Article Text
PDF
Abstract
Objectives OSKIRA-4 evaluated the efficacy of fostamatinib monotherapy versus placebo on the signs and symptoms of rheumatoid arthritis over 6 weeks by Disease Activity Score C reactive protein (DAS-28(CRP)) and assessed non-inferiority to adalimumab monotherapy at Week 24 by DAS-28(CRP).
Methods Overall, 279 patients not currently taking disease-modifying antirheumatic drugs were randomised to: (A) fostamatinib 100 mg twice daily for 24 weeks plus placebo injection every 2 weeks (PBOI); (B) fostamatinib 100 mg twice daily for 4 weeks, then 150 mg once daily up to Week 24, plus PBOI; (C) fostamatinib 100 mg twice daily for 4 weeks, then 100 mg once daily up to Week 24, plus PBOI; (D) adalimumab 40 mg every 2 weeks for 24 weeks, plus oral placebo twice daily; or (E) oral placebo twice daily for 6 weeks, plus PBOI, then a switch to arm A or B.
Results Fostamatinib demonstrated a significant improvement in DAS-28(CRP) score from baseline versus placebo at Week 6 for arms A and B, but not C. Fostamatinib was significantly less effective than adalimumab at Week 24 based on DAS-28(CRP). Adverse events observed with fostamatinib treatment were consistent with those reported in previous studies, including hypertension and diarrhoea.
Conclusions Fostamatinib demonstrated efficacy as monotherapy, showing superior DAS-28(CRP) score changes between baseline and 6 weeks when compared with placebo in treatment arms A and B. However, all fostamatinib regimens demonstrated inferior responses compared with adalimumab at Week 24.
Trial registration number Clinicaltrials.gov: NCT01264770.
- Rheumatoid Arthritis
- DMARDs (synthetic)
- Anti-TNF
Statistics from Altmetric.com
Introduction
Fostamatinib is a kinase inhibitor that blocks spleen tyrosine kinase (SYK) activity and has been investigated as an oral therapy for rheumatoid arthritis (RA).1 Fostamatinib is rapidly and extensively converted to the active metabolite R406 by gut enzymes.2 R406 blocks SYK-dependent immune cell activation mediated by FcεR, FcγR and BCR immunoreceptors.3 ,4 Thus, inhibition of SYK signalling may have potential to ameliorate RA disease activity. Since SYK activation is also implicated in production of inflammatory cytokines and metalloproteinases,5 interruption of these processes could provide a useful therapeutic strategy.
The present phase IIb study (OSKIRA-4) was a 24-week, multicentre, randomised, double-blind, placebo-controlled (for 6 weeks), parallel-group study of the efficacy and safety of fostamatinib monotherapy in active RA patients. The responses to three dosing regimens of fostamatinib were compared with placebo and the tumour necrosis factor-α (TNFα) inhibitor adalimumab. This is the only study to investigate fostamatinib as RA monotherapy.
Two studies in the phase II TASKi clinical programme demonstrated significant clinical improvements in RA with fostamatinib taken in combination with methotrexate or disease-modifying antirheumatic drugs (DMARDs) (NCT00326339 and NCT00665925),2 ,6 although fostamatinib failed to show efficacy in patients who had not responded to biologics (NCT00665626).7
Methods
Study population
The study (trial registration number NCT01264770) was approved by applicable institutional review boards and performed in accordance with Good Clinical Practice guidelines and the Declaration of Helsinki. All patients provided written informed consent before participation.
Patients aged ≥18 years with active RA (diagnosed after age 16 years) were eligible for enrolment if diagnosed within 5 years and had inadequate responses to ≤2 DMARDs, intolerance to DMARD therapy, or if they were DMARD-naive and diagnosed within 2 years prior. Additional eligibility criteria included ≥4 swollen joints and ≥4 tender joints (28-joint count) and either an erythrocyte sedimentation rate ≥28 mm/h or C reactive protein (CRP) levels ≥10 mg/L. Patients were only eligible if they had at least two of the following: a positive rheumatoid factor, radiographic erosion in the last 12 months or a positive anticyclic citrullinated peptide antibody.
