Efficacy and safety of golimumab in patients with non-radiographic axial spondyloarthritis: a withdrawal and retreatment study (GO-BACK)

Abstract

Objectives: The GO-BACK study was designed to evaluate the efficacy and safety of golimumab (GLM) treatment withdrawal in adults with non-radiographic axial spondyloarthritis (nr-axSpA) who demonstrate inactive disease during a 10-month open-label (OL) GLM run-in.

Methods: Eligible participants received OL GLM in period 1. In period 2, participants who achieved inactive disease were randomized 1:1:1 to receive double-blind (DB) treatment with monthly placebo (PBO, treatment withdrawal) or continued GLM treatment given monthly (GLM QMT) or every 2 months (GLM Q2MT). Participants who did not have a disease flare continued DB treatment for ∼12 months. Participants with a disease flare discontinued DB treatment and resumed monthly OL GLM. Primary endpoint compared the proportion of participants without a disease flare in the continued GLM treatment groups (QMT or Q2MT) vs PBO in a multiplicity-controlled, step-down fashion. Safety follow-up continued for ∼3 months after last treatment.

Results: A total of 188 patients, out of the 323 enrolled, were eligible for participation in period 2. Both GLM QMT and GLM Q2MT were superior to treatment withdrawal (PBO) in preventing disease flare (P < 0.001), with a treatment-difference vs PBO of 50.4% and 34.4% for the GLM QMT and GLM Q2MT groups, respectively. The time-to-first flare was significantly longer (log-rank P < 0.0001) with GLM treatment compared with PBO. Of 53 participants (in Q2MT or PBO) who had a confirmed disease flare, 51 (96.2%) attained a clinical response within 3 months of restarting OL GLM. Adverse events were consistent with the known GLM safety profile.

Conclusion: Among participants with active nr-axSpA who attained inactive disease after 10 months of GLM treatment, continued GLM treatment is well tolerated and provides superior protection against disease flares compared with GLM withdrawal. (EudraCT: 2015-004020-65, registered on 30 March 2022; NCT: 03253796, registered on 18 August 2017.).

Keywords: efficacy; golimumab; non-radiographic axial spondyloarthritis; reduced dosing; safety; withdrawal.

Graphical abstract

Figure 1.

Study design. An interactive voice/web response system was used to assign treatment, including double-blind randomization (1:1:1 ratio) in period 2. GLM 50 mg (or 100 mg for participants with a body weight ≥100 kg) was administered subcutaneously; only three participants received GLM 100 mg. ^aOnly participants who demonstrated clinical response (BASDAI improvement ≥2 or ≥50% relative to month 0) at month 4 and inactive disease (ASDAS <1.3) at both months 7 and 10 were eligible to participate in period 2. ^bParticipants with a confirmed disease flare (ASDAS from two consecutive assessments that both showed an absolute score ≥2.1 or a post-withdrawal increase of ≥1.1 relative to month 10) discontinued blinded treatment and resumed open-label GLM QMT for at least 3 months. ^cAll participants were followed for safety for approximately 3 months after the last dose of study treatment. GLM: golimumab; PBO: placebo; QMT: monthly dosing; Q2MT: every 2 months dosing

Figure 2.

Participant disposition. Completed study is defined as those participants who completed at least 22 months of study treatment (either double-blind or open-label after a disease flare) and had a safety follow-up contact not less than 76 days from the last dose of study treatment. AS: ankylosing spondyloarthritis; ASAS: Assessment of the SpondyloArthritis International Society; ASDAS: Ankylosing Spondylitis Disease Activity Score; CRP: C-reactive protein; GLM: golimumab; MRI: magnetic resonance imaging; NY: New York; QMT: monthly dosing; Q2MT: every 2 months dosing

Figure 3.

Proportion of participants without a disease flare (period 2, double-blind): full analysis set population. ^aPrimary endpoint was analysed using a stratified Miettinen and Nurminen test, with stratification based on CRP level (>6 mg/l or ≤6 mg/l) and a weighting scheme based on sample size weights. Type I error rate over the multiple treatment comparisons (GLM QMT vs PBO and GLM Q2MT vs PBO) was controlled by a sequential, step-down testing procedure. Period 2 participants who discontinued prior to a disease flare were counted as having a flare. All disease flare determinations were based on complete ASDAS assessments (BASDAI questions 2, 3, 6; PGD and CRP). GLM: golimumab; PBO: placebo; QMT: monthly dosing; Q2MT: every 2 months dosing

Figure 4.

Time-to-first disease flare (period 2, double-blind): full analysis set population. Start of period 2 represents day 1 for the Kaplan–Meier analysis. Participants who did not flare were censored at the time of discontinuation or at 13 months in double-blind treatment period 2. ^aP-value represents comparison of GLM QMT vs PBO and GLM Q2MT vs PBO based on log-rank test stratified by CRP level (>6 mg/l or ≤6 mg/l). GLM: golimumab; PBO: placebo; QMT: monthly dosing; Q2MT: every 2 months dosing