Differential Effects of Wildfire Smoke Fine Particulate Matter Exposure on Respiratory Disease Emergency Department Visits in the Western United States

Abstract

Rationale:

Wildfires significantly affect air quality in the western United States. Although prior research has linked wildfire smoke fine particulate matter (i.e., aerodynamic diameter ≤2.5 μm [PM_2.5]) to respiratory health outcomes, these studies typically have limited geographic and temporal coverage, lacking evidence from multiple states over extended periods.

Objectives:

To examine the differential acute effects of wildfire smoke and nonsmoke PM_2.5 exposure on respiratory disease emergency department visits in the western United States.

Methods:

We obtained data on more than 6 million emergency department visits for respiratory diseases, including asthma, chronic obstructive pulmonary disease, upper respiratory infections (URIs), and bronchitis, from five states in the western United States during 2007–2018. Daily exposure to wildfire smoke and nonsmoke PM_2.5 was estimated using a state-of-the-art model system. A time-stratified case cross-over design with conditional logistic regression models was used to assess the acute respiratory effects of smoke and nonsmoke PM_2.5 exposure.

Measurements and Main Results:

The odds ratios associated with a 1-μg/m³ increase in the 3-day average wildfire smoke PM_2.5 were 1.016 (95% confidence interval, 1.015–1.016) for asthma, 1.004 (1.003–1.005) for chronic obstructive pulmonary disease, 1.001 (1.000–1.011) for URIs, and 1.004 (1.004–1.004) for bronchitis. Wildfire smoke PM_2.5 had stronger estimated effects than nonsmoke PM_2.5, particularly for asthma (nonsmoke PM_2.5: odds ratio, 1.002; 95% confidence interval, 1.001–1.004). Stratified analyses showed greater vulnerability among women and adults. Sensitivity analyses confirmed robust associations across exposure windows, and concentration–response functions suggested no clear threshold for adverse effects.

Conclusions:

Wildfire smoke PM_2.5 was associated with increased risks of acute respiratory outcomes. Stronger effects were observed from smoke PM_2.5 than nonsmoke PM_2.5, particularly for asthma and URIs. These findings underscore the need for targeted public health interventions and further research into the unique toxicological properties of wildfire emissions.

Driven by the impacts of climate change, the frequency and intensity of wildfire events have surged in recent decades, establishing them as a global public health concern (1, 2). The western United States is among the regions most affected by increasingly severe wildfire events as a result of its distinct climate, unique vegetation, and escalating drought (3–5). From the 1950s to the 2010s, the frequency of wildfire events in this regions has increased exponentially, and the average burned area per wildfire increased more than doubled, from 1,204 acres to 3,474 acres (5). Smoke released from wildfire events contains a complex mixture of air pollutants, including significant amounts of fine particulate matter (i.e., with an aerodynamic diameter ≤2.5 μm [PM_2.5]) (6, 7). In the western United States, wildfire-emitted smoke PM_2.5 significantly contributes to increased ambient PM_2.5 concentrations (8, 9). Previous studies have suggested that wildfire smoke can account for as much as 76% of the summer mean PM_2.5 in the Pacific Northwest and 40% in the entire western United States (10, 11). In some areas of southern California, the annual average smoke-specific PM_2.5 concentration can reach 25 μg/m³, far exceeding the current annual U.S. national ambient air quality standard for PM_2.5 of 9 μg/m³ (9).

The health effects of ambient PM_2.5 are well documented, but evidence of wildfire smoke PM_2.5’s respiratory impacts remains inconsistent. Prior studies show conflicting results on respiratory infections and often focus on specific populations or limited spatiotemporal ranges, leading to uncertainty about smoke PM_2.5’s short-term effects on the general population (12–21). However, previous studies conducted in the United States have typically been limited to single states or specific wildfire events, resulting in a lack of comprehensive evidence derived from multiple states over an extended period. Exposure assessment of wildfire smoke PM_2.5 in prior studies has been challenging because of difficulties in isolating wildfire contributions from ground-level PM_2.5 measurements. To address this, many previous studies have applied satellite-derived models in assessing wildfire smoke PM_2.5 exposure. Despite the coarse spatial resolution and limited coverage of some earlier work, combining the satellite-observed wildfire smoke with machine-learning algorithms can effectively distinguish smoke PM_2.5 from other sources across large spatial and temporal scales (13, 14, 17, 22–25). Therefore, there is a critical need for comprehensive, long-period studies that leverage these advanced exposure assessment methods to evaluate the respiratory impacts of wildfire smoke PM_2.5 across large spatial areas.

