Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Apr 2.
Published in final edited form as: Headache. 2023 Jan 18;63(1):94–103. doi: 10.1111/head.14442

Wildfire smoke exposure and emergency department visits for headache: A case-crossover analysis in California, 2006–2020

Holly Elser 1,2, Sebastian T Rowland 3,4, Maksym S Marek 1, Mathew V Kiang 5, Brittany Shea 3, Vivian Do 3, Tarik Benmarhnia 6, Andrea L C Schneider 1,7, Joan A Casey 3
PMCID: PMC10066880  NIHMSID: NIHMS1883234  PMID: 36651537

Abstract

Objective:

To evaluate the association of short-term exposure to overall fine particulate matter of <2.5 μm (PM2.5) and wildfire-specific PM2.5 with emergency department (ED) visits for headache.

Background:

Studies have reported associations between PM2.5 exposure and headache risk. As climate change drives longer and more intense wildfire seasons, wildfire PM2.5 may contribute to more frequent headaches.

Methods:

Our study included adult Californian members (aged ≥18 years) of a large de-identified commercial and Medicare Advantage claims database from 2006 to 2020. We identified ED visits for primary headache disorders (subtypes: tension-type headache, migraine headache, cluster headache, and “other” primary headache). Claims included member age, sex, and residential zip code. We linked daily overall and wildfire-specific PM2.5 to residential zip code and conducted a time-stratified case-crossover analysis considering 7-day average PM2.5 concentrations, first for primary headache disorders combined, and then by headache subtype.

Results:

Among 9898 unique individuals we identified 13,623 ED encounters for primary headache disorders. Migraine was the most frequently diagnosed headache (N = 5534/13,623 [47.6%]) followed by “other” primary headache (N = 6489/13,623 [40.6%]). For all primary headache ED diagnoses, we observed an association of 7-day average wildfire PM2.5 (odds ratio [OR] 1.17, 95% confidence interval [CI] 0.95–1.44 per 10 μg/m3 increase) and by subtype we observed increased odds of ED visits associated with 7-day average wildfire PM2.5 for tension-type headache (OR 1.42, 95% CI 0.91–2.22), “other” primary headache (OR 1.40, 95% CI 0.96–2.05), and cluster headache (OR 1.29, 95% CI 0.71–2.35), although these findings were not statistically significant under traditional null hypothesis testing. Overall PM2.5 was associated with tension-type headache (OR 1.29, 95% CI 1.03–1.62), but not migraine, cluster, or “other” primary headaches.

Conclusions:

Although imprecise, these results suggest short-term wildfire PM2.5 exposure may be associated with ED visits for headache. Patients, healthcare providers, and systems may need to respond to increased headache-related healthcare needs in the wake of wildfires and on poor air quality days.

Keywords: case-crossover analysis, fine particulate matter of <2.5 μm (PM2.5), headache, wildfires

INTRODUCTION

Anthropogenic climate change is expected to drive increasingly frequent and intense wildfires in the Western United States (US) with implications for population health.1–3 In California, climate conditions conducive to wildfire that include high temperatures, changes in precipitation regimes, and accumulation of dried vegetation coinciding with strong downslope winds have produced historically destructive wildfire seasons.2,4–6 The consequences of wildfires include increased carbon emissions, periods of degraded air quality due to smoke, and negative impacts on municipal budgets.7 The total damages associated with the 2018 wildfire season alone were an estimated $149 billion.8

Smoke exposure from nearby wildfires is increasingly recognized as a health hazard. Fine particulate matter of <2.5 μm in diameter (PM2.5) is the primary air pollutant of concern,9,10 with wildfire-specific PM2.5 (hereafter “wildfire PM2.5”) potentially more health harmful than PM2.5 from fossil fuel combustion (e.g., traffic). Wildfire PM2.5 may result in worse health outcomes because the chemical compositions tend to have higher concentrations of oxidative and pro-inflammatory compounds.10–13 Exposure to wildfire PM2.5 has previously been associated with wide-ranging health outcomes including all-cause mortality14–16; respiratory events17; risk of hospitalization among those with cardiovascular disease18–20; and increased risk of low birth weight and preterm birth.21,22 Although short-term exposure to high levels of PM2.5 (from all sources) has previously been associated with increased risk of emergency department (ED) visits for headache including migraine,23–27 we know of no studies that consider whether wildfire PM2.5 exposure is associated with any headache-related outcomes. Primary headache disorders are extremely prevalent28 and associated with an estimated annual cost of up to $16.6 billion in lost productivity, work and school absences, and short-term disability for migraine alone.29–32 Identification of potential triggers or exacerbating factors for headache such as wildfire PM2.5 could therefore have implications for both patients and clinical practice.

Motivated by recent intensification of wildfires in the Western US and the profound societal burden associated with headache, we examined, for the first time, the association between short-term exposure to wildfire PM2.5 and ED visits for primary headache disorder in California. Our case-crossover analysis leverages a unique administrative database comprised of healthcare claims for privately insured Californians followed from 2006 to 2020 and separately considers exposure to wildfire and overall PM2.5 levels estimated using ensemble machine learning models. We hypothesized a priori that exposure to wildfire PM2.5 would be associated with increased risk of ED visits for primary headache disorders, and that the association would be stronger for wildfire versus overall PM2.5.

METHODS

The Optum Clinformatics® Data Mart Database is a large, de-identified commercial and Medicare Advantage claims database. Patient-level commercial claims data linked to patient residential zip code are available from January 1, 2006, through December 31, 2020. Member enrollment data; diagnostic codes from outpatient, ED, and inpatient visits; and pharmacy claims are deterministically linked across file types with a unique patient identifier.

