You are here

  • Home
  • Archive
  • Volume 78, Issue 10
  • Short-term air pollution exposure and exacerbation events in mild to moderate COPD: a case-crossover study within the CanCOLD cohort

Article Text

Article menu
Chronic obstructive pulmonary disease
Original research
Short-term air pollution exposure and exacerbation events in mild to moderate COPD: a case-crossover study within the CanCOLD cohort
  1. Bryan A Ross1,2,
  2. Dany Doiron1,
  3. Andrea Benedetti1,
  4. Shawn D Aaron3,
  5. Kenneth Chapman4,
  6. Paul Hernandez5,
  7. François Maltais6,
  8. Darcy Marciniuk7,
  9. Denis E O'Donnell8,
  10. Don D Sin9,
  11. Brandie L Walker10,
  12. Wan Tan9,
  13. Jean Bourbeau1,2
  14. for the CanCOLD Collaborative Research Group and the Canadian Respiratory Research Network
  1. 1 Respiratory Epidemiology and Clinical Research Unit, Centre for Outcomes Research and Evaluation, Research Institute of the McGill University Health Centre, Montreal, Québec, Canada
  2. 2 Medicine, McGill University Health Centre, Montreal, Québec, Canada
  3. 3 The Ottawa Hospital Research Institute, University of Ottawa, Ottawa, Ontario, Canada
  4. 4 Toronto General Hospital Research Institute, University of Toronto, Toronto, Ontario, Canada
  5. 5 Medicine, Dalhousie University Faculty of Medicine, Halifax, Nova Scotia, Canada
  6. 6 Institut universitaire de cardiologie et de pneumologie de Québec, Université Laval, Québec, Québec, Canada
  7. 7 Respiratory Research Centre, University of Saskatchewan, Saskatoon, Saskatchewan, Canada
  8. 8 Medicine, Queens University, Kingston, Ontario, Canada
  9. 9 Centre for Heart Lung Innovation, St. Paul's Hospital, University of British Columbia, Vancouver, British Columbia, Canada
  10. 10 Medicine, University of Calgary, Calgary, Alberta, Canada
  1. Correspondence to Dr Jean Bourbeau, McGill University, Montreal, Canada; jean.bourbeau{at}mcgill.ca

Abstract

Background Infections are considered as leading causes of acute exacerbations of chronic obstructive pulmonary disease (COPD). Non-infectious risk factors such as short-term air pollution exposure may play a clinically important role. We sought to estimate the relationship between short-term air pollutant exposure and exacerbations in Canadian adults living with mild to moderate COPD.

Methods In this case-crossover study, exacerbations (‘symptom based’: ≥48 hours of dyspnoea/sputum volume/purulence; ‘event based’: ‘symptom based’ plus requiring antibiotics/corticosteroids or healthcare use) were collected prospectively from 449 participants with spirometry-confirmed COPD within the Canadian Cohort Obstructive Lung Disease. Daily nitrogen dioxide (NO2), fine particulate matter (PM2.5), ground-level ozone (O3), composite of NO2 and O3 (Ox), mean temperature and relative humidity estimates were obtained from national databases. Time-stratified sampling of hazard and control periods on day ‘0’ (day-of-event) and Lags (‘−1’ to ‘−6’) were compared by fitting generalised estimating equation models. All data were dichotomised into ‘warm’ (May–October) and ‘cool’ (November–April) seasons. ORs and 95% CIs were estimated per IQR increase in pollutant concentrations.

Results Increased warm season ambient concentration of NO2 was associated with symptom-based exacerbations on Lag−3 (1.14 (1.01 to 1.29), per IQR), and increased cool season ambient PM2.5 was associated with symptom-based exacerbations on Lag−1 (1.11 (1.03 to 1.20), per IQR). There was a negative association between warm season ambient O3 and symptom-based events on Lag−3 (0.73 (0.52 to 1.00), per IQR).

Conclusions Short-term ambient NO2 and PM2.5 exposure were associated with increased odds of exacerbations in Canadians with mild to moderate COPD, further heightening the awareness of non-infectious triggers of COPD exacerbations.

  • COPD exacerbations
  • COPD exacerbations mechanisms
  • COPD epidemiology
  • Clinical epidemiology

Data availability statement

Data are available upon reasonable request.

https://doi.org/10.1136/thorax-2022-219619

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • While chronic obstructive pulmonary disease (COPD) exacerbations are often attributed to infection, there is growing literature supporting the role of short-term air pollution exposure. Prior studies investigating this association have relied on health administrative data to classify COPD disease status and to detect exacerbations and tended to include patients with predominantly severe disease despite the fact that mild to moderate COPD accounts for half of all patients diagnosed, one in four of which remain undiagnosed.

    WHAT THIS STUDY ADDS

    • In this multisite case-crossover study with a population sampling of participants with mild to moderate COPD diagnosed by spirometry, the inclusion of both ‘symptom-based’ as well as ‘event-based’ exacerbations provided a unique, sensitive and clinically relevant outcome. Positive associations were demonstrated between short-term nitrogen dioxide and fine particulate matter exposure (in warm and cool seasons) and ‘symptom-based’ exacerbations.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • These results show that seasonal short-term air pollutant exposure is a risk for COPD exacerbations among milder disease individuals and females in settings with relatively low annual average air pollutant concentrations, which may inform targeted public health interventions and clinical management strategies.

