Effects of TNFα-blockers tapering on occurrence of patient-perceived flares in rheumatoid arthritis: A subanalysis of strass

Background Flares in rheumatoid arthritis (RA) are a patient-perceived sign of disease activity which might be particularly important to assess in the context of treatment tapering. However assessment of flares in not yet well-defined. Objectives To explore the frequency of patient perceived flares during treatment tapering in RA, and to assess agreement between patient-perceived flares and other criteria including patient-reported outcomes and change in DAS28. Methods The STRASS study was a step-down randomized trial (ref). Patients had RA, were treated with adalimumab (ADA) or etanercept (ETN), and were in DAS 28-remission (DAS ≤2.6) for ≥6 months. Patients were randomized to either the “spacing”(S) arm (where the TNF blocker was tapered gradually) or “maintaining”(M) arm, over 18 months. Flares were evaluated through a patient-reported questionnaire every 3 months, asking: “Concerning the last 3 months, did you experience symptoms of a relapse of RA?”, with subquestions on pain and swollen joints. The frequency of visits where patients reported flares was compared in the M and S arms by Wilcoxon test. Disease characteristics were compared (t-test or Wilcoxon as appropriate) between visits with vs without a flare reported for both patient and physician reported outcomes. Agreement between patient-perceived-flares and DAS28 defined relapse (DAS 28>2.6 and an increase in DAS28 of at least 0.6 points at the same visit) was assessed by kappa. Results 137 patients were included in STRASS, 64 and 73 in the S and M arms respectively: age (mean ± SD) 55±11 yrs, females 78%, RA duration 9±8 years (ref). Over the 18 months of the study, the mean number of visits where the patient reported at least one flare (out of a possible total number of visits of 6 visits), was 1.87±1.74, with 2.44±1.68 visits with flares in the S arm, and 1.37±1.65 visits with flares in the M arm (p=0.0001). For the 256 visits with patient-perceived flares, the most frequent symptom (on subquestions) was pain (91.0%), rather than swollen joints (56.6%). Comparison between visits at which patients reported flares, and visits at which patients did not, showed statistically significant differences concerning all other outcomes (table). Agreement between patient-perceived flares and DAS 28-relapse was moderate: kappa was 0.40 to 0.49 [0.22-0.65] across visits. (Figure presented) Conclusions Patient-perceived flares discriminated well between the treatment arms in the STRASS study, and were related to worsenings in other usual outcomes in RA, indicating flares may be a useful study endpoint in tapering trials. However, agreement with DAS worsening was only moderate. Although flares are an important concept for patients with RA, more work is needed on the concept of flares.