Measuring spinal mobility in early axial spondyloarthritis: does it matter?

Abstract

To investigate the frequency and order of impairment of spinal mobility measures (SMMs) and their cross-sectional and longitudinal usefulness in early axial spondyloarthritis. SMMs measurements of patients from the DESIR (5-year data) and SPACE (2.6 (1.9) years of follow-up) cohorts were analysed. Cross-sectional (group level) and longitudinal (individual level) analyses were performed comparing SMMs to pre-defined cut-offs derived from healthy individuals. Subgroup analyses were used to study patient and disease characteristics potentially influencing spinal mobility. Reliability was analysed using intraclass correlation coefficients and the smallest detectable change. In 328 DESIR and 148 SPACE patients, lateral spinal flexion (LSF) and mSchober were the most impaired SMMs. If both (LSF and mSchober) were measured, 84% (DESIR) and 74% (SPACE) of the patients with impairment in ≥1 SMM would be captured. LSF and Bath AS Metrology Index best discriminated between subgroups of patients (higher impairment in patients ever treated with biologics, with higher disease activity and presence of baseline syndesmophytes): e.g. 31% of LSF impairment in patients with Ankylosing Spondylitis Disease Activity Score (ASDAS) < 2.1 in ≥2/3 visits vs 49% in those with ASDS ≥ 2.1. A high variability in SMMs within the same patient over time was observed, even when restricting the analysis to patients with low disease activity. Reliability of SMMs was 'fair' to 'good' (inter-reader intraclass correlation coefficients (2, 1): 0.55-0.84; intrareader intraclass correlation coefficients (2, 1): 0.49-0.72). Smallest detectable changes were in general high, e.g. 5.1 cm for LSF. Cross-sectional use of SMMs, at the group level, is informative in patients with early axial spondyloarthritis. However, the high variation of SMMs over time impairs their use, at the individual patient level.