Restoration of natural range and pattern of motion is the primary goal of joint replacement. In total ankle replacement, proper implant positioning is a major requirement to achieve good clinical results and to prevent instability, aseptic loosening, meniscal bearing premature wear and dislocation at the replaced ankle. The current operative techniques support limitedly the surgeon in achieving a best possible prosthetic component alignment and in assessing proper restoration of ligament natural tensioning, which could be well aided by computer-assisted surgical systems. Therefore the outcome of this replacement is, at present, mainly associated to surgeon's experience and visual inspection. In some of the current ankle prosthetic designs, tibial component positioning along the anterior/posterior (A/P) and medio/lateral axes is critical, particularly in those designs not with a flat articulation between the tibial and the meniscal or talar components. The general aim of this study was assessing in-vitro the effects of the A/P malpositioning of the tibial component on three-dimensional kinematics of the replaced joint and on tensioning of the calcaneofibular (CaFiL) and tibiocalcaneal (TiCaL) ligaments, during passive flexion. Particularly, the specific objective is to compare the intact ankle kinematics with that measured after prosthesis component implantation over a series of different positions of the tibial component.

Four fresh-frozen specimens from amputation were analysed before and after implantation of an original convex-tibia fully-congruent three-component design of ankle replacement (Box Ankle, Finsbury Orthopaedics, UK). Each specimen included the intact tibia, fibula and ankle joint complex, completed with entire joint capsule, ligaments, muscular structures and skin. The subtalar joint was fixed with a pin protruding from the calcaneus for isolating tibiotalar joint motion. A rig was used to move the ankle joint complex along its full range of flexion while applying minimum load, i.e. passive motion. In these conditions, motion at the ankle was constrained only by the articular surfaces and the ligaments. A stereofotogrammetric system for surgical navigation (Stryker-Leibinger, Freiburg, Germany) was used to track the movement of the talus/calcaneus and tibial segments, by using trackers instrumented with five active markers. Anatomical based kinematics was obtained after digitization by an instrumented pointer of a number of anatomical landmarks and by a standard joint convention. The central point of the attachment areas of CaFiL e TiCaL was also digitised. Passive motion and ankle joint neutral position were acquired, and the standard operative technique was performed to prepare the bones for prosthesis component implantation. The final component for the talus was implanted, the tibial component was initially positioned well in front of the nominal right (NR) position, the meniscal bearing was instrumented with an additional special tracker, and passive motion was collected again in passive flexion. Data collection was repeated for progressively more posterior locations for the tibial component, for a total of six different locations along the tibial A/P axis: three anterior (PA), the NR, and two more posterior (PP), approximately 3 to 5 mm far apart each. The following three-dimensional kinematics variables were analyzed: the three anatomical components of the ankle joint (talus-to-tibial) rotation (dorsi/plantar flexion, prono/supination and internal/external rotation respectively in the sagittal, frontal and transverse planes), the meniscal bearing pose with respect to the talar and tibial components, the ‘ligament effective length fraction’ as the ratio between the instantaneous distance between the ligament attachment points and the corresponding maximum distance, and the instantaneous and mean helical axes in the tibial anatomical reference frame.

In all specimens and in all conditions, physiological ranges of flexion, prono/supination and internal/external rotation were observed at the ankle joint. A good restoration of motion was observed at the replaced joint, demonstrated also by the coupling between axial rotation and flexion and the physiological location of the mean helical axis, in all specimens and in most of the component positions. Larger plantar- and smaller dorsi-flexion were observed when the tibial component was positioned more anteriorly than NR, and the opposite occurred for more posterior positions. In regards to the meniscal bearing, rotations were small and followed approximately the same patterns of the ankle rotations, accounted for the full conformity of the articulating surfaces. Translations in A/P were larger than in other directions, the bearing moving backward in plantarflexion and forward in dorsiflexion with respect to both components. It was observed that the closer to NR the position of the tibial component is, the larger this A/P motion is, accounted mainly to the associated larger range of flexion. The change of CaFiL and TiCaL effective length fraction over the flexion arc was found smaller than 0.1 in three specimens, smaller than 0.2 in the fourth, larger both in more anterior and more posterior locations of the tibial component. The simulated malpositioning did not affect much position and orientation of the mean helical axis in both the transversal and frontal planes.

The experimental protocol and measurements were appropriate to achieve the proposed goals. All kinematics variables support the conclusion that the ankle replaced with this original prosthesis behaves as predicted by the relevant computer models, i.e. physiological joint motion and ligament tension is experienced resulting in a considerable A/P motion of the meniscal bearing. These observations are particularly true in the NR postion for the prosthesis, but are somehow correct also in most of the tibial malpositions analysed, in particular those on the back.