The clinical development of SC abatacept is based on a phase II study of safety and 4 dose-finding phase III studies (Table 1).

The ACQUIRE study6 compared the efficacy of 24 weeks of abatacept 150mg SC weekly versus traditional IV administration. The study included 1457 patients with active RA despite treatment with MTX. The SC group received one IV loading dose on day one. Some patients with a biological DMARD failure (3.3% in the SC and 4.5% in the IV group) were allowed for inclusion. The primary objective was to demonstrate non-inferiority of the SC formulation against IV by ACR20 at 24 weeks. The characteristics of the patients were similar in both groups and showed significant clinical activity, since patients would be required to have a number of painful joints and swollen over 10 and a DAS28 exceeding 6 in both groups. The primary endpoint showed no significant differences, since the SC abatacept group had an ACR20 response of 76% and IV treated patients, 75.8%. Other secondary objectives, such as ACR50 and 70, were similar. There were no differences when stratifying patients by weight. Disability measured using the Health Assessment Questionnaire (HAQ) improved in both groups. HAQ response was seen in 68.2% versus 63.8% in SC and IV patients, respectively. The reduction in DAS28 was similar in both groups: 2.57 (SC) and 2.55 (IV). In both groups, similar percentages of patients with low activity were seen at 6 months (DAS28<3.2: 39.5 SC and 41.3 IV) and remission: 24.2 and 24.8 for SC and IV, respectively. There were also no differences in patient assessment of pain and disease activity.

Data from the open-label extension of the AIM and ACQUIRE4 trials6 serve as comparison of IV and SC abatacept administration.22 The analysis included 1372 patients from ACQUIRE, 378 from AIM in its open-label extension phase. The patients in both studies had high rates of activity, with swollen joint count at baseline 31 for AIM and 29.6 in ACQUIRE and DAS28 of 6.4 and 6.2, respectively. For comparison of both treatment group, the beginning of the days until the last visit was considered as 253, on day 897. The values of ACR20, 50, and 70 were similar in both studies, both at baseline and at the end of follow-up. An ACR20 efficacy of 81.8% and 80.1% at the start of the IV and SC groups, which remained to the end of the period of observation period (83.6% and 83.4%) was observed. The behavior was similar in the case of ACR50, 50.8% (IV) and 54.9% (SC) at the start of the comparison and 58.6% (IV) and 59.9% (SC) at the end. The ACR70 was 27.1% compared with 32% and 34.7% and 38.1% for IV and SC treatment at the beginning and end of treatment, respectively. The disease activity dropped to DAS 2.72 (IV) and 3.01 (SC) from baseline at the start of the comparison in ACQUIRE, with virtually identical values in AIM. The percentage of patients with low activity measured by DAS28 <3.2 and remission <2.6 was similar in both treatment groups. The low activity in AIM went from 40% at baseline to 50.2% at the end and a remission rate of 20.5 and 29.7. In ACQUIRE, there was low activity in 44% and 53.8%, remission in 27.3% at baseline and 35% at the end. In short, the results of SC abatacept are similar to those obtained with IV after comparing patients included in the AIM study, with similar populations, patients with very active disease, refractory to MTX. The degree of improvement and the percentages of ACR improvement and patients in remission or low activity were maintained during the observation period for two and a half years.

Some studies included an IV loading dose to rapidly achieve therapeutic concentrations of the drug. At the start of the ALLOW open trial, 167 patients received abatacept SC with an initial IV loading dose (about 10mg/kg body weight). In the long-term extension of ACCOMPANY, patients were stratified to SC abatacept, 125mg weekly without initial loading dose. In both studies, patients were refractory to DMARD or biologicals and similar in other demographics, although the severity of the disease was lower in patients who received an IV loading dose compared to than those who did not.