    Introduction

    Acute exacerbations of chronic obstructive pulmonary disease (AECOPD) contribute to the burden of COPD by accelerating lung function decline, impairing quality of life and increasing healthcare utilisation.1 More frequent and severe AECOPDs are associated with increased mortality2 and they are estimated to consume 60% of the costs of COPD care.3

    Although AECOPDs are widely understood to be triggered mainly by viral/bacterial infection,4 less well-recognised factors such as short-term increases in ambient air pollution may play a role in precipitating these events.5 Ongoing uncertainty regarding AECOPD aetiology is well reflected by national and international reports, which characterise AECOPDs by how they are treated rather than by cause or mechanism.6 7 Pollutants such as NO2, fine particulate matter (PM2.5) and ground-level ozone (O3) have irritant effects,8 which can acutely provoke cough, sputum production and bronchial hyperreactivity.9

    Beyond limited mechanistic studies,10 most prior reports investigating the relationship between short-term air pollution and AECOPDs have been epidemiological studies employing ecological and time series designs, which estimate associations at the population level and tend to rely on hospital administrative database (HAD) and International Classification of Disease (ICD) data to classify COPD disease status and detect AECOPD events. Previous studies have under-represented milder disease and lower air pollutant concentration settings. Patients with COPD may be particularly susceptible to relative acute increases in air pollutant concentrations, even in countries with relatively low air pollution concentrations.11 12 In North America, supralinear relationships (steepest at lower concentrations) were recently described between air pollutants and the incidence of COPD13 as well as for cardiovascular mortality in COPD.14 Thus, even low-level air pollutant concentrations may have a notable impact on AECOPD incidence.14 15

    The primary objective of this study was to estimate the relationship between short-term exposures to NO2, PM2.5 and O3 and exacerbations during the warm and cool seasons in a population-based sample of Canadians with mild to moderate COPD confirmed by spirometry. The secondary objectives were to investigate the effect of gender and disease severity on these associations. We hypothesised that short-term increases in each pollutant would be positively associated with symptom-based and event-based exacerbations during both warm and cool seasons, and that gender and disease severity would be relevant modifiers.

    Methods

    Study design

    This study conducted between 1 July 2012 and 31 December 2019 made use of data collected by the Canadian Cohort Obstructive Lung Disease (CanCOLD). CanCOLD is a prospective longitudinal cohort study in which a random population sampling of non-institutionalised adults aged ≥40 years, both with and without COPD, and split evenly between men and women, were invited to participate from each of nine study sites (cities with a total population of at least 250 000 people each—see online supplemental file) using random telephone digit dialling. A case-crossover design16 17 was used, whereby cases served as their own controls and inference was based on exposure distribution (rather than risk of disease).

    Supplemental material

    Study population

    Inclusion/exclusion criteria of the CanCOLD cohort have been previously published.18 Subjects in this study were selected if they had COPD diagnosed by spirometry (performed at three in-site visits; see online supplemental file) and if they also experienced at least one exacerbation subsequent to COPD diagnosis.

    Health outcome of interest

    Exacerbations19 were subcategorised into ‘symptom-based’ or ‘event-based’ events: increased dyspnoea or sputum volume or purulence lasting ≥48 hours (‘symptom-based’); or meeting ‘symptom-based’ criteria plus antibiotic or corticosteroid use or unscheduled doctor/emergency room visit or hospitalisation (‘event-based’). Symptom-based events do not necessitate healthcare contact, thereby providing a sensitive and clinically relevant outcome20 21 compared with traditionally defined AECOPDs.6 Since July 2012, exacerbation events were prospectively collected through telephone interviews using a standardised questionnaire19 every 3 months (see online supplemental file). Participants were instructed to record every event’s date of onset (first date of symptoms) as they occurred. Only those who reported at least one exacerbation within 4 months of a telephone visit were included/used in analyses.

    Air pollution exposure and covariates

    Hourly measurements of NO2, PM2.5 and O3 from National Air Pollution Surveillance fixed-location monitoring stations within the nine CanCOLD cities were aggregated into 24-hour averages, providing daily city-level air pollutant concentrations for each city. The composite (sum) of NO2, O3 and Ox was calculated and also aggregated into 24-hour averages. Temperature and relative humidity were retrieved from open-access databases (Applied Climatology/Meteorological Services of Canada) to provide city-level daily averages.

    Statistical analysis

    Daily average concentrations of PM2.5, NO2, O3 and Ox on the day of exacerbations (‘Lag0’ hazard period: for both symptom-based and event-based exacerbations, defined as the first date of symptom onset) were compared with daily average concentrations sampled from 3 to 4 separate reference (‘Lag0’ control period) days. Time-stratified referent selection of control periods was used in order to reduce overlap bias and in order to control for confounding by time trend22 23 since control periods on the same day-of-week, month and year as the hazard period control for day-of-week, subseasonal, seasonal and yearly trends.24 Due to a well-described lag effect of air pollutants in precipitating exacerbations, and following the recommendations of the largest meta-analysis performed to date,25–27 daily data for hazard/control periods from 1 to 6 days prior to exacerbations were collected. Given Canadian climatic seasonal extremes, whereby relatively hotter temperatures (in warmer months) and relatively colder temperatures (in cooler months) may precipitate exacerbations, all data were dichotomised into ‘warm’ (May–October) and ‘cool’ (November–April) seasons and analysed separately27 28 to establish temperature directionality. Results are provided separately for symptom-based and event-based exacerbations.

    Models were fitted for each Lag (Lag0 to Lag−6) using generalised estimating equations (GEE) accounting for within-participant event ‘clusters’ as well as for ‘clusters’ of study site (city) over the 7.5-year follow-up period.29 Cumulative effect models (across Lag0−6) were also fitted. ORs and their 95% CI per air pollutant IQR increase were estimated using PROC GENMOD in SAS V.9.4 (SAS Institute, Cary, North Carolina, USA), based on binary distribution, with logit link and repeated statement, accounting for within-subject correlations of repeated measures. The ‘Exchangeable’ correlation structure was used, along with ‘robust’ SE estimation, which is robust to structure misspecification.30

    In all adjusted analyses, models were adjusted for the time-varying exposure variables of mean temperature and relative humidity24 as well as for study site (city). In the primary analysis, single-pollutant models31 (NO2, PM2.5, O3 and Ox) were fitted. In secondary analyses, adjusted two-pollutant models were estimated. Only models with between-pollutant correlations of <0.5 were included to avoid multicollinearity. Subgroup differences by (1) self-reported gender and (2) COPD severity by spirometry6 were analysed separately. Effect modification was assessed using the Wald test with robust SEs. In sensitivity analyses, adjusted primary analyses were repeated using all exacerbations, regardless of recall period.

