Skip to main content

Long-term exposure to fine particulate matter and ozone and the onset of systemic autoimmune rheumatic diseases: an open cohort study in Quebec, Canada

Abstract

Objectives

To estimate associations between fine particulate matter (PM2.5) and ozone and the onset of systemic autoimmune rheumatic diseases (SARDs).

Methods

An open cohort of over 6 million adults was constructed from provincial physician billing and hospitalization records between 2000 and 2013. We defined incident SARD cases (SLE, Sjogren’s syndrome, scleroderma, polymyositis, dermatomyositis, polyarteritis nodosa and related conditions, polymyalgia rheumatic, other necrotizing vasculopathies, and undifferentiated connective tissue disease) based on at least two relevant billing diagnostic codes (within 2 years, with at least 1 billing from a rheumatologist), or at least one relevant hospitalization diagnostic code. Estimated PM2.5 and ozone concentrations (derived from remote sensing and/or chemical transport models) were assigned to subjects based on residential postal codes, updated throughout follow-up. Cox proportional hazards models with annual exposure levels were used to calculate hazard ratios (HRs) for SARDs incidence, adjusting for sex, age, urban-versus-rural residence, and socioeconomic status.

Results

The adjusted HR for SARDS related to one interquartile range increase in PM2.5 (3.97 µg/m3) was 1.12 (95% confidence interval 1.08–1.15), but there was no clear association with ozone. Indirectly controlling for smoking did not alter the findings.

Conclusions

We found associations between SARDs incidence and PM2.5, but no relationships with ozone. Additional studies are needed to better understand interplays between the many constituents of air pollution and rheumatic diseases.

Introduction

Ambient air pollution is the fifth leading cause of deaths worldwide [1]. Exposure to particulate matter, especially fine particles of diameter 2.5 µm or less (PM2.5), increases risk of developing or dying from cardiac, cerebrovascular, and chronic airway diseases [2]. Only a few studies have investigated the effects of air pollution on rheumatic diseases [3], and the evidence of an association from these studies is not persuasive [4].

Systemic autoimmune rheumatic diseases (SARDs), including systemic lupus erythematosus (SLE) and related conditions, are complex chronic rheumatic disorders that cause high personal and societal burdens [5]. The etiology of SARDs is poorly understood. Inhaled air pollutants (e.g., PM2.5 and ozone) can increase oxidative stress and inflammatory responses in the lung and potentially trigger systemic immune system changes [6]. However, to date, no studies of air pollution and the incidence of SARDs have been published. Our purpose was to determine whether the incidence of SARDs is associated with exposures to ambient PM2.5 and ozone.

Methods

We used provincial administrative health data from the province of Quebec (population of approximately 8.2 million in 2013) to create this population-based cohort [7]. In Canada, residents of each province are recipients of provincial universal health care plans, covering physician encounters (including all clinic, inpatient, and emergency room encounters) and hospitalizations. Information about these encounters (with International Classification of Diseases, ICD, billing claim diagnoses for physician encounters, and primary and up to 25 non-primary ICD diagnostic codes for hospitalizations) is maintained in provincial administrative databases. From these databases, we formed an open cohort initially of all Quebec residents who were 18 years of age and older as of April 1, 2000. Prior to cohort entry, subjects were required to have lived in the province of Quebec for at least 4 years and to be free of SARDs, based on no physician billings or hospitalizations. After 2000, residents who met these entry criteria were added to the cohort in each subsequent calendar year. Follow-up extended from cohort until the first of date of SARDs diagnosis, death, migration from the province, or end of study (i.e., December 31, 2012). Information on age, sex, six-character residential postal codes throughout the entire follow-up, and physician visits and hospitalizations was available for all subjects.

An incident SARD case was defined on the basis of two or more physician billing claims with a relevant ICD diagnostic code within 2 years, including at least one relevant physician claim from a rheumatologist, or one or more hospitalizations with a relevant ICD diagnostic code. The codes of relevance included ICD-9 446, 710, or 725 and ICD-10 diagnostic codes of M30, M31, M32.1, M32.8, M32.9, M33, M34, M35.0, M35.3, M35.5, or M35.9. These diagnostic codes include SLE, Sjogren’s syndrome, scleroderma, polymyositis, dermatomyositis, polyarteritis nodosa and related conditions, polymyalgia rheumatic, other necrotizing vasculopathies, and undifferentiated connective tissue disease. This approach to rheumatic disease case definition has been adapted for use within the Public Health Agency of Canada’s national chronic disease surveillance system, and our previous work has estimated high specificity and sensitivity for the use of administrative claims data for SARDs diagnoses in Quebec [8].

