ABSTRACT
Background
Particulate matter ⩽2.5 µm in diameter (PM2.5) and its role in chronic rhinosinusitis (CRS) pathogenesis have gained heightened attention. We previously demonstrated that PM2.5 exposure may bias the nasal mucosa in CRS toward a Type 2 inflammatory pathway. However, there are limited data comparing cytokine changes in CRS sinonasal tissue to non‐CRS patients as it relates to PM2.5 exposure. We hypothesized that long‐term exposure preferentially increases the risk of manifesting CRS with nasal polyposis (CRSwNP).
Methods
We performed a retrospective analysis of 376 patients (308 CRS, 68 controls) who underwent endoscopic sinus or skull base surgery. A spatiotemporal machine‐learning model estimated daily PM2.5 levels for 1 year prior to each patient's surgery date. Cytokines were quantified using a multiplex flow cytometric bead assay and compared to estimated PM2.5 exposure using Spearman correlation and multivariate regression. Patients with high and low 12‐month PM2.5 exposures were matched across age, sex, income, and rurality using a nearest neighbor algorithm. Multivariate adjusted logistic regression was used to estimate the odds of CRS based on PM2.5 exposure.
Results
Reduced IL‐10 levels were associated with higher PM2.5 exposures in control patients (β = −0.735, p = 0.0196). In exposure‐matched logistic regression analysis, high 12‐month PM2.5 exposure was an independent predictor of CRSwNP (β = 1.97, OR: 7.22, p = 0.0001) after adjustment for age, income, rurality, and comorbid asthma/allergic rhinitis. A similar relationship was not identified for CRSsNP.
Conclusions
PM2.5 exposure is associated with reduced IL‐10 in control patients compared to CRS and may increase odds of CRSwNP development.
Keywords: allergic rhinitis, air pollution, aeroallergens, chronic rhinosinusitis, particulate matter
1. Introduction
While heterogeneous in presentation and severity, chronic rhinosinusitis (CRS) is thought to be perpetuated by mucosal epithelial tight‐junction violation and dysfunctional mucociliary clearance [1, 2]. However, the underlying source of these impairments and resultant mucosal inflammation remains poorly characterized. The most recent theories point to multifaceted and synergistic interactions between the mucosal epithelium and microbiome, barrier defense mechanisms, and environmental exposures [3, 4].
Historically, CRS has been broadly categorized into phenotypic presentations based on the presence (CRS with nasal polyposis [CRSwNP]) or absence (CRSsNP) of polyps, with the initial assumption that these phenotypes demonstrate somewhat distinct disease entities defined by different pathologic triggers and processes [5]. With the popularization of CRS endotyping, researchers have observed a more intricate picture of interdependent processes across these two phenotypes. Type 1 inflammation is defined by elevated interferon (IFN)‐γ, interleukin (IL)‐1B, IL‐6, IL‐8, and IL‐12 with a neutrophilic predominance. Type 2 inflammation is characterized by IL‐4, IL‐5, and IL‐13 and eosinophilic predominance. Type 3 inflammation is linked to IL‐17 and IL‐22 driven innate and adaptive immunity [6]. Although Type 1 inflammation is classically thought to be the driver of CRSsNP and Type 2 of CRSwNP, many patients express markers across phenotypes or do not cleanly fit into one endotype [7]. In addition, regional heterogeneity in cytokine expression patterns across CRS patients living in different parts of the world supports the notion that environmental exposures significantly impact CRS pathogenesis [8, 9].
Environmental pollutants, including particulate matter less than or equal to 2.5 µg in diameter (PM2.5), have been well studied in lower airway diseases such as asthma and have been shown to cause reactive oxygen species (ROS) production, altered gene transcription, and inflammatory cytokine upregulation, among other cellular metabolic changes in humans [5, 10, 11]. These pollutants are emitted into the ambient air from natural sources, including fires and dust/particle storms, but the majority of PM2.5 come from anthropogenic sources, including fuel consumption, transportation, and industrial processes [12]. The nasal cavity and paranasal sinuses serve as the first location of pollutant‐mucosal exposure within the respiratory system, and researchers have utilized the strong literature on environment‐exacerbated lower airway pathology as a platform to better delineate the role of particulate matter in CRS pathogenesis.
PM2.5 exposure and subsequent eosinophilic nasal inflammation have been demonstrated in several animal models. For example, Ramanathan et al. observed decreased expression levels of claudin‐1 and E‐cadherin epithelial junction protein in mice who were exposed to PM2.5 over 16 weeks [13]. Electron microscopy studies in rabbit models showed ciliary structure disorganization, goblet cell hyperplasia, and collagen deposition in rabbit nasal tissue, further implicating the role of ambient particulate matter in tissue remodeling [14]. Within the human CRS population, research has largely been confined to clinical associations between increased PM2.5 exposure and CRS diagnosis in case–control studies. Recently, however, our group recently demonstrated a positive correlation between Type 2 cytokine levels, including IL‐5 and IL‐13, as well as IL‐2, 12, and 21 with estimated 12‐month preoperative PM2.5 exposure in over 300 CRS patients undergoing endoscopic sinus surgery (ESS) [15]. Additionally, eosinophil counts in sinonasal tissue were also higher in patients with increased exposure to PM2.5. These associations were still durable even after adjusting for age, income, body mass index (BMI), rurality, polyps, asthma, and allergic rhinitis, suggesting that PM2.5 exposure may be linked to a mixed, Type 2 predominant inflammatory response.
