ANCA-negative EGPA: only eosinophils without vasculitis? Insights from anti-T2 biologics

The pathogenic role of p-ANCA in eosinophilic granulomatosis with polyangiitis (EGPA) is a long-standing matter of debate. In this work, we report our real-life experience with EGPA patients, treated with biologics targeting type 2 (T2)-eosinophilic inflammation (Mepolizumab, Benralizumab, Dupilumab). Interestingly, we observed EGPA extrarespiratory relapses only in p-ANCA-positive patients (2/5 cutaneous vasculitis, 3/5 constitutional symptoms), with new rise of p-ANCA and normal eosinophil blood count. Notably, revising our cohort with the new ACR 2022 criteria, these five patients were the only ones to satisfy the entry criterion of vasculitis’ defined diagnosis at disease onset. These observations may suggest that biologics, selectively turning off T2 inflammation, may have unmasked p-ANCA exclusive role in the pathogenesis of vasculitis in EGPA. Therefore, we raise the question whether EGPA vasculitis exists only in p-ANCA-positive patients, and whether p-ANCA-negative disease is “only eosinophils without vasculitis”.

Perinuclear anti-neutrophil cytoplasm antibodies (p-ANCA) are observed only in ~30-35% of cases, and their pathogenetic role is a long-standing matter of debate, while the role of eosinophils is well established.Therapies include corticosteroids, immunosuppressants and Mepolizumab (3,4), an anti-IL-5 drug directed against type 2 (T2)-eosinophilic inflammation previously approved for the treatment of severe eosinophilic asthma and recently approved also for severe CRSwNP, hypereosinophilic syndrome (HES) and EGPA.Other therapies used in T2 respiratory diseases, such as Benralizumab (anti-IL-5-R, approved for severe eosinophilic asthma) and Dupilumab (anti-IL4/IL-13Ra, approved for severe T2-eosinophilic asthma and CRSwNP), showed efficacy also in the context of EGPA (5,6).In this work, we discuss about p-ANCA status, reporting our experience with 13 EGPA patients treated with biologics targeting T2-eosinophilic inflammation from 2/2018 to 1/2023 at Internal Medicine Department, Marche University Hospital, Ancona, Italy.Informed consent was obtained for using patients' data.

Patients' baseline characteristics
All patients were classifiable as EGPA according to the old American College of Rheumatology (ACR) 1990 criteria (7).Diagnosis was based on clinical symptoms and radiological, laboratory and histopathological findings.All patients had asthma, CRS, hypereosinophilia and non-fixed pulmonary infiltrates.10/13 patients had extra-respiratory manifestations.Organ biopsies were performed in 8/13 patients, all showing extravascular eosinophils, with evidence of necrotizing leukocytoclastic vasculitis in 4/8.5/13 patients were p-ANCApositive.Patients' details are reported in Table 1.
After EGPA diagnosis, all patients were treated with high-dose corticosteroids, and 10/13 with immunosuppressants (Table 1).Complete remission of all clinical manifestations was achieved, with normalization of eosinophils, C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), and p-ANCA negativization in positive patients.>However, despite concurrent immunosuppressants' use, every attempt to reduce/discontinuate steroids resulted in new eosinophils' increase associated with CRS-asthma relapses (steroiddependent patients).
In 2018 and 2019, respectively, Mepolizumab and Benralizumab were approved for severe eosinophilic asthma.Since all our patients were affected by this condition, they started Mepolizumab (100 mg/ month) or Benralizumab (30 mg/2 months) even if, at that time, these biologics were not approved for EGPA.For this reason, no approval from research ethics board (REB) was needed.
At Mepolizumab/Benralizumab initiation asthma and CRS were the only active clinical manifestations of EGPA, with Birmingham Vasculitis Activity Score (BVAS) = 2.All patients had normal CRP and ESR, and p-ANCA were negative in previously positive patients.Baseline use of oral corticosteroids (OCS) is reported in Table 2.In 12/13 patients, immunosuppressants have been discontinued because of lack of efficacy on respiratory symptoms, remission of extrarespiratory manifestations, side effects.

