24

THE ROLE OF FIBERS WITHIN THE TIBIAL ATTACHMENT OF THE ANTERIOR CRUCIATE LIGAMENT IN RESTRAINING TIBIAL DISPLACEMENT

Breck R. Lord FRCS1, PhD, Hadi El-Daou PhD1, Urszula Zdanovicz MD2, Robert Smigielski MD2, Andrew A. Amis FREng, DSc1,3

1Biomechanics Group, Mechanical Engineering Department, Imperial College London, London, UK

2Carolina Medical Centre, Warsaw, Poland

3Musculoskeletal Surgery Group, Imperial College London School of Medicine, London, UK

Corresponding author:

Prof Andrew A. Amis,

Mechanical Engineering Department, Imperial College London, London SW7 2AZ, UK

+44 (0)20 7594 7062

a.amis@imperial.ac.uk

Running title: ACL restraint of tibiofemoral laxity

IRB approval: Imperial College Tissue Bank Project R13058a under Human Tissues Authority licence 12275.

Acknowledgements

Mr Lord was supported by the Orthopaedic Research Fund of the Basingstoke and North Hampshire Hospital. The specimens were provided by the Carolina Clinic, Warsaw. The robot and Dr El-Daou were funded by the Centre of Excellence in Biomedical Engineering at Imperial College London, funded by the Wellcome Trust and the EPSRC.

THE ROLE OF FIBERS WITHIN THE TIBIAL ATTACHMENT OF THE ANTERIOR CRUCIATE LIGAMENT IN RESTRAINING TIBIAL DISPLACEMENT

Abstract

Purpose: To evaluate the load-bearing functions of the fibers of the ACL tibial attachment in restraining tibial anterior translation, internal rotation, and combined anterior and internal rotation laxities in a simulated pivot-shift test.

Methods: Twelve knees were tested using a robot. Laxities tested were: anterior tibial translation (ATT), internal rotation (IR), and coupled translations and rotations during a simulated pivot-shift (SPS). The kinematics of the intact knee was replayed after sequentially transecting 9 segments of the ACL attachment and fibers entering the lateral gutter (LG), measuring their contributions to restraining laxity. The ‘center of effort’ (COE) of the ACL force transmitted to the tibia was calculated. A blinded anatomical analysis identified the densest fiber area in the attachment of the ACL and thus its centroid (center of area). This centroid was compared to the biomechanical COE.

Results: The anterior-medial (AM) tibial fibers were the primary restraint of ATT (84% across 0 to 90o flexion) and IR (61%) during isolated and coupled displacements, except for the pivot-shift and ATT in extension. The LG resisted 28% of IR at 90o flexion. The AM fibers showed significantly greater restraint of SPS rotations than the central and posterior fibers (P<0.05). No significant differences ( all <2mm) were found between the anatomic centroid of the ‘C’-shaped attachment and the COE under most loadings.

Conclusions: The peripheral anteromedial fibers were the most important in the restraint of tibial anterior translation, internal rotation and simulated pivot-shift. These mechanical results matched the ‘C’-shaped anteromedial attachment of the dense collagen fibers.

Clinical Relevance: The most important fibers in restraining tibial displacements attach to the C-shaped anterior-medial area of the native ACL tibial attachment. This finding provides an objective rationale for ACL graft position to enable it to reproduce the physiological path of load transmission for tibial restraint.

INTRODUCTION

One aim of anterior cruciate ligament (ACL) reconstruction is to restore knee function, but up to 25% of patients have unsatisfactory functional scores [1-3]. The ACL is not a simple bundle of fibers, but consists of a complex of many fiber groups, or fascicles, and they each undergo lengthening and slackening that depends on their femoral and tibial attachments [4-7]; this complex fiber structure is commonly simplified into anteromedial (AM) and posterolateral (PL) bundles which correspond to their relative positions in the tibial attachment [8-11].

The femoral attachment has been the focus for optimising knee kinematics after ACL reconstruction, with superior rotational restraint reported with laterally-placed ‘mid-bundle’ graft positioning [12]. However, Kawaguchi et al [13] reported that the majority of restraint to tibial displacements was offered by a ridge of dense fibers across the anterior aspect of the femoral ACL attachment; fibers attaching posteriorly contributed minimal restraint. This result provided objective knowledge of where an ACL graft femoral tunnel should be placed if reproduction of the physiological path of load transmission were desired.

Assessments of the tibial plateau have reported a variable pattern of ACL fiber attachment [4,14],with a complex spiral relationship among the ACL fibers through the mid-substance [6] and little is known about their function at the tibial attachment. A recent anatomical study [15] reported a ‘ribbon-like’ appearance of the mid-substance of the ACL, corresponding to the anterior ridge of dense collagen fibers reported by Mochizuki et al [16] and their dominant load-bearing role reported by Kawaguchi et al [13]. Histological evaluation has shown a predominance of dense collagen across the anterior aspect of the tibial ACL attachment [17]. A further anatomical study [18] expanded this evidence, describing an anterior-medial C-shaped area of dense collagen fibers attaching to the tibia. However, no biomechanical evaluation has reported where loads are resisted during tibial displacements, in relation to the ACL fiber structure.

Knowledge of the anatomy and load transmission behavior of the native ACL could be used to guide tibial tunnel placement during ACL reconstruction so that its actions reproduce those of the native ACL in restraining tibiofemoral joint laxity. This data has been reported for the femoral attachment [13]; combination of that data with the present study would show the line of action of the ACL across the knee, when retraining tibiofemoral joint displacements. This line of action is known in engineering as the resultant force, and it is effectively the balanced summation of all the contributions of each sub-section of the ACL, to give the same overall tibial restraining effect at a single point within the whole ACL attachment area.