Although our previous work demonstrated a novel understanding of possible nasal inflammatory microenvironment changes in CRS patients with varying levels of PM2.5 exposure, a knowledge gap still exists in contextualizing these differences in comparison to non‐CRS patients with similar exposure levels. Thus, the purpose of this study was to investigate differences in nasal mucus cytokine levels as they relate to PM2.5 exposure in CRS and non‐CRS patients and use exposure‐matching methods from the causal inference literature to estimate the effect of chronic particulate matter exposure on the risk of developing sinonasal disease.
2. Methods
2.1. Patient Selection
The study was approved by Vanderbilt University Medical Center's (VUMC) institutional review board. Patients who presented to the Vanderbilt Rhinology/skull base surgery clinic with bilateral CRS who underwent ESS between 2015 and 2021 were included. In addition, healthy control patients were recruited from the population of patients undergoing skull base surgery for non‐secretory pituitary adenomas or colony stimulating factor (CSF) rhinorrhea. Exclusion criteria included patients with evidence of odontogenic rhinosinusitis or suspected mycetoma, current monoclonal antibiotic therapy, systemic steroid use within one month of surgery, cystic fibrosis, or known autoimmune disease. A CRS diagnosis was defined by the International Consensus Statement on Allergy and Rhinology Rhinosinusitis and American Academy of Otolaryngology—Head and Neck Surgery Clinical Practice Guidelines [4]. Patients included in this study underwent elective ESS after appropriate trial of medical management.
2.2. Middle Meatus Mucus Collection and Processing
Mucus samples were ascertained from the middle meatus of the nasal cavity at the start of the surgical procedure, with the mucus collection process and cytokine quantification methods extensively outlined in the previous literature [16, 17, 18]. In summary, polyurethane sponges (Summit Medical, St. Paul, MN) were placed in the middle meatus under endoscopic guidance and removed after five minutes for immediate processing via a multiplex cytometric bead assay (BD Sciences Franklin Lakes, NJ). Levels of IL‐1α, IL‐1β, IL‐2, IL‐4, IL‐5, IL‐6, IL‐7, IL‐8, IL‐9, IL‐10, IL‐12, IL‐13, IL‐17A, IL‐21, tumor necrosis factor (TNF‐α), IFN‐γ, eotaxin, regulated‐on‐action, normal T‐cell‐expressed and secreted (RANTES), and granulocyte‐macrophage CSF (GM‐CSF) were quantified and recorded in both control and CRS patients.
2.3. PM2.5 Exposure Estimation
We employed a validated, open‐source spatiotemporal PM2.5 exposure estimation model to predict the daily exposure for each participant near their home for the year up to each sample collection date. This model was previously used in a prior study by our group [15]. Participant's home zip codes were translated into latitude and longitude coordinates for geolocalization, which was then used with the addPMData package in R to obtain daily exposure estimates [19]. Published by Brokamp et al., this machine‐learning model is based on Environmental Protection Agency (EPA) data and incorporates land use, meteorology, emissions, and traffic, as well as other geospatial data in obtaining exposure estimates [20].
2.4. Statistical Methods
Data were collected and stored on REDCap (Vanderbilt University, Nashville, TN), a Health Insurance Portability and Accountability Act (HIPAA)‐compliant database. Categorical comparisons between CRS and control groups were performed using chi‐square tests. The potential association between mucus inflammatory cytokine levels and estimated 12‐month mean PM2.5 levels in control patients was initially investigated using a univariate linear regression analysis. Cytokine counts were log‐transformed to account for the skewed nature of inflammatory cytokines, similar to prior work in CRS patients [15, 16]. Finally, to account for the potential impact of various confounders such as age, BMI, allergic rhinitis/asthma, AGI, and rurality on cytokine profiles, a multivariate regression analysis was performed to estimate cytokine levels, including these covariates along with mean 12‐month PM2.5 levels. These associations in our control cohort were compared to the same analyses performed from our CRS cohort, which has been previously described by our group [15].
