Background/Purpose: Some patients do not experience clinical improvement after total knee replacement (TKR). Presence of structural abnormalities on MRI before surgery may inform pain prognosis following TKR. Our objective was to evaluate MRI-detected knee inflammation before TKR as a risk factor for failure to experience clinically important pain reduction two years later.

Methods: Osteoarthritis Initiative participants with MRI assessment at the clinic visit before TKR and 2 year follow-up post-TKR, were selected. Effusion-synovitis (ES) and Hoffa synovitis (HS) on 3T MRI were scored using the MRI Osteoarthritis Knee Score (MOAKS). Pain was assessed at the clinic visit before TKR as well as 2 years later, with WOMAC pain (0-20) and knee pain severity in the past 7 days (0-10). Mean WOMAC pain and knee pain severity before TKR and 2 years later were estimated using a mixed model for repeated measures. Clinically important reduction in WOMAC pain (-1.5) and pain severity on NRS (-1.7) were defined based on previous literature. Participants with no/small ES/HS and those with medium/large ES/HS were compared based on the proportion reporting no clinically important pain reduction; odds ratios were estimated with logistic regression, adjusted for age, sex, race, and BMI.

Results: Participants (n=156) were predominately white (87%), women (60%), with mean age 66.2 years (SD 8.6) and mean BMI of 29.9 kg/m² (SD 5.1). Before TKR, 62% of the knees had medium/large ES, with higher mean WOMAC pain score compared to those with no/small ES (7.9 [95%CI: 7.1, 8.6] vs. 6.3 [95%CI: 5.3, 7.3]). Two years following TKR, the mean WOMAC pain was modestly higher for knees that had no/small ES before TKR (1.8 [95%CI: 1.2, 2.4] vs 2.7 [95%CI: 1.9, 3.4]). Knees with medium/large ES prior to TKR had greater pain reduction compared to those with no/small ES (6.7 [95%CI: 5.5, 7.9] vs 4.0 [95%CI: 2.8, 5.3]), a significant difference in mean pain reduction of 2.7 [95%CI: 1.3, 4.1] p=0.002 (Figure 1A). Participants with no/small pre-operative ES had an increased odds of reporting no clinically important pain reduction compared to participants with medium/large ES (29% vs 13%; OR=2.98, 95%CI: 1.26, 7.04). Similarly, knees with no/small ES had less mean pain severity reduction (p=0.006; Figure 1B), and with increased odds of no clinically important pain reduction (27% vs 10%; OR=4.35, 95%CI: 1.69, 11.23). We observed a similar trend for HS, detected in 37% of knees, though pain reduction was not significantly different between the groups (p=0.07).

Conclusion: Participants with no/small ES on MRI before TKR reported less pain reduction 2 years later compared to those with medium/large ES, and had significantly greater odds of no clinically important pain reduction following TKR. Pre-operative ES, potentially in combination with other features of MRI-detected structural abnormalities, may serve as predictive biomarkers of pain reduction following TKR.