Specific populations are likely to be particularly vulnerable to the health impacts of wildfire smoke as a result of a combination of biological, socioeconomic, and behavioral factors (26). A recent systematic review found inconsistent evidence regarding differential vulnerability by age group, although it consistently reported that female individuals experience higher risks from wildfire smoke exposure (17, 21). Moreover, epidemiology studies suggest that wildfire-emitted smoke PM_2.5 may lead to stronger respiratory health effects than ambient PM_2.5 (17, 27, 28). Such differences may be attributed to variations in particle size distribution and unique chemical compositions, potentially leading to distinct biological responses (29–31). Particles of different sizes exhibit varying abilities to penetrate the alveolar barrier and enter the circulatory system, inducing oxidative stress and vascular dysfunction in the endothelium (32).

In this study, we aimed to extend the evidence base beyond previous studies limited to short durations, limited geographic coverage, insufficient exposure assessments, and inadequate differentiation between smoke and nonsmoke PM_2.5. Specifically, we leveraged robust emergency department (ED) visits data from five states in the western United States from 2007 to 2018 to examine the associations between acute wildfire smoke PM_2.5 exposure and respiratory diseases, including asthma, bronchitis, chronic obstructive pulmonary disease (COPD), and upper respiratory infections (URIs). Using a state-of-the-art exposure modeling system with a large, represented sample population, we investigated the differential acute respiratory effects of wildfire smoke PM_2.5 and nonsmoke PM_2.5, with a focus on understanding the differential impacts among sex, age, race, and ethnicity groups. Additionally, we investigated the concentration–response functions of wildfire smoke PM_2.5 exposure. Some of the results of these studies have been previously reported in the form of an abstract (33).

Methods

ED Visits

We obtained daily patient-level ED visit records from five western U.S. states (Arizona, California, Nevada, Oregon, and Utah) covering 57,303,211 residents across a total area of 1,480,262 km² during the wildfire seasons (May to October) from 2007 to 2018 (34). ED visits were defined as medical care provided by ED units encompassing outpatient visits and those subsequently admitted for inpatient services. Each record includes the date of admission, patient age group, sex, residential zip code, race, ethnicity, and the International Classification of Diseases, 9th and 10th Revisions (ICD-9 and ICD-10) codes for primary and secondary diagnoses. For this study, we focused solely on records with a primary diagnosis of specific respiratory outcomes, including asthma, bronchitis, COPD, URIs, and a total respiratory disease group. The ICD-9 and ICD-10 codes corresponding to these outcomes are detailed in Table E1 in the data supplement, and the detailed ED cases by states are shown in Table E2. The institutional review board at Emory University granted approval for this study and issued a waiver of informed consent requirements in view of the impracticality of obtaining consent from each individual patient and the minimal risk posed by the research (approval code STUDY00004823).

Exposure Assessment

To accurately differentiate and estimate wildfire smoke–specific PM_2.5 from ambient PM_2.5, we applied a two-stage, twin-model machine-learning system powered by satellite remote sensing data, meteorology reanalysis, land use information, and ground PM_2.5 observations. Detailed information on this framework is published elsewhere (9). These models generated daily smoke-specific PM_2.5 and nonsmoke PM_2.5 concentrations at a 1-km resolution, which were then aggregated to zip code level to match ED visit records. In Nevada, matching was based on the first four digits of the zip code because of ED visit data confidentiality restrictions. We further calculated 2-day to 6-day moving average concentrations of each PM_2.5 metric, reflecting multiple-day average levels prior to each day in the dataset.

Time-Relevant Covariates

Daily average temperature, relative humidity, day of the year, and federal holiday indicators were included in the analysis to control for potential confounding effects. We obtained Daymet data on daily maximum and minimum temperatures (in °C) and vapor pressure (in Pa) at a 1-km resolution from the Oak Ridge National Laboratory Distributed Active Archive Center for Biogeochemical Dynamics (35). We calculated the daily average temperature by averaging the daily maximum and minimum temperatures and used Magnus’s formula to estimate daily relative humidity (36). These meteorological variables were then aggregated to the zip code level to align with the ED visit records following the same process used for PM_2.5 concentrations.

