Objective To assess the efficacy and safety of abatacept plus methotrexate versus methotrexate alone in early erosive rheumatoid arthritis (RA).
Methods The AGREE was a 2-year phase IIIb multinational study in early (≤2 years) RA. During the double-blind period (year 1), patients were randomly assigned 1:1 to receive abatacept+methotrexate or methotrexate alone; all patients received open-label abatacept+methotrexate during year 2. Clinical outcomes assessed included 28-joint disease activity score (DAS28) defined remission, low disease activity score (LDAS), American College of Rheumatology (ACR) responses and physical function. Radiographic outcomes were assessed using the Genant-modified Sharp total score (TS). Safety was monitored throughout.
Results Of the 459 patients completing year 1, 433 patients (94.3%) completed year 2. DAS28-defined remission, LDAS, ACR and physical function were sustained through year 2 in the original abatacept+methotrexate group, with 55.2% in remission at 2 years. Upon introduction of abatacept in the methotrexate-alone group, additional patients achieved DAS28-defined remission (44.5% vs 26.9%), LDAS (60.4% vs 43.2%) and improved ACR 70 (49.8% vs 31.7%) for year 2 versus year 1. Less radiographic progression was observed at 2 years in the original abatacept+methotrexate group than the methotrexate-alone group (change in TS 0.84 vs 1.75, p<0.001). No new safety issues were seen. Similar rates of serious adverse events, serious infections and autoimmune events were observed in years 1 and 2.
Conclusions The AGREE trial was the first to examine the impact of T-cell co-stimulation modulation with abatacept in patients with early erosive RA. Early treatment with abatacept+methotrexate resulted in greater sustainable clinical, functional and radiographic benefits than methotrexate alone, with acceptable safety and tolerability.
Trial Registration: NCT00122382
Statistics from Altmetric.com
Patients with early rheumatoid arthritis (RA) with poor prognostic factors such as rheumatoid factor (RF), anti-cyclic citrullinated peptide type 2 (CCP2) antibodies, high disease activity and baseline erosions have poor outcomes.1,–,3 Initiation of early, intensive treatment is recommended for such patients; however, there is a lack of consensus on when to initiate biological therapy. When a biological agent is added early to methotrexate, significantly greater improvements in disease activity status, rates of remission and radiographic outcomes are observed compared with methotrexate alone.4,–,7
Abatacept is a disease-modifying antirheumatic drug (DMARD) biological agent that has demonstrated significant efficacy benefits in patients with early RA as well as those with long-standing RA with an inadequate response to methotrexate and anti-tumour necrosis factors.8,–,14
AGREE (Abatacept trial to Gauge Remission and joint damage progression in methotrexate-naive patients with Early Erosive rheumatoid arthritis) was a multinational, randomised, double-blind, 2-year trial. Results from the year 1 double-blind period in methotrexate-naive early RA patients demonstrated that abatacept plus methotrexate resulted in greater clinical and radiographic improvement compared with methotrexate alone, with a comparable safety profile.7 Here we present the efficacy and safety of abatacept treatment over 2 years.
Patients and methods
The AGREE trial design and patient eligibility criteria have been described previously.7 Patients were 18 years or older, diagnosed with RA15 for 2 years or less, seropositive for RF and/or anti-CCP2 with evidence of radiographic erosions. Patients were methotrexate-naive or had limited previous exposure to methotrexate (≤10 mg/week for ≤3 weeks) and had high disease activity (≥12 tender and 10 swollen joints, and a C-reactive protein (CRP) ≥0.45 mg/dl).
During the double-blind period (year 1), patients were randomly assigned 1:1 to receive abatacept plus methotrexate (∼10 mg/kg according to weight range) or placebo plus methotrexate. Methotrexate was initially dosed at 7.5 mg/week, increased to 15 mg (week 4) then to 20 mg (week 8), which was maintained until study completion. Remission defined as 28-joint disease activity score (DAS28) (CRP) less than 2.616 and radiographic progression as measured by the Genant-modified Sharp total score (TS)17,–,19 at year 1 were co-primary endpoints of the trial.