    Results

    A total of 449 participants were included in this study. Of the 1561 recruited CanCOLD participants (figure 1), 389 participants had a baseline diagnosis of COPD and also reported at least one exacerbation between 1 July 2012 and 31 December 2019 with a ≤4-month recall period. Sixty additional participants without COPD at baseline demonstrated fixed airway obstruction on follow-up spirometry and also subsequently experienced at least one eligible (≤4 month recall) exacerbation during the follow-up period. A total of 1400 symptom-based and 841 event-based exacerbations were included for analysis. In total, there were 2156.75 person-years of follow-up for the 449 participants, yielding event rates of 0.65 events/person-year (symptom-based exacerbations) and 0.39 events/person-year (event-based exacerbations). A complete-case analysis was used (see online supplemental file).

    Figure 1
    Figure 1

    Flow diagram of study participants. CanCOLD, Canadian Cohort Obstructive Lung Disease; NHANES, National Health and Nutrition Examination Survey.

    Demographic characteristics of participants recruited at baseline and all those recruited by end-of-study are presented in table 1. There is a roughly equal male-female representation among participants, with slightly more than half having mild (GOLD1) COPD.6 Demographic characteristics are very similar between the 389 baseline COPD participants and the 60 participants included subsequently following confirmation of COPD by spirometry.

    View this table:
    Table 1

    Characteristics of study participants recruited at baseline and by the end of the study

    Descriptive statistics for average daily air pollutant concentrations over the 2012–2019 study period (dichotomised by warm/cool seasons) are presented in table 2 and descriptive air pollutant concentration statistics subdivided by study site (city) are presented in online supplemental table S1. Overall, median NO2 concentrations were higher in the cool season, PM2.5 concentrations were similar between seasons, O3 concentrations were slightly higher in the cool season and Ox concentrations were higher in the cool season.

    View this table:
    Table 2

    Air pollutant concentrations during case and control periods in the warm and cool seasons

    Single-pollutant models

    Figures 2–5 present adjusted GEE models for each of Lags0−6 as well as cumulative models for NO2, PM2.5, O3 and Ox, respectively.

    Figure 2
    Figure 2

    Single-pollutant adjusted generalised estimating equations (GEE) models for NO2 and the odds of symptom-based and event-based exacerbations (A. warm, B. cool). Models adjusted for mean temperature and relative humidity as well as city. ORs and 95% CIs are presented per IQR increase in NO2.

    Figure 3
    Figure 3

    Single-pollutant adjusted generalised estimating equations (GEE) models for PM2.5 and the odds of symptom-based and event-based exacerbations (A. warm, B. cool). Models adjusted for mean temperature and relative humidity as well as city. ORs and 95% CIs are presented per IQR increase in PM2.5.

    Figure 4
    Figure 4

    Single-pollutant adjusted generalised estimating equations (GEE) models for O3 and the odds of symptom-based and event-based exacerbations (A. warm, B. cool). Models adjusted for mean temperature and relative humidity as well as city. ORs and 95% CIs are presented per IQR increase in O3.

    Figure 5
    Figure 5

    Single-pollutant adjusted generalised estimating equations (GEE) models for Ox and the odds of symptom-based and event-based exacerbations (A. warm, B. cool). Models adjusted for mean temperature and relative humidity as well as city. ORs and 95% CIs are presented per IQR increase in Ox.

    In the warm season adjusted model, an IQR NO2 increment was associated with an increased odds of symptom-based events (1.14 [1.01,1.29]) on Lag-3 (figure 2). In the cool season, no significant association was observed.

    In the warm season, no significant association was observed between PM2.5 exposure and the odds of exacerbations (figure 3). In the cool season, an IQR PM2.5 increment was associated with a statistically significant increased odds of symptom-based exacerbations (1.11 (1.03 to 1.20)) on Lag−1.

    In the warm season, an IQR O3 increment on Lag-3 was associated with a decreased odds of event-based exacerbations (0.73 (0.52 to 1.00)) (figure 4). In the cool season, no significant association was observed.

    No significant association was observed between Ox exposure and the odds of exacerbations in either the warm or cool seasons (figure 5).

    Similar results were observed between adjusted (figures 2–5) and unadjusted (online supplemental figures S2–S5) GEE models for Lags0−6 for NO2, PM2.5, O3 and Ox.

    Two-pollutant models

    Figure 6 and figure 7 present adjusted two-pollutant GEE models for NO2-PM2.5 and PM2.5-O3, respectively. Strong NO2-O3 correlation (R2=−0.61) precluded an NO2-O3 two-pollutant model (online supplemental table S2). Overall, the same associations and of a similar magnitude were observed in the NO2-PM2.5 and PM2.5-O3 two-pollutant models as those observed in the single-pollutant models (detailed summaries provided in online supplemental file). Notable differences observed include a decreased odds of warm season symptom-based (Lag-2) exacerbations, as well as an increased odds of cool season event-based (Lag-3) exacerbations with each PM2.5 IQR increment (NO2-PM2.5 model—figure 6); a decreased odds of cool season event-based exacerbations on Lag−2 with each PM2.5 IQR increment (PM2.5-O3 model—figure 7) and a decreased odds of cool season symptom-based (Lag−2) exacerbations with each IQR O3 increment (PM2.5-O3 model−figure 7).

    Figure 6
    Figure 6

    Two-pollutant generalised estimating equations (GEE) models for PM2.5 and NO2 and the odds of symptom-based and event-based exacerbations. All models adjusted for mean temperature, mean relative humidity, city, and both pollutants (A. and B. NO2 adjusted for PM2.5; C. and D. PM2.5 adjusted for NO2; A. and C. warm; B. and D. cool). ORs and 95% CIs are presented per IQR increase in NO2, and per IQR increase in PM2.5, respectively.

    Figure 7
    Figure 7

    Two-pollutant generalised estimating equations (GEE) models for PM2.5 and O3 and the odds of symptom-based and event-based exacerbations. All models adjusted for mean temperature, mean relative humidity, city, and both pollutants (A. and B. PM2.5 adjusted for O3; C. and D. O3 adjusted for PM2.5; A. and C. warm; B. and D. cool). ORs and 95% CIs are presented per IQR increase in PM2.5, and per IQR increase in O3, respectively.

