Background/Purpose: ACPAs are sensitive, highly specific markers of RA. Current tests cannot differentiate ACPA+ RA subtypes. Fine-specificity ACPAs (FS) can distinguish between ACPA+ RA subtypes to explore predictors of treatment (tmt) response. Higher vs lower baseline (BL) ACPA levels have been associated with improved response to abatacept (ABA).¹ In the Phase IIIb AVERT study, ABA + MTX was more effective than MTX in early RA: 61% vs 45% of patients (pts) achieved DAS28 (CRP) remission (< 2.6) at 12 mths.² We investigated FS as biomarkers to predict tmt response in pts with seropositive RA in AVERT.

Methods: Pts in AVERT (NCT01142726) were randomized to double-blind weekly SC ABA 125 mg + MTX, SC ABA 125 mg or MTX for 12 mths. In this post hoc analysis, predictive value of 29 BL FS and 10 non-citrullinated control antibodies (Ab) for efficacy measures (DAS28 [CRP], SDAI) was explored using logistic or linear regression models for remission rates (RR; SDAI remission ≤3.3) and change from BL (CfB) outcomes, respectively, with tmt arm, BL disease activity, age, sex, prior CS use, BL FS (continuous or categorized into tertiles [T; specific for each FS and control Ab]) and FS-by-tmt arm interaction as covariates. Linear mixed-effects model with repeated measures was used for time series analysis. Pairwise comparisons (unadjusted p values) between tmt arms were performed for DAS28 (CRP) and SDAI RR or adjusted mean CfB in each T. FS were analyzed with custom Bio-Plex^TM bead-based autoantibody assay.³

Results: BL characteristics were comparable in overall (N=351) and target biomarker (n=340) populations and across tmt arms (not shown). Of 29 BL FS, 15 significantly correlated with anti-CCP Ab (ACCP; surrogate of ACPA). Of 29 FS or ACCP, higher BL anti-fibrinogen B 62–81 cit 74 (ABA + MTX p=0.017; MTX p=0.99) and ACCP (ABA + MTX p=0.018; MTX p=0.89) levels were significantly associated with higher probability of DAS28 (CRP) remission in pts who received ABA + MTX but not MTX (Figure 1); levels of non-citrullinated control anti-fibrinogen B 62–81 did not correlate with tmt response (p=0.324). For DAS28 (CRP) RR, tmt differences (ABA + MTX vs MTX) were greater in medium (T2) and high (T3) subgroups vs overall population: ACCP (T2; p≤0.01), 28%; anti-fibrinogen B 62-81 cit 74 (T3; p≤0.01), 32%; overall population, 16% (Figure 2). Tmt differences in adjusted mean DAS28 (CRP) CfB for ABA + MTX vs MTX and ABA vs MTX were greater in medium (T2; Mth 6) and high (T3; Mths 6, 12) vs low (T1) anti-fibrinogen B 62–81 cit 74 subgroups (Figure 3); similar findings were seen for SDAI. Anti-fibrinogen B 62–81 cit 74 (vs ACCP) better predicted DAS28 (CRP) CfB improvement with ABA + MTX vs MTX (not shown).

Conclusion: BL FS predicted improved DAS28 (CRP), SDAI RR and CfB with abatacept + MTX vs MTX; higher BL T showed greater differentiation in tmt response with abatacept + MTX vs MTX. Anti-fibrinogen B 62–81 cit 74 (vs ACCP) better predicted DAS28 (CRP) improvement with abatacept + MTX vs MTX. These data support anti-fibrinogen B 62–81 cit 74 as a predictor of tmt response to abatacept.

References

1. Sokolove J, et al. Ann Rheum Dis 2016;75:709–714.
2. Emery P, et al. Ann Rheum Dis 2015;74:19–26.
3. Wagner CA, et al. Ann Rheum Dis 2015;74:579–586.

Medical writing: Katerina Kumpan (Caudex).

« Back to ACR Convergence 2020

ACR Meeting Abstracts - https://acrabstracts.org/abstract/higher-baseline-fine-specificity-acpas-predict-greater-treatment-response-with-abatacept-mtx-versus-mtx-monotherapy-in-seropositive-ra-a-post-hoc-analysis/