Wildfires have become a major global concern, and in the United States (US) there are hundreds of thousands to millions of acres burned [1
]. Consequently, wildland smoke emissions are progressively being recognized as a public health concern, due to large scale wildfire fire events [3
]. The increased number of these events are attributed to anthropogenic climate change, including warmer temperatures, early spring melt, and decreased winter precipitation [4
]. Lightning and human ignition of excess forest fuels from years of previous fire suppression activity, as well as forest management practices, have contributed to large scale wildland fire events [5
]. It has been projected that there will be an ∼50% increase in burned areas across the western US between 2009 and 2050 and future predictive models show that this area will continue to see rapidly growing fire activity with increases of 80% burned areas in the Pacific Northwest alone [6
]. While the western states (Washington, Oregon, Montana, Idaho, California, Wyoming, Nevada, Arizona) shoulder a majority of fires/acres burned (7 million+ in 2017), the Midwest and South had hundreds of thousands of acres of wildfires in 2017. Because of fire location and prevailing wind patterns, western Montana communities in the Northern Rockies are annually inundated with smoke from increasing seasonal wildfires [8
], and the region presents an opportunity to study the health effects of wildfire smoke exposures in historically at-risk communities.
Given recent climate trends, a growing incidence of ‘historic’ fires is appearing to represent the new normal in the western US. While most fire seasons result in significant levels of exposures, the 2017 Seeley Lake region in Montana experienced an unprecedented level of smoke exposure from nearby wildfires in terms of sustained PM concentrations. Due to numerous factors, including close proximity to multiple fires and the presence of significant inversions (atmospheric conditions that trap pollution closer to the ground), community residents were exposed to EPA-designated “very unhealthy” and “hazardous” PM2.5
levels for 35 of the 49 days of exposure (1 August–19 September) in the summer of 2017 with a median 24 h average of 220.9 mg/m3
for the entire period. Historically, studies attempting to assess the health effects of wildfire smoke in local communities have focused on medical records including emergency department visits, hospital admissions, or provider visits categorized with specific ICD codes for respiratory or cardiovascular diagnoses [11
]; however, the long-term human health implications of these exposures have not previously been assessed. Wildfire smoke exposure is ascribed to an average 339,000 deaths each year, and studies have reported that wildland smoke PM is associated with respiratory effects [16
]. The present study addresses this gap in knowledge with two-year follow up of evaluating community members exposed to these significant levels of wildfire smoke in Seeley Lake, MT.
During the 2017 wildfires, the Missoula City-County Health Department, Division of Environmental Quality (MCCHD-DEQ) contacted the University of Montana about the exposed community. A multi-disciplinary team headed by the IPHARM (ImProving Health Among Rural Montanans) program in the School of Pharmacy at the University of Montana was assembled to enroll and screen community members in Seeley Lake, MT and for comparison, the similarly sized town of Thompson Falls, MT whose smoke exposure during the same time period was five-fold less PM2.5. Participants were given multiple surveys in addition to screening of health parameters: blood pressure, pulse-oximetry, and spirometry. The present study focused on respiratory effects from exposure to wildfire smoke. Analysis of full spirometry testing on community members from Seeley Lake showed significant decreases in lung function parameters up to two years post exposure, with clinically significant decreases in FEV1 and changes in the FEV1/FVC ratio indicating obstruction. The present work represents one of the first of its kind to assess and follow an exposed cohort to determine potential long-term health effects of wildfire smoke.
2.1. Study Design
The study was designed to enroll and assess multiple health parameters of persons living in Seeley Lake, MT area following exposure to unprecedented levels of wildfire smoke during the summer of 2017. Using the IPHARM health screening program infrastructure, five screening visits were conducted: one initial screening in 2017 within 24 h following the last day of elevated smoke, two in 2018 and two in 2019. Screening tests included spirometry, blood pressure, heart rate, oximetry, subject survey data, and collecting blood and saliva samples for later epigenetic testing. An additional cohort was enrolled and similarly screened in July of 2018 in Thompson Falls, MT, USA.
