Autoimmunity

RMD Open: first published as 10.1136/rmdopen-2021-002055 on 15 March 2022. Downloaded from http://rmdopen.bmj.com/ on November 20, 2024 by guest. Protected by copyright.
ORIGINAL RESEARCH

Association between long-­term

exposure to air pollution and immune-­
mediated diseases: a population-­based
cohort study
Giovanni Adami ‍ ‍,1 Marco Pontalti,1 Giorgio Cattani,2 Maurizio Rossini,1
Ombretta Viapiana,1 Giovanni Orsolini ‍ ‍,1 Camilla Benini,1 Eugenia Bertoldo,1
Elena Fracassi,1 Davide Gatti,1 Angelo Fassio1

To cite: Adami G, Pontalti M, ABSTRACT

Cattani G, et al. Association Objective Environmental air pollution has been associated
Key messages
between long-­term with disruption of the immune system at a molecular level.
exposure to air pollution
The primary aim of the present study was to describe the
What is already known about this subject?
and immune-­mediated ⇒ Environmental air pollution can trigger adaptive
association between long-­term exposure to air pollution
diseases: a population-­based immunity.
cohort study. RMD Open and risk of developing immune-­mediated conditions.
Methods We conducted a retrospective observational ⇒ The association between long-­
term air pollution
2022;8:e002055. doi:10.1136/
study on a nationwide dataset of women and men. exposure and risk of autoimmune diseases is still
rmdopen-2021-002055
Diagnoses of various immune-­mediated diseases (IMIDs) unclear.
were retrieved. Data on the monitoring of particulate What does this study add?
matter (PM)10 and PM2.5 concentrations were retrieved ⇒ We found that long-­term exposure to air pollution
from the Italian Institute of Environmental Protection and was associated with higher risk of autoimmune
Research. Generalised linear models were employed to diseases.
determine the relationship between autoimmune diseases ⇒ In particular, exposure to particulate matter (PM)10
​rmdopen.​bmj.​com
prevalence and PM.
was associated with rheumatoid arthritis, while ex-
Results 81 363 subjects were included in the study. We
posure to PM2.5 was associated with rheumatoid
Received 25 October 2021 found a positive association between PM10 and the risk
arthritis, connective tissue diseases and inflamma-
Accepted 15 January 2022 of autoimmune diseases (ρ+0.007, p 0.014). Every 10 µg/
tory bowel diseases.
m3 increase in PM10 concentration was associated with
an incremental 7% risk of having autoimmune disease. How might this impact on clinical practice or
Exposure to PM10 above 30 µg/m3 and PM2.5 above further developments?
20 µg/m3 was associated with a 12% and 13% higher risk ⇒ Individuals chronically exposed to high levels of air
of autoimmune disease, respectively (adjusted OR (aOR) pollution might be at risk of developing autoimmune
1.12, 95% CI 1.05 to 1.20, and aOR 1.13, 95% CI 1.06 to diseases.
1.20). Exposure to PM10 was associated with an increased
risk of rheumatoid arthritis; exposure to PM2.5 was
associated with an increased risk of rheumatoid arthritis,
connective tissue diseases (CTDs) and inflammatory bowel tissue damage.1 IMIDs include a broad range
diseases (IBD). of rheumatic diseases such as rheumatoid
Conclusion Long-­term exposure to air pollution was arthritis, psoriatic arthritis, systemic lupus
associated with higher risk of developing autoimmune erythematosus, systemic sclerosis, connec-
diseases, in particular rheumatoid arthritis, CTDs and IBD. tive tissue diseases (CTDs), as well as gastro-
© Author(s) (or their
employer(s)) 2022. Re-­use Chronic exposure to levels above the threshold for human enterological diseases such as inflammatory
permitted under CC BY. protection was associated with a 10% higher risk of
bowel diseases (IBDs) and immune-­mediated
Published by BMJ. developing IMIDs.
1
neurological diseases such as multiple scle-
Rheumatology Unit, Department
of Medicine, University of
rosis. Epidemiological studies show that
Verona, Verona, Italy the incidence and the prevalence of such
2
Italian Institute for INTRODUCTION conditions are increasing steadily in the
Environmental Protection and Immune-­ mediated diseases (IMIDs) are a last decade.2 3 The reasons at the basis of
Research, Rome, Italy wide group of diseases characterised by the this increasing trend are not fully known,
Correspondence to dysregulation and uncontrolled activation of underlining the need for a more profound
Dr Giovanni Adami; the immune system, resulting in increased understanding of pathophysiology and aeti-
​adami.​g@​yahoo.​com systemic inflammation and immune-­mediated opathogenesis of such diseases. Unrevealing

