Chronic Rhinosinusitis in Elderly Patients is Associated with an Exaggerated Neutrophilic Pro-Inflammatory Response to Pathogenic Bacteria

Abstract

Background:

Potential effects of aging on chronic rhinosinusitis (CRS) pathophysiology have not been well defined, but may have important ramifications given a rapidly aging U.S. and world population.

Objective:

The goal of the current study was to determine whether advanced age is associated with specific inflammatory CRS endotypes or immune signatures.

Methods:

Seventeen mucus cytokines and inflammatory mediators were measured in 147 CRS patients. Hierarchical cluster analysis was used to identify and characterize inflammatory CRS endotypes as well as determine whether age was associated with specific immune signatures.

Results:

A CRS endotype with a pro-inflammatory, neutrophilic immune signature was enriched with older patients. In the overall cohort, patients 60 years and older had elevated mucus levels of IL-1β, IL-6, IL-8, and TNF-α when compared to their younger counterparts. Increases in pro-inflammatory cytokines were associated with both tissue neutrophilia and symptomatic bacterial infection/colonization in aged patients.

Conclusions:

Aged CRS patients have a unique inflammatory signature that corresponds to a neutrophilic pro-inflammatory response. Neutrophil-driven inflammation in aged CRS patients may be less likely to respond to corticosteroids and may be closely linked with chronic microbial infection or colonization.

CAPSULE SUMMARY:

The immune signature in elderly patients with CRS suggests a distinct pathophysiology marked by prevalence of bacterial colonization and an exaggerated but apparently ineffective neutrophilic proinflammatory response to microbial pathogens.

Graphical Abstract

INTRODUCTION

Chronic rhinosinusitis (CRS) is a heterogeneous inflammatory airway disease with poorly defined pathophysiology that affects between 4–12% of the U.S. population^1–3. Though generally more prevalent in patients with comorbid asthma and environmental allergies, CRS can affect individuals of any age group and is now recognized as one of the most economically burdensome physical health conditions in the U.S.⁴. The impact of CRS in the geriatric population is poorly understood; however, recent epidemiologic and cohort studies suggest that the disease may be more prevalent in older patients. For example, a Korean study found that the prevalence of CRS in patients aged 60 or older was almost twice that of patients under the age of 40⁵. Likewise, North American studies evaluating patients who presented with a diagnosis of CRS have suggested that advanced age may be associated with higher rates of nasal polyposis and comorbid asthma^{6, 7}. The U.S. population, consistent with demographic changes in other industrialized nations, is rapidly aging. Current projections estimate that individuals aged 65 years or older will represent 20% of the U.S population by 2050⁸. This population shift will likely burden the overall health care economy and drive increases in costs to delivering care, making it critical to understand the effects of age on chronic diseases.

Limited evidence suggests that CRS in the elderly may represent a distinct clinical entity with unique pathophysiology. Cho et al. reported that levels of S100A8/9, which are proteins important to epithelial barrier function, are reduced in elderly patients^{6, 7}. Aging may also be associated with reductions in sinonasal ciliary beat frequency and mucociliary clearance^9–13. General effects of aging on immune system function could also impact the presentation and evolution of CRS in the elderly. Gradual age-associated deterioration of the immune system, known as immunosenescence, can result in reduced phagocytosis of pathogenic bacteria and impaired innate defense mechanisms^14–17. While understudied in CRS, the impact of aging on immune function has been more extensively investigated in asthma. Early studies found that elderly asthmatics had more severe airway obstruction than their younger counterparts¹⁸. Older patients were also found to have marked sputum neutrophilia that seemed to suggest an age-related neutrophilic asthma phenotype^{18, 19}. A more recent prospective study found that asthma patients over the age of 60 had increased sputum levels of IL-1β and IL-6²⁰. This inflammatory pattern was associated with airway neutrophil recruitment and had an adverse impact on asthma control. Collectively, these studies suggest that a pro-inflammatory neutrophilic milieu may be unique to elderly asthmatics, with significant implications for disease management and treatment. Similar age-dependent factors may be equally important in CRS given the shared immune mechanisms and pathophysiology common to both upper and lower airway inflammatory disease.

There is an urgent need to develop more effective treatment options for CRS in the elderly population. While the mainstays of medical therapy for CRS include antibiotics and systemic corticosteroids, both of these options are associated with increased risks of adverse complications in elderly patients^21–25. Several studies also suggest that advanced age may be a risk factor for serious complications following sinus surgery^{26, 27}. We hypothesized that CRS in elderly patients may be associated with a unique inflammatory signature and distinct pathophysiology. The purpose of the current study was to identify characteristics distinct to aged CRS patients, with the potential to change treatment approaches for this at-risk population.

