LAUSR

To validate the 2019 EULAR/ACR classification criteria for SLE in Hong Kong Chinese patients and compare its performance with the 2012 Systemic Lupus International Collaborating Clinics (SLICC) and 1997 American college of rheumatology (ACR) criteria.We retrospectively reviewed the medical records of consecutive patients who attended the Rheumatology clinics in Tuen Mun and Pok Oi hospitals between May and September 2019. Patients with anti-nuclear antibody (ANA) ≥1:80 were included and patients with ANA <1:80 or no ANA results were excluded. Patients were evaluated and cross-checked for the fulfilment of the 1997 ACR, 2012 SLICC and 2019 EULAR/ACR criteria by two investigators (YKC,CL). Medical records were then reviewed by an expert panel consisting of 3 senior rheumatologists, who were blinded for the results of the criteria evaluation, for a diagnosis of SLE based on the clinical judgement and therapeutic decisions. Teleconferences were arranged by the panel to discuss the discrepancies of the final diagnosis and agreement was made by voting. The three SLE criteria were evaluated against the clinical diagnosis of SLE as judged by the expert panel on the sensitivity and specificity, which was calculated by 2x2 contingency tables (“condition positive” = clinical diagnosis of SLE; “test positive” = criteria positive for SLE) with standard formulas (sensitivity = true positive/[true positive + false negative]; specificity = true negative / [true negative + false positive]). Receiver operating characteristic (ROC) curve was used to study the optimal cut-off points from the EULAR/ACR criteria for the highest summation of specificity and sensitivity.3967 patients were screened; 1542 patients who were positive for ANA (≥1:80) were included (88.3% women). The mean age of these patients at first rheumatology clinic attendance was 45.6±15.0 years and the duration of follow-up was 7.5±7.0 years. A total of 567 patients were judged to have SLE by the expert panel (discrepancy of clinical diagnosis in 135 patients resolved with voting). The sensitivity and specificity of the three SLE classification criteria in our patients are listed in Table 1. ROC analysis showed that the best cut-off for a clinical diagnosis of SLE using the EULAR/ACR criteria was 10 points (area under the curve [AUC] 0.977; sensitivity 89.2% and specificity 89.6%). Similar figures were obtained for subgroups of patients stratified by gender and different age ranges.In our cohort of Hong Kong Chinese patients, the 2019 EULAR/ACR criteria is more sensitive but less specific when compared with 1997 ACR criteria for classifying SLE. On the other hand, the EULAR/ACR criteria is less sensitive but more specific than the 2012 SLICC criteria. The specificity of the EULAR/ACR criteria for SLE is higher in male than female patients. In our patients older than 50 years, the EULAR/ACR criteria is less sensitive but more specific for a classification of SLE. Overall, the performance of the EULAR/ACR criteria for a diagnosis of SLE in our study is similar to that reported in recent Asian studies although the sensitivity is lower, which may be related to the inclusion of ANA+ patients only.Classification criteriaSensitivitySpecificity1997 ACR85.9%94.4%2012 SLICC97.5%86.4%2019 EULAR/ACR with 10 points as cut-off89.2%89.6%2019 EULAR/ACR with 9 points as cut-off93.6%68.7%2019 EULAR/ACR with 11 points as cut-off86.9%92.4%2019 EULAR/ACR with 10 points as cut-off (men)88.9%94.5%2019 EULAR/ACR with 10 points as cut-off (women)89.2%88.8%2019 EULAR/ACR with 10 points as cut-off (age >50 years)78.7%94.1%2019 EULAR/ACR with 10 points as cut-off (age ≤50 years)91.7%84.1%None declared