A total of 167 patients participated in ALLOW, receiving SC abatacept plus an IV initial loading dose and 100 patients entered the ACCOMPANY study and received SC abatacept with or without MTX, without a loading dose. Improvements in DAS28 and HAQ were similar in both groups with or without the IV loading dose (Table 2). Although the data should be used with caution by people with different baseline severity, the magnitude of improvement was similar in both studies, regardless of whether an initial IV loading dose was administered.23

The pharmacokinetic data obtained in these studies (discussed at length in this paper) also indicate that the IV loading dose is not necessary, since SC abatacept reaches therapeutic concentrations by the time of the second weekly dose.17

Both the study of multiple dose phase II11 trial of weight-dependent compared with fixed doses of 125mg weekly of abatacept, as in ACQUIRE,6 pharmacokinetic results showed that a weekly dose of 125mg of SC abatacept provided serum concentrations within the therapeutic range, regardless of patient weight.

In the ACQUIRE trial, patients with active RA, stratified by body weight (<60kg, 60–100kg, >100kg) were randomized to receive abatacept IV 10mg/kg body weight every 4 weeks or SC abatacept 125mg. The SC abatacept group received, on day 1, an IV infusion (10mg/kg) 1h before the first SC dose. Pharmacokinetic and immunogenic studies showed that a fixed dose of 125mg/week were well tolerated, did not induce immunogenicity against CTLA-4 and produced concentrations within the therapeutic range. The effectiveness, measured as ACR20/50/70 response, DAS28, low disease activity and remission was similar in the IV and SC abatacept groups. Subgroup analysis according to body weight did not show differences in the ACR20 response that depended on the route of administration.

In the ATTUNE study,8 an open single treatment arm trial, 71 patients from the AIM extension study (abatacept in patients with failure to MTX4) and 52 from ATTAIN (abatacept in patients with anti-TNF failure5) were included in a 12-week study extended to a year, to assess the safety of changing the formulation of abatacept from IV to SC (both in combination with MTX). The percentage of patients with baseline remission or low disease activity remained stable and physical function was assessed by HAQ during the study period and at one year follow up. This occurred in all patients and when both populations, AIMs and ATTAINs, were analyzed separately.

In the ALLOW study,7 167 patients were treated with SC abatacept+MTX. Of these, 120 patients had reduction in DAS28 greater than 0.6 at 12 weeks and were randomized to continue SC abatacept+MTX or placebo+MTX until week 24. At the end of this period, all patients entered an extension study of treatment with SC abatacept along with 37 patients who had no response at week 12. The reintroduction of abatacept in the group of patients who had temporarily suspended recovered the efficacy rate and induced a clinical response of the same magnitude as the abatacept group that remained with uninterrupted treatment.

In a 4-month study, a group of 100 patients were stratified to receive SC abatacept+MTX (n=51) or SC abatacept monotherapy (n=49). Of these, 50 and 46, respectively, entered the extension study. Both treatment groups were similar, except that the monotherapy group had higher activity than the combination group (baseline DAS28 5.6 vs 5.0). After 4 months, 55.3% (SC abatacept+MTX) and 44.2% (abatacept monotherapy) patients had low disease activity. At the end of an observation period of 18 months, this situation was present in 57.5% of the combined group and 69.4% in the monotherapy arm. After 4 months, the DAS28 remission occurred in 29.8% of the combined-therapy group and 32.6% of those who received monotherapy. In the extension phase at 18 months, remission rates of 42.5% (combination therapy) and 58.3% (monotherapy) were attained. Dropout rates were also equivalents.

SC abatacept improves the functional capacity of RA patients equally than IV Abatacept. Functional capacity improvement is also maintained when patients receive SC abatacept after a period of withdrawal, or they receive it as a continuation of IV abatacept IV. Variations in physical function were also measured by HAQ with SC abatacept evaluated against IV abatacept in patients with inadequate response to MTX (ACQUIRE study),6 in adults with active RA (ALLOW study)7 and in patients switching treatment from IV to SC (ATTUNE study).8

The percentage of patients achieving an improvement in the HAQ in the ACQUIRE trial was 68.2% (95% CI [95% CI], 64.8–71.6) and 63.8% (95% CI, 60.3–67.3) for SC and IV abatacept groups, respectively (estimate of difference [4.5%, 95% CI, −0.4 to 9.4]). The change in adjusted mean±standard deviation from baseline to month 6 for SC abatacept was 0.69±0.02 and 0.70±0.02 for IV abatacept. The primary endpoint, non-inferiority of SC abatacept versus IV abatacept, based on ACR20 responses at 6 months was fulfilled.