The purpose of this study: To evaluate the load-bearing functions of the fibers of the ACL tibial attachment in restraining tibial anterior translation, internal rotation, and combined anterior and internal rotation laxities in a simulated pivot-shift test.

It was hypothesized: that most of the load transmitted by the ACL would be in the relatively narrow C-shaped peripheral anteromedial attachment which has the highest collagen density.

MATERIALS AND METHODS

Changes of the restraint of tibiofemoral joint laxity were measured using a robotic test system, in response to sequential cutting of the fibers of the ACL. The tested specimens were then examined at an anatomy facility blinded to the biomechanical data in order to identify the areas where the collagen density was greatest. Finally, the biomechanical and anatomical data were reconciled, in order to relate the force transmission to the anatomy of the ACL at its tibial attachment.

Specimen Preparation

Twelve fresh-frozen cadaveric knees were procured from a tissue bank after IRB approval from [blinded] (mean age 55 years; range 29-74; 4 male, 8 female; 5 left, 7 right). A power analysis based on a previous study [13] assuming a matched-pairs one-tailed test, showed that a change of restraint of 8% could be identified with data from 8 knees with power 0.815 and alpha 0.05. Physical examination and MRI were used by an orthopaedic surgeon to exclude previous surgery or disease. The femur was cut 190 mm from the joint line and the soft tissues resected from the proximal 80 mm. The tibia was cut 160 mm from the joint line and the soft tissue resected from the distal 60 mm. A tricortical screw was inserted across the neck of the fibula and into the tibia in order to maintain the anatomic position of the proximal tibiofibular joint. A lateral para-patella arthrotomy was used, the patellar tendon divided 10 mm proximal to the tibial tubercle, the extensor mechanism reflected and Hoffa’s fat pad excised; all other tissues remained intact. The tibia was secured within a stainless-steel cylinder using poly methylmethacrylate bone cement. The femur was potted at 0 ° knee flexion with the posterior condylar axis parallel to the base of the robot.

Testing Protocol

The rectangular area containing the ACL tibial attachment was marked with pins at its corners, and the dimensions of each attachment were recorded prior to testing (Figures 1 and 2). This area was defined by the anterior border of the ACL (1) and determining the medial border as a line from the summit of the medial tibial spine (2) along the junction of the most medial ACL fibers and the medial condyle articular cartilage (3); this enabled the anteromedial corner to be marked (4) using the intersection of (1) and a line perpendicular to the medial border (dashed line). The lateral border was defined as a line from the medial edge of the lateral tibial spine (5), perpendicular to the anterior border and parallel to the medial. The area thus defined was divided into 9 equal segments (Figure 1). This rectangular grid was used to avoid any subjective influence of anatomical observations influencing the formal cutting protocol during the biomechanical testing.

The tibia was mounted within the end effector of a 6 degrees-of-freedom robotic testing system (TX90, Stäubli Ltd, Switzerland) with repeatability of 0.03 mm in translation and 15 passive flexion-extension cycles were performed to minimize tissue hysteresis before securing the femur to the fixed base unit. A 6-axis force-moment sensor (Omega 85, ATI Industrial Automation) (force resolution 0.43 N for the Z axis and 0.29 N for the X and Y axes, and torque resolution 0.009 Nm about the Z axis and X-Y resolution 0.013 Nm) guided neutralization of forces and moments across the knee, establishing a neutral path of motion in 1 o increments from 0 to 90 ° flexion; that is position control in knee flexion and force/moment control in the other 5 DOF [9,19,20]. Simulated laxity tests were performed with an intact ACL for anterior tibial translation (ATT) in response to 90 N anterior draw force and internal tibial rotation (IR) in response to 5 Nm torque at 0 °, 30 °, 60 ° and 90 ° knee flexion. A simulated pivot-shift (SPS) was performed at 0 °, 15 °, 30 °, and 45 ° using coupled internal tibial and then valgus torques of 4 Nm and 8 Nm, respectively [20]. Each laxity test was repeated 3 times to check that tissue hysteresis had been eliminated. Larger loads were not used in order to avoid stretching-out of ligament remnants during the sequential cutting protocol described below.

In addition to the nine attachment segments shown in Figure 1, the ACL blends with the anterior horn of the lateral meniscus within the lateral gutter of the tibial ACL attachment [21,22]. The contribution of this attachment to tibial restraint has not been reported, so this extra cut was added. A digital micrometer and pin-point calipers (mean error 0.10 mm, SD 0.11 mm) were used to measure the size of the rectangular area containing the ACL attachment, as shown by the pin heads. These dimensions were each divided into three to define a grid of nine equal segments, and the micrometer and calipers were used to position each cut. Transection of the fibers in each segment was performed while the knee remained undisturbed in the robot at 90 ° flexion, using a fresh No.15 scalpel blade. Each segment of the tibial attachment was isolated from the remaining ACL by means of two cuts oriented parallel to the long axis of the tibia (that is: with the blade vertical to the tibial plateau) in both the coronal and sagittal planes. All attachments were cut from anterior to posterior: with the ACL fibres in the flexed knee being close to tangential to the bone surface, and passing posteriorly from their points of tibial attachment, a vertical cut in a coronal plane transected all ACL fibres attaching anterior to the line of the cut. The fibre configuration of the ACL meant that it was not possible to start cutting from the posterior edge of the attachment. The definition of the grid and the ten cutting stages, which sectioned the entire ACL, were all performed by a single orthopaedic surgeon in every knee.