MATERIALS AND METHODS

Study Design and Population

This study was approved by the Vanderbilt University Institutional Review Board. Patients presented to the Vanderbilt Asthma, Sinus, and Allergy Program (ASAP) and Otolaryngology clinic at the Vanderbilt Bill Wilkerson Center. CRS was diagnosed according to the European Position Paper on Rhinosinusitis and Nasal Polyps and the International Consensus Statement on Allergy and Rhinology and therefore were initially managed medically ^{28, 29}. CRS patients offered surgery had previously failed adequate medical therapy that included two or more weeks of oral prednisone, two or more weeks of broad-spectrum or culture-directed antibiotics, and a combination of other additional therapies including oral/topical antihistamines, topical nasal steroid sprays, oral decongestants, mucolytics, or saline rinses. Patients with continued symptoms who elected to undergo endoscopic sinus surgery were prospectively enrolled and signed informed consent. Only patients with diffuse, bilateral inflammatory CRS were included, and patients with odontogenic rhinosinusitis, fungus balls, and isolated osteomeatal complex obstruction were excluded. Control cases included patients undergoing pituitary or skull base surgery without a clinical or radiographic history of CRS. Patients were excluded if they had received systemic steroids within 4 weeks of surgery. Patients with cystic fibrosis, autoimmune, or granulomatous diseases or who were receiving immune-directed monoclonal antibodies were excluded. The presence of concomitant allergic rhinitis and asthma was recorded. Allergic rhinitis was diagnosed based on positive skin prick testing and/or prior physician diagnosis and clinical history suggestive of seasonal variation of atopic symptoms with improvement following use of topical nasal steroid or oral antihistamines. Asthma was diagnosed based on a positive methacholine challenge or consistent pulmonary function studies, or by prior diagnosis by a pulmonologist. Allergic fungal rhinosinusitis (AFRS) was diagnosed according to published criteria ³⁰, including presence of fungal elements and allergic mucin on pathology, characteristic findings on CT imaging, and concomitant positive allergy testing to fungal allergen(s). Aspirin exacerbated respiratory disease (AERD) was diagnosed based on presence of asthma and nasal polyposis, as well as a prior history of positive aspirin challenge or at least two episodes of respiratory reaction to aspirin or non-steroidal anti-inflammatory drugs. Patient reported symptom severity was measured utilizing the Sinonasal Outcome Test-22 (SNOT-22) ³¹. All patients underwent a high resolution CT scan of the paranasal sinuses within 3 months of surgery. Each scan was evaluated by two physicians who were blinded to subject identifiers and diagnosis. A standard Lund Mackay scoring system was used to assess overall extent of CRS. Subjects enrolled in the study also completed the 40item Smell Identification Test (SIT) immediately prior to surgery. The SIT has excellent sensitivity, correlates closely with scores attained via formal threshold testing, and has the advantage of being easily and quickly administered to subjects on the day of surgical intervention ³². Raw scores were adjusted for patient age and gender by subtracting the mean normative age- and sex-appropriate SIT score from the total SIT score for each subject ³³. Thus, a negative adjusted SIT score represents reduced sense of smell compared to the mean for that subject’s age and gender. Normative SIT scores were extracted from the Smell Identification Test Administration Manual (Sensonics International; Haddon Heights, NJ). Bacterial cultures were taken when mucopurulence was present. A standardized methodology using a mini-tip calcium alginate swab was used in order to reduce cross-contamination. Aerobic cultures were transported in a BD culture swab collection and transport system (BD Biosciences; San Jose, CA). After collection, cultures were immediately taken to the Vanderbilt University Medical Center microbiology laboratory for microbiological speciation.

Mucus Collection and Histopathologic Evaluation of Sinonasal Tissue

At the beginning of surgery, 9 × 24mm polyurethane sponges (Summit Medical; St. Paul, MN) were placed into the middle meatus or ethmoid cavity of each subject under endoscopic guidance as previously reported ³⁴. This approach has advantages over other methods for mucus collection, including standardization between subjects and avoidance of specimen dilution. Each sponge was removed after 5 minutes, placed in a sterile microcentrifuge tube and immediately processed. Sponges were placed into a microporous centrifugal filter device (MilliporeSigma; Billerica, MA) and centrifuged at 14,000 × g for 10 minutes to elute mucus. Samples were then gently vortexed and again centrifuged for 5 minutes to remove any cellular debris. Supernatants were removed, placed into a new microcentrifuge tube, and frozen at −80°C for later analysis.

Cytokine assays were performed using a multiplex cytokine bead assay (BD Biosciences; Franklin Lakes, NJ) according to the manufacturer’s protocol. Briefly, 50 μL of mucus was incubated with 50 μl of mixed capture beads for each measured inflammatory mediator and incubated for 1 hour. 50 μL of mixed detection reagent was then added to each sample and standard, and incubated for an additional 2 hours. After addition of 1 mL wash buffer, samples were centrifuged at 200 × g for 5 minutes and the supernatant was discarded. The beads were then resuspended in 300 μL wash buffer and analyzed on an LSR Fortessa flow cytometer (BD Biosciences; San Jose, CA). A total of 17 cytokines and inflammatory mediators were analyzed, excluding only mediators with overlapping detection wavelengths and most growth factors and chemokines. Data was analyzed using BD FCAP Array Software version 3.0.

Sinonasal tissue was collected from the ethmoid bulla or ethmoid sinus in all patients undergoing endoscopic sinus surgery for CRS. Tissue from healthy controls was collected from either the ethmoid sinus or sphenoethmoid recess. Histopathological evaluation of excised tissue was performed by a head and neck pathologist in a blinded fashion, and using a structured approach^{35, 36}. This standardized evaluation included both quantitative measurement of tissue eosinophils and neutrophils, and assessment of observable features associated with tissue architecture, including edema, papillary changes, squamous metaplasia, and fibrosis.

