Background
Abatacept is a fully humanized fusion protein that acts as a selective T-cell co-stimulation modulator. It is approved globally for the treatment of moderate-to-severe rheumatoid arthritis (RA) in patients with an inadequate response to one or more disease-modifying anti-rheumatic drugs (DMARDs; including methotrexate [MTX] or an anti-tumor necrosis factor [TNF]) [
1‐
4]. In Canada, abatacept can also be used in combination with MTX for the treatment of moderate-to-severe RA in patients who are MTX-naïve [
1‐
4]. Abatacept is available in subcutaneous and intravenous (IV) formulations [
1‐
4], and a favorable efficacy and safety profile has been demonstrated in randomized controlled trials [
5,
6]. IV abatacept received regulatory approval for the treatment of RA in 2006 in Canada and in 2007 in Europe. ACTION (
Abata
Cep
T In r
Outi
Ne clinical practice) is a large, real-world, prospective, observational cohort study of patients with RA from Europe and Canada who were receiving IV abatacept and were followed for a maximum of 2 years. Results from a 6-month interim analysis of the ACTION study suggest that abatacept is an effective and well-tolerated treatment option in patients with RA [
7].
The emergence of new therapies for the treatment of RA, the paucity of head-to-head studies, and the heterogeneous nature of responses to current biologics highlight the need for the identification of prognostic factors for treatment response and retention in clinical practice. The identification of prognostic factors may support individualized treatment strategies and could aid physicians in making informed therapeutic decisions [
8].
Real-world studies may include patients with a wide range of disease activity levels and encompass regional variations in treatment, and can therefore supplement the findings of randomized controlled trials with strict inclusion and exclusion criteria [
9]. Although some studies have identified prognostic factors for response to abatacept, few have been confirmed [
10]. In the French Orencia and Rheumatoid Arthritis (ORA) registry, positivity for anti-cyclic citrullinated peptide (CCP) antibodies was associated with European League Against Rheumatism (EULAR) response in a multivariate analysis, after adjustment for Disease Activity Score (DAS)28 [
11]. In an Italian study of 32 patients with RA who had been treated with abatacept, low levels of CD4 + CD28− and CD8 + CD28− T cells were associated with a greater likelihood of achieving remission at 6 months [
12].
Treatment discontinuation can result from safety concerns or a lack of efficacy, but prognostic factors for patient retention have not been explored thoroughly despite data for abatacept and other biologics being available from national registries. In patients within the ORA registry, anti-CCP positivity occurred more frequently in patients who had continued abatacept treatment after 6 months, compared with patients who had discontinued abatacept (72.5 % versus 62.4 %,
p = 0.02) [
11]. In the Swedish Rheumatology register, 6-month abatacept retention rates were higher in patients who were biologic naïve compared with patients who had received 1 or ≥2 previous biologics (94 % versus 78 % and 77 %, respectively) [
13]. In a pooled analysis of eight European RA registries, abatacept retention was strongly influenced by the number of previous biologics. Compared with patients who had not previously received biologic agents, retention was lower in patients who had received prior biologics (
p < 0.01) [
14].
This article reports prognostic factors for abatacept retention in patients with RA who had failed at least one prior biologic agent which were identified in an interim analysis from the international, non-interventional ACTION cohort study.
Discussion
This is the first international, prospective study to evaluate the prognostic factors for abatacept retention in a real-world setting in patients with RA who had received at least one prior biologic agent. In this interim analysis, the 6- and 12-month abatacept retention rates were 88.0 % and 69.9 %, respectively. Abatacept retention rates in ACTION were numerically within the range reported in other published studies [
11,
13,
22]. The 12-month retention rates for abatacept were within the range of those reported for tocilizumab [
22‐
25] and for anti-TNFs [
26].
Here, we report predictive factors for abatacept retention at 12 months identified from a cohort of patients from the ACTION study who had received at least one prior biologic agent: anti-CCP positivity, failing <2 prior anti-TNF agents, and cardiovascular comorbidity at initiation were associated with higher retention. Differences in retention between some countries were also observed. The percentage of patients receiving abatacept as monotherapy in this study (23.2 %) was broadly consistent with that observed in biologic registries (approximately 30 %) [
27].