Adami G, et al. RMD Open 2022;8:e002055. doi:10.1136/rmdopen-2021-002055    1

RMD Open

RMD Open: first published as 10.1136/rmdopen-2021-002055 on 15 March 2022. Downloaded from http://rmdopen.bmj.com/ on November 20, 2024 by guest. Protected by copyright.
the factors associated with the development of IMIDs heart failure and ischaemic heart diseases.13 14 Further-
is crucial in order to find out new and effective targets more, PM2.5 is associated with higher risk of developing
for therapies and, possibly, to set up appropriate public neurodegenerative and neurovascular conditions. PMs
health prevention strategies. can cross the brain–blood barrier and reach the central
Currently, it is largely accepted that the overt clinical nervous system, stimulating local immune responses and
expression of IMIDs results from the interaction between causing infarct of small vessels.15
genetic predisposition and environmental factors. Air However, despite the strong biological rationale
pollution is among the environmental factors that are linking air pollution exposure to the development of
thought to cause and exacerbate a number of IMIDs.4 various autoimmune diseases, such relationship is still a
Environmental air pollution is composed of solid matter of controversy. This debate mainly originates from
particles and gaseous substances which derive predomi- the dearth of epidemiological studies on large popula-
nantly from fossil fuel combustion coming from industry tions investigating the association between chronic expo-
production and vehicle exhaust.5 The main components sure to air pollution and IMIDs. The vast majority of the
of air pollution are particulate matters (PMs), a complex studies available in the literature concern rheumatoid
mixture of chemical elements such as heavy metal, carbo- arthritis.16–18 The exposure to air pollution and the prox-
naceous materials, persistent organic pollutants, vola- imity to streets have been associated with higher risk of
tile compounds and polycyclic aromatic hydrocarbons rheumatoid arthritis in various epidemiological studies.4
(PAHs), and a miscellany of gases (eg, carbon monoxide, Moreover, the exposure to air pollution was associated
nitrogen dioxide, ozone and sulfur dioxide). with higher chance of positivity of anticitrullinated anti-
Evidence shows that these substances can enhance bodies.19 Recently, Park and colleagues reported a higher
expression of several inflammatory pathways, stimulate risk of rheumatoid arthritis but not systemic lupus erythe-
the production of cytokines and upregulate the activa- matosus or ankylosing spondylitis in patients chronically
tion of genes involved in inflammatory response.6 PM2.5 exposed to ultrafine PM.20 However, another study by
exposure has been associated with an increased produc- Bernatsky et al reported that exposure to ultrafine PM was
tion of interleukin (IL)-­ 1, IL-­
6 and tumour necrosis associated with higher risk of immune diseases.21 Overall,
factor.7 Furthermore, PM frequently contains microbial a systematic review of eight studies published before 2016
components, for example, lipopolysaccharides, which showed that the relationship between immune-­mediated
can activate the NLPR3 inflammasome, resulting in an rheumatic diseases and air pollution was unclear.22
increased production of proinflammatory cytokines The primary objective of the present study was to deter-
that contribute to the local and systemic inflammation.8 mine and describe, using a nationwide cohort, the asso-
Moreover, PMs are able to promote the generation of ciation between long-­term exposure to air pollution and
reactive oxygen species which, in turn, activate nuclear IMIDs.
factor kappa B and stimulate the production of T-­helper
lymphocytes type 1 (Th1), resulting in DNA damage and
cell death.9 MATERIALS AND METHODS
In addition, PM exposure can negatively affect several Clinical and demographic data were extracted from the
organs and tissues, including the pulmonary, cardio- DeFRA dataset, which originated from the homonym web-­
vascular and central nervous systems. In the lungs, PMs based fracture risk assessment tool, widely diffused across
can directly interact with alveolar macrophages leading Italy. Data are entered in the DeFRA dataset directly by
to the release of Th1-­type cytokines (IL-­12 and inter- registered physicians on fracture risk calculation. DeFRA
feron gamma) that generate an intense local inflamma- users can calculate the 10-­year fracture risk in the setting
tion.10 Oxidative stress generated by PMs can also induce of osteoporosis evaluation by entering clinical features
dendritic cells to migrate to the local lymph nodes, cause of their patients on the website (https://defra-osteopo-
cell necrosis and apoptosis, and contribute to the release rosi.it/), similarly to what happened with the FRAX. We
and activation of neutrophil extracellular traps, which is retrieved data of men and women all over Italy collected
accompanied by increased IL-­17 and IL-­23 serum levels.11 from June 2016 to November 2020 by more than 3500
Moreover, local pulmonary inflammation induced by physicians (both family care practitioners and specialists).
environmental air pollution participates in the pathogen- The DeFRA dataset contains many clinical variables such
esis of rheumatoid arthritis through direct citrullination as age; weight; height; number and site of prior fragility
of the proteins, which stimulates the secretion of antici- fractures; parental history of hip and clinical vertebral
trullination peptide antibodies (ACPAs).12 Through the fractures; glucocorticoid (GC) intake (≥5 mg/day pred-
lungs, PMs may enter the circulation, reaching extrapul- nisone equivalent); treatment with adjuvant hormone
monary tissues. Evidence suggests a link between airborne therapy for breast or prostate cancer; diagnosis of rheu-
PM2.5 exposure and exacerbation of pre-­existing cardio- matoid arthritis, systemic lupus erythematosus, psoriatic
pulmonary diseases leading to increased morbidity and arthritis, systemic sclerosis, CTDs, IBDs, chronic obstruc-
mortality; in addition, chronic exposure to PMs early tive pulmonary disease, diabetes, neurological diseases
in life is directly link to the development of significant (including Parkinson’s disease, multiple sclerosis and
cardiovascular alterations, including hypertension, severe physical disability); and lumbar spine and femoral

