A Multicentre, Randomised, Double-Blind, Placebo-Controlled, Parallel-Group Study to Evaluate the Efficacy and Safety of Sirukumab in the Treatment of Giant Cell Arteritis

This was a multicentre, randomised, double-blind, placebo-controlled, two-part phase 3 trial: Part A was a 52-week double-blind treatment phase to establish efficacy and safety of sirukumab; Part B was a 104-week long-term extension phase with optional open-label sirukumab treatment (up to 52 weeks) for patients with active disease at the discretion of the investigator. The study was conducted at 56 hospitals and clinics in Australia, Belgium, Bulgaria, France, Germany, Hungary, Italy, the Netherlands, New Zealand, Spain, the UK, and the USA.

Following a ≤ 6-week screening phase, eligible patients were randomised (3:3:2:2:2) at baseline (week 0) to receive: sirukumab 100 mg q2w for 12 months plus a 6-month prednisone taper; sirukumab 100 mg q2w for 12 months plus a 3-month prednisone taper; sirukumab 50 mg q4w for 12 months plus a 6-month prednisone taper; placebo q2w for 12 months plus a 6-month prednisone taper; or placebo q2w for 12 months plus a 12-month prednisone taper (Fig. 1).

Randomisation was performed centrally via an Interactive Response Technology System according to a computer-generated schedule (generated prior to study commencement), stratified by baseline prednisone dose (< 30 mg/day; ≥ 30 mg/day). Enrolment of patients with relapsed/refractory disease would be capped at approximately 50% to ensure enough patients with new-onset disease were included.

To maintain blinding, patients in the sirukumab 50 mg q4w arm received placebo injections q2w (alternating between sirukumab and placebo), both administered subcutaneously in visually matched syringes using prefilled SmartJect^® autoinjectors. All patients received prednisone (≥ 20 mg/day) at the start of screening. At baseline, prednisone dose was 20–60 mg/day, with the starting dose chosen by investigators based on their best clinical judgement. Patients remained on this prednisone dose for the first week following baseline; subsequently, the prednisone taper was administered open-label for doses ≥ 20 mg/day and blinded for doses < 20 mg/day by providing prednisone and matching placebo in blister packs. Patients experiencing disease flare were treated with an investigator-defined open-label prednisone rescue regimen while continuing the double-blind injections of sirukumab/placebo. Investigators, patients, and the sponsor/study team were blinded to treatment throughout the 52-week treatment period (detailed in Supplementary Appendix 1). During Part B, open-label sirukumab treatment was administered as per investigator discretion.

Patients completing Part A were eligible to enter Part B. Patients in remission at week 52 discontinued blinded sirukumab/placebo and were observed for maintenance of remission; in the event of disease flare, patients could receive open-label sirukumab 100 mg q2w for up to 52 weeks (per investigator discretion). Patients not in remission at week 52 or unable to tolerate prednisone taper could also receive open-label sirukumab 100 mg q2w for up to 52 weeks.

This study (GSK study number: 201677; NCT02531633) was reviewed and approved by institutional review boards and local research ethics committees before commencement (listed in Supplementary Appendix 2). The study was conducted in accordance with International Council on Harmonisation of Technical Requirements for Pharmaceuticals for Human Use Good Clinical Practice ethical principles and the Declaration of Helsinki [33]. Patients provided written informed consent prior to study commencement. The full study protocol is available at https://​www.​clinicaltrials.​gov/​ProvidedDocs/​33/​NCT02531633/​Prot_​000.​pdf.

Patients aged ≥ 50 years with active GCA were included. GCA was diagnosed according to the following criteria: history of erythrocyte sedimentation rate (ESR) ≥ 50 mm/h and/or C-reactive protein (CRP) ≥ 2.45 mg/dl; plus unequivocal cranial GCA symptoms and/or unequivocal polymyalgia rheumatica symptoms; plus features of GCA by temporal artery biopsy or imaging (e.g., ultrasound, magnetic resonance angiography, computed tomography angiography, positron emission tomography-computed tomography). Active GCA was defined as the presence of unequivocal cranial GCA symptoms and/or unequivocal polymyalgia rheumatica symptoms and ESR or CRP elevation (ESR ≥ 30 mm/h or CRP ≥ 1 mg/dl) within 6 weeks of baseline. Patients with a major ischemic event (unrelated to GCA) within 12 weeks of screening, marked prolongation of QTc interval, or receiving certain medications were excluded. Major inclusion/exclusion criteria are detailed in Supplementary Appendix 1.