In this study, anti-CCP positivity was associated with improved abatacept retention, compared with anti-CCP negative status. In addition, RF seropositivity was significant in the univariate analysis but not in the multivariate analysis. Anti-CCP positivity has previously been shown to be a predictor of clinical response to abatacept in patients with RA [
11,
28]. The mechanisms underlying this association remain to be further elucidated, although some initial findings suggest that the very early effect of abatacept on T-cell modulation is an important mechanism of action in patients with anti-CCP positivity [
29]. In the present study, patients with exposure to at least two prior anti-TNFs were more likely to discontinue abatacept than patients with exposure <2 prior anti-TNFs. Patients with prior exposure to multiple anti-TNFs are likely to have advanced RA and may find it difficult to benefit from any therapy; furthermore, previous studies have shown that no or low prior exposure to biologic agents is associated with longer drug survival in patients receiving tocilizumab or anti-TNFs [
14,
23,
30]. The reason for discontinuation of the last biologic agent prior to abatacept initiation was a significant predictor of abatacept retention in the univariate but not the multivariate analysis. In a subgroup analysis of the ATTAIN (
Abatacept
Trial in
Treatment of
Anti-TNF
INadequate responders) trial, the reason for prior anti-TNF failure was not associated with differences in the efficacy of abatacept over 6 months [
31].
Cardiovascular comorbidity at initiation (including cardiac arrhythmia, cardiac valve disorders, coronary artery disorders, heart failure, and myocardial disorders) was associated with a decreased risk of abatacept discontinuation compared with not having cardiovascular comorbidity. The long-term safety profile of abatacept is well established in RA [
32]. Abatacept does not have any special warnings or contraindications in patients with cardiovascular diseases. In contrast, adalimumab, certolizumab pegol, golimumab and infliximab are contraindicated in patients with moderate-to-severe heart failure (New York Heart Association class III/IV) [
33‐
36] and etanercept has a special warning in patients with congestive heart failure [
37]; this may result in limited options for switching treatment in patients with a history of cardiovascular comorbidity. Channeling may have been introduced at abatacept initiation and comorbidities may have been underreported.
Differences in retention were observed between countries in the multivariate analysis. Patients in Greece and Italy were less likely to discontinue abatacept than patients in Germany and Canada; these differences in abatacept retention may represent differences in patient populations and disease characteristics, access to biologic agents, and national guidelines for the management of RA. Although covering other European countries, country was found to impact abatacept retention in a pooled analysis of 9 European registries [
38,
39]. Baseline demographics and disease characteristics were broadly comparable across the countries included in this analysis, although patients in Canada had higher swollen joint counts and were less likely to exhibit radiographic erosion than patients in Germany, Italy, and Greece. Patients in Italy were less likely to have received ≥2 prior biologic agents compared with patients in the other countries assessed, and rates of concomitant corticosteroid use differed markedly between countries (48.5 % in Canada compared with 69.1 %, 78.8 %, and 85.1 % in Greece, Italy, and Germany, respectively). In addition, several studies have reported a lower incidence and prevalence of RA, as well as disease activity, in southern European Union (EU) countries versus northern EU countries [
40,
41]. Access to biologic agents may be related to gross domestic product per capita and healthcare system structure, including the number of rheumatologists per inhabitant [
38,
42,
43]. Treatment guidance for the management of RA is available in most countries and may have been developed taking into account gross domestic product and disease parameters (e.g. disease activity, poor prognostic factors) [
44]. Finally, patient management is also dependent on the type of practice (evidence-based versus routine practice). Together, these elements may partly explain differences in retention between countries, independently of patient characteristics. However, such detailed information was not specifically collected and is probably aggregated in the variable ‘country’, including influence of genetic background and environmental factors.
The limitations of this study are similar to other studies using uncontrolled, real-world data. There was no active comparator in this study and there are few examples of real-world data comparing abatacept with other biologic agents, or comparing other biologic agents. Owing to the real-world design of ACTION, a selection bias based on disease severity or adverse events is plausible. To minimize patient selection bias, participating physicians enrolled subsequent patients who were eligible as per the inclusion and exclusion criteria. To ensure that ACTION did not interfere with a physician’s routine clinical practice, the decision to treat a patient with abatacept was made prior to enrollment in the study. A process of random selection of study investigators from a comprehensive list of rheumatologists was used to obtain a geographically balanced group of investigators in each country who were representative of rheumatologists who treat patients with IV biologics.
This analysis of patients from the ACTION study included a large number of patients in an international real-life setting (n = 865), which permitted inclusion of a comprehensive list of all relevant prognostic factors. Most previous studies of abatacept in clinical practice have explored only national registries from single countries [
11,
22,
45]. As this was an interim analysis, further analyses are warranted to confirm these findings.
