Institutional review board approval and written, informed consent forms were obtained before testing for all patients. A dynamic finite element knee model () (Appendix Figure A-1, available in the online version of this article at http://ajs.sagepub.com/supplemental/
) was created from computed tomography and magnetic resonance images of a skeletally mature, young adult, female athlete to determine articular cartilage pressure distributions during normal landing and injury simulations (Abaqus 6.10 software, Simulia, Providence, Rhode Island). Details of model development methods are described in the Appendix (Appendix Tables A1–3 and Figures A1–3, available online).39
demonstrates a flow chart of the methods sequence. The model was subjected to validity evaluations and model simulations to analyze articular pressure distributions during normal landing and injury.
Knee finite element model that demonstrates the femur, tibia, fibula, patella anterior cruciate ligament (ACL), and medial collateral ligament meshes.
Flow chart describing the methods of the study.
The finite element model was subjected to evaluations based on cadaveric experimental data available in the literature and from previous cadaveric tests conducted by some of the current investigators (Appendix Figures A-2 and A-3, available online, provides detailed results). The first criterion for model confirmation was that the Pearson correlation was good (r
= .50–.75) to excellent (r
> .75) for the model results compared with cadaveric data available in the literature.38,46
The second criterion for model confirmation was that the differences between finite element model simulations and cadaveric data were less than 2 standard deviations (SDs) from the mean of the cadaveric data.31
We observed an excellent correlation (r
= .97, P
< .001) between the cadaveric results and predicted laxity in the finite element model. The finite element model predicted knee laxity values similar to those of Markolf et al,30
with anterior-posterior, external-internal, and varus-valgus laxities at 0°, 20°, and 45° of knee flexion for the finite element model within 2 SDs of the predicted mean for the cadaveric investigations. We used data from 10 human cadaveric lower extremities tested by Sohn et al48
to confirm the articular cartilage pressure distributions for the finite element model during weightbearing conditions. We observed a correlation (r
= .888, P
< .01) between the finite element predictions and the cadaveric data for articular cartilage pressures, with the model results within 2 SDs of the mean of the cadaveric data (Appendix Table A-3, available online). Markolf et al28
also examined combined loading states that generate high ACL forces (varus-valgus, internal-external tibial rotation, and anterior tibial force). These loading conditions were repeated in the model (knee flexion angles of 0°–40°). Again, an excellent correlation (r
= .92, P
< .001) was observed between the finite element predicted forces and the cadaveric results (Appendix Figures A-2 and A-3, available online). The finite element data were within 2 SDs of the reported cadaveric means for each testing condition.
The femur and tibia were divided into 6 separate sections each in order to describe articular cartilage pressure distributions (). These sections were defined as anterior-lateral (AL), anterior-medial (AM), middle-lateral (ML), middle-medial (MM), posterior-lateral (PL), and posterior-medial (PM) by drawing a sagittal plane divider and 2 frontal plane dividers (). This technique was previously described in detail and was adapted to allow for comparisons to current clinical techniques for localizing articular cartilage abnormalities (International Cartilage Repair Society format).4,39
Because the PL tibial and ML femoral sections are the most common bone bruise locations after ACL injury, articular cartilage contact pressure locations in the PL and ML sections were considered likely ACL injury mechanisms ( and ).33
Peak ACL strains were identified for each condition that resulted in peak articular cartilage pressures in the PL tibia and ML femur. Peak ACL strain was determined by the midsub-stance ACL element that had the maximum critical value.
Figure 4 Illustration of the sections used to describe pressure locations that demonstrate articular cartilage pressures during a loading scenario. The various colors indicate the stresses in the menisci and articular cartilage (further description in (more ...)
Figure 5 Injury conditions that resulted in peak articular cartilage contact pressure locations in the posterolateral tibia and lateral femur. A, combined abduction tibial rotation and anterior tibial translation injury conditions. B, the articular cartilage pressure (more ...)
Kinematic and kinetic data during a drop vertical jump maneuver from female athletes before their athletic seasons were used as baseline input for the finite element model. Mean kinematic and kinetic data at initial contact during landing from a drop vertical jump for athletes who went on to subsequent injury (INJ) after testing (n = 9 knees) and control (CTRL) (n = 390 knees) athletes (who did not have subsequent ACL injury) were used as baseline input for the model.16
Methods for biomechanical data collection are described in Ford et al9
and Hewett et al.16
A static trial was used as the athlete’s neutral (zero) alignment with subsequent measures relative to this position.10
Three-dimensional coordinates for the tibial and femoral markers were identified from the in vivo data (see Ford et al for marker placement) and used to define baseline input rotational and translational boundary conditions for the finite element model at corresponding node locations.9,16
The coordinate system described by Grood and Suntay14
was used to define the knee joint coordinate system of the finite element model with respect to the 3-dimensional knee joint motions, and a geometric center axis was used to apply flexion rotations.34
A vertical ground-reaction force was applied to the finite element model as a compressive force across the tibiofemoral joint. A 600-N quadriceps tension and 400-N hamstrings tension were applied to the model to simulate in vivo muscle resistance and compressive joint forces.46,51
Baseline “normal” landing conditions included a knee abduction angle of 5° (INJ) or −3.4° (CTRL) with a 20° knee flexion angle and a vertical ground-reaction force (600 N) for both groups.16
We used the finite element model to simulate ACL injury mechanisms for each group (INJ, CTRL) by single planar loading conditions for abduction rotation (abduction of the tibia relative to the femur), internal tibial rotation, external tibial rotation, and anterior tibial translation (). Combined injury conditions were examined for combined abduction/anterior tibial translation, abduction/internal tibial rotation, abduction/external tibial rotation, anterior tibial translation/internal tibial rotation, and anterior tibial translation/external tibial rotation. Because video analyses indicate that female athletes have 4° higher abduction angles compared with their initial contact angle during the ACL injury event, abduction injury simulations consisted of a 4° increase in abduction angle relative to baseline initial contact landing conditions.22
Anterior tibial translation injury simulations consisted of 6-mm increases in anterior tibial translation. Normal anterior tibial translation laxity is approximately 3 mm (depending on flexion angle and applied force), and ACL deficiency leads to an increase of 3 mm (or more) of anterior tibial translation laxity compared with the healthy contralateral limb.6,29,30
Internal and external tibial rotation injury simulations included 9° increases in the respective rotations because weightbearing axial rotations are approximately 5° to 7° in each direction, and Olsen et al36
reported approximately 9° of internal or external rotation of the tibia during ACL injury.47
The articular cartilage contact pressure locations were determined for each injury scenario and compared with the common bone bruise locations after acute ACL injury.