Inflammatory Characteristics of Central Compartment Atopic Disease

Abstract

Background:

Central compartment atopic disease (CCAD) is an emerging phenotype of chronic rhinosinusitis with nasal polyposis (CRSwNP) characterized by prominent central nasal inflammatory changes. This study compares the inflammatory characteristics of CCAD relative to other phenotypes of CRSwNP.

Methods:

A cross sectional analysis of data from a prospective clinical study was performed on patients with CRSwNP who were undergoing endoscopic sinus surgery (ESS). Patients with CCAD, aspirin-exacerbated respiratory disease (AERD), allergic fungal rhinosinusitis (AFRS) and nontypeable CRSwNP (CRSwNP NOS) were included and mucus cytokine levels and demographic data were analyzed for each group. Chi-squared/Mann-Whitney U tests and partial least squares discriminant analysis (PLS-DA) were performed for comparison and classification.

Results:

A total of 253 patients were analyzed (CRSwNP, n=137; AFRS, n=50; AERD, n=42; CCAD, n=24). Patients with CCAD were the least likely to have comorbid asthma (p = 0.0004). The incidence of allergic rhinitis in CCAD patients did not vary significantly compared to patients with AFRS and AERD, but was higher compared to patients with CRSwNP NOS (p = 0.04). On univariate analysis, CCAD was characterized by less inflammatory burden, with reduced levels of IL-6, 8, IFN-γ, and eotaxin relative to other groups and significantly lower type 2 cytokines (IL-5, IL-13) relative to both AERD and AFRS. These findings were supported by multivariate PLS-DA, which clustered CCAD patients into a relatively homogenous low-inflammatory cytokine profile.

Conclusions:

CCAD has unique endotypic features compared to other patients with CRSwNP. The lower inflammatory burden may be reflective of a less severe variant of CRSwNP.

Introduction

Central compartment atopic disease (CCAD) is a recently described CRS phenotype characterized by polypoid degeneration of the central structures of the nasal cavity (nasal septum, middle turbinate, superior turbinate) and a lack of peripheral paranasal sinus involvement.¹ These endoscopic findings are associated with a centrally limited radiologic pattern in the paranasal sinuses and significantly lower Lund-Mackay scores.^3–5 Originally introduced by DelGaudio et al in 2017, CCAD is increasingly being recognized as a clinically relevant phenotype of CRSwNP. With greater clinical and investigative attention directed toward this disease process, several studies have started to develop links between CCAD and both comorbidities and postoperative outcomes. For example, multiple studies have drawn a link between CCAD and allergic rhinitis (AR), and 74–100% of CCAD patients have some form of allergy sensitization.^1,6–7 A similar relationship has been shown between polypoid changes of the middle turbinate and atopic disease, with a 95% specificity for inhalant allergy^2,8–9. It has been hypothesized that allergic particle deposition into these central structures may initiate an inflammatory cascade that results in obstructive edema and eventually development of CRS. This is thought to be mediated by a predominantly type 2 inflammatory response and IgE-mediated reactions.¹⁰

The relationship between CCAD and asthma has also been investigated. DelGaudio’s group reported a 17.1% prevalence of asthma among patients with CCAD, which was the lowest prevalence of any CRSwNP subtype.⁶ Similar to CRS, asthma is a highly heterogeneous disease that has a large subset of patients with associated IgE-mediated allergy and type 2 inflammation.¹² Especially when viewed from a unified airway perspective, it is unknown why this patient population has such a robust allergic response in the upper airway without a significant lower airway component.

With respect to post-operative outcomes, CCAD patients appear to have a comparably durable response to endoscopic sinus surgery (ESS) when compared to other patients with CRSwNP. In a cohort of patients in the U.S., rates of polyp recurrence and revision surgery were significantly lower in patients with CCAD compared to other CRSwNP subtypes following primary surgery with postoperative steroid irrigation.¹² Liang et al likewise evaluated Taiwanese patients with CCAD and found a significant improvement in post-operative outcomes on SNOT-22 and mucociliary clearance time (MCT) scores compared to other groups.¹³ Collectively, current evidence suggests that CCAD patients may have a more sustained response to surgical treatment.