Although many studies have identified predictive factors of clinical response to biologic agents, few have been confirmed, in particular for abatacept [
10]. Even fewer studies have identified prognostic factors of abatacept retention. Similar trends were found in ACTION and in the pooled analysis of 9 European registries [
39]. Given the range of biologic agents now available for the treatment of RA, the identification of patients who will benefit from a specific therapy is of interest and may aid realistic cost-effectiveness estimates. Therefore, the identification of prognostic factors of clinical response and retention are of growing importance, as highlighted by the 2013 EULAR recommendations [
8].
Acknowledgments
The authors would like to thank all the physicians who participated in the ACTION study. The ACTION study was funded by Bristol-Myers Squibb. Clinical research organizations involved in the ACTION study were Archemin BVBA, Inventiv Health Clinical, TFS Trial Form Support S.r.l, and Winicker Norimed. Statistical analysis support was provided by Florence Mercier (Stat Process) and Guillaume Desachy (Excelya). Gilbert L’Italien (Executive Director, Global Health Economics and Outcomes Research) and Nathalie Schmidely (Associate Director of Real World Research) from Bristol-Myers Squibb provided input into the design and interpretation of this study. Professor Xavier Mariette, Hôpital Bicêtre, Paris, France, provided input into statistical analysis and interpretation of this analysis. The first draft of the manuscript was prepared by academic and industry authors with professional medical writing and editorial assistance provided by Stephen Moore, PhD, of Caudex, and funded by Bristol-Myers Squibb. The academic authors vouch for the completeness and accuracy of the data and data analyses, and for the fidelity of the study to the protocol.
Competing interests
H.G. Nüßlein has received consulting fees and speaker honoraria from Bristol-Myers Squibb, Abbott, Chugai, UCB, Essex, Wyeth, Pfizer, MSD, Novartis, and Roche. R. Alten has received research grants and speaker honoraria from Bristol-Myers Squibb. M. Galeazzi has nothing to disclose. H.-M. Lorenz has received honoraria from Bristol-Myers Squibb, Abbott/Abbvie, Chugai, UCB, Pfizer, MSD, Novartis, sanofi-aventis, SOBI, Janssen-Cilag, and Roche for presentations and participation in advisory boards, and provided advice on planning of the ACTION study. M.T. Nurmohamed has received consultancy fees from Abbott, Roche, Pfizer, MSD, UCB, SOBI, and Bristol-Myers Squibb, and has received payment for lectures from Abbott, Roche, Bristol-Myers Squibb, and Pfizer. W.G. Bensen has attended advisory boards, presented data, and performed research for Bristol-Myers Squibb, Amgen, Abbott, UCB, Merck, Pfizer, Novartis, AstraZeneca, Roche, Janssen, Warner Chilcott, and sanofi-aventis. G.R. Burmester has received consulting fees from Abbott Immunology Pharmaceuticals, Bristol-Myers Squibb, Roche, Merck, and Pfizer; research grants from Abbott Laboratories, Abbott Immunology Pharmaceuticals, Bristol-Myers Squibb, and Roche; and speaker honoraria from Abbott Immunology Pharmaceuticals, Bristol-Myers Squibb, and Roche. H.-H. Peter has nothing to disclose. K. Pavelka has been a speaker for Pfizer, Amgen, MSD, Bristol-Myers Squibb, and Abbott. M. Chartier is a consultant for Bristol-Myers Squibb. C. Poncet is a consultant for Bristol-Myers Squibb. C. Rauch is an employee of Bristol-Myers Squibb. M. Le Bars is an employee of Bristol-Myers Squibb and holds stock options in the company.
Authors’ contributions
HGN made a substantial contribution to the conception and design of the study, to the acquisition of data, and to the analysis and interpretation of data. RA made a substantial contribution to the conception and design of the study, and to the analysis and interpretation of data. MG made a substantial contribution to the acquisition of data, and to the analysis and interpretation of data. H-ML made a substantial contribution to the conception and design of the study, to the acquisition of data, and to the analysis and interpretation of data. MTN made a substantial contribution to the acquisition of data. WGB made a substantial contribution to the acquisition of data. GRB made a substantial contribution to the acquisition of data, and to the analysis and interpretation of data. H-HP made a substantial contribution to the acquisition of data, and to the analysis and interpretation of data. KP made a substantial contribution to the acquisition of data. MC made a substantial contribution to the conception and design of the study, to the acquisition of data, and to the analysis and interpretation of data. CP made a substantial contribution to the conception and design of the study, and to the analysis and interpretation of data. CR made a substantial contribution to the conception and design of the study, and to the analysis and interpretation of data. MLB made a substantial contribution to the conception and design of the study, and to the analysis and interpretation of data. All authors have read and approved the final manuscript.