This study has been approved by the Institutional Review Board (IRB) of the Third Hospital of Hebei Medical University.
Fourteen fresh-frozen cadaveric lower limbs with intact soft tissue were autopsied (all cadavers were provided by the Department of Human Anatomy, Hebei Medical University), average height of the donors was 171 cm (range, 163 to 181 cm), average age was 55 years (range, 42 to 65 years). Each cadaveric lower limb had complete femur, tibia and knee joint, there were no gross deformities of the knee, i.e. hyperflexion, hyperextension, varus and valgus. The joint can passively flex and extend without restriction. Furthermore, inner knee structures were examined by X ray, pathological(osteoporosis, rheumatism, tuberculosis, or tumors) or anatomical deformities(unsymmetrical joint surface, bone hyperplasia, or other imaging abnormalities) were excluded(Fig. 1).
Then removed all muscular tissues, draw the anatomical axes along the length of the intramedullary canals of the femur and tibial based on the method proposed by Moreland et al. For this biomechanical experiment, we reserved approximately 25 cm of the distal femur and proximal tibia and fibula, and wrapped the dissected cadaveric knee with polyethylene films to prevent dehydration and cryopreserved at −20℃.
Establishment of rotatory fixation model
The cadaveric knee were thawed at room temperature for 12 hours before experiment. Cut a horizontal incision about 3-4 cm long at the level of the joint space, both sides of the patellar ligament. Separate the subcutaneous fat, cut the sac, and expose the joint space, reserve anterior and posterior cruciate ligaments, as meniscus is a weight-bearing structure that can buffer pressure and affect the expansion, so it has to be preserved. Then saw the femoral shaft transversely at distal 1/3, guarante each cut basically at the same level to eliminate heterogeneity. Compared with the previously drawn anatomical axes, neutral position (0°, anatomically reduced), the internal rotation 5°, 10°, 15° and the external rotation 5°, 10°, 15° were measured with an bone protractor, fixed the stumps with plates and screws. Repeat the above experimental steps to complete other rotatory fixation models.
Inserted pressure-sensitive film
An ultra-low-pressure sensitive film (LLW type, Fujifilm Investment Co. Ltd. Japan) (0.5–2.5 MPa) is used to measure the contact pressure on tibial plateau, in order to ensure the quality of the pressure-sensitive film, we set the room humidity to 35%RH and the temperature to 20°C. Trim the pressure-sensitive film into somehow match shape according to our preliminary experiment, then seal it with a polyethylene film bag, a total thickness must be less than 250μm, thereafter carefully insert it under the meniscus and fully accessed into the joint cavity, suture the capsule tightly, leakage, bending, breakage of the sealed bag mean failure (Fig. 2). In order to distinguish the anterior and posterior side of knee, the corresponding anterior side of knee on the pressure sensitive sheet is clamped with a hemostatic forceps in advance to make an impression.
Specimen Assembled to Biomechanical Testing Machine
Clamp the femur and tibial end perpendicularly and reinforce with the denture base resin and solution (type II self-setting dental powder and tray water) (Fig. 3-4). Then transfer and assemble the combination to the biomechanical testing machine (Electroforce 3520-AT, Bose company, USA). As the measurement work will be done dozens of times, so we are intended to ensure conformity between each step.
Start the biomechanical machine, load the test bench, pressurize to 200N at a speed of 10N/s to eliminate creep. After stabilizing, apply a vertical load to the specimen to 400N at a speed of 10N/s and uphold for 2 minutes, unload and get the pressure-sensitive film out.
FPD-305E density meter and FPD-306E pressure converter were used to read relative pressure value. We divided the contact pressure area (the red area) of each pressure-sensitive film into 4 quadrants (anterolateral, anterior medial, posterior medial and posterior lateral), each quadrant randomly and equally read 5 values, total 20 values in one film, take the average as final values.
The experimental data were organized and computed by SPSS 21.0 software (SPSS, Chicago, IL, USA). The normality is verified using the Shapiro–Wilk test and expressed as ±s, we used T-test of two independent samples to access difference between medial and lateral groups, the Student–Newman–Keuls test for pairwise comparisons between the multiple sample measurements. Using the Levene test for variance consistency, and analysis of variance (ANOVA) for random block groups. Data doesn’t fit normality expressed as the median (quartile) and using Mann-Whitney U test to access difference between medial and lateral groups. Kruskal-Wallis H test for random block groups, significance was P<0.05.