A theoretical limitation of sequential cutting is that load might be transferred across the width of the ACL, from a cut fibre area to a remaining intact area, skewing the results. Although that effect was shown to be minimal in a previous study [13], it was decided to study the possibility in this experiment. Therefore, keeping the cutting sequence of anterior row, middle row then posterior row of the nine ACL fibre attachments, each of these rows were cut from medial to lateral in 6 knees, and from lateral to medial in 6. Differences of forces in each area would allow identification of any load transfer from released to intact fibers.

The kinematics of the intact knee was replayed 3 times after each cut. The decrease in force/torque reflected the contribution of the fibers within that cut segment to restraining tibiofemoral joint laxity.

The resultant path of load transmission of the ACL to the tibia (the ‘center of effort’ COE) was calculated by combining the contributions of the nine segments for each test. The load carried by each segment was assumed to act at the center of that segment. The contribution of the LG was equally assigned to the adjacent anterolateral and centrolateral segments for this analysis.

After the biomechanical study, the specimens were sent to another center with extensive experience of studying ACL fiber structures [references blinded] for blinded anatomical characterization; the axial direction of the fiber cuts had preserved the morphology of the ACL attachment. The synovial envelope was meticulously resected to identify the morphology of the tibial attachment of the ACL (Figure 2 a). The femur was separated from the tibia and high-resolution scaled photographs of the tibial plateau were taken. A distinct area of dense collagen fibers was identified and shown in each digital image, and its centroid (‘center of area’) was calculated (Figure 2 b).

Finally, the mechanical and anatomical data were reconciled by overlaying the anatomically defined area of dense fibres (Figure 2b) over the photograph of the grid of cut ACL fibres at the end of the robotic study, for each knee. By using the dimensions of the attachment grid, the mean shape of the dense fibre area (+SD) for the set of 12 knees could be derived, and the mean +/- SD position of the centroid, then the coordinates of the mean +/-SD of the resultant force, for each clinical test examined, could be combined.

Data analysis

The mean endpoints of each set of 3 loading cycles were expressed as mean and standard deviation for the 12 specimens. When analysing the SPS, the IR torque endpoint was measured when the valgus torque was at its peak. The decrease in force or torque after each cut was converted into a percentage of the whole ACL load. To facilitate comprehension of the data from the ten cuts (That is: 9 ACL segments plus the LG), they have been presented firstly when grouped in a simple two-way plus LG split (Analysis 1), and then in a more complex set (Analysis 2). The percentage contributions from the five segments which comprised the anterior and medial segments were initially (Analysis 1) combined to represent the region of the anteromedial ‘C’-shaped [18,23] attachment area (‘A’, Figure 3) and the remaining four central, lateral and posterior fiber segments were similarly combined (‘B’, Figure 3), plus the lateral gutter LG. Although the areas A and B were centred approximately anteromedially and posterolaterally, the checkerboard of rectangular cut segments was not a split into the anatomical AM and PL bundles known in previous work, because they have variable shapes, but the analysis will provide guidance related to these. Analysis 2 (Figure 4) was developed to provide more detailed insights into the load transmission in the ACL, with 4 areas of interest plus the LG. Presentation of the raw data of 9 areas plus LG was judged to be too complex, hindering comprehension. The five parts of the attachment area were named: 1: two anteromedial segments, 2: two anterolateral segments, 3: two centrolateral segments, 4: the three posterior segments, 5: LG (Figure 4).

Two-way repeated-measures analysis of variance (ANOVA) was used to assess the main effects of each independent variable (knee flexion angle and cutting state of the ACL attachment). The dependent variable was the percentage load contribution of each segment during each test. To identify significant changes in fiber recruitment through knee motion, the percentage contribution of each segment was analysed using a one-way repeated-measures ANOVA for each tibial displacement. To investigate ‘load transfer’ between fibers following each transection, a two-way ANOVA was performed to compare the relative contributions of each of the 10 segments of the knees sequentially cut from medial to lateral with those cut from lateral to medial, across all flexion angles. An independent samples t test or the Mann-Whitney U test for non-normally distributed data was used to address statistical significance. Paired samples t tests compared the lateral-medial and anterior-posterior positions of the anatomical centroid of the dense fiber area with the center of effort during the ATT, IR and SPS at each flexion angle. Statistical analysis used SPSS v 21, IBM Corp. The level of significance was set at P < 0.05 for a single comparison.

RESULTS

Anterior Tibial Translation

The ACL cutting sequence caused a significant decrease in the force required to reach the anterior translation laxity of the intact knee at all angles of flexion (P< .001). The interaction with the flexion angle was significant (P< .001), therefore, the contribution of each segment varied across the arc of flexion (Table 1).

For the first analysis, the anteromedial fiber area (A in Figure 3) offered significantly greater restraint of anterior laxity than the posterolateral area B (P< .001) or LG (P< .001) at all angles of flexion, whilst B offered similar restraint to LG at all angles except 0 ° (P< .02) (Figure 5a). The contribution of segment A significantly increased from 0 ° to 30 ° (P< .02), while that of B significantly decreased from 0 ° to 30 ° flexion (P< .02).

For the second analysis, the anteromedial and anterolateral areas 1 and 2 (Figure 4) each offered significantly more restraint than the centrolateral and posterior fiber areas 3,4 and LG (P< .03) across all flexion angles except 0° (Figure 5b). The contribution of the anterolateral fibers (area 2) significantly increased from 0 ° to 30 ° (P< .001) and 30 ° to 60 ° (P< .01) whilst that of posterior area 4 significantly decreased from 0 ° to 30 ° (P< .05) of flexion. Thus, overall, the center of effort moved anteriorly when the knee flexed.

