The assessment of autoinflammatory disease classification criteria (Eurofever/PRINTO) in a real-life cohort.

Abstract

The aim of the study was to determine the sensitivity and specificity rates of Eurofever/PRINTO autoinflammatory recurrent fever classification criteria with real-life data in patients with an autoinflammatory disease. A total of 119 patients were included in the study. Based on clinical symptoms, they were divided into four subgroups: cryopyrin-associated periodic syndromes (CAPS), TNF receptor-associated periodic syndrome (TRAPS), mevalonate kinase deficiency (MKD), and syndrome of undifferentiated recurrent fever (SURF) using the Eurofever/PRINTO clinical classification criteria. In the last step, the patients were re-evaluated in the light of genetic results and their final diagnosis was reached. A total of 119 patients, including 37 CAPS, 13 TRAPS, 8 MKD, 39 SURF, 14 NLRP12-related autoinflammatory disease (NLRP12-AID), and 8 familial Mediterranean fever (FMF) patients were evaluated in the study. While the sensitivity of the new clinical Eurofever/PRINTO criteria was 48% for CAPS, 77% for TRAPS, 87.5%for MKD, and the specificity of the clinical criteria was 86% for CAPS, 85% for TRAPS, and 60% for MKD. The sensitivity of the new mixed (genetic plus clinical variables) Eurofever/PRINTO criteria was 27% for CAPS, 61% forTRAPS, 85% for MKD, and the specificity of the mixed criteria for each group was 100%. We found the sensitivity of the Eurofever/PRINTO classification criteria to be low as genotypic changes between populations cause phenotypic differences. For this reason, we think that patient-based evaluation is correct rather than standard classification criteria in real life. Key-points • In systemic autoinflammatory diseases, common variants in the populations may alter the phenotype, and making it difficult to classify some patients with the current classification criteria. • In populations with common genetic variants, the classification criteria should be modified according to the clinical phenotype.