KneeDaily

1.6

Role of the Anterior Cruciate Ligament in Rotational Stability

Yongtao Maov1.0Published 11 references
Chinese original
Contents

Cadaveric biomechanical testing and examination in sports medicine often seem to conflict. On a six-degree-of-freedom rig, an isolated internal-rotation moment of 5 N·m usually does not produce a statistically significant increase in isolated internal tibial rotation at any flexion angle after the anterior cruciate ligament (ACL) is sectioned alone [1,2]. During a sudden stop and cut, or during a pivot-shift test on the examination table, an ACL-deficient knee can show marked anterior subluxation of the lateral tibial plateau and a clunk of reduction as the knee flexes [3]. Treating “rotational stability” as a change in internal tibial rotation about the long axis of the tibia leaves that typical pivot shift unexplained, and it mixes up the mechanics behind a negative examination and a positive one. Isolated internal tibial rotation, lateral tibial plateau translation, and pivot-shift reduction during continuous flexion are three different physical motions. What the ACL contributes to each depends directly on the external load and the flexion angle.

1. Internal tibial rotation, lateral tibial plateau translation, and pivot-shift reduction

In three-dimensional knee kinematics, bony motion is defined by the change in position and orientation of the tibia relative to the femur [3]. Clinical discussions often group three distinct kinematic quantities under “rotational instability.” These quantities are not interchangeable.

Internal tibial rotation, measured here as the isolated axial internal-rotation angle, is the angle, in degrees, through which the tibia rotates internally about its own long axis when a pure internal-rotation moment is applied at a specified flexion angle [1,2]. It describes how far the knee yields under a single torsional moment in the transverse plane.

Lateral tibial plateau translation, the displacement underlying anterior subluxation of the lateral tibial plateau, is the anterior linear displacement, in millimeters, of the lateral tibial compartment relative to the lateral femoral condyle [2,3]. In a knee made of medial and lateral compartments, axial internal rotation and compartmental anterior translation do not rise and fall together in a fixed ratio under combined loading. When the lateral plateau translates farther anteriorly than the medial plateau, the geometry is anterior translation accompanied by internal tibial rotation. When an added anterior shear force drives the medial compartment farther forward, the lateral plateau may already be markedly subluxated anteriorly while net internal tibial rotation is smaller than in the intact loaded knee [2].

Noyes et al. separated axial rotation from compartmental translation in group 1, seven male cadaveric knees (mean age, 57 ± 18 years; range, 24–79 years), using a six-degree-of-freedom robot under four-degree-of-freedom loading [2]. With an isolated 5 N·m internal-rotation moment at 25°, 60°, and 90° of flexion (Table 3 of that paper), ACL sectioning changed isolated internal tibial rotation, relative to the intact knee, by 0.6° (95% confidence interval [CI], −0.3° to 1.4°), 0.0° (−0.1° to 0.2°), and 0.0° (−0.2° to 0.2°). In the same specimens, a 100 N anterior force, a 7 N·m valgus moment, and a 5 N·m internal-rotation moment were applied during flexion–extension and sampled at 25° (simulated pivot shift 1; Table 2 of that paper). ACL sectioning increased anterior translation of the lateral compartment by 5.8 mm (95% CI, 3.7–7.9 mm) and anterior translation of the medial compartment by 8.3 mm (4.5–12.2 mm). Net internal tibial rotation changed by −4.0° (−7.9° to −0.1°) relative to the intact loaded knee. With the internal-rotation moment reduced to 1 N·m (simulated pivot shift 2), anterior translation increased by 8.8 mm (3.9–13.7 mm) in the lateral compartment and by 11.8 mm (6.8–16.9 mm) in the medial compartment, and net internal tibial rotation changed by −4.6° (−8.9° to −0.2°) [2]. Once the ACL is absent, the 100 N anterior force also draws the medial plateau substantially forward. A change in the internal-rotation angle alone misses anterior subluxation of the lateral compartment. If net internal tibial rotation is negative, that angle can also be misread as a change in the opposite direction.

Pivot-shift reduction is a two-stage dynamic transition. Anterior subluxation of the lateral plateau near extension shifts suddenly backward as flexion deepens, and the tibia rotates externally with that shift. Internal tibial rotation does not increase continuously with flexion through the event [3]. Naendrup et al. recorded 520 manual pivot-shift maneuvers with a six-degree-of-freedom electromagnetic tracking system (displacement accuracy, 0.5 mm; angular accuracy, 0.5°) on two fresh-frozen whole-lower-limb cadaveric specimens. Twenty-five clinicians of differing experience performed the maneuvers (five categories, five examiners in each, 20 repetitions each), as did one senior orthopedic surgeon (10 repetitions on each knee, 20 in total) [3]. Both specimens were prepared by arthroscopic ACL sectioning combined with lateral meniscectomy. Specimen 1 had high-grade laxity with a gross clunk. Specimen 2 had low-grade laxity with a glide. The records are repeated manual maneuvers, without muscle resistance, on two specimens with combined injury. They do not represent 520 patients, or a population with an isolated ACL rupture [3].

The measured six-degree-of-freedom curves, and Table 1 of that paper, separate subluxation from reduction. Near extension, with internal-rotation and valgus stress applied, the lateral tibial plateau lies anterior to the lateral femoral condyle. In the positive tests that entered the analysis, abrupt reduction began at 34.9° ± 9.1° in the high-laxity specimen and at 34.7° ± 7.8° in the low-laxity specimen. The lateral compartment moved posteriorly within a mean of 0.18 s. The reduction displacements were 17.3 ± 8.8 mm and 8.8 ± 10.5 mm, with tibial external rotation of −19.3° ± 7.9° and −10.0° ± 9.2°, and with a return into varus [3]. The background of that paper cites the classic mechanism. The experiment itself did not measure iliotibial band (ITB) tension or the path of the band. On that cited account, reduction requires the examiner to release valgus and internal-rotation stress together during flexion, so that passive posterior pull from the ITB guides the lateral tibial plateau back [3]. The pivot-shift “clunk” felt by the examiner is this posterior glide of the lateral plateau with external rotation. The values of about 35° and 0.18 s are pooled results from the screened tests on two specimens with combined injury. They are not fixed clinical constants of angle or time. Strong correlations among the degrees of freedom mainly reflect geometric coupling inside that same reduction [3].