Statistical Analysis

We applied a case cross-over study design to assess the associations between short-term exposure to smoke PM_2.5 and nonsmoke PM_2.5 and ED visits for cause-specific respiratory disease ED visits during the wildfire seasons of the western United States. A time-stratified approach was used for selecting control days, whereby each ED visit (case) was matched to as many as four nonevent (control) days based on the same day of the week within the same calendar month and year as the case day. This method mitigates bias from temporal trends by ensuring that control days are chosen independently of the case day. Individual-level time-constant characteristics such as sex, age, and socioeconomic status are inherently controlled for in this design.

Conditional logistic regression was used to estimate the relationship between smoke PM_2.5 and nonsmoke PM_2.5 and each outcome. Both exposures were modeled together in a single formula to assess their independent effects. The model was specified as follows:

$$\begin{gathered} \text{logit}\left(P\right)={\mathrm{\beta }}_0+{\mathrm{\beta }}_1{\text{smokePM}}_{2.5} \\ +{\mathrm{\beta }}_2{\text{nnosmokePM}}_{2.5}+{\mathrm{\beta }}_3\text{holidays} \\ +\text{ns}\left(\text{RH}\right)+\text{ns}\left(\text{temp}\right)+\text{ns}\left(\text{DOY}\right) \end{gathered}$$

where $P$ represents the probability of a specific respiratory disease ED visit. Zip code–level averaged relative humidity (RH) and zip code–level averaged temperature (temp), as well as the day of the year (doy) were addressed using natural splines (ns) with 4, 4, and 6 degrees of freedom, respectively. Federal holidays were accounted for using an indicator variable. Our primary temporal exposure metric of interest was the 3-day (72-h) moving average, including the day of the ED visit or the matched control days, for smoke PM_2.5 and nonsmoke PM_2.5. Temperature and relative humidity temporal metrics were matched to those used for PM_2.5 and were therefore also 3-day moving averages in our primary analysis. Results are expressed as odds ratios (ORs) and 95% confidence intervals (CIs).

Stratified Analysis

We conducted stratified analyses to explore whether the associations between smoke PM_2.5 and nonsmoke PM_2.5 exposure with respiratory disease ED visits varied across demographic groups. These groups included age (children aged 5–17 yr, adults aged 18–64 yr, and adults aged ≥65 yr), sex (male and female), ethnicity (Hispanic and non-Hispanic), and race (White, Black, Asian [Asian and Pacific Islander], Native American, and other). Statistical differences in associations between subgroups were evaluated using pairwise two-sample z-tests, with a significance level (α) of 0.05. Additionally, we explored potential variation in associations of wildfire smoke PM_2.5 across states by conducting analyses stratified separately by each of the five states included in our study.

Sensitivity Analysis

We tested the effects of different exposure averaging windows, ranging from 2-day to 6-day moving averages, to evaluate whether associations remained consistent across varying temporal definitions of exposure. To explore potential nonlinear effects or thresholds in the exposure–response relationship, we estimated concentration–response curves by modeling smoke PM_2.5 as smooth terms using restricted cubic splines with three knots in multivariate-adjusted models as above. We used 0 μg/m³ as the reference exposure levels to plot the exposure–response functions.

Results

Study Population

In this study, we included a total of 6,448,525 ED visits for respiratory diseases during the wildfire seasons from 2007 to 2018. Of these, 1,112,758 visits (17.3%) were due to asthma, 368,108 (5.6%) for bronchitis, 502,324 (7.8%) for COPD, and the majority, 3,399,925 (52.7%), were attributed to URIs (Table 1). The demographic distribution of the study population highlights variations across sex, ethnicity, race, and age group. Overall, men and boys accounted for 3,010,428 of the total respiratory ED visits (46.7%), whereas women and girls accounted for 3,355,869 (52.0%).

Table 1.