Subjects were excluded if they had taken any DMARD within 6 weeks of screening or previous treatment with a TNFα antagonist or another biological, systemic inflammatory conditions (other than RA) or chronic pain disorders, uncontrolled hypertension (blood pressure (BP) ≥140/90 mm Hg), recent (≤6 months) cardiovascular events, liver disease or significant liver function test abnormalities, significant active or recent infection, severe renal impairment or neutropenia. Recruitment and follow-up took place at 91 hospitals in Russia, USA, Ukraine, Bulgaria, Poland, Czech Republic, South Africa, Hungary, UK and Slovakia between February 2011 and October 2012.
Study design
Patients were assigned an enrolment number after providing written informed consent and then randomised sequentially by central allocation via an interactive voice-response system (IVRS) to receive (1:1:1:1:1) one of three oral dosing regimens of fostamatinib tablets, adalimumab by subcutaneous injection or a matching placebo tablet regimen: (A) fostamatinib 100 mg twice daily for 24 weeks plus placebo injection every 2 weeks; (B) fostamatinib 100 mg twice daily for 4 weeks, followed by once-daily maintenance dosing with 150 mg up to Week 24, plus placebo injection every 2 weeks; (C) fostamatinib 100 mg twice daily for 4 weeks, followed by once-daily maintenance dosing with 100 mg up to Week 24, plus placebo injection every 2 weeks; (D) adalimumab 40 mg by subcutaneous injection every 2 weeks for 24 weeks, plus oral placebo twice daily; or (E) oral placebo twice daily for 6 weeks followed by switch to either (i) fostamatinib 100 mg twice daily up to Week 24 or (ii) fostamatinib 100 mg twice daily for 4 weeks, followed by fostamatinib 150 mg once daily for the remaining 14 weeks up to Week 24 plus placebo injection every 2 weeks, for 24 weeks.
Either adalimumab or placebo was administered by injection once every 2 weeks. Treatment administration was double-dummy blinded. Since an exact placebo match for adalimumab is not available, each site was required to appoint an unblinded administrator for the injection (independent of the on-site study team). Patients who completed the scheduled treatment period were offered the opportunity to receive fostamatinib therapy in a long-term extension study (OSKIRA-X: NCT01242514).
A reduced dosing regimen of fostamatinib 100 mg once daily was available in case of certain adverse events (AEs). Patients who had dose reduction continued taking this reduced dose during the study. Where fostamatinib dose reduction was required, the original treatment allocation remained blinded. Patients in treatment arm C who met dose reduction criteria after Week 4 were withdrawn since no further dose reduction was available. This entailed unblinding of the patient via IVRS. Dose reductions were not permitted for adalimumab injection; however, due to the blinded nature of the study, reduction of the placebo dose for patients taking regimen D may have occurred.
Efficacy assessments
The primary endpoints of this study were change from baseline in DAS-28(CRP) at Week 6 (fostamatinib vs placebo) and change in DAS-28(CRP) at Week 24 (fostamatinib vs adalimumab).8 Non-responder imputation was used for all efficacy data, by imputing baseline values after withdrawal for any reason or after initiation of any DMARD. Secondary endpoints included European League Against Rheumatism (EULAR) DAS-28(CRP) response criteria, DAS-28(CRP) <2.6 and DAS-28(CRP) <3.2; American College of Rheumatology 20% response criteria (ACR20), ACR50 and ACR70;9 and assessment of the physical function status of patients using the Health Assessment Questionnaire-Disability Index (HAQ-DI).10
Safety and tolerability
Safety outcomes included AEs, vital signs, ECG (with independent adjudication of cardiovascular events), clinical chemistry, haematology, urinalysis and physical examination. An independent safety review committee reviewed accumulating safety data on an ongoing basis.