2.2. Study Population
The study population consisted of male and female subjects living in Seeley Lake and Thompson Falls, MT, USA during the summer of 2017. Subjects were between 23 and 85 years of age. Exclusion criteria excluded persons under the age of 18, inability to answer survey questions, or inability to perform spirometry based on the following screening questions: 1. In the last 3 months have you had a chest injury or surgery involving the eye, ear, chest, abdomen, or been hospitalized for a heart attack? 2. Do you experience hemoptysis? 3. Have you had a respiratory infection, such as flu, pneumonia, bronchitis or chest cold, in the last 3 weeks? 4. Have you ever had a pneumothorax? 5. Do you experience regular chest pain? 6. Have you ever had thoracic, abdominal, or cerebral aneurysms? An affirmative answer to any of these questions precluded a participant from the spirometry testing. Additionally, participants undergoing current treatment for hypertension were evaluated for control (i.e., <130/80) before undergoing spirometry testing. The study protocol was approved by the Institutional Review Board (IRB) at the University of Montana and participants provided informed consent. Initial study approval was obtained by the University of Montana-Missoula Institutional Review Board on 14 September 2017 (#185-17), with annual continuation approval on 23 August 2018 and 26 August 2019.
2.3. Particle Exposures
The MCCHD-DEQ has monitoring stations in Seeley Lake and Thompson Falls, MT. Daily PM2.5 levels were chronicled using the EPA NowCast method which registered PM2.5 concentrations for every 12 h and then calculated a weighted average of those hours. In Seeley Lake, the air quality monitor is located outside and in close proximity to the elementary school athletic field in the town and is an average of 1.755 miles from each participant’s listed address. In Thompson Falls, the monitor is located near the parking lot of the high school and is an average of 4.74 miles from the participants’ listed addresses.
2.4. Study Procedures
2.4.1. Surveys and Participant Screening
Clinical history was obtained through surveys with special reference to smoking habits; asthma; allergies; systemic cardiovascular and respiratory diseases; ownership of wood burning stoves; COPD/Emphysema; bronchitis; and past and current health history in two weeks prior to the screening event. General physical examination included height, weight, and arterial and brachial blood pressure measurements (PulseWave analysis system; SphygmoCor AtCor Medical).
2.4.2. Pulmonary Function Tests
Pulmonary function tests were performed according to NIOSH approved guidelines using a NIOSH approved spirometer, the ndd EasyOne spirometer (ndd Medical Technologies Inc., Andover, MA, USA) or the Vitalograph asma-1 monitor (Vitalograph, Inc., Lenexa, KS, USA) with testing conducted in a seated position. After screening spirometry was performed with the acceptability of each test determined by the NIOSH-certified tester and spirometry software. The EasyOne™ spirometer has an inbuilt test quality grading system (A–D, F) that provides feedback to the operator on test acceptability and repeatability. A goal was set for three acceptable tests of A or B session quality with a limit of no more than eight attempts. This corresponded to three acceptable tests with between-test repeatability of 150 mL or less, as per ATS/ERS criteria. The forced vital capacity (FVC), forced expiratory volume in one second (FEV1
), and FEV1
/FVC ratio was calculated automatically by the spirometer as the percentage of predicted values based on age, height, and gender as defined by NHANES III prediction equation (third National Health and Nutrition Examination Survey, 1999) [16
]. In addition, the lower limit of normal (LLN) was calculated using z-scores ((measured-predicted)/standard deviation), where z-score = −1.64 (5th percentile) is defined as the LLN [17
]. Spirometry quality and lung function values were compared between testing events to assess the reliability and changes in results over time.