Internal Tibial Rotation

Cutting the ACL caused a significant decrease in the tibial rotation torque needed to reach the IR laxity of the ACL-intact state (P< .001). Whilst there was no significant overall interaction with knee flexion angle during either analysis, there was a significant increase in the contribution of the LG fibers between 0 ° and 90 ° when considered in isolation (P< .04 – Figure 6a and Table 2).

The anteromedial segment A offered significantly greater restraint of IR than the posterolateral area B at all angles of flexion (P< .03) and LG at 0 ° and 30 ° (P< .001), whilst B offered similar restraint to LG (Figure 6a).

Similar results were found when the smaller areas 1-4 plus LG were analyzed, with the anteromedial and anterolateral areas 1 and 2 dominant in restraint of tibial internal rotation near knee extension and with an increasing contribution from the LG as the knee flexed (Figure 6b).

Simulated Pivot-Shift Test

The ACL cutting sequence caused significant decreases in the force and torque required to reach the ACL-intact laxity under combined internal rotation and valgus moments (P < .001). No significant overall interaction with knee flexion angle was observed (P > .05, Table 3).

The anteromedial segment A offered significantly greater restraint of coupled ATT and IR than B and LG (P< .01) at all flexion angles except 0 ° whilst B offered similar restraint to LG at all angles of flexion (Figure 7a).

The anteromedial segment 1 offered significantly more restraint than the centrolateral and posterior areas 3 and 4, apart from when at 0 ° flexion (Figure 7b). The anteromedial area 1 resisted 32% of the SPS loading, along with the LG resisting 30%, at 0 ° flexion, while the anterolateral area 2 contributed more as the knee flexed, reaching 35% of the restraint at 90 ° flexion.

Effect of cutting sequence

A small but significant interaction was found between the coronal direction of segment transection (that is: cutting across the ACL attachment from medial to lateral versus lateral to medial) in the anterior segments (P< .04) during anterior translation, at the central anterior segment during internal rotation (P< .01), and at the anterolateral segment (P< .02) during the SPS; no significant interaction was found in the central or posterior segments. These small effects were absorbed by presenting the mean data from the two cutting sequences.

Anatomical Analysis

The coronal width of the tibial attachment of the ACL was 14±1 mm (mean±SD) with a mean sagittal depth of 19±2 mm. A distinct anterior-medial C-shaped dense fiber attachment area was found in all knees, with a mean area of 87±21 mm2. The mean centroid of the C-shape, in relation to the size of the ACL attachment, was 56±7% medial and 32±7% posterior from the anterolateral corner of the attachment. In relation to the overall size of the tibial plateau, it was 46±2% lateral from the medial border and 32±% posterior from the anterior border. This centroid was 19±2 mm anterior to the PCL and 8±1 mm lateral from the medial tibial spine.

The position of the center of effort of the ACL

The COE for the restraint of tibial displacements is shown in Figures 8 a-c. Across the range of knee flexion examined, the mean position of the COE did not differ significantly from that of the centroid of the C-shaped attachment site, being within ± 0.7 mm in the anterior-posterior direction (P=0.71 for anterior drawer, P=0.795 for internal rotation, P=0.744 for SPS), and within ± 1.1 mm in the lateral-medial direction (P=0.786 for anterior drawer, P=0.407 for internal rotation, P=0.351 for SPS). Tests at each angle of flexion found that the COE in resisting ATT was more posterior at 0 ° (P< 0.02) and more anterior at 60 ° and 90 ° (P< 0.01) flexion, and IR and SPS were more lateral at 60 ° (P< 0.03) and 45 ° (P< 0.04), respectively (Figure 8a-c) but, despite their statistical significance, these were all within 2 mm of the centroid.

DISCUSSION

The main finding of this study was that, in support of the hypothesis, the contributions of the fibers at the ACL tibial attachment to the restraint of tibial translation and rotation were consistent with a C-shaped morphology of dense collagen fibers at the anterior and medial edges of the attachment [15,18]. There was a coincidence of the centroid of the area of dense collagen fibers with the line of action of the ACL found in the mechanical tests: the mean position across the arc of knee flexion being within 0.7 mm in anterior-posterior and 1.1 mm in medial-lateral directions for the three loads imposed. Areas with high collagen fibre density are stiffer than those with low density, and so ligament elongation will cause the force to be resisted there. Previous studies of the attachment area [24] have not allowed for the variation of collagen density across it [17,25]; parts of the attachment have sparse collagen [6], reducing the role of the posterolateral area advocated previously [12,26,27,28].

The C-shaped area of dense collagen fibers contributed 68% to 92% of the resistance to ATT throughout knee flexion, with a significantly greater role at 90 ° compared to 0 ° flexion. The role of the central-posterior fibers diminished significantly with early knee flexion, consistent with ACL length change patterns – the more-posterior fibers slacken with knee flexion [4,7,29] - and with histological observations of the higher density of collagen fibers in the more anterior region of the ACL [17,25]. The changing contributions as the knee flexed explain the progressive anterior movement of the COE when the ACL restrained ATT, with knee flexion.

Further analysis found that 47% of the restraint of ATT in extension was from the most anteromedial fibers (area 1 in Figure 4) alone, decreasing with flexion. Conversely, the anterolateral fibers (area 2 in Figure 4) became the dominant restraint with flexion, contributing 54% at 90 ° flexion. This pattern is consistent with tibial internal rotation as the knee flexes [30,31]; the lateral plateau moves anteriorly and therefore tensions the more anterolateral fibers.

