Fractional exhaled nitric oxide as a novel biomarker of relapse in eosinophilic granulomatosis with polyangiitis

To the Editor Eosinophilic granulomatosis with polyangiitis (EGPA) is a ­multisystem syndrome characterized by vasculitis, asthma, and other manifestations secondary to peripheral and tissue eosinophilia.1 The recent success of the anti–IL-5/5R therapies in controlled trials of EGPA has confirmed the central role of the eosinophil and has offered the prospect of steroid-free disease remission.2-4 The blood eosinophil count (BEC) is recognized as a cheap, accessible biomarker of disease activity in EGPA, but its near-complete depletion following initiation of the anti-IL-5R benralizumab means an alternative, noninvasive biomarker of eosinophilic inflammation is needed to support treatment decisions. In asthma, fractional exhaled nitric oxide (FeNO) is a widely used and reliable biomarker of type 2 (T2) airway inflammation. Although levels can be partially suppressed on anti-IL-5/5R mAbs,5 increased levels provide clinicians with a means to identify residual steroid-responsive, T2 inflammation that may require additional treatment. To date, FeNO has not been studied as a potential biomarker in EGPA, and it remains unknown whether FeNO levels change during EGPA relapses. To investigate the potential utility of FeNO as a biomarker of EGPA activity in patients treated with benralizumab, we analyzed FeNO, BEC, and serum IL-5 levels in patients with EGPA established on benralizumab at our regional EGPA referral center at Guy’s Hospital, London, United Kingdom. Clinical assessments were collected at 3 time points: baseline (stable state), relapse, and recovery. Eosinophilic granulomatosis with polyangiitis relapse was defined as the recurrence of acute clinical signs or symptoms after a period of stability requiring an increase in daily oral corticosteroid (OCS) dose. The recovery visit reflected the subsequent clinical assessment. There was no consistent time interval between the 3 visits, but the relapse and recovery visits all occurred within a 12-month period. Some of the results of this analysis have been reported in abstract form.6 All patients had comorbid severe asthma and were receiving the asthma dose of benralizumab (30 mg subcutaneously every 8 weeks). Clinical data were collected from ­electronic health ­records, and patients provided written informed ­consent (REC ref: 14/LO/1699). Blood levels, including serum IL-5 (Bio-Techne SPCKA-PS005696), were measured at Guy’s Hospital. Data were analyzed using GraphPad Prism v10.1.1 (GraphPad Software) and reported as mean (SD) or median (IQR). Parametric and nonparametric data were analyzed using the paired t test and Wilcoxon test, respectively. Correlations of nonparametric variables were performed with the Spearman r (rs) test. Two-tailed ­values of P < .05 were considered statistically significant. Eighty-six patients with EGPA treated with benralizumab were identified. Biomarker measures were available for 46 EGPA relapses from 35 patients (mean age 48.7 years, female 46%, Antineutrophil Cytoplasmic Antibodies positive 34%). At stable state visit, 27 of 35 (77%) were on maintenance OCS (median prednisolone dose 7.0 mg/d [5.0-10.0]). Eleven of 35 (31%) were concomitantly receiving additional immunosuppressants (mycophenolate, methotrexate, or azathioprine). Of the 46 relapses, 30 involved worsening asthma/respiratory symptoms, of which 23 of 46 (50%) only involved a respiratory deterioration. Of the ­remaining 23 relapses, 20% involved worsening sinonasal disease: 15% cardiac, 15% nonspecific arthralgia, and 4% gastrointestinal. Three relapses required hospitalization. The increase in OCS dose at relapse was to 30.0 mg/d (20.0-40.0). Five relapses were treated with additional rituximab. The inhaled corticosteroids dose was stable throughout the observational period. At relapse, the median BEC did not change and remained ­fully suppressed on benralizumab at 0.0 (0.0-0.0) cells/µL. Median FeNO increased from 52 parts per billion (ppb) (31-116) to 84 ppb (40-162) at relapse (P < .01), returning to baseline levels at recovery: 51 ppb (31-88) (P < .01) (Figure 1). A rise in FeNO was observed for both airway-dominant (P = .008) and non-airway-dominant (P = .03) relapses. The OCS dose at recovery was similar to baseline (9.0 mg/d vs 7.0 mg/d, P = .19). Median serum IL-5 (n = 24) at baseline, relapse, and recovery was 9.8 pg/mL (2.9-17.4), 14.9 pg/mL (5.6-40.2), and 9.4 pg/mL (2.5-19.4), respectively, with a significant increase at relapse compared to baseline (P = .01). Overall, the median increase in FeNO from baseline to relapse was 22.5% (0-60). Serum IL-5 correlated with FeNO (rs = 0.43, P = .002) when ­combining baseline and relapse measurements (Figure 2). This correlation remained significant when relapse values were analyzed alone (rs = 0.42, P = .04) but not at baseline alone (P = .07). The changes in FeNO did not correlate with any changes in C-reactive protein, anti-MPO titers, or FEV1. Fractional exhaled nitric oxide at baseline (stable state on benralizumab), relapse, and recovery visits (n = 46). Correlation between fractional exhaled nitric oxide and serum IL-5 (both expressed as log10) at baseline (stable state on benralizumab) and relapses (n = 48). These data represent the first report of serial FeNO levels in EGPA to date. This extends our previously published finding that increased FeNO at 1 year on benralizumab is associated with a ­failure to achieve clinical remission and remain relapse-free.4 Taken together, our observations suggest clinically relevant T2 pathway activation in patients treated with benralizumab. Although FeNO most closely reflects IL-13, elevated levels ­likely reflect upstream cell activation, including ILC2s via epithelial alarmins such as TSLP and IL-33.7 We speculate that these cells are also the source of the increased IL-5 we observed at relapse. It is also possible that non-T2 pathways may, in part, underpin the rise in FeNO at relapse. Understanding the nature of this residual T2 and non-T2 (eg, Th1/Th17) inflammation is an important area of future research in EGPA as anti-TSLP and/or anti–IL-33 may prove effective options for patients who continue to relapse on anti–IL-5/5R therapies.8 A similar dissection of T2 inflammation into eosinophilic and noneosinophilic components has been recently ­demonstrated in asthma. As with EGPA, many patients with severe asthma achieve clinical remission following eosinophil depletion with benralizumab,9 whereas others require a switch to therapies that target noneosinophilic components of T2 inflammation, such as anti–IL-4R and anti-TSLP.10,11 As with our current analysis, elevated FeNO most clearly identifies responders to these treatments. While we acknowledge this report is limited by a relatively small sample size and retrospective design, we believe the significant correlation between FeNO levels and serum IL-5 adds further mechanistic plausibility to our observations. In summary, our data suggest that in the absence of a recordable BEC due to the mechanism of action of benralizumab, FeNO offers a cheap, noninvasive biomarker for EGPA activity. We believe FeNO may assist clinicians in determining whether uncontrolled T2 inflammation may be a key driver of residual disease ­activity and could support future biomarker-led treatment ­decisions for these patients. Larger, prospective studies are now needed to ­explore this possibility F.C.: conceptualization, writing-original draft and formal analysis; A.-C.M.-P., A.D.R., J.G., F.H. and P.K.: investigation; J.D., A.M.N. and D.J.J.: conceptualization, investigation, supervision and validation. Supplementary material is available at American Journal of ­Respiratory and Critical Care Medicine online. F.C. has received support to attend conferences or honoraria for lectures from AstraZeneca, Chiesi, GlaxoSmithKline, and Sanofi Regeneron. F.C. is recipient of the Italian Respiratory Society (IRS) and the European Academy of Allergy and Clinical Immunology (EAACI) Research Fellowship. A.-C.M.-P. reports consultation fees or research support from AstraZeneca,GlaxoSmithKline and Sanofi/Regeneron. J.G. reports honoraria for lectures and educational events from AstraZeneca and Sanofi and support for attending meetings from AstraZeneca. J.D. reports honoraria from ­AstraZeneca and Chiesi and support to attend conferences from AstraZeneca. A.M.N. reports payment or honoraria for lectures and educational events from AstraZeneca, GlaxoSmithKline, Chiesi and Sanofi and support for attending conferences from AstraZeneca, Chiesi, and GlaxoSmithKline. D.J.J. reports research grants to the department from AstraZeneca and GSK; consulting fees from ­AstraZeneca, GSK, and Sanofi Regeneron; and payments or ­honoraria for lectures and educational events from AstraZeneca, GSK, and Sanofi Regeneron. The other authors report no relevant ­conflicts of interest. Please see the ICMJE disclosure forms, which have been provided as supplementary material. No artificial intelligence tools were used in writing this manuscript. The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.