2.5. Data Analysis
Intergroup differences were evaluated using three-way ANOVA followed by Sydak’s Test for comparing multiple group means while controlling type I error. Odds ratios were generated using 2 × 2 contingency tables with one factor being the presence or absence of a clinically significant decrease in FEV1 over a year (>30 mL) and the other factor being the presence or absence of a health attribute (e.g., asthma, allergies, etc.). The dependent variable was simply the frequency of occurrence. Fisher’s Exact test was used to test for independence between factors with a probability of Type I error set at 5 percent (two-tailed). The 95% confidence intervals for the odds ratios were calculated using the Baptista-Pike method. Significance for odds ratios was determined by the confidence interval. If the 95% CI included ‘1′ within the interval it was not statistically significant. Statistical significance of the frequency of clinically significant decreases in FEV1 was determined by binomial probabilities, contrasting observed frequency of clinically deceased FEV1 with the expected frequency for any given 1-year period.
Wildfires are a growing and significant concern globally. Most studies assessing potential health effects of exposures to the resulting smoke have focused on historical data of emergency department visits, hospital admissions, or provider visits [11
]. The previous studies reported an increase in visits with ICD codes including cardiovascular and respiratory complications in the time frame following a wildfire event. In contrast, the present study was designed to evaluate and longitudinally follow a cohort of individuals in a community impacted with significant levels of smoke from wildfires. The cohort in Seeley Lake, MT is an older population (average age: 63 years) with a fairly even distribution of sexes (Table 1
), while the comparison community of Thompson Falls had an average age of 59 years with the majority of patients being female. Participants were screened for inclusion, given health and demographic surveys, and underwent spirometry testing to assess potential effects on respiratory function parameters from the exposures.
Residents living in Seeley Lake, MT during the summer of 2017 were exposed to extremely high levels of PM2.5
is a major component of air pollution and one of the criterion air pollutants designated by the EPA and has established the PM2.5
cutoff to be 35.4 µg/m3
for “unhealthy” designations. Seeley Lake had 35 consecutive days with PM2.5
levels at 150.5 µg/m3
(a designation of “very unhealthy”) and above and 9 days where levels were greater than 250.4 µg/m3
(“hazardous”). There were four fires burning within a 50-mile radius of Seeley Lake in 2017, contributing to the smoke exposure of the residents. The valley location of this community allowed for the smoke from the nearby wildfires to be trapped on the valley floor with temperature inversions, a weather phenomenon that occurs when cold air at night traps air pollution in a valley and prevents it from blowing away or rising higher in the atmosphere. In other recent studies examining health impacts of smoke from wildfire events, levels of PM2.5
had not reached the levels of those in the Seeley Lake exposure. In Australia, during a particularly significant wildfire period in 2006/2007, the daily average for the two-month timeframe was 15.81 µg/m3
(max. 294.95 µg/m3
]; while during the 2007 San Diego fires there was a five-day average of 89.1 µg/m3
(max. 803.1 µg/m3
]. Our study’s comparison community of Thompson Falls is also located in the Northern Rockies region, in the Clark Fork river valley. For the same time period in 2017 this community was also exposed to EPA designation of “unhealthy” levels of wildfire smoke (daily average of 47 µg/m3
), however, it was 5-fold less average PM2.5
than Seeley Lake, MT (Figure 1
). These exposures are unprecedented and have afforded researchers the opportunity to follow a cohort longitudinally.
The most likely health effects from wildfire smoke exposures are on the respiratory and cardiovascular systems. To this end studies have generally taken the form of historical evaluations of medical records and ICD codes for respiratory and cardiovascular outcomes [11
]. In the 2007 San Diego wildfires, increased respiratory medical encounters were found to correlate with peak smoke periods [13
]. Likewise, Alman, et al. noted an increase in hospitalizations and ED visits for cardiorespiratory codes during the 2012 Colorado wildfires [11
]. These studies are able to illustrate immediate effects of these exposures, but without individual longitudinal data we cannot appreciate, or identify, long-term complications of wildfire smoke exposures. The main physiological parameter assessed in the present study was lung function via spirometry testing. Spirometry is a common method of assessing pulmonary function that can be used to diagnose asthma, chronic obstructive pulmonary disease (COPD) and other respiratory pathologies. Spirometry is often used to evaluate lung physiology instead of X-rays and CAT scans because it can detect abnormalities in lung function even when no signs or symptoms of a disease are evident. When assessing lung function and the potential effects of environmental exposures, spirometry generates multiple parameters for comparison [19
]. The main values utilized from spirometry assessments include, but are not limited to, FEV1
(forced expiratory volume in the first second), FVC (forced vital capacity), and PEF (peak expiratory flow). The volume (FEV1
) is compared to the FVC volume, which is the total amount air exhaled during testing, and the FEV1
/FVC ratio is considered a reliable indicator of lung function. In analyses of these types of data, the ratios are age-matched and individuals are evaluated based on their lower limit of normal (LLN) [22
]. In addition, FEV1
values are used to determine whether lung function is declining at a normal rate based on age. According to the Mayo Clinic, the expected annual decline in pulmonary function in FEV1
is 30 mL for males and 25 mL for women.
