Abstract
During total knee replacement (TKR), knee surgical navigation systems (KSNS) report in real time relative motion data between the tibia and the femur from the patient under anaesthesia, in order to identify best possible locations for the corresponding prosthesis components. These systems are meant to support the surgeon for achieving the best possible replication of natural knee motion, compatible with the prosthesis design and the joint status, in the hope that this kinematics under passive condition will be then the same during the daily living activities of the patient. Particularly, by means of KSNS, knee kinematics is tracked in the original arthritic joint at the beginning of the operation, intra-operatively after adjustments of bone cuts and trial components implantation, and after final components implantation and cementation. Rarely the extent to which the kinematics in the latter condition is then replicated during activity is analysed. As for the assessment of the active motion performance, the most accurate technique for the in-vivo measurements of replaced joint kinematics is three-dimensional video-fluoroscopy. This allows joint motion tracking under typical movements and loads of daily living. The general aim of this study is assessing the capability of the current KSNS to predict replaced joint motion after TKR. Particularly, the specific objective is to compare, for a number of patients implanted with two different TKR prosthesis component designs, knee kinematics obtained intra-operatively after final component implantation measured by means of KSNS with that assessed post-operatively at the follow-up by means of three-dimensional video-fluoroscopy.
Thirty-one patients affected by primary gonarthrosis were implanted with a fixed bearing posterior-stabilized TKR design, either the Journey® (JOU; Smith&Nephew, London, UK) or the NRG® (Stryker®-Orthopaedics, Mahwah, NJ-USA). All implantations were performed by means of a KSNS (Stryker®-Leibinger, Freiburg, Germany), utilised to track and store joint kinematics intra-operatively immediately after final component implantation (INTRA-OP). Six months after TKR, the patients were followed for clinical assessment and three-dimensional video fluoroscopy (POST-OP). Fifteen of these patients, 8 with the JOU and 7 with the NRG, gave informed consent and these were analyzed. At surgery (INTRA-OP), a spatial tracker of the navigation system was attached through two bi-cortical 3 mm thick Kirschner wires to the distal femur and another to the proximal tibia. The conventional navigation procedure recommended in the system manual was performed to calculate the preoperative deformity including the preoperative lower limb alignment, to perform the femoral and tibial bone cuts, and to measure the final lower limb alignment. All these assessment were calculated with respect to the initial anatomical survey, the latter being based on calibrations of anatomical landmarks by an instrumented pointer. Patients were then analysed (POST-OP) by three-dimensional video-fluoroscopy (digital remote-controlled diagnostic Alpha90SX16; CAT Medical System, Rome-Italy) at 10 frames per second during chair rising-sitting, stair climbing, and step up-down. A technique based on CAD-model shape matching was utilised for obtaining three-dimensional pose of the prosthesis components. Between the two techniques, the kinematics variables analysed for the comparison were the three components of the joint rotation (being the relative motion between the tibial and femoral components represented using a standard joint convention, the translation of the line through the medial and lateral contact points (being these points assumed to be where the minimum distance between the femoral condyles and the tibial baseplate is observed) on the tibial baseplate and the corresponding pivot point, and the location of the instantaneous helical axes with the corresponding mean helical axis and pivot point.
In all patients and in both conditions, physiological ranges of flexion (from −5° to 120°), and ab-adduction (±5°) were observed. Internal-external rotation patterns are different between the two prostheses, with a more central pivoting in NRG and medial pivoting in JOU, as expected by the design. Restoration of knee joint normal kinematics was demonstrated also by the coupling of the internal rotation with flexion, as well as by the roll-back and screw-home mechanisms, observed somehow both in INTRA- and POST-OP measurements. Location of the mean helical axis and pivot point, both from the contact lines and helical axes, were very consistent over time, i.e. after six months from intervention and in fully different conditions. Only one JOU and one NRG patient had the pivot point location POST-OP different from that INTRA-OP, despite cases of paradoxical translation.
In all TKR knees analysed, a good restoration of normal joint motion was observed, both during operation and at the follow-up. This supports the general efficacy of the surgery and of both prosthesis designs. Particularly, the results here reported show a good consistency of the measurements over time, no matter these were taken in very different joint conditions and by means of very different techniques. Intra-operative kinematics therefore does matter, and must be taken into careful consideration for the implantation of the prosthesis components. Joint kinematics should be tracked accurately during TKR surgery, and for this purpose KSNS seem to offer a very good support. These systems not only supports in real time the best possible alignment of the prosthesis components, but also make a reliable prediction of the motion performance of the replaced joint. Additional analyses will be necessary to support this with a statistical power, and to identify the most predicting parameters among the many kinematics variables here analysed preliminarily.
