, 1.21)
||1.04 (0.96, 1.13)
||1.08 (0.94, 1.24)
||0.97 (0.87, 1.09)
||0.90 (0.80, 1.01)
||1.00 (0.85, 1.19)
||1.03 (0.90, 1.16)
||0.92 (0.81, 1.05)
||1.16 (1.07, 1.27)
||1.06 (0.98, 1.14)
||1.01 (0.93, 1.10)
||1.09 (0.98, 1.21)
||1.12 (1.03, 1.23)
||1.05 (0.95, 1.15)
||1.18 (1.05, 1.33)
||1.08 (0.98, 1.20)
||1.03 (0.93, 1.14)
||1.12 (0.98, 1.28)
||1.15 (1.03, 1.29)
||1.06 (0.94, 1.20)
aAdjusted for age, sex, season, race, employment status, alcohol status, smoking status, education, income, and ADI.
bAdjusted as in (a) and additionally adjusted for continuous 30-day average PM2.5.ppb indicates parts per billion.
Associations with PM2.5 appeared to be generally stronger, and to show more consistent exposure-response trends, among nonsmokers than among smokers (Supplemental Figure 2; http://links.lww.com/EE/A148 and Supplemental Table 6; http://links.lww.com/EE/A148). We did not observe heterogeneity by smoking for ozone exposure. Although interactions between season and PM2.5 exposure were not statistically significant (Supplemental Figure 3; http://links.lww.com/EE/A148 and Supplemental Table 7; http://links.lww.com/EE/A148), relationships between PM2.5 and neurologic outcomes were more consistent and stronger in fall/winter than in spring/summer. Conversely, the interaction between season and ozone suggested effect measure modification (P-interaction < 0.003 for all endpoints), but stratified associations showed no clear patterns in any season. We did not detect clear evidence of effect measure modification by sex (Supplemental Figure 4; http://links.lww.com/EE/A148 and Supplemental Table 8; http://links.lww.com/EE/A148), age (Supplemental Figure 5; http://links.lww.com/EE/A148 and Supplemental Table 9; http://links.lww.com/EE/A148), ADI (Supplemental Figure 6; http://links.lww.com/EE/A148 and Supplemental Table 10; http://links.lww.com/EE/A148), or race (Supplemental Figure 7; http://links.lww.com/EE/A148 and Supplemental Table 11; http://links.lww.com/EE/A148) with either exposure. It should be noted that, for PM2.5 exposure, we observed clearer exposure-response relationships among non-White participants than their White counterparts (Supplemental Figure 6; http://links.lww.com/EE/A148).
Using alternative parameters for ADI (i.e., ranked categories, cutpoints from regional and national distributions), other cutpoints for elevated ozone days, and lifetime smoking status did not change interpretations (data not shown). Results from peak exposure to PM2.5 and ozone (data not shown) were similar to those presented in main analyses of average exposure concentrations.
In this investigation of air pollution and neurologic effects, exposure to PM2.5 was associated with increasing prevalence of neurologic symptoms. We observed monotonic exposure-response relationships between 30-day PM2.5 and all neurologic symptom clusters. Average ozone levels, however, were not associated with most neurologic outcomes. While number of elevated ozone days (ozone >70 ppb) initially appeared to be associated with symptom prevalence, adjustment for co-pollutants suggested that PM2.5 was likely driving the associations observed with ozone exposure. Indeed, adjusting for PM2.5 exposure generally attenuated ozone-related associations, but adjusting for ozone exposure did not impact PM2.5-related effect estimates (Table 2 and Supplemental Tables 3 and 4; http://links.lww.com/EE/A148), suggesting that PM2.5 is the more relevant exposure in relation to prevalence of neurologic symptoms. PM2.5 was more strongly associated with neurologic symptoms among nonsmokers (compared with smokers) and during colder seasons (compared with spring/summer). We did not detect heterogeneity of effect by other social and demographic factors.
We observed a consistent relationship between 30-day average PM2.5 exposure and each cluster of neurologic symptoms, with the strongest associations for reporting multiple CNS symptoms. These findings are supported by previous research, as PM2.5 appears to be the most frequent component of air pollution associated with adverse CNS outcomes like neurodegenerative diseases and cognitive decline.10,19,20,22 While the ability of PM to promote initiation of neurodegenerative diseases remains unclear, its role in exacerbating disease severity is supported.21,26 In the absence of consensus regarding the relevant window of susceptibility for air pollution and neurologic effects, our findings suggest that expanding studies of air pollution’s neurologic impacts to earlier ages may be illuminating. Our observation that 30-day PM2.5, but not 7-day exposure, was associated with neurologic symptoms suggests that the 30-day metric may be more reflective of typical PM2.5 exposure, whereas the 7-day metric may introduce exposure misclassification because it is more heavily influenced by fluctuations in PM2.5 levels and therefore more likely to deviate from usual exposure.