2 Adami G, et al. RMD Open 2022;8:e002055. doi:10.1136/rmdopen-2021-002055

 Autoimmunity

RMD Open: first published as 10.1136/rmdopen-2021-002055 on 15 March 2022. Downloaded from http://rmdopen.bmj.com/ on November 20, 2024 by guest. Protected by copyright.
neck T-­scores (calculated from the bone mineral density
Table 1 General characteristics of the cohort
(BMD) reference range of young same-­sex individuals).
Overall cohort N=81 363
The dataset characteristics have been described in detail
elsewhere23–26; in this paper, the dataset has been updated Age (years) (±SD) 65 (11)
to November 2020. The DeFRA dataset has been used for Sex, female (%) 74 772 (91.9)
similar purposes of the present study in another recently Menopause, n (% of female) 71 654 (95.8)
published paper.25 Weight (±SD) 62.2 (12.0)
Data on the air pollution (PM10 and PM2.5 concen-
Height (±SD) 160 (8)
trations) were provided by the Italian Institute of Envi-
ronment Protection and Research, which collects daily Concomitant None 69 616 (85.6)
treatment (%) Treatment with adjuvant 6264 (7.7)
data from air quality stations across Italy. The long-­term
average PM concentrations were the exposure of interest. hormone therapy
Every study subject was linked to a PM exposure value, Glucocorticoids ≥5 mg/day of 5483 (6.7)
which resulted from the average concentration of urban, prednisone equivalent for ≥3
months
rural and near-­traffic stations of the subject province of
residency from January 2013 to November 2020. Patients Diseases (%) None 63 497 (78.0)
were linked to the nearest air quality station through zip Rheumatoid arthritis 3817 (4.7)
code centroids. Psoriatic arthritis 992 (1.2)
Group comparisons were performed using the Student Systemic lupus 384 (0.5)
t-­test and the Mann-­Whitney U test (for normally and erythematosus
non-­ normally distributed continuous variables, respec- Systemic sclerosis 374 (0.5)
tively). Associations between continuous variables
Other connective tissue 2571 (3.2)
were tested using Pearson correlation coefficients. diseases
Generalised linear models with robust estimators were
Inflammatory bowel diseases 1250 (1.5)
employed to identify determinants of IMIDs. Exposure
Multiple sclerosis 335 (0.4)
to PMs was analysed either as a continuous variable or
as a binary variable (exposure thresholds were 30 µg/ Diabetes 4184 (5.1)
m3 for PM10 and 20 µg/m3 PM2.5). We approached the Chronic pulmonary diseases 2097 (2.6)
analysis treating PM exposure both as a continuous vari- Severe physical handicap 1058 (1.3)
able and a dichotomous variable. We did so because we Parkinson’s disease 502 (0.6)
presumed that there might be a threshold effect of expo-
HIV infection 302 (0.4)
sure to PM. Such effect might be blunted or reduced in
Specialty of Rheumatology 17 310 (21.3)
intensity when treating the PM exposure as a continuous
the physician
variable and should be more apparent when considering Endocrinology 26 278 (32.3)
entering
thresholds. We chose the thresholds that are gener- patients’ data Family care practitioner 10 207 (12.5)
ally considered as harmful for human health. The fully (%) Internal medicine 5757 (7.1)
adjusted model included age, body mass index, meno- Orthopedy 5699 (7.0)
pause, glucocorticoid treatment, treatment with adjuvant
Physiatry 5420 (6.7)
hormone therapy for breast or prostate cancer, specialty
of the physician who entered the data and macroarea of Gynaecologist 2328 (2.9)
residency (stratified as a categorical variable: northern Other 8364 (10.3)
Italy, central Italy and southern Italy). Differences were
considered significant at a p value of <0.05. All statistical
analyses were performed using SPSS V.26. Data were (11.9% of the whole cohort) were diagnosed with an
anonymised in full compliance with the Italian Code of autoimmune disease. Table 1 shows the general charac-
Protection of Personal Data (Legislative Decree 196/03, teristics of the cohort. Data on air quality were obtained
http:// www.camera.it/parlam/leggi/deleghe/03196dl.​ from 617 air quality stations across 110 Italian provinces.
htm). No identifiers related to patients were provided Average long-­ term exposure in Italy (period of data
to the researchers. Results derived from all analyses collection 2013–2019) was 16.0 µg/m3 for PM2.5 and
were produced as aggregated summaries, which are not 25.0 µg/m3 for PM10 (figure 1). Exposure to air pollu-
possible to assign, either directly or indirectly, to the indi- tion was above average in northern Italy, Po Valley and
vidual patients. other near-­ city areas. We found a positive, yet small,
association between exposure to PM10 and the risk of
being diagnosed with an autoimmune disease (ρ 0.007,
RESULTS p 0.014; fitting of the fully adjusted model was good with
A total of 81 363 subjects were enrolled in the study. The deviance/df of 0.788). This finding translates into a 7%
vast majority were female (91.9%) and 17 866 (22%) higher risk of having any autoimmune disease every
presented at least one comorbidity. Among these, 9723 10 µg/m3 increase in PM10 concentration. As regards

Adami G, et al. RMD Open 2022;8:e002055. doi:10.1136/rmdopen-2021-002055 3

RMD Open

RMD Open: first published as 10.1136/rmdopen-2021-002055 on 15 March 2022. Downloaded from http://rmdopen.bmj.com/ on November 20, 2024 by guest. Protected by copyright.
Figure 1 (A) Prevalence of autoimmune diseases across Italy in the DeFRA database. (B) Long-­term exposure to PM of less
than 10 µm in Italy (2013–2019 average concentration in μg/m3). PM, particulate matter.

