Disease activity assessment in ankylosing spondylitis in a Chinese cohort: BASDAI or ASDAS?

Recently, the Ankylosing Spondylitis Disease Activity Score (ASDAS), a new index, has been shown to be validated and highly discriminatory in assessing ankylosing spondylitis (AS) disease activity. This study is to evaluate the performance of ASDAS in a local Chinese cohort of AS in a cross-sectional setting and to compare it with the existing instrument, the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI). Consecutive patients with AS were recruited from a local rheumatology clinic. Data, including BASDAI, Bath Ankylosing Spondylitis Functional Index (BASFI), Visual Analogue Scale (VAS) for spinal pain, and patient and physician global assessments were gathered during clinic visit. Inflammatory markers, including erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), and high-sensitivity (hs)-CRP were collected. ASDAS was calculated accordingly. The discriminatory capacity of BASDAI and ASDAS was compared by: (1) standardized mean difference statistics, (2) R2 in linear regressions, and (3) area under receiver operating characteristic curve (AUC) in logistic regression models. Both ASDAS and BASDAI showed satisfactory predictive value on disease activity with reference to patient and physician global assessment. R2 in linear regression models ranged from 0.6–0.7. Both indices also demonstrated good discriminatory capacity as evidenced by a relatively high AUC (> 0.8) under the logistic regression models using either patient or physician global assessment score ≥4 and <4 as cut off of high and low disease activity status, respectively. Although we could not demonstrate significant differences in the performance between them, subgroup analysis suggested better discriminatory ability of ASDAS in the high inflammatory marker subgroup. ASDAS and BASDAI showed similarly good performance in a cross-sectional setting in a local Chinese AS cohort. ASDAS performed better in subgroup with raised inflammatory markers.