In the ALLOW trial, the primary endpoint (immunogenicity) was fulfilled. Efficacy assessments were secondary objectives. At the end of the period I, all patients had similar reductions in the DAS28 score. The mean reductions in HAQ for patients who remained in the SC abatacept group were: in period I −0.74 (95% CI, −0.91 to −0.57) in period II −0.72 (95% CI, −0.95 to −0.50) and −0.86 in period III (95% CI, −1.04 to −0.67). For patients who suspended SC abatacept during period II, the mean reductions in HAQ were, in period I −0.63 (95% CI, −0.76 to −0.49), in period II of −0.50 (95% CI, −0.63 to −0.37) and −0.72 in period III (95% CI,−0.85 to 0.60).

The main objective of the ATTUNE study was to evaluate the safety of SC abatacept during the first 3 months after switching from IV abatacept. Changing from IV to SC abatacept was well tolerated, with a safety profile consistent with that observed during treatment with long-term IV abatacept in ATTAIN and AIM.4,5 At baseline, the mean±SD HAQ for patients with an inadequate response to MTX (n=71) was 0.91 (0.68) and for patients with inadequate response to anti-TNF (n=52) was 0.98 (0.71). The mean HAQ, based on an analysis as observed for patients with inadequate response to MTX at month 3 (n=71) was 0.82±0.69, and at month 12 (n=69) was 0.86±0.70. For patients with an inadequate response to anti-TNF, HAQ at 3 months (n=49) was 0.92±0.67, and at month 12 (n=45) was 0.97±0.66. It can be concluded that abatacept SC maintains the improvement in the HAQ previously obtained with IV abatacept.

Inhibition of radiographic progression with SC abatacept was evaluated in the AMPLE trial.10 This study conducted a comparison of SC abatacept versus adalimumab in patients, both with standard doses, with inadequate response to MTX. The primary endpoint was non-inferiority in the ACR20 at 12 months of treatment, resulting in a clinical response and similar kinetics with both drugs. An abstract presented at the last ACR congress (Washington, 2012) also showed that other secondary variables, such as pain, HAQ and the various components of the SF36,24 reached similar values. In addition, an important secondary outcome measure was the absence of radiographic progression assessed using the van der Heijde modified total Sharp Score (MTSS). Matched radiographic images were available at one year for 1290 and 289 patients in the abatacept and adalimumab groups, respectively. In both treatment groups, a similar inhibition of radiographic damage is observed. After one year, the mean changes from baseline in the abatacept+MTX and adalimumab+MTX groups were: total score, 0.58 and 0.38, erosion score, 0.29 and −0.01, and narrowing of the joint space, 0.28 and 0.39, respectively. The rate of absence of radiographic progression (defined as the smallest detectable change ≤2.8) was observed in 84.8% and 88.6% of patients in the abatacept+MTX and the adalimumab+MTX groups, respectively. In conclusion, it can be said that SC abatacept is similar in efficacy to adalimumab for inhibition of radiographic progression in RA.

All processed proteins, as biological agents, have the potential to develop an immune response, but their structure is one of the most important factors that determine the appearance of immunogenicity.25 Chimeric antibodies, such as infliximab, due to its murine part, are the most frequently associated to anti-drug antibodies, while protein compounds formed from the extracellular portion of a receptor coupled to an immunoglobulin are the least immunogenic, as is the case of etanercept and abatacept.