Using these data, we conducted a case-crossover analysis33–35 to examine the association of wildfire and overall PM2.5 concentration with ED visits for primary headache (i.e., headaches not caused by another disease or injury) in California. Because individuals serve as their own controls in case-crossover analysis, this approach eliminates confounding by time-invariant individual- and area-level confounders and is therefore well-suited to the study of short-term effects of environmental exposures.36–38 Individuals eligible for the present study were aged ≥18 years at the time of their ED visit, resided in California, and had at least one headache-related healthcare encounter within the study period. The sample size for this analysis was determined based on the available data, and thus no power calculation was conducted. There were no missing data. This study was approved by the Institutional Review Boards at Stanford University (Protocol #40974) and the University of Pennsylvania (Protocol #850741). Requirement for written informed consent was waived for this study.

Emergency department visits for primary headache

The primary outcome of interest for this analysis was an ED visit for primary headache identified using diagnostic codes from the International Classification of Diseases, Ninth Revision (ICD-9) and the Tenth Revision (ICD-10). We limited our analysis to ED visits with headache-associated diagnoses in the first, second, or third diagnostic position. Visits for primary headache were a priori further subclassified as related to tension-type headache, migraine headache, cluster headache, and “other” primary headache (Table S1) given differences in clinical severity, treatment, and risk factors.

Fine particulate matter concentrations

First, daily average concentrations of overall PM2.5 were estimated for each California zip code with ensemble models that integrated multiple machine learning algorithms (including random forest, gradient boosting, and deep learning) and predictor variables including outdoor PM2.5 measurements from the Environmental Protection Agency Air Quality System (AQS), aerosol optical depth, plume height, and meteorological variables (precipitation, minimum and maximum temperature, surface shortwave radiation, specific humidity, wind speed and direction) extracted from the high-resolution Gridded Surface Meteorological dataset (gridMET).39

Second, daily measures of wildfire PM2.5 for each zip code were estimated using a spatio-temporal multiple imputation approach. Briefly, PM2.5 concentrations attributable to wildfire smoke were estimated for zip-code-days identified by comparing the observed PM2.5 values to estimated counterfactual values in the absence of wildfire smoke. The wildfire PM2.5 ensemble model achieved high accuracy (R2 = 0.86, root mean square error [RMSE] = 3.48); for additional details, see Aguilera et al. 2021.40 Using daily, zip code-level estimates of exposure, we calculated the average concentrations of wildfire PM2.5 and overall PM2.5 in the 7-day window prior to each case and control date based on participant residential zip code.

Covariates

Additional patient-level demographic characteristics derived from insurance eligibility files included sex (female or male), age category (18–34; 34–54; 55–64; ≥65 years), and calendar year of ED visit. We additionally assigned season for all headache-related visits (winter: December–February; spring: March–May; summer: June–August; fall: September–November).

Statistical analysis

We first calculated the number and percentage of all ED visits for primary headache disorders that occurred by sex, age category, and across headache subtypes. We conducted a time-stratified case-crossover analysis to estimate the association of 7-day average PM2.5 concentrations with ED visits for primary headache. For each ED visit, we selected up to four control dates matched on day of week, month, and calendar year to account for secular trends and seasonality. We used conditional logistic regression analysis in which the unit of analysis is the matched set of ED visits and control dates.

The case-crossover design has five major assumptions for exchangeability between case and control periods: (i) variation in exposure over time; (ii) no selection bias; (iii) exchangeability of risk of exposure; (iv) exposure assessment accuracy is independent of case status; and (v) transient effect.41 Our study satisfies nearly all these assumptions. First, we observed variation of PM2.5 concentrations over time. Second, patient data was derived from ED encounters and inclusion was not predicated on air pollution levels. Third, area-level wildfire and overall PM2.5 concentrations are not impacted by individual headache cases and our time-stratified control selection process has been shown to maximize exchangeability for air pollution and weather exposures.42 Similarly, we estimated PM2.5 concentrations using a validated approach that is independent of any case status. Finally, although research on the effects of wildfire PM2.5 and headache remain limited, previous assessments suggest most health effects occur within the week following the exposure.43,44

We estimated odds ratios (ORs) for the associations of wildfire PM2.5 and overall PM2.5 by comparing concentrations in the 7-day period preceding ED visit dates versus control dates.

Our interpretation of study results extends beyond consideration of statistical significance alone to include the width of the 95% confidence intervals (CIs) and the range of values included therein, the magnitude of estimated ORs, and occurs within the context of previously published scientific research on headache and the environment. Main exposure variables re-scaled such that ORs corresponded to a 10 μg/m3 increase in the concentration of PM2.5. For context, the United States Environmental Protection Agency (EPA) has established a 24-h national air quality standard for fine particle of 35 μg/m345 and a 10 μg/m3 increase in daily total PM2.5 has been associated with ~1% increase in all-cause mortality.46 We first estimated the association with all primary headache diagnoses combined, and then separately for tension-type headache, migraine headache, cluster headache, and “other” primary headache. We fitted conditional logistic regression models to plot the predicted odds of ED visits for all primary headaches as a function of PM2.5 concentrations.

Secondary analyses

As headache risk and severity differs within populations by age and sex, we additionally conducted analyses within these subgroups.47–50 For subgroup analyses, we conducted Cochran’s Q test for heterogeneity to ascertain whether subgroup results differed from overall estimates.51 Cochran’s Q is a two-sided pairwise multiple comparison test with threshold for significance of p < 0.05. Finally, because wildfire activity in California is concentrated in the fall, we estimated the associations of wildfire PM2.5 and overall PM2.5 with ED visits for primary headache restricted to visits that occurred in September, October, and November.

