Some clinical pictures satisfy spondyloarthritis criteria without any detected imaging signs and the question sometimes arises in clinical daily practice if biologics should be started. Our aim was to evaluate anti-TNF efficacy in patients with clinical but not imaging (radiographic, CT-scan, MRI) signs of spondyloarthritis.

This retrospective study concerned patients with axial spondyloarthritis fulfilling European Spondyloarthritis Study Group (ESSG) criteria, treated with anti-TNF after failure of conventional therapies. Therapeutic responses, rated according to the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI)-50% or 20mm and ASAS-20 or -40 definitions, were evaluated after 12 months. Factors associated with those responses were also sought. Propensity score was used to check thereafter whether there were interactions with some baseline variables.

Among 385 patients included, 257 with imaging signs had significantly more frequent therapeutic responses (P=0.0005). About 40% of the spondyloarthritis patients without imaging signs responded to anti-TNF. The response rate was significantly higher in HLA-B27 carriers with initial imaging signs (P=0.028). Furthermore, responders were younger at biotherapy onset, with lower Bath Ankylosing Spondylitis Functional Index (BASFI) and pain visual analog scale score, and higher C-reactive protein (CRP), compared to non-responders. After weighted calculation, the prediction of response to TNF-blockers was quite similar in both groups.

The percentage of patients with exclusively clinical signs who responded to anti-TNF was far from negligible. Regardless the HLA-B27 status, having imaging signs of spondyloarthritis does not provide a superiority of response to anti-TNF compared to the absence of imaging sign. The absence of any imaging sign in patients with spondyloarthritis should therefore not lead to the exclusion of anti-TNF therapy.