Although the clinical characteristics of CCAD have been well described, little is known about the etiology and inflammatory mechanisms that contribute to the disease. Consequently, we aimed in this study to identify the inflammatory characteristics of CCAD by evaluating mucus cytokine levels and comparing them to other phenotypes of CRSwNP. We hypothesized that patients with CCAD would have a distinct cytokine signature that may explain some of its unique clinical characteristics.

Methods

This study was approved by the Institutional Review Board of Vanderbilt University Medical Center. Patients with CRSwNP who were undergoing endoscopic sinus surgery were enrolled in the study. CRS was diagnosed according to the European Position Paper on Rhinosinusitis and Nasal polyps and the International Consensus Statement on Allergy and Rhinology^10,14. CRSwNP was diagnosed by the presence of visible nasal polyps on nasal endoscopy or during endoscopic sinus surgery. AERD was diagnosed based on the clinical triad of nasal polyps, asthma with active disease, and at least two documented reactions to aspirin or other NSAIDs or a positive aspirin challenge. AFRS was diagnosed according to criteria established by Bent and Kuhn and included the presence of nasal polyps, allergic mucin, and a history of fungal atopy⁵. All AFRS patients had either skin prick testing or were clinically diagnosed with allergic rhinitis. CCAD was diagnosed based on the presence of centrally located polypoid disease on endoscopy or imaging and was independently confirmed by two fellowship-trained Rhinologists. Surgery was offered if patients failed maximal medical therapy, defined as at least 2 weeks of oral prednisone and, antibiotics, and additional therapies that included a combination of decongestants, topical corticosteroids, antihistamines, and saline irrigations. Control cases included patients undergoing anterior skull base or pituitary surgery without clinical or radiographic history of sinus disease. Patients were excluded if they had been treated with systemic steroids within four weeks of surgery, had comorbid cystic fibrosis, autoimmune, or granulomatous disease, or if they had received immune-directed monoclonal antibodies or other immunomodulators within three months of surgery. The presence of asthma and allergic rhinitis was recorded for all patients. Allergic rhinitis was diagnosed based on positive skin prick testing or physician diagnosis based on seasonal variation in atopic symptoms and relief after using an oral antihistamine or intranasal corticosteroids. Asthma was diagnosed based on a bronchodilator response on pulmonary function testing, methacholine challenge, or prior diagnosis by a pulmonologist and/or allergist. All patients underwent high-resolution CT scan of the paranasal sinuses within 3 months of surgery. A Lund McKay score was assigned to each scan to assess the severity of radiographic sinus disease.

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 and included standardized evaluation of both tissue eosinophilia and neutrophilia.^15,16

Mucus Collection and Cytokine Measurement

9 × 24mm polyurethane sponges (Summit Medical; St. Paul, MN) were placed into the middle meatus or ethmoid cavity of each subject under endoscopic guidance at the time of surgery, as has been previously reported^17,18. This method of mucus collection has the advantage of minimal specimen dilution and standardization between patients, and we have previously shown that mucus cytokine levels are highly consistent within different subsites of the sinonasal cavity¹⁹. Sponges were left in place for 5 minutes, after which they were 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. Samples were then gently vortexed and again centrifuged for 5 minutes to remove cellular debris. Supernatants were removed, placed into a new microcentrifuge tube, and frozen at −80°C.

A multiplex cytokine bead assay (BD Biosciences; Franklin Lakes, NJ) was used to analyze all mucus samples. In brief, 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. Samples were centrifuged at 200 × g for 5 minutes after the addition of 1 mL wash buffer, 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 seventeen cytokines and inflammatory mediators were analyzed. The assay data was analyzed using BD FCAP Array Software version 3.0.