Sensitivity analyses

We considered alternative windows for PM2.5 exposure and repeated our analysis to average 7-day PM2.5 concentration of ≤50 μg/m3 to examine whether the dose–response and overall findings persisted after excluding the highest concentrations of PM2.5. In additional sensitivity analyses, we further identified outpatient visits and inpatient hospitalizations for primary headache disorders. As outpatient visits are often scheduled in advance and because inpatient hospitalizations for primary headache are rare, we did not expect them to be associated with PM2.5 concentrations. We therefore conducted a negative outcome control analysis in which we examined the association of wildfire PM2.5 and overall PM2.5 concentrations with outpatient visits and inpatient hospitalizations for primary headache.52,53

All statistical analyses were conducted using R statistical software version 3.5.6 (R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Among 9898 unique individuals in California aged ≥18 years, we identified 13,623 unique ED encounters for primary headache disorders for which we selected 48,017 control dates matched on weekday, month, and calendar year. Most ED visits occurred among patients who were male (76.3%) and 40.5% were aged between 35 and 54 years. Migraine was the most frequently diagnosed headache (47.6%) followed by “other” primary headache (40.6%), whereas tension-type headache (10.4%) and cluster headache (1.3%) comprised the minority of diagnoses (Table 1). The average concentrations of wildfire PM2.5 and overall PM2.5 by zip code are depicted in Figure 1.

TABLE 1.

Emergency department encounters for primary headache disorders in California.

Variable N (%)
Total 13,623 (100.0)
Sex
 Male 10,395 (76.3)
 Female 3228 (23.7)
Age category, years
 18–34 3272 (24.0)
 35–54 5522 (40.5)
 55–64 1648 (12.1)
 ≥65 3181 (23.4)
Headache type
 Tension type headache 1418 (10.4)
 Migraine headache 5534 (47.6)
 Cluster headache 182 (1.3)
 “Other” primary headache 6489 (40.6)
Open in a new tab

FIGURE 1.

FIGURE 1

Open in a new tab

Average concentrations of wildfire-specific fine particulate matter of <2.5 μm (PM2.5) (A) and overall PM2.5 (B) in California by zip code from January 1, 2006, to December 31, 2020.

In our analysis of wildfire PM2.5 and all primary headache, we estimated that odds of an ED visit increased by 17% for every 10 μg/m3 increase in the 7-day average concentration (OR 1.17, 95% CI 0.95–1.44) (Figure 2, Table S2). By subtype, we observed positive associations between wildfire PM2.5 and ED encounters for tension-type headache (OR 1.42, 95% CI 0.91–2.22), “other” primary headache (OR 1.40, 95% CI 0.96–2.05), and cluster headache (OR 1.29, 95% CI 0.71–2.35), although the estimated association for cluster headache was imprecise due to its relatively rare nature (~1% of all headache cases). We did not observe clear evidence for migraine headache (OR 1.07, 95% CI 0.83–1.37). In our analysis of overall PM2.5 and ED encounters for all primary headache, we observed no clear evidence of an association. In our analysis of headache subtypes, we did not observe any evidence of an association with ED encounters for migraine headache or cluster headache, but we did observe a weak, positive association with “other” primary headache (OR 1.09, 95% CI 0.96–1.24) and a stronger association with tension-type headache (OR 1.29, 95% CI 1.03–1.62) (Figure 2, Table S2). Many of the associations estimated were imprecise, as evidenced by wide CIs.

FIGURE 2.

FIGURE 2

Open in a new tab

Association of wildfire-specific PM2.5 and overall PM2.5 with emergency department (ED) visits for primary headache. We used conditional logistic regression to estimate the association of 7-day average wildfire-specific fine particulate matter of <2.5 μm (PM2.5) concentration (A) and overall PM2.5 concentration with whether an individual had an ED encounter for any primary headache disorder. We additionally considered associations for headache subtypes classified as migraine headache, cluster headache, tension-type headache, and “other” primary headache disorders. Odds ratios (ORs) were estimated using conditional logistic regression models, where the OR reflects the association for a 10 μg/m3 increase in the 7-day average PM2.5 concentration.

Figure 3 depicts the predicted odds of an ED visit for primary headache as a function of wildfire and overall PM2.5 concentration. For both wildfire and overall PM2.5, the odds of an ED visit increase with rising 7-day average concentrations with consistently higher predicted odds for wildfire PM2.5 versus overall PM2.5, but with wider CIs.

FIGURE 3.

FIGURE 3

Open in a new tab

Predicted odds of emergency department (ED) visit as a function of the concentration of fine particulate matter of <2.5 μm (PM2.5). The associations of 7-day average wildfire-specific PM2.5 and overall PM2.5 with ED visits for primary headache disorders were estimated separately using conditional logistic regression. Fitted conditional logistic regression models were then used to plot the predicted odds of ED visits for primary headaches a function of continuous wildfire-specific PM2.5 (A) and overall PM2.5 (B) in μg/m3 with fitted values restricted below the 95th percentile of observed PM2.5.

Secondary analyses

In subgroup analyses, we did not observe meaningful differences between men and women (Cochrane’s Q = 0.86, p = 0.36) (Table S3). By age category, we observed the strongest association between wildfire PM2.5 and headache among patients aged 18–34 years (OR 1.13, 95% CI 0.80–1.58) and the strongest association between overall PM2.5 and headache among those aged ≥65 years (OR 1.18, 95% CI 0.99–1.40; Cochran’s Q = 354.69, p < 0.01) (Table S4). In analysis restricted to ED visits that occurred in the fall, associations were generally increased in magnitude for all primary headache and subtype groups in our analysis of both wildfire PM2.5 and overall PM2.5. (Table S5).