Dimensional inconsistency in the original threshold for a successful test. In the “Quantitative criteria...” paragraph of the Methods, Naendrup et al. counted a maneuver as a successful pivot-shift test only when the difference between maximum and minimum anterior translation of the lateral compartment exceeded the “square root of the standard deviation” of all lateral-compartment anterior-translation values. The translation difference is in millimeters. The square root of the standard deviation is in square-root millimeters (√mm). The two quantities do not share a dimension. The analysis code was not released, so the screening formula that the program actually ran cannot be confirmed, and the criterion should not be rewritten, without that code, as the square root of the variance. The 201 positive tests that entered the analysis, and the quantitative summaries of their starting angles and durations, all depend on this screening step. Those pooled values, and the so-called test “success rates” (54%, 30%, and 40% overall), cannot be used as a uniform standard for manual technique or as an index of clinical diagnostic performance. The direction of motion shown by the measured curves can still be used: anterior subluxation near extension, then posterior translation with external rotation as the knee flexes [3].

The ACL runs from the medial wall of the lateral femoral condyle obliquely forward, medially, and distally to the anterior intercondylar area of the tibia. For a moment arm about the central long axis of the tibia, the ACL lies near the joint center. At 30° of flexion and beyond, that moment arm is much shorter than the moment arms of the superficial and deep ITB and of the anterolateral capsule, which lie lateral to the axis of rotation [1,4]. A short axial moment arm does not imply that the ACL takes no part in rotational stability.

At full extension (0°), part of the ACL fiber bundles is loaded in tension and crosses the center of the joint obliquely, so the ligament can share a pure internal-rotation moment directly. In the six-degree-of-freedom robotic experiments of Willinger et al., the ACL contribution to the moment resisting 5 N·m of pure internal rotation at 0° was as high as 16% (Figure 2 marks P < .05) [1]. In the ACL-intact robotic series of Kittl et al. (eight knees), under 5 N·m of pure internal rotation at 0°, only the ACL reached statistical significance as an individual contribution to the resisting moment (19.7% ± 7.9%, P < .01) [4].

Anterolateral rotatory instability requires anterior subluxation of the lateral tibial plateau on the lateral femoral condyle near extension [1-3]. In the sagittal plane the lateral tibial plateau is slightly convex or flat. Near extension, valgus compression and an internal-rotation moment generate a shear component at the lateral tibiofemoral surfaces that drives the lateral plateau forward. The ACL is the primary passive restraint to anterior tibial translation (see Chapter 1.5). From 0° through low and moderate flexion it directly limits anterior glide of the tibial plateau [1,2,5]. After the ACL ruptures, that primary passive anterior restraint near extension is lost. Under combined internal rotation and valgus, the lateral plateau readily subluxates anteriorly, and a reduction shift can then appear during flexion [1,3].

2. Loading, flexion angle, and the measured quantity

Numbers reported from rotational biomechanics often differ widely. Specimens and tissue handling differ, but the experiments also do not measure the same mechanical quantity. Load control in a laboratory “simulated pivot shift” differs from a dynamic pivot-shift examination at the bedside.

Cadaveric studies use three classes of measurement. They cannot be converted into one another.

  1. In situ force, or resultant ligament force measured directly (N). While an intact or reconstructed knee carries an external load, tension in the whole ACL, in a bundle, or in a graft is measured with a load cell on an isolated bone cap at the tibial insertion [5], a graft-force sensor [2], or six-degree-of-freedom robotic path replay [6,7]. The value is the load carried at that posture and external load. It does not give the free displacement, or the increase in rotation, after the ligament is cut.
  2. Contribution ratio of resisting force or resisting moment during replay of the baseline path (%). A six-degree-of-freedom robot first records the baseline three-dimensional path under a prescribed load. Target structures are then cut in sequence, the same bony path is replayed, and the percentage is calculated from the change in force (N) or moment (N·m) [1,4,8]. The denominator is not defined in the same way in every study. It cannot be read throughout as “the sum of the resisting forces in every tissue that was cut.” Kittl et al. used the baseline force or moment of the corresponding initial reference state, ACL-intact or ACL-deficient, as the denominator (p. 349) [4]. Parsons et al. divided the magnitude of each ligament’s in situ force vector by the magnitude of the intact force vector in the initial prepared state, and stated on p. 672 that the structural percentages do not sum to 100% [8]. The four percentages of internal-rotation resisting moment in Figure 2 on p. 1139 of Willinger et al. likewise do not sum to 100%. The Methods do not give a formula that normalizes those four structures [1]. A percentage of this kind is the share of resisting force or resisting moment along one baseline path. It is a different quantity from the change in kinematic laxity when the joint is allowed to move.
  3. Increment in free kinematic laxity under a constant external load (mm or °). After the target structure is cut, the same external force (N) or moment (N·m) is applied again and the tibia is allowed to settle at a new equilibrium. The increase in anterior translation or in rotation is then measured [1,2]. Soft-tissue load–displacement behavior is nonlinear. A structure that carried 15%–20% of the resisting moment on the baseline path may, once cut, increase free rotation by less than 1°–2° if the remaining structures tighten and take up load. In a sequential-sectioning model, free displacement or rotation can increase markedly when the primary anterior restraint is cut, and again when the last parallel lateral restraint is cut [1].