    Subgroups, effect modifiers and sensitivity analyses

    Evidence of effect modification was observed for gender and disease severity, with stronger associations between air pollutant exposure and exacerbations observed among women and participants with GOLD1 COPD. Patterns of associations observed in the overall population between exposure to each air pollutant (in particular, NO2 and PM2.5) and exacerbations were observed in women but not in men (online supplemental tables S3 and S4) and were observed in participants with GOLD1 but not GOLD2+COPD (online supplemental tables S5 and S6).

    Sensitivity analyses using all exacerbations regardless of recall period (which included 1646 symptom-based and 993 event-based exacerbations) revealed similar patterns in associations to the patterns observed in primary analyses (online supplemental table S7) with an additional decreased odds of negative cool season event-based (Lag−2) exacerbations with each IQR PM2.5 increment.

    Discussion

    This study shows that short-term exposure to ambient NO2 and PM2.5 was associated with increased odds of exacerbations in a Canadian multisite cohort of participants with mild to moderate COPD. The specific pattern of positive associations for both NO2 (warm, Lag−3) and PM2.5 (cool, Lag−1) was observed consistently across primary, secondary and sensitivity analyses. Furthermore, this pattern was observed specifically in women but not in men, and in mild (GOLD1) but not in predominantly moderate (GOLD2) COPD, and gender and disease severity were both observed to be effect modifiers on the association between air pollutant exposure and exacerbations. The pattern of negative association between ambient O3 exposure and exacerbations (warm, Lag−3) was observed inconsistently (on different Lags) across primary and sensitivity analyses, and no consistent associations were observed across analyses between Ox and exacerbations. Furthermore, some negative associations were observed between PM2.5 exposure and exacerbations in some sensitivity analyses and secondary analyses but were not observed in the primary (adjusted or unadjusted) analyses. These associations were estimated in spirometry-confirmed mild to moderate COPD, using a sensitive study outcome, across many cities in a country with low-level air pollution concentrations, which cumulatively reflect the novelty of this study.

    The air pollutant concentrations reported in this study are notably lower in comparison to prior descriptions in this literature26; however, the maximal hazard period concentrations observed did approach or exceed Canadian Ambient Air Quality standards.11 Thus, even chronic lung disease patient populations living in countries with relatively ‘cleaner’ air12 may be susceptible to short-term/episodic air pollutant concentration spikes, which may be acutely detrimental to their respiratory status, of which the cumulative effects are unknown. Likewise, the observed event rates in this study are notably lower in comparison to previous observational COPD cohorts, and this is likely attributable to the milder spectrum of disease of participants in this study. When equivalent subgroups (eg, ‘moderate’ COPD subgroups across different cohorts) are compared with one another, comparable event rates are observed.19 32 The present study results would suggest that even patients living with milder forms of disease may be susceptible to these air pollutant concentration spikes.

    Very few similar case-crossover studies have been performed in the North American setting,27 33 or in a well-characterised cohort of participants with COPD confirmed by spirometry.29 33 The main results from past studies are also conflicting. For example, short-term increases in PM2.5 concentration were found to be associated with an increased odds,27 with no change,28 and even with a reduced odds 33 of exacerbations among past case-crossover studies. This PM2.5 discordance was also observed in the present study (primary vs secondary/sensitivity analyses findings), for reasons which remain to be explained.

    In the present case-crossover study, NO2 (Lag−3 warm) and PM2.5 (Lag−1 cool) exposure was associated with increased odds of exacerbations in a seasonal manner. Warm season-specific NO2 effects have been reported previously,28 and the absence of cool season associations may relate to less time spent outdoors in the Canadian wintertime. Conversely, the observed cool season-specific PM2.5 associations may relate to known seasonal and urban properties of PM2.5 composition and concentration. PM2.5 composition is highly heterogeneous. Some reports have demonstrated seasonal and geographic patterns of increased caustic compounds in the makeup of PM2.5 during winter months and in urban settings.34 Since the present study was conducted across nine urban centres, possible distinct urban and wintertime PM2.5 characteristics may offset any influences on time spent outdoors and might support a possible mechanism for these cool season-specific PM2.5 associations.

    An inverse correlation was observed between daily ambient O3 and NO2 concentrations, and exposure to ambient O3 was associated with a decreased odds of exacerbations. Strong inverse O3-NO2 correlations have been reported previously,35 and likely stem from known typical geographical distributions of NO2 (higher in city core) and O3 (higher in nearby suburbs).11 The association observed between O3 and exacerbations was unexpected. In the largest published systematic review and meta-analysis,26 Lag0 exposure to NO2 was associated with a significantly increased odds of AECOPD while O3 was associated with a significantly reduced odds of AECOPD.26 This same paradoxical association was observed between short-term increases in O3 concentration and all ‘COPD visit’ types in prior Canadian time series study results.25 Associations between O3 concentrations and other acute events, such as myocardial infarction, have also uniquely differed from all other main pollutants.36 These seemingly paradoxical ‘protective’ O3 associations, reported here and previously,26 27 29 may be explained in part by O3 as a ‘secondary bystander’36 as one of many components in a complex mixture of other harmful pollutants.

    Gender and disease severity were effect modifiers on the association between air pollutant exposure and exacerbations. Positive short-term air pollution exposure associations with exacerbations were observed in women and not men. Short-term27 and long-term37 gender-specific effects of air pollution have been reported previously. Gender differences in hormonally-mediated immune responses, T-cell expression patterns and pro-inflammatory cytokine release observed in COPD may help explain these findings.38 Furthermore, the principal study findings were observed only in GOLD1 COPD but not in GOLD2+ participants. Individual-level behaviours including exposure avoidance (staying indoors on ‘high-pollutant’ days) may occur disproportionately in subgroups with more severe disease. The GOLD2+ subgroup consisted mainly of moderate (‘GOLD2’) disease, with under-representation/non-representation of severe (‘GOLD3’: n=31) and very severe (‘GOLD4’: n=1) disease, which limits the ability to further investigate this pattern across the spectrum of disease. Overall, this gender-specific and disease severity-specific pattern is interesting since women appear to be at a higher risk of developing early-onset COPD than men.39

