Background/Purpose: Cartilage volume loss is considered a key feature of knee osteoarthritis (OA) and increased loss over two years is associated with future knee replacement surgery. Previous studies have identified structural predictors of cartilage loss over 2–3 years including baseline cartilage volume, cartilage defects, bone marrow lesions (BMLs), meniscal pathology and bone area. These are now considered important risk factors for osteoarthritis progression. However no study has examined whether these features predict loss over longer timeframes. The aim of this study was to examine structural predictors of cartilage volume loss over ten years.

Methods: 220 participants [mean age 45 (26-61); 57% female] were studied at baseline, two, and ten years. Approximately half were adult offspring of subjects who had a knee replacement performed for knee OA and the remaining were randomly selected controls that were initially matched by age and sex. Cartilage volume (tibial, patella), cartilage defects (tibial, femoral), BMLs (tibial, femoral, patella), meniscal extrusion (medial, lateral) and bone area (tibial) were measured using MRI. Predictors of ten year cartilage loss were examined using linear regression and all models were adjusted for age, sex, BMI, and offspring-control status. Interactions between offspring-control status and knee structures on cartilage loss were explored.

Results: Cartilage change over ten years was -2.5% per annum (pa), -1.1% pa, and -2.1% pa for the medial and lateral tibial and patella sites respectively.

BMLs predicted tibial (β = -433 mm³ per 1 unit increase in BML size, P < 0.01) but not patella cartilage loss. Meniscal extrusion predicted lateral tibial cartilage loss (β = -872 mm³ in those with a baseline extrusion versus those without, P < 0.01) but not medial tibial or patella cartilage loss. Cartilage volume at baseline predicted tibial and patella ten year cartilage loss (β = ‑0.26 to ‑0.41 mm³ per unit increase in baseline volume, P< 0.01).

There was interaction between offspring-control status for cartilage defects. Baseline defect score predicted lateral tibial cartilage loss in offspring (B= -242 mm³ per unit increase in defect score, P < 0.01); but, surprisingly defects appeared to be protective against medial and lateral tibial cartilage loss in controls (β = 276 to 327 mm³ per unit increase in defect score, P<0.01).

Both change in cartilage volume (r = -0.58 to -0.62, all P < 0.01) and progression of cartilage defects (r = -0.18 to -0.58, all P< 0.01) over two years predicted ten year tibial cartilage loss.

Bone area did not predict ten year change in cartilage.

Conclusion: Structural features which have been identified as risk factors for cartilage loss over 2–3 years appear to predict long-term cartilage changes with the most consistent data for BMLs. This was seen in a largely pre‑radiographic cohort and indicates that bone marrow lesions, presence of meniscal extrusion; and to a lesser extent, cartilage volume and cartilage defects, may be used to identify individuals who will lose the most cartilage long-term. Additionally this study demonstrates that cartilage loss and cartilage defect progression over two years are reasonable predictors of long-term cartilage changes.

---

« Back to 2013 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/structural-predictors-of-ten-year-knee-cartilage-volume-loss/