We randomly selected 300 patients with systemic lupus erythematosus, rheumatoid arthritis and scleroderma, as well as 50 with primary SS. SS diagnosis was established by two independent rheumatologists and was based on the combination of symptoms, signs, diagnostic tests and medical chart review. We evaluated the fulfillment of the 2002 AECG, 2012 ACR and 2016 ACR/EULAR criteria, and their performance using as the gold standard the clinical diagnosis.
We identified 154 patients with a clinical (definitive/probable) SS diagnosis, 95 patients (61.7%) fulfilled the AECG, 96 patients (62.3%) the ACR and 90 (58.4%) the 2016 ACR-EULAR criteria. Among the subset with definitive SS clinical diagnosis (n=99), 83 patients (83.8%) fulfilled the AECG, 77 (77.7%) the ACR and 79 (79.7%) the 2016 ACR-EULAR criteria. The concordance rate between the clinical diagnosis (definitive/probable) and the AECG, ACR and 2016 ACR-EULAR criteria was Κ=0.58, Κ=0.55 and Κ=0.60, respectively. The 2016 ACR-EULAR criteria showed the best AUCs results (0.87 definitive/probable diagnosis, 0.90 definitive diagnosis), followed by the AECG (0.82 definitive/probable diagnosis, 0.85 definitive diagnosis) and ACR (0.80 definitive/probable diagnosis, 0.79 definitive diagnosis) criteria. As a sensitivity analysis, the results were similar when excluding patients with primary SS.
Our study provides further evidence that the 2016 ACR/EULAR criteria are applicable in the setting of secondary SS.