Background/Purpose:

We
evaluated the predictive value of the multi-biomarker disease activity (MBDA)
score for clinical outcomes and radiographic progression in a TNFi tapering
study in RA patients with sustained low disease activity.

Methods:

DRESS
is an 18-month randomised trial on non-inferiority of tapering of adalimumab or
etanercept compared with usual care (UC). TNFi was tapered every 3 months until
stopping or clinical flare (ΔDAS28-CRP >1.2 or >0.6 if current DAS28-CRP
≥3.2; major flare: >3 months). Patients were assessed every 3 months
and at flare. For flares, TNFi was restarted or escalated. X-rays of hands and
feet were scored with modified Sharp-van der Heijde score (mSHS) at
baseline (BL) and 18 months. Correlations between MBDA score and DAS28-CRP were
determined at BL, 9, 18 months and first flare. MBDA scores at BL, 9 and 18
months and ΔMBDA score from BL
to 9 months were analysed for successful stopping, successful tapering, and no tapering
possible. AUROC evaluated the predictive value of: 1) BL MBDA score for a) successful tapering vs. no tapering
possible, b) stopping vs. no tapering possible, c) occurrence of (major)
flares, d) incidence of radiographic progression (RP); 2) ΔMBDA score from BL to 9 months for
clinical outcomes and RP; 3) MBDA scores at first and second visits preceding a
flare for occurrence of that flare; 4) ΔMBDA score between
first and second visits preceding a flare for occurrence of that flare.

Results:

Of
180 patients, 171 (115
tapering, 56 UC) had
serum and month 18 outcomes available: 64% female, mean disease duration 12.1
(SD 8.3) years, 73.3% ACPA positive. TNFi had been successfully stopped in 19%,
tapered in 44%, and re-escalated to baseline dose in 37%. MBDA scores and
DAS28-CRP are summarized in Figure 1. Correlation of DAS28-CRP and MBDA score
was greatest at 18 months (Spearman’s r=0.45, p<0.01) and lowest at BL
(r=0.19, p<0.01). AUROCs for predicting successful stopping, tapering and
flare by BL MBDA score were not significant. AUROC for major flare by BL MBDA
score was significant in the UC group (10% flared; AUROC 0.72, 95% CI
0.56-0.88) and not the taper group (12% flared). RP (ΔmSHS >0.5 from BL to 18
months) occurred in 26% of 167 patients with available data and was not
predicted by MBDA score. ΔMBDA score from BL
to 9 months was not predictive for clinical outcomes or RP. MBDA scores at
first and second visit preceding a flare were not predictive for any flare.
AUROC for ΔMBDA score between
first and second visit preceding a flare was borderline significant (0.56, 95%
CI 0.50-0.63).

Conclusion:

Neither
BL MBDA score nor ΔMBDA score from BL to
9 months was predictive for successful stopping; tapering; occurrence of flare;
or radiographic progression. However, exploratory analyses showed that: 1) BL
MBDA score was predictive for flare in patients who received UC and 2) ΔMBDA score between visits
preceding a flare was borderline significant for predicting flare.