During year 2, patients in the original abatacept plus methotrexate group continued treatment, while abatacept was initiated in the methotrexate-alone group. No concomitant biological agents were permitted. Decreases in daily corticosteroids, weekly methotrexate, and any additional DMARD added during year 1 were permitted during year 2 at the discretion of the investigator. No adjustments to abatacept dose or schedule were permitted.
Primary analyses were performed at 1 year; however, the subjects, sites, and central laboratories remained blinded until the end of the study.
Clinical and imaging assessments
Clinical outcomes evaluated during year 2 included: DAS28 remission, low disease activity score (LDAS, defined as DAS28 (CRP) score ≤3.2) and American College of Rheumatology (ACR) responses including major clinical response (defined as 6 months of consecutive ACR70 response).
Physical function was assessed using the health assessment questionnaire disability index (HAQ-DI). HAQ-DI response was defined as a reduction from baseline of 0.3 units or greater. Radiographic assessments and outcomes were assessed using the Genant-modified Sharp scoring method. Radiographic non-progression was defined as a change from baseline in TS of 0 or less. Baseline and year 1 radiographs were re-read concurrent with year 2 follow-up films by readers blinded to sequence and treatment.
Adverse events (AE) and serious AE were classified using MEDRA version 11.1. AE of special interest included infections, autoimmune events, malignancy and infusional events.
Efficacy and safety analyses were based on a modified intent-to-treat (ITT) population including all patients who completed year 1 and received at least one infusion of abatacept in the year 2 open-label period. Efficacy data are reported according to the original year 1 double-blind treatment group, with patients who discontinued considered non-responders.
For safety analyses, data from all patients in the modified ITT population arms were pooled. Descriptive statistics with two-sided 95% CI were provided for clinical variables. Analysis over time using imputation for missing data was performed for ACR response, HAQ-DI response, DAS28 remission and LDAS.
Radiographic analyses were based on as-observed data and included no imputation. A non-parametric analysis of covariance model was used to assess changes from baseline in Genant-modified Sharp scores for TS, erosion score (ES) joint-space narrowing (JSN). The observed cumulative proportion was plotted against the actual change in TS from baseline to year 2.
Post-hoc analyses evaluated: (1) the proportion of patients who achieved simplified disease activity index (SDAI) and Boolean-defined remission20; (2) radiographic outcomes in patients who were in sustained remission from year 1 to year 2; (3) the number of active joints (defined as tender or swollen or both) in patients who were in remission; (4) normalisation of physical function, defined as HAQ-DI of 0.5 or less;21,–,23 and (5) an ITT analysis of DAS28 (CRP) response on all randomly assigned and treated patients by treatment group (abatacept plus methotrexate and placebo plus methotrexate).
Baseline characteristics and patient disposition
Of 459 patients who completed year 1, 94.3% (433) completed year 2 (figure 1). Baseline demographics (table 1) for all the patients treated in year 2 were similar to those of the original treatment groups at the start of the trial.24 Baseline SDAI scores were comparable between treatment groups (abatacept plus methotrexate, 48.6; methotrexate alone, 48.0).
Discontinuation rates during year 2 were low, with 2.4% of the patients (11/459; seven from the original abatacept plus methotrexate arm, and four from the methotrexate-alone arm) discontinuing due to AE and 0.7% of the patients (3/459; all from the methotrexate-alone group) discontinuing due to lack of efficacy.
The mean weekly dose of methotrexate at the end of year 2 was similar between the original abatacept plus methotrexate (17.9 mg) and methotrexate-alone (17.3 mg) groups. Few patients in the original abatacept plus methotrexate (1.7%) and methotrexate (4.0%) groups had an additional non-biological DMARD added during year 2; the addition of two or more non-biological DMARD was infrequent (0% and 0.4%).