Characteristics of Emergency Department Visits with a Primary Diagnosis of Respiratory Diseases

Characteristic	Asthma	Bronchitis	COPD	URI	Total Respiratory
Patient sample	1,112,758 (17.3%)	368,108 (5.6%)	502,324 (7.8%)	3,399,925 (52.7%)	6,448,525 (100.0%)
Sex
Male	523,070 (47.0%)	158,924 (43.2%)	221,021 (44.0%)	1,576,906 (46.4%)	3,010,428 (46.7%)
Female	578,531 (52.0%)	208,710 (56.7%)	250,954 (50.0%)	1,817,418 (53.5%)	3,355,869 (52.0%)
Missing	11,157 (1.0%)	474 (0.1%)	30,349 (6.0%)	5,601 (0.2%)	82,228 (1.3%)
Ethnicity
Non-Hispanic	607,719 (54.6%)	221,871 (60.3%)	292,402 (58.2%)	1,712,170 (50.4%)	3,300,126 (51.2%)
Hispanic	441,271 (39.7%)	135,018 (36.7%)	180,355 (35.9%)	1,576,778 (46.4%)	2,823,702 (43.8%)
Missing	63,768 (5.7%)	11,219 (3.0%)	29,567 (5.9%)	110,977 (3.3%)	324,697 (5.0%)
Race
White	564,032 (50.7%)	221,043 (60.0%)	380,005 (75.6%)	1,871,538 (55.0%)	3,709,121 (57.5%)
Black	227,409 (20.4%)	45,469 (12.4%)	61,392 (12.2%)	392,290 (11.5%)	813,343 (12.6%)
American Indian/Alaska Native	9,863 (0.9%)	3,207 (0.9%)	3,227 (0.6%)	37,444 (1.1%)	63,040 (1.0%)
Asian/Pacific Islander	57,903 (5.2%)	17,149 (4.7%)	19,137 (3.8%)	139,204 (4.1%)	303,048 (4.7%)
Other	233,431 (21.0%)	74,246 (20.2%)	33,213 (6.6%)	889,609 (26.2%)	1,438,456 (22.3%)
Missing	20,120 (1.8%)	6,994 (1.9%)	5,350 (1.1%)	69,840 (2.1%)	121,517 (1.9%)
Age group
5–17 yr	426,704 (38.3%)	86,916 (23.6%)	1,055 (0.2%)	1,868,547 (55.0%)	2,720,282 (42.2%)
18–64 yr	595,767 (53.5%)	229,822 (62.4%)	218,887 (43.6%)	1,367,741 (40.2%)	2,800,858 (43.4%)
≥65 yr	89,365 (8.0%)	50,884 (13.8%)	282,375 (56.2%)	131,420 (3.9%)	886,052 (13.7%)
Missing	922 (0.1%)	486 (0.1%)	7 (0.0%)	32,217 (0.9%)	41,333 (0.6%)

Definition of abbreviations: COPD = chronic obstructive pulmonary disease; URI = upper respiratory infection.

Ethnically, Hispanic individuals made 2,823,702 of the visits (43.8%) and non-Hispanic individuals made 3,300,126 (51.2%). When considering race, White individuals made up 57.5% of the cases, followed by Black individuals (12.6%), with other racial groups contributing smaller percentages. Respiratory ED visits were particularly high among individuals aged 5–17 years (42.2%) and those aged 18–64 years (43.4%), with the elderly population (age ≥65 ys) accounting for 13.7% of the visits. For COPD, most ED visits were made by patients aged ≥65 yr (56.2%), whereas ED visits for other conditioned were more common among children and adults. Our study had low levels of missingness (1.3%, 5.0%, 1.9%, and 0.6%) in the age, ethnicity, race, and age group data, respectively. During the study period, the average concentrations of 3-day average smoke PM_2.5 exposure during case days and control days were 0.74 μg/m³ (IQR, 0.81) and 0.76 μg/m³ (IQR, 0.82), respectively. For nonsmoke PM_2.5, the average concentrations were 10.16 μg/m³ (IQR, 5.81) during case days and 10.11 μg/m³ (IQR, 5.77) during control days. The distributions of the exposure are presented as Figure E1.

