ACL forces and knee kinematics produced by axial tibial compression during a passive flexion–extension cycle

Authors

  • Keith L. Markolf,

    Corresponding author
    1. Biomechanics Research Section, Department of Orthopaedic Surgery, David Geffen School of Medicine at UCLA, Los Angeles, California
    Search for more papers by this author
  • Steven R. Jackson,

    1. Biomechanics Research Section, Department of Orthopaedic Surgery, David Geffen School of Medicine at UCLA, Los Angeles, California
    Search for more papers by this author
  • Brock Foster,

    1. Biomechanics Research Section, Department of Orthopaedic Surgery, David Geffen School of Medicine at UCLA, Los Angeles, California
    Search for more papers by this author
  • David R. McAllister

    1. Biomechanics Research Section, Department of Orthopaedic Surgery, David Geffen School of Medicine at UCLA, Los Angeles, California
    Search for more papers by this author

Abstract

Application of axial tibial force to the knee at a fixed flexion angle has been shown to generate ACL force. However, direct measurements of ACL force under an applied axial tibial force have not been reported during a passive flexion–extension cycle. We hypothesized that ACL forces and knee kinematics during knee extension would be significantly different than those during knee flexion, and that ACL removal would significantly increase all kinematic measurements. A 500 N axial tibial force was applied to intact knees during knee flexion–extension between 0° and 50°. Contact force on the sloping lateral tibial plateau produced a coupled internal + valgus rotation of the tibia, anterior tibial displacement, and elevated ACL forces. ACL forces during knee extension were significantly greater than those during knee flexion between 5° and 50°. During knee extension, ACL removal significantly increased anterior tibial displacement between 0° and 50°, valgus rotation between 5° and 50°, and internal tibial rotation between 5° and 15°. With the ACL removed, kinematic measurements during knee extension were significantly greater than those during knee flexion between 5° and 45°. The direction of knee flexion–extension movement is an important variable in determining ACL forces and knee kinematics produced by axial tibial force. © 2013 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 32:89–95, 2014.

It is well recognized that the ACL is a key structure in maintaining knee stability, and injury to the ACL often occurs during weight-bearing activities. Although some ACL injuries are thought to be associated with dynamic forces, torques and moments applied to the knee joint, there is experimental evidence that application of axial tibial force alone, at a fixed angle of flexion, is sufficient to produce ruptures of the ACL in the laboratory setting.[1, 2] These tests have been performed in the absence of stabilizing knee muscle forces, and the observed effects with applied axial tibial force can be attributed to condylar geometry of the articulating surfaces acting in conjunction with the stabilizing functions of the cruciate ligaments, collateral ligaments, and joint capsule. However most in vivo knee loadings do not occur at a static knee flexion angle, and involve flexion–extension movement of the knee. In vivo ACL strains have been measured without and with an axial compression force during knee flexion and extension against a resistance torque.[3] However, resultant ACL forces generated by application of axial tibial force to an unconstrained cadaveric knee have not been measured during a knee flexion–extension cycle.

ACL deficient patients frequently complain of instability that they describe as giving way. This sensation is associated with weight bearing activities that involve knee flexion–extension movements. The symptom of giving way is an important component of many functional outcome scores.[4] Although this symptom is difficult for patients to describe in precise kinematic terms, it may represent a combination of rotational and anterior instabilities that occur as weight is applied to the affected limb. Changes in knee kinematics produced by application of axial tibial force to an ACL deficient knee have not been measured during a flexion–extension cycle.

We hypothesized that (1) ACL force and knee kinematics generated by application of axial tibial force to an intact unconstrained knee would be significantly affected by the direction of knee flexion–extension movement, and (2) removal of the ACL would significantly increase the magnitude of recorded kinematic variables.