Improvements in disease activity observed at the end of year 1 were maintained through year 2 in the original abatacept plus methotrexate group. The mean DAS28 score (SD) at 2 years for patients in the original abatacept plus methotrexate group was 2.54 (1.08) (see supplementary table 1, available online only). At the end of year 2, 55.2% (95% CI 48.8% to 61.6%) were in DAS28 (CRP)-defined remission, compared with 46.1% (39.7% to 52.5%) at the end of year 1 (figure 2A). Likewise, 71.1% (65.3% to 77.0%) were in LDAS at year 2, compared with 60.8% (54.5% to 67.1%) at year 1 (figure 2B) (see supplementary table 2, available online only).
An ITT analysis at year 2 on all the randomly assigned and treated patients (n=509) showed that 50% of the patients in the original abatacept plus methotrexate group and 39.9% in the methotrexate-alone group were in DAS 28 (CRP)-defined remission on day 729.
While 81.3% (73.9% to 88.7%) of the patients in the original abatacept plus methotrexate group who were in remission at year 1 remained in remission at year 2, 72.1% (60.9% to 83.4%) of the patients in the methotrexate-alone group achieved sustained remission at year 2.
Among patients in remission at year 2, 63.5% (55.1% to 71.9%) had no active joints, 11.9% (6.3% to 17.6%) had one active joint and 7.1% had two active joints; 17.5% had three or more active joints. After abatacept was added to the methotrexate-alone group, DAS28 remission and LDAS response rates increased to 44.5% (38.0% to 51.0%) and 60.4% (54.0% to 66.7%), respectively (figure 2). The corresponding mean DAS28 score (SD) in this group at the end of year 2 was 2.77 (1.24) (see supplementary table 1, available online only).
SDAI remission analysis at the end of year 1 showed 32.1% (26.0% to 38.3%) of the patients in the original abatacept plus methotrexate group were in remission, compared with 12.5% (8.1% to 16.9%) in the methotrexate-alone group. While remission was maintained in the original abatacept plus methotrexate group (38.6%, 31.9% to 45.3%) at year 2, the proportion of patients achieving remission upon initiation of abatacept in the methotrexate-alone group increased to 35.5% (28.6% to 42.4%), similar to the original abatacept plus methotrexate group.
When analysed using the Boolean definition, at the end of year 1, 4.5% (1.8% to 7.2%) of the patients in the original abatacept plus methotrexate group were in remission, compared with 1.9% (0.5% to 4.7%) in the methotrexate-alone group. In year 2, 8.7% (4.9% to 12.6%) of the patients who continuously received abatacept plus methotrexate were in remission, compared with 3.2% (0.7% to 5.6%) of the patients in the methotrexate-alone group who were initiated to receive abatacept.
The ACR responses observed in year 1 were sustained in year 2 (figure 3). At year 2, ACR50, ACR70 and ACR90 response rates were 74.1% (68.5% to 79.8%), 53.9% (47.5% to 60.3%) and 22.0% (16.7% to 27.3%) in the original abatacept plus methotrexate group (see supplementary table 2, available online only). More patients achieved a major clinical response in year 2 (52.2%, 45.7% to 58.6%) compared with year 1 (30.2%, 24.3% to 36.1%). A total of 89.7% (83.6% to 93.7%) of patients who demonstrated an ACR50 response at year 1 sustained this response at year 2. Similarly, from year 1 to year 2 a majority of patients sustained ACR70 (78.6%, 71.0% to 86.2%) and ACR90 responses (62.8%, 48.3% to 77.2%).
When abatacept was added to the methotrexate-alone group, improvements in ACR responses were observed by month 3 of year 2 (figure 3). ACR response rates at year 2 were similar to those of the original abatacept plus methotrexate group. The proportion achieving major clinical response increased from year 1 (13.7%, 9.2% to 18.1%) to year 2 (48.5%, 42.0% to 55.0%), approximating that of the original abatacept plus methotrexate group by the end of the trial.