To estimate the odds of developing CRS with chronic PM2.5 exposure, we employed an exposure‐matching approach utilizing a nearest neighbor algorithm implemented in the MatchIt package in R to simulate random assignment of high PM2.5 exposure as would occur during a hypothetical randomized trial of PM2.5 exposure. This exposure‐matching approach has been validated and cited in the statistical and public health literature and is generally used in settings where conducting a randomized trial would be logistically or ethically unfeasible [19, 21, 22, 23]. High PM2.5 was defined as a 12‐month PM2.5 value above the mean value for the entire cohort. Briefly, this method takes a set of input matching variables that could be potential confounders of high PM2.5 exposure and identifies the most similar matched patient with low PM2.5 exposure based on these variables using a scaled Euclidean distance. Note that this approach is distinct from most matched case–control studies that seek to match participants based on the outcome of interest and retrospectively assess for differences in exposures; a recent such study with a large cohort of patients indeed found strong evidence of an association between PM2.5 and severe CRS [24]. Instead, participants were matched with respect to age, sex, AGI, and RUCA1 scores to high and low PM2.5 exposure groups, from which the relative distribution of CRS patients and controls was assessed using 2 × 2 contingency tables and chi‐square tests as the outcome measure. This resulted in an equal number of exposure‐matched participants in each PM2.5 exposure category, with 119 participants in both high and low PM2.5 groups in the CRSwNP/control comparison and 88 participants in each group for CRSsNP. A multivariate logistic regression model was then used on the matched data to estimate odds ratios for developing CRS based on exposure to high PM2.5 levels, with adjustments for age, AGI, RUCA1, asthma, and allergic rhinitis. All analyses were conducted in R version 4.2.1 (R Core Team, Vienna, Austria) as hosted on the Advanced Computing Center for Research and Education (ACCRE) compute cluster at Vanderbilt University.
3. Results
A total of 376 control and CRS patients met inclusion criteria; characteristics of these patient groups are noted in Tables 1 and 2, respectively. As anticipated, there were significantly more participants with allergic rhinitis and asthma in the CRS group (p < 0.0001) compared to controls. CRS patients also had a statistically lower BMI compared to controls (p < 0.0001), likely indicative of known association between BMI and CSF rhinorrhea. Estimated mean PM2.5 exposure over 12 months for the entire combined cohort was 8.12 µg/m3 (range: 5.13–13.03 µg/m3). The estimated mean PM2.5 exposure over 12 months for CRS cohort was 8.12 µg/m3. This is in‐line with PM2.5 concentrations across the United States, which was measured to be 8.1 µg/m3 in all of the United States between 2020 and 2022, as well as similar to prior studies on PM2.5 exposure in humans [25]. Exposures can fluctuate dramatically on a microenvironment scale based on time of year and geospatial location, with the average maximum 1‐h average of 16.7 µg/m3 across the United States [26].
TABLE 1.
Baseline patient demographics, comorbid conditions, and disease‐specific measures for patients with CRSwNP + control cohort by PM2.5 exposure (n = 238; 196 CRSwNP, 42 controls).
| Low PM2.5 exposure (n = 119) | High PM2.5 exposure (n = 119) | |||||
|---|---|---|---|---|---|---|
| Characteristics | N (%) | Mean (SD) | Range | N (%) | Mean (SD) | Range |
| Age at enrollment (years) | — | 49 [±15.02] | (16–81) | — | 47 [±15.65] | (18–78) |
| Male | 64 (54) | — | — | 68 (57) | — | — |
| White/Caucasian | 101 (85) | — | — | 94 (79) | — | — |
| African American | 14 (12) | — | — | 18 (15) | — | — |
| Hispanic/Latino | 0 (0) | — | — | 3 (3) | — | — |
| Asian | 1 (1) | 1 (1) | ||||
| Other | 3 (3) | — | — | 3 (3) | — | — |
| Control patients | 36 (30) | — | — | 6 (5) | — | — |
| Asthma a | 46 (39) | — | — | 63 (53) | — | — |
| Allergic rhinitis b | 64 (54) | — | — | 81 (68) | — | — |
| Body mass index | — | 30.26 [±6.98] | (17.80–54.70) | — | 29.41 [±5.98] | (16.70–53.50) |
| CT score (Lund–Mackay) | — | 12.80 [±8.07] | (0–24) | — | 15.63 [5.81] | (0–24) |
Abbreviations: AERD, aspirin exacerbated respiratory disease; CT, computed tomography. AERD, aspirin exacerbated respiratory disease; CT, computed tomography.
aDiagnosed by pulmonologist.
bConfirmed by skin prick or radioallergosorbent testing.
TABLE 2.
Baseline patient demographics, comorbid conditions, and disease‐specific measures for CRSsNP cohort by PM2.5 exposure (n = 176; 112 CRSsNP, 64 controls).
| Low PM2.5 exposure (n = 88) | High PM2.5 exposure (n = 88) | |||||
|---|---|---|---|---|---|---|
| Characteristics | N (%) | Mean (SD) | Range | N (%) | Mean (SD) | Range |
| Age at enrollment (years) | — | 51 [±15.55] | (18–78) | — | 50 [±15.9] | (19–79) |
| Male | 31 (35) | — | — | 35 (40) | — | — |
| White/Caucasian | 72 (82) | — | — | 70 (80) | — | — |
| African American | 12 (14) | — | — | 13 (15) | — | — |
| Hispanic/Latino | 1 (1) | — | — | 0 (0) | — | — |
| Asian | 1 (1) | 4 (5) | ||||
| Other | 2 (2) | — | — | 1 (1) | — | — |
| Control patients | 36 (41) | — | — | 28 (32) | — | — |
| Asthma a | 14 (16) | — | — | 24 (27) | — | — |
| Allergic rhinitis b | 41 (47) | — | — | 37 (42) | — | — |
| Body mass index | — | 32.35 [±8.87] | (17.9–77.3) | — | 29.5 [±7.30] | (20.6–56.8) |
| CT score (Lund–Mackay) | — | 6.91 [±5.76] | (0–20) | — | 8.88 [±5.63] | (0–22) |
Abbreviations: AERD, aspirin exacerbated respiratory disease; CT, computed tomography. AERD, aspirin exacerbated respiratory disease; CT, computed tomography.