The anteromedial fiber area contributed >50% of the restraint from the ACL to IR across the arc of knee flexion examined. The position of the COE for the restraint of tibial IR did not differ significantly from that of the centroid of the C-shaped attachment at any angle of flexion (Figure 8b). The contribution of the LG fibers to resisting IR reached 28% at 90 ° flexion, indicating load transfer to/from the anterior horn of the lateral meniscus when resisting IR in the flexed knee. Similarly, the LG attachment contributed 30% of the resistance to the SPS in the extended knee. The data did not support a significant role for the posterolateral fibers to restraint of tibial IR.

The fibers within the anteromedial C-shaped dense fiber attachment were the primary restraint of coupled tibial displacements during the simulated pivot-shift (SPS). The position of the COE in the sagittal plane did not differ significantly from that of the centroid of the C-shaped area, across the range of angles where the SPS was tested (0 to 45 ° flexion), reaching 1.5 mm lateral at 45 ° flexion.

Hwang et al. [24] proposed tunnel positions based on their systematic review of the centers of the fiber bundle attachments. The centers of the AM and PL bundles were reported as 20 and 12 mm anterior to the PCL, respectively. The present study found the centroid of the ‘C’-shaped attachment 19±2 mm anterior to the PCL, matching the reported center of the AM bundle [14,24,32]. A similar difference was found between the mean centers of the femoral attachments of the AM and PL bundles [11], versus the COE of the ACL acting on the femur [13]. The COE findings confirmed the primary role of the ‘direct’ dense fiber attachment on the lateral intercondylar ridge [16,33,34] at the AM bundle position. The data in this study and that of Kawaguchi et al [13] reflect that the collagen density varies greatly across the ACL, so it is inappropriate to assume that the centroid of the whole attachment area corresponds to where the resultant load in the ACL acts on the bone. This is an important basic principle: the combined mechanical and histological evidence shows clearly that not all mm2 of the ACL attachment areas are equal in terms of carrying load, and this may apply to other ligaments where the collagen density varies across the width of the ligament [25]. For example, the PCL has been split into two fibre bundles: AL and PM, and it was found that the AL bundle had much higher ultimate tensile stress than the PM [35].

The COE data suggest that anatomic (central) graft tunnel positions are more posterior than the main load-bearing zone in the tibial attachment. Hara et al. [6] reported an absence of collagen fibers in the postero-central-lateral region of the ACL tibial attachment, with only fat and vascular tissue present. Other studies [18,23] observed similar features, reporting a ‘direct’ C-shaped attachment from the medial spine to the anterior root of the lateral meniscus, consistent with the findings of the present study (Figure 9).

The present study shows how fibers attaching to specific areas within the ACL tibial attachment restrained tibial displacements. Although Gabriel et al. [9] reported significantly higher in-situ forces within the mid-substance of the PL bundle than those of the AM whilst resisting ATT in extension, there is very little evidence of PL bundle deficiency leading to increased tibial rotation [36]. Lorbach et al [37] found only 1 or 2 degrees increase of tibial rotation at a torque of 10 Nm, which is above the torque that can be applied by one hand during clinical examination. In contrast, the present study found that central and posterior fibers combined contributed less than half of the restraint offered by the anterior fibers, with the posterior fibres’ contribution dropping rapidly with knee flexion. This can be reconciled by the complex structure of the ACL, with its fascicular bundles changing orientation towards their attachments [6]. Also, ‘in-situ’ forces act along the fiber orientations, and are not the same as the restraint of tibial displacements which were measured in the present study; for example, a high in-situ force in a fiber acting in a proximal-distal direction would have little effect on ATT. The fibers that are anterior within the attachment are tight throughout knee flexion and are close to isometric. The more-posterior fibers slacken with early knee flexion which is characteristic of the PL bundle [4,7,29]. Kawaguchi et al. [13] reported that the proximal anterior region of the ‘direct’ ACL femoral attachment (usually described as ‘the AM bundle’ attachment) was the most important in the control of both ATT and IR. In an anatomical study of 20 fascicular bundles of the ACL, it was observed that the fibers that attach in this region on the femur correspond with the anterior and antero-medial areas of the tibial attachment [6]. This supports the findings of the present study, of the line of action of the ACL, and is further evidence that the PL bundle has a lesser role in tibial restraint than has sometimes been suggested.

Limitations

The current study had several limitations. The age of specimens was higher than the patient demographic who experience ACL rupture, but comparable to similar cadaveric studies [12,26]. It is possible that specimen age influenced the data, if there had been degeneration of the more posterolateral fibers, but we have no data to support that idea. Testing simulated clinical manual examination and not the loading expected during gait. However, the ACL is not loaded beyond linear elastic behavior in activities of daily living, and so the position of the line of action would not change significantly with scaled-up forces. The clinical pivot shift is a dynamic examination, but we were unable to replicate this with a single robotic manipulator, so this and other studies [26,38,39] have not mimicked the in vivo kinematics but only the coupled laxities. The sequential sectioning was performed from both medial and lateral directions, finding a significant interaction between the direction of transection across only the anterior segments. Any load transfer to uncut fibers would have overestimated the role of the more-posterior fibers and, therefore, not have influenced the conclusions of the study. The effect of cutting posterior fiber areas first could not be tested, because of the inability to access them while the anterior fibers were intact.

The importance of this study is that, coupled with the matching data for the femur from Kawaguchi et al [13], we have shown the line of restraining action of the ACL. However, extrapolation of the findings of these studies to clinical ACL reconstruction should be approached with care. For example, an ACL graft tunnel placed in the anterior-medial zone of the natural ACL attachment may be subjected to impingement by the roof of the femoral intercondylar notch when the knee reaches full extension [40].

CONCLUSION

The peripheral anteromedial fibers were the most important in the restraint of tibial anterior translation, internal rotation and simulated pivot-shift. These mechanical results matched the ‘C’-shaped anteromedial attachment of the dense collagen fibers.

