In the case-crossover analysis, each patient serves as their own control; the exposure of interest is compared in 2 different periods in the same group of patients followed longitudinally. Abbreviations: AUC, area under receiver operating curve; C6H6, benzene; CO, carbon monoxide; NO2, nitrogen dioxide; NOX, oxides of nitrogen; PM10, coarse particulate matter (2.5-10.0 μm in diameter); PM2.5, fine particulate matter (<2.5 μm in diameter).
Mean concentration in the 60 days before assessment. Abbreviations: PM10, coarse particulate matter (2.5-10.0 μm in diameter); PM2.5, fine particulate matter (<2.5 μm in diameter).
eTable 1. Air pollutants concentrations (as mean and AUC) in 15 days before the flare and the control visit.
eTable 2. Air pollutants concentrations (as mean and AUC) in 30 days before the flare and the control visit.
eTable 3. Air pollutants concentrations (as mean and AUC) in 45 days before the flare and the control visit.
eTable 4. Case-crossover sensitivity analysis: air pollutants concentrations (as mean and AUC) in 15 days before the flare visit vs the 15 days before the same visit separated by time frame of 30 days.
eTable 5. Air pollutants concentrations (as AUC) in the 60 days before the flare and the control visit in patients aged <50 vs ≥50 years.
eTable 6. Odds of having PASI ≥5 at different threshold of exposures of PM10 and PM2.5 (mean concentration in the 15 days before assessment).
eTable 7. Odds of having PASI ≥5 at different threshold of exposures of PM10 and PM2.5 (mean concentration in the 30 days before assessment).
eTable 8. Odds of having PASI ≥5 at different threshold of exposures of PM10 and PM2.5 (mean concentration in the 45 days before assessment).
eTable 9. Odds of having PASI ≥5 at different threshold of exposures of PM10 (mean concentration in the 60 days before assessment) stratified for trimester.
eTable 10. Odds of having PASI ≥5 at different threshold of exposures of PM2.5 (mean concentration in the 60 days before assessment) stratified for trimester.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Bellinato F, Adami G, Vaienti S, et al. Association Between Short-term Exposure to Environmental Air Pollution and Psoriasis Flare. JAMA Dermatol. 2022;158(4):375–381. doi:10.1001/jamadermatol.2021.6019
Does short-term exposure to environmental air pollution increase the risk of psoriasis flares?
In this case-crossover and cross-sectional study involving 957 patients affected by chronic plaque psoriasis with 4398 follow-up visits, the concentrations of air pollutants were significantly higher in the period before psoriasis flare compared with the control visit.
These findings support that air pollution may be a trigger factor for psoriasis flares.
Psoriasis is a chronic inflammatory disease with a relapsing-remitting course. Selected environmental factors such as infections, stressful life events, or drugs may trigger disease flares. Whether air pollution could trigger psoriasis flares is still unknown.
To investigate whether short-term exposure to environmental air pollution is associated with psoriasis flares.
Design, Setting, and Participants
This observational study with both case-crossover and cross-sectional design retrospectively analyzed longitudinal data from September 2013 to January 2020 from patients with chronic plaque psoriasis consecutively attending the outpatient dermatologic clinic of the University Hospital of Verona. For the case-crossover analysis, patients were included who had at least 1 disease flare, defined as Psoriasis Area and Severity Index (PASI) increase of 5 or greater between 2 consecutive assessments in a time frame of 3 to 4 months. For the cross-sectional analysis, patients were included who received any systemic treatment for 6 or more months, with grade 2 or higher consecutive PASI assessment.
Main Outcomes and Measures
We compared the mean and cumulative (area under the curve) concentrations of several air pollutants (carbon monoxide, nitrogen dioxide, other nitrogen oxides, benzene, coarse particulate matter [PM; 2.5-10.0 μm in diameter, PM10] and fine PM [<2.5 μm in diameter, PM2.5]) in the 60 days preceding the psoriasis flare and the control visits.
