Background/Purpose: The MOBILITY study, a study on spinal mobility measures in healthy individuals, showed that the Bath Ankylosing Spondylitis Metrology Index (BASMI) often overestimates spinal mobility as these measures also depend on height, sex and age, with a significant impact on older and taller individuals. To better address this, we developed a Corrected AxSpA Metrology Index (CASMI) and compared its construct validity with that of BASMI in patients with established axSpA.

Methods: Data from the MOBILITY study on spinal mobility in healthy individuals were used to develop a corrected version of the BASMI (CASMI). Age-, sex-, and height-specific normal limits were derived for each component (lateral spinal flexion -LSF-, tragus-to-wall -TTW-, intermalleolar distance -IMD-, Schobert’s test, cervical rotation -CR-) from regression models or trigonometric calculations. Components were anchored using the 50th percentile of healthy individuals (MOBILITY study, score = 0) and the 95th percentile of axSpA patients (cohort study of axSpA used, score = 10). TTW was corrected using fixed thresholds. Construct validity of several CASMI versions were compared to BASMI in an axSpA population included in the OASIS cohort. Construct validity was assessed by comparing means, minimum and maximum values, floor and ceiling effects of BASMI and CASMI, and correlations with age, height and disease outcomes (BASFI, BASDAI, ASDAS, mSASSS and number of syndesmophytes). It was hypothesized that CASMI would reduce correlations with age and height while maintaining correlations with disease measures. Differences between BASMI and CASMI in the score distributions were visualized with density plots. Known group discrimination, as part of construct validity, was tested by comparing standardized mean differences (SMD) between known groups, stratified by: i) BASFI (≤3 vs ≥6), ii) mSASSS (≥1 vs >10), and iii) syndesmophyte count (≤1 vs 5). Higher absolute SMD values indicate better discrimination, with >0.8 considered as ‘good’.

Results: Table 1 presents the formulae for the five components of BASMI, as well the formulae for the CASMI components. In the OASIS cohort, mean values for some corrected components were lower, whereas minimum values remained unchanged, and were slightly lower for CASMI. Floor effects increased in individual items (e.g., TTW: 0.5% to 52%) but remained 0% in the composite scores. Maximum values and ceiling effects increased in individual components but not in CASMI. As intended, correlation with age and height decreased in CASMI and several components, but no major changes in correlations with disease outcomes, confirming the hypotheses made (Table 2). The density plots showed that CASMI systematically yields lower scores and has a higher floor effect, suggesting better calibration and better sensitivity to capture subtle impairments in spinal mobility. Discrimination between known groups was good across all score versions and external constructs (SMD >0.80).

Conclusion: An age-, sex- and height-corrected mobility score, the CASMI, has been developed as a more personalised and truthful measure of mobility in axSpA, compared to BASMI.

Figure 1. Density plots for BASMI and CASMI. CASMI represented here is CASMI_Ma which formula was adjusted using age, sex and hight correction factors derived from the regression coefficients obtained in the MOBILITY study.

« Back to ACR Convergence 2025

ACR Meeting Abstracts - https://acrabstracts.org/abstract/development-and-validation-analysis-of-a-corrected-axial-spondyloarthritis-metrology-index/