Background/Purpose: Early diagnosis and treatment initiation improve outcomes for patients with rheumatoid arthritis (RA). Delays to diagnosis and treatment affect patients with seronegative vs seropositive RA disproportionately. Magnetic resonance imaging (MRI) is superior to clinical examination for detecting synovitis in patients with RA and can predict progression to clinical RA from undifferentiated arthritis. We hypothesized that MRI utilization is higher in seronegative vs seropositive patients with early RA, and is increasing over time in recent years.

Methods: A retrospective study was performed of a population-based cohort of patients with incident RA between 1999-2014 who fulfilled either the 1987 or 2010 ACR/EULAR classification criteria. Patients who underwent MRI within (+/-) 1 year of fulfillment of the earlier of either classification criteria for RA were identified using Current Procedural Terminology codes corresponding to MRI studies of the extremities (hand, wrist, elbow, shoulder, foot, ankle, knee, hip). Studies were confirmed via manual chart review. MRIs performed for indications unrelated to arthritis were excluded. MRI utilization was compared between seropositive (RF+ and/or CCP+) and seronegative (RF-, CCP-) patients. Factors associated with undergoing MRI within one year of meeting criteria were analyzed. Comparisons between groups were made using Kruskal-Wallis and Chi-square tests. Utilization of MRI over time was evaluated with a test for trend.

Results: 789 patients with incident RA who fulfilled 1987 and/or 2010 ACR/EULAR classification criteria for RA were included. Of these, 554 (70%) were female, 709 (90%) Caucasian, 534 (68%) RF and/or CCP positive, mean age 55.9 (SD 15.2) at time of meeting criteria. At least one MRI was performed within one year of fulfilling criteria in 165 (21%) of patients over the study interval. Age, sex, race/ethnicity, and smoking status did not differ between groups who did and did not undergo MRI.

RF and CCP autoantibodies were negative in 77 (47%) of those who underwent MRI, versus 178 (29%) who did not (p< 0.001). At time of meeting classification criteria, presence of morning stiffness rheumatoid nodules, erosive disease, and abnormal sedimentation rate or C-reactive protein were similar among patients who did and did not undergo MRI. Patients with symptom duration ≥ 6 weeks (64% vs. 46%, p< 0.001) and involvement of 4-10 (22% vs. 13%) small joints or >10 (39% vs. 30%) joints (p< 0.001) were more likely to have an MRI performed. Prevalence of obesity (BMI ≥ 30 kg/m²) was similar between MRI (45%) and no MRI (39%) (p=0.18) groups.

The median year of fulfilling RA criteria was later in the MRI (2010; IQR 2005-2013) vs no MRI (2008; IQR 2003-2011) (p< 0.001) groups. MRI utilization overall increased over time by 9% per calendar year (OR:1.09; 95% CI: 1.05-1.14; p< 0.001).

Conclusion: In this population-based cohort of patients with incident RA, MRI was used for evaluation of arthritis within one year of fulfilling classification criteria in 21% of patients. MRI was more commonly used in seronegative than seropositive patients, which may reflect diagnostic uncertainty in the seronegative group. MRI utilization increased over time during the 15-year study interval.

« Back to ACR Convergence 2021

ACR Meeting Abstracts - https://acrabstracts.org/abstract/utilization-of-magnetic-resonance-imaging-in-early-rheumatoid-arthritis-results-from-a-population-based-cohort-1999-2014/