REFERENCES

1. Ardern CL, Webster KE, Taylor NF, Feller JA. Return to sport following anterior cruciate ligament reconstruction surgery: a systematic review and meta-analysis of the state of play. Brit J Sports Med. 2011;45(7):596-606.

2. Freedman KB, D'Amato MJ, Nedeff DD, et al. Arthroscopic anterior cruciate ligament reconstruction: a metaanalysis comparing patellar tendon and hamstring tendon autografts. Am J Sports Med. 2003;31(1):2-11.

3. Harner CD, Giffin JR, Dunteman RC, et al. Evaluation and Treatment of Recurrent Instability After Anterior Cruciate Ligament Reconstruction.Instr Course Lect. 2001; 50:463-474.

4. Amis AA, Dawkins GP. Functional anatomy of the anterior cruciate ligament. Fibre bundle actions related to ligament replacements and injuries. J Bone Jt Surg Br. 1991;73(2):260-267

5. Friederich N, O’Brien W. Functional anatomy of the cruciate ligaments. In: The knee and the cruciate ligaments. Springer, Heidelberg. 1992:78-91

6. Hara K, Mochizuki T, Sekiya I, et al. Anatomy of normal human anterior cruciate ligament attachments evaluated by divided small bundles. Am J Sports Med. 2009;37(12):2386-2391.

7. Hefzy MS, Grood ES, Noyes FR. Factors affecting the region of most isometric femoral attachments Part II: The anterior cruciate ligament. Am J Sports Med. 1989;17(2):208-216.

8. Dodds JA, Arnoczky SP. Anatomy of the anterior cruciate ligament: a blueprint for repair and reconstruction. Arthrosc: J Arthrosc Rel Surg. 1994;10920:132-139.

9. Gabriel MT, Wong EK, Woo SL, et al. Distribution of in situ forces in the anterior cruciate ligament in response to rotatory loads. J Orthop Res. 2004;22(1):85-89.

10. Harner CD, Baek GH, Vogrin TM, et al. Quantitative analysis of human cruciate ligament insertions. Arthrosc: J Arthrosc Relat Surg. 1999;15(7):741-749.

11. Piefer JW, Pflugner TR, Hwang MD, Lubowitz JH. Anterior cruciate ligament femoral footprint anatomy:systematic review of the 21st century literature. Arthrosc: J Arthrosc Relat Surg. 2012;28(6):872-881.

12. Kondo E, Merican AM, Yasuda K, Amis AA. Biomechanical comparisons of knee stability after anterior cruciate ligament reconstruction between 2 clinically available transtibial procedures: anatomic double bundle versus single bundle. Am J Sports Med. 2010;38(7):1349-1358.

13. Kawaguchi Y, Kondo E, Takeda R, et al. The Role of Fibers in the Femoral Attachment of the Anterior Cruciate Ligament in Resisting Tibial Displacement. Arthroscopy: The Journal of Arthroscopic & Related Surgery, 2015;31(3):435-444.

14. Edwards A, Bull AM, Amis AA. The attachments of the anteromedial and posterolateral fibre bundles of the anterior cruciate ligament: Part 1: tibial attachment. Knee Surg Sports Traumatol Arthrosc, 2007. 15(12): p. 1414-1421.

15. Śmigielski R, Zdanowicz U, Drwięga M, et al. Ribbon like appearance of the midsubstance fibres of the anterior cruciate ligament close to its femoral insertion site: a cadaveric study including 111 knees. Knee Surg, Sports Traumatol, Arthrosc. 2015;23(11):3143-3150.

16. Mochizuki T, Fujishiro H, Nimura A, et al. Anatomic and histologic analysis of the mid-substance and fan-like extension fibres of the anterior cruciate ligament during knee motion, with special reference to the femoral attachment. Knee Surg, Sports Traumatol, Arthrosc 2014;22(2):336-344.

17. Petersen W, Tillmann B. Structure and vascularization of the cruciate ligaments of the human knee joint. Anat Embryol. 1999;200(3):325-334.

18. Siebold R, Schumacher P, Fernandez F, et al. Flat midsubstance of the anterior cruciate ligament with tibial “C”-shaped insertion site. Knee Surg, Sports Traumatol, Arthrosc. 2015;23(11):3136-3142.

19. Kittl C, El-Daou H, Athwal KK, et al. The Role of the Anterolateral Structures and the ACL in Controlling Laxity of the Intact and ACL-Deficient Knee. Am J Sports Med. 2016;44(4):345-354.

20. Lord BR, El-Daou H, Sabnis BM, et al. Biomechanical comparison of graft structures in anterior cruciate ligament reconstruction. Knee Surg, Sports Traumatol, Arthrosc 2016;24(1):1-10.

21. Arnoczky SP. Anatomy of the anterior cruciate ligament. Clin Orthop Relat Res 1983;172:19-25.

22. Girgis FG, Marshall JL, JEM Al Monajem. The Cruciate Ligaments of the Knee Joint: Anatomical, Functional and Experimental Analysis. Clin Orthop Relat Res 1975;106:216-231

23. Śmigielski R, Zdanowicz U, Drwięga M, et al. The anatomy of the anterior cruciate ligament and its relevance to the technique of reconstruction. Bone Jt J. 2016:98-B(8);1020-1026.

24. Hwang MD, Piefer JW, Lubowitz JH. Anterior cruciate ligament tibial footprint anatomy: systematic review of the 21st century literature. Arthrosc : J Arthrosc Relat Surg. 2012;28(5):728-734.

25. Mommersteeg T, Kooloos J, Blankevoort L, et al. The fibre bundle anatomy of human cruciate ligaments. J Anat. 1995;187(2):461-471.

