Background/Purpose

Some observations have suggested that the effectiveness of abatacept (ABA) may be decreased in rheumatoid arthritis (RA) patients (pts) who previously failed rituximab (RTX).(1) The objective of this study was to compare the effectiveness of ABA started after prior inadequate response (IR) to RTX (RTX-IR) versus prior IR to anti-TNF agents only (aTNF-IR) in routine care.

Methods

This is a pooled observational database analysis of 9 prospective registries of RA pts (Czech Republic, Denmark, France, Italy, Norway, Portugal, Spain, Sweden, Switzerland). We included all RA pts treated with ABA with information on prior use of specific bDMARDs and who experienced either a RTX-IR or a aTNF-IR . The primary outcome was drug retention of ABA, defined as the time between first and last administration plus one dispensation interval, and analyzed using a Cox proportional hazards model. A secondary endpoint was EULAR good or moderate response rate at one year, estimated by longitudinal interpolation and corrected for drug retention (Lundex(2)). All analyses were adjusted for potential confounders, such as calendar year, demographics, country, number of prior bDMARDs and other disease characteristics.

Results

We identified 1994 pts initiating ABA with 3105 pt-years of follow-up. Of these, 486 pts (24%) received ABA after failing RTX and 1508 pts (76%) after failing > 1aTNFs, but never RTX. RTX-IR pts had significantly higher disease activity at baseline, longer disease duration, more functional disability, more prior bDMARDs and used more concomitant glucocorticoids than had aTNF-IR pts. (Table)

Baseline characteristics *	RTX-IR (N=486)	aTNF-IR (N=1508)	p-value
Age, (yrs)	57	57	0.08
Female, (%)	84	82	0.41
DAS28	5.2	4.9	<0.001
Dis. Duration (yrs)	13.9	12.1	<0.001
HAQ	1.4	1.2	<0.001
BMI	26	26	0.91
CRP (mg/L)	24.8	22.7	0.11
ESR (mm/1h)	34	32	0.25
RF+, (%)	70	74	0.10
N° prior bDMARDs, med (IQR)	3[2-4]	2[1-2]	<0.001
Oral Glucocorticoid use, (%)	74	59	<0.001

* Values are the means, unless stated otherwise

The crude median retention time of ABA after RTX-IR was 1.65 y (IQR: 1.42 – 2.13) compared to 2.05 y (IQR: 1.87 – 2.27) after aTNF-IR (p = 0.11). After adjustment for potential confounders no difference between ABA retention rates was found (Hazard Ratio (HR) RTX-IR vs aTNF-IR: 1.00 (95%CI: 0.83 – 1.21)). Similar results were found when examining only ABA treatment discontinuations due to ineffectiveness (HR: 1.04 (95%CI: 0.84 – 1.29)). However, response rates at one year were slightly lower in pts with RTX-IR compared to pts with aTNF-IR (73% EULAR good or moderate responses versus 83% in aTNF-IR (p=0.001); Lundex-adjusted EULAR good or moderate response 45% versus 56% in aTNF-IR, p=0.003, respectively).

Conclusion

The results of this large pooled RA population of inadequate responders to bDMARDs suggest that the slightly decreased effectiveness of ABA in patients having experienced a RTX-IR may be largely driven by a selection of pts with more treatment refractory disease.

References: 1. Das S. Et al. Ann Rheum Dis 2014;73:909-12 ; 2. Kristensen LE. et al. Arthritis Rheum. 2006 Feb;54(2):600-6

---

« Back to 2014 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/abatacept-after-rituximab-in-rheumatoid-arthritis-a-pan-european-collaboration-of-ra-registries/