Statistical Analysis

Descriptive statistics for demographic/clinical characteristics and inflammatory mediators were presented as the median with interquartile (IQR) or mean with standard deviation (SD) for continuous variables and the frequency with percentages for categorical variables. Differences between defined CRS subgroups or clusters were evaluated using the Kruskal-Wallis test for continuous variables or c2 test for categorical variables followed by Dunn’s test for multiple comparisons. Pairwise mean comparisons using Mann-Whitney U tests was performed for all cytokines, cell counts, and CT scores. Statistical significance was defined as a P value < 0.05.

Sparse Partial Least Squares Discriminant Analysis (sPLS-DA)

As inflammatory cytokine variables are highly skewed and correlated, alternative approaches to standard regression must be used when comparing cytokines in a multivariate fashion. sPLS-DA is a supervised, multivariate approach that identifies linear cytokine groupings that best differentiate or classify between previously known or specified groups. As such, it can be thought of as a supervised version of principal component analysis, which our group has used previously.^20,21 For the present analysis, sPLS-DA was implemented using the mixOmics package in R, using inflammatory nasal mucus levels alone to identify potentially different cytokine patterns between AERD, AFRS, CCAD, and non-typed CRSwNP.²² Cytokine component loadings and PLS-DA plots were then used to assess the relative distribution of patients in each group across the three sPLS-DA components.²³

Hierarchical cluster analysis

Sample size for cluster analysis was estimated by establishing a subject to variable ratio of greater than 5:1 as recommended by Gorsuch and Hatcher^{24, 25}. 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 log-transformation, resulting in elimination or significant reduction of skewing for all variables. Hierarchical cluster analysis was performed using Ward’s method on squared Euclidian distances. 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 repeating the estimation procedure on 1000 resampled datasets to ensure that the clustering results were stable and not unique to the original dataset. Variances of each biological variable were compared using the coefficient of variation to create a unitless measure of comparison between groups. Omnibus hypothesis testing for differences in variance was performed using the asymptotic test for the equality of coefficients of variation.

Results

Two hundred and fifty-three patients were enrolled in the study, with 24 meeting diagnostic criteria for CCAD (Table 1). Patients across CRSwNP phenotypes were similar with respect to gender and BMI. CCAD patients had significantly fewer prior surgeries compared to AERD (p = 0.0002), AFRS (p = 0.0003), and CRSwNP NOS (p = 0.011). There was a lower percentage of patients with asthma within the CCAD population (p = 0.0004) but notably the rates of allergic rhinitis within the group were on par with AERD/AFRS patients and significantly higher than the CRSwNP NOS group (p=0.04).

Table 1.

Demographics and clinical characteristics of study population.

Variable	CCAD	AERD	AFRS	CRSwNP NOS	p-values (vs. CCAD)
Number	24	42	50	137	*
Age, years	48.3 +/− 15.6	48.5 +/− 12.5	39.2 +/− 14.1	51.4 +/− 15.1	AERD 0.964, AFRS 0.023, CRSwNP NOS 0.376
Sex, no. female (%)	6 (26.1)	24 (57.1)	21 (42)	45 (32.4)	0.0174
Race (% white)	14 (63.6)	36 (87.8)	35 (70)	124 (89.7)	0.0004
Asthma, no. (%)	7 (30.4)	35 (83.3)	27 (54)	64 (46.4)	0.0001
Allergic rhinitis, no. (%)	17 (73.9)	30 (71.4)	39 (78.0)	81 (58.2)	0.045
Prior surgery, no. (%)	4 (17.4)	24 (60.0)	26 (56.5)	45 (41.7)	AERD <0.0001, AFRS <0.0001 CRSwNP NOS 0.011
BMI (kg/m2)	48.3 ± 6.8	28.9 ± 5.4	31.4 ± 8.3	29.1 ± 5.9	AERD 0.774 AFRS 0.116 CRSwNP NOS 0.661
CT score	12.9 ± 1.6	17.6 ± 4.9	17.5 ± 5.4	16.7 ± 9.5	AERD <0.0001 AFRS <0.0001 CRSwNP NOS <0.0001
Eosinophil hpf, max	95.4 ± 72.3	280.5 ± 217.3	194.7 ± 163.3	165.5 ± 197.8	AERD 0.033 AFRS 0.101 CRSwNP NOS 0.738
Eosinophil hpf, mean	72.5 ± 73.0	188.7 ± 141.9	142.5 ± 126.2	105.7 ± 121.9	AERD 0.047 AFRS 0.073 CRSwNP NOS 0.724