The percentage of patients in the original abatacept plus methotrexate group who achieved a HAQ-DI response at year 1 was maintained to year 2 (81.5%, 76.5% to 86.5%). Among year 1 HAQ-DI responders, 94.1% (90.8% to 97.5%) sustained their response to year 2. The mean change from baseline in the HAQ-DI score reported at the end of year 1 (−1.01, SE 0.05) was sustained during year 2 (−1.06, SE 0.05).
After the addition of abatacept to methotrexate in year 2, the proportion of HAQ-DI responders in year 2 was 78.4% (73.1% to 83.8%). The mean change from baseline in the HAQ-DI score at year 2 in this group was −1.01 (0.05).
Among the original abatacept plus methotrexate group, 54.7% (48.2% to 61.2%) had normalisation of physical function at 2 years (figure 2C). After the addition of abatacept to the methotrexate-alone group in year 2, 47.6% (40.8% to 54.4%) of patients in the original methotrexate group achieved HAQ normalisation, a greater proportion compared with year 1.
Radiographic data were available at baseline, year 1 and year 2 in 86.7% of patients. At year 1, significantly less progression from baseline in TS (0.65 vs 1.48, p=0.004) and ES (0.50 vs 1.26, p=0.004) was observed in the abatacept plus methotrexate group (n=211) compared with the methotrexate-alone group (n=199, figure 4A), while no significant difference was observed in progression of JSN (0.16 vs 0.22, p=0.205). In an additional 207 patients in the original abatacept plus methotrexate group with radiographs at baseline, year 1 and year 2, significantly less progression was observed during year 2 compared with year 1 (mean change in TS 0.18 vs 0.66 units, p<0.001).
Upon initiation of abatacept to the methotrexate-alone group (n=192), less progression was observed during year 2 than year 1 (mean change in TS 0.25 units vs 1.49 units, p<0.001). However, the cumulative structural damage observed to year 2 was higher in the methotrexate-alone group than in those continuously treated within the abatacept plus methotrexate group (mean change from baseline to year 2 in TS of 1.75 vs 0.84 (p<0.001) and ES of 1.40 vs 0.59 (p=0.003), respectively; figure 4A). Radiographic non-progression from baseline to year 2 was observed in 56.8% (50.2% to 63.5%) of patients in the original abatacept plus methotrexate group compared with 43.8% (36.7% to 50.8%) in the methotrexate-alone group (figure 4B). Among year 1 non-progressors, 112/123 (91.1%, 86.0% to 96.1%) and 81/97 (83.5%, 76.1% to 90.9%) in the original abatacept plus methotrexate and methotrexate-alone groups, respectively, remained non-progressors to year 2.
For patients in sustained remission from year 1 to year 2, the cumulative structural damage observed was comparable between the abatacept plus methotrexate group and the original methotrexate-alone group initiated to receive abatacept (mean change (SD) from baseline to year 2 in ES of 0.46 (1.07) vs 0.56 (1.22); JSN of 0.14 (0.47) vs 0.08 (0.33) and TS of 0.59 (1.31) vs 0.64 (1.47)). In these patients, radiographic non-progression was observed in 55.0% (44.1% to 65.9%) of patients who continuously received abatacept plus methotrexate and 52.6% (36.8% to 68.5%) in the methotrexate-alone group who were initiated to abatacept at year 2.
AE were reported in 345/459 (75.2%) patients during year 2 of treatment (table 2). No opportunistic infections or cases of tuberculosis were reported. Incidence rates (per 100 patient-years) observed during year 2 were comparable to those seen during year 1 for serious AE (6.42 vs 8.35), serious infections (1.73 vs 2.04), infections (66.8 vs 78.37), malignant neoplasms (0 vs 0.81) and autoimmune events (1.30 vs 2.47).
Acute infusional AE were infrequent (12 patients, 2.6%) during year 2. Increased blood pressure (three patients) and hypertension (two patients) were the most frequent events. One patient experienced dizziness and dyspnoea; all other events were reported in one patient each. All acute infusional AE were mild or moderate, except one serious event (anaphylactic shock), in which the drug was discontinued and the patient recovered with treatment.
