Arthritis & Rheumatism, Volume 62,
November 2010 Abstract Supplement

Abstracts of the American College of
Rheumatology/Association of Rheumatology Health Professionals
Annual Scientific Meeting
Atlanta, Georgia November 6-11, 2010.

MRI-Defined Risk Factors for Cartilage Loss over a 6 Months Period.

Roemer1,  Frank W., Kwoh2,  C. Kent, Hannon2,  Michael J., Green2,  Stephanie M., Jakicic2,  John M., Boudreau2,  Robert, Crema1,  Michel D.

Boston University
University of Pittsburgh


Cartilage morphology may only be directly assessed by MRI. Cartilage loss over short observational periods is rare but not uncommon. Aim of the study was to assess several non-MRI and MRI-features as baseline (BL) risk factors that may predict patello-femoral (PF) and tibio-femoral (TF) cartilage loss over a 6 months follow-up (FU) period.

Materials and Methods:

The JOG study is a double-blind randomized trial investigating the effect of oral glucosamine supplementation over a 6 months period. 177 subjects aged 35–65 with chronic, frequent knee pain were included. 3 T MRI of both knees was performed at BL and 6 months FU using the same pulse sequence protocol as in the Osteoarthritis Initiative. Knees were semiquantitatively assessed according to the WORMS system. Cartilage status was scored on a scale from 0–6. A change of +0.5 in any subregion, indicating within-grade progression, was defined as the minimum requirement for cartilage loss. Bone marrow lesions (BMLs) in the same subregion, meniscal damage or meniscal extrusion in the TF compartment, synovitis and effusion were scored in addition and assessed as baseline risk factors.

All MR joint morphologic features were dichotomized into present (>= 1) vs. absent (0). Logistic regression models were applied to assess the risk of cartilage loss for knees exhibiting risk factors when compared to knees without the risk factor at baseline. We performed a subregion-based analysis using general estimating equations (GEE) to account for the clustering of subregions within a knee and knees within an individual. Multivariate models were adjusted for age, gender, treatment (oral glucosamine) and BMI. In addition MRI-based risk factors were adjusted for each other in a multi-adjusted model.


51.2 % of participants were men, mean age was 52.3 (± 6.2), mean BMI was 29.1 (± 4.1). Baseline Kellgren/Lawrence grades (for worst knee) were: K/L 0: 37 (20.9%) knees, K/L 1: 14 (7.9%) knees, K/L 2: 26 (14.7%) knees, K/L 3: 81 (45.8%) knees K/L 4: 19 (10.7%). Of the 353 knees, 304 knees (87.9%) and 1,153 subregions (23.8%) exhibited prevalent cartilage damage at baseline. 79 (1.6%) subregions showed incident or worsening cartilage damage at 6 months FU. Predictors for PF cartilage loss at 6 months were baseline presence of effusion and prevalent cartilage damage in the same subregion. Risk factors for TF cartilage loss were baseline ipsi-compartmental meniscal extrusion, prevalent BMLs and cartilage damage in same subregion.


Cartilage loss over 6 months is rare, but may be detected semiquantitatively by MRI. The strongest predictors of PF cartilage loss were presence of baseline effusion and prevalent cartilage damage in the same subregion. The greatest predictor of TF cartilage loss was also prevalent cartilage damage, but prevalent BMLs and meniscal extrusion were also significant predictors of compartment-specific TF cartilage loss.

Table 1. Risk factors for patello-femoral and tibio-femoral cartilage loss over 6 months of follow-up

Risk factorReferenceAdjusted Odds Ratio3 (95% confidence intervals)
Effusion (WORMS >= 1)1Effusion absence (WORMS score = 0)3.54 (1.30–9.64)*
Synovitis (modified WORMS >= 1)1Synovitis absence (modified WORMS score = 0 in both synovitis subregions)0.79 (0.28–2.07)
Prevalent cartilage damage (WORMS >=2)2No cartilage damage in subregion (WORMS score = 0 or 1)4.32 (1.35–13.85)*
BML (WORMS >= 1)2No BML in subregion (WORMS score = 0)1.61 (0.67–3.84)
Effusion1 (WORMS >= 1)1Effusion absence (WORMS score = 0)1.79 (0.76–4.23)
Synovitis1 (modified WORMS >= 1)1Synovitis absence (modified WORMS score = 0)0.68 (0.32–1.45)
Meniscal damage4No meniscal damage1.98 (0.76–5.15)
Meniscal extrusion4No meniscal extrusion3.62 (1.29–10.12)*
Prevalent cartilage damage (WORMS >= 2)2Absence of cartilage damage in subregion (WORMS = 0 or 1)15.90 (5.08–49.79)*
BML(WORMS >= 1)2No BML in subregion (WORMS = 0)4.58 (1.08–19.44)*
1cartilage loss at 6 months in any of 4 PF/10 TF subregions
2cartilage loss or worsening of BML in same subregion at 6 months follow-up
3multi-adjusted GEE model accounting for correlations within and between knees, and adjusted for age, gender, BMI, treatment, and the other MR features
4cartilage loss in same compartment as meniscal damage or extrusion (5 subregions medial or lateral)
*statistically significant at p <= 0.05

To cite this abstract, please use the following information:
Roemer, Frank W., Kwoh, C. Kent, Hannon, Michael J., Green, Stephanie M., Jakicic, John M., Boudreau, Robert, et al; MRI-Defined Risk Factors for Cartilage Loss over a 6 Months Period. [abstract]. Arthritis Rheum 2010;62 Suppl 10 :193
DOI: 10.1002/art.27962

Abstract Supplement

Meeting Menu