A laboratory “simulated pivot shift” and a clinical manual pivot shift are also different tests. The clinical test produces subluxation near extension and then continues into flexion, looking for a rapid clunk as the lateral plateau reduces posteriorly with external rotation [3]. Laboratory protocols are limited by the fixtures and the control mode. They generally lack axial compression from body weight, active muscle contraction, and dynamic proximal pull on the ITB, and they do not use the timing or the acceleration of the clinical clunk as an endpoint [1,2,4,6,7]. Kittl et al. applied an 8 N·m valgus moment plus a 4 N·m internal-rotation moment only at fixed angles of 15°, 30°, and 45°. They did not test 0°, applied no external anterior force, and reported only the proportional drop, at peak valgus, in the moment resisting 4 N·m of internal rotation [4]. Willinger et al., on an optical kinematic rig, moved the knee passively from 0° to 100° under a 5 N·m internal-rotation moment plus an 8 N·m valgus moment, with no external anterior force and with the tibia hanging vertically, and recorded quasi-static envelopes of anterior translation and internal rotation [1]. Noyes et al. applied a 100 N anterior force, a 7 N·m valgus moment, and a 5 or 1 N·m internal-rotation moment during flexion–extension, and reported compartment translation and rotation only at 25° [2]. Millimeters or degrees from a laboratory “simulated pivot shift” are not a clinical pivot-shift grade.

Markolf et al. mechanically isolated a subchondral bone cap at the tibial ACL attachment in 14 fresh-frozen cadaveric knees from donors aged 50–78 years and connected the cap to a three-degree-of-freedom load cell. The cell was zeroed at 30° of flexion, slightly posterior to the midpoint of tibial anteroposterior travel. Zero degrees was defined by a 2.5 N·m extension moment. The tibia could rotate about its long axis and could angulate into varus and valgus in the horizontal plane. The knee was extended by hand, passively, from 90° to 5° of hyperextension at about 10°/s, and the resultant force in the whole ACL was recorded continuously. Under high load the bone cap occasionally pulled out of elderly cancellous bone. Figures 2A and 3A of that paper therefore contain 12 knees, Figure 3B contains 13 knees, and the other figures contain 14 knees [5].

Resultant ACL forces from single external loads cannot be added to predict combined loading, and the direction of the added internal-rotation effect reverses with flexion [5]. Under a 100 N anterior tibial force alone, the mean resultant ACL force near 30° was about equal to the applied anterior force, and at 0° it rose to about 150% of that force. When a 10 N·m internal-rotation moment was added (Figure 2A, 12 knees), the resultant ACL force was significantly higher than under the 100 N anterior force alone at less than 20° of flexion, including 0° and 5° of hyperextension. From 20° to 40° the difference was not statistically significant. Above 40°, the resultant ACL force was significantly lower than under the anterior force alone [5]. A 100 N anterior force combined with a 10 N·m external-rotation moment did the opposite: beyond 10° of flexion it significantly lowered the resultant ACL force, which approached zero near 90° (Figure 2B, 14 knees). From 10° to 70°, a 10 N·m internal-rotation moment combined with a 10 N·m valgus moment produced a resultant ACL force significantly higher than a 10 N·m internal-rotation moment alone. Near full extension that difference did not reach significance (Figure 3A, 12 knees) [5]. Added load from internal rotation is therefore concentrated near extension when anterior shear is also present, and in mid-flexion when valgus is combined with internal rotation. The measurements were passive extension of elderly specimens, without active muscle force or ground-reaction force, and bone-cap failure depended on the strength of elderly cancellous bone. They cannot be used to derive an in vivo load at which the ligament fails, a ranking of injury probability, or a safe flexion angle for rehabilitation after surgery.

Kittl et al. used six-degree-of-freedom robotic path replay. They tested 18 knees and included 16 fresh-frozen knees (mean age, 69 years; range, 52–93 years), eight with the ACL intact and eight with the ACL deficient. In a fixed superficial-to-deep order they resected the superficial ITB and related supporting tissues (sITT), the deep and capsulo-osseous ITB (dcITT), the extracapsular anterolateral ligament (ALL), the anterolateral capsule (Cap), and the ACL. The percentage of resisting moment was calculated against the moment in the corresponding initial state [4].

Under a 5 N·m pure internal-rotation moment (Table 2 of that paper), the ACL-intact group at 0° had only one statistically significant contribution to the internal-rotation resisting moment, from the ACL (19.7% ± 7.9%, P < .01). At 30°, 60°, and 90° the ACL contribution fell to 8.9% ± 7.7%, 1.0% ± 1.5%, and 0.9% ± 0.8%, none of them significant. The ITB system became the main restraint to internal rotation. Together, sITT and dcITT accounted for 44% ± 10%, 76% ± 11%, and 71% ± 15% at 30°, 60°, and 90° [4]. In the simulated pivot shift at 15°, 30°, and 45° (8 N·m valgus plus 4 N·m internal rotation; Table 3 of that paper), the ACL contribution in the ACL-intact group was reported only as 16.8% ± 10.9%, 10.4% ± 9.9%, and 5.1% ± 5.6%, none significant [4]. That lower share fits the protocol. Zero degrees was not tested, no anterior force was applied, the ACL was resected last after all of the anterolateral tissues, and the only statistic was the drop, at peak valgus, in the moment resisting 4 N·m of internal rotation.

Willinger et al. included 22 unpaired cadaveric knees (mean age, 57 years; range, 47–65 years). Ten were used for six-degree-of-freedom robotic path replay at fixed angles of 0°, 30°, 60°, and 90°. Another 12 were mounted on a six-degree-of-freedom optical kinematic rig that allowed the knee to displace freely. Two were excluded for technical error, and 10 were analyzed during passive flexion–extension from 0° to 100°. Within one study, resisting-moment contribution on the replayed path was set against the increment in free kinematic laxity [1].