Abatacept is a fully human recombinant protein, comprising the extracellular domain of the CTLA4 molecule and a fragment of the Fc portion of human IgG1, with just a difference of four amino acids in the connection or hinge region, genetically modified to reduce binding to Fc receptors and prevent complement activation; except the binding region or hinge protein, it is fully human and needs not be recognized as foreign by the immune system. Additionally, it has also been proposed that abatacept can reduce ADA formation due to its mechanism of action.

The clinical consequences of immunogenicity depends on the intensity of the immune reaction and the affinity for the ligand and may range from no clinical consequence, if the amount of antibody is small and of low affinity, to a significant alteration in the drugs pharmacokinetics, efficacy or safety. Non-neutralizing anti-drug antibody formation leads to small immune complexes, which are slowly removed by the liver and spleen, but, theoretically, the drug still retains the ability to bind to the ligand and therefore does not completely lose its effectiveness. The formation of neutralizing antibodies will prevent the drug from binding to its therapeutic target, so its clinical efficacy will be directly affected. In the case of abatacept, for example, there is a theoretical risk for the development of neutralizing antibodies that can react with the endogenous CTLA4 molecule expressed on the T lymphocytes, neutralizing the activity of the endogenous protein and lead to an immunostimulatory state, which potentially may worsen the underlying disease or help develop new autoimmune diseases.

One of the most important aspects of immunogenicity is the development of methodology to determine it, and because there is still no standardization of methods, the results can not be compared or contrasted.25 At present, there is only abatacept immunogenicity data reported by Bristol–Myers Squibb and no independent groups has studied it. It is important to note that, often, the data provided by the companies in the registration studies and data arising in clinical practice are dissimilar, primarily due to patient selection, different methodologies and longer follow-up all patients in observational studies, whereas in the extension studies only those continuing earn the benefit of treatment, which are probably those who are not developing anti-drug antibodies.

Two methods were used to study the immunogenicity of SC abatacept, based on the first tests used in the studies of IV abatacept, with some improvements that increase sensitivity. Initially, we developed an ELISA directed against the entire abatacept molecule, but to avoid false positives and the signal produced by the presence of rheumatoid factor and nonspecific anti-Fc antibodies, a more specific assay was developed using only the portion of the CTLA4 molecule used as a positive control and a pool of polyclonal antiabatacept antibodies obtained by inoculating a monkey.26 All ELISA are exposed to results interfered with by the presence of free drug, which can not detect anti-drug antibodies.27 To minimize this, all samples were collected just before administration of the next dose and have also been studied in patients who had skipped some infusions or after the treatment period.

The second method is based on another ELISA, called electroluminiscence (ELS), and has proven to be more sensitive and therefore does not suffer interference from circulating drug levels. Samples are first treated with acid to dissociate immune complexes and to minimize the effects of other nonspecific antibodies. This assay can detect antibodies against all regions of abatacept.8 All positive sera with anti-CTLA4 antibody in ESL and ELISA were studied to identify the presence of neutralizing antibodies through a bioassay for determining whether the serum of patients with anti-drug antibodies is capable of inhibiting the immunosuppressive effect of abatacept in lymphocyte culture (LE 5).

The results of immunogenicity studies with IV abatacept, both during blinded treatment as long-term extension, overall show very low immunogenicity. From a total of 2237 patients studied, 62 (2.8%) demonstrated anti-abatacept or CTLA4 antibodies. Using ESL, the figure increased to 3% and no protective effect of MTX (2.3% with MTX vs 1.4% without MTX) was demonstrated. Patients who discontinued treatment had slightly higher frequencies of anti-drug antibodies than those who continued treatment (7.4% vs 2.6%) and only 8 of 13 patients studied showed neutralizing activity. No clinical relevance was demonstrated in terms of loss of efficacy, a need to increase the dose, side effects or increased development of autoimmune diseases.26

Development of SC abatacept should achieve efficacy and safety results similar to those of the IV route, with similar immunogenicity levels, taking into account that the SC route of administration is more immunogenic than IV when forming aggregates at the injection site, process antigens differently and more efficiently and transports most of the volume administered by the lymphatic route to the bloodstream.1 The study was based on four general principles, a study comparing the two routes of administration, another after the suspension and the reintroduction of the drug, one that investigates the relationship with concomitant medication with MTX and finally the development of immunogenicity in patients moving from IV to SC route.