Disease remission has been proposed as the primary goal for the management of early RA.25,–,27 More intensive therapies including the early use of biological agents in combination with methotrexate have been suggested as treatment strategies for achieving remission, especially in patients with poor prognostic factors.25 28,–,30 The AGREE trial was the first to examine the impact of T-cell co-stimulation modulation with abatacept in patients with early RA, reporting DAS remission as an endpoint for the trial. This paper reports the long-term sustainability of clinical efficacy responses in patients treated continuously for 2 years with abatacept, with no new safety issues.
In AGREE, early treatment of patients with poor prognostic factors with abatacept plus methotrexate resulted in significant increases in the number of patients with disease remission after 1 year compared with patients on methotrexate alone and, by the end of the trial, over half of patients with continuous abatacept plus methotrexate combination treatment had achieved remission. While two-thirds of the patients in DAS remission had no active joints, 17.5% had three or more active joints. Further analysis with the more stringent SDAI and Boolean definitions showed a greater proportion of patients in remission, in those treated with abatacept plus methotrexate compared with methotrexate alone, with more than one-third reaching remission by the end of the trial using the SDAI. A similar observation was seen with the Boolean, with lower proportions of patients in remission, compared with DAS remission criteria.
Improvements in disease activity, physical function and radiographic outcomes observed during year 1 of treatment with abatacept plus methotrexate were maintained during year 2. These improvements are consistent with those observed in similar early RA populations treated for 2 years with methotrexate in combination with adalimumab5 or etanercept.31
Recent joint European League Against Rheumatism/ACR guidelines stress the importance of reporting not only proportions of patients achieving a given disease state, but also proportions sustaining that response over time.32 In AGREE, ACR70 and DAS28-defined remission were sustained from year 1 to year 2 in approximately 80% of patients; radiographic non-progression and HAQ responses were sustained in over 90% of patients. These data illustrate the sustainability of treatment response to abatacept in an early RA cohort, and are consistent with sustained treatment responses observed in methotrexate-inadequate responders treated with abatacept plus methotrexate for up to 7 years.24 33 34
As progression of structural damage has been shown to correlate longitudinally with loss of function,35,–,37 early intensive treatment may alter the course of RA in the long term.38 39 A recent meta-analysis reported that patients who initiated DMARD therapy early had reduced rates of long-term joint damage.38 As radiographic progression can determine physical function in both early and late disease,36 38 40 41 early intensive treatment that decreases progression may prevent subsequent functional and work disability.33 42 43 In this study, during year 1, less radiographic progression was observed in patients receiving abatacept plus methotrexate compared with patients receiving methotrexate alone.7 Moreover, significantly less radiographic progression was observed during year 2 than during year 1 in the original abatacept plus methotrexate group. These results provide further evidence of the increasing inhibition of structural damage over time observed with abatacept in the AIM trial.19
While clinical responses (DAS28 remission, ACR responses) in patients originally randomly assigned to methotrexate approximated those of the original abatacept plus methotrexate group after initiating abatacept during year 2, significantly greater cumulative structural damage was observed at the end of 2 years in that group compared with patients in whom combination therapy was not delayed. Taken together, these data underline the benefit of early, intensive therapy in patients with early RA who have poor prognostic factors.
Safety is an important consideration when initiating early intensive treatment, especially in populations that can expect to receive long-term therapy. At the end of year 1, the safety profile of abatacept plus methotrexate in this early RA cohort was comparable to that of methotrexate alone.7 Importantly, the safety profile during year 2 was similar to that seen in year 1, with no unexpected safety issues reported with longer exposure. Low incidence rates of serious AE and autoimmune events were observed during year 2. Also during year 2, no malignancies were reported.7 No cases of tuberculosis or opportunistic infections were reported during this trial, consistent with the low incidence observed in other abatacept trials.24 33 Further supporting the safety and tolerability profile of abatacept was the low frequency of serious infections and discontinuations due to AE observed during year 1 (2.0–3.1%) and year 2 (1.7–2.4%).
