The alignment of prostheses components has a major impact on the longevity of total knee protheses as it significantly influences the biomechanics and thus also the load distribution in the knee joint. Knee joint loads depend on three factors: (1) geometrical conditions such as bone geometry and implant position/orientation, (2) passive structures such as ligaments and tendons as well as passive mechanical properties of muscles, and (3) active structures that are muscles. The complex correlation between implant position and clinical outcome of TKA and later in vivo joint loading after TKA has been investigated since 1977. These investigations predominantly focused on component alignment relative to the mechanical leg axis (Mikulicz-line) and more recently on rotational alignment perpendicular to the mechanical axis. In general four different approaches can be used to study the relationship between implant position and knee joint loads: In anatomical studies (1), the influence of the geometrical conditions and passive structures can be analyzed under the constraint that the properties of vital tissue are only approximated. This could be overcome with an intraoperative load measurement approach (2). Though, this set up does not consider the influence of active structures. Although post-operative in vivo load measurements (3) provide information about the actual loading condition including the influence of active structures, this method is not applicable to investigate the influence of different implant positions. Using mathematical approaches (4) including finite element analysis and multi-body-modeling, prostheses positions can be varied freely. However, there exists no systematical analysis of the influence of prosthesis alignment on knee loading conditions not only in axial alignment along and rotational alignment perpendicular to the mechanical axis but in all six degrees of freedom (DOF) with a validated mathematical model. Our goal was therefore to investigate the correlation between implant position and joint load in all six DOF using an adaptable biomechanical multi-body model. A model for the simulation of static single leg stance was implemented as an approximation of the phase with the highest load during walking cycle. This model is based on the AnyBody simulation software (AnyBody Technology A/S, Denmark). As an initial approach, with regard to the simulation of purely static loading the knee joint was implemented as hinge joint. The patella was realised as a deflection point, a so called “ViaNode,” for the quadriceps femoris muscle. All muscles were implemented based on Hill's muscle model. The knee model was indirectly validated by comparison of the simulation results for single and also double leg stance with in-vivo measurements from the Orthoload database (www.orthoload.de). For the investigation of the correlation between implant position and knee load, major boundary conditions were chosen as follows: Flexion angle was set to 20° corresponding to the position with the highest muscle activity during gait cycle. Muscle lengths and thereby also muscle loads were adapted to the geometrical changes after each simulation step representing the situation after post-operative rehabilitation. As input parameters, the tibial and femoral components' positions were independently translated in a range of ±20mm in 10 equally distant steps for all three spatial directions. For the rotational alignment in adduction/abduction as well as flexion/extension the tibial and femoral components' positions were varied in the range of ±15° and for internal/external rotation within the range of ±20°, also in 10 equally angled steps. Changes in knee joint forces and torques as well as in patellar forces were recorded and compared to results of previous studies. Comparing the simulation results of single and double leg stance with the in-vivo measurements from the Orthoload database, changes in knee joint forces showed similar trends and the slope of changes in torques transmitted by the joint was equal. Against the background of unknown geometrical conditions in the Orthoload measurements and the simplification (hinge joint) of the initial multi-body-model compared to real knee joints, the developed model provides a reasonable basis for further investigations already – and will be refined in future works. As influencing parameters are very complex, a non-ambiguous interpretation of force/torque changes in the knee joint as a function of changes in component positions was in many cases hardly possible. Changes in patella force on the other hand could be traced back to geometrical and force changes in the quadriceps femoris muscle. Positional changes mostly were in good agreement with our hypotheses based on literature data when knee load and patellar forces respectively were primarily influenced by active structures, e.g. with regard to the danger of patella luxation in case of increased internal rotation of the tibial component. Whereas simulations also showed results contradicting our expectations for positional changes mainly affecting passive structures, e.g. cranial/caudal translation of the femoral component. This shows the major drawback of the implemented model: Intra-articular passive structures such as cruciate and collateral ligaments were not represented. Additionally kinematic influences on knee and patella loading were not taken into account as the simulations were made under static conditions. Implementation of relative movements of femoral, tibial and patella components and simulation under dynamic conditions might overcome this limitation. Furthermore, the boundary condition of complete muscle adaptations might be critical, as joint loads might be significantly higher shortly after operation. This could lead to a much longer and possibly ineffective rehabilitation process.
