Objective: To evaluate the ACR/EULAR 2022 criteria for ANCA-associated vasculitides (AAV) classification and compare them with the European Medicines Agency (EMA) algorithm and with classification based only on ANCA serology.
Methods: In the analysis, 374 cases (47% female) were classified according to the EMA algorithm, ANCA serology and ACR/EULAR criteria. The agreement rate was calculated using the kappa (κ) statistic.
Results: Under EMA, 192 patients were classified as granulomatosis with polyangiitis (GPA), 159 as microscopic polyangiitis (MPA) and 23 as eosinophilic granulomatosis with polyangiitis (EGPA). The ACR/EULAR criteria classified 199 patients as GPA, 136 as MPA and 22 as EGPA. Four patients (1.1%) met criteria of two disease categories, and 13 (3.5%) were unclassifiable. The observed agreement between EMA and ACR/EULAR was 85% for GPA, 75% for MPA and 96% for EGPA. The unweighted κ statistic was 0.66 (95% CI: 0.60, 0.74). Of the 188 PR3-ANCA positive patients, 186 (98.9%) were classified as GPA using ACR/EULAR criteria, and 135 of 161 (83.9%) MPO-ANCA positive patients were classified as MPA. With a classification solely based on ANCA specificity, agreement with ACR/EULAR was 99% for GPA and 88% for MPA.
Conclusions: EMA and ACR/EULAR classification give similar results. A small proportion of patients cannot be classified or fall into two categories. Some patients exhibiting granuloma, a key feature of GPA, are nevertheless classified as MPA, conflicting with the current view of histopathology of AAV. There is high agreement of ANCA-based classification with that of ACR/EULAR, reflected in the considerable weight granted to ANCA in the new criteria. These crucial elements within the new criteria necessitate a consensus discussion among field experts.
Keywords: ANCA-vasculitis; GPA; MPA; classification.
© The Author(s) 2023. Published by Oxford University Press on behalf of the British Society for Rheumatology.