  • Pure internal-rotation moment of 5 N·m. The robot tested four structures: the ACL, the Kaplan fibers (KFs), the anterolateral capsule/ligament (C/ALL), and the lateral meniscus posterior root (LMPR). The ACL contributed significantly to the internal-rotation resisting moment only at 0° (as high as 16%; Figure 2 marks P < .05). At 30° and above, that contribution was not significant. At 60° and 90° the KFs were significantly higher than the C/ALL. At 90° the KFs reached 44% ± 23% (P < .01) and the C/ALL was 14% ± 13% (P < .05) [1]. On the optical rig, isolated ACL sectioning did not significantly increase the isolated internal-rotation angle at any position from 0° to 100° [1].
  • Quasi-static simulated pivot shift, 5 N·m internal rotation plus 8 N·m valgus. After the valgus moment was added, and with no external anterior force, isolated ACL sectioning still significantly increased anterior tibial translation from 0° to 50° of flexion (Figure 5A). It also produced a small but significant increase in internal tibial rotation from 0° to 40° (mean ≤ 2°; Figure 5B marks P < .05) [1]. Combined sectioning of the lateral structures (KFs + C/ALL + LMPR) then increased anterior translation and internal rotation further, and significantly, from 10° to 100° (P < .001 relative to isolated ACL sectioning). From 50° to 100°, internal rotation increased by an additional mean of 7°. After the KFs were sectioned, the increase was at most 5°. At 100°, total internal rotation was 8° greater than in the intact knee [1]. Under combined valgus and internal rotation, isolated ACL sectioning appears mainly as anterior translational instability from 0° to 50°, rather than as a large increase in isolated internal tibial rotation.

The statistical methods and the results statement on p. 1139 do not agree. On that page, Willinger et al. state in the statistical-methods paragraph that laxity increments after sectioning a single secondary restraint in the kinematic experiment were not tested and were reported only descriptively, with 3 or 4 knees in each of three secondary sectioning orders. The results paragraph on the same page says that isolated sectioning of the KFs, the C/ALL, or the LMPR in the ACL-deficient knee “did not significantly increase” anterior tibial translation. That phrase was not backed by a test. It cannot be used as statistical evidence that a single lateral structure has no effect. The effect of sectioning the whole ACL, and the effect of combined sectioning of the lateral structures, were analyzed by repeated-measures analysis of variance and can still be used [1].

Noyes et al. tested reconstruction and extra-articular tenodesis in 15 male cadaveric knees with a six-degree-of-freedom robot and four-degree-of-freedom loading [2]. The two groups were not built the same way. Group 1 (7 knees; age, 57 ± 18 years) was tested in order: intact knee, ACL sectioning, 10 mm single-bundle bone–patellar tendon–bone (BPTB) reconstruction in the central footprint, sectioning of the ALL and the capsulo-osseous ITB, and manual-tension ITB tenodesis, with 13.3–22.2 N applied at 10° of flexion in neutral rotation. Group 2 (8 knees; age, 51 ± 6 years; range, 42–62 years) kept the ACL intact throughout. Only after the ALL/ITB had been sectioned in that intact knee was a low-tension ITB tenodesis added, with 8.9 N applied at 30° of flexion in neutral rotation [2]. The groups differ at once in ACL status, tenodesis construct, fixation angle, and tension. They are not a randomized comparison of fixation tension alone.

In group 1, under both simulated pivot shifts sampled at 25° during flexion–extension loading, central BPTB reconstruction reduced the lateral-compartment anterior-translation difference after ACL sectioning to 0.7 mm (95% CI, −0.6 to 1.9 mm) and 1.0 mm (−0.5 to 2.6 mm) relative to the intact knee. In situ force in the BPTB graft was 153.5 N (95% CI, 108.0–199.0 N) and 162.6 N (145.1–180.0 N) [2]. After the ALL and the capsulo-osseous ITB were also sectioned, lateral-compartment anterior translation in the 25° simulated pivot shift changed by 0.3 mm (0.1–0.5 mm) and 0.1 mm (−0.3 to 0.6 mm). Isolated internal tibial rotation under 5 N·m increased by 2.9° at both 60° and 90° (95% CIs, 0.9°–4.8° and 0.6°–5.2°). After the additional manual-tension ITB tenodesis at 13.3–22.2 N, isolated internal tibial rotation at 25°, 60°, and 90° decreased by 7.3°, 16.1°, and 21.9° relative to the BPTB-reconstructed state [2].

These means are not the same statement as “BPTB reconstruction restored statistical equivalence for every measure.” Equivalence was tested with two paired one-sided t tests (TOST). The prespecified bounds were ±2 mm for displacement and ±2° for rotation, and statistical equivalence required P < .05. A result of P = .82 or P = .99 means equivalence was not shown. These P values must not be interpreted as “no difference” by applying the logic of conventional tests for a difference. By the P values in the text on p. 2688, the comparisons of compartment anterior translation after BPTB reconstruction in group 1 under the 25° simulated pivot shift (P = .07, .13, .14, and .18) all failed the P < .05 equivalence criterion. Under a 100 N anterior pull at 25° (Lachman loading), a medial-compartment anterior-translation difference of 2.3 mm remained (−0.7 to 5.3 mm), with an accompanying change in internal tibial rotation of −4.3° (−8.6° to −0.1°) [2].

