A total of 8 fresh-frozen human cadaver legs were utilized for the biomechanical testing, the specimens were thawed to room temperature (24ºC) . The mean age of the cadavers was 71.4 (7.4) years old, 4 of the 8 were men and 4 of the 8 were right foot. X-rays were obtained for each specimen and none of them had malalignment of the tibia, hindfoot, nor preexisting subtalar joint osteoarthritis. All specimens had normal range of motion of both ankle and subtalar joint. The anterior soft tissue (including skin, subcutaneous tissue, anterior joint capsule, tendons and neurovascular bundles) of the ankle joint were dissected for the access to the ankle joint. Both the medial and lateral ankle ligaments were well preserved.
The tibia and fibula was cut at 20cm above the ankle joint. For each specimen, the proximal tibia and fibula was potted securely into a custom made shell, then mounted on a custom made fixture. The tibia and fibula was embedded and securely fixed into the shell using dental gymsum. Load was applied to the tibia and fibula via the custom made shell. Each specimen must be potted in neutral position, no plantarflexion or dorsiflexion of the ankle joint in sagittal plane, no varus or valgus malalignment of the hindfoot in coronal plane, no internal or external rotation of the foot in horizontal plane.
The custom designed fixture was utilized for the testing. Spirit levels were utilized to make sure both the working table and the top plate was horizontal throughout the entire testing process. The varus malalignment of tibia (0º,2º,4º,6º,8º,10º,12º,14º,16º,18º,20º of tibial varus) was simulated by a custom made apparatus. Each hole on the apparatus represented a specific varus angle. A bolt was used to fix the specimen at a desired varus angle. (Figure 1)
The four threaded polyethylene pillars were utilized to connect the top plate, the compressive forces were exerted via the four pillars. Sensor cells were placed in each pillar, a monitor was connected to each sensor cell, displaying the real-time force. Springs were placed right above each force sensor, then followed by nuts, compressive forces could be generated by twisting the nut on the spring.
The sensor pads (Model 6900, TekScan, Inc., South Boston, MA) , with each pad measuring 14*14mm, each pad had 121 senels (11*11 sensels) , the column and row spacing was 1.3mm, resulting in a spatial resolution of 0.62 mm2 per sensel. Two pads were put side by side within the ankle joint for the measurement of the ankle joint pressure. The sensor pads were inserted into the ankle joint from anterior and secured by thumbtacks to the distal tibial metaphysis and the foot in order to avoid sensor motion during testing.[7, 8] The sensor pads were connected to the handle which could be further connected to a personal computer, data including pressure, force was collected using I-Scan software. (Figure 2)
The baseline ankle joint pressure distribution was initially collected for each specimen. The specimen was fixed at 0º of tibial varus by inserting a bolt, the foot was then placed onto the floor freely. A compressive force was generated through the 4 pillars by twisting the nuts. Making sure the top plate was maintained horizontally throughout the testing. A 600N compressive force was applied to simulate the normal load within the ankle joint during ambulation. Both the medial and lateral ankle joint pressure data was collected. Then free the specimen by removing the bolt and then the top plate of the fixture was lifted upward, followed by 2º of tibial varus, then 4º,6º,8º,10º,12º,14º,16º,18º,20º. Both the medial and lateral aspect of the ankle joint pressure data were collected for each alignment.
Statistical analysis
SPSS V.23 software (IBM Inc. , New York) was used for the data analysis. Matlab was used for the calculation of shift of center of force (COF) , the peak pressure (Pmax) and the mean pressure (Pmean) . The lateral shift of COF relative to the ankle joint was defined as positive, the medial shift was defined as negative. S-W test was used for the test of normality. Paired student t test was utilized to determine the significant differences for the COF shift, Pmax and Pmean at different working conditions. The level of significance was set to a p value <0.05.