Statistics

Principal component analysis and hierarchical clustering was performed as previously reported, with some minor variations related to data normalization and using additional approaches for accurately estimating the number of clusters and verifying cluster stability³⁷. Briefly, sample size for principal component analysis and subsequent clustering was estimated by establishing a subject to variable ratio of greater than 5:1 as recommended by Gorsuch and Hatcher ^{38, 39}. Adequacy of the sample size was verified post hoc by assessing variable communality (heavy loading of variables in retained components). Descriptive statistics and frequency distributions were examined for each biological variable and all were positively skewed. In order to normalize data for subsequent analysis, values were transformed by taking the square root, resulting in elimination or significant reduction of skewing for all variables. A principal component factor analysis with varimax rotation was then performed on the transformed biological variables. The appropriate number of factors was selected by analysis of the Scree plot, with a requirement that retained factors explain at least 70% of data variance, and that each factor have an eigenvalue > 1.0. The regression method was then used to calculate a factor score for each subject in each of the five factors. Hierarchical cluster analysis was performed using Ward’s method on squared Euclidian distances using the five factor scores. Ward’s method determines which clusters can be merged by comparing dissimilarity between subjects and by minimizing total within-cluster variance⁴⁰. This method has the ability to reduce a complex dataset with multiple subjects and variables into a more manageable format, and finds natural groupings of subjects based on those input variables. The hierarchical structure and taxonomic relationships between subjects was visualized using a dendogram. The appropriate number of clusters (k) was selected using the Elbow method. This approach calculates the total within sum of squared error (SSE) for between 1 and 10 clusters and determines k by identifying the break point where adding additional clusters does not substantially change the SSE. Cluster stability was verified using bootstrap analysis. This involved randomly resampling the data 1000 times to ensure that the clustering results were stable and not unique to the original dataset. All clusters had a stability of between 0.7 and 0.9 (values closer to 1 indicate stable clusters).

Differences between clusters were then assessed for demographic and clinical data, and for the individual biological variables themselves. For comparison between groups, normality of data was assessed using the D’Agostino-Pearson omnibus test. Variables with a normal distribution were compared using a student’s t-test or analysis of variance, while nonparametric data was analyzed using the Mann-Whitney test or Kruskal-Wallis test followed by Dunn’s test for multiple comparisons. Comparative data was presented as means +/− standard deviation or medians with interquartile range, respectively. A p value less than 0.05 was considered statistically significant for all comparisons.

Age-specific effects of cytokines on SNOT-22 scores and frequency of revision surgery were assessed using linear regression models. Continuous inflammatory marker variables for each patient were classified into dichotomous “high” or “low” variables based on whether the measured level was higher or lower than the median value of the overall group. To assess the relationship between age, inflammatory markers, and continuous outcomes of interest, individual multivariate linear regression models were fitted for each marker, including an interaction term for age. A significant interaction term in this model implies that the effect of the cell type or cytokine on the outcome is dependent on age category. For categorical outcomes (revision surgery), a similar approach was used with multiple, individual logistic regression models fitted for each inflammatory marker of interest, again with incorporation of an interaction term for age. In the context of logistic regression, a significant interaction term implies that the effect of the inflammatory marker on the odds ratio of the outcome varies by age category. In both approaches, hypothesis testing between groups was conducted using t-tests of the regression coefficients. Overall model significance was computed using the F-test for overall significance. For all hypothesis tests, the alpha-level was set at p < 0.05. All data management and analysis was conducted using R version 3.4 (R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org/) and Prism 6 software (Graphpad; La Jolla, CA).

RESULTS

Study Population and Demographics

A total of 147 CRS patients undergoing endoscopic sinus surgery and 30 non-CRS controls were enrolled in this ongoing prospective study (Table E1), with a subset of patients having being previously characterized elsewhere^{37, 41}. Among all CRS patients, 59% had nasal polyps, 65% had allergic rhinitis, and 39% had comorbid asthma. Overall disease severity was significant, with a mean SNOT-22 score of 45.3 and a median CT score of 15.0. Thirty-eight percent of patients had undergone prior endoscopic sinus surgery.

Validation of CRS Endotypes and Role of Age in Disease Clustering

In a recent pilot study we characterized inflammatory CRS endotypes by performing hierarchical cluster analysis, and using mucus cytokine levels as input biological variables³⁷. Clusters had distinct characteristics, including variability in polyp status, asthma prevalence, and prior surgery. Conversely, most demographic factors, including age and sex, did not vary between disease clusters. We did, however, observe that one CRS cluster was enriched in older patients, leading us to suspect that age may affect inflammatory signatures, and perhaps be associated with CRS pathophysiology. A small number of previous studies have demonstrated age-dependent differences in CRS prevalence or phenotype^5–7. In the process of validating clusters identified in our initial study, we repeated disease clustering in an updated cohort of 147 patients. Similar to our pilot study, cluster analysis again suggested either 5 or 6 inflammatory CRS endotypes, and we settled on a 5 cluster model after statistical modeling and assessment of cluster stabilities (Table E2). Age was significantly different between clusters (p=0.04), particularly among patients of advanced age, herein defined as those 60 years of age or older (p=0.009). Hierarchical cluster analysis of our 147 patient cohort also showed that patients of advanced aged were concentrated within specific clusters (Figure 1). Most noticeably, cluster 2 which was most strongly associated with elevations in the pro-inflammatory cytokines IL-1β, IL-6, IL-8, and TNF-α (Table E2), was composed almost entirely of patients aged 60 or older.

Figure 1. Dendogram representing hierarchical cluster analysis of CRS patients and relationship with patient age.

Hierarchical cluster analysis was performed using Ward’s method on squared Euclidian distances using 17 cytokines and inflammatory mediators as biological variables. Age is recorded in the right panel for individual study subjects in the adjacent dendogram and as a continuous mean.