Statistics in the original abstract, text, and tables do not agree (pp. 2683, 2688–2690, and 2692). The paper contains three sets of inconsistent results. (1) After BPTB reconstruction in group 1, the lateral-compartment anterior-translation difference in simulated pivot shift 1 was 0.7 mm. The abstract on p. 2683 lists P = .70. The text on p. 2688 lists P = .07. The paper does not say which figure is a typographical error. As printed, neither meets P < .05 for equivalence. A mean that lies close to the intact knee, and equivalence that has been demonstrated, have to be stated separately. (2) In group 2, the ACL-intact knees, low-tension ITB tenodesis left a difference of −4.2° in isolated internal tibial rotation at 60° relative to the intact knee. The text on p. 2690 and Table 5 on p. 2692 give a 95% CI of −9.8° to 1.4°. The abstract on p. 2683 gives the 95% CI of 4.2° as 1.9°–6.8°, which is the interval belonging to the preceding increase of 4.3° after ALL/ITB sectioning. That abstract interval cannot be used to confirm mild overconstraint at 60°. (3) The abstract on p. 2683 says that manual-tension tenodesis reduced lateral anterior translation in simulated pivot shift 1 by 4.8 mm, with a 95% CI of 1.4–8.1 mm. Table 2 on p. 2688 and the text on p. 2689 give −8.8 to −0.9 mm, a magnitude of 0.9–8.8 mm. The abstract also gives P = .99 for the 2.9° increase in internal rotation at 90° after ALL/ITB sectioning in group 1. The text on p. 2688 gives P = .82. Interpretation sets aside the conflicting abstract P values and the mismatched intervals. The means in the text and in the original tables are used to describe the separation of compartment translation from isolated internal tibial rotation, and the way this time-zero model carried load [2].

Table 1. Methodologic features and the main differences in direction of effect in four key cadaveric experiments on rotation

Study, sample, and loading Measurements and ACL findings
Markolf et al., 1995 [5] (14 knees; 12–13 knees in some combinations). Loading and motion control: Continuous passive extension from 90° to 5° of hyperextension; 100 N anterior force combined with 10 N·m rotation or varus–valgus Measured quantity (none included body-weight compression or recording of the clinical clunk): Resultant force in the whole ACL (N), measured directly from an isolated tibial bone cap. Principal difference related to the ACL: Internal rotation added to an anterior force increased load below 20° and decreased it above 40°; valgus added to internal rotation gave a higher force than internal rotation alone from 10° to 70°
Kittl et al., 2016 [4] (16 knees analyzed; 8 intact and 8 deficient). Loading and motion control: Robot at fixed angles: 5 N·m pure internal rotation; valgus plus internal rotation at 15°–45° (0° not tested; no anterior force) Measured quantity (none included body-weight compression or recording of the clinical clunk): Share of resisting moment (%) on replay, relative to the baseline of the initial state. Principal difference related to the ACL: Under pure internal rotation the ACL resisting moment was significant only at 0°; at peak valgus from 15° to 45° the share of internal-rotation resisting moment was not significant
Willinger et al., 2023 [1] (10 robotic knees; 10 further knees on the optical rig). Loading and motion control: Robot at fixed angles; optical rig, passive flexion–extension from 0° to 100°, for pure internal rotation and for valgus plus internal rotation Measured quantity (none included body-weight compression or recording of the clinical clunk): Robotic share of internal-rotation resisting moment (%); free displacement and angle on the optical rig (mm/°). Principal difference related to the ACL: Under pure internal rotation, resisting moment significant only at 0° and free internal rotation not significantly increased; after valgus was added, anterior translation increased significantly from 0° to 50°
Noyes et al., 2018 [2] (group 1, reconstruction in 7 knees; group 2, ACL retained in 8 knees). Loading and motion control: Six-degree-of-freedom robot, four-degree-of-freedom flexion–extension loading; simulated pivot shift sampled at 25° Measured quantity (none included body-weight compression or recording of the clinical clunk): Compartment anterior translation (mm), internal-rotation angle (°), and BPTB force (N); equivalence by TOST. Principal difference related to the ACL: After ACL sectioning, pure internal rotation barely changed, while compartment anterior translation increased markedly in the 25° simulated pivot shift; BPTB reduced the mean anterior-translation difference but did not meet equivalence

3. Anteromedial and posterolateral bundle forces under combined rotation

Clinical and teaching papers once used a mechanical split: the anteromedial bundle (AMB) of the ACL was said to restrain anterior tibial translation, and the posterolateral bundle (PLB) to control knee rotation and the pivot shift. In situ force measurements by Gabriel et al. [6] and Wu et al. [7], made with a six-degree-of-freedom robot and a universal force–moment sensor (UFS), show that “one bundle for anterior translation, one bundle for rotation” does not match the measured forces. Anatomic attachment of the two bundles is treated in Chapter 1.1. Their full tension behavior under an isolated anterior load and under muscle load is treated in Chapter 1.8.

Gabriel et al. first recorded the six-degree-of-freedom path of 10 fresh-frozen intact knees under load. They then removed the other soft tissues, leaving an isolated femur–ACL–tibia complex, separated the ligament into functional bundles, removed the femoral-insertion bone blocks in random order, replayed the original path, and calculated in situ forces [6]. The combined rotational load was a 10 N·m valgus moment plus a 5 N·m internal-rotation moment, tested only at 15° and 30° of flexion.

  • At 15°, with accompanying anterior translation of 3.7 ± 2.2 mm, the resultant in situ force of the whole ACL was 46 ± 21 N. AMB in situ force was 30 ± 15 N and PLB in situ force was 21 ± 11 N. AMB force was significantly higher than PLB force (P < .05) [6].
  • At 30°, with accompanying anterior translation of 5.7 ± 2.7 mm, the resultant in situ force of the whole ACL was 43 ± 19 N. AMB in situ force was 35 ± 16 N and PLB in situ force was 14 ± 6 N. AMB force was again significantly higher than PLB force (P < .05) [6].