A total of 957 patients with plaque psoriasis with 4398 follow-up visits were included in the study. Patients had a mean (SD) age of 61 (15) years and 602 (62.9%) were men. More than 15 000 measurements of air pollutant concentration from the official, open-source bulletin of the Italian Institute for Environmental Protection and Research (ISPRA) were retrieved. Among the overall cohort, 369 (38.6%) patients with psoriasis flare were included in the case-crossover study. We found that concentrations of all pollutants were significantly higher in the 60 days before psoriasis flare (median PASI at the flare 12; IQR, 9-18) compared with the control visit (median PASI 1; IQR, 1-3, P < .001). In the cross-sectional analysis, exposure to mean PM10 over 20 μg/m3 and mean PM2.5 over 15 μg/m3 in the 60 days before assessment were associated with a higher risk of PASI 5 or greater point worsening (adjusted odds ratio [aOR], 1.55; 95% CI, 1.21-1.99; and aOR, 1.25; 95% CI, 1.0-1.57, respectively). Sensitivity analyses that stratified for trimester of evaluation, with various lag of exposure and adjusting for type of treatment, yielded similar results.
Conclusions and Relevance
The findings of this case-crossover and cross-sectional study suggest that air pollution may be a trigger factor for psoriasis flare.
Environmental air pollution is defined by the World Health Organization (WHO) as the contamination of the ambient air by any chemical, physical, or biological agent that modifies the natural characteristic of the atmosphere.1 The main pollutant compounds include coarse particulate matter (PM; 2.5-10.0 μm in diameter, PM10) and fine PM (<2.5 μm in diameter, PM2.5); gaseous products, ie, carbon monoxide (CO) and oxides of nitrogen (NOx); and volatile organic products such as benzene (C6H6). These compounds are produced mostly by fossil fuel combustion from vehicle and industry emissions. Additional secondary pollutions are derived from the interaction of volatile organic compounds and NOx upon UV radiation photoactivation.2 Air pollution exerts negative effects on human health, and has been associated with stroke, heart disease, chronic obstructive pulmonary disease, osteoporosis, and infertility.3-6 After inhalation, pollutants can circulate in the blood stream, exerting oxidative damage and causing inflammation. Moreover, air pollutants can directly come into contact with the skin.7 Exposure to diesel exhaust particles could activate skin resident T-cells, resulting in abnormal production of pro-inflammatory cytokines including tumor necrosis factor α (TNF-α) and interleukins (ILs), including IL-1α, IL-1β, IL-6, and IL-8.8 Indeed, worsening of different inflammatory cutaneous diseases including atopic dermatitis, photoaging, and acne have been associated with exposure to air pollution.9-13 Psoriasis is a chronic inflammatory disease with a relapsing-remitting course and selected environmental factors such as infections and/or drugs may trigger disease flares.14 Whether air pollution could trigger psoriasis flares is not known.15 The objective of this study was to investigate whether short-term exposure to environmental air pollution is associated with psoriasis flares.
We retrospectively analyzed data on patients with chronic plaque psoriasis from the electronic medical records of the Dermatology Unit of the University Hospital of Verona, Italy, from September 2013 to January 2020. Inclusion criteria were (1) patients with any continuous systemic treatment for at least 6 months before dermatological visit, (2) patients with at least 2 consecutive measurements of Psoriasis Area and Severity Index (PASI) at 3 to 4 month intervals, (3) patients with residency (deemed by zip code) within 10 km of the air pollutant collection point, and (4) patients with at least 1 disease flare, defined as PASI increase of 5 or greater between 2 dermatological visits separated by a time frame of 3 months.16,17 The following clinical and demographic parameters were considered: sex; disease duration; zip code of residency; and systemic treatment with conventional and biological therapies including methotrexate, cyclosporine, oral retinoids, dimethyl fumarate, infliximab, etanercept, adalimumab, ustekinumab ixekizumab, secukinumab, brodalumab, and apremilast.
The study was conducted according to the protocol BIOREVE 534CESC and reviewed and approved by the University of Verona local Ethics Committee, in accordance with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Exemption of informed consent of study participants was granted by the local Ethics Committee.