Elevation of Pro-Inflammatory Cytokines in Aged CRS Patients

We sought to further explore potential relationships between age and CRS inflammatory signatures by comparing patients of advanced age with their younger counterparts (Table 1). We separated the study population into younger (< 60 years) and aged (≥60 years) patients, largely for the purpose of maintaining consistency with prior studies^{6, 7, 18, 19, 42, 43}, which have likewise differentiated patients with this cut-off. Aged and younger patients were demographically similar in all assessed variables, excepting allergic rhinitis, which was more common in younger patients (p < 0.05). In contrast to a prior study that reported increased frequency of nasal polyps and asthma in aged patients ^{6, 7}, this observation was not present in our cohort. Instead, aged and younger patients had a remarkably similar frequency of both nasal polyps (51% vs. 62%, p = 0.33) and asthma (38% vs. 39%, p = 1.00). We likewise did not observe any significant difference in measures of disease severity, including SNOT-22 or CT scores. There was a moderately increased likelihood of prior endoscopic sinus surgery in aged patients, but this difference did not reach statistical significance (p =0.17).

Table 1. Differences in demographic and clinical characteristics between aged and younger CRS patients.

Values are presented as either the mean +/− standard deviation or median with interquartile range, depending on the normalcy of the data. Differences between groups were assessed using the Kruskal-Wallis test or Chi-square analysis. BOLD, p < 0.05. AERD, aspirin-exacerbated respiratory disease; AFRS, allergic fungal rhinosinusitis; BMI, body mass index; CT, computed tomography; NCS, nasal corticosteroid; LTR, anti-leukotriene; SIT, smell identification test; SNOT-22, sinonasal outcome test-22.

	Younger (18–59)	Aged (>60)	P value
No.	110	37
Age (years)	43.3 +/− 10.2	66.4 +/− 4.7	<0.0001
Sex, no. (% female)	51 (46)	16 (43)	0.57
Race, no. (% white)	93 (85)	34 (92)	0.41
Current smoker, no. (%)	6 (6)	0 (0)	0.34
BMI (kg/m2)	29.64 +/− 8.1	31.2 +/− 5.6	0.11
Nasal polyps, no. (%)	68 (62)	19 (51)	0.33
Asthma, no. (%)	43 (39)	14 (38)	1.00
Allergic Rhinitis, no. (%)	77 (70)	19 (51)	<0.05
AERD, no. (%)	11 (10)	3 (8)	1.00
AFRS, no. (%)	15 (14)	3 (8)	0.56
Taking NCS, no. (%)	90 (82)	29 (78)	0.64
Taking LTR, no. (%)	28 (25)	10 (27)	0.83
SNOT-22 score	46.3 +/− 19.6	42.2 +/− 18.4	0.39
CT score	15.8 (11.0–20.0)	14.0 (11.0–18.8)	0.28
SIT score	−7.0 (−23.3–−2.0)	−8.0 (−22.5–−3.0)	0.92
Prior surgery, no. (%)	38 (35)	18 (49)	0.17
# of prior surgeries (no.)	0.6 +/− 0.10	0.78 +/− 0.18	0.18

We next explored whether mucus cytokine levels varied between younger and aged patients. We observed a 4- to 6-fold increase in pro-inflammatory cytokines in aged vs. younger CRS patients (Table 2, Figure 2), including IL-1β (p < 0.0001), IL-6 (p < 0.0001), IL-8 (p < 0.0001), and TNF-α (P < 0.0001). This observation was maintained when patients were separately evaluated by polyp status (data not shown). Aging has been associated with increased circulating levels of pro-inflammatory cytokines, a finding commonly referred to as “inflamm-aging”^{44, 45}. We therefore sought to verify the pathophysiological relevance and specificity of the differences identified in our study by comparing aged CRS patients to age-matched healthy controls (Table 2). CRS in aged patients was associated with elevated levels of IL-1β (p=0.0008), IL-6 (p=0.01), IL-8 (p=0.006), and TNF-α (p=0.009) compared to age-matched healthy control patients. No significant differences in any measured cytokines were identified in aged control patients when compared to younger control patients, suggesting that the elevation of pro-inflammatory cytokines in CRS is indicative of disease pathophysiology and not to a general elevation of circulating pro-inflammatory cytokines in older patients. To further characterize the association of age and pro-inflammatory mediators in CRS patients we performed Spearman correlation for each cytokine with age as a continuous variable (Figure 3). Advancing age was strongly correlated with mucus levels of IL-1β (p<0.0001), IL-6 (p=0.003), IL-8 (p=0.0003), and TNF-α (p=0.003), however, the correlation was entirely driven by patients over the age of 60. No age-dependent correlation was observed for patients under the age of 60, which included the majority of patients in the cohort. Conversely, cytokine levels steadily increased with advancing age starting at approximately 60 years old (IL-1β, r = 0.40, p = 0.01; IL-6, r = 0.34, p = 0.04; IL-8, r = 0.32, p < 0.05; TNF-α, r = 0.39, p = 0.02).

Table 2. Mucus cytokine levels in aged and younger patients.

Median cytokine levels of control and CRS subjects are shown for all 17 assayed biological variables. Significant differences between aged and younger patients were identified using the Kruskal-Wallis test. Data is represented as medians with interquartile interquartile range. BOLD, p < 0.05.