    This study has several strengths. The case-crossover study design inherently controls for all individual-level, non-time-varying confounders.24 Furthermore, the majority of prior studies have relied on ICD codes and HADs to classify both COPD disease status and outcome (ie, AECOPD hospitalisation). Within our study, COPD status was confirmed by spirometry at three time points and exacerbation outcomes were recorded prospectively over a 7.5-year period. This approach helped minimise disease misclassification and outcome misclassification. This multicity study with participants from nine cities across six provinces is a unique and representative sampling of the Canadian mild to moderate COPD population (many of them being undiagnosed) and of typical Canadians’ air pollution exposure. Finally, the inclusion of ‘symptom-based’ exacerbations, a unique, sensitive and clinically relevant outcome,20 21 in combination with a population sampling of mild to moderate COPD, allowed for a novel approach to estimate these associations beyond what is possible when only more severe AECOPD case definitions and/or administrative health data-derived outcomes are used.

    This study has several limitations. First, the case-crossover method does not control for time-varying confounders such as daily medication use/adherence and daily activity/time spent outdoors. Since it is most likely that COPD medication use and outdoor avoidance behaviours would both increase around hazard periods, both of these uncontrolled confounders would be expected to lead to effect underestimation. The study was able to control for confounding by all non-time-varying demographic confounders including comorbid conditions (though effect modification by comorbidity was not performed); by time trends through time-stratified control period sampling and for two time-varying confounders, mean temperature and relative humidity. Second, even though participants were instructed to prospectively record event dates, those events recalled further from the time of telephone questionnaires are a potential source of error. To minimise recall bias, an exclusion criterion was set to only include events recalled within 4 months; however, this may have itself created a selection bias. Reassuringly, sensitivity analyses in which all events were included led to similar overall patterns. Third, it was assumed that participants were in their study site city during exacerbations, and a single level of exposure was assigned to all participants within each city. This area-level estimation approach is in keeping with methodology used in prior studies25 27 28 33 40 and may have contributed to the small effect sizes observed; the potential for exposure measurement error is non-differential since bias between hazard and control period day sampling is equal, which would lead to an estimate closer to the null and to wider CIs.41 Fourth, this study did not include rural areas but instead only included large urban cities as study sites. In addition to this being a limitation of this study, the under-representation of rural areas represents a persisting gap in this literature. Fifth, due to the well-established lag effect of air pollutants in precipitating exacerbations and in keeping with recommendations from prior work,26 a large number of associations were tested (Lags0−6). It is important to consider this in the interpretation of the study results, although they were consistent within subgroups and with previous reports. Finally, because this was a case-crossover study from an observational cohort, the sample size is much smaller than that of time series-based studies previously performed.26 Likewise, both more advanced forms of disease as well as the number of event-based exacerbations observed were under-represented.

    These results demonstrating short-term air pollutant exposure as a risk for an increase in COPD exacerbations are relevant to patients, clinicians and health systems alike. These findings are transferrable to mild to moderate COPD patient populations (accounting for half of all patients diagnosed), and often undiagnosed COPD, living in other high-income countries. Many prior studies were limited by the use of administrative health data, which studied this association primarily in clinically diagnosed COPD with severe disease and may have missed clinically relevant milder events. Many other studies were limited to only one city or state/province rather than reporting on a national sampling. This study provides additional support for public health measures such as air pollution controls, and for communicating these risks to target patient populations. Future work on the interactions between air pollution and climate factors such as extremes in temperature is necessary. Further studies in well-characterised subclinical and clinical COPD cohorts are also needed to assess non-hospitalised and hospitalised AECOPDs.

    Data availability statement

    Data are available upon reasonable request.

    Ethics statements

    Patient consent for publication

    Ethics approval

    This study involves human participants and was approved by McGill University Health Centre (MUHC) Research Ethics Board (REB): 2010-1897 (09-025-BMB-t). Participants gave informed consent to participate in the study before taking part.

    Acknowledgments

    We thank all CanCOLD participants, the CanCOLD collaborative research group, the open-access Government of Canada databases and the grant foundations of this research for making this work possible. Executive Committee: Jean Bourbeau (McGill University, Montreal, QC, Canada); Wan C Tan, J Mark FitzGerald, Don D Sin (UBC, Vancouver, BC, Canada); Darcy D Marciniuk (University of Saskatoon, Saskatoon, SK, Canada); Denis E O'Donnell (Queen's University, Kingston, ON, Canada); Paul Hernandez (Dalhousie University, Halifax, NS, Canada); Kenneth R Chapman (University of Toronto, Toronto, ON, Canada); Brandie Walker (University of Calgary, Calgary, AB, Canada); Shawn Aaron (University of Ottawa, Ottawa, ON, Canada); François Maltais (University of Laval, Quebec City, QC, Canada). International Advisory Board: Jonathon Samet (the Keck School of Medicine of USC, California, USA); Milo Puhan (John Hopkins School of Public Health, Baltimore, USA); Qutayba Hamid (McGill University, Montreal, QC, Canada); James C Hogg (University of British Columbia, James Hogg Research Centre, Vancouver, BC, Canada). Operations Centre: Jean Bourbeau (PI), Dany Doiron, Palmina Mancino, Pei Zhi Li, Dennis Jensen, Carolyn Baglole (University of McGill, Montreal, QC, Canada), Yvan Fortier (Laboratoire telematique Respiratory Health Network, FRQS); Wan C Tan (co-PI), Don Sin, Julia Yang, Jeremy Road, Joe Comeau, Adrian Png, Kyle Johnson, Harvey Coxson, Miranda Kirby, Jonathon Leipsic, Cameron Hague (University of British Columbia, James Hogg Research Centre, Vancouver, BC, Canada). Economic Core: Mohsen Sadatsafavi (University of British Columbia, Vancouver, BC). Public Health Core: Teresa To, Andrea Gershon (University of Toronto). Data management and Quality Control: Wan C Tan, Harvey Coxson (UBC, Vancouver, BC, Canada); Jean Bourbeau, Pei-Zhi Li, Zhi Song, Andrea Benedetti, Dennis Jensen (McGill University, Montreal, QC, Canada); Yvan Fortier (Laboratoire telematique Respiratory Health Network, FRQS). Field Centres: Wan C Tan (PI), Christine Lo, Sarah Cheng, Elena Un, Cynthia Fung, Wen Tiang Wang, Liyun Zheng, Faize Faroon, Olga Radivojevic, Sally Chung, Carl Zou (UBC James Hogg Research Centre, Vancouver, BC, Canada); Jean Bourbeau (PI), Palmina Mancino, Jacinthe Baril, Laura Labonté (McGill University, Montreal, QC, Canada); Kenneth Chapman (PI), Patricia McClean, Nadeen Audisho (University of Toronto, Toronto, ON, Canada); Brandie Walker, (PI), Curtis Dumonceaux, Lisette Machado (University of Calgary, Calgary, AB, Canada); Paul Hernandez (PI), Scott Fulton, Kristen Osterling, Denise Wigerius (Dalhousie University and Queen Elizabeth II Health Sciences Centre, Halifax, NS, Canada); Shawn Aaron (PI), Kathy Vandemheen, Gay Pratt, Amanda Bergeron (University of Ottawa, Ottawa, ON, Canada); Denis O'Donnell (PI), Matthew McNeil, Kate Whelan (Queen's University, Kingston, ON, Canada); François Maltais (PI), Cynthia Brouillard (University of Laval, Quebec City, QC, Canada); Darcy Marciniuk (PI), Ron Clemens, Janet Baran, Candice Leuschen (University of Saskatchewan, Saskatoon, SK, Canada).