Air pollutant daily concentrations (μg/m3) of the Verona area related to the same period were collected from the official, open-source, bulletin of the Italian Institute for Environmental Protection and Research (ISPRA).18 In particular, the following air pollutants were available: CO, NO2, NOx, C6H6, PM10, and PM2.5. Short-term exposure was defined as mean and area under the curve (AUC) concentration in the 60 days immediately before the dermatological visit and the PASI assessment. Patients’ data and air pollution exposure were linked through zip code centroids.
A case-crossover analysis was designed to investigate the temporal association between psoriasis flare and short-term exposure to air pollution (Figure 1). In the case-crossover analysis each patient serves as their own control; the exposure of interest is compared in 2 different periods in the same group of patients followed longitudinally. The case-crossover study has been commonly applied to investigate the association between short-term exposures to potential trigger factors and the risk of acute disease and/or flare of chronic diseases. We examined the concentrations of air pollutants as cumulative AUC exposure in the 60 days preceding the 2 consecutive dermatologic visits (“control visit” and “flare visit”). These 60-day intervals were anchored to the 2 PASI measurements (day 0 and day −60, respectively). Therefore, every patient contributed to a hazard period (the 60 days preceding the flare visit) and a control period (the 60 days preceding the control visit) (Figure 1). Shorter times of exposure to air pollution (ie, 15, 30, and 45 days) and other criteria of psoriasis flare (ie, 50%-100% increase in PASI) were considered in sensitivity analyses. To assess the influence of the cumulative age damage, we ran the main analysis stratifying the population by the age of the patients (< 50 vs ≥ 50 years). As sensitivity analysis, we assessed the concentrations of air pollutants as mean and cumulative AUC exposure during the 15 days hazard period before the flare visit vs the 15 days before the same visit separated by a time frame of 30 days.
A cross-sectional analysis was also performed. Inclusion criteria were (1) patients with any continuous systemic treatment for at least 6 months before dermatological visit, (2) patients with 2 or more consecutive measurements of PASI, and (3) patients with residency (deemed by zip code) within 10 km of the air pollutant collection point. The PASI measurements of each visit of each patient were correlated with the pollutant concentrations of the previous 60 days and this analysis was adjusted for seasonality (trimester) and for systemic anti-psoriatic treatment (conventional or biological).
Association between continuous variables was tested using Pearson correlation coefficients and multivariate linear regression. Binary logistic regression was used to determine the association between exposure to pollutants and psoriasis flare (ie, PASI increase ≥5 points, considered as categorical variable) in the case-crossover analysis. Linear and binary logistic regression was used to assess the association between different thresholds of PMs (ie, ≥15, 20, 25, 30 μg/m3 of PM10 and ≥10, 15, 20, 25 μg/m3 of PM2.5)18 and PASI score 5 or greater in the cross-sectional analysis. Air pollutant concentrations of the flare and control visit were compared with t-student and Mann-Whitney U tests (for normally and non-normally distributed continuous variables, respectively). All differences were considered significant when P value was inferior to .05. All statistical analyses were performed using SPSS Version 26 (SPSS Inc).
Data on 957 patients with psoriasis with 4398 follow-up visits and more than 15 000 measurements of air pollutant concentrations were retrieved. The mean (SD) age of the patients was 61 (15) years, of whom 602 (62.9%) were men, with a mean (SD) disease duration of 24 (17) years. The patients were visited every 3 to 4 months for a median follow-up of 2.7 years (IQR, 1.1-5.2). During this follow-up period, the patients were exposed to an average concentration of 0.34 μg/m3 CO, 28.87 μg/m3 NO2, 44.66 μg/m3 NOx, 0.57 μg/m3 C6H6, 29 μg/m3 PM10, and 21 μg/m3 PM2.5.