	CRS			Control
	Younger (18–59)	Aged (>60)	P value	Younger (18–59)	Aged (>60)	P value
No.	110	37		20	10
IL-1β	100.5 (16.5–299.5)	672.9 (344.1–2754.0)	<0.0001	73.6 (15.1–344.6)	131.9 (38.8–317.6)	0.68
IL-2	3.7 (0.2–15.0)	1.4 (0.2–15.6)	0.53	1.3 (0.2–10.5)	13.5 (0.2–31.2)	0.35
IL-4	0.3 (0.3–1.7)	0.6 (0.3–3.5)	0.11	0.3 (0.3–1.4)	0.3 (0.3–3.6)	0.61
IL-5	19.3 (1.4–67.6)	7.7 (0.4–45.7)	0.20	0.4 (0.2–5.9)	1.7 (0.4–10.7)	0.28
IL-6	132.3 (29.0–399.2)	633.6 (233.3–1493.0)	<0.0001	103.2 (19.7–468.1)	140.1 (58.3 –319.2)	0.62
IL-7	8.4 (2.4–20.7)	14.1 (2.8–28.0)	0.21	4.0 (1.5–12.9)	10.1 (3.3–33.7)	0.21
IL-8	5551 (2486–16480)	21082 (11003–120997)	<0.0001	4613 (1314–20417)	5557 (3605–14492)	0.42
IL-9	2.0 (0.6–9.0)	1.4 (0.6–9.5)	0.96	0.6 (0.6–2.7)	1.2 (0.6–17.0)	0.17
IL-10	4.3 (1.1–15.3)	12.9 (4.5–27.3)	0.005	3.3 (0.1–9.1)	6.0 (1.6–21.1)	0.41
IL-12	45.5 (6.3–154.1)	72.3 (19.7–185.0)	0.27	77.2 (20.3–173.5)	56.6 (25.6–237.6)	0.79
IL-13	17.3 (1.5–72.7)	9.3 (0.3–59.3)	0.67	4.5 (0.4–22.3)	12.3 (2.6–41.2)	0.47
IL-17A	0.1 (0.1–2.0)	0.7 (0.1–5.3)	0.08	0.1 (0.1–0.2)	0.1 (0.1–3.5)	0.26
IL-21	32.0 (6.8–182.9)	146.8 (6.8–245.0)	0.26	101.9 (6.8–227.8)	215.0 (46.3–557.4)	0.16
TNF-α	6.4 (1.6–12.1)	24.0 (6.8–117.6)	<0.0001	5.9 (0.8–17.1)	4.7 (2.1–6.8)	0.56
IFN-γ	0.1 (0.1–1.5)	0.1 (0.1–4.0)	0.20	0.4 (0.4–0.5)	0.9 (0.4–7.4)	0.05
Eotaxin	19.1 (6.4–46.8)	27.0 (10.5–66.6)	0.09	23.8 (16.2–79.8)	65.3 (27.0–261.5)	0.13
RANTES	1628 (305–6633)	645 (146–2818	0.01	1297 (763–2741)	2121 (596–8313)	0.42

Figure 2. Mucus pro-inflammatory cytokines in aged and younger patients.

Mucus levels of IL-1β, IL-6, IL-8, and TNF-α are elevated in aged vs. younger patients. Cytokine values are plotted on a log scale. Solid lines indicate the median with interquartile range.

Figure 3. Correlation of age and mucus pro-inflammatory cytokine levels in aged patients.

Mucus cytokine levels were compared against patient age and analyzed by Spearman correlation. Analyses were performed for the entire patient cohort, and for younger and aged patient groups separately.

Histologic Changes Associated with Aging

We next evaluated whether age had any effect on tissue architecture or cellular infiltrates. In addition to counting tissue eosinophils and neutrophils, we also performed a global evaluation of tissue inflammation, fibrosis, and other markers using a structured histopathologic approach ^{35, 36}. Younger and aged patients did not differ in most histologic variables, including inflammation score, subepithelial edema, papillary changes, squamous metaplasia, and fibrosis (Table E3). In contrast, aged patients had higher mean (p=0.01) and maximum (p=0.02) tissue neutrophil counts compared to their younger counterparts (Figure 4B). A generalized reduction in eosinophil counts was also evident, but did not reach statistical significance (Figure 4A). Tissue neutrophilia closely correlated with mucus levels of IL-1β (r=0.30, p=0.0005), IL-6 (r=0.33, p<0.0001), and IL-8 (r=0.32, p=0.0001) (Figure 4C-F). These results further support a pro-inflammatory, neutrophilic inflammatory response in older CRS patients.

Figure 4. Association of tissue neutrophilia and pro-inflammatory cytokines in aged CRS patients.

Mean tissue eosinophils (A) and neutrophils (B) were compared between younger and aged patients. Bars represent the median and error bars represent the interquartile range. Tissue neutrophil counts correlate with mucus levels of pro-inflammatory cytokines (C-F). Mucus cytokine levels and tissue cells/HPF are plotted on a log scale. The correlation coefficient and associated p-value are presented for each cytokine.

