Evaluating performance of the 2019 EULAR/ACR, 2012 SLICC, and 1997 ACR criteria for classifying adult-onset and childhood-onset systemic lupus erythematosus: A systematic review and meta-analysis

Introduction: The American College of Rheumatology (ACR) 1997, Systemic Lupus International Collaborating Clinics (SLICC) 2012, and European League Against Rheumatism (EULAR)/ACR 2019 SLE criteria are often used to classify patients with adult-onset and childhood-onset systemic lupus erythematosus (SLE) in clinical practice because there are currently no diagnostic criteria for SLE. However, there is scarce evidence regarding which criteria are best for diagnosing patients with adult-onset and childhood-onset SLE.Methods: We searched Medline and Scopus databases for English-language articles from inception through October 2021. Data were extracted from the included publications by two independent reviewers. We performed bivariate meta-analysis with a random-effects model to pool diagnostic parameters. Meta-regression and subgroup analyses were performed to explore heterogeneity sources. We used network meta-analysis to compare diagnosis performance among the three criteria and ranked them in descending order. Publication bias was assessed using Deeks' funnel plot.Results: We included 29 studies for systematic review and meta-analysis. Of these, 18 studies involved adult-onset SLE and 11 studies involved childhood-onset SLE. The pooled sensitivities of the three criteria for diagnosis of adult-onset SLE were comparable between SLICC 2012 and EULAR/ACR 2019 [95.86, 95% confidence interval (CI) 92.28-97.81 vs. 94.79, 95% CI 92.03-96.63]; pooled specificity was highest in ACR 1997 (92.24, 95% CI 87.06-95.46). In childhood-onset SLE, pooled sensitivity was highest in SLICC 2012 (93.76, 95% CI 89.45-96.39), and pooled specificity was highest in ACR 1997 (95.89, 95% CI 91.73-98.00). In network meta-analysis, the pooled diagnostic odds ratio ranked highest for EULAR/ACR 2019 (131.570, 95% CI 61.50-281.47) in adult-onset SLE and ranked highest for SLICC 2012 (191.07, 95% CI 76.06-480.01) in childhood-onset SLE. Deeks' funnel plot showed no publication bias.Conclusion: We found that the diagnostic performance of the ACR 1997, SLICC 2012, and EULAR/ACR 2019 criteria differed between adult-onset and childhood-onset SLE. EULAR/ACR 2019 performed best for adult-onset SLE and SLICC 2012 was best for childhood-onset SLE in classifying patients with SLE.