Statistical methods
The full analysis set was used as the primary population for reporting efficacy and safety data. This comprised all randomised patients who received at least one dose of investigational product and was analysed according to randomised treatment (intention-to-treat principle).
All efficacy endpoints were tested at a two-sided significance level of 10% for each dose regimen versus placebo. A sample size of ∼250 patients in total (∼50 each arm) was expected to provide at least 85% power to detect a difference (between fostamatinib and placebo) in the mean change from baseline in DAS-28(CRP) of 0.7 at Week 6, and to confirm non-inferiority between fostamatinib and active comparator adalimumab at Week 24 (margin of 0.6 in the change from baseline DAS-28(CRP)).
CIs for efficacy comparisons between fostamatinib and placebo are presented as 90% CIs. However, when comparing fostamatinib and adalimumab on DAS-28(CRP) at Week 24, 80% CIs were calculated (allowing assessment of non-inferiority by comparing the lower confidence limit with the non-inferiority margin).
As a phase IIb trial, no adjustments were made for multiplicity. Primary endpoints were analysed using an analysis of covariance (ANCOVA) model on the change from baseline. Analyses of ACR20, ACR50 and ACR70 at each time point were performed using a test of treatment difference in proportion of responders with a Mantel–Haenszel approach. DAS-28(CRP) EULAR response at each time point was analysed using a proportional odds model. The proportion of patients classified as achieving low disease activity, DAS-28(CRP) remission and HAQ responders at each time point were each analysed using logistic regression. HAQ-DI scores were analysed using an ANCOVA model on the change from baseline.
Results
Demography and baseline disease characteristics
In all, 452 subjects were screened; 167 did not meet inclusion criteria and six withdrew prior to randomisation. In total, 279 patients were randomised, 265 (95%) received study treatment (figure 1) and 201 (72%) completed treatment.
Patient disposition. Percentages are based on the number of patients randomised to each treatment group. The full analysis set was used as the primary population for reporting efficacy and safety data (this comprised all randomised patients who received at least one dose of investigational product). *The number of patients who completed treatment includes those who had a dose reduction.
Demography was generally well balanced across treatment arms (table 1). In total, 70.4%–84.2% of patients were women and >80% were Caucasian. There were no notable imbalances in baseline disease characteristics across randomised treatment arms (table 1), though previous and current hypertension (controlled at study entry with BP ≤140/90) were slightly more prevalent in Group A (previous hypertension, 1.9%; current hypertension, 46.3%) versus Groups B (0.0%; 37.5%), C (0.0%; 35.1%) and D (0.0%; 35.2%).
Baseline disease characteristics
Efficacy
Fostamatinib demonstrated a statistically significant improvement in DAS-28(CRP) from baseline versus placebo at 6 weeks for treatment arms A (0.56 difference (90% CI 0.23 to 0.90); p=0.006) and B (0.49 difference (90% CI 0.14 to 0.84); p=0.022), with least squared (LS) mean improvements of 1.09 and 1.02, respectively (figure 2A). However, arm C (fostamatinib 100 mg twice daily for 4 weeks then 100 mg once daily) failed to show significant improvements versus placebo for this primary outcome measure (LS mean improvement 0.75 vs 0.53; 0.22 difference (90% CI –0.12 to 0.56); p=0.280). There was notable reduction in DAS-CRP score at Week 1 for fostamatinib 100 mg twice daily versus placebo (LS mean improvement 0.69 vs 0.29; 0.41 difference (90% CI 0.21 to 0.60); p<0.001).
Primary endpoint: (A) fostamatinib versus placebo at 6 weeks (DAS-28(CRP)); (B) fostamatinib versus adalimumab at Week 24 (DAS-28(CRP)). DAS-28(CRP), Disease Activity Score based on a 28 joint count and CRP measurement; CRP, C reactive protein.