Values are presented as means +/− SDs or means with interquartile ranges. Differences between groups were assessed by using the Kruskal-Wallis test or X² analysis. Boldface text indicates a P value of less than 0.05. CCAD, central compartment atopic disease; AERD, Aspirin-exacerbated respiratory disease; AFRS, allergic fungal rhinosinusitis; CRSwNP NOS, chronic rhinosinusitis not otherwise specified; BMI, body mass index; hpf, high powered field.

Average preoperative Lund-Mackay scores were 17.6 ± 4.9 for AERD, 17.5 ± 5.4 for AFRS, 16.7 ± 9.5 for CRSwNP NOS, and 12.9 ± 1.6 for CCAD, indicating lower overall radiographic burden in the CCAD population (Table 1). Tissue Eosinophilcell counts were performed for all CRSwNP subtypes (AERD, 280.5 ± 141.9; AFRS, 194.7 ± 126.2; CCAD, 95.4 ± 73.0; CRSwNP NOS, 165.5 ± 121.9). The number of eosinophils/HPF varied significantly between CCAD and AERD (p=0.033). No significant difference in tissue neutrophils were identified between the groups.

Univariate Cytokine Signatures in CCAD

CCAD was generally characterized by reduced inflammatory burden relative to other types of CRSwNP (Table 2). Compared to AERD, CCAD had significantly lower eotaxin (p=0.001), IFN-γ (p=0.024), IL-6 (p=0.006), IL-10 (p=0.002), and IL-13 (p=0.002). Similarly, compared to AFRS, CCAD had reduced mucus levels of RANTES (p=0.024), TNF-α (p=0.0009), IFN-γ (p=0.011), IL-4 (p=0.016), IL-8 (p=0.008), IL-9 (p=0.002), IL-10 (p=0.011), and IL-13 (p=0.00025). CCAD could also be differentiated from CRSwNP NOS by significantly lower eotaxin (p=0.038) and IL-6 (p=0.019) (Figure 1).

Table 2.

Cytokine levels in CCAD, AERD, AFRS, and CRSwNP NOS patients.