Sensitivity analyses

In sensitivity analysis of alternative exposure windows ranging from same-day to 7-day average, we observed increasing strength of association with increased length of exposure window (Table S6). Analysis with values of PM2.5 restricted to concentrations of ≤50 μg/m3 were consistent with those of our main analysis (Table S7). Finally, in our negative outcome control analysis we observed reduced odds of inpatient hospitalizations for primary headache associated with a 10 μg/m3 increase in 7-day average wildfire PM2.5 (OR 0.97, 95% CI 0.96–0.98) and no relation with outpatient visits (OR 0.98, 95% CI 0.90–1.08). For overall PM2.5 we observed no evidence of an association with inpatient hospitalizations but reduced odds of outpatient visits (OR 0.98, 95% CI 0.97–0.99) for primary headache (Table S8).

DISCUSSION

In this case-crossover study, we present a novel analysis of the association between short-term exposure to wildfire PM2.5 and ED visits for headache in California. Many of the estimated associations we present do not reach the threshold for statistical significance as the 95% CI includes the null value. Our results may nevertheless provide evidence for an underlying relation between wildfire smoke exposure and ED visits for primary headache disorders. First, the estimated ORs deviate meaningfully from the null value,54 particularly in the context of prior studies population health effects of environmental exposure in which associations are often modest in magnitude.55–57 Second, we interpret our results in the context of some prior research that suggests common pollutants may trigger headache. Finally, we consider our results in the context of accumulating evidence that wildfire smoke exposure negatively affects a wide range of health outcomes.58 Our findings thus provide preliminary evidence that short-term exposure to wildfire PM2.5 may be associated with ED visits for primary headache disorders and points to the need for public health action to mitigate these exposures.

In our analysis, associations were stronger in magnitude for wildfire PM2.5 versus overall PM2.5 for primary headache disorders overall and generally across headache subtypes. Whereas background urban particulate matter is produced by fossil fuel combustion (e.g., traffic emissions), wildfire produces PM2.5 through combustion of organic materials at substantially higher temperatures.59 Particulate matter from wildfires has been previously demonstrated to contain more pro-inflammatory and oxidative compounds10 and on average has smaller particle size.60 These features may collectively increase the likelihood of headache exacerbation.61 The composition of wildfire PM2.5 also differs by ecoregion, and smoke from several wildfires can combine to produce distinct mixtures with possibly varying health impacts.62–64

We found an association between wildfire PM2.5 and all primary headache, estimating that odds of an ED visit increased by 17% for every 10 μg/m3 increase in the 7-day average concentration. Prior studies that examine the association between particulate matter generally and primary headache have showed mixed results. A study by Mukamal et al.65 did not find clear association between ED encounter for headache and PM2.5 levels in the days (0–2 days) prior. Vodonos et al.66 found a slightly increased risk of ED headache visits with increased levels of PM10 on the same day in autumn (fall). A study by Özsaraç et al.67 found an association between increased cumulative monthly number of ED visits for migraine and “other” primary headache disorders as the chief complaint and increased monthly mean levels of PM10. Chiu et al.68 found an association between short-term, average levels on the same and previous 2 days, exposure to air pollution including PM10 and clinic visits for headache.

Another study by Dales et al.69 found an increased risk of headache hospitalizations from acute increases (24-h mean concentrations) in air pollution including PM2.5 and PM10 in Santiago Province of Chile using optimal lags of 1 day for all relative risks. Chang et al.70 found an increased risk of outpatient headache visits associated with days with higher PM2.5 levels in Taipei, Taiwan. Using data from Edmonton, Canada ED visits from 1992 to 2002, Szyszkowicz et al.24 found increased frequency of visits for general headache, as well as migraine, associated with exposure to increased concentrations of PM2.5 lagged by 0, 1, and 2 days only among female patients during the cold season (October–March). A study by Szyszkowicz and Kousha71 from April 2004 to 2011 in nine cities in Ontario, Canada found an association between air pollutant levels, including PM2.5, lagged by 0–8 days and the number of ED visits for headache, in addition to migraine, among females.

Prior studies of short-term PM2.5 exposure have focused primarily on migraine. A second study by Szyszkowicz et al.23 found an association between SO2 and PM2.5 levels, lagged by 2 days, and ED visits in a time-series analysis of ED visits across five cities in Canada. In their case-crossover analysis of data from Taipei, Taiwan from 2006 to 2011, Chen et al.26 found that outpatient clinic visits for migraine were only associated with increased concentrations of PM2.5 on warm days (i.e., mean daily temperature ≥23°C). Among 18,921 patients in Seoul who visited the ED with migraine as the primary complaint, Lee et al.25 found each interquartile range increase in short-term PM2.5 concentrations was associated with a 3% increase in ED visits for migraine. Finally, Elser et al.72 found that annual average PM2.5 concentrations were linearly associated with measures of migraine severity including outpatient neurology visits and migraine-specific urgent care visits among Sutter Health patients in Northern California from 2014 to 2018. Although notably Li et al.73 did not find an association between particulate matter exposure and headache among 98 individuals in the Greater Boston area assessed via electronic survey from 2016 to 2017.

In contrast with some of this prior research, we found no clear evidence of an association between wildfire PM2.5 and ED visits for migraine. One possible explanation is that wildfire PM2.5 exposure does not precipitate migraine headache, although this is less likely given prior research that links pollution with migraine onset. Another potential explanation is that individuals with known migraine headache disorder can avoid being treated in the ED by taking prophylactic medications or at-home abortive therapies (i.e., non-steroidal anti-inflammatory drugs and triptans)74 or by changing their behaviors such as by wearing an N95 mask or using air filtration systems. Finally, it is possible that migraine headache is frequently misclassified as “other” primary headache disorder.

