Résumé
Background: Anti-TNF treatments (TNFi) have shown high efficacy in axial spondyloarthritis (ax-SpA) with inadequate response to non-steroidal anti-inflammatory drugs (NSAIDs). However their effect remains predominantly symptomatic, and their long-term tolerance as well as significant societal cost justify investigation about a potential reduction in drug dosage, or –most feasible and comfortable for the patient– increase in intervals between doses. Objectives: To assess if a progressive and monitored reduction of administered TNFi by increase of intervals between injections results in a comparable proportion of patients remaining after 12 months (m) in low disease activity state despite a decreased cumulative treatment dose received. Methods: Non-inferiority randomized controlled trial, having included adult patients with ax-SpA fulfilling ASAS criteria, already treated by anti-TNF, and in stable low disease activity for at least 6 m (current and at least 6 m old BASDAI<4/10), who were randomized into 2 groups: either keeping on their usual treatment with stable doses (“unchanged” group), or progressive spacing of injections of their treatment (“spacing” group). Follow-up was done every 3 m during 12 m, with regular monitoring of disease activity and, in patients from the group “spacing”, modification of the rhythm of injections according to disease activity and predefined standardized protocol (either increase or decrease (step-back) of intervals between injections). The primary endpoint was the difference of proportions of patients having a low disease activity state (BASDAI<4/10) after 12 m of follow-up between the 2 groups. It was estimated on the ITT population after multiple imputation. The 90% confidence interval associated was calculated using the Farrington-Manning method and the lower bound was compared to the non-inferiority margin of -20%. With an expected proportion of 85% patients remaining in low disease activity in the unchanged group, and α and β risks at respectively 5% and 90%, the required number of patients was calculated at 358, and thus 398 had to be included with a 10% expected proportion of patients with unavailable data. Results: 398 patients were randomized in 23 French rheumatology units (197 and 201 in the spacing and unchanged groups respectively), and 389 included in analyses (9 did not receive the allocated treatment). Mean (SD) age was 44.3 (12.4) years, 71.2% were males. Mean (SD) BASDAI at inclusion was 1.45 (1.02). TNFi used were etanercept (35.7%), adalimumab (33.9%), infliximab (20.6%), golimumab (9.3%) and certolizumab (0.5%). For the 373 patients with complete follow-up (93.7%), 162/184 (88.0%) had a low disease activity in the “spacing” group vs. 173/189 (91.5%) in the “unchanged” group at 12 m. After multiple imputation for the 16 patients with missing data, the difference of proportion between the two groups was estimated to -4.18% [CI90% -10.0; 1.7], thus confirming the non-inferiority of the “spacing” procedure. In the “spacing” group at 12 m, 134/162 (82.7%) patients in low disease activity were still receiving a lowered TNFi dose. Conclusion: In ax-SpA patients with BASDAI<4 for at least 6 months under TNFi, it is possible to increase intervals between injections while maintaining a low disease activity by adjusting treatment with quarterly monitoring of SpA activity.