For the primary endpoint at Week 24 (figure 2B), fostamatinib arms A, B and C demonstrated significantly lower efficacy than adalimumab based on DAS-28(CRP). Treatment with adalimumab resulted in LS mean improvement from baseline of 1.78 compared with improvements of 1.06, 1.17 and 1.06 for fostamatinib arms A, B, and C respectively (A −0.72 difference (80% CI −1.04 to −0.40) p=0.005; B −0.61 (−0.94 to −0.27) p=0.020; C −0.72 (−1.04 to −0.40) p=0.004) at Week 24. Therefore, as the lower CIs were below non-inferiority margins of −0.6 in each case, all three fostamatinib arms failed to demonstrate non-inferiority to adalimumab.
Similar results were found for the secondary endpoints. All active treatments resulted in comparable, significant increases in the percentage of patients achieving the ACR20 criteria versus placebo by Week 6. Adalimumab demonstrated the highest ACR50 response rates at Week 6. In fostamatinib comparisons against placebo, only treatment arms A and B showed differences for ACR50 at this time point (13.0%, p=0.059 and 12.5%, p=0.071, respectively) versus 3.8% for placebo (figure 3A). No difference between Group C (7.0% response rate, p=0.758) and placebo was observed (figure 3A). Additionally, very few patients in any treatment arm achieved an ACR70 response at Week 6 (figure 3A). At Week 24, adalimumab generally showed higher response rates for ACR20 (59.3%), ACR50 (31.5%) and ACR70 (20.4%) than the fostamatinib-based treatments (figure 3B). Compared with adalimumab, significantly lower response rates were observed for ACR20 in fostamatinib arms A (40.7%, p=0.049) and C (35.1%, p=0.003); ACR50 in arm C (12.3%, p=0.003); and ACR70 in arm C (3.5%, p=0.002). Furthermore, though a similar ACR20 response rate at Week 24 was achieved for fostamatinib arm B versus adalimumab (56.3%, p=0.803), non-significant differences were seen for ACR50 in arms A (20.4%, p=0.114) and B (18.8%, p=0.078); and ACR70 in arms A (9.3%, p=0.082) and B (10.4%, p=0.092). Consistent findings were generally seen across ACR component analyses.
Secondary endpoint: (A) ACR20, ACR50 and ACR70 at 6 weeks versus placebo; (B) ACR20, ACR50 and ACR70 at Week 24 versus adalimumab. Only significant p values <0.1 for the comparison of fostamatinib groups versus placebo (A) or adalimumab (B) are displayed. ACR20/ACR50/ACR70, American College of Rheumatology 20%/50%/70% response criteria.
Across DAS28-based secondary endpoints, results were generally consistent with the primary endpoints; fostamatinib showed improvement when compared with placebo, but efficacy was lower than with adalimumab. Similar results were shown for HAQ-DI improvement over time, with adalimumab-treated patients showing the best improvement.
Safety and tolerability
Up to Week 6, 31.5%, 27.1% and 28.1% of patients experienced an AE in fostamatinib treatment arms A, B and C, respectively, versus 19.2% of patients receiving placebo. By study end (Week 24), the proportion of patients reporting an AE was 72.2%, 60.4% and 59.6% for fostamatinib arms A, B and C, respectively. Among patients receiving adalimumab, 55.6% had reported an AE by Week 24. Serious AEs (SAEs) were reported at similar rates for fostamatinib treatment arms A and C, and the adalimumab treatment arm (9.3%, 7.0% and 7.4%, respectively). A slightly lower SAE rate was observed in fostamatinib treatment arm B (2.1%). The incidence of discontinuations due to AEs was 16.7%, 12.5% and 14.0% in fostamatinib arms A, B and C, respectively. No discontinuations were reported due to AEs in patients taking adalimumab. The incidence of AEs leading to dose reduction was 11.1% and 4.2% in fostamatinib arms A and B, respectively, and 3.7% in the adalimumab arm.
No patients died during the study, and the only malignancy was a case of multiple myeloma in the 6-week placebo period of arm E.