Cytokines	CCAD	AERD	AFRS	CRSwNP NOS	p-values (vs. CCAD)
Eotaxin	26.9 +/− 32.4	95.4 +/− 121.2	70.2 +/− 113.8	62.6 +/− 98.5	AERD 0.001 AFRS 0.099 CRSwNP NOS 0.038
Gm csf	1.31 +/− 2.05	10.8 +/− 42.8	6.80 +/− 16.5	2.14 +/− 4.09	AERD 0.249 AFRS 0.266 CRSwNP NOS 0.749
Rantes	3424.6 +/− 6847.1	2821.0 +/− 4542.3	1146.2 +/− 1954.1	2870.1 +/− 6847.1	AERD 0.462 AFRS 0.024 CRSwNP 0.173
TNF-α	7.59 +/− 6.34	35.1 +/− 74.1	74.5 +/− 164.6	62.0 +/− 274.9	AERD 0.438 AFRS 0.0009 CRSwNP NOS p=0.353
IFN-γ	1.70 +/− 2.69	5.56 +/− 10.6	35.2 +/− 165.4	18.1 +/− 91.2	AERD 0.024 AFRS 0.011 CRSwNP NOS p=0.18
IL-1 α	121.0 +/− 193.1	120.0 +/− 135.5	154.4 +/− 142.4	121.7 +/− 176.4	AERD 0.677 AFRS 0.216 CRSwNP NOS 0.818
IL-1 β	213.4 +/− 259.4	712.7 +/− 1443.8	445.5 +/− 571.9	1128.4 +/− 2503.7	AERD 0.442 AFRS 0.03 CRSwNP NOS 0.236
IL-2	19.1 +/− 36.1	23.1 +/− 72.1	101.5 +/− 346.4	152.8 +/− 924.5	AERD 0.504 AFRS 0.931 CRSwNP NOS 0.136
IL-4	1.08 +/− 1.72	2.28 +/− 3.59	20.8 +/− 72.1	2.71 +/− 10.2	AERD 0.228 AFRS 0.016 CRSwNP NOS 0.764
IL-5	63.7 +/− 107.3	267.1 +/− 466.1	267.3 +/− 362.8	103.3 +/− 233.6	AERD 0.059 AFRS 0.003 CRSwNP NOS 0.454
IL-6	181.7 +/− 168.9	2760.5 +/− 9143.0	2200.8 +/− 3462.2	2163.5 +/− 5729.1	AERD 0.0006 AFRS <0.0001 CRSwNP NOS 0.019
IL-7	21.3 +/− 15.1	19.3 +/− 22.9	20.3 +/− 23.8	19.4 +/− 21.9	AERD 0.136 AFRS 0.245 CRSwNP NOS 0.153
IL-8	15619.8 +/− 15883.6	99059.4 +/− 415112.9	53265.9 +/− 92994.2	108492.0 +/− 489713.9	AERD 0.755 AFRS 0.008 CRSwNP NOS 0.075
IL-9	3.05 +/− 8.15	10.6 +/− 27.9	66.0 +/− 164.9	6.95 +/− 16.6	AERD 0.534 AFRS 0.002 CRSwNP NOS 0.566
IL-10	13.4 +/− 21.4	30.8 +/− 50.6	19.2 +/− 20.1	55.6 +/− 312.0	AERD 0.002 AFRS 0.011 CRSwNP NOS 0.249
IL-12	146.5 +/− 185.1	110.7 +/− 122.9	128.2 +/− 170.3	126.3 +/− 202.4	AERD 0.759 AFRS 0.828 CRSwNP NOS 0.339
IL-13	52.8 +/− 70.8	153.3 +/− 179.4	179.4 +/− 182.3	66.9 +/− 128.8	AERD 0.002 AFRS 0.0003 CRSwNP NOS 0.583
IL-17a	3.50 +/− 5.9	3.56 +/− 3.95	35.5 +/− 127.6	4.29 +/− 5.86	AERD 0.385 AFRS 0.068 CRSwNP NOS 0.909
IL-18	1625.4 +/− 609.5	1349.5 +/− 449.0	1457.8 +/− 819.3	1358.9 +/− 731.0	AERD 0.857 AFRS 0.548 CRSwNP NOS 0.583
IL-21	219.8 +/− 205.9	148.4 +/− 196.0	208.4 +/− 286.0	214.2 +/− 328.1	AERD 0.117 AFRS 0.279 CRSwNP NOS 0.343

Mucous levels of cytokines are shown for each group. Values are presented as means +/− SDs. Differences between groups were assessed by using the Kruskal-Wallis test followed by pairwise T-test. All cytokine levels are presented as ng/mL. Boldface text indicates P value of less than 0.05. CCAD, central compartment atopic disease; AERD, Aspirin-exacerbated respiratory disease; AFRS, allergic fungal rhinosinusitis; CRSwNP NOS, chronic rhinosinusitis not otherwise specified; gm csf, Granulocyte-macrophage colony-stimulating factor; TNF-α, Tumor necrosis factor alpha; IFN-γ, Interferon gamma; IL, interleukin

Figure 1. Mucus cytokines in patients with CCAD, AERD, AFRS, patients with CRSwNP NOS.

Cytokine values are plotted on a log scale for each CRSwNP subgroup. Solid lines indicate medians with interquartile ranges.

