Background/Purpose:

Kellgren-Lawrence grade (KL) 1, defined by a questionable osteophyte and/or doubtful joint space narrowing, is thought to be an intermediate stage from non-OA (i.e., KL 0) to definite radiographic osteoarthritis (i.e., KL ≥2, ROA). MRI allows the identification of specific structural changes in joint tissues that precede the development of ROA. We examined whether the morphologic changes on MRI that precede ROA differ for knees that are KL 0 vs. those that are KL 1.

Methods:

The participants in this nested case-control study were from the Osteoarthritis Initiative (OAI) cohort with risk factors for OA but not ROA in the target knee (i.e., KL 0 or 1) at baseline. Participants were assessed by knee radiographs and non-contrast 3T MRI using the OAI protocol. Baseline MRIs were assessed for Hoffa-synovitis, effusion-synovitis, bone marrow lesions (BMLs), meniscal damage or meniscal extrusion using the MRI Osteoarthritis Knee Score (MOAKS). Case knees were those that developed definite ROA on the PA fixed flexion radiographs at the 12M, 24M, 36M or 48M OAI visits. The control knees (i.e., did not develop incident ROA over the four-year period) were 1:1 matched by age (within five years), sex, and baseline KL in the target knee and the contralateral knee (e.g., KL 0/0, 0/1/, 1/1, 0/2, or 1/2).  Conditional Logistic Regression was used to estimate the odds ratios (OR) of subsequent incident ROA for each of the MRI features.

Results:

355 knees that developed ROA (67% female) were matched to 355 controls with a mean age of 60.1 (SD 8.5). The majority were overweight: mean body mass index was 28.9 (SD 4.5) in cases and 27.7 (SD 4.4) in controls. For knees that were KL 0 at baseline (n=133 cases/controls pairs), the MRI predictors of subsequent ROA were Hoffa synovitis (OR 1.80 and 95% CI [1.10-2.94]); meniscal damage (tear or maceration) in the posterior medial horn (OA 4.11 [1.98-8.52] or the anterior lateral horn (OR 9.00 [1.14-71.04]); and number of subregions in the medial compartment with meniscal tear or extrusion (n=1 subregions, OR 4.14 [1.86-9.25], ≥2, OR 4.42 [1.46-13.43]. For knees that were KL 1 at baseline (n=222 cases/controls pairs), the MRI predictors of subsequent ROA were Hoffa synovitis (OR 1.97 [1.31-2.96]; effusion-synovitis (OR 1.65 and [1.11-2.44]; lateral meniscal extrusion (OR 2.80 [1.01-7.77]; and number of BMLs (=1, OR 1.98 [1.22-3.20]; =2, OR 2.20 [1.31-3.68]; and ≥3, OR 2.21 [1.22-4.02]). Only 27/133 (20%) of KL 0 knees at baseline were KL1 at any reading before being KL ≥2.  

Conclusion:

The altered joint determinants on MRI that precede the development of ROA differ for KL 0 knees vs. KL1 knees suggesting that different tissues are involved at the different stages of OA development.   These findings have important implications for the development of potential preventative and/or disease modifying OA drugs (DMOADs). 

---

« Back to 2013 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/kellgren-lawrence-grade-kl-0-vs-kl-1-knees-differ-in-their-mri-predictors-of-subsequent-radiographic-osteoarthritis/