We used estimates of concentrations of PM2.5 for 2000–2012 derived from complex models including the GEOS (Goddard Earth Observing System) Chemical Transport Model (GEOS-Chem) and remotely sensed measurements of aerosol optical depth. GEOS-Chem (http://geos-chem.org) is a 3-D model simulating atmospheric chemistry by meteorological factors. These relatively coarse PM2.5 image products (at the 10 km resolution) were further resampled to a 1-km (km) spatial resolution using geographically weighted regression modeling. The PM2.5 estimates hold a high agreement (R2 = 0.81) with PM2.5 concentrations from ground monitoring stations [9]. Hourly concentrations of ozone simulated by the Canadian Hemispheric and Regional Ozone and NOx System Chemical Transport Model and measured by ground monitoring stations were combined using an optimal interpolation scheme [10]. The combined ozone concentrations have been further averaged for each calendar year between 2002 and 2012 and exported as a gridded dataset at a spatial resolution of 21 km. Ozone data for 2002 was used for 2000 and 2001. We assigned for the days of follow-up of each subject, annual average daily concentrations of PM2.5, and ozone at their residential address according to their postal code of residence. Thus, a person followed for the 13-year period of observation (i.e., 2000–2012) and residing at the same location for that period would have the same yearly exposure average applied for all days of each year. If a person moved within a year, two different exposure values were assigned, one for the days at the old address and one for the days at the new address.

Using a 10% random sample and the Bayesian approach developed by Nasari et al. [11], we assessed possible linear and nonlinear relationships between SARDs onset and exposures to the two individual air pollutants. This preliminary assessment showed that linear functions described the relationships best (i.e., with the lowest Akaike information criterion). Given the Bayesian approach is rather inefficient in terms of computing time, we used Cox proportional hazards models with time-varying exposure levels instead of the Bayesian approach for all individuals followed in the cohort. Multi-variable Cox proportional hazards models for time to SARD onset were developed separately for each of the two continuous variables of the air pollutants, adjusting for age and sex with time since cohort entry as the time axis. We did not combine PM2.5 and ozone in the same models because air pollution concentrations are closely correlated with each other [12]. Previous data suggest that socioeconomic variables may also influence risk of SARDs, and thus, we adjusted for sex, age, and socioeconomic status (SES) at cohort entry. SES was estimated using the Pampalon material deprivation index (indicating lack of financial resources) [13], from the least (Quintile-1) to the most (Quintile-5) deprived (see Table 1 for the detailed categories). These index values are assigned at the level of dissemination area, which consists of one or more neighboring blocks of houses with a population of 400 to 700. Since regional variations in SES may potentially confound associations between disease onset and air pollution [3], we adjusted for sex, age, socioeconomic status, and urban-versus-rural residence in another group of models. To classify urban-versus-rural residence, we used the Statistics Canada classification system to define major census metropolitan areas (CMA), other CMAs, census agglomerations, and rural areas. Given that women have a higher incidence of SARDs [14] and that air pollution has been shown to influence outcomes differently in men and women [15], we performed a sensitivity analysis stratifying by sex.

Table 1 Distribution of selected characteristics of the entire cohort, Quebec, 2000–2013

Previous studies have suggested that smoking may increase risks of certain SARDs (e.g., SLE) [17], and it is possible that local smoking habits may vary systematically with regional variations in air pollution levels. However, smoking information is not collected in Canada’s administrative health data. Thus, as others have done [18], we performed an additional indirect adjustment of our HRs for smoking. The method requires information on linear associations between the variables included in the survival models and the variable indirectly adjusted for (i.e., smoking). To derive these associations, we used data on the percentage of Quebec residents living in postal code regions who were current/former vs. never smokers, based on the 2005–2015 Canadian Community Health Survey (CCHS) and the published HR for current/former smoking and SARDs onset [17]. Because ozone data has not been linked to CCHS smoking data, we did not conduct this indirect adjustment for the analyses of ozone and SARDs.

