In office sampling of eosinophil peroxidase to diagnose eosinophilic chronic rhinosinusitis

Abstract

Background: Practical biomarkers for endotypic characterization of chronic rhinosinusitis (CRS) remain elusive, hindering clinical utility. Eosinophil peroxidase (EPX) is an enzyme released by activated eosinophils. The objective of this study was to evaluate a clinic EPX assay as a marker of eosinophilic CRS.

Methods: Subjects with and without CRS presenting to a tertiary care rhinology clinic were prospectively enrolled, and nasal cytology brushings were collected from the middle meatus during in-clinic nasal endoscopy. ELISA assay was used to quantify EPX levels, and a customized multiplex immunoassay was used to quantify inflammatory cytokine mediators. Findings were correlated with clinical data.

Results: Forty-two subjects were enrolled, including 31 CRS subjects and 11 controls. Median EPX levels were 125.0 ng/mL (standard deviation [SD] 1745.8) and 6.5 ng/mL (SD 99.0) for CRS group and controls, respectively (p = 0.003). EPX levels were associated with history of asthma (p = 0.015), allergies (p = 0.028), polyps (p = 0.0006), smell loss (p = 0.006), and systemic eosinophilia or elevated immunoglobulin E (p ≤ 0.0001). Twenty-eight subjects from both the CRS and control groups had prior pathology for comparison, with histologic confirmation of local tissue eosinophilia (>10 eosinophils/hpf) in 11 subjects. This subgroup had a median EPX level of 967.5 ng/mL compared to 10.6 ng/mL in 17 subjects without local tissue eosinophilia (p = 0.0008). EPX levels were positively correlated to interleukin-5 levels (p = 0.0005).

Conclusion: EPX levels can be measured via well-tolerated in-clinic collection of nasal mucus. EPX levels are associated with clinical markers of type 2 inflammation and tissue eosinophilia and may provide a valuable diagnostic tool to delineate eosinophilic CRS.

Keywords: chronic rhinosinusitis; endotype; eosinophil peroxidase; nasal polyp; type 2 inflammation.

FIGURE 1

(A) Eosinophil peroxidase (EPX) levels in subjects with prior pathology correspond to eosinophilic chronic rhinosinusitis (eCRS) as defined by >10 eosinophils per high powered field (p = 0.0011). On Dunn's multiple comparisons test, non‐CRS control versus non‐eCRS (p = 1.0), non‐CRS control versus eCRS (p = 0.0032), and non‐eCRS versus eCRS (p = 0.024). (B) This receiver operating characteristic (ROC) curve illustrates the diagnostic ability of EPX levels to predict the presence of eCRS, when eCRS is defined by >10 eosinophils per high powered field on histopathology. The curve is plotted with the true positive rate (sensitivity) on the y‐axis and the false positive rate (1 ‒ specificity) on the x‐axis. The diagonal line represents the performance of a random classifier, serving as a baseline. The area under the curve (AUC) is 0.89, indicating a good level of discrimination between eCRS and non‐eCRS cases. The optimal cutoff point, determined using Youden's J statistic, is an EPO level of 118.35. At this threshold, the sensitivity and specificity are balanced, providing a reliable diagnostic marker for eCRS.

FIGURE 2

(A) Eosinophil peroxidase (EPX) concentration levels within all subjects correspond to percent eosinophils (Spearman r = 0.55, p = 0.0005) and percent macrophages (Spearman r = 0.46, p = 0.0041) within nasal cytology specimens. Percent neutrophils did not correlate with EPX level (Spearman r = 0.11, p = 0.5020). (B) When grouped by EPX categorization (control, low, or high EPX), the high EPX group was associated with higher percent eosinophils than the non‐chronic rhinosinusitis (CRS) control group (p = 0.004), and the low EPX group was associated with higher percent neutrophils than the non‐CRS control group (p = 0.041).

FIGURE 3

(A) Eosinophil peroxidase (EPX) concentration levels in nanograms per millileter (ng/mL) within all subjects is associated with interleukin‐5 (IL‐5) concentration in pg/mL (Spearman r = 0.52, p = 0.0005). (B) Nine cytokines associated with chronic rhinosinusitis (CRS) endotypes were evaluated for association with EPX categorization (control, low, or high EPX). The high EPX CRS group represents CRS patients with EPX levels ≥118.35 ng/mL and the low EPX CRS group represents CRS patients with EPX levels <118.35 ng/mL. Only interferon‐gamma (IFN‐γ) and IL‐5 were associated with EPX category (p = 0.022 and 0.012, respectively). On Dunn's multiple comparisons test, IFN‐γ was significantly different between low and high EPX CRS groups (p = 0.012). IL‐5 was significantly different between non‐CRS control and high EPX CRS groups (p = 0.010) as well as between low and high EPX CRS groups (p = 0.037). Asterisk (*) indicates decreased values by the power of 1000 to represent graphically with other cytokines.