In the present study, associations between ozone and neurologic symptoms were suggestive, although not as compelling as those with PM2.5. In contrast to a previous investigation that reported associations between ambient ozone levels and increased headache, fatigue, and difficulty concentrating in adults,33 we did not observe associations between ozone levels and CNS symptoms (including headache). We did, however, find that the highest ozone levels were associated with increased PNS symptoms. Additionally, exposure to ozone exceeding NAAQS was associated with all neurologic symptoms studied, although it is not clear whether this relationship persists independent of PM exposure.
The magnitude of associations between PM2.5 and neurologic symptoms was notably higher among nonsmokers than among smokers. The dose of related air pollutants received from cigarette smoke among smokers may so much exceed the ambient exposure levels reflected in our data that they obscure the potential air pollution-related associations. Evaluating the nonsmoking population separately eliminates this noise and reveals a clear dose-response between PM2.5 and neurologic symptoms. Associations between PM2.5 and neurologic symptoms were stronger in colder seasons. Seasonal differentials of criteria air pollutant levels tend to be most pronounced in regions with sharper seasonal contrast than is typically seen in the Gulf states.15,47,48 Indeed, we did not observe a strong seasonal gradient of PM2.5 in our study population (although levels were subtly higher in summer). An alternative explanation for the seasonal variability of associations is that exposure to heat may confound associations because some of the neurologic symptoms (e.g., dizziness, nausea, sweating, palpitations, and stumbling) may also indicate heat-related effects.49 During colder seasons, there would be less potential confounding by heat, resulting in a clearer relationship between air pollution exposure and neurologic symptoms.
Inhaled PM and ozone impact the CNS primarily by inducing systemic inflammation and damaging the blood-brain barrier. Additionally, ultrafine PM (<0.1 μm in diameter) can enter the bloodstream through the lungs and cross the blood-brain barrier or access the brain directly via. the olfactory nerve.22,50 PM2.5 leads to neurotoxicity by activating microglia, resulting in neuroinflammation and oxidative stress.14,20,21,51,52 A proinflammatory CNS environment can trigger microglial phagocytosis,50 which may also generate reactive oxygen species and contribute to oxidative stress.53 Other hypothesized mechanisms underlying air pollution’s adverse neurologic effects include a possible neuroendocrine response through activation of the hypothalamic-pituitary-adrenal axis or PM-mediated epigenetic changes, including reduced deoxyribonucleic acid (DNA) methylation.22,50,51,54,55
Strengths of this study include the large sample size, detailed covariate information, and demographic characteristics of the cohort. Together, these assets allowed us to evaluate several different effect measure modifiers with reasonable precision. While most research evaluating the relationships of air pollution and neurologic effects is in children or aging populations, our study uniquely considers this question in a population that consists primarily of middle-aged adults. Using USEPA’s downscaler estimates of exposure was advantageous because this approach leverages monitored concentrations to mitigate potential calibration bias, generating better estimates of daily ambient air pollutant levels than those obtained from models alone. Further, this resource outputs daily concentrations at a fine geographic resolution, allowing us to assign moving average exposures corresponding to two time periods (7- and 30-day) matched to each participant’s individual symptom reporting window and residential location.
This study also has limitations. The exposure data were available only for PM2.5 and ozone, so we could not evaluate other criteria pollutants. Although PM2.5 and ozone are the air pollutants most strongly linked to neurologic health, future research could improve on our approach by considering more realistic exposure scenarios, such as air pollution mixtures, or evaluating ultrafine particulate matter, which has demonstrated neurotoxicity.10,56–58 Further, attributing health impacts of PM2.5 and ozone to the specific precursor criteria pollutants that lead to their formation is a necessary step towards targeted control of emissions.12 Another limitation is that our exposure assessment relied solely on ambient concentrations because we lacked information on indoor concentrations and on participants’ daily activity patterns. We attempted to mitigate potential recency bias of self-reported outcomes by analyzing both 7- and 30-day exposure metrics. Because of inherent overlap between the CNS and PNS, some of the symptoms we analyzed could plausibly indicate dysfunction of either component (or both). To minimize the impacts of possible cluster misclassification on interpretations, we also analyzed the presence of “any” neurologic symptoms as the primary endpoint. Finally, because this study was cross-sectional, we could not interpret exposure-response relationships as causal.