The primary endpoint was the proportion of patients in sustained remission at week 52. Sustained remission was defined as achievement of all the following: remission (absence of clinical signs and symptoms of GCA and normalisation of ESR [< 30 mm/h] and CRP [< 1 mg/dl]) by week 12; absence of disease flare (recurrence of symptoms attributable to active GCA with or without elevations in ESR/CRP) following remission at week 12 through week 52; completion of assigned prednisone taper protocol; no requirement for rescue therapy at any time through week 52.

Secondary endpoints in Part A included the proportion of patients with disease flare, cumulative prednisone-equivalent dose at week 52, patient- and physician-reported outcomes (Supplementary Appendix 1), change in serum ESR and CRP from baseline over time (collected at every visit), and safety of sirukumab compared with placebo.

Safety assessments included incidence of treatment-emergent AEs (TEAEs; from first dose to 16 weeks post last dose), treatment-emergent serious AEs (TESAEs), AEs of special interest (AESIs; listed in Supplementary Appendix 1), and changes in clinical chemistry parameters. The National Cancer Institute Common Terminology Criteria for AEs version 4.03 [34] were used to grade laboratory abnormalities. Additional secondary endpoints are listed in Supplementary Appendix 1.

Key endpoints in Part B were maintenance of disease remission (proportion of patients remaining in remission without requiring rescue therapy or treatment change over time) at week 24 (Part B) and after cessation of 12-month sirukumab treatment, and safety (as per Part A).

Due to early study termination, sample size was the number of patients randomly assigned to treatment at termination. The original study sample size was calculated assuming a 30% sustained remission rate in the placebo plus 6-month prednisone taper arm, versus a 70% sustained remission rate in the sirukumab plus 6-month prednisone taper arms at week 52. In order to detect a difference with a 5% significance level, the following sample sizes provided > 91% power: sirukumab 100 mg q2w, n = 51; sirukumab 50 mg q4w, n = 34; placebo, n = 34; all arms in combination with a 6-month prednisone taper. However, given early study termination, the actual sample size was the number of patients randomly assigned to treatment at study termination. A planned interim futility analysis was not performed due to early study termination.

No statistical analyses were performed due to limited patient numbers as a result of early study termination. All data interpretations were based on listed data summarised by treatment group. The intent-to-treat (ITT) population included all randomised patients who received ≥ 1 dose of sirukumab/placebo, analysed according to treatment allocated at randomisation. The primary endpoint was summarised in the revised ITT population: patients who received ≥ 1 dose of sirukumab/placebo and completed 52 weeks or discontinued prior to sponsor announcement of study termination. For the revised primary endpoint, patients who withdrew early were considered treatment failures; missing components, including absence of disease flare, were imputed with the assumption that they have not been met. The safety population included all patients who received ≥ 1 dose of sirukumab/placebo, analysed according to treatment received. The randomised population included all randomised patients. No patients were excluded from ITT or safety populations; therefore, these three populations comprised the same patients.

Post hoc analyses were performed as described in Supplementary Appendix 1. For patients in the revised primary analysis, patient-level data review was performed after study completion to evaluate presence of investigator-reported disease flare, without imputation. Patient-level data review after study completion was also performed to evaluate disease flare rate up to week 12 in all randomised patients.

In total, 246 patients were screened and 161 (of 204 patients originally planned) were randomised (sirukumab: n = 107; placebo: n = 54) at the time of study termination. The first patient was randomised 18 November 2015, study termination was 10 October 2017, and the last patient last visit was 21 March 2018. Overall, 28/161 (17.4%) patients completed week 52 and 133/161 (82.6%) discontinued early, largely due to sponsor request (early study termination) (Fig. 2). Demographic characteristics were similar across treatment arms (Table 1). Proportions of patients with baseline prednisone doses < 30 mg/day and ≥ 30 mg/day were similarly balanced across treatment arms.