26. Kim D, Asai S, Moon C-W, et al. Biomechanical evaluation of anatomic single-and double-bundle anterior cruciate ligament reconstruction techniques using the quadriceps tendon. Knee Surg Sports Traumatol Arthrosc. 2015; 23(3):687-695.

27. Morimoto Y, Ferretti M, Ekdahl M, et al. Tibiofemoral joint contact area and pressure after single- and double-bundle anterior cruciate ligament reconstruction. Arthrosc : J Arthrosc Relat Surg 2009;25(1):62-69.

28. Musahl V, Voos JE, O'Loughlin PF, et al. Comparing stability of different single- and double-bundle anterior cruciate ligament reconstruction techniques: a cadaveric study using navigation. Arthrosc: J Arthrosc Relat Surg. 2010;26(9 Suppl):S41-S48.

29. Sidles JA, Larson RV, Garbini JL, et al. Ligament length relationships in the moving knee. J Orthop Res. 1988;6(4):593-610.

30. Amis AA. The functions of the fibre bundles of the anterior cruciate ligament in anterior drawer, rotational laxity and the pivot shift. Knee Surg, Sports Traumatol, Arthrosc. 2012;20(4):613-620.

31. Lord BR, Amis AA. The Envelope of Laxity of the Pivot Shift Test. In: Rotatory Knee Instability. Springer, Heidelberg. 2017:223-234.

32. Amis AA, Jakob RP. Anterior cruciate ligament graft positioning, tensioning and twisting. Knee Surg, Sports Traumatol, Arthrosc. 1998;6(Suppl 1): S2-12

33. Iwahashi T, Shino K, Nakata K, et al. Direct anterior cruciate ligament insertion to the femur assessed by histology and 3-dimensional volume-rendered computed tomography. Arthrosc : J Arthrosc Relat Surg. 2010; 26(9 Suppl):S13-S20.

34. Sasaki N, Ishibashi Y, Tsuda E, et al. The femoral insertion of the anterior cruciate ligament: discrepancy between macroscopic and histological observations. Arthrosc: J Arthrosc Relat Surg. 2012;28:1135-1146.

35. Race A, Amis AA. The mechanical properties of the two bundles of the human posterior cruciate ligament. J Biomechs 1994;27: 13-24.

36. Amis AA. The functions of the fibre bundles of the anterior cruciate ligament in anterior drawer, rotational laxity, and the pivot-shift. Knee Surg, Sports Traumatol, Arthrosc. 2012;20:613-620.

37. Lorbach O, Pape D, Maas S, et al. Influence of the anteromedial and posterolateral bundles of the anterior cruciate ligament on external and internal tibial rotation. Am J Sports Med 2001;38:721-727.

38.Goldsmith MT, Jansson KS, Smith SD, et al. Biomechanical Comparison of Anatomic Single-and Double-Bundle Anterior Cruciate Ligament Reconstructions An In Vitro Study. Am J Sports Med. 2013;41:1595-1604.

39. Yagi M, Wong EK, Kanamori A, et al. Biomechanical analysis of an anatomic anterior cruciate ligament reconstruction. Am J Sports Med. 2002;30:660-666.

40. Howell SM, Taylor MA. Failure of reconstruction of the anterior cruciate ligament due to impingement by the intercondylar roof. J Bone Joint Surg Am 1993;75:1044-1055.

Captions for illustrations

Figure 1. Drawing of the tibial plateau of a right knee. The area containing the tibial attachment was defined by the anterior border of the ACL (1) and determining the medial border as a line from the summit of the medial tibial spine (2) along the junction of the most medial ACL fibers and the medial condyle articular cartilage (3); this enabled the anteromedial corner to be marked (4) using the intersection of (1) and a line perpendicular to the medial border (dashed line). The lateral border was defined as a line from the medial edge of the lateral tibial spine (5), perpendicular to the anterior border and parallel to the medial. LM: lateral meniscus; MM: medial meniscus.

Figure 2. Typical tibial ACL attachments. (a): Specimen as received for anatomical measurement, following the cutting study. In this left knee, the dense fibre area is seen passing lateral-medial below the mm scale, then curving posteriorly along the medial aspect of the tibial intercondylar space. (b): The dense fiber area was distinct, and was defined (solid line) while ‘blinded’ to the mechanical testing. The area containing the tibial attachment as defined in Figure 1 is shown by the interrupted lines. ‘X’ is the centroid of the dense fiber attachment area.

Figure 3. The first (relatively simple) analysis (Analysis 1) grouped together the data for the contributions of the anterior and medial fiber area (‘A’), the central and posterior-lateral (‘B’) and the fibers within the lateral gutter (LG). The roles of these three areas in resisting anterior tibial translation, internal tibial rotation torque and coupled moments (internal rotation and valgus) during a simulated pivot shift were examined.

Figure 4. The second analysis (Analysis 2) examined the roles of five areas of fiber attachment: the anteromedial (1), the anterolateral (2) and centrolateral (3), the posterior (4) and the fibers within the lateral gutter (LG) in resisting anterior tibial translation, internal tibial rotation torque and coupled moments (internal rotation and valgus) during a simulated pivot shift.

Figure 5a. Analysis 1: The percentage contribution of the combination of ACL tibial attachment segments A, B, and LG to the restraint of anterior tibial translation. * Significant difference observed (P < .05); A: anteromedial, B: posterolateral, LG: lateral gutter.

Figure 5b. Analysis 2: The percentage contribution of the combination of ACL tibial attachment segments 1-4 and LG to the restraint of anterior tibial draw. * Significant difference observed (P < .05). The five parts of the attachment area: 1: anteromedial, 2: anterolateral, 3: centrolateral, 4: posterior, 5: lateral gutter.