Purulence and Bacterial Colonization in Aged CRS Patients

Subjects enrolled in our ongoing prospective study are routinely evaluated for purulence on nasal endoscopy, with most receiving culture-directed antibiotics as part of their initial medical therapy. When reviewing our updated cohort, we found that clustering of patients with culture-positive purulence overlapped significantly with patients of advanced age (Figure 5A) and that the allocation of culture-positive patients varied significantly between individual clusters (Figure 5B). The prevalence of culture-positive purulence in aged patients was more than twice that among younger patients (p=0.002) (Figure 5C). Sinus cultures from aged patents showed a variety of gram positive, gram negative, and anaerobic bacteria (Table E4). At the individual pathogen level, aged patients were more likely to have cultures positive for P. aeruginosa (p=0.03), while no significant difference was noted for S. aureus (Figure 5C). Analysis of cytokine levels in the entire 147 patient cohort revealed that culture-positive purulence was most highly associated with elevated pro-inflammatory cytokines, including IL-6 (p=0.03), IL-8 (p=0.01), TNF-α (p=0.007), and most significantly IL-1β (p<0.0001) (Table 3). Additionally, elevations in IFN-γ (p<0.05) and eotaxin (p=0.04) were also associated with purulence. Interestingly, associations were driven almost entirely by patients over the age of 60 for IL-1β (p=0.007) (Figure 5D) and IL-8 (p=0.02), while associations with elevated IFN-γ (p=0.02) and the chemokine eotaxin (p=0.04) were driven almost entirely by younger patients. Though not significant in the overall cohort, purulence in aged patients was also associated with reduced levels of IL-2 (p<0.05). In addition to cytokines, the association between purulence and types of infiltrating inflammatory cells was also examined. Among all 147 patients, those with culture positive purulence had elevated mean tissue neutrophil counts (p=0.007), and a trend toward lower mean (p=0.06) and maximum (p=0.05) eosinophil counts. This association was not apparent in younger patients when analyzed separately. In contrast, purulence in aged patents was strongly associated with elevated neutrophil counts (p=0.01) and decreased eosinophil counts (p=0.001). Collectively, these results suggest that younger and aged patients may have differential innate immune responses to colonizing bacteria mediated by specific granulocyte subsets and distinct cytokines.

Figure 5. Prevalence of Culture positive purulence in older and younger patients and between individual CRS endotypes.

Advanced patient age and presence of culture (+) sinonasal purulence are overlapping features (A), particularly within Cluster 2 (B). The dendogram shows the location of individual patients, with age and culture status in the panels to the right. The bar graph shows the frequency of culture (+) purulence within each cluster. The presence of sinonasal purulence varied by age (C) and among aged pateints was associated with elevations in specific pro-inflammaotry cytokines, chiefly IL-1β (D). Solid lines represent the median with error bars indicative of the interquartile range.

Table 3. Effect of purulent bacterial infection on mucus cytokine levels in aged and younger patients.

Median cytokine levels of all CRS patients are shown with comparisons between patients with no purulence and those with culture positive purulence. Significant differences between groups were identified using the Kruskal-Wallis test. Data is represented as medians with interquartile range. BOLD, p < 0.05.

	All CRS			Younger (18–59)			Aged (>60)
	No purulence	Culture (+) purulence	P value	No purulence	Culture (+) purulence	P value	No purulence	Culture (+) purulence	P value
No.	103	44		85	25		18	19
IL-1β	116.3 (16.4–389.5)	462.3 (105.2–2744.0)	<0.0001	82.52 (13.7–82.5)	129.6 (62.6–460.4)	0.09	456.1 (146.8 (800.5)	2441.0 (578.1–5263.0)	0.0007
IL-2	3.4 (0.2–15.0)	2.0 (0.2–15.6)	0.55	3.3 (0.2–13.6)	6.3 (0.2–18.2)	0.45	7.8 (0.2–20.8)	0.2 (0.2–7.9)	<0.05
IL-4	0.3 (0.3–1.7)	0.3 (0.3–2.5)	0.19	0.3 (0.3–1.5)	0.3 (0.3–2.4)	0.27	0.6 (0.3–3.5)	0.6 (0.3–4.2)	0.86
IL-5	18.9 (2.0–65.8)	7.9 (0.2–64.6)	0.18	20.8 (1.3–66.1)	15.3 (1.9–81.2)	0.97	10.9 (1.6–175.3)	5.1 (0.2–20.3)	0.23
IL-6	148.8 (42.2–524.5)	315.6 (88.5–1267.0)	0.03	118.1 (21.5–394.0)	198.1 (65.7–563.5)	0.22	420.6 (109.8–1182.0)	1014.0 (285.9–3217.0)	0.31
IL-7	9.2 (2.5–22.4)	7.5 (2.3–23.7)	0.71	8.4 (2.0–21.2)	8.2 (2.8–18.3)	0.83	15.6 (9.2–25.0)	7.3 (1.5–28.4)	0.22
IL-8	8051 (2492–20971)	14636 (3685–77289)	0.01	5680 (2069–16597)	5363 (2957–21946)	0.55	12658 (8890–34042)	46279 (18802–335288)	0.02
IL-9	2.0 (0.6–9.3)	1.4 (0.6–9.0)	0.72	2.8 (0.6–9.0)	0.6 (0.6–9.7)	0.76	1.1 (0.6–11.2)	1.5 (0.6–9.1)	0.89
IL-10	4.6 (1.1–17.8)	7.1 (2.0–22.8)	0.34	3.7 (0.8–14.7)	5.9 (2.0–16.6)	0.31	12.9 (9.1–27.6)	10.3 (2.1–27.4)	0.55
IL-12	59.9 (5.8–152.5)	66.8 (11.1–183.8)	0.56	45.3 (5.1–153.6)	50.3 (11.5–171.4)	0.64	74.0 (22.0–133.7)	69.9 (3.4–306.8)	0.95
IL-13	16.7 (1.6–66.9)	10.9 (0.3–78.4)	0.64	16.7 (1.5–55.1)	21.9 (1.7–98.4)	0.52	21.0 (4.1–136.7)	5.7 (0.1–47.1)	0.14
IL-17A	0.1 (0.1–2.4)	0.1 (0.1–3.2)	0.86	0.1 (0.1–1.9)	0.1 (0.1–2.1)	0.26	0.8 (0.1–4.7)	0.7 (0.1–8.2)	0.78
IL-21	71.6 (7.7–219.0)	25.5 (6.8–163.2)	0.11	56.8 (7.6–210.8)	17.0 (6.8–119.4)	0.11	170.1 (12.5–247.3)	118.9 (6.8–243.9)	0.31
TNF-α	6.9 (1.6–14.8)	10.0 (4.8–69.1)	0.007	5.3 (0.8–11.9)	8.7 (4.8–14.5)	0.05	19.0 (7.3–54.1)	30.3 (3.4–174.5)	0.36
IFN-γ	0.1 (0.1–1.5)	0.6 (0.1–3.2)	<0.05	0.1 (0.1–1.1)	0.8 (0.1–3.5)	0.02	0.1 (0.1–5.5)	0.1 (0.1–2.2)	0.90
Eotaxin	18.5 (6.4–45.2)	29.0 (9.8–81.6)	0.04	13.4 (5.8–43.4)	24.2 (9.9–96.9)	0.04	26.4 (10.7–66.0)	30.9 (9.2–70.0)	0.97
RANTES	1246 (307–6299)	1791 (159–5789)	0.4	1512 (304–6833)	2698 (281–6225)	0.87	759 (291–1916)	356 (73–3513)	0.68
Neut/HPF (mean)	0.7 (0.0–4.2)	4.0 (0.1–20.8)	0.007	0.6 (0.0–4.0)	0.9 (0.0–12.9)	0.39	2.7 (0.0–9.0)	7.5 (1.7–22.9)	0.01
Neut/HPF (max)	1.0 (0.0–9.5)	6.5 (0.3–35.0)	0.01	1.0 (0.0–8.0)	2.0 (0.0–27.5)	0.43	5.0 (0.0–13.5)	15.5 (4.3–53.3)	0.02
Eos/HPF (mean)	49.4 (4.5–128.8)	4.0 (0.7–95.8)	0.06	50.0 (4.3–125.3)	32.9 (1.4–236.9)	0.91	42.6 (7.5–168.0)	1.4 (0.2–10.9)	0.001
Eos/HPF (max)	88.0 (11.3–196.0)	7.5 (1.0–199.5)	0.05	88.0 (11.0–175.0)	40.5 (2.5–332.5)	0.98	93.0 (9.5–219.5)	4.0 (0.3–20.3)	0.004