Beyond migraine, our analysis considers multiple subtypes of primary headache. We observed an association between wildfire PM2.5 exposure and ED encounter for tension-type and “other” primary headache. We also find an association between wildfire PM2.5 exposure and ED visits for cluster headache. Cluster headache is typified by recurrent, unilateral pain that involves the orbital or periorbital region with associated lacrimation, rhinorrhea, conjunctival injection, ptosis or miosis, and periorbital edema.75 Individual attacks are rapid in onset characterized by excruciating pain that typically last 15–180 min and tend to “cluster” together occurring as frequently as every other day for several weeks separated by periods of remission.76,77

Because of the severe nature of these headaches, management is focused primarily on pharmacological prevention, e.g., with calcium channel blocking agents (e.g., verapamil).78 To our knowledge, no prior studies have assessed the association of short-term wildfire PM2.5, or any air pollutant, with cluster headache. Although imprecise, our finding is potentially important given the severe and highly morbid nature of cluster headaches. While studying this outcome is complicated due to the long time-period over which the attacks can occur, if wildfire smoke is indeed associated with cluster headache, these individuals may therefore take extra care to mitigate their exposure to avoid subsequent attacks. More broadly, examination of the role of various environmental exposures in the onset of cluster headache including PM2.5 may be an important area for future research.

LIMITATIONS

Data for this study were derived from a database of private and Medicare Advantage claims. Our results therefore do not reflect the experiences of the poorest Californians. Yet, headache follows a social gradient and is less common among wealthier individuals.79,80 Similarly, wildfires may disproportionally affect vulnerable populations including indigenous and migrant populations, older and low-income individuals.3,81,82 Further examination of the implications of wildfires for headache that specifically considers potential modifications by individual, and community-level social class and material resources is needed. Although such individual- and area-level covariates were not available in these data, we control for any confounding by time-invariant characteristics with the case-crossover design because individuals serve as their own controls.

In the setting of wildfires, PM2.5 co-occurs with several other toxicants. Additionally, major wildfire events may stoke significant anxiety and fear among residents.83,84 It is therefore possible that the associations presented here reflect some combination of these factors rather than the effects of PM2.5 alone. These data did not include measures of factors such as time spent indoors, housing quality and infiltration of smoke, or whether the individual commuted to work outside of their residential zip code, all of which could confound the association between wildfire smoke exposure and headache. In the present study, confounding by these factors should be minimized by the case-crossover design in which control periods are matched to dates of ED visits on weekday, month, and calendar year within an individual. Further, we assigned PM2.5 exposure metrics at the patient residential zip code level based on data availability. This method assumes constant PM2.5 concentrations within zip codes, while heterogeneity may in fact exist. Future studies could improve on our design by using address-level healthcare and patient-level mobility data to reduce exposure misclassification. Our study lacked sufficient statistical power to examine heterogeneity in the association between wildfire PM2.5 and headache-related ED visits by sex and age group. As previously mentioned, the estimated association between wildfire PM2.5 and ED visits for cluster headache was also imprecise, likely due to the relatively rare nature of this headache disorder. Because most outpatient visits are scheduled in advance, we considered outpatient visits for primary headache disorders as a negative outcome control. However, some percentage of outpatient visits are comprised of urgent care and same-day appointments, which could plausibly be expected to occur more frequently among symptomatic individuals in response to wildfire smoke exposure.

Finally, we know of no studies that have validated ED claims for primary headache disorders, and thus the outcome measure is potentially subject to misclassification particularly across subtypes of primary headache disorders. However, we have no reason to suspect this misclassification varies systematically based on levels of PM2.5. Moreover, ED visits capture only the subset of individuals experiencing headache severe enough that it prompts treatment seeking in the acute care setting. Our outcome measure is meaningful insofar as it reflects the immediate headache-related burden imposed on local healthcare systems by wildfire PM2.5. However, there is an important role for future research that considers implication of wildfire for headache wherever more sensitive measures of headache, such as self-report, are available.

CONCLUSION

This case-crossover study presents a novel analysis of the association between exposure to wildfire PM2.5 exposure and ED visits for primary headache disorders in California, where anthropogenic climate change, fire management practices, and population shifts have produced historically devastating wildfires in the recent past. Our findings provide the first preliminary evidence that short-term exposure to wildfire PM2.5 may be associated with ED visits for primary headache disorders. Providers and healthcare systems may need to respond to increased headache-related healthcare needs. Individuals with headache may need to avoid going outdoors in the wake of wildfires and on poor air quality days.

Supplementary Material

Supplement

FUNDING INFORMATION

Mathew V. Kiang was supported by the National Institutes on Drug Abuse R00DA051534. Andrea L.C. Schneider was supported by the National Institute of Neurological Disorders and Stroke K23NS123340, and Joan A. Casey was supported by the National Institute of Environmental Health Sciences R00ES027023 and the National Institute on Aging RF1AG071024. The funders had no role in study design, data collection, analysis, decision to publish, or preparation of the manuscript.

Abbreviations:

CI

confidence Interval

ED

emergency department

OR

odds ratio

PM2.5

fine particulate matter of <2.5 μm

Footnotes

CONFLICTS OF INTEREST

Holly Elser, Sebastian T. Rowland, Maksym S. Marek, Mathew V. Kiang, Brittany Shea, Vivian Do, Tarik Benmarhnia, Andrea L.C. Schneider, and Joan A. Casey declare no conflicts of interest.

SUPPORTING INFORMATION

Additional supporting information can be found online in the Supporting Information section at the end of this article.

