Evaluating the Applicability of the EULAR/ACR 2019, SLICC 2012, and ACR 1997 Classification Criteria for Systemic Lupus Erythematosus in Children: A Multicenter Study.

Abstract

Systemic lupus erythematosus (SLE) is a chronic autoimmune disease that affects approximately 15% to 20% of patients during childhood at initial onset. In childhood-onset SLE (cSLE), clinical manifestations are not typical in the early stages; therefore, cSLE-specific classification criteria are lacking, making diagnosis difficult. To evaluate suitable classification criteria for these patients, we conducted a multicenter cohort study to compare the practicability of the European Alliance of Associations for Rheumatology (EULAR)/American College of Rheumatology (ACR) 2019, Systemic Lupus International Collaborating Clinics (SLICC) 2012, and the ACR1997 classification criteria for SLE. There were patients from 3 different regions, including children with cSLE (n = 348), a pediatric control group (n = 59), adults with SLE (n = 80), and an adult control group (n = 76). Sensitivity, specificity, and area under the curve (AUC) values for the EULAR/ACR 2019, SLICC 2012, and ACR 1997 classification criteria were calculated. Serial and parallel tests were conducted, and data were compared after adjusting for the EULAR/ACR 2019 classification criteria score thresholds. There were 348 cases with a firmly established clinical diagnosis of cSLE (83.62% female) included. Among children with SLE, the ACR 1997 criteria showed the highest specificity (98.31%, 95% CI 90.9-100%), whereas SLICC 2012 criteria had the highest sensitivity (94.54%, 95% CI 91.6-96.7%). In comprehensive comparisons, the EULAR/ACR 2019 criteria yielded the highest AUC (0.944) and Youden index (0.89) values. Parallel testing using the SLICC 2012 or EULAR/ACR 2019 criteria for cSLE achieved the highest AUC (0.962) and increased sensitivity to 99.14%. Finally, a EULAR/ACR 2019 score of 10 (cSLE cutoff) produced the highest AUC (0.953, 95% CI 0.93-0.97). The EULAR/ACR 2019 criteria were the most appropriate for diagnosing cSLE. Moreover, parallel testing using the SLICC 2012 or EULAR/ACR 2019 criteria enhanced diagnostic sensitivity. In addition, a total EULAR/ACR 2019 score of ≥ 10 was appropriate for classifying cSLE. Our findings provide a basis for determining the most appropriate diagnostic strategy for children with SLE.