Among the overall study population (n = 957), we included in the case-crossover analysis 369 (38.6%) patients with at least 2 consecutive visits, 1 psoriasis flare and a continuous systemic treatment for at least 6 months (Table 1). In these selected patients, the median PASI at the flare visit (12; IQR, 9-18) was significantly higher than PASI at the control visit (1; IQR, 1-3), P < .001. Notably, we found that the air pollutant concentrations (as mean and AUC) were higher in the 60 days before the flare compared with the control visit (Table 2). We performed sensitivity analyses applying different definitions of psoriasis flare such as 50% and 100% increase in PASI. A total of 515 (35.8%) patients had at least a 50% increase and 452 (47.2%) had at least a 100% increase in PASI compared with the control visit, respectively. Even adopting different definitions of flare, we found that all air pollutant concentrations remained significantly higher before the flare visit compared with the control visit. In addition, we conducted further sensitivity analysis restricting the lag of exposures to 15, 30, and 45 days that confirmed similar results (eTables 1, 2, and 3 in the Supplement). To mitigate the impact of seasonality, even restricting the assessment to a trimester period, we observed that the air pollutant concentrations were still significantly different comparing the 15 days before the flare visit vs the 15 days before the same visit separated by a time frame of 30 days (eTable 4 in the Supplement). Interestingly, as far as patients younger than 50 years old, we did not find significantly higher air pollutants concentrations in the 60 days before the flare compared with the control visit (eTable 5 in the Supplement).
A total of 4072 visits from 957 patients were included for the cross-sectional analysis. We found that the short-term air pollutant concentrations were positively associated with absolute PASI, independently from the type of treatment (conventional vs biologic treatment) and the season of the visit. Linear regression model β coefficients for PM10, PM2.5, NO2, NOx, CO, and C6H6 were 0.052 (P = .006), 0.052 (P = .006), 0.092 (P < .001), 0.071 (P < .001), 0.116 (P < .0001), and 0.025 (p 0.043), respectively. Sensitivity analyses using AUC exposure considering the 15-day, 30-day, and 45-day periods before visiting yielded similar results (eTables 6, 7, and 8 in the Supplement). We found that the mean and AUC air pollutant concentrations were higher in 60 days before the visits with PASI 5 or greater compared with the visits with PASI below 5 (Table 3). Exposure to mean PM10 above 20 μg/m3 and mean PM2.5 above 15 μg/m3 were associated with 50% and 25% higher risk of having an absolute PASI score of 5 or more points, respectively (adjusted odds ratio [aOR] 1.55; 95% CI, 1.21-1.99; and aOR 1.25; 95% CI, 1.0-1.57; respectively). At thresholds of 15 μg/m3 for PM10 and 10 μg/m3 for PM2.5 the association was no longer evident. Binary logistic regression model results showing the association between different thresholds of PM concentrations and PASI score 5 or greater are reported in Figure 2. Finally, we conducted subgroup analyses by stratifying the cohort based on trimester of the visit in order to adjust for the season of the visit; these investigations produced similar findings (eTables 9 and 10 in the Supplement).
We found that higher concentration of different air pollutants was associated with psoriasis flares in patients living in an industrialized city of the Po valley (Verona). The concentrations of air pollution were higher in the 2 months preceding a flare visit, compared with the 2 months preceding the control dermatologic assessment. In addition, we found that PASI correlated with air pollutants concentrations during a median 2.7-year period follow-up and such association was independent from seasonality and anti-psoriasis treatment applied. The present study results have important clinical reflections, suggesting that environmental air pollutant fluctuations may affect psoriasis course.