PM2.5, we did not find any association with the risk of mentioned threshold of PMs on the risk of autoimmune
autoimmune diseases (ρ 0.001, p 0.751). Table 2 shows diseases. Exposure to high levels of PM10 were associ-
the results of the generalised linear model for each ated with an increased risk of rheumatoid arthritis (aOR
immune-­mediated condition considered separately. We 1.408, 95% CI 1.271 to 1.560) but no other autoimmune
then analysed the exposure to air pollution as categor- diseases, whereas exposure to high levels of PM2.5 was
ical variables (threshold of chronic exposure was set at associated with an increased risk of rheumatoid arthritis
30 µg/m3 for PM10 and 20 µg/m3 for PM2.5). In the (aOR 1.559, 95% CI 1.401 to 1.734), CTDs (aOR 1.147,
adjusted binary logistic regression model, we found that 95% CI 1.024 to 1.286) and IBDs (1.206, 95% CI 1.028 to
subjects chronically exposed to levels of PM10 above 1.415) but no other autoimmune diseases.
30 µg/m3 had a 13% higher risk of having any autoim-
mune disease (adjusted OR (aOR) 1.122, 95% CI 1.052
to 1.196). As regards chronic exposure to PM2.5, we DISCUSSION
found that patients exposed to levels higher than 20 µg/ Herein we presented an observational, nationwide cohort
m3 had a 13% higher risk of developing autoimmune study on the association between long-­term exposure to
diseases (aOR 1.128, 95% CI 1.056 to 1.205). Figures 2 fine PM and prevalence of autoimmune diseases. We
and 3 show the effects of chronic exposure above the found that exposure to air pollution was associated with

Table 2 Association between air pollution exposure and risk of autoimmune diseases
Diseases PM10, β (95% CI) P value PM2.5, β (95% CI) P value
All autoimmune diseases 0.007 (0.001 to 0.013) 0.014 0.001 (−0.006 to 0.008) NS
Rheumatoid arthritis 0.005 (−0.004 to 0.014) NS 0.022 (0.011 to 0.034) 0.0001
Psoriatic arthritis 0.003 (−0.002 to 0.005) NS 0.004 (−0.016 to 0.023) NS
Systemic lupus erythematosus −0.007 (−0.033 to 0.019) NS −0.014 (−0.046 to 0.017) NS
Systemic sclerosis 0.013 (−0.011 to 0.038) NS 0.013 (−0.017 to 0.042) NS
Other connective tissue diseases −0.004 (−0.014 to 0.007) NS 0.015 (0.003 to 0.028 0.013
Multiple sclerosis 0.025 (−0.001 to 0.051) 0.063 0.036 (0.006 to 0.067) 0.019
NS, not significant; PM, particulate matter.

