Background/Purpose: Both zygapophyseal joint (ZJ) fusion and syndesmophytes contribute to structural spine damage in Ankylosing Spondylitis (AS). However, because ZJs are difficult to visualize on radiographs, little is known about the prevalence of ZJ fusion and its relationship to other features of spinal damage in AS. In particular it is not known if ZJ fusion occurs before syndesmophytes and which feature contributes most to spinal rigidity. We used computed tomography (CT) to visualize the ZJ and investigate those questions.

Methods: 50 patients (84% men) underwent thoracolumbar CT scans (Table 1). Six vertebral levels were examined (T10-T11 to L3-L4). Anonymized scans were read for the presence of ZJ fusion at each level after digitally masking the vertebral bodies to blind readers to the presence of syndesmophytes. Using a validated semi-automated computer algorithm, we quantitated syndesmophytes at the same vertebral levels. We measured syndesmophyte height in 72 angular sectors of 5 degrees around the vertebral rim. We then summed the results of the 72 angular sectors to yield the circumferential height. We used Pearson correlations to evaluate the association between ZJ fusion and the lumbar range of motion and multiple regression analysis to evaluate the relative contributions of ZJ fusion and syndesmophytes to spinal motion.

Results: Agreement on the presence of ZJ fusion between the two physicians’ readings was good (kappa=0.84). 56% of patients had ZJ fusion at 1 or more vertebral levels.  The median number of vertebral levels with ZJ fusion was 1 (range 0 -6).  The 3 most superior vertebral levels were more often fused (39% of patients) than the lower 3 (15% of patients). Fusion was most often bilateral. Syndesmophytes were often present in vertebral levels without ZJ fusion. In all levels with syndesmophytes, only 35% had ZJ fusion. ZJ fusion was present in only 7% of vertebral levels without syndesmophytes.   Bridging was also often present in levels without ZJ fusion.  In levels with an extent of bridging < 50 degrees, only 35% had ZJ fusion. These results suggest that ZJ fusion most often follows the development of syndesmophytes.  The number of levels with ZJ fusion correlated with the Schober test (r=−0.58, p<0.0001) and lateral thoracolumbar flexion (r=−0.58, p<0.0001). Syndesmophytes and ZJ fusion were equally associated with the Schober test (Table 2). Syndesmophytes were more strongly associated with lateral thoracolumbar movement.

Conclusion: ZJ fusion is common in AS, particularly in the T10-T11to T12-L1 levels. Although syndesmophytes seem to occur first, ZJ fusion is also important and contributes to spinal rigidity.