    References

    1. ↵
      2. Seemungal TA ,
      3. Donaldson GC ,
      4. Paul EA , et al
      . Effect of exacerbation on quality of life in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1998;157:1418–22. doi:10.1164/ajrccm.157.5.9709032
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      2. Soler-Cataluña JJ ,
      3. Martínez-García MA ,
      4. Román Sánchez P , et al
      . Severe acute exacerbations and mortality in patients with chronic obstructive pulmonary disease. Thorax 2005;60:925–31. doi:10.1136/thx.2005.040527
      OpenUrlAbstract/FREE Full Text
    3. ↵
      2. Chapman KR ,
      3. Mannino DM ,
      4. Soriano JB , et al
      . Epidemiology and costs of chronic obstructive pulmonary disease. Eur Respir J 2006;27:188–207. doi:10.1183/09031936.06.00024505
      OpenUrlFREE Full Text
    4. ↵
      2. Ritchie AI ,
      3. Wedzicha JA
      . Definition, causes, pathogenesis, and consequences of chronic obstructive pulmonary disease exacerbations. Clin Chest Med 2020;41:421–38. doi:10.1016/j.ccm.2020.06.007
      OpenUrl
    5. ↵
      2. Celli BR ,
      3. Fabbri LM ,
      4. Aaron SD , et al
      . An updated definition and severity classification of chronic obstructive pulmonary disease exacerbations: the Rome proposal. Am J Respir Crit Care Med 2021;204:1251–8. doi:10.1164/rccm.202108-1819PP
      OpenUrl
    6. ↵
      Global strategy for the diagnosis, management and prevention of chronic obstructive pulmonary disease (2022 report). 2022.
    7. ↵
      2. Bourbeau J ,
      3. Bhutani M ,
      4. Hernandez P , et al
      . Canadian thoracic society clinical practice guideline on pharmacotherapy in patients with COPD – 2019 update of evidence. Can J Resp Crit Care Sleep 2019;3:210–32. doi:10.1080/24745332.2019.1668652
      OpenUrl
    8. ↵
      2. Ritz B ,
      3. Hoffmann B ,
      4. Peters A
      . The effects of fine dust, ozone, and nitrogen dioxide on health. Dtsch Arztebl Int 2019;51–52:881–6. doi:10.3238/arztebl.2019.0881
      OpenUrl
    9. ↵
      2. Schraufnagel DE ,
      3. Balmes JR ,
      4. Cowl CT , et al
      . Air pollution and noncommunicable diseases: a review by the forum of international respiratory societies’ environmental committee, part 2: air pollution and organ systems. Chest 2019;155:417–26. doi:10.1016/j.chest.2018.10.041
      OpenUrlCrossRefPubMed
    10. ↵
      2. Ryu MH ,
      3. Afshar T ,
      4. Li H , et al
      . Impact of exposure to diesel exhaust on inflammation markers and proteases in former smokers with chronic obstructive pulmonary disease: a randomized, double-blinded, crossover study. Am J Respir Crit Care Med 2022;205:1046–52. doi:10.1164/rccm.202104-1079OC
      OpenUrl
    11. ↵
      Canadian environmental sustainability indicators: population exposure to outdoor air pollutants environment and climate change Canada. 2021. Available: https://www.canada.ca/en/environment-climate-change/services/environmental-indicators/population-exposure-outdoor-air-pollutants.html
    12. ↵
      1. Health Effects Institute
      . State of global air 2020. Special report. Boston, MA: Health Effects Institute, 2020.
    13. ↵
      2. Shin S ,
      3. Bai L ,
      4. Burnett RT , et al
      . Air pollution as a risk factor for incident chronic obstructive pulmonary disease and asthma. A 15-year population-based cohort study. Am J Respir Crit Care Med 2021;203:1138–48. doi:10.1164/rccm.201909-1744OC
      OpenUrl
    14. ↵
      2. Alexeeff SE ,
      3. Deosaransingh K ,
      4. Liao NS , et al
      . Particulate matter and cardiovascular risk in adults with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2021;204:159–67. doi:10.1164/rccm.202007-2901OC
      OpenUrl
    15. ↵
      1. World Health Organization
      . WHO global air quality guidelines: particulate matter (PM2.5 and PM10), ozone, nitrogen dioxide, sulfur dioxide and carbon monoxide. Geneva, 2021.
    16. ↵
      2. Jaakkola JJK
      . Case-crossover design in air pollution epidemiology. European Respiratory Journal 2003;21:81S–85s. doi:10.1183/09031936.03.00402703
      OpenUrlAbstract/FREE Full Text
    17. ↵
      2. Maclure M
      . The case-crossover design: a method for studying transient effects on the risk of acute events. Am J Epidemiol 1991;133:144–53. doi:10.1093/oxfordjournals.aje.a115853
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      2. Bourbeau J ,
      3. Tan WC ,
      4. Benedetti A , et al
      . Canadian Cohort Obstructive Lung Disease (canCOLD): fulfilling the need for longitudinal observational studies in COPD. COPD 2014;11:125–32. doi:10.3109/15412555.2012.665520
      OpenUrlCrossRefPubMed
    19. ↵
      2. Labonté LE ,
      3. Tan WC ,
      4. Li PZ , et al
      . Undiagnosed chronic obstructive pulmonary disease contributes to the burden of health care use. Data from the cancold study. Am J Respir Crit Care Med 2016;194:285–98. doi:10.1164/rccm.201509-1795OC
      OpenUrlCrossRefPubMed
    20. ↵
      2. Langsetmo L ,
      3. Platt RW ,
      4. Ernst P , et al
      . Underreporting exacerbation of chronic obstructive pulmonary disease in a longitudinal cohort. Am J Respir Crit Care Med 2008;177:396–401. doi:10.1164/rccm.200708-1290OC
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      2. Bourbeau J ,
      3. Ford G ,
      4. Zackon H , et al
