Background Anti-citrullinated protein antibodies (ACPA) are a major risk factor for bone destruction in rheumatoid arthritis (RA). ACPA may directly induce bone loss by stimulating osteoclast differentiation . Accordingly, the trabecular bone microarchitecture at peripheral sites is significantly compromised in ACPA(+) subjects irrespective of joint inflammation [2,3]. Whether ACPA positivity also influences systemic bone changes is however unknown.
Objectives To investigate the relationship between generalised bone loss and the ACPA status in patients with early RA.
Methods We evaluated 102 consecutive untreated patients with early RA (symptoms duration <12 months). Bone mineral density (BMD) was assessed at the lumbar spine (L1-L4) and hip by dual X-ray absorptiometry. Demographic and clinical data were collected, including: age, gender, menopausal status, body max index (BMI), current smoking and alcohol intake, other risk factors for secondary osteoporosis. Disease variables evaluated included symptoms duration, the disease activity score 28 (DAS28), the swollen and tender joint count, acute phase reactants, and ultrasonographic (US) scores for Gray Scale and Power Doppler signals at the hands and wrists.
Results The study population included 73 women (63% in post-menopause) and 29 men, with a mean age of 57 (±14) years and a median disease duration of 3 (IQR 1.9-6) months. All patients had active disease, with a mean DAS28 of 4.41 (±1.11). Thirty-nine patients (38.2%) were ACPA(+), with ACPA values >3 times the upper limit of normal (ULN) in 89.7% of the cases. In the overall group, BMD measurements at baseline were almost comparable to the normal range when corrected for age and sex. Z scores were 0.03 (±1.31) at the lumbar spine, and -0.11 (±0.97) at the femoral neck. Z scores were not significantly influenced by disease variables such as symptoms duration, overall disease activity, the clinical and the US joint count, and acute phase reactants. In contrast, ACPA(+) patients had lower Z scores at the lumbar spine compared to ACPA(-) patients (-0.33 [±1.11] vs 0.26 [±1.38], p=0.03), whilst no differences were found at the femoral neck. Furthermore, in ACPA(+) patients, the lumbar Z score was inversely related to ACPA titers (r -0.33, p=0.04 for log-transformed ACPA). Accordingly, 40% of the patients with ACPA values >3 times the ULN had a Z score <-1 compared with 17.5% of ACPA(-) patients (p=0.03). Despite similar disease duration and activity, ACPA(+) patients showed a non-significant increase in C-reactive protein levels (p=0.07). However, at multiple regression analysis, ACPA emerged as the only disease variable significantly predicting the BMD at the spine independent of gender, age, BMI and other risk factors for secondary osteoporosis (p=0.04).
Conclusions Vertebral BMD in patients with early RA is significantly reduced in relation with ACPA positivity. This finding supports the hypothesis that autoimmunity associated with RA may have a significant role in bone remodeling, either directly by stimulating osteoclasts or indirectly by eliciting subclinical inflammation.
Harre U, et al. J Clin Invest 2012;122:1791-802.
Kocijan R, et al. Ann Rheum Dis 2013;doi: 10.1136/annrheumdis-2013-203791.
Kleyer A, et al. Ann Rheum Dis 2013;doi:10.1136/annrheumdis-2012-202958.
Disclosure of Interest None declared