Application of the SLERPI scoring system as a classification tool in systemic lupus erythematosus and its performance in early disease recognition.

The SLE Risk Probability Index (SLERPI) was developed as diagnostic aid in early disease. Previous studies applied it as a classification tool in established SLE, even when analyzing its performance in early disease. We assessed the performance of SLERPI as a classification tool but also at initial assessment, in antinuclear antibody positive patients. In a cross-sectional setting, we enrolled 104 consecutive SLE patients and 104 ANA-positive controls with inflammatory rheumatic conditions. The performance of SLERPI and the EULAR/ACR-2019 criteria was analyzed cumulatively and at initial assessment. Different SLERPI positivity thresholds were evaluated. The cumulative SLERPI score was higher in the SLE group (14.5; IQR 12.0-18.6) compared to controls (7.0; IQR 5.0-8.3). It exhibited a sensitivity of 98.1%, slightly higher compared to EULAR/ACR-2019 (97.1%), and a low specificity (54.8%), lower compared to EULAR/ACR-2019 (56.7%). At initial assessment, SLERPI was higher among SLE patients (median 10.0; IQR 8.5-12.6) compared to controls (median 5.0, IQR 4.5-7.0). Its sensitivity was 83.7% and specificity 76.0%, both higher compared to the EULAR/ACR-2019 (80.8 and 75.0%). Areas under the curve (AUC) for both criteria were 0.898 initially. Cumulatively, the AUC of SLERPI was 0.961, and the AUC of EULAR/ACR-2019 0.956. SLERPI misclassified 20.2% (cumulatively) and 5.8% (initially) of controls as "definite SLE". Higher thresholds of SLERPI (8 initially, 8.5 cumulatively) were associated with the highest Youden index. SLERPI retained a high sensitivity for early recognition and classification purposes, even in ANA-positive patients, with higher cut-off values of its positivity performing somewhat better.