      . Impact on patients’ health status following early identification of a COPD exacerbation. Eur Respir J 2007;30:907–13. doi:10.1183/09031936.00166606
      OpenUrlAbstract/FREE Full Text
    22. ↵
      2. Janes H ,
      3. Sheppard L ,
      4. Lumley T
      . Case-crossover analyses of air pollution exposure data: referent selection strategies and their implications for bias. Epidemiology 2005;16:717–26. doi:10.1097/01.ede.0000181315.18836.9d
      OpenUrlCrossRefPubMedWeb of Science
    23. ↵
      2. Janes H ,
      3. Sheppard L ,
      4. Lumley T
      . Overlap bias in the case-crossover design, with application to air pollution exposures. Stat Med 2005;24:285–300. doi:10.1002/sim.1889
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      2. Mittleman MA ,
      3. Mostofsky E
      . Exchangeability in the case-crossover design. Int J Epidemiol 2014;43:1645–55. doi:10.1093/ije/dyu081
      OpenUrlCrossRefPubMed
    25. ↵
      2. To T ,
      3. Feldman L ,
      4. Simatovic J , et al
      . Health risk of air pollution on people living with major chronic diseases: a Canadian population-based study. BMJ Open 2015;5:e009075. doi:10.1136/bmjopen-2015-009075
    26. ↵
      2. Li J ,
      3. Sun S ,
      4. Tang R , et al
      . Major air pollutants and risk of COPD exacerbations: a systematic review and meta-analysis. Int J Chron Obstruct Pulmon Dis 2016;11:3079–91. doi:10.2147/COPD.S122282
      OpenUrlPubMed
    27. ↵
      2. Szyszkowicz M ,
      3. Kousha T ,
      4. Castner J , et al
      . Air pollution and emergency department visits for respiratory diseases: a multi-city case crossover study. Environ Res 2018;163:263–9. doi:10.1016/j.envres.2018.01.043
      OpenUrlCrossRefPubMed
    28. ↵
      2. Lin M-T ,
      3. Kor C-T ,
      4. Chang C-C , et al
      . Association of meteorological factors and air no2 and o3 concentrations with acute exacerbation of elderly chronic obstructive pulmonary disease. Sci Rep 2018;8:10192. doi:10.1038/s41598-018-28532-5
    29. ↵
      2. Peacock JL ,
      3. Anderson HR ,
      4. Bremner SA , et al
      . Outdoor air pollution and respiratory health in patients with COPD. Thorax 2011;66:591–6. doi:10.1136/thx.2010.155358
      OpenUrlAbstract/FREE Full Text
    30. ↵
      2. Ghisletta P ,
      3. Spini D
      . An introduction to generalized estimating equations and an application to assess selectivity effects in a longitudinal study on very old individuals. Journal of Educational and Behavioral Statistics 2004;29:421–37. doi:10.3102/10769986029004421
      OpenUrlCrossRefWeb of Science
    31. ↵
      2. Stafoggia M ,
      3. Breitner S ,
      4. Hampel R , et al
      . Statistical approaches to address multi-pollutant mixtures and multiple exposures: the state of the science. Curr Environ Health Rep 2017;4:481–90. doi:10.1007/s40572-017-0162-z
      OpenUrl
    32. ↵
      2. Hurst JR ,
      3. Vestbo J ,
      4. Anzueto A , et al
      . Susceptibility to exacerbation in chronic obstructive pulmonary disease. N Engl J Med 2010;363:1128–38. doi:10.1056/NEJMoa0909883
      OpenUrlCrossRefPubMedWeb of Science
    33. ↵
      2. DeVries R ,
      3. Kriebel D ,
      4. Sama S
      . Low level air pollution and exacerbation of existing COPD: a case crossover analysis. Environ Health 2016;15:98. doi:10.1186/s12940-016-0179-z
    34. ↵
      2. Hand JL ,
      3. Schichtel BA ,
      4. Pitchford M , et al
      . Seasonal composition of remote and urban fine particulate matter in the United States. J Geophys Res 2012;117:n doi:10.1029/2011JD017122
    35. ↵
      2. Stieb DM ,
      3. Szyszkowicz M ,
      4. Rowe BH , et al
      . Air pollution and emergency department visits for cardiac and respiratory conditions: a multi-city time-series analysis. Environ Health 2009;8:25. doi:10.1186/1476-069X-8-25
      OpenUrlCrossRefPubMed
    36. ↵
      2. Mustafic H ,
      3. Jabre P ,
      4. Caussin C , et al
      . Main air pollutants and myocardial infarction: a systematic review and meta-analysis. JAMA 2012;307:713–21. doi:10.1001/jama.2012.126
      OpenUrlCrossRefPubMedWeb of Science
    37. ↵
      2. Doiron D ,
      3. Bourbeau J ,
      4. de Hoogh K , et al
      . Ambient air pollution exposure and chronic bronchitis in the lifelines cohort. Thorax 2021;76:772–9. doi:10.1136/thoraxjnl-2020-216142
      OpenUrlAbstract/FREE Full Text
    38. ↵
      2. Kamil F ,
      3. Pinzon I ,
      4. Foreman MG
      . Sex and race factors in early-onset COPD. Curr Opin Pulm Med 2013;19:140–4. doi:10.1097/MCP.0b013e32835d903b
      OpenUrlCrossRefPubMed
    39. ↵
      2. Foreman MG ,
      3. Zhang L ,
      4. Murphy J , et al
      . Early-Onset chronic obstructive pulmonary disease is associated with female sex, maternal factors, and African American race in the COPDGene study. Am J Respir Crit Care Med 2011;184:414–20. doi:10.1164/rccm.201011-1928OC
      OpenUrlCrossRefPubMedWeb of Science
    40. ↵
      2. Gao N ,
      3. Xu W ,
      4. Ji J , et al
      . Lung function and systemic inflammation associated with short-term air pollution exposure in chronic obstructive pulmonary disease patients in Beijing, China. Environ Health 2020;19:12. doi:10.1186/s12940-020-0568-1
    41. ↵
      2. Hutcheon JA ,
      3. Chiolero A ,
      4. Hanley JA
      . Random measurement error and regression dilution bias. BMJ 2010;340:c2289. doi:10.1136/bmj.c2289