METHODS

Nine fresh-frozen unpaired cadaveric knee specimens were used for this study. The mean age was 32.2 years (SD 10.0 years, range 19–45 years). The tibia and femur were sectioned mid-shaft and dissected free of soft tissue to within 10 cm of the joint line. The bone ends were potted in cylindrical molds of polymethylmethacrylate acrylic cement (PMMA) for gripping in test fixtures. The angle between the tibia and femur was defined as 0° flexion (full extension) when a 2.5 N-m extension moment was applied to the knee. This definition of full extension has been used in our prior published knee studies.[5, 6] Internal–external rotation, varus–valgus rotation, and anterior–posterior tibial displacement were defined as zero with the knee at full extension (0° flexion).

The ACL's femoral insertion site was mechanically isolated using a cylindrical coring cutter, and a cap of bone containing the entire femoral footprint was attached to a load cell that recorded resultant tensile force developed in the ligament. This technique was similar to that described previously for installation of a tibial load cell to measure ACL force.[5]

The knee was mounted in a fixture that allowed manual flexion and extension of the femur between 0° and 50° about a fixed axis; the tibia remained horizontal during the test (Fig. 1). The femur was clamped in its fixture such that as the knee was flexed and extended under no-load conditions (no applied forces or torques) the tibia remained stationary with minimal varus–valgus, proximal–distal, or anterior–posterior movements. Alignment of this flexion–extension axis was determined by trial and error for each individual knee specimen. This process involved three principal adjustments of the femur relative to the horizontally resting tibia: proximal–distal positioning of the femoral fixture along its slotted mounting arm, tilt of the femoral fixture relative to the slotted arm, and rotation of the potted femur within the femoral fixture. With proper femoral alignment, the combined tibial movements described above were minimized as best possible. Due to the coupled nature of tibial motions, reducing one tibial movement often increased another movement. Varus–valgus tibial rotation, internal–external tibial rotation, and AP tibial displacement were defined as zero with the knee in full extension. All kinematic and ACL force measurements following application of axial tibial force were referenced relative to this no load condition.

Figure 1.

Test apparatus for application of axial tibial force during a knee flexion–extension cycle.

For knee tests with 500 N resultant axial tibial force, a large housing, clamped to the outer race of a thrust bearing, was attached to the end of the potted tibia (Fig. 1). The inner race of the bearing was fixed to the tibial potting acrylic by set screws. Two horizontal cables, connected to the medial and lateral sides of the bearing housing, passed over pulleys and down to a pneumatic air cylinder (not shown) that applied an equalized 250 N “follower” force to each cable (Fig. 1). That is to say, the piston of the air actuator moved during the test so as to maintain a constant 250 N tensile force on each side of the bearing housing as the tibia underwent proximal–distal displacements during the flexion–extension cycle. The line of action of these cables in the sagittal plane could be adjusted to apply the resultant axial tibial force at the approximate point of tibiofemoral contact (roughly along the axis of the potted tibia). In the frontal plane, the medial and lateral cables were equidistant from the center of the knee. The thrust-bearing allowed unconstrained tibial rotation as cable forces were applied to the bearing housing. Varus–valgus rotation of the tibia relative to the femur, and AP displacement of the tibia relative to the femur were also unconstrained with this force-equalizing pneumatic cable system. Internal–external rotation of the tibia was recorded by the shaft of a rotary potentiometer connected to the end of the tibia; the potentiometer housing was mounted on a roller assembly that followed varus–valgus rotation of the tibia (Fig. 1). Varus–valgus rotation was recorded by a potentiometer connected to the end of the tibia with a four-bar linkage (Fig. 1). AP displacement of the tibia was recorded by a linear variable differential transformer (LVDT) connected to a wire attached to the undersurface of the bearing housing (not shown). Further details of this loading apparatus can be found in our prior publication.[5]

For all tests, the femur was flexed at approximately 20°/s from 0° to 50°, and then the knee was extended back to 0°. Flexion–extension tests with 500 N axial tibial force were first performed on the intact knee, while recording knee kinematics and ACL force. This test was then performed with the tibial held in neutral tibial rotation during the flexion–extension cycle. Then the femoral bone cap was disconnected from the load cell (creating an ACL deficient condition) and the flexion–extension test was repeated with free tibial rotation, while recording knee kinematics alone. During tests with a disconnected bone cap, the cap was free to move in the tunnel without binding on the tunnel wall, thereby preventing development of tension in the ligament.

