Abstract
Introduction:
Despite over 95% long-term survivorship of TKA, 14–39% of patients express dissatisfaction due to anterior knee pain, mid-flexion instability, reduction in range of flexion, and incomplete return of function. Changing demographics with higher expectations are leading to renewed interest in patient-specific designs with the goal of restoring of normal kinematics.
Improved imaging and image-processing technology coupled with rapid prototyping allow manufacturing of patient-specific cutting guides with individualized femoral and tibial components with articulating surfaces that maximize bony coverage and more closely approximate the natural anatomy. We hypothesized that restoring the articular surface and maintaining medial and lateral condylar offset of the implanted knee to that of the joint before implantation would restore normal knee kinematics. To test this hypothesis we recorded kinematics of patient-specific prostheses implanted using patient-specific cutting guides.
Methods:
Preoperative CT scans were obtained from nine matched pairs of human cadaveric knees. One of each pair was randomly assigned to one of two groups: one group implanted with a standard off-the-shelf posterior cruciate-retaining design using standard cutting guides based on intramedullary alignment; the contralateral knee implanted with patient-specific implants using patient-specific cutting guides, both manufactured from the preoperative CT scans. Each knee was tested preoperatively as an intact, normal knee, by mounting the knee on a dynamic, quadriceps-driven, closed-kinetic-chain Oxford knee rig (OKR), simulating a deep knee bend from 0° to 120° flexion. Following implantation with either the standard or patient-specific implant, knees were mounted on the OKR and retested. Femoral rollback, tibiofemoral rotation, tibial adduction, patellofemoral tilt and shift were recorded using an active infrared tracking system.
Results:
To reduce the effect of variability, change in each kinematic measure was quantified as the absolute difference between the normal kinematic measure and the same measure after implantation (10° flexion increments). The cumulative difference from normal kinematics was calculated by summing the area beneath the curve (Fig 2). Cumulative differences in kinematics from normal were statistically lower for the patient-specific group compared to the standard group for all measures except patellar shift (Fig 2, paired t-test).
Discussion:
Knee kinematics with the patient-specific design more closely approximated normal femoral rollback and tibial adduction than knees with the standard design.
Femoral rollback is significantly closer qualitatively and quantitatively to normal in specimens implanted with patient-specific designs (Figs 1). The tibia rotated internally with flexion; however, the patient-specific group more closely approximated normal rotation. The patient-specific group more closely approximated normal tibial adduction suggesting ligament balance was better restored.
Due to substantial differences in articular morphology among genders, races and patients, it is impossible to provide multiple sized implants to address the full range of inter-patient variability. Patient-specific designs that remove this variation, restore normal articular geometry, and maintain alignment are more likely to result in normal kinematics.
Our results support the hypothesis that knees with patient-specific implants generate kinematics more closely resembling normal knee kinematics than standard knee designs. Clinical outcome studies are necessary to determine if our results translate into better outcomes.