Results

After excluding 294,667 individuals with missing data (4.6%), we followed 6,104,859 adults (51.4% female) without SARDs for a total of 63,042,096 person-years (average 10 years), with 32,179 new SARDs cases during this time (5.1 cases per 10,000 person-years). Among the 32,179 new SARDs patients, diffuse diseases of connective tissue and polymyalgia rheumatica were the most frequent (35.1% and 25.5% of cases respectively these codes). The number of cases defined by each ICD diagnostic code is shown in Table S1. Statistics on the entire cohort are shown in Table 1. The average PM2.5 and ozone concentrations across Quebec were 7.5 (standard deviation, SD 2.5) µg/m3 and 24.0 (SD 4.1) parts per billion (ppb) respectively. The interquartile ranges of PM2.5 and ozone were 4.0 µg/m3 and 5.6 ppb respectively. The annual average PM2.5 concentrations for Quebec decreased steadily from 8.9 to 6.6 µg/m3 during 2000 to 2012 while the annual ambient ozone concentration increased from 21.6 to 31.8 ppb. The average exposure of PM2.5 was weakly correlated to that of ozone with a Spearman correlation coefficient of 0.22 (p < 0.01). More detailed distribution of population weighted PM2.5 and ozone exposures are shown in Table S2.

Table 2 shows the adjusted HR and 95% confidence intervals (CIs) for the linear associations between the two air pollutants and SARDs incidence. With the overall subjects in our sample, the adjusted HRs suggested a positive association of SARD onset with PM2.5 exposure, but no clear associations were found for ozone. Similar results were observed in the male and female subgroups. After additional indirect adjustment for smoking, the adjusted HR for the association between SARDs and PM2.5 was not greatly change (see Table 2).

Table 2 Adjusted HRs (95% CIs) from Cox proportional hazards models for the linear association between SARDs and per interquartile increment in PM2.5 or ozone exposure.a, adjusting for different covariates, Quebec, 2000–2013

Discussion

Our findings of a positive association between PM2.5 exposure and SARDs incidence are consistent with our previous work using prevalent SARD diagnoses in Quebec and in Alberta [3, 16]. Ours is the first population-based study of incident SARDs and air pollution. We did not find a clear association between ozone and SARD onset. No one has previously studied ozone and SARD onset, although a positive association between ozone and rheumatoid arthritis onset was observed in one study from British Columbia, Canada [19]. However, we used more precise methods to estimate ozone exposure.

The use of a large population-based sample is a great strength of our study. Also, we evaluated incident disease, in contrast to previous work using prevalent cases [16]. Moreover, in our previous studies [3, 16], each subject was assigned a fixed exposure value. In our study, we used time-varying air pollutants levels for the entire period of follow up of the individuals to better reflect exposures over time and fully consider the residential mobility.

We should note that, in using administrative health data, we are only studying individuals who have presented for (and received) health care. However, SARDs are most often symptomatic (thus patients are likely to present for care), and our previous analyses have suggested that an average of 10 years of follow-up is adequate to detect SLE, one of the more common SARDs [20].

As most existing studies did, we assigned to each participant concentrations of PM2.5 and ozone at the central point of the postal code of his/her residence. However, people are mobile in their cities for different daily activities while locations of those activities cannot be fully recorded [21]. Thus, we admit that some exposure misclassifications result from the current assignment method. Additionally, a postal code in a rural region tends to cover a larger area of land than that in an urban region. However, we do not believe this leads to important misclassification of exposures. Variance in air pollutant concentrations in rural areas is relatively small, even though a rural participant’s house and the central point of the postal code of the participant within a region may be relatively far away.

The average PM2.5 and ozone concentrations across Quebec are lower than the standard air pollutant levels of 8.8 µg/m3 and 62 ppb (for 8-h average) respectively defined by the Canadian Ambient Air Quality Standards (CAAQS) [22]. In fact, health effects below the CAAQS standards have been demonstrated in many previous studies [23, 24].

Not having exposure data prior to entering the cohort is a limitation of this study. We do acknowledge the relatively coarse spatial resolution of the ozone data, which could have masked a true significant association between ozone and SARDs onset. Moreover, the significance of the relationship between ozone exposure and SARDs may be different after adding more covariates such as individual smoking habit to the model. Ambient ozone levels have great seasonal variations and ozone levels may spike to exceptionally high a few days each year. Although we did not observe a clear association between long-term exposure to ozone and SARDs, we did not model effects of short-term very high ozone levels on SARDs. Also, our analyses do no preclude the possibility that jurisdictions with higher ozone concentrations might yield different results.

Conclusions

PM2.5 exposures were associated with higher risks of SARDs onset in Quebec. Indirectly controlling for smoking did not alter the association. We did not observe conclusive associations between SARDs onset and ozone exposure in Quebec. However, this does not preclude the possibility that ozone concentrations might be associated with rheumatic disease in other settings or jurisdictions. Additional studies are needed to better understand interplays between the many constituents of air pollution and rheumatic disease.

