Ability of the multi-biomarker disease activity score to identify rheumatoid arthritis patients in remission at risk of relapse after tnf-blocker tapering. An ancillary study of the strass trial

Background: Rheumatoid arthritis (RA) patients in sustained remission are candidates for tapering of disease modifying anti-rheumatic drugs (DMARDs). However, no predictors of relapse have been clearly identified so far. The STRASS trial included 137 RA patients in sustained DAS28 remission with TNF blockers (TNFb) who were randomized to 2 strategy arms: maintenance of TNFb at full dose or progressive DAS28-driven spacing of TNFb injections 1. RA relapses were significantly more common with TNFb spacing in STRASS but no baseline characteristics predicted the risk of relapse in either strategy arm. The multi-biomarker disease activity (MBDA) blood test measures 12 serum protein biomarkers and uses a validated algorithm to produce a score for RA disease activity on a scale of 1-100. In some studies, the MBDA score was a significant predictor of flare following DMARD tapering in well controlled RA patients. Objectives: To test the ability of the MBDA score to identify rheumatoid arthritis patients who will relapse following continuation or tapering of TNF blocker (TNFb) treatment in the STRASS study Methods: MBDA scores were determined in a central laboratory (Crescendo Biosciences, San Francisco, USA) for available archived serum samples that had been obtained at baseline from patients in STRASS (N=133). The ability of the MBDA score to predict relapse (defined as DAS28 >2.6 and increased >0.6 from previous study visit) in each arm of the 18-month STRASS trial was assessed by: 1) comparing MDBA scores in relapsing vs. non-relapsing patients by Wilcoxon rank sum or Fisher's test, and 2) determining the ability of the MBDA score to discriminate relapse, based on receiver operating characteristic (ROC) analysis summarized by C-statistic (i.e., area under the ROC curve [AUC]), and its 95% confidence interval (CI). Analyses were performed for intention-to-treat (ITT) and completer populations Results: At 18 months, 48% and 77% of the patients relapsed in the full dose maintenance and the spacing arms, respectively. In each study arm, mean MBDA scores and the percentages of patients with low, moderate or high MBDA scores at baseline were not significantly different in relapsing vs. non-relapsing patients (Table 1). The ROC analyses displayed no statistically significant discriminating ability for the MBDA score in terms of relapse prediction over the 18-month period of the trial in either arm of the ITT population (Figure 1). However, a significant ROC result was obtained in the maintenance group in the completer analysis (AUC 0.638 [95% CI.: 0.502, 0.774], p=0.01). Conclusion: The MBDA score did not display quantification of the risk of relapse in RA patients in remission, when assessed, in the context of the DAS28-driven TNFb tapering strategy in STRASS.