TABLE 2.

Linear regression models aiming to explain factors associated with chronic postoperative pain 6 months after total knee arthroplasty.

Model	Postoperative parameters	Standardized coefficient	p	Adjusted R 2
1				0.645
	Cuff pressure pain threshold	0.105	0.474
	Temporal summation of pain	0.245	0.025
	Conditioned pain modulation	0.030	0.795
	PCS scores	0.380	<0.001
	Hoffa synovitis grade	0.276	0.032
	Effusion grade	0.365	0.008
	Bone marrow lesion a	−0.101	0.285
2				0.653
	Temporal summation of pain	0.251	0.013
	PCS scores	0.392	<0.001
	Hoffa synovitis grade	0.263	0.036
	Effusion grade	0.397	0.003

Note: Clinical pain intensity was the dependent variable for all models and postoperative magnetic resonance imaging findings (synovitis, effusion and bone marrow lesions), pain sensory profiles (cuff pressure pain thresholds, temporal summation of pain and conditioned pain modulation) and pain catastrophizing were the independent parameters. Model 1 explained 64.5% of the postoperative pain intensity and model 2 explained 65.3% of the postoperative pain intensity following four iterations of backwards elimination. Bold indicate significant findings.

Abbreviation: PCS, Pain Catastrophizing Scale.

^a Postoperative bone marrow lesions are graded from 0 to 18 (and not from 0 to 45 as per recommendations; Hunter et al., 2011) as only the tibia and patella were included as the entire femur was replaced by the femoral implant.