The most commonly reported AEs from randomisation to study end (Week 24) were diarrhoea and hypertension (table 2). Diarrhoea was more commonly reported in the fostamatinib treatment arms (table 2), although dose ordering was not observed. Hypertension was more prominent in the 100 mg twice daily fostamatinib treatment arm (arm A) than in fostamatinib arms B and C (13.0%, 6.3% and 8.8%, respectively). The rate of hypertension with adalimumab was 9.3%. No new safety signals were identified.
Most common adverse events (AEs) (reported with a frequency ≥5%) from randomisation to end of study (Week 24)
Mean systolic BP (SBP) at baseline was similar across treatment arms. Fostamatinib was associated with an elevation in SBP (figure 4). By Week 4, there was a mean increase in SBP of 5.5 mm Hg for patients receiving fostamatinib 100 mg twice daily (n=153). By Week 8, the change from baseline was greatest in both treatment arm A (5.7 mm Hg) and in the postplacebo switchers receiving fostamatinib 100 mg twice daily and then 150 mg once daily (9.4 mm Hg). However, all fostamatinib-induced elevations in SBP were substantially attenuated by the end of the study in a setting in which treatment for hypertension was allowed. SBP increases were also observed in patients receiving adalimumab treatment. Patients receiving placebo did not experience significant changes in SBP. Plots for diastolic BP over time followed a similar pattern to those for SBP.
Change in systolic blood pressure over time in each treatment group.
From randomisation to study end, the incidence of infection was 24.1%, 10.4% and 10.5% in fostamatinib arms A, B and C, respectively, and 14.8% in the adalimumab arm. There were no opportunistic infections (Pneumocystis jirovecii pneumonia, tuberculosis, serious fungal infection). No absolute neutrophil counts (ANC) of <0.5×109/L were reported, though five patients on fostamatinib had ANC <1.0×109/L. All of these patients improved, three without a change in dose, one following dose reduction and one following dose interruption.
Alanine aminotransferase (ALT) increases ≥10× upper limit of normal (ULN) were seen in two patients in fostamatinib arm A. ALT increases from ≥5 to <10× ULN were seen in one patient in fostamatinib arm A and one in placebo arm E. All findings were reversible, and none met criteria for Hy's law.11 There were no patterns of clinical concern for changes in total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol or triglycerides.
Discussion
Previous placebo-controlled, phase II studies (NCT00326339 and NCT00665925) demonstrated that fostamatinib, when taken in combination with methotrexate/DMARDs, is associated with significant clinical benefits in RA patients.2 ,6 OSKIRA-4—the present study—is the only investigation to date of the performance of fostamatinib monotherapy, and aimed to determine the efficacy and safety of three dosing regimens in patients with relatively early RA who were either DMARD-naive or had an inadequate response/intolerance to DMARDs. Comparisons were made with placebo (up to Week 6) and the TNFα inhibitor adalimumab (up to Week 24).
OSKIRA-4 results show that fostamatinib monotherapy has a clinical effect; the primary endpoint of a significantly superior change from baseline DAS-28(CRP) score versus placebo was met at Week 6 for the two higher fostamatinib doses. A difference between arms was evident within 1 week of treatment initiation. However, the study did not meet the primary endpoint of non-inferiority to adalimumab at Week 24 (fostamatinib treatment arms demonstrated significantly lower efficacy than adalimumab in change from baseline in DAS-28(CRP)). Results for secondary efficacy endpoints were generally consistent with the primary endpoints for comparisons with placebo and adalimumab. Fostamatinib was, therefore, more efficacious than placebo, but less efficacious than adalimumab. The performance of adalimumab in this trial was at the upper end of that observed in previous studies.12–15
Safety and tolerability findings for fostamatinib in this study were generally consistent with those observed in the TASKi phase II studies2 ,6 and included liver transaminase elevations, hypertension, diarrhoea and neutropenia (table 2). A recent meta-analysis and systematic review reported similar findings.16 However, fostamatinib-associated hypertransaminasaemia and neutropenia appear to be reversible, dose-related phenomena that respond to interruption of or reduction in fostamatinib dose.7 ,17 This is supported by the current data, showing that neutropenia only occurred in patients receiving the highest fostamatinib dose. Furthermore, hypertension occurred more frequently in arm A than in any other treatment arm. This difference may be explained by the higher prevalence of ‘hypertension history’ in this group at baseline. Previous investigation also found fostamatinib-associated increases in BP more commonly in patients with prior hypertension.7
Fostamatinib has a number of off-target effects that may be responsible, in part, for the observed AEs.17 Fostamatinib effectively inhibits SYK, but also has a broader range of inhibition across the kinome.17 For example, it has been suggested that off-target effects of R406 on vascular endothelial growth factor receptor 2 may account for increases in BP.17 However, it is not possible to determine the exact contribution of any off-target effects to safety findings.
