Background Subgroup analyses can be used to investigate the size and direction of treatment effects across a range of demographic and disease characteristics.
Objectives To explore response consistency in subgroups of patients with nonradiographic axial spondyloarthritis (nr-axSpA) who received golimumab (GLM) or placebo for 16 wks.
Methods GO-AHEAD was a double-blind, randomized, placebo (PBO)-controlled trial of GLM in patients with active nr-axSpA (ASAS criteria and centrally read sacroiliac (SI) joint X-rays and MRIs, disease duration ≤5 years, chronic back pain, high disease activity [total back pain ≥40 mm on a 0–100 mm Visual Analogue Scale and BASDAI ≥4 cm], and inadequate response/intolerance to NSAIDs). Patients were randomized 1:1 to subcutaneous GLM 50 mg or PBO every 4 wks. Estimated between-group treatment differences and 95% CIs for ASAS 20 and ASAS 40 response at wk 16 were calculated for prespecified patient subgroups. Treatment and subgroup differences were compared by stratified Miettinen–Nurminen methods with baseline inflammation by SI MRI and screening C-reactive protein (CRP) level as stratification factors. No multiplicity control was applied.
Results A total of 197 patients were treated (GLM=97, PBO=100). Overall, ASAS 20 at wk 16 was achieved by 71.1% of GLM patients and 40.0% of PBO patients (P <.0001). Although size of response differed somewhat across the subgroups, responses were greater to GLM than PBO in most patient subgroups (P <.05; table). Because of the small numbers of patients within subgroups and because subgroups may not represent randomized samples, results should be considered exploratory and interpreted with caution. Treatment group differences appeared to be larger in patients with objective signs of inflammation (MRI SI or CRP>ULN). The results for ASAS 40 were very similar to those for ASAS 20.
Conclusions Overall, patients with active nr-axSpA who received GLM were more likely to achieve ASAS 20 at wk 16 than those treated with PBO across a variety of baseline characteristics, including those with baseline objective signs of inflammation (eg, MRI SI or CRP >ULN).
Disclosure of Interest M. Dougados Grant/research support from: AbbVie, Lilly, Novartis, Pfizer, Roche, Sanofi, UCB, Consultant for: AbbVie, Lilly, Novartis, Pfizer, Roche, Sanofi, UCB, G. Bergman Employee of: Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Kenilworth, NJ, W. Maksymowych Grant/research support from: AbbVie, Janssen, Pfizer, Consultant for: AbbVie, UCB, Pfizer, Merck, Janssen, Eli Lilly, Celgene, Synarc, S. Curtis Employee of: Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Kenilworth, NJ, S. Huyck Employee of: Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Kenilworth, NJ, A. Tzontcheva Employee of: Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Kenilworth, NJ, J. Sieper Consultant for: AbbVie, Eli Lilly, Janssen Biologics, Merck, Novartis, Pfizer, Roche, UCB