Background/Purpose: The SELENA flare index (SFI) has been used in clinical and clinical trial settings since 2005 (1), but has the limitation that it classifies only severe versus combined mild/moderate flares. Since mild flares are often clinically insignificant, and it is useful to distinguish flares by organ system, a revision of the SELENA flare index (2) was carried out to define severe, moderate and mild flares separately, in terms of clinical and/or treatment variables, and by organ system. This revised flare index (rSFI) is independent of SLEDAI and was compared to the original SFI in the Phase 2 (ROSE) trial of rontalizumab (anti-interferon alpha) in extrarenal lupus (3).

Methods: The ROSE study enrolled patients with moderate to severe active SLE disease (3). Study investigators were trained on the use of the rSFI and other instruments before the start of the trial. rSFI was administered at scheduled study visits in the ROSE trial along the original SELENA flare index (SFI) and BILAG index. The new BILAG 2004 flare definition was used (severe = new or worse A, moderate = ≥2 B new or worse, mild = ≥1B or ≥3C new or worse). The proportion of patients with flares from week 4 to 24 was tabulated. Inter-instrument agreement between the different flare classifications was evaluated with Cohen’s kappa/weighted kappa coefficient (4).

Results: A total of 235 efficacy-evaluable patients were included in the analysis.   The assignment of no flare/ mild-moderate flare/ severe flare was compared for the rSFI versus SFI.  In general, there was substantial agreement between the rSFI and SFI (kappa 0.70, weighted kappa 0.73). The agreement between rSFI and SFI was better for the clinical part of the rSFI (kappa 0.69) than the treatment part of rSFI (kappa 0.35). The SFI showed fair agreement (kappa 0.30; weighted 0.39) and the rSFI showed moderate agreement (kappa 0.31; weighted 0.51) with the BILAG flare index. Many of the significant discrepancies between instruments (e.g., mild flare by rSFI, but severe by the BILAG flare index) were determined, after further review, to be correctable data reporting errors.

Conclusion: The revised SFI (rSFI) has substantial agreement with the previously validated SFI, and the clinical portion of the rSFI, tested independently, showed the best agreement. The rSFI has better agreement with the BILAG flare index than the original SFI. Some discrepancies were due to data entry errors that can be avoided with improved site training. One possibility to improve flare endpoints would be to combine moderate and severe flares, as those are most likely to be clinically important.

(1)  Petri M. et al Combined oral contraceptives in women with systemic lupus erythematosus. NEJM 2005 353(24): 2550-58.

(2)  Bayon J. et al. Revision of the SELENA flare index. Arthritis Rheum 2009. 60:S339

(3)  ACR 2012 Abstract # 2622

(4)  Cohen, J.  Weighed kappa: Nominal scale agreement with provision for scaled disagreement or partial credit. Psychological Bulletin 1968.  70 (4): 213–220.

ACR Meeting Abstracts - https://acrabstracts.org/abstract/validation-of-the-revised-selena-flare-index-in-systemic-lupus-erythematosus/