Particulate matter’s ability to delay neurodevelopment, alter neurobehavior, and exacerbate neurodegeneration is supported by a growing literature,10,21,59 although questions related to exposure timing, exposure threshold, persistence of effects, and mechanisms of action remain. We observed that short-term (30-day) exposure to PM2.5 concentrations below NAAQS levels is associated with increases in neurologic symptoms in adults. Although the magnitude of effect estimates was modest, associations were consistent across analyses and increasing exposure-response relationships were clear. Given the ubiquity of these neurologic symptoms, and exposure to PM2.5 at levels comparable to those observed in our study, PM2.5 could meaningfully and adversely impact health across broad populations. Continued research is needed to determine whether the observed associations between these air pollutants and neurologic symptoms have implications for subsequent development of more severe or chronic neurologic health problems.
Conflicts of interest statement
The authors declare that they have no conflicts of interest with regard to the content of this report.
This research was supported by the Intramural Research Program of the National Institutes of Health (NIH) and the National Institute of Environmental Health Sciences/NIH (Z01-ES049030).
1. Forouzanfar MH, Afshin A, Alexander LT, et al. Global, regional, and national comparative risk assessment of 79 behavioural, environmental
and occupational, and metabolic risks or clusters of risks, 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet. 2016;388:1659–1724.
2. Landrigan PJ, Fuller R, Acosta NJR, et al. The Lancet Commission on pollution and health. Lancet. 2018;391:462–512.
3. Lelieveld J, Evans JS, Fnais M, Giannadaki D, Pozzer A. The contribution of outdoor air pollution sources to premature mortality on a global scale. Nature. 2015;525:367–371.
4. Goodkind AL, Tessum CW, Coggins JS, Hill JD, Marshall JD. Fine-scale damage estimates of particulate matter air pollution reveal opportunities for location-specific mitigation of emissions. Proc Natl Acad Sci U S A. 2019;116:8775–8780.
5. Pope CA 3rd, Ezzati M, Dockery DW. Fine-particulate air pollution and life expectancy in the United States. N Engl J Med. 2009;360:376–386.
6. United States Environmental
Protection Agency. The Plain English Guide to the Clean Air Act
. 2007. Available at: https://www.epa.gov/sites/production/files/2015-08/documents/peg.pdf
. Accessed 16 December 2020.
7. Di Q, Rowland S, Koutrakis P, Schwartz J. A hybrid model for spatially and temporally resolved ozone exposures in the continental United States. J Air Waste Manag Assoc. 2017;67:39–52.
8. Suh HH, Bahadori T, Vallarino J, Spengler JD. Criteria air pollutants and toxic air pollutants. Environ Health Perspect. 2000;108(suppl 4):625–633.
9. Anderson JO, Thundiyil JG, Stolbach A. Clearing the air: a review of the effects of particulate matter air pollution on human health. J Med Toxicol. 2012;8:166–175.
10. Costa LG, Cole TB, Dao K, Chang YC, Coburn J, Garrick JM. Effects of air pollution on the nervous system and its possible role in neurodevelopmental and neurodegenerative disorders. Pharmacol Ther. 2020;210:107523.
11. Wei Y, Wang Y, Wu X, et al. Causal effects of air pollution on mortality rate in Massachusetts. Am J Epidemiol. 2020;189:1316–1323.
12. Dedoussi IC, Eastham SD, Monier E, Barrett SRH. Premature mortality related to United States cross-state air pollution. Nature. 2020;578:261–265.
13. Likhvar VN, Pascal M, Markakis K, et al. A multi-scale health impact assessment of air pollution over the 21st century. Sci Total Environ. 2015;514:439–449.
14. Salvi A, Salim S. Neurobehavioral consequences of traffic-related air pollution. Front Neurosci. 2019;13:1232.
15. Flores-Pajot MC, Ofner M, Do MT, Lavigne E, Villeneuve PJ. Childhood autism spectrum disorders and exposure to nitrogen dioxide, and particulate matter air pollution: a review and meta-analysis. Environ Res. 2016;151:763–776.
16. Guxens M, Ghassabian A, Gong T, et al. Air pollution exposure during pregnancy and childhood autistic traits in four European population-based cohort studies: the ESCAPE project. Environ Health Perspect. 2016;124:133–140.
17. Volk HE, Perera F, Braun JM, et al.; Environmental
influences on Child Health Outcomes. Prenatal air pollution exposure and neurodevelopment: a review and blueprint for a harmonized approach within ECHO. Environ Res. 2021;196:110320.
