Background/Purpose: Osteophyte (OP) formation is one of the clinical features of osteoarthritis (OA), so early detection of OP formation can be of diagnostic value. OP can be visualised using magnetic resonance imaging (MRI) that are not easily visualised by conventional radiography, and at greater sensitivity than radiographs for detection of early and central OP. Although some studies have examined relationships between MRI-detected OP and clinical features of knee OA, longitudinal studies are rare. Furthermore, clinical significance of x-ray- undetectable OP is unknown. Therefore, our study aims to describe cross-sectional and longitudinal associations between MRI-detected OP and knee pain and structural changes in older adults; and to evaluate the longitudinal associations between x-ray- undetectable OP and knee pain and structural changes.

Methods: 837 participants (mean age 62 years, 50% female) were randomly selected from local community at baseline. T1- or T2-weighted fat suppressed magnetic resonance imaging (MRI) was used to assess knee OP, cartilage volume (CV), cartilage defects (CD) and bone marrow lesions (BMLs) at baseline and after 2.6 years. Knee pain was assessed by self-administered Western Ontario and McMaster Osteoarthritis (WOMAC) Index questionnaire at baseline and after 5 years. Radiographic osteoarthritis (ROA) was assessed at baseline with a standing anteroposterior semiflexed radiograph scored using the Osteoarthritis Research Society International atlas. OPs detected by MRI but not by X-ray at baseline were defined as x-ray- undetectable OP. Analyses were performed using linear regression models and log-binominal regression models.

Results: 86.6% of participants had MRI-detected OPs at baseline, while only 10% of participants had radiographic OPs. Cross-sectionally, MRI-detected OPs at medial tibiofemoral, lateral tibiofemoral and patellar compartments were significantly and site-specifically associated with a higher prevalence of CD and BMLs, and reduced CV after adjustment for common covariates (all p<0.01). MRI-detected OPs in whole compartment were significantly associated with higher prevalence of total knee pain (p<0.01). Longitudinally, baseline MRI-detected OP site-specifically predicted increases in CD and BMLs and loss of CV (all p<0.01) in multivariable analyses. Compared to participants without any OP, participants with x-ray- undetectable OP and definite OP (both radiographic and MRI-detected OP) had greater CV loss and increased CD and BMLs longitudinally. Presence of x-ray- undetectable OP predicted decreases in total knee pain over 5 years, while participants with definite OP predicted increases in knee pain, after adjustment for BMLs, CD and NSAIDs usage.

Conclusion: MRI-detected OPs were associated with knee structural abnormalities and knee pain cross-sectionally and longitudinally. Although x-ray- undetectable OP is associated with knee abnormal structural changes, it predicts decreases in knee pain over time suggesting an adaptive response. 

« Back to 2016 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/association-between-mri-detected-osteophytes-and-changes-in-knee-pain-and-structures-in-older-adults-a-population-based-cohort-study/