Pertrochanteric femoral fractures are common and intramedullary nailing with a proximal femoral nail (PFNA®) is an accepted method for the surgical treatment. Accurate guide wire and subsequent hardware placement in the femoral neck is believed to be essential in order to avoid mechanical failure. Malpositioned implants may lead to rotational or angular malalignment or “cut out” in the femoral neck. Hip and knee arthritis might be a potential long-term consequence. The conventional technique might require multiple guidewire passes, and relies heavily on fluoroscopy. A computer-assisted surgical planning and navigation system based on 2D-fluoroscopy was developed in-house as an intraoperative guidance system for navigated guide wire placement in the femoral neck and head. To support the image acquisition process, the surgeon is supported by a so-called “zero-dose C-arm navigation” module. This tool enables a virtual radiation-free preview of the X-ray images of the femoral neck and head. The aim of this study was to compare PFNA® insertion using this system to conventional implantation technique. We hypothesised that guide wire and subsequent implant placement using our software decreases radiation exposure to the minimum of two images and reduces the number of drilling attempts. Furthermore, accuracy of implant placement in comparison to the conventional method might be improved and operation time shortened. We used 24 identical intact left femoral Sawbones® to simulate reduced pertrochanteric femoral fractures. First, we performed placement of the PFNA® into 12 Sawbones using the conventional fluoroscopic technique (group 1). Secondly, we performed placement of the PFNA® into 12 Sawbones guided by the computer-assisted surgical planning software (group 2). In each group, we first performed open and secondly minimal-invasive intramedullary nailing in six sawbones each. For minimal-invasive guide wire placement, a surgical drape imitated soft tissue coverage. Conventional and navigated technique used a C-arm fluoroscope (Siemens IsoC 3D®, Erlangen, Germany) in conventional 2D mode. Guidewire and subsequent blade placement in the femoral neck was evaluated. We documented: 1: the number of fluoroscopic images; 2: the total number of drilling attempts; 3: implant placement accuracy (3.1. Tip apex distance (TAD); 3.2. visible penetrations of the femoral neck and head; 3.3. blade-corticalis bone distance in the anteroposterior and lateral plane) and the 4: operation time. The number of fluoroscopic single shots taken to achieve an acceptable PFNA®-blade position was reduced significantly with computer-assistance by 71.5% (p<0.001) in the open and by 72,4% (p<0.001) in the minimally invasive technique. In each operation two X-rays for final documentation were taken. The average number of drilling attempts for the computer-guided system was significantly (p<0.05) less than that of the conventional technique in the minimally invasive procedure. The average number of drilling attempts showed no difference between the computer-assisted and conventional techniques in the open procedure. Accuracy of implant placement showed no difference between the computer-assisted and the conventional group. Computer assistance significantly increased the mean operation time for fixation of pertrochanteric femoral fractures with a PFNA® by 79.8% (p<0.001) in the open technique and by 54.4% (p<0.001) in the minimally invasive technique. Use of our computer-guided system for fixation of pertrochanteric femoral fractures by a PFNA® decreases the number of fluoroscopic single shots and of suboptimal guide wire passes while maintaining blade placement accuracy that is equivalent to the conventional technique. Computer-assisted surgery with our system increases the operation time and has just been tested in non-fractured sawbones. Although these results are promising, additional studies including fractured sawbones and cadaver models with extension of the navigation process to all steps of PFNA® introduction and with the goal of reducing the operation time are indispensable before integration of this navigation system into the clinical workflow.