A series of special tests to measure the tibial torque produced by applied axial tibial force were also performed with the ACL bone cap disconnected. For these tests, the femoral fixture was clamped at pre-selected flexion positions (10°, 20°, 30°, and 45°) and 500 N axial tibial force was applied to the tibia. This always produced an internal tibial torque, accompanied by an internal tibial rotation. To measure this torque, weights were applied to the end of a rod attached to the tibia, at a fixed distance from the axis of tibial rotation, to rotate the tibia externally back to 0°. The external torque required to return the tibia to 0° rotation was calculated as the product of the weight times the distance to the center of rotation. This external torque was equal in magnitude to the internal torque produced by the applied axial tibial force.

A two-way repeated measures ANOVA model was used to compare internal tibial rotations, valgus tibial rotations, and anterior tibial displacements between ACL intact and ACL deficient conditions. A similar analysis was used to compare ACL forces and tibial motions generated during knee flexion to those generated during knee extension. A one-way repeated measures ANOVA model was used to compare internal tibial torques generated by applied axial tibial force between knee flexion angles. Pairwise comparisons were made using the Student–Neuman–Keuls procedure. The level of significance was p < 0.05.

RESULTS

With 500 N axial tibial force, release of the tibia from a neutral rotation position significantly increased ACL force at flexion angles greater than 20° (during knee flexion) and at flexion angles greater than 0° degrees during knee extension (Fig. 2). With the tibia held in neutral rotation, ACL forces during knee extension were not significantly different than those during knee flexion. With free tibial rotation, ACL forces during knee extension were significantly greater than corresponding forces during knee flexion between 5° and 50°.

Figure 2.

ACL force generated by 500 N axial tibial force applied to intact knees during a flexion–extension cycle. Arrows on the mean curves indicate the direction of knee flexion–extension movement. Mean ACL forces (with sample standard deviations) are shown for knee flexion and knee extension with the tibia free to rotate, and with the tibia held in neutral rotation. Indicated differences between means are significant at p < 0.05.

Anterior tibial displacements produced by applied axial tibial force were significantly greater for ACL deficient knees than corresponding values for intact knees at all flexion angles greater than 0°; this was true during knee flexion and knee extension (Fig. 3). Tibial displacements were significantly greater during knee extension than during knee flexion from 0° to 35° (intact knees), and at all flexion angles for ACL deficient knees (Fig. 3).

Figure 3.

Anterior displacement of the tibia generated by 500 N axial tibial force during a flexion–extension cycle. Arrows on the mean curves indicate the direction of knee flexion–extension movement. Mean displacements (with sample standard deviations) are shown for knee flexion and knee extension before and after ACL removal. Indicated differences between means are significant at p < 0.05.

Valgus tibial rotations produced by applied axial tibial force were significantly greater for ACL deficient knees than corresponding values for intact knees between 0° and 50° during knee flexion and between 5° and 50° during knee extension (Fig. 4). For intact knees, valgus rotations during knee extension were not significantly different than corresponding values during knee flexion (Fig. 4). For ACL deficient knees, valgus rotations during knee extension were significantly greater than corresponding values during knee flexion at flexion angles greater than 0° (Fig. 4).

Figure 4.

Valgus rotation of the tibia generated by 500 N axial tibial force during a flexion–extension cycle. Arrows on the mean curves indicate the direction of knee flexion–extension movement. Mean rotations (with sample standard deviations) are shown for knee flexion and knee extension before and after ACL removal. Indicated differences between means are significant at p < 0.05.