Multivariate Cytokine Analysis

Three components were used in the sPLS-DA model to identify CCAD-specific cytokine profiles relative to other CRSwNP groups, based on the convention that in sPLS-DA the optimal number of components should be the minimum of p, the number of variables in the dataset, or K-1, where K is the number of classes or groupings of interest.²⁶ While sharing some conceptual similarity to PCA, sPLS-DA is optimized to identify latent variables that maximize the difference between the classification groups of interest, rather than the unsupervised PCA approach that only seeks to maximize variance in the data. Thus, sPLS-DA is an ideal approach to identify potential cytokine differences between CCAD, AERD, AFRS, and CRSwNP NOS. In our cohort, Component 1 was primarily characterized by type-2 cytokine loadings, including IL-13, IL-9, IL-5, and IL-4, while component 2 was weighted towards markers of Type 1/3 inflammation, with primary contributions from IL1b, TNF-α, and IFN-γ. The third component was dominated by contributions from IL-21 and IL-12. For components 1 and 2, CCAD was characterized by the lowest mean values for all 12 included cytokines except for RANTES, which was lowest in AFRS. This was visually confirmed on the PLS-DA plot, which clustered CCAD patients into a relatively homogenous low-inflammatory group centered near the origin for each component axis (Figure 2). In comparison, AFRS, AERD, and CRSwNP NOS were more dispersed along components, with AFRS cytokine profiles showing notable variability along the type 2 / component 1 axis while the inverse was seen in non-typeable CRSwNP, which was more variable along the type 1/3 / component 2 axis.

Figure 2. PLS-DA plot of patients with CCAD, AERD, AFRS, patients with CRSwNP NOS.

partial least squares linear discriminant analysis was used to separate classes based on cytokines and investigate differential cytokine patterns between groups. Component 1 is primarily weighted towards type 2 inflammatory markers while component 2 is weighted more towards type 1 and 3 inflammatory markers. Note clustering of CCAD around low values of both components, suggestive of generally low inflammatory burden. Ellipses represent the 95% bounds for each CRSwNP phenotype, with the area of each circle measuring the overall heterogeneity of cluster members.

Endotypic Features of CCAD Identified Using Hierarchical Cluster Analysis

Our group has previously used unstructured statistical approaches to identify CRS endotypes using mucus cytokine levels.^{20, 27–30} Here we sought to use a similar approach to determine whether patients with CCAD have any unique endotypic features. We identified six CRSwNP inflammatory disease clusters and validated their stability using bootstrap analysis. Most CCAD patients fell into cluster 2 (n=9, 41%) and cluster 5 (n=8, 36%) (Figure 3). These clusters were characterized primarily by reduced levels of type 2 mediators (IL-5, 13) (Table 3).

Figure 3. Identification of inflammatory disease clusters in CRSwNP patients. * represents patients with CCAD.

Dendogram representing hierarchical cluster analysis of patients with CRSwNP. Analysis was performed by using the Ward method on squared Euclidian distances with seventeen cytokines as biological variables.

Table 3.

