Synovitis detected on magnetic resonance imaging and its relation to pain and cartilage loss in knee osteoarthritis

Objective

To examine the relationship between longitudinal fluctuations in synovitis with change in pain and cartilage in knee osteoarthritis.

Methods

Study subjects were patients 45 years of age and older with symptomatic knee osteoarthritis from the Boston Osteoarthritis of the Knee Study. Baseline and follow‐up assessments at 15 and 30 months included knee magnetic resonance imaging (MRI), BMI and pain assessment (VAS) over the last week. Synovitis was scored at 3 locations (infrapatellar fat pad, suprapatellar and intercondylar regions) using a semiquantitative scale (0–3) at all 3 time points on MRI. Scores at each site were added to give a summary synovitis score (0–9).

Results

We assessed 270 subjects whose mean (SD) age was 66.7 (9.2) years, BMI 31.5 (5.7) kg/m²; 42% were female. There was no correlation of baseline synovitis with baseline pain score (r = 0.09, p = 0.17). The change in summary synovitis score was correlated with the change in pain (r = 0.21, p = 0.0003). An increase of one unit in summary synovitis score resulted in a 3.15‐mm increase in VAS pain score (0–100 scale). Effusion change was not associated with pain change. Of the 3 locations for synovitis, changes in the infrapatellar fat pad were most strongly related to pain change. Despite cartilage loss occurring in over 50% of knees, synovitis was not associated with cartilage loss in either tibiofemoral or patellofemoral compartment.

Conclusions

Change in synovitis was correlated with change in knee pain, but not loss of cartilage. Treatment of pain in knee osteoarthritis (OA) needs to consider treatment of synovitis.The cause of pain in knee osteoarthritis remains elusive as the primary site of pathology in OA, cartilage, has no pain fibres.¹ Many other structures around the knee such as the periosteum, subchondral bone, the fat pad, capsule and, inconsistently, the synovium have been shown to contain nociceptive fibres.¹ In addition, inflammation itself appears to play a role in increasing input from peripheral nociceptors.² Biopsies of patients with both early and late knee OA have shown low‐grade chronic synovitis with production of pro‐inflammatory cytokines.^3,4Magnetic resonance imaging (MRI) allows evaluation of multiple structures within the knee, including synovium, cartilage, menisci, bone marrow lesions and effusion. In cross‐sectional studies of MRI in knee osteoarthritis (OA), bone‐marrow lesions, periarticular lesions, knee effusions and synovitis have been shown to be more often present in persons with knee pain than in persons with a comparable amount of radiographic knee osteoarthritis but without pain.^5,6,7,8 Fernandez‐Madrid et al demonstrated that synovial thickening seen on non‐contrast enhanced MRI in the infrapatellar region of knees with OA showed low‐grade synovial inflammation on biopsy. This feature was present in 73% of knees with early OA.⁹We have previously shown that this synovial thickening is present in persons with knee pain and OA much more often than in persons with OA but without pain.⁷ Among those with pain, the presence of MRI synovial thickening identified those with more severe pain. While this evidence suggests that synovial thickening may affect pain, these data are cross‐sectional, making it impossible to evaluate the temporal relation of pain with synovial thickening. More persuasive evidence would emanate from a longitudinal study in which fluctuations in synovial thickening could be tied to fluctuations in the severity of knee pain. Herein, we provide such evidence. Our aims were to study the association between baseline and longitudinal changes in MRI‐detected synovitis and changes in knee pain, and also to study the association between baseline and longitudinal changes in synovitis and cartilage loss in patients with symptomatic knee OA.