aDiagnosed by pulmonologist.
bConfirmed by skin prick or radioallergosorbent testing.
3.1. Univariate Linear and Multivariate Regression Modeling
There was no statistically significant relationship between PM2.5 and any cytokines in our control cohort under univariate regression, although IL‐10 was just above the pre‐specified alpha threshold on univariate analysis (β = −0.5701, p = 0.0635). Next, we performed multivariate regression to adjust for potential confounders of this association, including age, BMI, diagnosis of asthma or allergic rhinitis, rurality, and income. In this analysis, IL‐10 was significantly reduced with higher PM2.5 exposure (β = −0.735, p = 0.0196). No other cytokine had a significant relationship with estimated PM2.5 levels in control patients. A scatterplot and fitted regression line for the identified inverse IL‐10/PM2.5 relationship are seen in Figure 1. The univariate and multivariate analyses linking PM2.5 exposure and nasal mucus cytokines in our CRS cohort have been previously described by our group [15].
FIGURE 1.
IL‐10 in controls as a function of estimated 12‐month PM2.5. A scatterplot with univariate regression line estimating the relationship between IL‐10 as a function of mean 12‐month PM2.5 estimates.
3.2. Matched Logistic Regression Analysis
For study participants with nasal polyps and healthy controls, a total of 238 patients were included in the matched analysis, with 119 patients each in the high and low PM2.5 exposure groups. Chi‐square testing of the 2 × 2 contingency table comparing the distribution of CRS with PM2.5 exposure showed a statistically higher percentage of CRS patients in the exposed group (94% vs. 69%, p < 0.0001). On subsequent multivariate regression analysis, being in the high 12‐month PM2.5 exposure group was independently associated with an odds ratio of 7.22 (β = 1.97, p = 0.0001) for presenting with nasal polyps. A history of asthma (β = 2.04; OR 7.73, p < 0.0001) and allergic rhinitis (β = 2.16; OR 8.70, p < 0.0001) was similarly associated with high odds of presenting with CRSwNP. Age, AGI, and RUCA1 scores were not statistically significant covariates. The full model with coefficients and confidence intervals is presented in Table 3.
TABLE 3.
Estimated odds ratios for CRSwNP in exposure‐matched logistic regression analysis CRSwNP (n = 196) and controls (n = 42).
| Variable | Odds ratio | 95% Confidence interval | p value |
|---|---|---|---|
| Mean PM2.5 (high) | 7.23 | (2.64–19.81) | 0.00012 |
| AGI | 0.99 | (0.99–1.01) | 0.558 |
| RUCA | 0.90 | (0.76–1.08) | 0.269 |
| Allergic rhinitis | 7.73 | (3.07–19.51) | 0.0001 |
| Asthma | 8.70 | (2.68–28.19) | 0.00031 |
| Age | 1.00 | (0.97–1.03) | 0.925 |
Note: Statistically significant covariates bolded.
Abbreviations: AGI, adjusted gross income; CRSsNP, chronic rhinosinusitis without nasal polyps; RUCA, rural–urban commuting area. AGI, adjusted gross income; CRSsNP, chronic rhinosinusitis without nasal polyps; RUCA, rural–urban commuting area.
A similar analysis method was performed for CRS patients presenting without nasal polyps and healthy controls. For this model, 88 patients were included from the overall cohort in the high and low PM2.5 arms for a total sample of 176 patients. There was a higher percentage of CRSsNP patients in the high exposure group compared to the low exposure group (68% vs. 59%), but the difference was not statistically significant (p = 0.273). On multivariate analysis, high estimated PM2.5 exposure was not associated with increased odds of CRSsNP (β = 0.355, OR 1.42, p = 0.337), but a similar pattern to CRSwNP was seen for allergic rhinitis (β = 1.85; OR 6.41, p < 0.0001) and asthma (β = 1.18; OR 3.26, p = 0.0254). Again, age, AGI, and RUCA1 scores were not statistically significant covariates; the full model is presented in Table 4. Respective power calculations based on the sample size identified for each simulated study showed an 80% power to detect a difference in CRS/control proportions of 16% in the CRSwNP analysis and 19% in the CRSsNP analysis.
TABLE 4.
