Cut-off value to identify a flare using the Flare Assessment in Rheumatoid Arthritis (FLARE-RA) questionnaire: analysis of the TOSCA study

Abstract

Abstract Objective The Flare Assessment in RA (FLARE-RA) self-administered questionnaire aims to identify patients who had flare in the interval between two consultations. This study aimed to establish a threshold for FLARE-RA score to identify RA flare. Methods The Tocilizumab SubCutAneous study evaluated the efficacy and safety of s.c. tocilizumab (TCZ) to patients with active RA. Disease activity was assessed with the DAS28ESR at baseline and at week 2 (W2), W4, W12 and W24. The FLARE-RA questionnaire was administered at W12 and W24. Patient satisfaction, assessed at baseline and W24 with the Patient Acceptable Symptom State (PASS), was used as a surrogate marker of no flare. A correlation was sought between the FLARE-RA score at W12 and W24 and the area under the receiver operating characteristic (ROC) curve (AUC) for monthly DAS28ESR. The optimal FLARE-RA cut-off below which patient satisfaction reached the PASS was explored with an ROC curve. Results A total of 139 patients were included (mean age 57.3 ± 13.8 years, 74.1% women, mean RA duration 10.8 ± 9.2 years, mean DAS28ESR 5.8 ± 1.1). The correlation between the FLARE-RA score and DAS28ESR AUC was moderate at all times: ρ = 0.41 at W12 (P < 0.0001) and 0.51 at W24 (P < 0.0001). The optimal cut-off for the FLARE-RA score to identify absence of flare (i.e. an acceptable situation based on the PASS) was 2.3 with an AUC of 0.81. Conclusion FLARE-RA and DAS28ESR assessment differ; we propose a FLARE-RA cut-off of 2.3, below which the situation (i.e. without flare) is acceptable for patients.