Effects of Wildfire Smoke and Nonsmoke PM_2.5

During wildfire seasons from 2007 to 2018, after adjusting for concurrent nonsmoker PM_2.5 concentrations, we found that a 1-μg/m³ increase in the 72-hour average wildfire smoke PM_2.5 concentration was associated with higher odds of total respiratory ED visits (OR, 1.004; 95% CI, 1.004–1.004). When disaggregated by specific outcomes, the effect of smoke PM_2.5 with asthma (OR, 1.016; 95% CI, 1.015–1.016) was notably stronger than those with other respiratory conditions. Associations were also observed between smoke PM_2.5 exposure and COPD (OR, 1.004; 95% CI, 1.003–1.005). Among communicable outcomes, associations were generally attenuated, with a slightly stronger effects for bronchitis (OR, 1.002; 95% CI, 0.999–1.004) than for URIs (OR, 1.001; 95% CI, 1.000–1.001). Wildfire smoke PM_2.5 exerted stronger effects on asthma, URIs, and total respiratory ED visits than nonsmoke PM_2.5, of which the ORs are 1.002 (95% CI, 1.001–1.004), 0.999 (95% CI, 0.999–1.001), and 1.001 (95% CI, 1.001–1.002), respectively. In contrast, nonsmoke PM_2.5 contributed similarly or slightly more to increasing the risks for bronchitis (OR, 1.005; 95% CI, 1.002–1.007) and COPD (OR, 1.005; 95% CI, 1.003, 1.008) compared with wildfire smoke PM_2.5 (Figure 1).

Figure 1.

Associations between 3-day average wildfire smoke and nonsmoke fine particulate matter (i.e., aerodynamic diameter ≦ 2.5 μm) exposures and respiratory emergency department visits across western U.S. states during wildfire seasons (May through October) in 2007–2018. Odds ratios per 1-μg/m³ increase in fine particulate matter are presented with corresponding 95% confidence intervals. COPD = chronic obstructive pulmonary disease; URI = upper respiratory infection.

Stratified Analysis

The detailed results of demographic-stratified analyses are presented in Figure 2 and Figure E2 and Table E3. Wildfire smoke PM_2.5 exposure was positively associated with all outcomes across both sexes. Notably, female individuals faced significantly higher risks of asthma, COPD, and overall respiratory ED visits following short-term wildfire smoke exposure. This effects were most pronounced for asthma, with an OR of 1.018 (95% CI, 1.017–1.020) in female individuals compared with 1.013 (95% CI, 1.012–1.014) in male individuals, and the difference was statistically significant (P < 0.001). Conversely, nonsmoke PM_2.5 exposure had significantly higher impacts on total respiratory ED visits in male individuals than in female individuals, with ORs of 1.002 (95% CI, 1.001–1.003) for male individuals and 1.001 (95% CI, 0.999–1.002) for female individuals (P < 0.001).

Figure 2.

Associations between 3-day average wildfire smoke PM_2.5 and respiratory emergency department visits stratified by sex, age group, ethnicity, and race across western U.S. states during wildfire seasons (May through October) in 2007–2018. Odds ratios per 1-μg/m³ increase in smoke PM_2.5 are shown with corresponding 95% confidence intervals. COPD = chronic obstructive pulmonary disease; PM_2.5 = fine particulate matter (i.e., ≤2.5 μm in aerodynamic diameter); URI = upper respiratory infection.

The age-stratified analysis reveals that adults had statistically higher ORs between smoke PM_2.5 and noncommunicable respiratory ED visits. Adults were the only group showing a significant increase in URI visits with exposure to smoke PM_2.5, with an OR of 1.001 (95% CI, 1.000–1.002). In contrast, nonsmoke PM_2.5 exposure did not contribute to a strong age-specific pattern for total respiratory ED visits. However, URI risk increased progressively with age under nonsmoke PM_2.5 exposure, and COPD risk was higher in younger adults than in older adults.

We observed no consistent pattern in the effects of smoke PM_2.5 in the analyses stratified by racial and ethnic group. However, Hispanic individuals exhibited higher ORs of bronchitis and URI visits associated with smoke PM_2.5 exposure compared with non-Hispanic individuals (P = 0.0036 and P = 0.0500, respectively). Black Americans exhibited statistically significantly higher associations for total respiratory visits compared with other racial groups, excluding Native Americans. Conversely, Asian individuals had statistically significantly lower odds of COPD compared with other racial groups, excluding Native Americans. In terms of nonsmoke PM_2.5, exposure did not display significant differences in ORs across racial and ethnic groups.