4 Adami G, et al. RMD Open 2022;8:e002055. doi:10.1136/rmdopen-2021-002055

 Autoimmunity

RMD Open: first published as 10.1136/rmdopen-2021-002055 on 15 March 2022. Downloaded from http://rmdopen.bmj.com/ on November 20, 2024 by guest. Protected by copyright.
Indeed, smoking habits has been associated with a twofold
higher risk of seropositive rheumatoid arthritis.30 Smoke
can directly elicit the post-­translational citrullination of
proteins in the lower and upper airways; these citrulli-
nated proteins are, in turn, responsible of the production
of ACPA.12 Remarkably, cigarette smoking shares various
toxics with emission from fossil fuels. As a matter of fact,
diesel exhaust has been shown to induce peptidyl argi-
nine deiminase activity in the bronchial epithelium.31
These preclinical findings have been supported by
several, real-­life, clinical observations as well. Indeed, air
pollution exposure has been associated with an increased
risk of developing rheumatoid arthritis16 and, more
recently, with a higher risk of experience flares.32 33 More-
over, the exposure to traffic-­derived pollutants has been
Figure 2 Risk of immune-­mediated conditions at chronic linked with poorer response to biological treatments in
exposure to PM10 ≥30 µg/m3. PM, particulate matter. chronic inflammatory arthritides.34 35 With our nation-
wide experience, we further confirmed such results. We
found that the prevalence of rheumatoid arthritis was
higher prevalence of IMIDs. In particular, we witnessed
nearly 50% higher in subjects exposed to high levels of
an association with rheumatoid arthritis, CTDs and IBDs
PMs. This finding is line with a meta-­analysis of previ-
in patients exposed to high levels of PM2.5.
ously published studies on the topic.17 Interestingly, the
The biological rationale supporting our results is
authors of the meta-­analysis found that PM2.5 appeared
strong. Our study has been inspired by numerous preclin-
to confer a higher risk of rheumatoid arthritis as
ical studies and several clinical observational studies.
compared with PM10 or other toxics. The incidence and
Exposure to air pollution has been linked with dysreg-
prevalence of IBDs are increasing worldwide, with accel-
ulation of the adaptive immune system with impairment
erating incidence in highly industrialised countries.36
of host response and abnormal cytokine secretion.7 As
Remarkably, only the exposure to gaseous pollutants and
an example, exposure to PAHs, assessed at individual
ultrafine PMs and not coarse PMs were associated with
level by measuring urinary 1-­hydroxypyrene, has been
an increased risk of IBDs,37 a result that is somehow in
associated with decreased number of circulating regula-
line with our findings. Indeed, we found that exposure
tory T cells and lower levels of IL-­10.27 This detrimental
to PM2.5 but not PM10 was associated with a significant
effect on the immune system has been proved to be
increase in the prevalence IBDs. The incidence of CTDs
mediated by the oxidative stress induced by air pollu-
is increasing worldwide, and the larger increase has been
tion, which is a well-­known proinflammatory stimulus.28
witnessed in western and industrialised countries,38 a
Numerous studies have also demonstrated that cigarette
result that is, again, consistent with our findings.
smoking can trigger adaptive immunity and lead to the
Of note, PM2.5 molecules, given their small diameter,
development of autoimmune diseases. A paradigmatic
are less affected by rain and weather conditions compared
example comes from rheumatoid arthritis, in which ciga-
with larger PMs.39 Indeed, PM2.5 concentrations tend to
rette smoking represents a major predisposing factor.29
do not fluctuate in response to rain and might represent
a more accurate proxy for chronic exposure to air pollu-
tion than PM10. This evidence might explain the discrep-
ancy that we found in our analysis.
We found that the prevalence of some immune-­
mediated conditions was higher in patients exposed to
fine PM. However, we did not find any significant asso-
ciation between psoriatic arthritis, systemic sclerosis,
systemic lupus erythematosus or multiple sclerosis and
exposure to air pollutants. To date, there is uncertainty
regarding the relationship between such diseases and
exposure to air pollution. As an example, there are
no clinical studies linking air pollution and psoriatic
arthritis, despite the robust association between cigarette
smoking and psoriasis that has been demonstrated,40 and
a few studies have shown a possible detrimental effect of
air pollution on psoriasis extension.41 Another example
Figure 3 Risk of immune-­mediated conditions at chronic comes for systemic sclerosis. Air pollution exposure has
exposure to PM2.5≥20 µg/m3. PM, particulate matter. been linked with higher burden of Raynaud phenomenon