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Twitter @kchapmanairways

    • Collaborators CanCOLD Collaborative Research Group and Canadian Respiratory Research Network.

    • Contributors BAR, DD and JB developed the protocol, analysed the data and drafted the original version of the manuscript. BAR, DD, AB, SDA, KC, PH, FM, DM, DEO'D, DDS, BLW, WT and JB made substantial contributions to the design of the study, interpretation of the data, revision of the manuscript for important intellectual content and approved the final version submitted for publication. All authors have given agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the study are appropriately investigated and resolved. JB is the guarantor for the study. JB accepts full responsibility for the finished work and/or the conduct of the study, had access to the data, and controlled the decision to publish.

    • Funding The Canadian Cohort Obstructive Lung Disease (CanCOLD; NCT00920348) study is currently funded by the Canadian Respiratory Research Network and the industry partners AstraZeneca Canada Ltd, Boehringer Ingelheim Canada Ltd, GlaxoSmithKline (GSK) Canada Ltd, and Novartis. Researchers at RI-McGill University Health Centre Montreal and iCAPTURE Centre Vancouver lead the project. Previous funding partners were the Canadian Institutes of Health Research (CIHR; CIHR/Rx&D Collaborative Research Program Operating Grants- 93326), the Respiratory Health Network of the Fonds de la recherche en santé du Québec (FRQS), and industry partners: Almirall; Merck Nycomed; Pfizer Canada Ltd; and Theratechnologies. This work was supported with funding from the Canadian Institutes of Health Research (CIHR: #453225) and the Réseau en Santé Respiratoire du FRQS (RSRQ).

    • Competing interests BAR reports grants/contracts from the Canadian Institutes of Health Research (CIHR), Réseau de Recherche en Santé Respiratoire du Québec (RSRQ), Research Institute of the MUHC (RI MUHC), Ministère de l'Économie et de l'Innovation (MEI) Québec, McGill University Health Centre (MUHC) Foundation Grant, Fonds de Recherche Santé Québec (FRSQ), and CHEST Foundation Grant; and payments/honoraria from the Canadian Thoracic Society (CTS), CHEST/ACCP, Respiplus (non-profit), Alberta Kinesiology Association (AKA), and McGill University Continuing Professional Development (CPD). SDA reports payments/honoraria from AstraZeneca, GSK; and participation on Data Safety Monitoring/Advisory Board for AstraZeneca, GSK, and Sanofi. PH reports grants/contracts from Boehringer Ingelheim, Cyclomedica, Grifols, Vertex; consulting fees from Acceleron, AstraZeneca, Boehringer Ingelheim, Covis, GlaxoSmithKline, Janssen, Novartis, Sanofi, Teva, Takeda, Valeo; and leadership/fiduciary role in the Canadian Thoracic Society. FM reports grants/contracts from GlaxoSmithKline, AstraZeneca, Sanofi, Novartis, Boehringer Ingelheim, Grifols; consulting fees from AstraZeneca; payment/honoraria from GlaxoSmithKline, Boehringer Ingelheim, Grifols, Novartis; and stock/stock options from Oxynov. DM reports grants/contracts from AstraZeneca, Boehringer Ingelheim, Canadian Institute of Health Research, GlaxoSmithKline, Grifols, Lung Association—Saskatchewan, Novartis, Sanofi, Saskatchewan Health Research Foundation, Schering-Plough; consulting fees from Alberta Health Services, Canadian Foundation for Healthcare Improvement, Health Canada, Lung Association—Saskatchewan, Ontario Ministry of Health and Long-Term Care, Saskatchewan Health Authority, Yukon Health and Social Services; payment/honoraria from the Lung Association—Saskatchewan, American College of Chest Physicians; leadership/fiduciary role in the CHEST journal, Canadian Thoracic Society, American Thoracic Society and AARC; and is an employee of the University of Saskatchewan. DE O’D reports grants/contracts from AstraZeneca, Lung Health Foundation and Boehringer Ingelheim Canada; and payment/honoraria from GSK and Viajes Pacifico. BLW reports payment/honoraria from AstraZeneca, GSK, Sanofi; and Data Safety Monitoring/Advisory Board participation for AstraZeneca, GSK, and Sanofi. JB reports grants/contracts from the Canadian Institute of Health Research (CIHR), Réseau en santé respiratoire du FRQS, McGill University, McGill University Health Centre Foundation, AstraZeneca Canada Ltd, Boehringer Ingelheim Canada Ltd, GSK, Grifols, Novartis, Sanofil, Trudell Canada Ltd; and payment/honoraria from AstraZeneca Canada Ltd, Boehringer Ingelheim Canada Ltd, GSK, Pfizer Canada Ltd, and Trudell Canada Ltd.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.