Background/Purpose: We aimed to quantify the magnitude of biologic effect on sacroiliitis in juvenile SpA by comparing the change in the Spondyloarthritis Research Consortium of Canada (SPARCC) sacroiliac joint inflammation score (SIS) over time in TNF inhibitor (TNFi)-exposed versus TNFi-na•ve patients.

Methods: This retrospective multi-center cohort study included 34 children with SpA.  Children were included if they met ESSG SpA or ILAR juvenile arthritis criteria for enthesitis-related arthritis and had at least 2 pelvic MRI studies separated by at least 12 weeks between 2005 and 2018. Images were de-identified, randomized and reviewed independently by 3 radiologists and assigned a SPARCC SIS, with mean score used for analysis.  The SIS divides the joint into quadrants and scores presence, depth, and intensity of bone marrow edema on STIR MRI (total score 0-72). We used a recency-weighted cumulative exposure model to quantify TNFi exposure in which higher weights were given to more recent exposure relative to the imaging date. The cumulative exposure scores were applied to a linear mixed-effects  regression to test the association of TNFi exposure and change in SIS over time, adjusted for baseline SIS score. Kruskal-Wallis and chi-squared tests were used to assess relationships between clinical findings and SIS scores, as appropriate. SpearmanÕs correlation was used to test association of change in clinical factors with the outcome.

Results: Baseline clinical features are in Table 1. The median time between images was 22 months (IQR 13-34.1). Ten (29%) children received a TNFi prior to baseline imaging, with a median exposure of 10 mos (IQR 3.9-24.4). Of the 26 patients (76%) with TNFi exposure between images, median exposure was 15.6 months (IQR 9-25.8). The mean SIS change score was 6.9 (SD 16.4; range -36 to 47).  Patients with baseline SIS >1 had higher physician global assessments (p= 0.02), fewer tender entheses (p= 0.02), and shorter disease duration (p= 0.05). Cumulative TNFi exposure was significantly associated with a decrease in SIS over time (b=-0.02, p=0.01). Of those with progression in SIS from 0 to >0 (n = 4), two were exposed to a TNFi, but were off therapy for at least six months preceding the second MRI. There was no significant correlation between change in pain score and SIS (p=0.09).  

Conclusion: Biologics have a quantifiable effect in reducing inflammation at the sacroiliac joints in children with SpA. Future studies should evaluate the rate of change in SIS in response to biologics to help guide appropriate timing for follow-up imaging in both clinical practice and efficacy studies.