Adami G, et al. RMD Open 2022;8:e002055. doi:10.1136/rmdopen-2021-002055 5

RMD Open

RMD Open: first published as 10.1136/rmdopen-2021-002055 on 15 March 2022. Downloaded from http://rmdopen.bmj.com/ on November 20, 2024 by guest. Protected by copyright.
and pulmonary manifestations,42 but no solid evidence since the dataset collects prevalently data of women at
has been produced regarding the association between high risk of fracture, the prevalence of autoimmune
exposure to traffic-­ derived pollutants and the risk of diseases might not reflect the real prevalence in the
developing systemic sclerosis. As regards systemic lupus general population of such diseases, possibly affecting
erythematosus, the clinical data are still controversial, but the generalisability of our results. Moreover, the date of
there is some evidence that air pollution is associated with the diagnosis and onset of autoimmune diseases were
this disease.43 In contrast, we did not find any association not available; hence, we cannot rule out with certainty
between systemic lupus erythematosus and PM expo- a possible reverse causality bias. The chronic exposure
sure, but the number of patients at risk was too small to pollution has been estimated from the mean concentra-
draw strong conclusions. Finally, many studies have been tion of PMs between 2013 and 2019. The exposure of
published on the association between multiple sclerosis interest might be imprecise and might just represent a
and air pollution. However, these studies yielded contro- rough estimation of the long-­term exposure. However,
versial results, with some indicating an association44 and this method is the simplest available and has been largely
others rejecting this hypothesis.45 In our analysis, we used in the literature for similar purposes.47 Indeed, we
found that multiple sclerosis was significantly associated assumed that the variation of PM was constant within
with air pollution exposure only when the PM exposure Italy in the decades before the measurement. We cannot
was treated as a continuous variable. rule out local variations of PM exposure prior to the
In our analysis, we found different results when consid- period of 2013–2020 due to expanding cities or indus-
ering PM exposure as a continuous variable or as a trial centres. However, we did not have access to pollu-
dichotomous variable. This seemingly controversial result tion models for estimating historic concentrations that
might be related to a sort of deterministic effect of PM covered the age of the study population. In addition, we
exposure on the risk of autoimmune diseases. Determin- did not have access to scholarity, climate, socioeconomic
istic phenomena are usually characterised by thresholds status, smoking status and other relevant covariates such
below which the outcome does not occur. In contrast, as occupational factors that can affect the prevalence of
stochastic effects are characterised by increasing chances IMIDs. Nevertheless, the large sample size and adjust-
of the effects as the exposure increases. In this scenario, ment for macroarea of residency should have attenuated
we might hypothesise that treating the exposure as a such confounding. Moreover, smoking and socioeco-
continuous variable might have, somehow, diluted the nomic status might not represent major confounders;
effect. indeed, it has been demonstrated that, conversely to what
There are three types of approaches to measure the was expected, regions with higher concentrations of PM
exposure to environmental air pollution. The first (the had higher income and less smokers.48 This inverse asso-
one that we used) is the air monitoring through air ciation is particularly true in Italy, where the Po Valley
quality stations; the second is direct exposure through represents the most polluted area of Italy by far but has
personal monitors; and the third requires the measure- the lowest prevalence of smokers and the highest gross
ment of biological markers that are related to exposure domestic product.49 Thus, smoking and socioeconomical
(eg, chemicals directly measured in the blood). All these status might only attenuate our results, without affecting
three methods have shortcomings and advantages. Air the overall validity. Nevertheless, the lack of information
monitoring might be inaccurate in assessing personal regarding socioeconomic factors and smoking represents
exposure, yet it allows an estimation of the long-­term a major limitation of the study. Indeed, patients with
exposure. Personal monitor represents the more precise higher scholarity and income might be more prone
method but it is limited by the ‘Hawthorne effect’ avoiding highly polluted areas and might be more likely
(ie, changing behaviour of subjects in study related to seek care for their autoimmune diseases, possibly
to increased awareness). Finally, the measurement of introducing selection bias. Furthermore, we did not have
biological metabolites of air pollutants is costly and, in access to other types of pollutants, which might be also
most of the cases, it can only examine the acute exposure associated with higher risk of immune diseases, intro-
to pollutants. To the best of our knowledge, all the epide- ducing a possible confounder. However, water and soil
miological studies that linked immune-­mediated diseases pollutants are not collected extensively, making it difficult
with air pollution were conducted using data from air to reduce the impact of such confounding. In addition,
monitoring stations or similar approaches (eg, proximity we should acknowledge some limitations common to all
to street).46 large datasets that originated from physician-­ reported
Our study should be interpreted in light of some data. For example, misclassifications and incorrect diag-
strengths and limitations. The key strength of our study nosis might have affected our findings. Moreover, the
is the large sample size and access to the diagnosis of dataset was originally designed to study determinants of
many IMIDs. Moreover, the cohort was evenly distributed fracture. The way the data are recorded may introduce a
across provinces that largely differed on PM exposure. recruitment bias, or some characteristics of these patients
Although our results are not generalisable to the entire (who are essentially women) may confound the observed
Italian population, our sample is highly representative of association. Specifically, the dataset over-­represents post-
the female population aged 50 years or more. However, menopausal women, who are also more prone to some