At 15°, the scalar sum of the two bundle-force magnitudes (30 N and 21 N) was larger than the resultant in situ force of the whole ACL (46 N). The AMB and PLB in situ forces are three-dimensional vectors with different directions, and the sum of the magnitudes is not the magnitude of the resultant. Separately, the authors recorded a small mechanical interaction when the bundles were bluntly separated. The maximum was 7.8 ± 7.0 N during anterior loading at 60° [6].

Angular range in the original figure, the donor-age statistics, and an extension in the Discussion (pp. 85–88). The anterior-loading bar graph in Figure 2 on p. 87 of Gabriel et al. shows data and significance markers at 120° of flexion. The abstract and the Methods on p. 86 say only that testing went to 90°. How the knee was loaded at 120° is not described, so those points cannot be cited as a valid result. Pages 85–86 also give the ages of the 10 knee donors as 50 ± 10 years, range 46–66 years. If ±10 is the sample standard deviation, then with n = 10, a mean of 50, and extremes of 46 and 66, the largest sample standard deviation that is mathematically possible is about 8.43 years. The three statistics cannot all be true. The paper gives neither ages knee by knee nor a definition of ±, so any citation of donor age has to note the contradiction. The Discussion on p. 88 says that the trend stayed the same when the internal-rotation moment was changed. The paper tested only one combination, a 5 N·m internal-rotation moment plus a 10 N·m valgus moment, so that extension across other moments is not supported by this series. The measured bundle forces under combined rotation at 15° and 30° are unaffected [6].

Wu et al. used eight fresh-frozen knees with the soft tissues around the knee preserved (mean age, 55 years; range, 47–60 years; seven male and one female). Through a small medial incision they sectioned the AMB or the PLB at the femoral attachment, alternating which bundle was cut first. After each sectioning they closed the incision in layers and replayed the intact-knee path [7]. The combined rotational load was again a 10 N·m valgus moment plus a 5 N·m internal-rotation moment. The results text and Figure 4 report only 0° and 30°.

  • At 0°, AMB in situ force was 59.9 ± 27.5 N and PLB in situ force was 40.9 ± 23.7 N. Both bundles carried a high load, and the difference between them did not reach statistical significance [7].
  • At 30°, AMB in situ force was 75.5 ± 42.5 N and PLB in situ force was 35.9 ± 31.4 N. AMB force was significantly greater than PLB force (P < .05) [7].

The two studies come from different laboratories and preserved the soft tissues differently. They agree on the loading. Under valgus plus internal rotation near extension and in slight flexion, the PLB does take a substantial share, the AMB shares the load at every angle tested, and from 15° to 30° the mean in situ force in the AMB is significantly higher than the mean in situ force in the PLB [6,7]. Valgus combined with internal rotation near extension produces a clear accompanying anterior translation of the tibial plateau, such as the 3.7–5.7 mm measured by Gabriel et al. [6]. The tibial insertion of the whole ACL shifts anteromedially with that translation, and the AMB and PLB are loaded together [6,7].

Both studies have a defined range of use. Absolute bundle forces in Wu et al. were higher than those in Gabriel et al. The studies differed at the same time in donor age, in preservation of periarticular soft tissue (tissue around the knee retained and the incision closed, versus an isolated bone–ligament complex), in the residual-force tolerance of the initial passive path (5 N / 0.5 N·m versus 0.5 N / 0.2 N·m), and in how the bundles were separated. They were not cross-tested on the same specimens. Any of those protocol differences may change the magnitude of the measured bundle force. Both papers measured force by replaying the intact-knee path. Neither measured free kinematic laxity after a single bundle was cut, and neither studied graft reconstruction. The forces correct the one-sided view that “the PLB alone governs rotation.” They are not evidence that anatomic double-bundle reconstruction is superior to anatomic single-bundle reconstruction for clinical rotational stability or for longer-term outcome [6,7].

4. Anterolateral soft tissue and the lateral meniscus after loss of the ACL

Near extension, the ACL limits anterior translation of the lateral tibial plateau and shares the internal-rotation moment at 0°. After the ACL ruptures, loading of the anterolateral soft tissues and of the lateral meniscus changes. Injury to those structures, if it is also present, directly affects rotational laxity and anterior laxity. Attachments and histologic relations of the anterolateral complex are described in Chapter 1.7. The concern here is their mechanical cooperation with the ACL, and the disputes over how the experiments were done.

Load transfer after ACL deficiency, and dependence on sectioning order

Kittl et al. compared eight ACL-intact knees with eight ACL-deficient knees. Loss of the ACL changes how the internal-rotation resisting moment is distributed at extension [4]. Under a 5 N·m pure internal-rotation moment at 0° (Table 2 of that paper), the ACL-intact group assigned 4.3% ± 3.9% of the internal-rotation resisting moment to the deep and capsulo-osseous ITB. That dcITT includes the femoral supracondylar attachment, the attachment beside the intermuscular septum, and the retrograde attachment, which is the Kaplan-fiber and capsulo-osseous system. The 4.3% ± 3.9% share was not statistically significant. The ACL was the only significant restraint (19.7% ± 7.9%). In the ACL-deficient group, the dcITT share at 0° rose to 18.3% ± 11.4% (P = .027; significantly higher than in the ACL-intact group, P < .01). At 0° it was the only structure resisting internal rotation that reached statistical significance on its own [4].