Estimated odds ratios for CRSsNP in exposure‐matched logistic regression analysis CRSsNP (n = 112) and controls (n = 64).
| Variable | Odds ratio | Confidence interval | p value |
|---|---|---|---|
| Mean PM2.5 (high) | 1.43 | (0.69–2.95) | 0.3376 |
| AGI | 1.00 | (1.00–1.01) | 0.2496 |
| RUCA | 0.87 | (0.73–1.04) | 0.1254 |
| Allergic rhinitis | 6.41 | (2.88–14.26) | 0.0001 |
| Asthma | 3.26 | (1.16–9.21) | 0.0254 |
| Age | 1.01 | (0.99–1.03) | 0.3501 |
Note: Statistically significant covariates bolded.
Abbreviations: AGI, adjusted gross income; CRSsNP, chronic rhinosinusitis without nasal polyps; RUCA, rural–urban commuting area.
4. Discussion
There is increasing evidence for a potential link between chronic air pollutant exposure and sinonasal inflammation, including recent work by our group showing a distinct endotype of CRS associated with PM2.5 exposure that is characterized by a mixed Type 2 dominant cytokine profile and elevated eosinophils [15, 24, 25]. However, as exposure to PM2.5 is ubiquitous in the modern industrialized world, designing experimental studies to determine the longitudinal effect of pollutants on sinus health and disease is challenging. In the present study, we sought to investigate this question by first exploring patterns of nasal mucus cytokine activation in association with estimated PM2.5 exposure in non‐CRS control patients and then conducting an exposure‐matched analysis to simulate a randomized trial of PM2.5 and obtain effect estimates. Our initial results suggest that the patterns of correlation between PM2.5 exposure and cytokine levels noted in CRS patients [15] may not extend to healthy controls, with no statistically significant correlation between any cytokine and mean PM2.5 exposure beyond 6 months on univariate analyses. Particulate matter estimates, however, are highly correlated with socioeconomic variables, so isolating the potential impact of PM2.5 requires adjustment for potential confounders. Indeed, after controlling for BMI, age, AGI, rurality, and a history of asthma or allergic rhinitis, we found that IL‐10 was significantly reduced at higher PM2.5 exposure levels.
Although this study is unable to make causal statements among cytokine levels, PM2.5 exposure, and CRS diagnosis, the distinct association of IL‐10 reduction with higher mean PM2.5 exposure in control, non‐CRS patients identifies a possible molecular pathway to further examine. A pleiotropic cytokine and crucial mediator of both innate and adaptive immunity, IL‐10 primarily serves as an anti‐inflammatory immune modulator, inhibiting the production of pro‐inflammatory cytokines, including IL‐1b, IL‐4, IL‐5, IL‐6, IL‐12, IL‐18, and TNF‐a [27, 28]. Although the relationship between IL‐10 and particulate matter exposure had not been previously examined in CRS patients until recently, its role in lower airway pathology has been well studied. For example, systemic IL‐10 production has been found to be reduced in asthmatic patients [29]. In animal models of PM2.5 exposure, concurrent administration of recombinant IL‐10 in rats exposed to PM2.5 improved acute lung injury, decreased mitochondrial damage, and reduced inflammation and apoptosis, suggesting that IL‐10 may have a therapeutic benefit in reducing lower airway damage from particulate matter [30].
Observational human studies of CRS and IL‐10 have shown conflicting evidence for IL‐10 reduction associated with CRSwNP specifically; some studies have reported reduced serum and nasal tissue levels of IL‐10 compared to that of healthy controls [31, 32], whereas another study found no difference in IL‐10 in nasal polyp tissue and turbinate tissue in control patients [33]. Our prior investigation of human CRS patients did not demonstrate any significant association between PM2.5 exposure and nasal mucus IL‐10 levels, highlighting a potentially unique cytokine inflammation pattern specific to non‐CRS patients exposed to increased levels of ambient pollution [15]. In addition, it is important to note IL‐10's highly multifaceted and context‐dependent role in immunologic homeostasis and hyperinflammation reduction. For example, one in vitro study of human mononuclear cells showed that when exposed to diesel exhaust particles (DEP) directly, IL‐10 transcription and protein production increased, whereas if those cells were first primed with lipopolysaccharides (LPS), a known pro‐inflammatory agent, the same DEP exposure resulted in a significant reduction in IL‐10 production and mRNA transcription [34]. Another study found that reduced levels of IL‐10 are associated with treatment‐resistant disease [16]. The findings from this study suggest the need for further in‐depth investigation into the role of ambient particulate exposure on IL‐10 expression and production on a cellular level and its potential downstream impact on CRS pathogenesis.
