Validation of the PEDiatric Behçet’s Disease classification criteria: an evidence-based approach

Objectives: To validate the PEDiatric Behcet's Disease classification criteria (PEDBD) with an evidence-based approach.Methods: A total of 210 pediatric patients [70 Behcet's disease (BD), 40 periodic fever, aphthous stomatitis, pharyngitis, adenitis, 35 familial Mediterranean fever, 26 hyper-IgD syndrome, 22 TNF-receptor associated periodic fever syndrome, 17 undefined recurrent fevers] were randomly selected from the Eurofever Registry. A set of 11 experienced clinicians/researchers blinded to the original diagnosis evaluated the patients. Using the table consensus as gold standard (agreement >= 80%), the PEDBD, ISG and ICBD criteria were applied to BD patients and to confounding diseases with other autoinflammatory conditions in order to define their sensitivity, specificity and accuracy.Results: At the end of the third round, a consensus was reached in 139/210 patients (66.2%). The patients with a consensus >= 80% were classified as confirmed BD (n = 24), and those with an agreement of 60-79% as probable BD (n = 10). When comparing these patients with the confounding diseases group, an older age at disease onset, the presence of oral and genital ulcers, skin papulo-pustular lesions, a positive pathergy test and posterior uveitis were BD distinctive elements. The ISG, ICBD and PEDBD criteria were applied to confirmed BD and to the confounding disease group, showing a sensitivity of 0.50, 0.79 and 0.58, a specificity of 1.00, 0.97 and 0.99, and an accuracy of 0.91, 0.94 and 0.92, respectively.Conclusions: The PEDBD criteria were very specific, while the ICBD were more sensitive. The complexity of childhood BD suggests larger prospective international cohorts to further evaluate the performance of the criteria.