18. Tsai TL, Lin YT, Hwang BF, et al. Fine particulate matter is a potential determinant of Alzheimer’s disease: a systemic review and meta-analysis. Environ Res. 2019;177:108638.
19. Kioumourtzoglou MA, Schwartz JD, Weisskopf MG, et al. Long-term PM2.5 exposure and neurological hospital admissions in the Northeastern United States. Environ Health Perspect. 2016;124:23–29.
20. Zanobetti A, Dominici F, Wang Y, Schwartz JD. A national case-crossover analysis of the short-term effect of PM2.5 on hospitalizations and mortality in subjects with diabetes and neurological disorders. Environ Health. 2014;13:38.
21. Babadjouni RM, Hodis DM, Radwanski R, et al. Clinical effects of air pollution on the central nervous system; a review. J Clin Neurosci. 2017;43:16–24.
22. Shou Y, Huang Y, Zhu X, Liu C, Hu Y, Wang H. A review of the possible associations between ambient PM2.5 exposures and the development of Alzheimer’s disease. Ecotoxicol Environ Saf. 2019;174:344–352.
23. Shah AS, Lee KK, McAllister DA, et al. Short term exposure to air pollution and stroke: systematic review and meta-analysis. BMJ. 2015;350:h1295.
24. Nzwalo H, Guilherme P, Nogueira J, et al. Fine particulate air pollution and occurrence of spontaneous intracerebral hemorrhage in an area of low air pollution. Clin Neurol Neurosurg. 2019;176:67–72.
25. Bazyar J, Pourvakhshoori N, Khankeh H, Farrokhi M, Delshad V, Rajabi E. A comprehensive evaluation of the association between ambient air pollution and adverse health outcomes of major organ systems: a systematic review with a worldwide approach. Environ Sci Pollut Res Int. 2019;26:12648–12661.
26. Nunez Y, Boehme AK, Weisskopf MG, et al. Fine particle exposure and clinical aggravation in neurodegenerative diseases in New York State. Environ Health Perspect. 2021;129:27003.
27. Weuve J, Puett RC, Schwartz J, Yanosky JD, Laden F, Grodstein F. Exposure to particulate air pollution and cognitive decline in older women. Arch Intern Med. 2012;172:219–227.
28. Cleary EG, Cifuentes M, Grinstein G, Brugge D, Shea TB. Association of low-level ozone with cognitive decline in older adults. J Alzheimers Dis. 2018;61:67–78.
29. Wellenius GA, Boyle LD, Coull BA, et al. Residential proximity to nearest major roadway and cognitive function in community-dwelling seniors: results from the MOBILIZE Boston Study. J Am Geriatr Soc. 2012;60:2075–2080.
30. Braithwaite I, Zhang S, Kirkbride JB, Osborn DPJ, Hayes JF. Air pollution (particulate matter) exposure and associations with depression, anxiety, bipolar, psychosis and suicide risk: a systematic review and meta-analysis. Environ Health Perspect. 2019;127:126002.
31. Fan SJ, Heinrich J, Bloom MS, et al. Ambient air pollution and depression: a systematic review with meta-analysis up to 2019. Sci Total Environ. 2020;701:134721.
32. Clifford A, Lang L, Chen R, Anstey KJ, Seaton A. Exposure to air pollution and cognitive functioning across the life course–a systematic literature review. Environ Res. 2016;147:383–398.
33. Apte MG, Buchanan IS, Mendell MJ. Outdoor ozone and building-related symptoms in the BASE study. Indoor Air. 2008;18:156–170.
34. Xu X, Ha SU, Basnet R. A review of epidemiological research on adverse neurological effects of exposure to ambient air pollution. Front Public Health. 2016;4:157.
35. Kwok RK, Engel LS, Miller AK, et al.; GuLF STUDY Research Team. The GuLF STUDY: a prospective study of persons involved in the deepwater horizon oil spill response and clean-up. Environ Health Perspect. 2017;125:570–578.
36. Engel LS, Kwok RK, Miller AK, et al. The Gulf Long-Term Follow-Up Study (GuLF STUDY): biospecimen collection at enrollment. J Toxicol Environ Health A. 2017;80:218–229.
37. Reff A, Phillips S, Eyth A, Mintz D. Bayesian Space-time Downscaling Fusion Model (Downscaler)-Derived Estimates of Air Quality for 2014. U.S. Environmental
Protection Agency Office of Air Quality Planning and Standards Air Quality Assessment Division; 2018.
38. United States Environmental
Protection Agency. Fused Air Quality Surface Using Downscaling (FAQSD
). 2020. Available at: https://www.epa.gov/hesc/rsig-related-downloadable-data-files
.Accessed 16 December 2020.