Internal tibial rotations produced by applied axial tibial force were significantly greater for ACL deficient knees than corresponding values for intact knees at flexion angles between 5° and 40° during knee flexion and between 0° and 15° during knee extension (Fig. 5). Internal rotations during knee extension were significantly greater than corresponding values during knee flexion between 5° and 25° for intact knees and between 0° and 45° for ACL deficient knees (Fig. 5).

Figure 5.

Internal rotation of the tibia generated by 500 N axial tibial force during a flexion–extension cycle. Arrows on the mean curves indicate the direction of knee flexion–extension movement. Mean rotations (with sample standard deviations) are shown for knee flexion and knee extension before and after ACL removal. Indicated differences between means are significant at p < 0.05.

Application of 500 N axial tibial force generated an internal tibial torque that produced the ACL forces shown in Figure 2, and the tibial motions shown in Figures 3-5. With the ACL removed, the external tibial torque required to return the tibia to neutral rotation (which was equal and opposite to the internal torque acting to produce the rotation) varied with flexion angle (Fig. 6), ranging from 1.7 N-m ± 1.0 N-m (at 10° flexion) to 4.3 N-m ± 2.9 N-m (at 45° flexion). All generated torques were significantly different from each other (p < 0.05).

Figure 6.

Mean values of the external tibial torque required to return the tibia to neutral rotation with 500 N applied axial tibial force. This is equal and opposite to the internal torque produced by applied axial tibial force. Sample standard deviations are shown with error bars. All mean torques were significantly different from each other (p < 0.05).

DISCUSSION

This study measured tibial motions and ACL forces generated by application of axial tibial force to the knee during a flexion–extension cycle. We found that ACL forces and knee kinematics during knee extension were significantly higher than those during knee flexion, which could have clinical implications for in vivo activities. The ACL was shown to be important for controlling the tibial motions produced by applied axial tibial force, and the increased motions after ACL removal could be related to the giving-way instability experienced by ACL deficient patients under weight-bearing conditions.

The ability of joint compression force to produce ACL ruptures has been previously demonstrated by Meyer and Haut.[1, 2] They reported that application of high compressive force to an intact knee caused the tibia to displace anteriorly and rotate internally, thereby producing ACL ruptures. Our directions of tibial motions produced by axial tibial force are consistent with theirs, though it should be noted that we investigated flexion–extension movements between 0° and 50°, while their compression loading tests were performed at fixed flexion angles between 30° and 120°.

Liu-Barba et al.[7] studied tibial rotations and displacements produced by applied compression force, but did not measure ACL forces. These authors reported anterior tibial translation with applied compressive load, which agrees with our findings and those of Meyers and Haute cited above. At 30° of knee flexion, they found that compressive load produced valgus rotation of the tibia, which is in agreement with our findings, and external tibial rotation, which is opposite to our findings and those of Meyers and Haute cited above. They found no significant changes in coupled rotations or anterior tibial displacements after removal of the ACL, which disagrees with our findings. The exact reasons for these differences are unknown, but could be related to testing methodology. Their tests were performed at a fixed knee flexion angle, while our measurements were recorded while the knee was in motion.

Fleming et al.[3] had patients perform knee flexion–extension movements between 10° and 90° without and with a tibial-femoral compression force of 40% body weight. These movements were performed with no external resistance torque, knee flexion against external resistance torque, and knee extension against external resistance torque. They found no significant reduction in peak strain values for either exercise type when the compressive load was applied to the foot with the 0 N-m resistance condition. This is in contrast to our findings where axial tibial compression force produced a significant ACL force over this flexion range. These authors did not do a direct comparison of ACL strain between flexion and extension movements. A comparison of our findings to theirs is not possible for resistance torque conditions because muscle activity was not simulated in our tests.

