Objective: To investigate the clinicopathological features of eosinophilic granulomatosis with polyangiitis (EGPA)-associated neuropathy with a focus on the presence or absence of anti-neutrophil cytoplasmic antibodies (ANCAs). Methods: We examined the clinical features and pathological findings of sural nerve biopsy specimens from 82 patients with EGPA-associated neuropathy. Of these patients, 32.9% were myeloperoxidase (MPO)-ANCA-positive, and 67.1% were MPO-ANCA-negative. PR3-ANCA was negative in all of 78 examined patients. Results: Upper-limb symptoms were more frequently reported as initial neuropathic manifestations in the MPO-ANCA-positive group than in the MPO-ANCA-negative group (44.4% vs. 14.6%, p < 0.01). The serum levels of C-reactive protein were significantly higher in the MPO-ANCA-positive group than in the MPO-ANCA-negative group (p < 0.05). Sural nerve biopsy specimens showed findings suggestive of vasculitis (i.e., destruction of vascular structures) in epineurial vessels; these results were more frequently seen in the MPO-ANCA-positive group than in the MPO-ANCA-negative group (p < 0.0001). Conversely, the numbers of eosinophils in the lumen of the epineurial vessels (p < 0.01) and epineurial vessels occluded by intraluminal eosinophils (p < 0.05) were higher in the MPO-ANCA-negative group than in the MPO-ANCA-positive group. Furthermore, the incidence of eosinophil infiltration in the endoneurium was higher in the MPO-ANCA-negative group than in the MPO-ANCA-positive group (p < 0.01). Conclusions: This study suggests that the pathogenesis of EGPA comprises at least two distinct mechanisms: ANCA-associated vasculitis resulting in ischemic effects and inflammation, which is prominent in MPO-ANCA-positive patients, and eosinophil-associated vascular occlusion leading to ischemia and eosinophil-associated tissue damage, which is conspicuous in MPO-ANCA-negative patients.

Supplementary figure e-1 

Methods for quantitative analysis of the sural nerve.

Transverse sections. Hematoxylin and eosin staining. (A) A whole transverse section of the sural nerve scanned by a virtual slide system (VS120-S5, Olympus, Tokyo, Japan) using the ×20 objective lens is shown. (B) Epineurial area and endoneurial area measured using ImageJ software version 1.50i (NIH, Bethesda, Maryland, USA) are shown as Ep and En, respectively. (C) Epineurial vessels counted using the same software are shown as black dots. Scale bars = 500 μm.