Background In patients with rheumatoid arthritis (RA) bone marrow edema (BME) as observed by magnetic resonance imaging (MRI) represents osteitis with infiltration of leucocytes and an increased number of osteoclasts. Both BME and anti-citrullinated protein antibodies (ACPAs) are predictors of radiographic progression in RA and recent data indicate that ACPA can directly activate osteoclasts. These findings together lead to the hypothesis that ACPA is associated with BME; indeed, two small studies observed an association between BME and ACPA. Thus far the association of ACPA with other forms of MRI-detected inflammation (synovitis and tenosynovitis) has not been thoroughly explored. Furthermore, it is unknown if the association with MRI-detected inflammation is only present for ACPA or also for other RA-related auto-antibodies.
Objectives This study aimed to investigate the associations of ACPA, RF and antiCarP with BME, synovitis and tenosynovitis.
Methods A total of 589 early arthritis patients, included in the Leiden Early Arthritis Clinic cohort, underwent contrast-enhanced 1.5T MRI of the unilateral wrist, metacarpophalangeal and metatarsophalangeal joints at baseline. BME, synovitis and tenosynovitis were scored according to the RAMRIS-method. ACPA, rheumatoid factor (RF) and anti-carbamylated protein (antiCarP) antibodies were determined at baseline.
Results BME, synovitis and tenosynovitis were concomitantly present on MRI. In univariable analyses, ACPA-positive patients had higher BME-scores than ACPA-negative patients (median scores 4.5 vs. 2.0, p<0.001), but not more synovitis and tenosynovitis. Besides ACPA, also RF (median scores 3.75 vs 2.0, p<0.001) and antiCarP antibodies (median scores 3.5 vs 2.5, p=0.012) were associated with higher BME-scores in univariable analyses. To explore the association of the different antibodies with BME the BME-scores of patients with different auto-antibody combinations were compared. ACPA+RF-antiCarP- patients did not have higher BME-scores than ACPA-RF-antiCarP- patients, however ACPA+RF+antiCarP- and ACPA+RF+antiCarP+ patients had higher BME-scores (median 5.0 and 4.5 versus 2.0 respectively, p<0.001 and p<0.001). ACPA and RF levels were not associated with BME-scores.
Conclusions The presence and the level of ACPA alone were not associated with BME-scores. However, the combined presence of ACPA and RF did associate with more BME. This suggests an additive role of RF to ACPA in mediating osteitis.
Disclosure of Interest None declared