This study allowed comparison of the efficacy of various fostamatinib regimens with placebo over a 6-week period and with adalimumab monotherapy over a 24-week period. The requirement for patients taking conventional DMARDs (cDMARDs) to withdraw them at least 6 weeks prior to randomisation might be considered a study limitation as it is possible that such patients could have been in a state of flare during the trial period. However, the randomisation procedure, which was stratified by cDMARD naivety/prior inadequate response or intolerance, means any such patients are equally distributed among the treatment arms; the proportion of patients with previous cDMARD use was noted to be similar in all treatment arms.
In conclusion, in this parallel-group, phase IIb clinical study, fostamatinib monotherapy demonstrated modest clinical efficacy by comparison with placebo after 6 weeks of treatment in RA. Fostamatinib was significantly less effective than adalimumab monotherapy after 24 weeks of treatment. Safety and tolerability findings for all fostamatinib arms were generally consistent with those observed in prior phase II studies.2 ,6 The modest clinical results in all of the phase III studies have led to a decision by the companies developing fostamatinib to not submit a regulatory filing for its use in RA.18 ,19 The relevance of these findings for other more selective Syk inhibitors remains to be seen.
Acknowledgments
The clinical study was sponsored by AstraZeneca. We acknowledge the editorial services of Shelley Lindley, PhD, from PAREXEL, which were funded by AstraZeneca. PCT thanks Arthritis Research UK for their funding of Arthritis Research UK Early Arthritis Treatment Centre at the University of Oxford and the National Institute of Health Research for their funding of The NIHR Biomedical Research Centre in Musculoskeletal Disease at Oxford University Hospitals NHS Trust and the University of Oxford. The views expressed are those of the author(s) and not necessarily those of ARUK, the NHS, the NIHR or the Department of Health.
References
Footnotes
Handling editor Tore K Kvien
Contributors PCT: study design, patient recruitment, data interpretation, preparation of manuscript. MCG and MW: interpretation of data, review and approval of paper. MG: statistical analysis, interpretation of trial data, review and approval of paper. MH: study design, project management, interpretation of trial data, review and approval of paper. EN, RS and JV: patient recruitment, interpretation of data, review and approval of paper. BO: study design, interpretation of trial data, review and approval of paper.
Competing interests PCT has received grant/research support from UCB and is a consultant for AstraZeneca, Pfizer, Lilly and Celgene. MCG has received grant/research support from AstraZeneca and is a consultant for AstraZeneca. MG and MH are employees of AstraZeneca and hold stocks/shares in AstraZeneca. BO is an ex-employee of AstraZeneca. EN has received grant/research support from Servier and speaker's fees from Roche, Schering-Plough Corporation and MSD. JV has received grant/research support from UCB, GSK, MSD, BMS and Novo Nordisk, and speaker's fee from UCB, GSK and MSD. MW is a consultant for AstraZeneca and AbbVie (formerly Abbott).
Ethics approval West of Scotland Research Ethics Service.
Provenance and peer review Not commissioned; externally peer reviewed.