Article Text

THU0425 Associations between radiographic and clinical osteoarthritis features and MRI-defined bone marrow lesions in the finger joints
  1. I.K. Haugen1,
  2. P. Bøyesen1,
  3. B. Slatkowsky-Christensen1,
  4. S. Sesseng2,
  5. D. van der Heijde1,3,
  6. T.K. Kvien1
  1. 1Rheumatology
  2. 2Radiology, Diakonhjemmet Hospital, Oslo, Norway
  3. 3Rheumatology, Leiden University Medical Center, Leiden, Netherlands


Background Bone marrow lesions (BMLs) are associated with pain in hand OA (1), but it is currently not known whether they are related to biomechanical or inflammatory processes.

Objectives To explore whether radiographic OA features and inflammation are associated with BMLs in a cross-sectional study of hand OA patients.

Methods We included 108 patients (98 women, mean (SD) age 68.8 (5.6) years) from the Oslo hand OA cohort with available MRIs (coronal/axial STIR images, 1.0T), radiographs (posteroanterior view) and clinical joint examination of soft tissue swelling of the 2nd-5th distal (DIP) and proximal interphalangeal (PIP) joints of the dominant hand. One reader read the STIR images for presence of BMLs according to the proposed Oslo hand OA MRI scoring system (2). The hand radiographs were scored for presence of radiographic osteophytes, joint space narrowing (JSN), erosions, cysts, sclerosis and malalignment according to the OARSI atlas. We examined whether radiographic hand OA features (markers of structural pathology) and clinical soft tissue swelling of the joints (marker of inflammation) were associated with presence of MRI-defined BMLs using logistic regression with Generalized Estimating Equations. Features that were associated (p<0.25) with BMLs in univariate analyses (adjusted for age and sex) were included in a multivariate model. The multivariate model included features that were associated with BMLs (p<0.10) after backward selection (adjusted for age and sex).

Results Patients with hand OA had low prevalence MRI-defined BMLs in the DIP and PIP joints of the dominant hand; the median (interquartile range; IQR) number of joints with BMLs was 1 (0-2). Radiographic JSN (grade 1-3) were present in the majority of joints (median 7, IQR 6-8), and we therefore defined presence of JSN as grade 2-3. All features except sclerosis were associated with presence of BMLs in the adjusted univariate analyses and were included in the multivariate model (table). The final multivariate model included JSN, malalignment, clinical soft tissue swelling and cysts (table), of which the latter was borderline significant.

Table 1. Median number of affected DIP and PIP joints and the association to presence of BMLs (logistic regression analyses with Generalized Estimating Equations)

Conclusions Radiographic JSN and malalignment and clinical soft tissue swelling were strongly associated with presence of MRI-defined BMLs in the DIP and PIP joints in this cross-sectional study. Whether BMLs in hand OA are caused by increased loading and bone trauma (as in knee OA) and/or inflammation needs to be proven in future longitudinal studies.

  1. Haugen et al. ARD 2011 Nov 25 [Epub ahead of print].

  2. Haugen et al. ARD 2011;70:1033.

Disclosure of Interest None Declared

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.