Age-specific Roles of Pro-Inflammatory Cytokines in CRS Disease Severity

Finally, we sought to determine whether levels of neutrophilic pro-inflammatory cytokines affected measures of disease severity or incidence of prior endoscopic sinus surgery. We were specifically interested in determining whether there were any age-specific effects of any cytokines assessed in our study. We first determined whether there were associations between cytokine levels and disease-specific quality of life, which was assessed using the sinonasal outcomes test-22 (SNOT-22) questionnaire (Table E5). Elevated mucus IL-1β levels were associated with significantly worse SNOT-22 scores among older patients at baseline, with a clinically relevant difference of 20.7 points (p < 0.05). A similar trend was noted for IL-8, with an 18.9 point difference, though this did not reach statistical significance (p = 0.08). In younger patients, cytokine levels were not associated with SNOT-22 scores. An interaction model suggested a possible age-specific effect of IL-1β on SNOT-22 scores, though this association did not reach statistical significance (p = 0.07). We next assessed the effect of cytokines on rates of revision surgery (Table E6). In younger patients several cytokines were associated with revision surgery, including IL-4 (OR 2.47, p=0.03), IL-6 (OR 2.80, p=0.01), IL-13 (OR 2.31, p=0.05), IL-17A (OR 3.69, p=0.002), and eotaxin (OR 2.58, p=0.02). In aged patients, elevated IL-2 (OR 0.21, p=0.03) and IL-21 (OR 0.21, p=0.04) were associated with a reduced likelihood of prior surgery. The cytokine associated with the highest odds ratio of prior surgery in aged patients was IL-1β (OR 6.07), however this association did not reach statistical significance (p=0.12). The regression model showed that IL-2 had an age-specific association with revision surgery (p=0.01).

DISCUSSION

The role of aging in CRS pathophysiology remains largely undefined despite the increases in health care utilization associated with an aging population. We hypothesized that aging in CRS may be associated with unique immune signatures and/or inflammatory endotypes. Our data suggests that aged patients with CRS have elevated levels of select cytokines, including IL-1β, IL-6, IL-8, and TNF-α. Aging was likewise associated with increases in bacterial colonization and/or infection that correlated with levels of pro-inflammatory cytokines. Our data supports a neutrophil-driven mechanism of chronic inflammation in aged patients, irrespective of polyp status, that may also be associated with an inability to clear pathogenic bacteria from the sinonasal cavity. Clinically, these patients frequently present with persistent mucosal inflammation and recalcitrant bacterial colonization that has a poor long-term response to antibiotics. Patients with this age-dependent, pro-inflammatory immune signature may be less likely to respond to glucocorticosteroids and may ultimately benefit from alternative and perhaps undeveloped treatment approaches.

Support for unique disease mechanisms among older patients is supported by a small number of prior studies. A single institution North American cohort study reported an increased prevalence of nasal polyps and asthma among elderly CRS patients^{6, 7}. Advanced age was not associated with tissue eosinophil or neutrophil counts, but did correlate with a reduction in S100A8/9 protein expression⁶. S100A8/9 are epithelial barrier-associated proteins that also have antimicrobial activity⁴⁶. A subsequent study by this group did not show any association between age and tissue IL-6 levels, which is in direct contrast to the results from our patient cohort. It should be noted, however, that this study only measured IL-6 levels in patients with nasal polyps, and none of the patients were older than 60 years of age. These findings are actually quite consistent with our results, which also failed to show any significant correlation between mucus cytokines and age in patients under the age of 60, but found a steady and progressive increase in several pro-inflammatory cytokines after age 60.