State-stratified analyses revealed notable variability in the associations between wildfire smoke PM_2.5 exposure and respiratory ED visits. Positive associations were consistently observed across all five states for asthma and total respiratory ED visits, although the associations in Utah were marginally significant. Larger effect sizes were identified in Arizona and Nevada. Conversely, Utah demonstrated comparatively lower ORs. However, the CIs for these state-specific associations substantially overlapped, indicating that observed differences among states were not statistically significant. No protective effects were observed in any state. Detailed state-specific associations are presented in Figure E3.

Sensitivity Analysis

As shown in Figure 3, we observed overall stable associations for wildfire smoke PM_2.5 and respiratory ED visits across different exposure windows. As presented in Figure 4, our analysis revealed an almost linear relationship, with no thresholds, between smoke PM_2.5 and ED visits for total respiratory disease and specific outcomes.

Figure 3.

Associations between wildfire smoke PM_2.5 exposure and respiratory emergency department visits across varying exposure windows (2-d to 6-d averages) in western U.S. states during wildfire seasons (May through October) in 2007–2018. Odds ratios per 1-μg/m³ increase in smoke PM_2.5 are presented with corresponding 95% confidence intervals. COPD = chronic obstructive pulmonary disease; PM_2.5 = fine particulate matter (i.e., ≤2.5 μm in aerodynamic diameter); URI = upper respiratory infection.

Figure 4.

Concentration–response curve illustrating the association between 3-day average wildfire smoke PM_2.5 and total respiratory emergency department visits across western U.S. states during wildfire seasons (May through October) in 2007–2018. The distribution of exposure concentrations is shown beneath the curve, with 0 μg/m³ serving as the reference level. Shaded areas represent 95% confidence intervals. PM_2.5 = fine particulate matter (i.e., ≤2.5 μm in aerodynamic diameter).

Discussion

This study investigated the differential associations between short-term exposure to wildfire smoke PM_2.5 and nonsmoke PM_2.5 with respiratory disease ED visits across the western United States during wildfire seasons from 2007 to 2018. Using a sample size of more than 6 million respiratory ED visits, our findings provide compelling evidence that short-term exposure to wildfire-specific PM_2.5 increases the risk of respiratory ED visits, particularly for asthma. Our study, with its broader spatiotemporal coverage and novel exposure assessment methods, strengthens the evidence of wildfire smoke’s short-term effects on asthma and COPD, corroborating findings from prior case cross-over studies conducted in Alaska, Colorado, Oregon, and Washington (22, 23, 37, 38). Importantly, our research provides novel evidence that wildfire smoke PM_2.5 also impacts communicable respiratory diseases, including URIs and bronchitis. This finding is particularly significant because previous literature has predominantly focused on noncommunicable conditions like asthma and COPD, leaving the impacts on communicable outcomes inconclusive (17). These gaps are likely attributable to limited sample sizes and exposure misclassification in earlier studies. For example, a prior study in Washington State reported positive but statistically insignificant associations between wildfire smoke PM_2.5 and hospitalizations for pneumonia (n = 3,165) and bronchitis (n = 298) (37). Similarly, Stowell and coworkers found ORs of 1.008 (95% CI, 0.995–1.021) for URIs and 1.018 (95% CI, 0.984–1.052) for bronchitis using a case cross-over design with 0–2-day exposure windows and sample sizes of 70,714 and 9,398 cases, respectively (38). In contrast, our study included 3,399,925 URI cases and 368,108 bronchitis cases, leveraging advanced exposure assessments capable of distinguishing wildfire smoke PM_2.5 from total ambient PM_2.5 at a 1-km resolution, compared with the 15-km and 4-km resolutions used in earlier studies.