The evidence about the risks of the exposure to air pollution on human health is increasing at an alarming rate.19,20 Oxidative stress and inflammation (ox-inflammation) are the 2 central recognized physio-pathological mechanisms of the damage caused by air pollutants. Reactive oxygen species (ROS) and reactive nitrogen species (RNS) are generated during phagocytosis of the PMs, as well as gaseous pollutants and UV light exposure.21-23 When the production of ROS/RNS overcomes the antioxidant defenses, lipid peroxidation products (ie, 4-hydroxynonenal) are formed and proteolytic activity occurs. Oxidative species can activate different redox-sensitive factors, such as proinflammatory nuclear factor kappa-light chain-enhancer of activated B cells (NF-κB), activator protein-1 (AP-1), Nrf2, and heat shock proteins.24 Creating an imbalance in the production of ROS/RNS, air pollutants can finally activate the inflammasome, a cytosolic multiprotein oligomer complex of the innate immune system.24 When activated, the inflammasome initiates the production of pro-inflammatory cytokines, such as IL-1β and IL-18. These cytokines can further induce the production ROS/RNS by endogenous enzyme activation, leading to a positive feedback loop.24 After inhalation, air pollutants can cause oxidative stress, not only in the airway’s epithelia but also reaching peripheral tissues, such as skin, because pro-inflammatory cytokines are transferred into the bloodstream.25,26 The cutaneous barrier is primarily exposed to air pollutants that are capable of reaching the dermis and subcutaneous tissue directly through the epidermis or via hair follicles and glands.27 Ultrafine carbon nanoparticles can interfere with the differentiation of keratinocytes affecting the expression of different genes, particularly upregulating early-stage keratinocyte markers (KRT8 and KRT18) and downregulating late-stage markers (KRT14 and KRT 5). Cheng et al28 found that increasing concentration of carbon nanoparticles upregulates not only pro-inflammatory chemokines (CCL20, CXCL1, CXCL2, and CXCL3) and cytokines (IL-6 and IL-1β), but also 2 psoriasis-related genes (S100A9 and S100A7). Various environmental factors are known to exert a critical role in the flares of psoriasis. For example, organic components of pollutants can exacerbate pruritus via activation of the transcription factor aryl hydrocarbon receptor.29 Exposure to cadmium can predispose a patient to the worsening of psoriasis.30 Heavy smoking can decrease the treatment effect of TNF-α inhibitors and induce psoriasis exacerbations.31 Of note, cigarette smoke is a complex mixture of ROS, CO, reactive nitrogen species, and aldehydes, closely resembling traffic- and industry-derived air pollution.32
In the case-crossover study, we found that the concentrations of air pollutants were considerably higher before a disease flare, as compared with the control visit. Disease flares, defined as a PASI increase of 5 or more points from the previous visit, were registered in 369 patients, approximately 40% of the cohort, over a median follow-up time of about 3 years. This flare rate is in line with other real-life cohort studies on psoriatic patients; indeed, Egeberg et al33 reported that 50% of psoriasis patients had a self-reported flare of the disease in the previous 12 months. Interestingly, we found that the risk of having a PASI score 5 or greater was elevated even at thresholds of exposure that are largely considered as safe (ie, below the threshold suggested by European Union air quality directive).34 Indeed, the risk for having a PASI score of 5 or greater was 40% to 50% higher at exposures as low as 20 μg/m3 of PM10 and 15 μg/m3 of PM2.5 (60-day period before assessment). A clear dose–response relationship between higher pollution and greater risk of psoriasis flare is scarce and cannot strengthen the causal inference of the results. A possible explanation might be found in the reduced sample size of subjects exposed to higher levels of pollution. Indeed, the number of days above a given threshold progressively decreases with the increase of the threshold. Similar conclusions have been previously drawn by Adami et al35 in a large cohort of patients with rheumatoid arthritis. The authors also found that environmental air pollution was a determinant of poor response to biological disease-modifying antirheumatic drugs in patients with chronic inflammatory arthritides.36
The present study has strengths and limitations. The main strength of the study is the wide cohort of patients with chronic psoriasis followed longitudinally for more than 7 years. Moreover, we had access to a comprehensive data set of daily measurements of toxic air compounds. We corroborated the results by multiple sensitivity analyses with different definitions of flares and temporal threshold of exposure. We did not need to adjust for the influence of systemic treatment because of the case-crossover design, where every patient is compared with themself. A limitation of the study is the definition of flare, based on a clinical score potentially influenced by different measurements, rather than an objective laboratory parameter, such as C-reactive protein.37 Second, we did not have access to annual weather conditions to adjust for UV sunlight exposure, which may be both correlated with psoriasis severity and pollutant concentrations. To overcome a possible bias introduced by weather conditions, we stratified the analyses for trimesters of visit and found similar results. Moreover, in the cross-sectional analysis, we conducted a logistic regression analysis that corrected for seasonality and systemic anti-psoriatic treatment. In addition, it has been demonstrated that UV sunlight can potentiate the pro-inflammatory effect of air pollution, possibly attenuating the bias introduced by seasonality.37-39 We chose the 60-day period of exposure for the main analysis to better represent the cumulative burden of the acute exposition to air pollution. Nonetheless, such lag of exposure is arbitrary and, since there is a dearth of publications on psoriasis severity and air pollution, is not supported by data. Notwithstanding that, we conducted multiple sensitivity analyses with various exposure definitions that produced comparable results. We did not consider the potential influence of the application of topical treatments or stressful and infective events, but we assumed that such eventualities were homogenously distributed in the whole period of interest. In addition, we did not have access to indoor pollution, initiation/cessation of smoking habit, and respiratory comorbidities. Nevertheless, the case-crossover design controls for time-invariant characteristics such as comorbidities, genetic characteristics, and chronic smoking habits. Finally, we assumed that zip code reflected the permanent residence, without considering unreported change of habitation. However, we admitted just a minimal rate of relocations in our cohort, probably not affecting the overall picture of the study. Finally, the results may not be generalizable to other psoriasis cohorts owing to the peculiar environmental setting of the Po valley. Indeed, we conducted this study in a highly polluted and scarcely ventilated area.