6 Adami G, et al. RMD Open 2022;8:e002055. doi:10.1136/rmdopen-2021-002055

 Autoimmunity

RMD Open: first published as 10.1136/rmdopen-2021-002055 on 15 March 2022. Downloaded from http://rmdopen.bmj.com/ on November 20, 2024 by guest. Protected by copyright.
autoimmune diseases. In this regard, systemic lupus 8 Uh S-­T, Koo SM, Kim Y, et al. The activation of NLRP3-­inflammsome
by stimulation of diesel exhaust particles in lung tissues from
erythematosus, which generally affects younger women, emphysema model and RAW 264.7 cell line. Korean J Intern Med
does not appear to be increased by air pollution, but, by 2017;32:865–74.
design, there are too few exposed young women in this 9 Morgan MJ, Liu Z-­gang, Liu Z. Crosstalk of reactive oxygen species
and NF-κB signaling. Cell Res 2011;21:103–15.
dataset, especially for a relatively rare disease. In addition, 10 Park E-­J, Roh J, Kim Y, et al. Pm 2.5 collected in a residential area
we did not have information regarding the seropositivity induced Th1-­type inflammatory responses with oxidative stress in
mice. Environ Res 2011;111:348–55.
of rheumatoid arthritis patients. Indeed, noxious airway 11 Stark MA, Huo Y, Burcin TL, et al. Phagocytosis of apoptotic
agents have been associated with the production of ACPA neutrophils regulates granulopoiesis via IL-­23 and IL-­17. Immunity
and seropositive rheumatoid arthritis. 2005;22:285–94.
12 Valesini G, Gerardi MC, Iannuccelli C, et al. Citrullination and
In conclusion, we found that the exposure to fine PMs autoimmunity. Autoimmun Rev 2015;14:490–7.
was associated with an increased risk of some autoim- 13 Yang S, Lee S-­P, Park J-­B, et al. PM2.5 concentration in
the ambient air is a risk factor for the development of high-­
mune diseases. Chronic exposure to traffic and industrial risk coronary plaques. Eur Heart J Cardiovasc Imaging
derived pollutants was associated with approximately 2019;20:1355–64.
40% higher risk of rheumatoid arthritis, 20% higher risk 14 Calderón-­Garcidueñas L, Vincent R, Mora-­Tiscareño A, et al.
Elevated plasma endothelin-­1 and pulmonary arterial pressure
of IBDs and 15% higher risk of CTDs. in children exposed to air pollution. Environ Health Perspect
2007;115:1248–53.
Contributors Conceptualisation, GA; data curation, GA, GC and MR; formal 15 Calderón-­Garcidueñas L, Solt AC, Henríquez-­Roldán C, et al. Long-­
analysis, GA; investigation, MR; supervision and validation: GA, MP and MR; Term air pollution exposure is associated with neuroinflammation,
writing–original draft, GA and MR. All authors contributed to the project an altered innate immune response, disruption of the blood-­brain
administration and writing–review and editing. barrier, ultrafine particulate deposition, and accumulation of amyloid
beta-­42 and alpha-­synuclein in children and young adults. Toxicol
Funding The authors have not declared a specific grant for this research from any Pathol 2008;36:289–310.
funding agency in the public, commercial or not-­for-­profit sectors. 16 Hart JE, Källberg H, Laden F, et al. Ambient air pollution exposures
and risk of rheumatoid arthritis: results from the Swedish EIRA case-­
Map disclaimer The inclusion of any map (including the depiction of any
control study. Ann Rheum Dis 2013;72:888–94.
boundaries therein), or of any geographical or locational reference, does not imply 17 Di D, Zhang L, Wu X, et al. Long-­Term exposure to outdoor air
the expression of any opinion whatsoever on the part of BMJ concerning the legal pollution and the risk of development of rheumatoid arthritis:
status of any country, territory, jurisdiction or area or of its authorities. Any such a systematic review and meta-­analysis. Semin Arthritis Rheum
expression remains solely that of the relevant source and is not endorsed by BMJ. 2020;50:266–75.
Maps are provided without any warranty of any kind, either express or implied. 18 Hart JE, Källberg H, Laden F, et al. Ambient air pollution exposures
Competing interests None declared. and risk of rheumatoid arthritis. Arthritis Care Res 2013;65:1190–6.
19 Bernatsky S, Smargiassi A, Joseph L, et al. Industrial air emissions,
Patient consent for publication Not applicable. and proximity to major industrial emitters, are associated with anti-­
Ethics approval This study was approved by the University of Verona ethic citrullinated protein antibodies. Environ Res 2017;157:60–3.
committee (prot.1876). Informed consent was not required using encrypted 20 Park JS, Choi S, Kim K, et al. Association of particulate matter with
retrospective information. autoimmune rheumatic diseases among adults in South Korea.
Rheumatology 2021;60:5117–26.
Provenance and peer review Not commissioned; externally peer reviewed. 21 Bernatsky S, Smargiassi A, Barnabe C, et al. Fine particulate air
pollution and systemic autoimmune rheumatic disease in two
Data availability statement Data are available upon reasonable request.
Canadian provinces. Environ Res 2016;146:85–91.
Open access This is an open access article distributed in accordance with the 22 Sun G, Hazlewood G, Bernatsky S, et al. Association between air
Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits pollution and the development of rheumatic disease: a systematic
others to copy, redistribute, remix, transform and build upon this work for any review. Int J Rheumatol 2016;2016:1–11.
purpose, provided the original work is properly cited, a link to the licence is given, 23 Adami G, Giollo A, Rossini M, et al. Different fracture risk profile in
and indication of whether changes were made. See: https://creativecommons.org/​ patients treated with anti-­osteoporotic drugs in real-­life. Reumatismo
licenses/by/4.0/. 2020;72:71–4.
24 Adami G, Gatti D, Rossini M, et al. Factors associated with referral
ORCID iDs for osteoporosis care in men: a real-­life study of a nationwide
Giovanni Adami http://orcid.org/0000-0002-8915-0755 dataset. Arch Osteoporos 2021;16:56.
Giovanni Orsolini http://orcid.org/0000-0003-4119-3016 25 Adami G, Cattani G, Rossini M, et al. Association between exposure
to fine particulate matter and osteoporosis: a population-­based
cohort study. Osteoporos Int 2022;33:169-­176.
26 Adami G, Gatti D, Rossini M, et al. Risk of fragility fractures in
obesity and diabetes: a retrospective analysis on a nation-­wide
cohort. Osteoporos Int 2020;31:2113–22.
REFERENCES 27 Yao Y, Wang D, Ma H, et al. The impact on T-­regulatory cell related
1 Rose NR. Prediction and prevention of autoimmune disease in the immune responses in rural women exposed to polycyclic aromatic
21st century: a review and preview. Am J Epidemiol 2016;183:403–6. hydrocarbons (PAHs) in household air pollution in Gansu, China: a
2 Cross M, Smith E, Hoy D, et al. The global burden of rheumatoid pilot investigation. Environ Res 2019;173:306–17.
arthritis: estimates from the global burden of disease 2010 study. 28 Li Z, Liu Q, Xu Z, et al. Association between short-­term exposure to
Ann Rheum Dis 2014;73:1316–22. ambient particulate air pollution and biomarkers of oxidative stress:
3 Dinse GE, Parks CG, Weinberg CR, et al. Increasing prevalence a meta-­analysis. Environ Res 2020;191:110105.
of antinuclear antibodies in the United States. Arthritis Rheumatol 29 Prisco LC, Martin LW, Sparks JA. Inhalants other than personal
2020;72:1026–35. cigarette smoking and risk for developing rheumatoid arthritis. Curr
4 Zhao C-­N, Xu Z, Wu G-­C, et al. Emerging role of air pollution in Opin Rheumatol 2020;32:279–88.
autoimmune diseases. Autoimmun Rev 2019;18:607–14. 30 Ishikawa Y, Terao C. The impact of cigarette smoking on risk of
5 Ambient (outdoor) air pollution. Available: https://www.who.int/news-​ rheumatoid arthritis: a narrative review. Cells 2020;9:475.
room/fact-sheets/detail/ambient-(outdoor)-air-quality-and-health 31 Colasanti T, Fiorito S, Alessandri C, et al. Diesel exhaust particles
6 Gawda A, Majka G, Nowak B, et al. Air pollution, oxidative stress, induce autophagy and citrullination in normal human bronchial
and exacerbation of autoimmune diseases. Cent Eur J Immunol epithelial cells. Cell Death Dis 2018;9:1073.
2017;42:305–12. 32 Adami G, Viapiana O, Rossini M, et al. Association between
7 Chen R, Li H, Cai J, et al. Fine particulate air pollution and the environmental air pollution and rheumatoid arthritis flares.
expression of microRNAs and circulating cytokines relevant to Rheumatology 2021;60:4591-­4597.
inflammation, coagulation, and vasoconstriction. Environ Health 33 Ingegnoli F, Ubiali T, Schioppo T, et al. Potential short-­term air
Perspect 2018;126:017007. pollution effects on rheumatoid arthritis activity in metropolitan areas