While likely understudied in CRS, the association between aging and inflammation has been more extensively evaluated in asthma, which shares many demographic and mechanistic features with CRS. Advanced age and an associated neutrophilic airway cellular infiltrate has been associated with more severe asthma phenotypes^{18, 19, 47–50}. Airway neutrophilia in these patients has likewise been linked to an elevation in sputum pro-inflammatory cytokines that directly or indirectly regulate neutrophilic inflammation, including IL-1α, IL-1β, IL-6, and IL-8^{50, 51}. Severe neutrophilic asthma has also been directly linked to upregulation of several IL-1 receptor family members^{48, 52}. Finally, a recent study found that severe, treatment-resistant asthma associated with neutrophilic inflammation was closely linked to elevations in the neutrophil chemokine IL-8, and to related elevations in pathogenic airway bacteria⁵⁰.

We found that aged CRS patients were much more likely to harbor symptomatic bacterial infection or colonization than their younger counterparts. Several cytokines were associated with mucopurulent infection, in a pattern that varied largely by age. Bacterial infection in aged patients was associated with neutrophilic, rather than eosinophilic tissue inflammation, and was likewise associated with elevated mucus levels of both IL-1β and IL-8. These relationships were not observed in younger patients, suggesting a mechanism uniquely associated with the aging process. Mucopurulent infection was also associated with reduced mucus IL-2. This cytokine is generally considered to be an essential growth factor for activated T-cells and has important roles in the immune response to microbial pathogens⁵³. Elevation of IL-2 in CRSsNP has previously been reported^{54, 55}, however, prior studies have not specifically assessed IL-2 levels in CRS patients of advanced age. Low levels of IL-2 in older patients could potentially result in dysfunctional T-cell maturation and a poor innate and adaptive immune response to sinonasal bacteria.

Bacteria and other pathogens are strong regulators of mucosal immunity and are among the most potent inducers of pro-inflammatory cytokine production and secretion. The innate immune response induced by bacteria is mediated by pattern recognition receptors, and ultimately results in neutrophil chemotaxis and secondary tissue infiltration by macrophages and other effector cells^{56, 57}. It is well established that resident bacteria can affect cytokine release and other immune responses in the human gut and airway^58–65, however such host-pathogen interactions have not been extensively evaluated in CRS. A recent in vitro study using tissue explants from CRS patients found that gram-positive and gram-negative bacteria can stimulate mucosal expression of both IL-6 and RANTES⁶⁶. Likewise, Chalermwatanachai et al. found that intramucosal presence of P. aeuruginosa was associated with elevated tissue TNF-α in CRSsNP patients⁶⁷. To our knowledge, the current study is the first to report a role for aging in the CRS host-pathogen immune response. Interestingly, current evidence suggests that aging may be associated with alterations to the sinonasal microbiome itself. A small study in healthy non-CRS individuals identified advanced age as an independent predictor of middle meatal microbiome composition at both the species and phylum level⁶⁵. A more recent analysis of CRS microbial community structure found that, of all demographic factors, only age was associated with significant changes to bacterial abundance, with changes most evident in the Pseudomonas genus⁴². While somewhat limited by the use of standard culture-based approaches, the results of the current study similarly suggest that advanced age may be associated with changes to microbial community structure or abundance, and points to a close relationship between microbes and the host immune response.

It should be noted that our study had some limitations that may prevent the generalizability of findings. Only surgical patients were enrolled, meaning that the patients evaluated may carry greater overall disease burden than the general CRS population. This was also a single-institution study that may not capture regional variations in disease characteristics. This concern may be partially alleviated given that patients were enrolled at a tertiary care institution with a large referral radius. Patients in the study were residents of 3–4 contiguous states in the southeastern United States, rather than being derived entirely from a single metropolitan area. Cytokines analyzed in our study were carefully chosen in order to capture a broad representation of Th1-, Th2-, and Th17-associated inflammation. However, several pro-inflammatory cytokines and mediators commonly associated with neutrophil recruitment or the innate immune response were not included in our analysis, with examples including GM-CSF, IL-1α, IL-18, and IL-23. Use of an unstructured proteomic or transcriptomic approach may ultimately augment our findings and provide greater understanding of inflammatory changes unique to older patients. Finally, associations identified in aged patients between the pro-inflammatory immune response and resident bacteria were identified through the use of standard culture-based approaches, with their inherent limitations. Future studies that incorporate culture-independent analysis of the sinonasal microbiome may help to further delineate connections between microbial community structure and abundance, and the host immune response.

In conclusion, we have used hierarchical cluster analysis to identify age-specific changes to the CRS inflammatory signature. Aging in CRS was associated with elevations in pro-inflammatory cytokines, neutrophilic tissue inflammation, and increased prevalence of bacterial infection or colonization. Our results suggest that older patients may be less responsive to corticosteroids and may benefit from therapeutic approaches that target the innate immune response to microbial pathogens.

CLINICAL IMPLICATIONS:

Aging in CRS is associated with a neutrophilic immune signature that is less likely to respond to corticosteroids and suggests a need to develop alternative therapies in this at-risk population.

ABBREVIATIONS

AERD

aspirin-exacerbated respiratory disesase

AFRS

allergic fungal rhinosinusitis

BMI

body mass index

CRSsNP

chronic rhinosinusitis without nasal polyps

CRSwNP

chronic rhinosinusitis with nasal polyps

CT

computed tomography

LTR

anti-leukotriene

SIT

smell identification test

SNOT-22

22 item sinonasal outcome test