Availability of data and materials

The dataset used and analyzed during the current study is available from Institut National de Santé Publique du Québec (INSPQ) on reasonable request.

Abbreviations

CAAQS:

Canadian Ambient Air Quality Standards

CCHS:

Canadian Community Health Survey

CI:

Confidence interval

CMA:

Census metropolitan areas

GEOS-Chem:

Goddard Earth Observing System Chemical Transport Model

ICD:

International Classification of Diseases

HR:

Hazard ratios

PM2.5 :

Fine particulate matter

SARDs:

Systemic autoimmune rheumatic diseases

SES:

Socioeconomic status

SD:

Standard deviation

SLE:

Systemic lupus erythematosus

References

  1. Cohen AJ, Brauer M, Burnett R, Anderson HR, Frostad J, Estep K. Estimates and 25-year trends of the global burden of disease attributable to ambient air pollution: an analysis of data from the Global Burden of Diseases Study 2015. Lancet. 2017;389:1907–18.

    Article  Google Scholar 

  2. Kim KH, Kabir E, Kabir S. A review on the human health impact of airborne particulate matter. Environ Int. 2015;74:136–43.

    CAS  Article  Google Scholar 

  3. Bernatsky S, Smargiassi A, Johnson M, Kaplan GG, Barnabe C, Svenson L, et al. Fine particulate air pollution, nitrogen dioxide, and systemic autoimmune rheumatic disease in Calgary. Alberta Environment Res. 2015;140:474–8.

    CAS  Article  Google Scholar 

  4. Sun G, Hazlewood G, Bernatsky S, Kaplan GG, Eksteen B, Barnabe C. Association between air pollution and the development of rheumatic disease: a systematic review. Int J Rheumatol. 2016;2016:5356307.

    PubMed  PubMed Central  Google Scholar 

  5. Carter EE, Barr SG, Clarke AE. The global burden of SLE: prevalence, health disparities and socioeconomic impact. Nat Rev Rheumatol. 2016;12:605–20.

    Article  Google Scholar 

  6. Bernatsky S, Fournier M, Pineau CA, Clarke AE, Vinet E, Smargiassi A. Associations between ambient fine particulate levels and disease activity in patients with systemic lupus erythematosus (SLE). Environ Health Perspect. 2011;119:45–9.

    CAS  Article  Google Scholar 

  7. Blais C, Jean S, Sirois C, Rochette L, Plante C, Larocque I, et al. Le Système Intégré de Surveillance des Maladies Chroniques du Québec (SISMACQ): une approche novatrice. Chron Dis Injuries Canada. 2014;34:226–35.

    CAS  Article  Google Scholar 

  8. Broten L, Aviña-Zubieta AJ, Lacaille D, Joseph L, Hanly JG, Lix L, et al. Systemic autoimmune rheumatic disease prevalence in Canada: updated analyses across 7 provinces. J Rheumatol. 2014;41:673–9.

    Article  Google Scholar 

  9. van Donkelarr A, Martin RV, Brauer M, Hsu NC, Kahn RA, Levy RC, et al. Global estimation of fine particulate matter using a combined geophysical-statistical method with information from satellites, models, and monitors. Environ Sci Technol. 2016;50:3762–72.

    Article  Google Scholar 

  10. Robichaud A, Menard R. Multi-year objective analyses of warm season ground-level ozone and PM2.5 over North America using real-time observations and Canadian operational air quality models. Atmospheric Chem Phys. 2014;14:1769–800.

    Article  Google Scholar 

  11. Nasari MM, Szyszkowicz M, Chen H, Crouse D, Turner MC, Jerrett M. A class of non-linear exposure-response models suitable for health impact assessment applicable to large cohort studies of ambient air pollution. Air Qual Atmos Health. 2016;9(8):961–72.

    CAS  Article  Google Scholar 

  12. Zhao N, Smargiassi A, Hudson M, Fritzler MJ, Bernatsky S. Long-term exposures to NO2, O3, and PM2.5 using a Bayesian kernel machine regression approach. Environ Int. 2020;136:1054725.

    Google Scholar 

  13. Pampalon R, Hamel D, Gamache P, Raymond G. A deprivation index for health planning in Canada. Chron Dis Canada. 2009;29:178–91.

    CAS  Article  Google Scholar 

  14. Berghen N, Vulsteke J-B, Westhovens R, Lenaerts J, De Langhe E. Rituximab in systemic autoimmune rheumatic diseases: indications and practical use. Acta Clin Belg. 2019;74:272–9.