One shortcoming of our study is the magnitude of axial tibial force (500 N) utilized for our tests. This loading level was intentionally limited to prevent excessive ACL forces that could produce failure of the ACL bone cap's attachment to the load cell. Direct measurement of ACL forces was a unique and essential aspect of this study. Although the 500 N force used in our tests was less than would expected during in vivo loadings, this load level was sufficient to demonstrate differences in ACL forces and kinematics between knee flexion and extension movements, and the kinematic changes resulting from ACL removal. As such, the magnitudes of our recorded ACL forces and knee kinematic variables may be lower than those that would have been produced by a higher applied axial tibial force. It should be noted that load levels of 500 N or less have also been used in prior cadaveric studies with applied axial tibial force.[8-11]

Another shortcoming of our study is the relatively slow angular rate of knee flexion–extension used for these tests, which did not reproduce the higher rates expected during some in vivo activities. Also missing from our tests was a simulation of muscle forces acting on the knee as the joint was flexed and extended. Furthermore, lower limb dynamics that might occur during in vivo activities were not simulated in our study. Therefore, our results and conclusions are confined to the influences of condylar geometry and stabilizing ligaments on ACL forces and knee kinematics produced by axial tibial force. The lack of simulated muscle forces in our experiments could have altered the position of the tibia relative to the femur during our tests, and inclusion of quadriceps and hamstrings forces to simulate active knee flexion–extension against resistance could have affected our results. As such, our findings demonstrate the interactions between applied axial tibial force, condylar geometry, and ligamentous restraints during a flexion–extension cycle. Therefore our results may have limited application to the in vivo condition.

We believe the following mechanism may help to explain our findings related to the direction of knee flexion–extension movement. The lateral tibial plateau is sloped more posteriorly relative to the medial plateau. With the tibia held in neutral rotation, contact force on the lateral plateau (from applied axial tibial force) acted to sublux the lateral tibia anteriorly, thereby generating an internal tibial torque (Fig. 6). When the tibia was released and allowed to rotate freely, this internal torque produced coupled internal + valgus rotations of the tibia and anterior displacement of the tibia (Figs. 3-5). These tibial motions are known to generate ACL force.[5] As the knee was flexed under load, the lateral femoral condyle was essentially “sliding down” the posterior slope of the lateral plateau as the tibia rotated internally and displaced anteriorly. During knee extension, the lateral femoral condyle had to “climb up” the slope of the lateral plateau as the lateral tibia sought a reduced position. This caused the lateral plateau to remain in a more internally rotated and anteriorly subluxed position during knee extension than during knee flexion (Figs. 3-5), which in turn produced relatively higher ACL forces during knee extension (Fig. 2). It should be noted that when the tibia was held in neutral rotation during a flexion–extension test, this condyle climbing mechanism was not operative, and ACL forces generated during knee extension were not significantly different from those during knee flexion (Fig. 2).

Prior studies have demonstrated an association of an increased posterior slope of the lateral tibial plateau with increased anterior tibial force[12] and a higher risk for ACL injury.[13, 14] We found that high ACL forces due to applied axial tibial force were accompanied by large internal and valgus rotations of the tibia when axial tibial force was applied. Both of these tibial motions are known to generate ACL force, and it would stand to reason that both motions would increase with an increase in posterior tibial slope. Furthermore, our results are also consistent with the findings of Oh et al.,[15] who also demonstrated coupled valgus + internal rotations of the tibia in response to applied axial force.

There are factors related to the configuration and function of testing apparatus that merit discussion. Since there were no tibial rotations or displacements observed during flexion–extension tests with no applied axial tibial force, we have concluded that the observed effects were due to applied axial tibial force alone, and not due to the flexion–extension pathway. Mechanical friction within elements of the testing rig (such pulley friction and the axial thrust bearing) was another possibility, but these effects would be operative during both flexion and extension portions of the testing sequence. Finally, it should be emphasized that all ACL force and kinematic measurements were referenced to the “no load” test conditions, minimizing any effects related to the axis of flexion–extension motion.

