Costovertebral joint involvement in patients with axial spondyloarthritis

To evaluate costovertebral joint involvement in patients with axial spondyloarthritis (axSpA) and to assess its association with disease features. We included 150 patients from the Incheon Saint Mary's axSpA observational cohort who underwent whole spine low-dose computed tomography (ldCT). Costovertebral joint abnormalities were scored by two readers on a scale of 0–48 based on the presence or absence of erosion, syndesmophyte, and ankylosis. The interobserver reliability of costovertebral joint abnormalities was assessed using intraclass correlation coefficients (ICCs). Associations between costovertebral joint abnormality scores and clinical variables were evaluated using a generalized linear model. Two independent readers found costovertebral joint abnormalities in 74 (49%) patients and 108 (72%) patients. The ICCs of scores for erosion, syndesmophyte, ankylosis, and total abnormality were 0.85, 0.77, 0.93, and 0.95, respectively. For both readers, total abnormality score was correlated with age, symptom duration, Ankylosing Spondylitis Disease Activity Score (ASDAS), Bath AS functional index (BASFI), CT syndesmophyte score (CTSS), and number of bridging spines. Multivariate analyses showed age, ASDAS, CTSS to be independently associated with total abnormality scores in both readers. The frequency of ankylosed costovertebral joint was 10.2% (reader 1) and 17.0% (reader 2) in patients without radiographic syndesmophytes (n = 62), and 10.3% (reader 1) and 17.2% (reader 2) in patients without radiographic sacroiliitis (n = 29). Costovertebral joint involvement was common in patients with axSpA, even in the absence of radiographic damage. LdCT is recommended for evaluating structural damage in patients with clinically suspected costovertebral joint involvement.