Background: In the management of patients with Systemic Lupus Erythematosus (SLE), it is of utmost importance to accurately identify lupus flares. There is a conceptual consensus definition of lupus flares [1]; however, the instruments used to identify flares in clinical trials, such as the SELENA Flare Index (SFI) are too cumbersome to apply in daily clinical practice. The SLE disease activity score (SLE-DAS) is a validated continuous measure of disease activity with higher sensitivity to change and validity in predicting damage accrual when compared to SLEDAI-2K. SLE-DAS is quickly scored with its online calculator. An increase in SLE-DAS ≥1.72 was validated as a clinically meaningful worsening of SLE disease activity. [2]

Objectives: To compare the performance of SLE-DAS, classic SFI (c-SFI), revised SFI (r-SFI) and SLEDAI-2K in the identification of lupus flares in a real-life clinical setting.

Methods: We included patients with SLE fulfilling classification criteria [ACR (1997) and/or SLICC and/or EULAR/ACR], followed at an academic lupus clinic from January 2017 to June 2020, and presenting with lupus low disease activity score (LLDAS) at baseline.

Flares occurring after baseline were identified as fulfillment of the conceptual definition of flare, as assessed by the senior lupus expert at time of each outpatient visit. For each flare event, we evaluated the fulfillment of flare criteria according to c-SFI, r-SFI, SLEDAI-2K (score increase ≥4 points from baseline), and SLE-DAS (score increase ≥1.72 from baseline). As control visits without flare, we considered the first visit after baseline, where we assessed the four tools, excluding those where a flare was identified by the gold-standard expert evaluation. Sensitivity and specificity of the four flare tools were estimated and McNemar’s test applied to assess differences with the gold-standard flare definition. The inter-instrument agreement with the gold-standard was assessed through Cohen’s Kappa.

Results: We included 297 patients (female: 86.2%; mean age: 48.9±14.6 years, mean disease duration: 12.5±9.0 years). At baseline, all patients were in LLDAS, receiving ongoing antimalarials, immunosuppressants, and/or glucocorticoids in 91.0%, 43.8% and 33.6%, respectively. During follow-up, 22.2% developed flares. The analysis included 92 flares [musculoskeletal (40.2%); renal (23.9%), mucocutaneous (18.5%), haematological (5.9%), serositis (3.3%); multisystemic (8.7%)], with increase or change of treatment in 80.4% of these episodes, and 292 visits without flare considered as control.

There was no statistically significant difference between either SLE-DAS flare or c-SFI and the gold-standard expert flare definition (p=0.41 and p=0.82, respectively), while r-SFI and SLEDAI-2K flare were different from the gold-standard (Table 1). There was a strong agreement between SLE-DAS flare, c-SFI, r-SFI and the expert definition (Cohen’s kappa, Table 1).

Conclusion: The c-SFI and SLE-DAS showed the best performance in identifying SLE flares. The SLE-DAS flare definition is easier to apply and hence might be considered as an optimal tool to be used in daily clinical practice.

References: [1]Ruperto N, Hanrahan LM, Alarcón GS, et al. International consensus for a definition of disease flare in lupus. Lupus. 2011;20(5):453-62.

[2]Jesus D, Matos A, Henriques C, et al. Derivation and validation of the SLE Disease Activity Score (SLE-DAS): a new SLE continuous measure with high sensitivity for changes in disease activity. Ann Rheum Dis 2019; 78:365-71.

Disclosure of Interests: None declared.