39. Berrocal VJ, Gelfand AE, Holland DM. Space-time data fusion under error in computer model output: an application to modeling air quality. Biometrics. 2012;68:837–848.
40. United States Environmental
Protection Agency. National Ambient Air Quality Standards (NAAQS) for Particulate Matter
. 2020. 40 CFR Part 50. 46. EPA-HQ-OAR-2015-0072. Available at: https://www.epa.gov/pm-pollution/national-ambient-air-quality-standards-naaqs-pm
. Accessed 16 December 2020.
41. United States Environmental
Protection Agency. National Ambient Air Quality Standards (NAAQS) for Ozone
. 2020. 40 CFR Part 58. 47. EPA-HQ-OAR-2020-0279. Available at: https://www.epa.gov/ground-level-ozone-pollution/ozone-national-ambient-air-quality-standards-naaqs
. Accessed 16 December 2020.
42. Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research. Epidemiology. 1999;10:37–48.
43. Neighborhood Atlas. University of Wisconsin School of Medicine and Public Health
. 2013. Area Deprivation Index 2.0. Available at: https://www.neighborhoodatlas.medicine.wisc.edu/
. Accessed 20 May 2019.
44. Kind AJH, Buckingham WR. Making neighborhood-disadvantage metrics accessible — the neighborhood atlas. New Engl J Med. 2018;378:2456–2458.
45. Singh GK. Area deprivation and widening inequalities in US mortality, 1969-1998. Am J Public Health. 2003;93:1137–1143.
46. Tessum CW, Paolella DA, Chambliss SE, Apte JS, Hill JD, Marshall JD. PM2.5 polluters disproportionately and systemically affect people of color in the United States. Sci Adv. 2021;7:eabf4491.
47. Peng RD, Dominici F, Pastor-Barriuso R, Zeger SL, Samet JM. Seasonal analyses of air pollution and mortality in 100 US cities. Am J Epidemiol. 2005;161:585–594.
48. Clark DPQ, Son DB, Bowatte G, et al. The association between traffic-related air pollution and obstructive sleep apnea: a systematic review. Sleep Med Rev. 2020;54:101360.
49. Epstein Y, Yanovich R. Heatstroke. N Engl J Med. 2019;380:2449–2459.
50. Boda E, Rigamonti AE, Bollati V. Understanding the effects of air pollution on neurogenesis and gliogenesis in the growing and adult brain. Curr Opin Pharmacol. 2020;50:61–66.
51. Thomson EM. Air pollution, stress, and allostatic load: linking systemic and central nervous system impacts. J Alzheimers Dis. 2019;69:597–614.
52. Anderson FL, Coffey MM, Berwin BL, Havrda MC. Inflammasomes: an emerging mechanism translating environmental
toxicant exposure into neuroinflammation in Parkinson’s disease. Toxicol Sci. 2018;166:3–15.
53. Gómez-Budia M, Konttinen H, Saveleva L, et al. Glial smog: interplay between air pollution and astrocyte-microglia interactions. Neurochem Int. 2020;136:104715.
54. Gondalia R, Baldassari A, Holliday KM, et al. Methylome-wide association study provides evidence of particulate matter air pollution-associated DNA methylation. Environ Int. 2019;132:104723.
55. Rose M, Filiatreault A, Guénette J, Williams A, Thomson EM. Ozone increases plasma kynurenine-tryptophan ratio and impacts hippocampal serotonin receptor and neurotrophic factor expression: role of stress hormones. Environ Res. 2020;185:109483.
56. Calderón-Garcidueñas L, Reynoso-Robles R, González-Maciel A. Combustion and friction-derived nanoparticles and industrial-sourced nanoparticles: the culprit of Alzheimer and Parkinson’s diseases. Environ Res. 2019;176:108574.
57. Cory-Slechta DA, Allen JL, Conrad K, Marvin E, Sobolewski M. Developmental exposure to low level ambient ultrafine particle air pollution and cognitive dysfunction. Neurotoxicology. 2018;69:217–231.
58. Cory-Slechta DA, Sobolewski M, Marvin E, et al. The impact of inhaled ambient ultrafine particulate matter on developing brain: potential importance of elemental contaminants. Toxicol Pathol. 2019;47:976–992.
59. Delgado-Saborit JM, Guercio V, Gowers AM, Shaddick G, Fox NC, Love S. A critical review of the epidemiological evidence of effects of air pollution on dementia, cognitive function and cognitive decline in adult population. Sci Total Environ. 2021;757:143734.