Background/Purpose: Understanding factors underlying initial development of knee OA is crucial to effective prevention strategy design.  Our goals were to: 1) determine extent of tissue damage on MRI in knees of persons at higher risk for knee OA and in whom both knee x-rays were normal (KL 0); and 2) evaluate whether presence of any bone marrow lesions (BML), meniscal tears (MT), meniscal extrusion (ME), or hand OA is associated with risk of incident cartilage damage.

Methods: The Osteoarthritis Initiative (OAI) is a cohort study of men and women, 45-79 years, all with or at increased risk to develop knee OA.  850 participants had had bilateral KL 0 at 12 months (baseline for this study) by centralized reading.  On their right knee MRIs, we undertook assessment of cartilage morphology, BML, MT, and ME using a modified MOAKS scoring system.  Readers were blinded to hypotheses, clinical data, and KL grade.  The definition of hand OA relied upon number of bony enlargements.  12 and 48 month image assessments occurred within an ancillary study; in addition, OAI clinical data V6.2.1 and BU x-ray reading data V1.5 were used.  Mulitiple logistic regression models were used to evaluate associations between baseline data and incident cartilage damage by 3 year follow-up; results are reported as adjusted odds ratios (aORs) with 95% CIs.

 Results: 850 persons met criteria for inclusion [mean age 59.6 years (8.8, SD), mean BMI 26.7 kg/m² (4.2), 475 (56%) women].  Among the 850 KL 0 knees, the number with abnormal tissue in one or more subregions at baseline was: 483 (57%) with cartilage damage (full-thickness in 67); 353 (42%) with BML; 180 (21%) with MT; and 117 (14%) with ME.  In only 56 (7%) knees were all of these tissues normal.  367 persons [age 58.5 (8.8), BMI 26.4 (4.2), 226 (62%) women] contributed 367 knees with normal cartilage morphology in all regions at baseline.  Of the 367, 80 had BML, 45 had MT (32 horizontal, 9 vertical, 4 partial maceration), and 25 had ME.  Lesions coexisted in some knees (11 with BML+MT, 7 with BML+ME, and 10 with MT+ME).  In separate models, each adjusting for age, gender, BMI, and hand OA, BML [aOR 2.24, 95% CI (1.15, 4.34)] and ME [aOR 2.83, 95% CI (1.03, 7.79)] were each associated with incident cartilage damage.  In a model including all covariates, BML continued to be significant [aOR 2.19, 95% CI (1.12, 4.28)], while ME was not quite significant [aOR 2.71, 95% CI (0.97, 7.58)].  Of note, hand OA presence was consistently associated with incident cartilage damage, e.g., in the fully adjusted model, aOR 2.30, 95% CI (1.15, 4.60) for this variable.

 Conclusion: Cartilage damage was already present in 57% of right knees from persons without OA in either knee by protocolized x-ray, but at higher risk, suggesting that radiographic studies of incident OA in this population, even restricted to KL 0, are frequently evaluating not incidence but early progression.  Among knees with normal cartilage morphology, hand OA, BML, and ME were each associated with an increased risk of incident cartilage damage.  While MRI is superior to x-ray, the optimal window for cohort studies to capture the onset of knee OA by MRI may fall earlier in the lifetime of individuals at higher risk than is being evaluated in many current studies. Prevention strategies may be most powerful at this time point.

---

« Back to 2012 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/tissue-lesions-in-osteoarthritis-initiative-participants-with-normal-x-rays-and-risk-factors-for-incident-cartilage-damage/