Background Tumor necrosis factor (TNF) inhibitors, including etanercept (ETN), have demonstrated efficacy in patients with non-radiographic axial spondyloarthritis (nr-axSpA) and an inadequate response to NSAIDs. However, not all patients respond to TNF-inhibitors. Thus, it is of interest to determine whether responders and non-responders have different baseline (BL) characteristics and/or early response patterns. This information could aid clinicians in deciding which patients to treat with a TNF-inhibitor.
Objectives To compare demographics, BL characteristics, and response to ETN treatment for the wk 24 ASAS40 responders and non-responders in the EMBARK study.
Methods Patients had axSpA per ASAS classification criteria without meeting modified NY radiographic criteria, symptoms for >3 months and <5 yrs, BASDAI score ≥4, and failed ≥2 NSAIDs. Patients were randomized to ETN 50 mg/wk or placebo (PBO) for 12 wks; then all patients received open-label ETN 50 mg/wk. This analysis includes the open-label modified intent-to-treat (mITT) population randomized to ETN. Patients were grouped according to ASAS40 response at wk 24. Clinical efficacy endpoints were assessed at 2, 4, 8, 12, 16 and 24 wks. Comparison of BL characteristics used one-way analysis of variance for continuous variables and Fisher's exact test or chi-square for all yes/no or categorical variables. Comparison of clinical responses used analysis of covariance with BL as covariate or Fisher's exact test.
Results 205 patients were in the open-label mITT population (ETN/ETN, n=100; PBO/ETN, n=105). For the 100 ETN/ETN patients in this analysis, mean (SD) age was 31.8 (7.8) yrs, 64.0% were male, symptom duration was 2.4 (2.0) yrs; 66% were HLA-B27+, and 84% were MRI sacroiliitis+. Significantly more males than females were 24-wk ASAS40 responders (77.3% vs 22.7%, table); there were no statistically significant differences in age, race, height/weight, or symptom duration. There were positive associations between 24-wk ASAS40 response and several BL disease characteristics related to back pain and also the Spondyloarthritis Research Consortium of Canada (SPARCC) MRI sacroiliac joint score; there was a negative association with enthesitis (table). According to 24-wk ASAS40 response, there were no significant differences in BL BASDAI, BASFI, ASDAS, physician global assessment, SPARCC MRI spinal score, swollen joint count, MRI sacroiliitis(+/-), HLA-B27(+/-), or several health-related patient-reported outcomes. Although higher BL CRP was significantly associated with ASAS40 response at 12 wks,1 it was not at 24 wks (BL mean CRP 7.3 for 24-wk ASAS40 responders, 7.0 for non-responders; p=ns). Notably, for endpoints with a significant difference according to 24-wk ASAS40, the distinction was apparent at wk 2 for all but ASDAS inactive disease, which differentiated by wk 8 (table). There were no significant differences in change in BASMI, CRP or ESR per 24-wk ASAS40 result.
Conclusions Categorizing patients according to 24-wk ASAS40 result demonstrated significant differences in several BL characteristics and clinical results; many clinical endpoints differentiated after only 2 wks of treatment. This information may aid clinicians in choosing which patients to treat with a TNF-inhibitor.
Dougados M, et al. Arthritis Rheumatol. 2014;66:2091-102
Disclosure of Interest J. Sieper Consultant for: Böhringer-Ingelheim, Janssen, Novartis, AbbVie, Merck, Pfizer, UCB, H. Jones Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, J. Bukowski Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, R. Pedersen Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, E. Mahgoub Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, L. Marshall Shareholder of: Pfizer Inc, Employee of: Pfizer Inc