In conclusion, short-term air pollution exposure is associated with increased psoriasis activity and likelihood of having a psoriasis flare. Further study is needed to examine whether these findings generalize to other populations and to better understand the mechanisms by which air pollution may affect psoriasis disease activity.
Accepted for Publication: December 22, 2021.
Published Online: February 16, 2022. doi:10.1001/jamadermatol.2021.6019
Corresponding Author: Francesco Bellinato, MD, Section of Dermatology and Venereology, Department of Medicine, University of Verona, Piazzale A. Stefani 1, 37126 Verona, Italy (email@example.com; firstname.lastname@example.org).
Author Contributions: Drs Bellinato and Adami had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Bellinato and Adami equally contributed to the manuscript.
Concept and design: Bellinato, Adami, Benini, Fassio, Girolomoni, Gisondi.
Acquisition, analysis, or interpretation of data: Bellinato, Adami, Vaienti, Gatti, Idolazzi, Rossini.
Drafting of the manuscript: Bellinato, Adami, Vaienti, Benini, Gisondi.
Critical revision of the manuscript for important intellectual content: Bellinato, Adami, Benini, Gatti, Idolazzi, Fassio, Rossini, Girolomoni, Gisondi.
Statistical analysis: Bellinato, Adami, Idolazzi.
Obtained funding: Bellinato.
Administrative, technical, or material support: Bellinato, Gatti, Idolazzi, Girolomoni.
Supervision: Bellinato, Adami, Gatti, Idolazzi, Fassio, Rossini, Girolomoni, Gisondi.
Other – data obtainment and analysis: Vaienti.
Conflict of Interest Disclosures: Dr Adami reported personal fees from Thermex outside the submitted work. Dr Idolazzi reported personal fees from AbbVie, Amgen, Biogen, Merck Sharp & Dohme, Eli Lilly, Novartis, Celgene, Sandoz, Janssen, and UCB outside the submitted work. Dr Girolomoni reported personal fees from AbbVie, Almirall, Amgen, Biogen, Boehringer-Ingelheim, Bristol Meyers Squibb, Eli Lilly, Galderma, Genzyme, Leo Pharma, Pfizer, Regeneron, Samsung, and Sanofi outside the submitted work. Paolo Gisondi received personal fees from AbbVie, Almirall, Amgen, Biogen, Eli Lilly, Jansenn, Novartis, Sanofi, and UCB. Dr Rossini received personal fees from AbbVie, Amgen, Bristol Myers Squibb, Eli Lilly, Galapagos, Novartis, Pfizer, Sandoz, Theramex, and UCB. No other disclosures were reported.
Additional Information: Compliance with ethical standard: The study was conducted according to the protocol BIOREVE 534CESC approved by the Ethics Committee of the University of Verona Hospital, in accordance with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Patient and Public Involvement statement: This research was done without patient involvement. Patients were not invited to comment on the study design and were not consulted to develop patient relevant outcomes or interpret the results. Patients were not invited to contribute to the writing or editing of this document for readability or accuracy. Transparency declaration: Dr Bellinato (the manuscript’s guarantor) affirms that the manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.