In the ACQUIRE study, few patients who developed antiabatacept antibodies or tested positive for them in the anti-CTLA4 trial, which failed to demonstrate any effect on the efficacy or safety in these patients.6 These results support the fact that immunogenicity is not increased when the SC route is used compared to IV. In a similar study in Japan,28 after 6 months of treatment no anti-drug antibodies were observed in any of the 59 SC patients versus one of 59 with IV treatment.

I the ALLOW7 study, the effect of drug suspension and rechallenge on immunogenicity in patients receiving MTX was investigated. In the first 3 months, all patients were treated with SC abatacept openly and end of the period, 2 (1.7%) of 118 patients developed anti-drug antibodies. In the open period, none of the patients who continued treatment developed immunogenicity while the other developed seven (9.6%) among the 73 who had stopped the drug for 3 months. In phase III of rechallenge, anti-drug antibodies were demonstrated in 1 (2.6%) of the 38 patients who remained on treatment throughout the period and in 2 (2.7%) of 73 patients who reintroduced treatment. These results may be due to a direct effect of the drug inhibiting the development of anti-drug antibodies during the administration or loss of interference of ELISA in the presence of the drug. Whatever the reasons, it is confirmed, as in the IV development, that the frequency of immunogenicity after discontinuing the drug and reintroducing it is very low.

In ACCOMPANY, the effect of concomitant treatment with MTX on immunogenicity was investigated when not using an IV loading dose.29 At 4 months of treatment, 3 (6.2%) of 48 patients receiving combination therapy showed anti-drug antibodies versus 2 (4.2%) of 47 in monotherapy. In the extension period, which reached 16 months, ADA were demonstrated in one (2.1%) patient treated in combination versus 0 in monotherapy. No effects were demonstrated in efficacy or safety.

Finally, in the ATTUNE trial8 investigating the development of immunogenicity in patients switching from IV treatment from AIM and ATTAIN to SC treatment, at 3 months of SC treatment, anti-drug antibodies were demonstrated in 8 (6.6%) of 122 patients in the ELISA while only 1 (0.8%) of the 122 patients were detected using the ESL assay, confirming the very low rates of immunogenicity of SC abatacept (LE 2b).

In the pooled safety analysis of30 Phase II trials for dose-finding, ACCOMPANY9 ACQUIRE 6, ALLLOW7 and ATTUNE8 with an exhibition of 4214 patient-years for 1879 patients, the incidence rates of adverse events serious per 100 patient-years for SC abatacept were 9.25 (95% CI, 7.46–11.48), 10.75 (95% CI, 8.72–13.24), 8.21 (95% CI, 6.29–10.72) and 8.85 (95% CI, 6.26–12.51) in the 0–6, 6–12, 12–18, and 18–24 month periods. In the31 AMPLE trial at one year, 1.3% of patients treated with SC abatacept+MTX and 3.0% of patients treated with SC adalimumab+MTX discontinued medication for serious adverse events.

In the ACQUIRE trial,6 SC abatacept and IV abatacept percentages of suspension for serious adverse events was 1.1% and 1.9%, respectively. In the ACCOMPANY9 trial at 4 months, these rates were 5.9% for SC abatacept+MTX and 2.0% for SC abatacept monotherapy. In the ATTUNE8 trial, suspension rates for serious adverse events were 0.8% at 3 months and 3.3% in the accumulated period. These percentages were higher in patients who had previously failed TNF antagonists.

