The SLE-DAS provides an accurate and feasible flare tool in the clinical setting: a validation study

Abstract Objective To assess the criterion validity of the SLE disease activity score (SLE-DAS) flare tool and compare its performance in identifying flares against other instruments. Methods Patients with SLE fulfilling SLE-DAS low disease activity at baseline were included from two academic lupus clinics. During follow-up, flares were identified by the senior attending clinician, applying the expert-consensus-based definition as gold-standard. The first clinical flare from flaring patients, and the first visit after baseline in patients without flares were analysed. In each no flare/flare visits, we assessed flares by SLE-DAS (score increase ≥1.72), classic-SELENA Flare Index (c-SELENA FI), revised-SELENA FI (r-SELENA FI), and SLEDAI-2K (score increase ≥4). We estimated the sensitivity, specificity, and Cohen’s Kappa agreement of each flare tool against the gold-standard. Results A total of 442 patients were included and followed-up for 22.9 (14.2) months. Incidence of flares was 8.19/100 patient-years, with 69 patients experiencing flares. The SLE-DAS identified 96.6% of the expert-defined flares implying a treatment change and classified 28.0% of those as moderate/severe. Sensitivity and specificity for the gold-standard flare definition were: SLE-DAS 97.1% and 97.3%, c-SELENA FI 88.4% and 98.1%, r-SELENA FI 88.4% and 96.8%, SLEDAI-2K 56.5% and 99.2%, respectively. Kappa coefficients of these instruments were 0.902 (95% CI: 0.847, 0.957), 0.870 (95% CI: 0.805, 0.935), 0.832 (95% CI: 0.761, 0.903), and 0.663 (95% CI: 0.557, 0.769), respectively. The number of flare misclassifications was lowest with the SLE-DAS, and highest with the SLEDAI-2K. Conclusion The SLE-DAS accurately identifies and categorizes flares as mild or moderate/severe. It is feasible and, thus, may help the physicians’ treatment decisions in the clinical practice setting.