It is unclear why ACL rupture increases osteoarthritis risk, regardless of ACL reconstruction. Our aims were: 1) to establish the reliability and accuracy of a direct method of determining tibiofemoral contact in vivo with UO-MRI, 2) to assess differences in knees with ACL rupture treated nonoperatively versus operatively, and 3) to assess differences in knees with ACL rupture versus healthy knees.

We recruited a convenience sample of patients with prior ACL rupture. Inclusion criteria were: 1) adult participants between 18–50 years old; 2) unilateral, isolated ACL rupture within the last five years; 3) if reconstructed, done within one year from injury; 4) intact cartilage; and 5) completed a graduated rehabilitation program culminating in return to sport or recreational activities. Participants were excluded if they had other ligament ruptures, osteoarthritis, an incompletely rehabilitated injury, were prohibited from undergoing MRI, or had a history of ACL re-rupture. Using the UO-MRI, we investigated tibiofemoral contact area, centroid location, and six degrees of freedom alignment under standing, weightbearing conditions with knees extended. We compared patients with ACL rupture treated nonoperatively versus operatively, and ACL ruptured knees versus healthy control knees. We assessed reliability using the intra-class correlation coefficient, and accuracy by comparing UO-MRI contact area with a 7Tesla MRI reference standard. We used linear mixed-effects models to test the effects of ACL rupture and ACL reconstruction on contact area. We used a paired t test for centroid location and alignment differences in ACL ruptured knees versus control knees, and the independent t test for differences between ACL reconstruction and no reconstruction. Analyses were performed using R version 3.5.1. We calculated sample size based on a previous study that showed a contact area standard deviation of 13.6mm2, therefore we needed eight or more knees per group to detect a minimum contact area change of 20mm2with 80% power and an α of 0.05.

We recruited 18 participants with ACL rupture: eight treated conservatively and 10 treated with ACL reconstruction. There were no significant differences between the operative and nonoperative ACL groups in terms of age, gender, BMI, time since injury, or functional knee scores (IKDC and KOOS). The UO-MRI demonstrated excellent inter-rater, test-retest, and intra-rater reliability with ICCs for contact area and centroid location ranging from 0.83–1.00. Contact area measurement was accurate to within 5% measurement error. At a mean 2.7 years after injury, we found that ACL rupture was associated with a 10.4% larger medial and lateral compartment contact areas (P=0.001), with the medial centroid located 5.2% more posterior (P=0.001). The tibiae of ACL ruptured knees were 2.3mm more anterior (P=0.003), and 2.6° less externally rotated (P=0.010) relative to the femur, than contralateral control knees. We found no differences between ACL reconstructed and nonreconstructed knees.

ACL rupture was associated with significant mechanical changes 2.7 years out from injury, which ACL reconstruction did not restore. These findings may partially explain the equivalent risk of post-traumatic osteoarthritis in patients treated operatively and nonoperatively after ACL rupture.