REFERENCES

  • 1.Abatzoglou JT, Williams AP. Impact of anthropogenic climate change on wildfire across western US forests. Proc Natl Acad Sci U S A. 2016;113(42):11770–11775. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Williams AP, Abatzoglou JT, Gershunov A, et al. Observed impacts of anthropogenic climate change on wildfire in California. Earths Future. 2019;7(8):892–910. [Google Scholar]
  • 3.Burke M, Driscoll A, Heft-Neal S, Xue J, Burney J, Wara M. The changing risk and burden of wildfire in the United States. Proc Natl Acad Sci U S A. 2021;118(2):e2011048118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Westerling AL, Hidalgo HG, Cayan DR, Swetnam TW. Warming and earlier spring increase western US forest wildfire activity. Science. 2006;313(5789):940–943. [DOI] [PubMed] [Google Scholar]
  • 5.Seager R, Hooks A, Williams AP, Cook B, Nakamura J, Henderson N. Climatology, variability, and trends in the US vapor pressure deficit, an important fire-related meteorological quantity. J Appl Meteorol Climatol. 2015;54(6):1121–1141. [Google Scholar]
  • 6.Li S, Banerjee T. Spatial and temporal pattern of wildfires in California from 2000 to 2019. Sci Rep. 2021;11(1):1–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Liao Y, Kousky C. The fiscal impacts of wildfires on California municipalities. J Assoc Environ Resour Econ. 2022;9(3):455–493. [Google Scholar]
  • 8.Wang D, Guan D, Zhu S, et al. Economic footprint of California wildfires in 2018. Nat Sustain. 2021;4(3):252–260. [Google Scholar]
  • 9.Stone SL, Anderko L, Berger M, Butler C, Cascio W, Clune A. Wildfire Smoke: A Guide for Public Health Officials. US Environmental Protection Agency; 2019. [Google Scholar]
  • 10.Verma V, Polidori A, Schauer JJ, Shafer MM, Cassee FR, Sioutas C. Physicochemical and toxicological profiles of particulate matter in Los Angeles during the October 2007 southern California wildfires. Environ Sci Technol. 2009;43(3):954–960. [DOI] [PubMed] [Google Scholar]
  • 11.Balmes JR. The changing nature of wildfires: impacts on the health of the public. Clin Chest Med. 2020;41(4):771–776. [DOI] [PubMed] [Google Scholar]
  • 12.Masselot P, Sera F, Schneider R, et al. Differential mortality risks associated with PM2.5 components: a multi-country, multi-city study. Epidemiology. 2021;33(2):167–175. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Wegesser TC, Pinkerton KE, Last JA. California wildfires of 2008: coarse and fine particulate matter toxicity. Environ Health Perspect. 2009;117(6):893–897. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Reid CE, Brauer M, Johnston FH, Jerrett M, Balmes JR, Elliott CT. Critical review of health impacts of wildfire smoke exposure. Environ Health Perspect. 2016;124(9):1334–1343. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Black C, Tesfaigzi Y, Bassein JA, Miller LA. Wildfire smoke exposure and human health: significant gaps in research for a growing public health issue. Environ Toxicol Pharmacol. 2017;55:186–195. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Liu C, Chen R, Sera F, et al. Ambient particulate air pollution and daily mortality in 652 cities. N Engl J Med. 2019;381(8):705–715. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Reid CE, Maestas MM. Wildfire smoke exposure under climate change: impact on respiratory health of affected communities. Curr Opin Pulm Med. 2019;25(2):179–187. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.DeFlorio-Barker S, Crooks J, Reyes J, Rappold AG. Cardiopulmonary effects of fine particulate matter exposure among older adults, during wildfire and non-wildfire periods, in the United States 2008–2010. Environ Health Perspect. 2019;127(3):037006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Gan RW, Ford B, Lassman W, et al. Comparison of wildfire smoke estimation methods and associations with cardiopulmonary-related hospital admissions. GeoHealth. 2017;1(3):122–136. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Bhatnagar A Cardiovascular effects of particulate air pollution. Annu Rev Med. 2022;73:393–406. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Holstius DM, Reid CE, Jesdale BM, Morello-Frosch R. Birth weight following pregnancy during the 2003 Southern California wildfires. Environ Health Perspect. 2012;120(9):1340–1345. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Cândido da Silva AM, Moi GP, Mattos IE, Hacon SS. Low birth weight at term and the presence of fine particulate matter and carbon monoxide in the Brazilian Amazon: a population-based retrospective cohort study. BMC Pregnancy Childbirth. 2014;14(1):1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Szyszkowicz M, Kaplan G, Grafstein E, Rowe B. Emergency department visits for migraine and headache: a multi-city study. Int J Occup Med Environ Health. 2009;22(3):235–242. [DOI] [PubMed] [Google Scholar]
  • 24.Szyszkowicz M, Stieb DM, Rowe BH. Air pollution and daily ED visits for migraine and headache in Edmonton, Canada. Am J Emerg Med. 2009;27(4):391–396. [DOI] [PubMed] [Google Scholar]
  • 25.Lee H, Myung W, Cheong H-K, et al. Ambient air pollution exposure and risk of migraine: synergistic effect with high temperature. Environ Int. 2018;121:383–391. [DOI] [PubMed] [Google Scholar]
  • 26.Chen C-C, Tsai S-S, Yang C-Y. Association between fine particulate air pollution and daily clinic visits for migraine in a subtropical city: Taipei, Taiwan. Int J Environ Res Public Health. 2015;12(5):4697–4708. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Chiu H-F, Yang C-Y. Air pollution and daily clinic visits for migraine in a subtropical city: Taipei, Taiwan. J Toxicol Environ Health A. 2015;78(9):549–558. [DOI] [PubMed] [Google Scholar]