Figure 6a. Analysis 1: The percentage contribution of the combination of ACL tibial attachment segments to the restraint of internal tibial rotation torque. * Significant difference observed (P < .05); A: anteromedial, B: posterolateral, LG: lateral gutter.

Figure 6b. Analysis 2: The percentage contribution of the combination of ACL tibial attachment segments to the restraint of internal tibial rotation torque. * Significant difference observed (P < .05). The five parts of the attachment area: 1: anteromedial, 2: anterolateral, 3: centrolateral, 4: posterior, 5: lateral gutter.

Figure 7a. Simulated pivot-shift: kinematic analysis 1: The contribution of the fibers within each group of ACL tibial attachment segments to the restraint of coupled tibial internal rotation (4 N-m) and valgus displacement (8 N-m) expressed as a percentage of the combined moments (12 N-m). * Significant difference observed (P < .05); A: anteromedial, B: posterolateral, LG: lateral gutter.

Figure 7b. Kinematic analysis 2: The contribution of the fibers within each group of ACL tibial attachment segments to the restraint of coupled tibial internal rotation (4 N-m) and valgus displacement (8 N-m) expressed as a percentage of the combined moments (12 N-m). * Significant difference observed (P < .05). The five parts of the attachment area: 1: anteromedial, 2: anterolateral, 3: centrolateral, 4: posterior, 5: lateral gutter.

Figure 8. The shaded area is the mean extent of the dense ACL fiber attachment, as identified in the ‘blinded’ anatomical examination, shown as percent of the size of the ACL attachment. The origin of the graph is the anterolateral corner of the grid in Figure 1. The solid line is the centerline along the area, from lateral to medial. The variability of the area is shown by the SD bars extending from the shaded area.

(a) The mean ± SD of the center of effort in the ACL tibial attachment in the restraint of anterior tibial translation, across 0 to 90o knee flexion. O: the centroid of the dense ACL fiber attachment, mean ± SD; * significant difference from the centroid in the sagittal plane. The center of effort migrated anteriorly as the knee flexed, as the posterior ACL fibers slackened.

(b) The mean ± SD of the center of effort in the ACL tibial attachment in the restraint of tibial internal rotation torque. * significant difference from the centroid in the coronal plane.

(c) The mean ± SD of the center of effort in the ACL tibial attachment in the restraint of tibial displacement during the coupled moments of a simulated pivot shift. * significant difference from the centroid in the coronal plane.

Figure 9: In this ACL-deficient right cadaveric knee, viewed through an anterolateral portal, the tip of the drill guide is located near the centroid of the C-shaped area of dense collagen fibers, which form a distinct ‘wall’ anterior to it and – behind it in the photograph - medially. An ACL graft tunnel placed here will cover the data shown in Figure 8, for the line of action of the native ACL tension restraining the tibia.

Tables

Table 1 Contribution (in Percentages) of Each Segment in Restraining 90-N of Anterior Tibial Draw (mean +/- SD)

Flexion Angle

Segment

0°

30°

60°

90°

A

68 ± 18

85 ± 10

91 ± 5

90 ± 5

B

26 ± 17A

10 ± 9A

6 ± 4A

6 ± 4A

LG

6 ± 9A,B

5 ± 6A

3 ± 4A

4 ± 4A

1

46 ± 24

46 ± 22

42 ± 22

36 ± 25

2

17 ± 16

38 ± 24

49 ± 25

54 ± 26

3

15 ± 13

7 ± 91,2

4 ± 41,2

3 ± 41,2

4

16 ± 14

4 ± 31,2

3 ± 41,2

3 ± 31,2

LG

6 ± 91

5 ± 61,2

2 ± 41,2

4 ± 41,2

Superscript denotes a significant difference (P < .05) from the stated segment.

Table 1 Contribution (in Percentages) of Each Segment Restraining a 5 N-m Internal Rotation Torque (Mean +/- SD)

Flexion Angle

Segment

0°

30°

60°

90°

A

67 ± 28

68 ±18

53 ± 27

61 ± 21

B

25 ± 21A

19 ± 13A

21 ± 15A

12 ± 14A

LG

8 ± 14A

13 ± 18A

26 ± 26

27 ± 27

1

49 ± 29

33 ± 20

19 ± 18

24 ± 23

2

17 ± 11

32 ± 15

30 ± 15

30 ± 13

3

12 ± 111

12 ± 112

14 ± 8

5 ± 82

4

14 ± 19

10 ± 92

12 ± 112

13 ± 15

LG

8 ± 141

13 ± 18

25 ± 26

28 ± 27

Superscript denotes a significant difference (P < .05) from the stated segment.

Table 2 Contribution (in Percentages) of Each Segment in Restraining Coupled 4 N-m Internal Rotation 8 N-m Valgus Torques (Mean +/- SD)

Flexion Angle

Segment

0°

15°

30°

45°

A

51 ± 28

67 ± 13

72 ± 13

66 ± 25

B

19 ± 23A

12 ± 7A

16 ± 13A

19 ± 16A

LG

30 ± 28

20 ± 17A

12 ± 11A

15 ± 12A

1

32 ± 24

38 ± 21

34 ± 19

21 ± 16

2

14 ± 12

24 ± 23

31 ± 21

35 ± 29

3

10 ± 11

7 ± 61

9 ± 51

10 ± 13

4

14 ± 21

10 ± 121

14 ± 17

20 ± 28

LG

30 ± 28

20 ± 17

12 ± 11

15 ± 12

Superscript denotes a significant difference (P < .05) from the stated segment.