Twenty-six patients continued to Part B; 8/26 (30.8%) received ≥ 1 dose of open-label sirukumab and 22/26 (84.6%) completed the 16-week safety follow-up visit. No patients had completed Part B at study termination. One patient was lost to follow-up.

Treatment duration was shorter than planned across all treatment arms (Table 2).

The primary analysis was revised due to early study termination and included only patients who had completed or discontinued the study prior to the study termination announcement to avoid potential bias. In the revised ITT population (N = 55), 28 patients completed week 52 and 27 patients discontinued prior to study termination. The sirukumab 100 mg q2w plus 3-month prednisone arm had the highest proportion of withdrawals (8/13 [61.5%]) (Table 3). A high proportion of patients failed to achieve sustained remission at week 52 across all treatment arms; the most common reason was non-completion of prednisone taper due to early study termination. Six patients, all in sirukumab arms, fulfilled sustained remission criteria at week 52 (Table 3). Sustained remission was not achieved by 82.4–88.9% patients in sirukumab arms and all patients in placebo arms due to: not being in remission at week 12; disease flares from week 12 to week 52; receiving glucocorticoid rescue therapy.

Using the imputation rule for primary endpoint components (where patients who had withdrawn from the study early were counted as having had a flare), 52.9–69.2% and 71.4–88.9% patients in sirukumab and placebo arms, respectively, were considered to have experienced disease flare from week 12–52 due to early study withdrawal (Table 3). An ad hoc review of patient-level data (without imputation) showed the proportion of patients experiencing disease flare from week 12–52 was 5.9–11.1% in sirukumab arms and 28.6–55.6% in placebo arms, lower than in the primary analysis in all treatment groups (Table 3). Glucocorticoid rescue therapy use was consistent with the non-imputed flare results from week 12 to week 52 (Table 3).

In randomised patients (n = 161) with ≥ 1 flare assessment from week 2–52, the observed proportion of patients with disease flare was lower in sirukumab (100 mg plus 6-month prednisone: 7/38 [18.4%]; 100 mg plus 3-month prednisone: 11/39 [28.2%]: 50 mg plus 6-month prednisone: 8/26 [30.8%]) than placebo arms (placebo plus 6-month prednisone: 10/25 [40.0%]; placebo plus 12-month prednisone 10/27 [37.0%]). An ad hoc patient-level data review showed 11.9–23.1% and 14.8–18.5% patients in sirukumab and placebo arms, respectively, experienced ≥ 1 flare between week 2 and week 12; this was highest in the sirukumab 100 mg q2w plus 3-month prednisone taper arm (23.1%) (Table 4). In an ad hoc patient-level data review of patients completing week 52 in Part A (completer analysis; N = 28), 11.1–20.0% and 50.0–80.0% patients in sirukumab and placebo arms, respectively, experienced disease flare between week 12 and week 52 (Table 5).

Mean cumulative prednisone-equivalent dose for patients completing week 52 was 2418–2974 mg and 3157–3603 mg in sirukumab and placebo arms, respectively (Supplementary Appendix 3; Supplementary Table 1).

In sirukumab arms, mean CRP and ESR decreased at week 2 compared with baseline; sustained suppression was observed throughout the study. In placebo arms, mean ESR and CRP increased or remained stable over time (Fig. 3).

Excluding decreased lymphocytes, grade 3 or 4 laboratory abnormalities in haematology, lipid and liver function tests were reported only in sirukumab arms (Table 8).

Five patients in sustained remission at week 52 (Part A) entered Part B; 3/5 (60%) maintained remission at week 4 and did not receive open-label sirukumab, and 2/5 (40%) withdrew prior to week 4. None of these patients experienced disease flare.

Overall, 34 AEs were reported in 8 patients who received open-label sirukumab. The only AE experienced by ≥ 2 patients was headache (one patient in the sirukumab 50 mg q4w plus 6-month prednisone arm and 1 in the placebo plus 6-month prednisone arm). One patient (in the placebo plus 12-month prednisone arm) permanently discontinued open-label sirukumab due to abnormal liver function tests. No TESAEs or AESIs were reported in Part B.