Segmental spinal damage predicts mobility restriction and functional decline in radiographic axial spondyloarthritis: a longitudinal study.

Abstract

To quantify the longitudinal association between radiographic spinal damage and spinal mobility and physical function in radiographic axial SpA (r-axSpA) using disco-vertebral unit (DVU)-level analyses. This longitudinal cohort study included 477 patients with r-axSpA contributing 1263 paired clinical-radiographic assessments. Structural damage was assessed using the modified Stoke AS Spinal Score (mSASSS) and classified at each DVU as no damage, syndesmophyte or bridging syndesmophyte. Linear mixed-effects models with patient-level random intercepts were used to evaluate continuous and categorical damage measures, adjusting for time since baseline, disease duration, CRP, BASDAI, biologic therapy use, sex and HLA-B27 status. Standardized beta coefficients (β) and Cohen's d were calculated to quantify effect magnitude. Radiographic damage increased over follow-up and was associated with worse mobility and function. Each one-point increase in cervical damage was associated with a 1.27° reduction in cervical rotation and each point increase in lumbar damage was associated with a 0.29-cm reduction in lumbar lateral flexion and a 0.06-cm reduction in Schober's test. Bridging syndesmophytes showed substantially larger effects than non-bridging lesions; in adjusted models, compared with no damage at the same level, bridging at C2-C3 was associated with a 35.5° lower cervical rotation, and bridging at L3-L4 with a 7.3 cm lower lumbar lateral flexion (both P < 0.0001). In standardized models, higher damage was associated with lower cervical rotation (β = -0.53), lower lumbar lateral flexion (β = -0.40) and worse BASFI (β = 0.31). Longitudinal DVU-level modelling translates radiographic spinal damage into anatomically resolved, clinically interpretable losses in mobility and function, extending beyond global damage scores in r-axSpA.