    Article  Google Scholar 

  15. Kan H, London SJ, Chen G, Zhang Y, Song G, Zhao N, et al. Season, sex, age, and education as modifiers of the effects of outdoor air pollution on daily mortality in Shanghai, China: the public health and air pollution in Asia (PAPA) study. Environ Health Perspect. 2008;116:1183–8.

    CAS  Article  Google Scholar 

  16. Bernatsky S, Smargiassi A, Barnabe C, Svenson LW, Brand A, Martin RV, et al. Fine particulate air pollution and systemic autoimmune rheumatic disease in two Canadian provinces. Environ Res. 2016;146:85–91.

    CAS  Article  Google Scholar 

  17. Costenbader KH, Kim DJ, Peerzada J, Lockman S, Nobles-Knight D, Petri M, et al. Cigarette smoking and the risk of systemic lupus erythematosus: a meta-analysis. Arthritis Rheum. 2004;50:849–57.

    Article  Google Scholar 

  18. Shin HH, Cakmak S, Brion O, Villeneuve P, Turner MC, Goldberg MS, et al. Indirect adjustment for multiple missing variables applicable to environmental epidemiology. Environ Res. 2014;134:482–7.

    CAS  Article  Google Scholar 

  19. De Roos AJ, Koehoorn M, Tamburic L, Davies HW, Brauer M. Proximity to traffic, ambient air pollution, and community noise in relation to incident rheumatoid arthritis. Environ Health Perspect. 2014;122:1075–80.

    Article  Google Scholar 

  20. Ng R, Bernatsky S, Rahme E. Observation period effects on estimation of systemic lupus erythematosus incidence and prevalence in Quebec. J Rheumatol. 2013;40(8):1334–6.

    Article  Google Scholar 

  21. Zhao N, Pinault L, Toyib O, Vanos J, Tjepkema M, Cakmak S. Long-term ozone exposure and mortality from neurological diseases in Canada. Environ Int. 2021;157:106817.

    CAS  Article  Google Scholar 

  22. Government of Canada. Population exposure to outdoor air pollutants. Available from: https://www.canada.ca/en/environment-climate-change/services/environmental-indicators/population-exposure-outdoor-air-pollutants.html#Table1, last Accessed on 12 June 2022.

  23. McClellan RO. Setting ambient air quality standards for particulate matter. Toxicology. 2002;181–182:329–47.

    Article  Google Scholar 

  24. Stafoggia M, Renzi M, Forastiere F, Ljungman P, Davoli M, de Donato F, et al. Short-term effects of particulate matter on cardiovascular morbidity in Italy: a national analysis. Europ J Prev Cardiol. 2022;29:1202–11.

    Article  Google Scholar 

Download references

Acknowledgements

The authors wish to thank Céline Plante for her assistance in computing the indirect adjustment for smoking and the three reviewers for their helpful comments and suggestions.

Funding

This work was funded by the Canadian Institutes of Health Research (CIHR) (PJT-159682). Some analyses were also conducted with financial support provided to the Canadian Network of Research Data Centers (CCRSR) by the Social Sciences and Humanities Research Council (SSHRC), the CIHR, the Canada Institute for Innovation (CFI), and Statistics Canada.

Author information

Authors and Affiliations

Authors

Contributions

NZ: methodology, writing. AS: data curation, methodology, writing. SJ: data curation, writing. PG: statistical analysis. EALSE: statistical analysis. HC: methodology, writing. MSG: methodology, writing. SB: methodology, supervision project administration, writing. The authors read and approved the final manuscript.

Corresponding author

Correspondence to Sasha Bernatsky.

Ethics declarations

Ethics approval and consent to participate

The study was reviewed by the McGill University, Faculty of Medicine, ethics review committee, and was given full approval to conduct (ethics #: A04-M46-12B).

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1:

Table S1. International Classification of Diseases (ICD) diagnostic codes for defining systemic autoimmune rheumatic diseases cases and their corresponding numbers of cases. Table S2. Distribution of population weighted concentrations of ambient PM2.5 and ozone in Quebec. SD: standard deviation.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Zhao, N., Smargiassi, A., Jean, S. et al. Long-term exposure to fine particulate matter and ozone and the onset of systemic autoimmune rheumatic diseases: an open cohort study in Quebec, Canada. Arthritis Res Ther 24, 151 (2022). https://0-doi-org.brum.beds.ac.uk/10.1186/s13075-022-02843-5

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://0-doi-org.brum.beds.ac.uk/10.1186/s13075-022-02843-5

Keywords

  • Systemic autoimmune rheumatic diseases
  • Air pollution
  • Fine particulate matter
  • Ozone