Adami G, et al. RMD Open 2022;8:e002055. doi:10.1136/rmdopen-2021-002055 7

RMD Open

RMD Open: first published as 10.1136/rmdopen-2021-002055 on 15 March 2022. Downloaded from http://rmdopen.bmj.com/ on November 20, 2024 by guest. Protected by copyright.
in the North of Italy: a cross-­sectional study. Int J Environ Res Public 42 Schioppo T, De Lucia O, Murgo A, et al. The burden of air pollution
Health 2021;18:8490. and temperature on Raynaud's phenomenon secondary to systemic
34 Adami G, Rossini M, Viapiana O, et al. Environmental air pollution sclerosis. Epidemiol Prev 2020;44:218–27.
is a predictor of poor response to biological drugs in chronic 43 Jung C-­R, Chung W-­T, Chen W-­T, et al. Long-­term exposure to
inflammatory arthritides. ACR Open Rheumatol 2021;3:451–6. traffic-­related air pollution and systemic lupus erythematosus in
35 Zhao N, Bernatsky S. Is air pollution linked with poor response to Taiwan: a cohort study. Sci Total Environ 2019;668:342–9.
biologics? Nat Rev Rheumatol 2021;17:583–4. 44 Bergamaschi R, Monti MC, Trivelli L, et al. PM2.5 exposure as a risk
36 Ng SC, Shi HY, Hamidi N, et al. Worldwide incidence and prevalence factor for multiple sclerosis. An ecological study with a Bayesian
of inflammatory bowel disease in the 21st century: a systematic mapping approach. Environ Sci Pollut Res Int 2021;28:2804–9.
review of population-­based studies. Lancet 2017;390:2769–78. 45 Bai L, Burnett RT, Kwong JC, et al. Long-­term exposure to air
37 Kaplan GG, Hubbard J, Korzenik J, et al. The inflammatory bowel pollution and the incidence of multiple sclerosis: a population-­based
diseases and ambient air pollution: a novel association. Am J
cohort study. Environ Res 2018;166:437–43.
Gastroenterol 2010;105:2412–9.
46 Sigaux J, Biton J, André E, et al. Air pollution as a determinant of
38 Smith E, Hoy DG, Cross M, et al. The global burden of other
rheumatoid arthritis. Joint Bone Spine 2019;86:37–42.
musculoskeletal disorders: estimates from the global burden of
disease 2010 study. Ann Rheum Dis 2014;73:1462–9. 47 Han B, Hu L-­W, Bai Z. Human exposure assessment for air pollution.
39 Borge R, Requia WJ, Yagüe C, et al. Impact of weather changes on Adv Exp Med Biol 2017;1017:27–57.
air quality and related mortality in Spain over a 25 year period [1993-­ 48 Villeneuve PJ, Goldberg MS, Burnett RT, et al. Associations between
2017]. Environ Int 2019;133:105272. cigarette smoking, obesity, sociodemographic characteristics
40 Armstrong AW, Harskamp CT, Dhillon JS, et al. Psoriasis and and remote-­sensing-­derived estimates of ambient PM2.5: results
smoking: a systematic review and meta-­analysis. Br J Dermatol from a Canadian population-­based survey. Occup Environ Med
2014;170:304–14. 2011;68:920–7.
41 Abolhasani R, Araghi F, Tabary M, et al. The impact of air pollution on 49 Gallus S, Pacifici R, Colombo P, et al. Prevalence of smoking and
skin and related disorders: a comprehensive review. Dermatol Ther attitude towards smoking regulation in Italy, 2004. Eur J Cancer Prev
2021;34:e14840. 2006;15:77–81.

8 Adami G, et al. RMD Open 2022;8:e002055. doi:10.1136/rmdopen-2021-002055