Response to anti-T2 biologics
Both Mepolizumab and Benralizumab normalized eosinophils in all patients.A significative improvement in asthma control was observed, as documented by asthma control test and lung function tests (data not reported), and by OCS reduction/discontinuation (reported in Table 2).
For incomplete CRSwNP response, negatively impacting on asthma, three patients were switched to Dupilumab (300 mg/2 weeks) with complete control of both upper and lower respiratory symptoms and OCS withdrawal (Table 2).Even for Dupilumab treatment, no approval from REB was needed, since patients were affected by severe T2 asthma-CRSwNP (approved indications for Dupilumab prescription).Dupilumab-induced hypereosinophilia was observed, but without related clinical symptoms and not exceeding the described alert cut-off of 3000/mmc (8), except for patient #2 (see below).
During treatment with anti-T2 biologics, we observed EGPA extrarespiratory relapses only in the 5 formerly p-ANCA-positive patients, with new rise of p-ANCA, CRP and ESR.
At the time of relapse, patient #2 was under Dupilumab treatment, patients #4, #12 and #13 were under Benralizumab treatment, and patient #5 was under Mepolizumab treatment.Patients #4 and #5 experienced arthromyalgias under Benralizumab and Mepolizumab treatment, respectively, with eosinophils within the normal range.Since the symptoms were mild, episodic, and self-limiting, and values of p-ANCA, ESR and CRP were low, a "wait-and-see" strategy was adopted.Patient 4 maintained Benralizumab therapy, while patient 5, for incomplete CRSwNP response negatively impacting on asthma, was switched from Mepolizumab to Benralizumab and finally to Dupilumab.Over time, in both patients, symptoms remained stable (mild, episodic and self-limiting), while p-ANCA, ESR, CRP mantained at low levels, often with spontaneous normalization at subsequent checks.Patients #12 and #13 relapsed under Benralizumab treatment, with eosinophils within the normal range.Patient #12 experienced arthromyalgias and purpura of the lower limbs, while patient #13 experienced fever, arthromyalgias, and livedo racemose of the arms.Both patients were then treated with steroids and Rituximab (RTX), with clinical remission and p-ANCA persistent negativization, even after steroids withdrawal.
Patient #2 experienced three extrarespiratory relapses under Dupilumab treatment.At first relapse (arthromyalgias), after OCS withdrawal, both p-ANCA and eosinophils were elevated; particularly, Dupilumab-induced hypereosinophilia was severe (5000/mmc).Dupilumab was prudentially stopped and the patient was treated with steroids and resumption of Benralizumab.However, at steroid reduction, Benralizumab again proved not to be effective on CRSwNP, so a new attempt with Dupilumab was made, concurrently using medium-dose OCS (prednisone 10 mg/day) to cover a new possible severe hypereosinophilia.After two months, gradual OCS tapering was started, strictly checking the blood count.In this context, two subsequent relapses (arthromyalgias and fever) occurred.During these relapses, eosinophils were stable at a value of 500-1000 cells/ mmc, under low-dose OCS (prednisone 5 mg/day), while p-ANCA increased again.The second relapse was treated with temporary re-increase of OCS (10 mg/day) and azathioprine.For the third relapse, Rituximab was administered, with clinical remission and p-ANCA persistent negativization, allowing steroids discontinuation.After OCS withdrawal, eosinophils rose to 1500-2000/mmc, but the patient maintained asymptomatic.Relapses' managements are reported in Table 2.