At 30°, 60°, and 90°, the ITB system carried the main internal-rotation resisting moment whether or not the ACL was intact. The system is superficial sITT together with deep dcITT. The superficial resection included the adjacent aponeurotic expansion of the biceps and the lateral retinaculum, so it is not an isolated narrow band. Under 5 N·m pure internal rotation in the ACL-deficient knee, the combined share was 58% ± 10% at 30°, 80% ± 13% at 60°, and 76% ± 14% at 90°. Under the simulated pivot shift at 45° (8 N·m valgus plus 4 N·m internal rotation), the share was 72% ± 14% [4]. The extracapsular ALL, which lies superficial to the lateral collateral ligament, accounted for 2.3%–4.9% under pure internal rotation at each angle in the ACL-deficient knee (4.4%, 4.9%, 2.3%, and 3.3% at 0°, 30°, 60°, and 90°). Under the simulated pivot shift from 15° to 45° it accounted for 2.9%–6.9% [4].

The optical kinematic experiment of Willinger et al. shows cooperative restraint and dependence on sectioning order in the increment of free laxity [1]. After the ACL had been sectioned, 10 knees were assigned to three orders of secondary-structure sectioning, with 3 or 4 knees in each subgroup (Figures 6 and 7 of that paper). The subgroups are too small for a quantitative ranking. The three sets of curves still show the same order effect. Under the quasi-static simulated pivot shift, if either the KFs or the C/ALL was still intact, cutting the other produced a smaller increase in laxity. The largest increases in anterior translation and in internal rotation occurred when the KFs (Figures 6B and 7B) or the C/ALL (Figures 6C and 7C) was cut as the last anterolateral restraint [1]. Parallel soft tissues share the load. While one set of anterolateral fibers crossing the joint remains intact, displacement of the lateral tibial plateau is still limited to some degree. After the ACL and the parallel superficial and deep anterolateral structures have all been cut, lateral anterior translation and internal-rotation laxity in this model increase markedly [1].

Where the dispute over the mechanical contribution of the ALL comes from

Papers have drawn two conclusions about the ALL that look opposed: that it is a primary restraint to anterolateral rotational stability, and that it is a secondary one. Parsons et al., in a 2015 robotic study, reported that under 5 N·m of internal rotation the in situ force contribution of the ALL above 35° of flexion reached 30%–45% and exceeded that of the ACL. The authors later published a response stating that the superficial ITB had been reflected before testing and that the incision included part of the deep lateral capsule [8,9]. Hohmann et al., in a 2025 systematic review, stated that time-zero cadaveric studies have not provided conclusive evidence that the ALL significantly limits anterior tibial translation or internal tibial rotation [10].

Parsons et al. dissected 12 fresh-frozen cadaveric knees and excluded one in which the ALL was not identified (pp. 670, 672, and 673). They tested 11 knees from eight cadavers (mean age, 76.3 years; range, 35–92 years). On a six-degree-of-freedom robot they sectioned the ACL, the lateral collateral ligament (LCL), and the ALL in randomized-block order, replayed the baseline path, and divided the magnitude of each structure’s in situ force vector by the magnitude of the intact force vector in the initial prepared state (p. 671). Page 672 states that the structural percentages do not sum to 100% [8]. Under a 5 N·m pure internal-rotation moment (Table 4 of that paper), the ACL in situ force share fell with flexion from 36% at 0°, 35% at 15°, and 30% at 25° to 22%–15% from 35° to 90°. The lowest share was 15% at 60°, and the share at 90° was 18%. The “ALL” share rose from 5% at 0° to 30% at 35°, 37% at 45°, 44% at 60°, 45% at 75°, and 39% at 90°. Under a 134 N anterior drawer, the ACL accounted for 49%–60% and the ALL for 1%–4% [8].

Preparation, the extent of the cut, and the normalization all differ from Kittl et al. [4] and Willinger et al. [1], and those differences affect comparison across studies. The Methods of Parsons et al. on pp. 670–671, and the authors’ response on p. NP22, state the conditions. Before testing, the investigators reflected the superficial ITB from proximal down to Gerdy’s tubercle and explicitly left the deep capsulo-osseous layer intact. They also left the medial parapatellar capsular incision open. When the ALL was cut at the level of the lateral tibial plateau, part of the closely applied deep lateral capsule was opened with it, the cut passing below the lateral meniscus. The contribution ratio was a ratio of force-vector magnitudes, not the proportional drop in the internal-rotation resisting moment [8,9]. The 30%–45% figures are therefore ratios of in situ force-vector magnitude for the ALL together with part of the deep lateral capsule, in a knee whose superficial ITB had been reflected and whose deep capsulo-osseous layer had been kept. They are not the share of internal-rotation resisting moment of an isolated ALL in a knee with an intact superficial ITB [8,9].

Error in the original identification rate (p. 672). The first sentence of the Results on p. 672 says that the ALL was identified in “83%” of specimens. That conflicts with the Methods on p. 670, Table 2 on p. 671, the Discussion on p. 673, and the legends of Figures 3 and 4, which agree on the count: 12 knees were dissected, 1 was excluded because the ALL was not seen, and the ALL was identified and tested in 11. Eleven of 12 is about 91.7%. Ten of 12 is 83.3% [8]. The figure “83%” is not used. The mechanical measurements from the 11 knees can still be used for methodologic comparison under that preparation: superficial ITB reflected, deep capsulo-osseous layer retained, and part of the deep lateral capsule opened with the ALL [8,9].

By contrast, the 2025 systematic review by Hohmann et al. included 22 cadaveric biomechanical studies and reached a negative conclusion. The Methods on p. 4883 excluded studies of the ITB, the Kaplan fibers, and lateral extra-articular tenodesis. Pages 4884 and 4896 state that no overall pooled kinematic effect size was calculated [10]. The title is written as “ACL-deficient knees and ACL-reconstructed knees.” The abstract and the conclusion in the text are narrower. They say that time-zero cadaveric testing did not show conclusive evidence that the ALL significantly limits anterior translation or internal rotation in ACL-intact knees and ACL-reconstructed knees. That claim does not refute cooperative instability when ACL deficiency is combined with anterolateral injury. It also does not represent the ITB and Kaplan-fiber system, which the review excluded [10].

