THU0356 The rank of administration and not the molecule has an impact on the anti tnf treatment response in axial spa

Background Five different sub-cutaneous TNF alpha blockers (TNFb) are available for the treatment of NSAID-refractory and active axial spondyloarthritis (axSpA). The efficacy of these drugs in axSpA has been well reported in several RCT and real-life studies separately or as a group, but only very few head-to-head studies have compared each molecule9s treatment effect in axSpA. Also, some data suggest that TNFb treatment effect in patients failing to a first TNFb is poorer than in TNFb-naive patients, but some other suggest that response to a second TNFb might be comparable. Objectives to describe TNFb prescription and treatment response of each TNFb, and to compare treatment response in TNFb- naive/not naive axSpA patients in a real-life setting. Methods Prospective, multicenter, usual care study with 2 visits (baseline and 12 weeks after TNFb initiation, Predict-SpA study, ClinicalTrials.gov: NCT03039088)). Patients: axSpA patients (diagnosis according to treating rheumatologist) initiating a TNFb. Data collection: Patients and disease characteristics at baseline. Previous exposure to TNFb was collected. Disease activity and function measures were collected at both visits. The choice of the TNFb to be started during the trial was at the treating rheumatologist9s discretion. Statistical analysis: effectiveness was defined by the BASDAI50 response. Non-responder imputation and baseline observation carried forward imputation (for binary and continuous outcome variables, respectively) were performed for patients who discontinued the TNFb treatment between baseline and the follow-up visit. Results Among the 527 patients enrolled in the study, 508 patients were included in the analysis (1 patients was excluded due to missing data on all disease activity measures at baseline, and 18 patients were excluded due loss of follow-up between baseline and the follow-up visit). Mean age was 41.4 (±11.6), 237 (47%) were women, with a 6.1±8.5 mean disease duration. 377/508 (74%) were TNFb-naive while 66 (13%), 35 (7%),19 (3.7%),8 (2%) and 3 (1%) had previously received, 1,2, 3, 4 and >4 TNFb, respectively. In the whole study population the most frequently prescribed TNFb was etanercept (197 (39%)), followed by adalimumab (131 (26%)), golimumab (88 (17%)), infliximab 53 (10%) and certolizumab (39 (8%)); the same order was observed in TNFb naive patients; however, in patients previously exposed to TNFb, the most frequently prescribed TNFb was infliximab (31 (24%)) followed by golimumab (29 (22%)), adalimumab (29 (22%)), certolizumab (22 (17%)) and etanercept (20 (15%)). BASDAI50 responses were comparable for all TNFb molecules: 52%, 52%, 48%, 53% and 49% for etanercept, adalimumab, golimumab, infliximab and certolizumab, respectively. BASDAI 50 response was comparable in patients TNFb-naive and patients previously exposed to only 1 previous TNFb (53% vs. 52%, for naive vs. non-naive patients) but significantly decreased after (i.e. 37% and 36.% for patients previously exposed to 2 and to 3 or more TNFb) Conclusions this study suggests that the TNFb treatment response is similar across the different available molecules but this treatment response decreased in case of a previous use of two TNFb. Acknowledgements This study was conducted thanks to an unrestricted grant from MSD Disclosure of Interest None declared