In the pooled safety analysis,30 incidence rates per 100 patient-years of serious infections for SC abatacept in periods of 0–6, 6–12, 12–18, and 18–24 months were 2.42 (95% CI, 1.60–3.68), 2.12 (95% CI, 1.33–3.36), 1.28 (95% CI, 0.66–2.46) and 1.51 (95% CI, 0.68–3.36). Pneumonia and urinary tract infections were the most frequent serious infections. Perhaps these figures are somewhat lower than those reported in clinical trials with other biologics. These incidents show that the risk of serious infection does not increase over time. In the31 AMPLE trial at one year, 2.2% of patients treated with SC abatacept+MTX and 2.7% of patients treated with SC adalimumab+MTX had a serious infection.

In the pooled safety analysis,30 3% of patients treated with SC abatacept developed some skin reaction at the injection site. The rate of adverse events per 100 patient-years was 2.2. These percentages are somewhat lower than for etanercept (6.5% at 3 months and 10% per year) and adalimumab (19.5%–26.1% at 1 year).32–34 For the 0–6, 6–12, 12–18, and 18–24 months periods of exposure, the incidence rates per 100 patient-years for skin reaction at the injection site were 5.59 (95% CI, 4.24–7.38), 0.72 (95% CI, 0.32–1.59), 1.40 (95% CI, 0.77–1.59) and 0.30 In the AMPLE one year trial, 3.2% of patients treated with SC abatacept+MTX developed some kind of skin reaction at the injection site compared with 9.1% of patients treated with SC adalimumab+MTX.

In the ACQUIRE6 trial at 6 months, 2.6% of patients in the group of patients treated with SC abatacept had some kind of skin reaction at the injection site compared with 2.5% in the IV group treated with abatacept. The most common reactions in the SC abatacept group were pruritus and erythema. In the ACCOMPANY9 trial at 4 months, 5.9% of patients treated with SC abatacept+MTX developed cutaneous reactions at the site of injection compared to 8.2% of patients treated with SC abatacept. Systemic reactions 24h after the injection appeared in 7.8% of patients treated with combination therapy and 8.2% of patients treated with abatacept monotherapy.

In the pooled safety analysis,30 the rate of autoimmune events per 100 patient-years was 1.28 (95% CI, 0.9–1.75) in patients treated with SC abatacept and 1.99 (95% CI, 1.74–2.26) in patients treated with IV abatacept. Incidence rates for SC abatacept in the 0–6, 6–12, 12–18, and 18–24 month periods were 1.54 (95% CI, 0.91–2.60), 0.94 (95% CI, 0.47–1.88), 0 1.5 (95% CI, 0.85–2.63) and 2.00 (95% CI, 1.20–3.31). The AMPLE one year trial showed that 3.1% of patients treated SC abatacept+MTX developed some type of autoimmune event compared with 1.2% of patients treated with SC adalimumab+MTX.

In the ACQUIRE6 trial, 1% of patients treated with SC abatacept+MTX and 0.8% of patients treated with IV abatacept+MTX, respectively, developed autoimmune events. In ACCOMPANY9 trial at 4 months, none of the patients developed autoimmune events. In the ATTUNE8 trial, in the cumulative period, 1.6% of patients developed autoimmune events.

In the pooled safety analysis,30 incidence rates per 100 patient-years for malignant tumors with SC abatacept in periods 0–6, 6–12, 12–18, and 18–24 months were 0.33 (95% CI, 0.11–1.02), 1.05 (95% CI, 0.55–2.01), 0.37 (95% CI, 0.12–1.14) and 0.65 (95% CI, 0.27–1.57). In the31 AMPLE trial at one year, 1.6% of patients treated with SC abatacept+MTX developed some kind of tumor compared with 1.2% of patients treated with SC adalimumab+MTX.

In the ACQUIRE6 trial at 6 months, 0.4% of patients in the group of patients treated with SC abatacept had a malignant or benign tumor compared with 0.7% in the IV abatacept group. The most common reactions in the SC abatacept group were pruritus and erythema. In the accumulated period of the ATTUNE8 trial, 1.6% of patients developed a tumor.

In conclusion, the safety profile in clinical trials with SC abatacept is similar to IV abatacept. Keep in mind that exposure in these trials is still short-term and it is necessary that this profile is complete with information that will come from observational studies generated by registries.