Discussion
In patients undergoing Mepolizumab/Benralizumab, eosinophils were within the normal range during extrarespiratory relapses.Therefore, we hypothesized that extrarespiratory relapses might be due to the raised p-ANCA.In patient #2, undergoing  Dupilumab, both p-ANCA and eosinophils were elevated during the first relapse, making difficult to establish symptoms' cause (p-ANCA, hypereosinophilia, or both); particularly, Dupilumabinduced hypereosinophilia was severe (5000/mmc).However, during subsequent relapses, eosinophils were stable at a value of 500-1000/mmc, under low-dose OCS, while p-ANCA increased again.Furthermore, p-ANCA negativization following Rituximab coincided with clinical remission, although eosinophils rose to 1500-2000/mmc after OCS discontinuation.Therefore, also in patient #2, we finally hypothesized that extrarespiratory relapses might be due to the raised p-ANCA.Nevertheless, Dupilumabinduced hypereosinophilia could have played a concurrent role in first relapse, so that data relating this patient should be evaluated with great caution.Based on the above, in our p-ANCA-positive patients, EGPA seemed to have two distinct pathogeneses, sensitive to different therapies: T2-eosinophilic inflammation and p-ANCAmediated inflammation, responsive to anti-T2 biologics and immunosuppressants, respectively.This concept has been reported in literature: while T2eosinophilic inflammation would be responsible for asthma, CRS and eosinophilic organ damage (e.g.myocarditis, gastroenteritis), p-ANCA would underlie vasculitic manifestations (1, 2).However, the verbatim application of this strict dualism cannot always be translated into routine clinical practice (1), and the pathogenic role of p-ANCA is a long-standing matter of debate.An International Consensus reported their poor sensitivity/specificity in differentiating "vasculitic" from "eosinophilic" EGPA (9), stating that p-ANCA cannot guide treatment decision (9).
Nevertheless, in our experience, biologics, selectively turning off T2-eosinophilic inflammation and allowing OCS reduction/ withdrawal, may have unmasked p-ANCA exclusive role in the pathogensis of vasculitis in EGPA.Indeed, even if extrarespiratory relapses were mainly represented by aspecific inflammatory constitutional symptoms, 2/5 patients relapsed with frank cutaneous vasculitis (purpura, livedo racemosa).Moreover, the 5 relapsing p-ANCA positive patients were the only ones to have a defined diagnosis of vasculitis at disease onset (4 bioptic, 1 radiologic), so being the only ones to satisfy the entry criterion of the new ACR 2022 criteria (10).
These observations may raise the question whether EGPA vasculitis exists only in p-ANCA-positive patients, and whether p-ANCA-negative disease is only T2-eosinophilic inflammation without vasculitis.
Assuming this, our p-ANCA-negative patients could be considered as having "only" eosinophilic asthma+CRS and/or hypereosinophilic syndrome (HES).The likely bias in their EGPA misdiagnosis would lie in the old ACR 1990 criteria (7), not considering vasculitis' defined diagnosis.
In literature, however, evidence of organ vasculitis was reported (albeit rarely) also in p-ANCA negative patients.For these cases, some possible explanations could be hypothesized: 1) confounding manifestations, like neuropathy, that are strongly suggestive of vasculitis but can also be exclusively provoked by T2-eosinophilic inflammation (11-13); unfortunately, neuropathy-involved nerves are not always susceptible to diagnostic biopsy in routinary clinical practice; 2) vasculitis not ANCA-associated (drugs, infections, cryoglobulins) that may occur in patients with eosinophilic asthma and/or HES, leading to a clinical picture misdiagnosed as EGPA; 3) vasculitis-like histopathology of involved organs, due to eosinophil-derived vascular damage and thrombosis, but not properly to an autoimmune vasculitis; 4) false negativity of p-ANCA when patients are already undergoing steroid therapy.
A great limit of our hypotheses is that they derive from observations in a small cohort of EGPA patients.However, similar cases have been recently reported, describing the new onset of p-ANCA vasculitis in patients undergoing anti-T2 biologics for severe asthma.In these patients, despite remission of T2-eosinophilic inflammation, development of vasculitis was observed, accompanied by the new finding of p-ANCA positivity (14-18).

Conclusion
Biologics, selectively turning off T2-eosinophilic inflammation and allowing OCS reduction/withdrawal, may have unmasked p-ANCA exclusive role in the pathogenesis of vasculitis in EGPA.If confirmed by larger studies, these observations may have important therapeutic implications, possibly leading to a pathogenesis-based therapy of EGPA.

TABLE 2
Response to biological therapies.
At last follow-up (1/2023) or at drug stop.‡ With frequent courses at higher doses.§ During the previous biological treatment.The symbol "→" represents the switch from one therapy to another.AZA, azathioprine; ENT, ear, nose and throat; OCS, oral corticosteroids, expressed in prednisone dose equivalent; RTX, rituximab.