The high retention rates in the AGREE trial (90.6% and 94.3% for years 1 and 2, respectively), are consistent with those observed in other abatacept trials in patients with longer-standing disease.24 33 Few patients (<1%) discontinued due to lack of efficacy. Retention rates may reflect the impact of a study design providing open-label biological treatment to patients initially blinded to treatment and/or may be a surrogate for the overall effectiveness of treatment.
A potential limitation of this trial is that the methotrexate-alone treatment arm was not continued in year 2; therefore, comparative treatment outcomes over 2 years could not be fully assessed. Furthermore, the study population had moderate to severe early erosive disease with poor prognostic features (96.5% RF positive and 90.2% anti-CCP positive),7 making the extrapolation of findings to patients with milder disease more difficult. The data presented here should be interpreted within the context of the trial. These data are encouraging because poor prognostic status has been associated with poor long-term outcomes and an aggressive disease course. Future analyses in patients with milder disease would be of interest. Although this trial was not designed to treat patients ‘to tight disease control’, analysis of patient-level data revealed that over 80% of patients achieving DAS remission at year 1 were still in remission at the end of year 2. This provides some confidence that remission is sustainable in most patients.
These findings demonstrate that early treatment with abatacept plus methotrexate resulted in greater sustainable, clinical, functional and radiographic benefits in methotrexate-naive early RA patients than methotrexate alone, with no new unique safety issues. These data support the safety and efficacy of abatacept at an early stage of the RA disease continuum.
The authors would like to thank the investigators, personnel and patients at all study sites. They also wish to thank Lee Bagenstose, PhD, Linda Felcone, MA and Anu Santhanagopal, PhD for their editorial assistance.
Funding Bristol-Myers Squibb funded the AGREE trial. Editorial assistance was funded by Bristol-Myers Squibb, Princeton, New Jersey, USA.
Competing interests ACX was a member of the national board as a consultant for Bristol-Myers Squibb and Abbott. JW has received consulting fees from Abbott, Bristol-Myers Squibb, Shering-Plough/Essex, UCB and Wyeth. JG-R is on the advisory boards of Abbott, Bristol-Myers Squibb, Pfizer, Roche, Schering-Plough and UCB, has received lecture fees from Abbott, Bristol-Myers Squibb, Roche, Schering Plough and Wyeth, and research grants from Roche and Schering Plough. WG is a consultant for Bristol-Myers Squibb, Abbott Immunology, General Electric, Esaote and Schering-Plough, has received honoraria from Bristol-Myers Squibb, Abbott Immunology, General Electric, Schering-Plough and Wyeth, and has received research support from Abbott Immunology and Wyeth. BH is a consultant and has received grant/research support from Abbott Canada, Amgen Canada, Bristol-Myers Squibb Canada, Roche and Schering-Plough. WS is a speaker and principal investigator for Amgen, Wyeth, Abbott, Bristol-Myers Squibb, Centocor, Genentech and Biogen Idec. HG is a consultant for CCBR-SYNARC, Bristol-Myers Squibb, Wyeth, Roche, Servier, GlaxoSmithKline, Merck, Biogen Idec and Genentech and has stock in CCBR-SYNARC. CP is owner and an employee of Spire Sciences, LLC, has stocks in Synarc, and consults for Abbott, Amgen, Biogen Idec, Celgene, Genentch, Lilly, Merck, Novartis, Roche, Servier, Simpirica and Wyeth. J-CB, DMR, ACX and RH are employees and stockholders of Bristol-Myers Squibb. RW has received consulting fees and speaker's bureau fees from Bristol-Myers Squibb, Roche, and Schering-Plough and research support from UCB. The remaining authors (JB, MR, SN, PD and S-HP) have nothing to disclose.
Patient consent Obtained.
Ethics approval The study was conducted in accordance with the Declaration of Helsinki and was consistent with International Conference on Harmonisation good clinical practice. The protocol and patient's informed consent received institutional review board/independent ethics committee approval before initiation of the study.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.