The increased risks we observed for communicable respiratory diseases may be attributed to the unique immunomodulatory effects of wildfire smoke. Wildfire smoke contains a complex mixture of reactive oxygen species and particulate-bound organic compounds, such as polycyclic aromatic hydrocarbons and aldehydes, which are known to impair mucociliary clearance and compromise the integrity of the epithelial barrier in the respiratory tract (39). These disruptions weaken the body’s primary defenses against pathogens, making individuals more susceptible to bacterial and viral infections. Additionally, components of wildfire smoke can trigger oxidative stress and inflammatory responses, further impairing immune functions (27). Toxicological evidence has associated wildfire smoke with transient immune suppression, characterized by decreased macrophage activity and alterations in cytokine production (31). Specifically, reduced levels of IFN-γ,a cytokine crucial for antiviral defense, has been found following wildfire smoke PM_2.5 exposure (12).

Our study found significantly stronger health effects of wildfire smoke PM_2.5 compared with nonsmoke PM_2.5, particularly for asthma, consistent with findings from previous research. In southern California, Aguilera and coworkers reported a 10% increase (95% CI, 3.5–16.5%) in daily respiratory admission rates per 10-μg/m³ increase in wildfire smoke PM_2.5, compared with a 0.72% increase (95% CI, 0.36–1.1%) for nonfire PM_2.5 (14). Similarly, Stowell and coworkers observed an OR of 1.021 (95% CI, 1.012–1.031) per 1-μg/m³ increase in wildfire smoke PM_2.5 for respiratory outcomes, with no significant effects associated with nonsmoke PM_2.5 (38). The polycyclic aromatic hydrocarbons and reactive oxygen species in smoke PM_2.5 can exacerbate oxidative stress and inflammation in the respiratory system through mechanisms such as the depletion of antioxidant defenses and the activation of proinflammatory pathways (31). Additionally, wildfire smoke particles are often coated with organic compounds that can enhance their toxicity, a feature less commonly observed in nonsmoke PM_2.5 (29). The smaller size of wildfire smoke particles also facilitates deeper penetration into the alveoli, where they can trigger localized inflammation and systemic oxidative stress (40, 41). This may explain the stronger associations observed for asthma, a condition characterized by heightened airway hyperreactivity and sensitivity to oxidative damage.

With a well-represented study population, our analysis robustly captured the differential vulnerabilities among various demographic groups, contributing valuable insights to the ongoing discussion of sex- and age-stratified effects. Women were found to have higher odds of asthma and COPD exacerbations compared with men following wildfire smoke PM_2.5 exposure, consistent with findings from most previous studies (17). This disparity may be attributed to hormonal influences on immune regulation, with estrogen potentially amplifying inflammatory responses in the airways (42, 43). Behavioral factors, such as differences in time spent outdoors, caregiving roles, and occupational exposure patterns, may further exacerbate this risk (44). Additionally, women might have greater baseline susceptibility to oxidative stress and inflammation due to sex-based physiological differences, which could magnify the impact of wildfire smoke PM_2.5 on respiratory outcomes (45). Age-specific results revealed that adults exhibited the highest vulnerability to smoke-related asthma and URIs. This is possibly due to overall more outdoor activities and inevitable outdoor work during wildfire events. In terms of race and ethnicity, we observed higher vulnerability to wildfire smoke among Black and Hispanic populations compared with other racial groups and non-Hispanic individuals under smoke PM_2.5 exposure, respectively. These disparities may reflect underlying differences in baseline respiratory health, healthcare access, and occupational exposures, highlighting the critical need for targeted public health strategies and interventions aimed at reducing the disproportionate impact of wildfire smoke in historically underserved communities (26).

Our state-stratified analyses revealed differences in the associations between smoke PM_2.5 exposure and respiratory ED visits, although wide and overlapping CIs suggest that these differences were not statistically significant. Nevertheless, the observed higher risks in Arizona, California, and Nevada underscore the need for heightened public health attention to wildfire smoke exposure in these states. Furthermore, the wide CIs at the state level reinforce the value of our multiple-state study design, as combining data across multiple states substantially increased the sample size, thereby enhancing the precision of effect estimates.

