Background A multi-biomarker disease activity (MBDA) score (Vectra® DA) has been validated to assess disease activity, on a scale of 1-100, by measuring 12 serum biomarkers associated with inflammation in RA.
Objectives To evaluate the correlation of MBDA score with DAS28-CRP and modified Sharp van der Heijde (mSvdH) progression score in a strategy study of TNF inhibitor (TNFi) tapering for patients with RA.
Methods The DRESS (Dose REduction Strategies of Subcutaneous TNFi) study is a randomised controlled trial comparing dose reduction of adalimumab or etanercept with usual care. Patients were selected based on DAS28-CRP <3.2 or rheumatologist clinical judgement. TNFi dosing was reduced every 3 months until discontinuation or flare (defined as DAS28-CRP increase >1.2 or >0.6 if DAS28-CRP ≥3.2). For flares, TNFi was restarted or escalated. Data and serum were collected every 3 months and at flare. Radiographs were assessed with mSvdH at BL and 18 months and progression (ΔmSvdH >0.5) was calculated. Correlations between MBDA score and clinical measures were determined at baseline (BL), 9, 18 months and first flare visits. MBDA values were compared between patients with and without a flare (untreated) at 9 months. DAS28-CRP and MBDA score were cross-classified for BL and for BL, 9 and 18 months combined according to 4 disease activity categories. Linear weighted Cohen's kappa was calculated. Difference in mean time-averaged (MTA) MBDA score for progressors vs. non-progressors was described.
Results 180 patients (121 dose reduction, 59 usual care) had serum available at BL, 9 and 18 months for 172, 167 and 169 patients respectively: 64% female, mean disease duration 12.1 (SD 8.3) years, 73.3% ACPA positive. For all patients, mean MBDA score and DAS28-CRP were lowest at BL, and greatest at flare (Figure 1). At 9 months, by when most flares had occurred, mean MBDA scores were 43.7 (SD 11.2) for flares (n=14) vs. 38.1 (SD 11.4) for non-flares (n=134) (unpaired t-test p=0.04); mean DAS28-CRP values were 3.6 (SD 0.7) vs. 2.0 (SD 0.6) (p<0.01). Moderate or high values were observed for 65% of MBDA scores and 21% of DAS28-CRP at BL, and for 71% and 25% respectively for all visits. Moderate or high MBDA scores were observed for 64% of patients with remission/low DAS28-CRP at BL and 68% for all visits. Linear weighted Cohen's kappa were 0.02 for BL and 0.08 for BL, 9 and 18 months combined. Correlation between DAS28-CRP and MBDA score was greatest at 18 months (Spearman's rho=0.45, p<0.01) and lowest at BL (rho=0.19, p<0.01)(Figure 1). Among DAS28-CRP components, correlation with MBDA score was greatest for CRP and lowest for tender joint count. Spearman's correlation for MTA DAS28-CRP and MTA MBDA score was 0.41 (p<0.01). Radiographic progression was observed in 26% of patients. MTA MBDA score was 38.2 (SD 7.9) for progressors versus 36.7 (SD 10.0) for non-progressors (NS).
Conclusions Correlation between MBDA score and DAS28-CRP in this TNFi withdrawal study was low at baseline, when MBDA scores tended to be lowest, and greatest at 18 months. MBDA score detected high disease activity more often than DAS28-CRP. These results suggest that MDBA score measures a somewhat different domain than DAS28-CRP. Radiographic progression was limited in this study and not associated with BL MBDA score.
Disclosure of Interest C. Bouman: None declared, A. van der Maas: None declared, O. Segurado Shareholder of: Myriad Genetics, Inc., Employee of: Crescendo Bioscience, a wholly owned subsidiary of Myriad Genetics, Inc., E. Sasso Shareholder of: Myriad Genetics, Inc., Employee of: Crescendo Bioscience, a wholly owned subsidiary of Myriad Genetics, Inc., F. van den Hoogen: None declared, A. den Broeder: None declared