  • 28.Stovner LJ, Nichols E, Steiner TJ, et al. Global, regional, and national burden of migraine and tension-type headache, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Neurol. 2018;17(11):954–976. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Lofland J Impact of migraine headache in the United States. Adv Stud Pharm. 2007;4:8–10. [Google Scholar]
  • 30.Porter JK, Di Tanna GL, Lipton RB, Sapra S, Villa G. Costs of acute headache medication use and productivity losses among patients with migraine: insights from three randomized controlled trials. Pharmacoeconomics Open. 2019;3(3):411–417. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Berg J, Ramadan NM. Societal Burden of the Headache. The Headaches. Lipencott & Williams; 2006:35–42. [Google Scholar]
  • 32.Gilligan AM, Foster SA, Sainski-Nguyen A, Sedgley R, Smith D, Morrow P. Direct and indirect costs among United States commercially insured employees with migraine. J Occup Environ Med. 2018;60(12):1120–1127. [DOI] [PubMed] [Google Scholar]
  • 33.Maclure M The case-crossover design: a method for studying transient effects on the risk of acute events. Am J Epidemiol. 1991;133(2):144–153. [DOI] [PubMed] [Google Scholar]
  • 34.Maclure M, Mittleman MA. Should we use a case-crossover design? Annu Rev Public Health. 2000;21(1):193–221. [DOI] [PubMed] [Google Scholar]
  • 35.Armstrong BG, Gasparrini A, Tobias A. Conditional Poisson models: a flexible alternative to conditional logistic case cross-over analysis. BMC Med Res Methodol. 2014;14(1):1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Wichmann J, Ketzel M, Ellermann T, Loft S. Apparent temperature and acute myocardial infarction hospital admissions in Copenhagen, Denmark: a case-crossover study. Environ Health. 2012;11(1):1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Kyobutungi C, Grau A, Stieglbauer G, Becher H. Absolute temperature, temperature changes and stroke risk: a case-crossover study. Eur J Epidemiol. 2005;20(8):693–698. [DOI] [PubMed] [Google Scholar]
  • 38.Buckley JP, Richardson DB. Seasonal modification of the association between temperature and adult emergency department visits for asthma: a case-crossover study. Environ Health. 2012;11(1):1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Abatzoglou JT. Development of gridded surface meteorological data for ecological applications and modelling. Int J Climatol. 2013;33(1):121–131. [Google Scholar]
  • 40.Aguilera R, Luo N, Basu R, Wu J, Gershunov A, Benmarhnia T. Using Machine Learning to Estimate Wildfire PM2.5 at California ZIP Codes (2006–2020). ChemRxiv; 2021. [Google Scholar]
  • 41.Mittleman MA, Mostofsky E. Exchangeability in the case-crossover design. Int J Epidemiol. 2014;43(5):1645–1655. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Mittleman MA. Optimal referent selection strategies in case-crossover studies: a settled issue. Epidemiology. 2005;16(6):715–716. [DOI] [PubMed] [Google Scholar]
  • 43.Chen G, Guo Y, Yue X, et al. Mortality risk attributable to wildfire-related PM2.5 pollution: a global time series study in 749 locations. Lancet Planet Health. 2021;5(9):e579–e587. [DOI] [PubMed] [Google Scholar]
  • 44.Gutiérrez-Avila I, Rojas-Bracho L, Riojas-Rodríguez H, Kloog I, Just AC, Rothenberg SJ. Cardiovascular and cerebrovascular mortality associated with acute exposure to PM2.5 in Mexico City. Stroke. 2018;49(7):1734–1736. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.United States Environmental Protection Agency. National Ambient Air Quality Standards (NAAQS) for PM. EPA; 2022. Accessed October 1, 2022. https://www.epa.gov/pm-pollution/national-ambient-air-quality-standards-naaqs-pm [Google Scholar]
  • 46.Achilleos S, Kioumourtzoglou M-A, Wu C-D, Schwartz JD, Koutrakis P, Papatheodorou SI. Acute effects of fine particulate matter constituents on mortality: a systematic review and meta-regression analysis. Environ Int. 2017;109:89–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Stewart WF, Lipton RB, Celentano DD, Reed ML. Prevalence of migraine headache in the United States: relation to age, income, race, and other sociodemographic factors. JAMA. 1992;267(1):64–69. [PubMed] [Google Scholar]
  • 48.Rasmussen BK. Epidemiology of headache. Cephalalgia. 1995;15(1):44–67. [DOI] [PubMed] [Google Scholar]
  • 49.Peterlin BL, Gupta S, Ward TN, MacGregor A. Sex matters: evaluating sex and gender in migraine and headache research. Headache. 2011;51:839–842. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Silberstein S The role of sex hormones in headache. Neurology. 1992;42(3 Suppl 2):37–42. [PubMed] [Google Scholar]
  • 51.Kaufman JS, MacLehose RF. Which of these things is not like the others? Cancer. 2013;119(24):4216–4222. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Shi X, Miao W, Tchetgen ET. A selective review of negative control methods in epidemiology. Curr Epidemiol Rep. 2020;7(4):190–202. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Lipsitch M, Tchetgen ET, Cohen T. Negative controls: a tool for detecting confounding and bias in observational studies. Epidemiology. 2010;21(3):383–388. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Greenland S, Senn SJ, Rothman KJ, et al. Statistical tests, P values, confidence intervals, and power: a guide to misinterpretations. Eur J Epidemiol. 2016;31(4):337–350. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Wei Y, Wang Y, Di Q, et al. Short term exposure to fine particulate matter and hospital admission risks and costs in the Medicare population: time stratified, case crossover study. BMJ. 2019;367:16258. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.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. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Cheng J, Xu Z, Bambrick H, et al. Cardiorespiratory effects of heat-waves: a systematic review and meta-analysis of global epidemiological evidence. Environ Res. 2019;177:108610. [DOI] [PubMed] [Google Scholar]