We found that ACL removal significantly increased the tibial motions produced by applied axial tibial force (Figs. 3-5). It is possible that these increased motions could be related to the instability and giving way experienced by ACL deficient patients as they bear weight on the injured limb. Normally, active knee musculature would act to help control these tibial motions in vivo, but during unguarded moments the internal torque developed by joint load could produce the kinematic instabilities we observed with ACL deficient specimens. In patients, this could manifest as an episode of giving way.

Traditionally, the Lachman test for anterior knee stability has been used to detect ACL injury by measuring side to side differences in anterior tibial displacement. However, recent clinical studies have suggested that absence of a pivot shift correlates better with good functional performance than straight anterior knee stability alone.[4, 16, 17] It is interesting to note that ACL deficient patients frequently remark that giving way under weight bearing is similar to the instability experienced during a pivot shift exam.[18] The reasons for this similarity have remained unclear. The results of the present study suggest a possible explanation.

When performing a pivot shift test, the examiner typically applies a valgus moment to the knee by applying a lateral force to the tibia. As the knee is flexed, contact force on the posteriorly sloping lateral tibial plateau causes it to sublux anteriorly. This produces a coupled internal + valgus rotations of the tibia, accompanied by anterior tibial displacement. As the knee is flexed, passive tension developed in the iliotibial band spontaneously reduces the anterior subluxation of the tibia, producing an external tibial rotation.[19] This reduction event is graded clinically as a pivot shift. When performing the exam, some examiners also apply an internal tibial torque in addition to the valgus moment, in order to accentuate the tibial subluxation that occurs prior to the reduction event.

In prior simulated pivot shift studies from our laboratory,[20, 21] we demonstrated that the coupled internal + valgus rotations and anterior tibial displacement produced by an applied valgus moment increased significantly after ACL removal. In those studies, we attributed the observed motions to contact forces on the posteriorly sloping lateral tibial plateau. In the present study, application of axial tibial force also produced contact force on the lateral tibial plateau, and similar tibial motions were recorded (that also increased after ACL removal). We believe these kinematic similarities between the two separate test conditions may help explain why an ACL deficient patient may perceive the instability elicited during a pivot shift exam to be similar to that occurring during a giving way episode under weight-bearing.

Our findings could have implications related to ACL forces generated during common activities. In vivo, the quadriceps is an important muscle group for producing acceleration movements of the lower limb during knee extension or controlling limb deceleration movements during knee flexion. During a stop and jump movement, compression force produced at the joint surface from both quadriceps activation force and ground reaction force act to load the ACL, as the knee first undergoes knee flexion followed by knee extension. Our test results with this same flexion–extension movement sequence would suggest that ACL forces generated during the extension part of this task would be higher than during the flexion portion.

This mechanism could also be operative during other in vivo activities as well. It is currently thought that some ACL injuries occur during the weight-acceptance (or knee flexion) phase, such as landing from a jump, that involve lower limb deceleration. However, it is also possible that other lower limb motions, such as cutting movements or changes in direction, could produce ACL injury during the acceleration phase when the knee has already flexed and has begun to extend. As non-contact ACL injury mechanism become more fully understood, our findings may have increased clinical relevance to more complex injury scenarios.

In conclusion, we found that application of axial tibial force to an intact knee generated an internal tibial torque that produced internal + valgus rotations of the tibia and anterior tibial displacement. These tibial motions were restrained by forces developed in the ACL. ACL forces generated during knee extension were significantly greater than those generated during knee flexion. ACL removal significantly increased the magnitudes of tibial motions produced by applied axial tibial force, producing a pattern of instability that could be related to the giving way sensation reported by ACL deficient patients with weight bearing activities.

ACKNOWLEDGEMENTS

Tissues for this study were provided by the Musculoskeletal Transplant Foundation.

Ancillary