Both short-term and long-term studies of populations exposed to pollution have found significant correlation between fine particle pollutants and respiratory morbidity and mortality [25
]. Exposures to PM2.5
have also been consistently associated with decreases in pulmonary function in epidemiological studies [26
]. Participants in our Seeley and Thompson Falls cohorts were assessed by an OSHA-certified staff scientist. The data from these studies suggest a significant decrease in lung function one year following the exposure and the decrease was maintained up to two years post smoke exposure (Seeley Lake). In Seeley Lake, a decrease in FEV1
/FVC (Figure 2
) was observed in both 2018 and 2019 with a significant difference in males. In addition, there was an increase in the number of individuals that dropped below the LLN for this parameter (Figure 3
). This type of change suggests an obstruction (as opposed to a restriction) that is in the category of asthma or COPD (chronic obstructive pulmonary disease). Participants were also evaluated for clinically significant annual changes in FEV1
(>25 mL in females and >30 mL in males) where at least half of the participants showed a clinically significant decrease in FEV1
with the largest average drop in the first year after exposure (Table 2
). The decreased FEV1
values in the Seeley Lake cohort at the initial screening suggests this parameter is more sensitive (temporally) to the smoke exposure (Figure 5
A), and presents in a more vulnerable population (>65 years; Figure 5
B). This decrease in FEV1
means that the lung is restricted from filling to its normal capacity. These lung function changes, while being statistically significant are, more importantly, clinically significant as depicted in the annual decrease of FEV1
volumes and the increase in the number of participants dropping below the LLN for the FEV1
/FVC ratio. In fact, the combination of these results are considered key in diagnosing obstructive changes [27
]. In addition, while previous studies found that increases in 9–10 µg/m3
increased the risk of asthma emergency department visits following exposure [13
], the present study suggests a long-term implication for asthmatics. Because all seven asthmatics in the Seeley Lake cohort presented with decreased FEV1
two years after the wildfire event, and 6 of 7 were clinically significant decreases, it suggests asthma as a risk factor for long-term complications and not just in the context of a short-term trigger.
Exposures to significant levels of wildfire smoke may result in obstructive lung pathology [28
]. Past studies in air pollution have focused on the effects of pollutants on the airway epithelial lining and subsequent activation of the innate immune system. Recent studies have shown that ozone, a major toxic air pollutant, induces IL-33 production by airway epithelial cells that results in activation of type 2 innate lymphoid cells (ILC2) [30
]. The ILC2 are important sources of IL-13 [32
] and have been linked to asthma, an obstructive lung pathology. Additionally, studies in both firefighters and in vitro studies reported increased IL-6, a contributor to inflammatory lung pathology [33
], in response to wildfire or wildfire smoke extract, respectively [35
]. The firefighters were assessed in the acute phase and found increased serum levels of IL-6, IL-8 and decreased IL-10, while the lung epithelial cultures presented with increased IL-6 production in addition to other markers of COPD including dysfunction of tight junctions. Therefore, the present study showing an increase in obstructive pathology based on spirometry results suggests a model of increased acute inflammation and activation of the innate immune system from wildfire smoke that results in tissue remodeling and decreased lung function (Figure 6