Performance of the 2016 ACR-EULAR classification criteria for primary Sjogren's syndrome in a Korean cohort

Abstract

This study compared the performance of the newly proposed 2016 American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) criteria to the 2002 American-European Consensus Group (AECG) and 2012 ACR classification criteria for primary Sjogren's syndrome (pSS) in well-characterized Korean patients. Patients with pSS from 12 university-affiliated hospitals in Korea were enrolled from October 2013 to January 2017. Clinical and laboratory data were reviewed. For the validation set, patients who underwent evaluation tests to rule out pSS at Seoul St. Mary's hospital from November 2016 to December 2017 were analyzed. Baseline registry data were available in 458 patients, and 328 patients had sufficient data to determine the fulfillment of each criteria set. All three sets of criteria were met by 307 patients (93.6%). The newly proposed 2016 ACR/EULAR criteria were met by 325 patients (99.1%). The 2002 AECG and 2012 ACR criteria were met by 325 (99.1%) and 310 patients (94.5%), respectively. In a validation cohort consisting of 161 patients with pSS-related symptoms/signs, the sensitivity and specificity of the 2016 ACR/EULAR criteria were 100% [95% confidence interval (CI), 96.11-100.00] and 81.8% [95% CI, 76.15-94.26], respectively. Agreement between the 2016 criteria and 2012 or 2002 criteria was high (Cohen's kappa 0.736 and 0.769, respectively). The newly proposed 2016 ACR/EULAR criteria were met by most patients diagnosed with pSS according to previous criteria and showed higher sensitivity and lower specificity compared with both previous criteria sets.