One novel finding of our study is the concentration–response relationship of wildfire smoke PM_2.5 on respiratory health, a topic that has been rarely explored in previous studies in the United States. Our study provides new and compelling evidence that short-term exposure to wildfire smoke PM_2.5 follows a linear concentration–response relationship, with no discernible threshold, highlighting the continuous health risks even at lower exposure levels and reinforcing the need for vigilance during wildfire events. Concurrently, we found no lag effects for the wildfire smoke PM_2.5 exposure. Unlike other health outcomes, the respiratory system experiences an immediate response to increased PM_2.5 levels due to direct exposure during inhalation. PM_2.5 rapidly induces oxidative stress, inflammation, and immune activation in the airway, disrupting epithelial integrity and exacerbating respiratory conditions (39). Consistent with previous research on the progressive respiratory damage from extended PM_2.5 exposure, our finding indicated that the effect of smoke PM_2.5 persists with longer exposure windows, suggesting possible respiratory damage from long-term smoke PM_2.5 exposure (46, 47).

This study has several strengths. To our knowledge, it is the first large-scale study to use wildfire-specific PM_2.5 estimations to assess the short-term health effects of wildfire smoke on respiratory diseases. Leveraging 6 million ED visit records across a wide geographic area, we were able to detect the impacts of wildfire smoke and nonsmoke PM_2.5 on acute respiratory outcomes. Our advanced exposure assessment method accurately distinguishes between smoke and nonsmoke PM_2.5, providing a robust basis for analysis. Finally, we are the first to estimate the concentration–response function for wildfire smoke PM_2.5 exposure and acute respiratory outcomes, offering valuable insights into the dose–response relationship.

Several limitations must also be acknowledged. First, our findings are limited to acute health impacts and do not account for potential long-term effects of wildfire smoke PM_2.5 exposure. Second, personal exposure estimates could not be precisely determined because our health data resolution was restricted to the zip-code level. Third, respiratory infections, such as bronchitis and upper respiratory conditions, may exhibit subtype-specific responses to pollution, which could not be differentiated using the current data. Finally, although our exposure assessment achieved high predictive accuracy, future research comparing respiratory health effects across different wildfire smoke PM_2.5 exposure models is warranted to determine whether and how exposure modeling choices impact epidemiological findings.

This study highlights the acute respiratory risks posed by wildfire smoke PM_2.5, strengthening existing evidence on its respiratory effects and adding critical findings on its impact on respiratory infections. By leveraging a large, diverse population and advanced exposure modeling, we identified heightened risks among vulnerable groups, including women, adults, and Black and Hispanic communities. Our results reveal the linearity of short-term respiratory responses to smoke PM_2.5, with increasing effects observed over extended exposure durations and nonlatency impacts. As wildfires become more frequent and severe as a result of climate change, this study underscores the urgency of targeted public health interventions and the need for further research into the long-term health impacts, toxicological pathways, and chemical compositions of wildfire smoke PM_2.5.

Supplementary Material

A data supplement for this article is available via the Supplements tab at the top of the online article.

At a Glance Commentary.

Scientific Knowledge on the Subject:

Wildfires are increasing in frequency and intensity worldwide, with the western United States heavily impacted. Fire smoke substantially increases concentrations of fine particulate matter (i.e., particles with an aerodynamic diameter ≤2.5 μm [PM_2.5]) and has been linked to respiratory morbidity. Previous studies often covered localized areas or short periods and could not reliably separate wildfire-related PM_2.5 from other sources at a high spatial resolution across extensive regions or longer time frames, leaving evidence on the differential effects of smoke-derived PM_2.5 versus total PM_2.5 inconclusive. Evidence on the effects of smoke PM_2.5 on communicable respiratory diseases, such as bronchitis and upper respiratory infection, remains sparse.

What This Study Adds to the Field:

Leveraging data from more than 6 million respiratory emergency department visits in the Western United States from 2007 to 2018, this study employs a novel machine-learning-based exposure model at 1-km spatial resolution to distinguish wildfire smoke PM_2.5 from non-smoke PM_2.5. Smoke PM_2.5 is associated with increased risks of combined and specific respiratory diseases, with stronger effects for asthma and upper respiratory infections compared with non-smoke PM_2.5. Susceptibility varies by groups, with women, adults, Black and Hispanic individuals showing greater vulnerability. Additionally, smoke PM_2.5 shows a near-linear association with respiratory outcomes, with no apparent thresholds. These findings support targeted public health interventions that address wildfire smoke’s differential toxicity and motivate further research into its mechanisms.