Structural damage and inflammation on radiographs or magnetic resonance imaging are associated with cortical interruptions on high-resolution peripheral quantitative computed tomography: a study in finger joints of patients with rheumatoid arthritis and healthy subjects

Objectives: To study the relationship between structural damage and inflammatory features on magnetic resonance imaging (MRI) or radiography and other risk factors [anti-citrullinated protein antibody (ACPA) and/or rheumatoid factor (RF) seropositivity, hand dominance, disease duration] and the presence or number of cortical interruptions in finger joints on high-resolution peripheral quantitative computed tomography (HR-pQCT). Method: Finger joints of 38 healthy subjects and 39 patients with rheumatoid arthritis (RA) were examined through radiographs, MRI, and HR-pQCT. Radiographs were scored according to the Sharp/van der Heijde (SvH) method; MRI for the presence of cortical interruptions, bone marrow oedema (BMO), and synovitis; and HR-pQCT images for cortical interruptions. Descriptive statistics were calculated and associations examined using generalized estimating equations. Results: Cortical interruptions were found in healthy subjects and patients with RA on HR-pQCT (mean ± sd 0.33 ± 0.63 vs 0.38 ± 0.64 per joint quadrant, respectively, p < 0.01). Structural damage on MRI (cortical interruptions) or radiographs (SvH ≥ 1) was associated with the presence of cortical interruptions on HR-pQCT [odds ratio (OR) 12.4, 95% confidence interval (CI) 7.5–21.4, p < 0.01 and OR 4.8, 95% CI 1.9–11.7, respectively, p < 0.01]. The presence of BMO or synovitis was associated with more cortical interruptions on HR-pQCT (β 0.47, 95% CI 0.4–0.6, p < 0.01 and β 1.9, 95% CI 0.6–3.1, p < 0.01). In patients with RA, ACPA, and/or RF seropositivity, hand dominance and disease duration were not associated with more cortical interruptions on HR-pQCT. Conclusion: Structural damage and inflammatory features on MRI and radiographs are associated with cortical interruptions on HR-pQCT. No association between other risk factors and cortical interruptions was demonstrated.