• systemic vasculitis
• granulomatosis with polyangiitis
• epidemiology

The new classification criteria for antineutrophil cytoplasmic antibodies (ANCA)-associated vasculitis (AAV) have been eagerly anticipated since the publication of the provisional criteria in 20171 and thus they are quite welcomed.2 3 Limitations of the previous criteria used for these conditions include the low sensitivity of the 1990 American College of Rheumatology (ACR) classification criteria, the fact that the patient’s ANCA status had not been included, that there were no previous classification criteria for microscopic polyangiitis (MPA), and the fact that the number of criteria rather than a weighted criteria were used. Furthermore, there was lack of satisfaction among professionals about the 1990 ACR criteria and the Chapel Hill Consensus Conference (CHCC) definitions: 43% and 68% were dissatisfied with these criteria for granulomatosis with polyangiitis (GPA), and 76% and 36% for eosinophilic GPA (EGPA), respectively.4

Given these considerations, we read with great interest the articles on the new classification criteria for GPA and MPA by Robson et al and Suppiah et al, respectively,2 3 which have shown to have good sensitivity and specificity. These criteria have been validated by the Diagnostic and Classification Criteria in Vasculitis (DCVAS) project which recruited 6991 participants (among patients with AAV and comparator diseases) from 132 sites in 32 countries around the world. Surprisingly, however, only two centres from Latin America, both from Mexico, were included in this validation process. This limited representation of the subcontinent in the validation process is of some concern. Latin America is a large and heterogeneous region in terms of socioeconomic, environmental and genetic features, which might impact on the presentation of some diseases including the AAV. Of interest, we have noted some of these features in a recent systematic literature review: although Peru and Mexico have similar ethnic ancestral background, MPA was found to be more common in Peru, and GPA in Mexico.5

We thus decided to apply the new criteria to patients from our Almenara vasculitis cohort, a predominantly Mestizo Peruvian cohort to assess their performance.6 To this end, we classified all patients using the ‘former criteria’ (which included the 1990 ACR criteria for GPA and the 1994 CHCC definition for MPA) and, then classified all patients using the new criteria. Of note, EGPA was not included given its infrequency in our cohort. The sensitivity and specificity of these new criteria are depicted in table 1.

As to MPA, our findings are very similar with those of Suppiah et al’s; however, for GPA, the sensitivity in our study was lower than Robson et al’s (80.5% vs 93.0%). This could be explained by the fact that some of the data in our cohort (clinical features, ANCA status, images, biopsy results) could have been missing given that these were gathered from the medical records rather than from actual visits; likewise, some patients may not have undergone adequate imaging or histopathological studies and we used the clinician’s diagnosis as the gold standard. On the other hand, the new criteria had better agreement (κ: 0.653) than the former criteria (κ: 0.305) for the clinical diagnosis of AAV. Furthermore, in our study the sensitivity of the 1990 ACR criteria for GPA was low (68.3%).

In conclusion, the new criteria for AAV, particularly for MPA and GPA, have good performance in a Latin-American population, mainly constituted by Mestizo patients, with comparable findings to those from Suppiah et al and Robson et al. We think many Latin-American centres as ours could be considered for new studies and protocols addressed by DCVAS and thus reduce the knowledge gap about AAV in this region.