Descriptive statistics by cytokine clusters

	Cluster 1 N=18	Cluster 2 N=66	Cluster 3 N=18	Cluster 4 N=50	Cluster 5 N=47	Cluster 6 N=3
IL-1β	63.9 (54.9–84.8)	10.0 (4.2–21.0)	13.2 (9.0–18.4)	11.8 (7.8–23.6)	10.6 (5.2–17.2)	16.9 (14.9–20.4)
IL-2	1.5 (1.4–2.7)	1.9 (1.3–4.0)	0.0 (0.0–1.5)	3.3 (1.5–4.7)	1.5 (0.0–2.9)	24.0 (23.7–35.0)
IL-4	1.02 (0.13–1.51)	0.53 (0.15–1.33)	0.00 (0.00–0.85)	1.17 (0.63–1.91)	0.00 (0.00–0.53)	13.06 (13.02–16.86)
IL-5	5.9 (2.1–8.4)	3.7 (1.8–6.5)	5.3 (3.3–10.1)	19.5 (15.4–22.3)	5.5 (2.4–8.1)	16.5 (14.5–21.8)
IL-6	71.3 (48.7–95.4)	12.9 (7.1–18.6)	13.9 (10.8–23.0)	31.5 (16.3–52.2)	11.3 (7.7–19.7)	74.5 (61.5–84.9)
IL-7	4.2 (2.9–6.4)	2.3 (1.5–3.9)	3.5 (2.6–4.2)	3.6 (2.5–4.5)	5.0 (4.1–6.4)	7.9 (6.0–9.2)
IL-8	361 (260–541)	93 (58–164)	138 (113–168)	107 (65–218)	116 (86–164)	231 (201–295)
IL-9	0.18 (0.00–0.79)	0.79 (0.00–2.26)	0.00 (0.00–0.00)	3.80 (1.90–6.16)	0.00 (0.00–2.01)	25.7 (15.7–26.9)
IL-10	5.0 (4.1–7.5)	2.4 (1.7–2.6)	3.1 (2.6–4.3)	4.3 (3.5–5.3)	2.2 (1.1–3.4)	9.2 (8.4–9.3)
IL-12	10.1 (8.0–15.3)	6.9 (2.5–10.7)	20.0 (18.1–21.8)	6.1 (3.2–10.5)	5.3 (2.8–8.3)	19.1 (15.4–22.8)
IL-13	7.0 (3.5–10.2)	3.9 (1.7–6.0)	12.9 (9.6–16.4)	13.2 (10.2–15.8)	3.8 (2.5–6.0)	20.5(16.5–25.6)
IL-17A	1.52 (0.24–2.72)	0.90 (0.24–1.80)	0.27 (0.00–1.70)	1.84 (0.84–2.33)	0.65 (0.12–1.97)	19.5 (17.9–23.3)
IL-21	14.0 (3.4–16.7)	8.3 (2.6–14.3)	25.2 (23.1–27.0)	7.5 (3.1–11.9)	9.8 (5.4–13.6)	28.6 (27.1–30.6)
TNF-α	8.56 (5.70–13.46)	2.68 (1.48–3.93)	2.44 (1.84–2.44)	4.23 (2.82–7.96)	1.79 (0.39–2.70)	23.27 (22.03–26.62)
IFN-γ	1.63 (0.60–3.09	0.98 (0.60–2.52)	0.30 (0.00–1.21)	1.68 (0.94–2.40)	0.60 (0.60–1.76)	14.4 (10.1–23.7)
Eotaxin	9.8 (7.9–13.1)	4.5 (2.8–6.8)	3.5 (1.7–5.5)	7.0 (3.3–10.3)	5.1 (3.3–9.5)	13.7 (11.8–14.9)
RANTES	15.0 (10.7–18.1)	62.4 (32.4–90.7)	16.1 (13.7–22.4)	18.1 (13.0–34.5)	13.0 (8.6–18.1)	11.2 (10.4–17.4)

Values are presented as medians with interquartile ranges in parenthesis. All cytokine levels are presented as ng/mL. Gm csf, Granulocyte-macrophage colony-stimulating factor; TNF-α, Tumor necrosis factor alpha; IFN-γ, Interferon gamma; IL, interleukin

Discussion

There has recently been a renewed focus on understanding and classifying the diversity of immune responses in CRS by measuring granular inflammatory biomarkers of disease. In the present study, we utilized nasal mucus to compare the inflammatory cytokine profile of CCAD to other known CRSwNP disease states such as AFRS and AERD. Our data suggests that CCAD is characterized by a relatively homogenous inflammatory signature with low overall burden compared to other phenotypes of CRSwNP.

Previous studies have evaluated CCAD’s phenotypic differences when compared to other more established subtypes. CCAD has a strong association with inhalant allergies and low concurrent prevalence of asthma.^6,7 In our study, the CCAD cohort likewise had a significantly lower rate of asthma compared to other CRSwNP phenotypes. Rates of allergic rhinitis were similar to those seen in AERD/AFRS patients and significantly higher than in the CRSwNP NOS group. CCAD patients in our study also had significantly fewer prior surgeries compared to other CRSwNP subtypes. This finding is consistent with the previously discussed report by Steehler et al., which also showed a low incidence of prior surgery in this group.¹²

