Introduction: Following the work of Freeman et al. on the medial pivot and roll back in the normal knee and several other studies on tibial slope we got interested in the subject. In most studies tibial slope is always measured on standard radiographs (lateral view) and therefore the medial and lateral side are superimposed.

Materials & methods: We studied the lateral view of the medial and lateral tibial plateau on a magnetic resonance scan (subchondral line). The study group (N=80) consisted of young patients (18–40 y) all consulting for patellofemoral problems with a non arthritic and stable knee. A neutral tibial axis was determined on the lateral view. Perpendicular to this axis the posterior slope of the medial and lateral compartment was measured. Statistical analysis was done.

Results: This analysis showed a mean posterior slope of – 5 ° ( range 0 ° – 12 °) on the lateral side, but an upslope on the medial side of + 7 ° ( range 5 °– 10 °). A significant statistical difference was noted between both.

Discussion: These results suggest an upslope on the medial side of the knee which could be important for deep knee flexion since this increases the posterior condylar clearance. Roll back on the medial side after 120° of flexion could be roll up of the condyle (2 mm). This could also explain the femoral external rotation (or tibial internal rotation) in natural knee flexion since the medial condyle rides up the medial meniscus and plateau allowing the lateral femoral condyle to roll down the lateral plateau during internal rotation of the tibia around the medial pivot point. This observation could explain paradoxical motion in total knee arthroplasty, since until now we made an equally sloped cut in both compartments.