Our second objective for this study was to estimate the effect of chronic PM2.5 on the development of CRS utilizing novel exposure‐matching approaches to account for the non‐random nature of our observational sample. Several prior studies have used traditional matched case–control studies to investigate this question, with findings that corroborate our result showing an odds ratio of 7.22 (β = 1.97, p = 0.0001) for developing CRSwNP with higher than average PM2.5 exposure. For example, Zhang et al. demonstrated long‐term exposure to PM2.5 increases the odds of CRS diagnosis in over 6000 subjects [24]. Additionally, a recent case–control study in the military population demonstrated a 5.99 OR of CRS development following PM2.5 exposure, though interestingly, they noted a 7.19 OR of CRSsNP development and only 3.98 OR of CRSwNP development when adjusting for age, gender, and diagnosis year [25]. In patients with a CRSwNP diagnosis, short‐term 30‐day air pollution, including PM2.5, PM10, SO2, NO2, CO, and O3, was associated with worsened Lund–Mackay scores, increased VAS headache/facial pain scores, and increased the risk of eosinophilic CRSwNP endotype [35]. In our study, the estimated odds of high PM2.5 exposure on the risk of CRSwNP were comparable to the risk associated with a history of asthma (β = 2.04; OR 7.73, p < 0.0001) and allergic rhinitis (β = 2.16; OR 8.70, p < 0.0001), providing a relative clinical context for understanding the effect size. A similar relationship was not seen in CRSsNP, which is consistent with our prior finding of high eosinophil counts and a mixed, Type 2 dominant cytokine profile with increasing exposure.
Although this study adds to the growing body of literature illustrating a clear relationship between particulate matter exposure and CRS development, there are several limitations to this study that are important to address. First, our CRS and control patient samples were recruited at one tertiary academic medical center in the southeastern United States. Given the heterogeneous regional variation in CRS phenotypes and endotypes [36, 37, 38], the results may not be as generalizable to patients in other regions or practice settings. Second, our relatively small control group of 68 patients limits the power of our statistical analyses and the conclusions we can draw from them for cytokine analyses. Furthermore, the calculation of odds ratios in this study is intended to serve as a statistical tool for contextualizing CRS development and particulate matter exposure across a non‐random group of CRS and control participants, with limited firm epidemiological conclusions. These limitations can be mitigated by future multi‐center collaborative research across more diverse geographic regions and ethnic populations. Third, the CRS patients used in the matched analysis were those undergoing surgical management of their CRS and had already failed medical management, introducing a sampling bias with a more severe, medically recalcitrant cohort of CRS patients. Fourth, the accuracy of our PM2.5 exposure estimations is inherently linked to the fidelity of the machine‐learning algorithm published by Brokamp et al. [16]. This model has been cross‐validated across different geographic regions in the United States, has shown to be superior or similar in accuracy to pre‐existing models, and was successfully implemented in our prior study of CRS patients [15]. However, we utilized patient home zip codes for PM2.5 estimations and do not take into account occupation‐related changes in environment, commuting distances and modes of transportation, and other individualistic location changes. Additional studies with direct measurements of PM exposure over a given period using wearables or means may improve the accuracy of the exposure data acquired in these studies. In addition, increased study participants may help us to understand if the associations discussed in this study hold true for larger populations. Ultimately, despite the use of causal inference methods for effect estimation, the findings from this study should be considered exploratory until replicated in a larger, geographically and racially diverse population, especially considering the wide confidence intervals for all significant effect estimates.
In conclusion, this study provides additional evidence that PM2.5 is an independent risk factor for the development of CRSwNP, with estimated odds ratios comparable to the impact of comorbid asthma or allergic rhinitis. The identified negative association among IL‐10, a classic anti‐inflammatory cytokine, and PM2.5 exposure levels in control patients suggests a plausible mechanism through which chronic PM2.5 exposure could increase susceptibility to chronic nasal mucosal inflammation and CRS, though additional experimental evidence from in vitro and longitudinal multi‐omics studies is needed to confirm this hypothesis.
Conflicts of Interest
The authors declare no conflicts of interest.
Lubner R. J., Krysinski M., Li P., Chandra R. K., Turner J. H., and Chowdhury N. I., “Long‐Term Particulate Matter Exposure May Increase Risk of Chronic Rhinosinusitis WIth Nasal Polyposis: Results From an Exposure‐Matched Study.” International Forum of Allergy & Rhinology 15, no. 9 (2025): 15, 926–933. 10.1002/alr.23589
Funding: This study was supported by NIH (Grant R01AG065550) to Justin H. Turner, Burroughs Wellcome Fund (PSIA grant) to Naweed I. Chowdhury, and ARS (CORE grant) to Rory J. Lubner.