  • 58.Xu R, Yu P, Abramson MJ, et al. Wildfires, global climate change, and human health. N Engl J Med. 2020;383(22):2173–2181. [DOI] [PubMed] [Google Scholar]
  • 59.Vardoulakis S, Jalaludin BB, Morgan GG, Hanigan IC, Johnston FH. Bushfire smoke: urgent need for a national health protection strategy. Med J Aust. 2020;212(8):349–353.e1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Makkonen U, Hellén H, Anttila P, Ferm M. Size distribution and chemical composition of airborne particles in South-Eastern Finland during different seasons and wildfire episodes in 2006. Sci Total Environ. 2010;408(3):644–651. [DOI] [PubMed] [Google Scholar]
  • 61.Shaposhnikov D, Revich B, Bellander T, et al. Mortality related to air pollution with the Moscow heat wave and wildfire of 2010. Epidemiology. 2014;25(3):359–364. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62.Liu JC, Peng RD. The impact of wildfire smoke on compositions of fine particulate matter by ecoregion in the Western US. J Expo Sci Environ Epidemiol. 2019;29(6):765–776. [DOI] [PubMed] [Google Scholar]
  • 63.Alves C, Vicente A, Nunes T, et al. Summer 2009 wildfires in Portugal: emission of trace gases and aerosol composition. Atmos Environ. 2011;45(3):641–649. [Google Scholar]
  • 64.Korontzi S, Ward DE, Susott RA, et al. Seasonal variation and ecosystem dependence of emission factors for selected trace gases and PM2.5 for southern African savanna fires. J Geophys Res Atmos. 2003;108(D24). doi: 10.1029/2003JD003730 [DOI] [Google Scholar]
  • 65.Mukamal KJ, Wellenius GA, Suh HH, Mittleman MA. Weather and air pollution as triggers of severe headaches. Neurology. 2009;72(10):922–927. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 66.Vodonos A, Novack V, Zlotnik Y, Ifergane G. Ambient air pollution, weather and daily emergency department visits for headache. Cephalalgia. 2015;35(12):1085–1091. [DOI] [PubMed] [Google Scholar]
  • 67.Özsaraç M, Uluer H, Ersel M, Kıyan S, Yürüktümen A, Ersan E. Effects of particulate air pollution on emergency department visits for headache as chief complaint. Turk J Emerg Med. 2009;9(3):105–108. [Google Scholar]
  • 68.Chiu H-F, Weng Y-H, Chiu Y-W, Yang C-Y. Air pollution and daily clinic visits for headache in a subtropical city: Taipei, Taiwan. Int J Environ Res Public Health. 2015;12(2):2277–2288. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 69.Dales RE, Cakmak S, Vidal CB. Air pollution and hospitalization for headache in Chile. Am J Epidemiol. 2009;170(8):1057–1066. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 70.Chang C-C, Chiu H-F, Yang C-Y. Fine particulate air pollution and outpatient department visits for headache in Taipei, Taiwan. J Toxicol Environ Health A. 2015;78(8):506–515. [DOI] [PubMed] [Google Scholar]
  • 71.Szyszkowicz M, Kousha T. Air pollution and emergency department visits for headache and migraine. Health Scope. 2016;5(3):e35122. [Google Scholar]
  • 72.Elser H, Morello-Frosch R, Jacobson A, et al. Air pollution, methane super-emitters, and oil and gas wells in northern California: the relationship with migraine headache prevalence and exacerbation. Environ Health. 2021;20(1):1–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 73.Li W, Bertisch SM, Mostofsky E, Buettner C, Mittleman MA. Weather, ambient air pollution, and risk of migraine headache onset among patients with migraine. Environ Int. 2019;132:105100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 74.Gelfand AA, Goadsby PJ. A neurologist’s guide to acute migraine therapy in the emergency room. Neurohospitalist. 2012;2(2):51–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 75.Nesbitt AD, Goadsby PJ. Cluster headache. BMJ. 2012;344:e2407. [DOI] [PubMed] [Google Scholar]
  • 76.Arnold M Headache Classification Committee of the International Headache Society (IHS). The International Classification of Headache Disorders. Cephalalgia. 2018;38(1):1–211. [DOI] [PubMed] [Google Scholar]
  • 77.Bahra A, May A, Goadsby PJ. Cluster headache: a prospective clinical study with diagnostic implications. Neurology. 2002;58(3): 354–361. [DOI] [PubMed] [Google Scholar]
  • 78.Leone M, D’amico D, Frediani F, et al. Verapamil in the prophylaxis of episodic cluster headache: a double-blind study versus placebo. Neurology. 2000;54(6):1382–1385. [DOI] [PubMed] [Google Scholar]
  • 79.Lipton RB, Bigal ME, Diamond M, Freitag F, Reed M, Stewart WF. Migraine prevalence, disease burden, and the need for preventive therapy. Neurology. 2007;68(5):343–349. [DOI] [PubMed] [Google Scholar]
  • 80.Lipton RB, Stewart WF, Diamond S, Diamond ML, Reed M. Prevalence and burden of migraine in the United States: data from the American migraine study II. Headache. 2001;41(7):646–657. [DOI] [PubMed] [Google Scholar]
  • 81.Kizer KW. Extreme wildfires—a growing population health and planetary problem. JAMA. 2020;324(16):1605–1606. [DOI] [PubMed] [Google Scholar]
  • 82.Masri S, Scaduto E, Jin Y, Wu J. Disproportionate impacts of wildfires among elderly and low-income communities in California from 2000–2020. Int J Environ Res Public Health. 2021;18(8):3921. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 83.Belleville G, Ouellet M-C, Lebel J, et al. Psychological symptoms among evacuees from the 2016 Fort McMurray wildfires: a population-based survey one year later. Front Public Health. 2021;9:444. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 84.Brown MR, Pazderka H, Agyapong VI, et al. Mental health symptoms unexpectedly increased in students aged 11–19 years during the 3.5 years after the 2016 Fort McMurray wildfire: findings from 9,376 survey responses. Front Psych. 2021;12:676256. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement
NIHMS1883234-supplement-Supplement.docx (40KB, docx)

RESOURCES