These clinical differences between CCAD and other CRSwNP groups are supported by our finding that CCAD may have a unique inflammatory signature characterized by reduced cytokine levels and relative homogeneity compared to other CRSwNP subtypes. When comparing univariate cytokines levels across CRSwNP subtypes, CCAD had lower IL-6, 8, INF-y, and eotaxin relative to all other groups and significantly lower type 2 cytokines (IL-5, IL-13) relative to both AERD and AFRS (Figure 1). CCAD patients were also found to have lower tissue eosinophil counts relative to AERD. While interpreting cytokines in isolation provides initial insights, the interaction and clustering of cytokines into various inflammatory pathways provides an opportunity to utilize multivariate approaches to understand how different disease states may arise as a result of known archetypal inflammatory programs. In the present study, we used a supervised discriminant analysis approach to perform dimension reduction of cytokines into three components that maximized separation between CCAD, AERD, AFRS, and CRSwNP NOS (Figure 2). In this analysis, CCAD was notably characterized by low values along both Component 1 and 2, which are associated with type 2 and type 1/3 cytokines, respectively. Interestingly, most of the variation in CCAD was along Component 3, which was primarily characterized by IL-12 and 21. In contrast, AFRS was relatively homogenous with respect to loadings along the type-1/3 axis, but was variable along the type-2 component, with the opposite finding for CRSwNP NOS. Consistent with prior work by our group, AERD patients showed notable variability along both axes, making this the most heterogenous CRSwNP subtype. CCAD, on the other hand, was the most homogenous group, with these patients generally lacking the inflammatory diversity that has been observed in CRSwNP.^31–37 Collectively, these results further support the hypothesis that CCAD is a clinically distinct diagnosis with a potentially unique pathomechanism.

As the inflammatory fingerprint of CRS continues to be revealed, clinically meaningful classification has become a major target of investigation. CRSwNP is not a homogenous inflammatory disease, but rather a complex clinical syndrome characterized by activation of multiple inflammatory pathways. Putative inflammatory CRS endotypes have recently been described and have the potential to guide targeted treatment decisions by directing patients towards targeted therapies based on their unique inflammatory characteristics. Improved understanding of the CCAD phenotype/endotype may help further characterize disease etiology and pathophysiology, while also assisting with the appropriate selection of effective therapies. Cluster analysis allows for potential identification of these endotypes. Our analysis identified six CRSwNP inflammatory disease clusters of which CCAD patients predominantly fell into those with low levels of type 2 inflammatory cytokines (Table 3). Lower overall inflammatory burden and a high rate of surgical success suggest that these patients should potentially be directed towards surgery and concomitant allergy management rather than costly interventions.

There are several pertinent limitations of this study that warrant discussion. First, all enrolled participants were seen at a single academic tertiary medical center within the southeastern part of the United States, which may not be representative of CCAD patients from other geographic regions. Secondly, our study group is entirely surgical patients and potentially not reflective of broader CRS groups that do not require surgical management. Finally, while our CCAD cohort represents one of the largest published studies to date, the small sample size limits the ability to detect smaller differences on univariate analysis. A follow-up multi-institutional study that includes patients from different regions and populations may aid in addressing these limitations.

Conclusion

CCAD is characterized by lower IL-6, IL-8, INF-g, and eotaxin relative to all other CRS subtypes and significantly lower levels of type 2 cytokines (IL-5, IL-13) relative to both AERD and AFRS. This suggests that CCAD may be a distinct subtype of CRSwNP with lower inflammatory burden that is relatively homogenous compared to the other defined phenotypes of CRSwNP.

ABBREVIATIONS

AERD

aspirin-exacerbated respiratory disesase

AFRS

allergic fungal rhinosinusitis

AR

allergic rhinitis

BMI

body mass index

CCAD

central compartment atopic disease

CRSwNP

chronic rhinosinusitis with nasal polyps

CRSwNP NOS

chronic rhinosinusitis not otherwise specified

CT

computed tomography

CV

coefficient of variation

ESS

endoscopic sinus surgery

IFN

interferon

IL

interleukin

MCT

Mucociliary clearance time

SNOT-22

22 item sinonasal outcome test

TNF

tumor necrosis factor