References
- 1. M. B. Soyka,, Wawrzyniak P., and Eiwegger T., “Defective Epithelial Barrier in Chronic Rhinosinusitis: The Regulation of Tight Junctions by IFN‐γ and IL‐4,” Journal of Allergy and Clinical Immunology 130, no. 5 (2012): 1087–1096.e10. [DOI] [PubMed] [Google Scholar]
- 2. Gudis D., Zhao K. Q., and Cohen N. A., “Acquired Cilia Dysfunction in Chronic Rhinosinusitis,” American Journal of Rhinology & Allergy 26, no. 1 (2012): 1–6. [Google Scholar]
- 3. Kato A., Schleimer R. P., and Bleier B. S., “Mechanisms and Pathogenesis of Chronic Rhinosinusitis,” Journal of Allergy and Clinical Immunology 149, no. 5 (2022): 1491–1503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Orlandi R. R., Kingdom T. T., Smith T. L., et al., “International Consensus Statement on Allergy and Rhinology: Rhinosinusitis 2021,” International Forum of Allergy & Rhinology 11, no. 3 (2021): 213–739. [DOI] [PubMed] [Google Scholar]
- 5. Fokkens W. J., Lund V. J., Hopkins C., et al., “European Position Paper on Rhinosinusitis and Nasal Polyps 2020,” Rhinology 58, no. S29 (2020): 1–464. [Google Scholar]
- 6. Staudacher A. G., Peters A. T., Kato A., and Stevens W. W., “Use of Endotypes, Phenotypes, and Inflammatory Markers to Guide Treatment Decisions in Chronic Rhinosinusitis,” Annals of Allergy, Asthma & Immunology 124, no. 4 (2020): 318–325. [Google Scholar]
- 7. Leland E. M., Zhang Z., Kelly K. M., and Ramanathan M., “Role of Environmental Air Pollution in Chronic Rhinosinusitis,” Current Allergy and Asthma Reports 21, no. 8 (2021): 42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Wang X., Zhang N., Bo M., et al., “Diversity of T(H) Cytokine Profiles in Patients with Chronic Rhinosinusitis: A Multicenter Study in Europe, Asia, and Oceania,” Journal of Allergy and Clinical Immunology 138, no. 5 (2016): 1344–1353. [DOI] [PubMed] [Google Scholar]
- 9. Stevens W. W., Peters A. T., Tan B. K., et al., “Associations between Inflammatory Endotypes and Clinical Presentations in Chronic Rhinosinusitis,” Journal of Allergy and Clinical Immunology: In Practice 7, no. 8 (2019): 2812–2820.e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Wu W., Jin Y., and Carlsten C., “Inflammatory Health Effects of Indoor and Outdoor Particulate Matter,” Journal of Allergy and Clinical Immunology 141, no. 3 (2018): 833–844. [DOI] [PubMed] [Google Scholar]
- 11. Ma Q. Y., Huang D. Y., Zhang H. J., Wang S., and Chen X. F., “Exposure to Particulate Matter 2.5 (PM2.5) Induced Macrophage‐Dependent Inflammation, Characterized by Increased Th1/Th17 Cytokine Secretion and Cytotoxicity,” International Immunopharmacology 50 (2017): 139–145. [DOI] [PubMed] [Google Scholar]
- 12. Ryou H. G., Heo J., and Kim S. Y., “Source Apportionment of PM(10) and PM(2.5) Air Pollution, and Possible Impacts of Study Characteristics in South Korea,” Environmental Pollution 240 (2018): 963–972. [DOI] [PubMed] [Google Scholar]
- 13. Ramanathan M. Jr., London N. R. , Tharakan A., et al., “Airborne Particulate Matter Induces Nonallergic Eosinophilic Sinonasal Inflammation in Mice,” American Journal of Respiratory Cell and Molecular Biology 57, no. 1 (2017): 59–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Zhao R., Guo Z., Dong W., et al., “Effects of PM2.5 on Mucus Secretion and Tissue Remodeling in a Rabbit Model of Chronic Rhinosinusitis,” International Forum of Allergy & Rhinology 8, no. 11 (2018): 1349–1355. [DOI] [PubMed] [Google Scholar]
- 15. Lubner R. J., Rubel K., Chandra R. K., Turner J. H., and Chowdhury N. I., “Particulate Matter Exposure Is Associated with Increased Inflammatory Cytokines and Eosinophils in Chronic Rhinosinusitis,” Allergy 79 (2024): 1219–1229. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Longino E. S., Labby A. B., Wu J., et al., “Association of Cytokine Profile with Prior Treatment Failure and Revision Surgery in Chronic Rhinosinusitis,” International Forum of Allergy & Rhinology 13 (2023): 5–14. [DOI] [PubMed] [Google Scholar]
- 17. Turner J. H., Li P., and Chandra R. K., “Mucus T Helper 2 Biomarkers Predict Chronic Rhinosinusitis Disease Severity and Prior Surgical Intervention,” International Forum of Allergy & Rhinology 8, no. 10 (2018): 1175–1183. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Morse J. C., Shilts M. H., Ely K. A., et al., “Patterns of Olfactory Dysfunction in Chronic Rhinosinusitis Identified by Hierarchical Cluster Analysis and Machine Learning Algorithms,” International Forum of Allergy & Rhinology 9, no. 3 (2019): 255–264. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Ho D., Imai K., King G., and Stuart E. A., “MatchIt: Nonparametric Preprocessing for Parametric Causal Inference,” Journal of Statistical Software 42, no. 8 (2011): 1–28. [Google Scholar]
- 20. Brokamp C., “A High Resolution Spatiotemporal Fine Particulate Matter Exposure Assessment Model for the Contiguous United States,” Environmental Advances 7 (2022): 100155. [Google Scholar]
- 21. Stuart E. A., “Matching Methods for Causal Inference: A Review and a Look Forward,” Statistical Science 25, no. 1 (2010): 1–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Chen H., Kaufman J. S., Olaniyan T., et al., “Changes in Exposure to Ambient Fine Particulate Matter after Relocating and Long Term Survival in Canada: Quasi‐Experimental Study,” Bmj 375 (2021): n2368. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Greifer N. and Stuart E. A., “Matching Methods for Confounder Adjustment: An Addition to the Epidemiologist's Toolbox,” Epidemiologic Reviews 43, no. 1 (2021): 118–129. [Google Scholar]
- 24. Zhang Z., Kamil R. J., London N. R., et al., “Long‐Term Exposure to Particulate Matter Air Pollution and Chronic Rhinosinusitis in Nonallergic Patients,” American Journal of Respiratory and Critical Care Medicine 204, no. 7 (2021): 859–862. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Elam T., Raiculescu S., Biswal S., Zhang Z., Orestes M., and Ramanathan M., “Air Pollution Exposure and the Development of Chronic Rhinosinusitis in the Active Duty Population,” Military Medicine 188 (2022): e1965–e1969. [Google Scholar]
- 26. Agency E. P., Overview of Particulate Matter (PM) Air Quality in the United States. In: EPA, ed. (2023).
- 27. Ouyang W. and O'Garra A., “IL‐10 Family Cytokines IL‐10 and IL‐22: From Basic Science to Clinical Translation,” Immunity 50, no. 4 (2019): 871–891. [DOI] [PubMed] [Google Scholar]
- 28. Jutel M., Akdis M., Blaser K., and Akdis C. A., “Mechanisms of Allergen Specific Immunotherapy–T‐Cell Tolerance and More,” Allergy 61, no. 7 (2006): 796–807. [DOI] [PubMed] [Google Scholar]
- 29. Ogawa Y., Duru E. A., and Ameredes B. T., “Role of IL‐10 in the Resolution of Airway Inflammation,” Current Molecular Medicine 8, no. 5 (2008): 437–445. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Zhang N., Li P., Lin H., et al., “IL‐10 Ameliorates PM2.5‐Induced Lung Injury by Activating the AMPK/SIRT1/PGC‐1α Pathway,” Environmental Toxicology and Pharmacology 86 (2021): 103659. [DOI] [PubMed] [Google Scholar]
- 31. Wu D., Wang J., and Zhang M., “Altered Th17/Treg Ratio in Nasal Polyps with Distinct Cytokine Profile: Association with Patterns of Inflammation and Mucosal Remodeling,” Medicine 95, no. 10 (2016): e2998. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Lucas B. R., Voegels R. L., do Amaral J. B., Bachi A. L. L., and Pezato R., “BMP‐7, MMP‐9, and TGF‐β Tissue Remodeling Proteins and Their Correlations with Interleukins 6 and 10 in Chronic Rhinosinusitis,” European Archives of Oto‐Rhino‐Laryngology 278, no. 11 (2021): 4335–4343. [DOI] [PubMed] [Google Scholar]
- 33. Bachert C., Wagenmann M., Hauser U., and Rudack C., “IL‐5 Synthesis Is Upregulated in Human Nasal Polyp Tissue,” Journal of Allergy and Clinical Immunology 99, no. 6 Pt 1 (1997): 837–842. [DOI] [PubMed] [Google Scholar]
- 34. Pacheco K. A., Tarkowski M., Sterritt C., Negri J., Rosenwasser L. J., and Borish L., “The Influence of Diesel Exhaust Particles on Mononuclear Phagocytic Cell‐Derived Cytokines: IL‐10, TGF‐Beta and IL‐1 Beta,” Clinical and Experimental Immunology 126, no. 3 (2001): 374–383. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Yang X., Shen S., Deng Y., Wang C., and Zhang L., “Air Pollution Exposure Affects Severity and Cellular Endotype of Chronic Rhinosinusitis with Nasal Polyps,” Laryngoscope 132, no. 11 (2022): 2103–2110. [DOI] [PubMed] [Google Scholar]
- 36. Turner J. H., Chandra R. K., Li P., Bonnet K., and Schlundt D. G., “Identification of Clinically Relevant Chronic Rhinosinusitis Endotypes Using Cluster Analysis of Mucus Cytokines,” Journal of Allergy and Clinical Immunology 141, no. 5 (2018): 1895–1897.e7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Tomassen P., Vandeplas G., Van Zele T., et al., “Inflammatory Endotypes of Chronic Rhinosinusitis Based on Cluster Analysis of Biomarkers,” Journal of Allergy and Clinical Immunology 137, no. 5 (2016): 1449–1456.e4. [DOI] [PubMed] [Google Scholar]
- 38. Divekar R., Rank M., Squillace D., Kita H., and Lal D., “Unsupervised Network Mapping of Commercially Available Immunoassay Yields Three Distinct Chronic Rhinosinusitis Endotypes,” International Forum of Allergy & Rhinology 7, no. 4 (2017): 373–379. [DOI] [PubMed] [Google Scholar]