Serious errors in transcription, group assignment, and citation matching in the systematic review (pp. 4886, 4897–4898, and 4899.e1). Hohmann et al. made five serious errors in arranging data from the key primary experiments. (1) Table 2 on p. 4886 prints the 11 knees actually tested by Parsons et al. as 41, drops a character and prints “45°: 37%” as “4°: 37%,” and places the in situ force share, measured by replaying the path of the initial intact prepared state, in the “ACL Deficient” column. (2) On the same page, Table 2 records 25.6% ± 9.0% for the dcITT at 30° of flexion in the ACL-intact knee, from Table 2 of Kittl et al. 2016, as an ALL contribution of “30°: 26%.” In the original table the ALL was 11.0% ± 9.7%. Reference 7, as marked in the review table, is the 2018 review by Kittl et al. in the reference list on p. 4898, not the 2016 primary experiment. (3) The Discussion on p. 4897 and the appendix on p. 4899.e1 rewrite 35° of flexion in Parsons et al. (30%) as 30°, assign the ALL shares in the ACL-deficient knees of Kittl et al. under pure internal rotation at 0°, 30°, 60°, and 90° (2.3%–4.9%, summarized as 2%–5%) to the simulated pivot-shift angles (15°–45°), and omit that Kittl et al. did test 5 N·m of pure internal rotation [4,8,10]. Quantitative restatement of those studies in the review, and its attribution of their protocols, cannot be used. The methodologic limits the review does identify can still be consulted: time-zero cadaveric experiments generally have small samples, the donors are elderly, and the loading protocols differ widely [10].

The primary experiments show that the anterolateral soft tissues are not a single narrow band [1,4,8,9]. When the superficial and deep ITB and its distal femoral attachments remain intact, the main extra-articular restraint to internal rotation above 30° of flexion comes from the ITB–Kaplan fiber–capsulo-osseous complex. The ALL and the anterolateral capsule are parallel secondary structures and share that restraint. When ACL deficiency is combined with injury of the superficial and deep anterolateral restraints, anterior translation and internal-rotation laxity in the cadaveric model are both greater than after isolated ACL sectioning [1,4].

The lateral meniscus posterior root with the ACL intact and with the ACL deficient

The intra-articular lateral meniscus also affects anterolateral rotational stability. Willinger et al. simulated complete loss of function of the LMPR by transecting the posterior lateral meniscus lateral to the attachment of the meniscofemoral ligament. That cut also removes the suspensory pull of the meniscofemoral ligament on the posterior horn. It is not a central-to-medial partial tear with the meniscofemoral ligament intact, and it is not every clinical type of posterior-root tear [1].

The LMPR findings differ between robotic path replay and free loading on the optical rig. Its effect depends on whether the ACL is intact and on the direction of the load [1].

  • ACL intact. On replay of the baseline path, the LMPR contribution to the moment resisting 5 N·m of pure internal rotation was not statistically significant. The study did not section the LMPR alone in the ACL-intact knee to measure the change in free internal tibial rotation. At every tested angle from 0° to 90°, the LMPR did contribute significantly to the moment resisting valgus (P < .001). In the latter five specimens, the superficial medial collateral ligament (sMCL) was also sectioned. In those specimens the sMCL was the primary restraint to valgus, and the LMPR had a significant secondary role against valgus. From 0° to 90° the LMPR also significantly limited posterior tibial translation. Its highest share was 12% at 90° (P < .01) [1]. A valgus moment compresses the lateral tibiofemoral compartment. With the posterior root intact, the meniscus can turn vertical and valgus compression into circumferential fiber tension, or hoop stress, and resist collapse of the lateral joint space and tilt into valgus.
  • ACL deficient, under a simulated pivot-shift load. Once the primary anterior restraint of the ACL is lost, the valgus moment compresses the lateral compartment and pushes the convex lateral femoral condyle across a lateral plateau that no longer has the circumferential enclosure of the meniscus. Meniscal extrusion and loss of hoop tension then make anterior and lateral glide of the lateral plateau easier. Descriptive curves from the three secondary sectioning orders of Willinger et al. show that incising the LMPR in an ACL-sectioned knee adds to injury of the KFs and the C/ALL and further increases anterior translation and internal rotation under the simulated pivot shift [1]. Naendrup et al. also combined ACL sectioning with lateral meniscectomy when they prepared the high-grade and low-grade laxity specimens [3].

5. Bedside tests and residual instability after reconstruction

Anterior-laxity tests used by orthopedic and sports-medicine surgeons, the Lachman test and the anterior-drawer test at 90°, assessment of isolated axial internal rotation, and the pivot-shift test each apply a different external load and call on different primary passive restraints (Table 2).

Table 2. Load, principal restraints, and corresponding findings in three bedside tests of knee stability

Test, load, and observation Restraints and corresponding findings
Anterior-laxity tests (Lachman, anterior drawer at 90°). Principal load and what is observed: Sagittal anterior shear; anterior tibial translation and the endpoint Principal passive restraints: From 0° to 90°, the ACL is the primary restraint, with the AMB and PLB sharing the load [1,4,6,7]; at large displacement, secondary structures share the load. Cadaveric findings and the clinical counterpart: An isolated complete ACL rupture is enough to increase anterior translation and soften the endpoint (see Chapter 1.5)
Assessment of isolated axial internal rotation. Principal load and what is observed: A pure internal-rotation moment in the transverse plane; the internal tibial rotation angle Principal passive restraints: At 0°, the ACL resisting moment is significant; from 30° to 90°, the superficial and deep ITB and the KFs carry most of it, with the C/ALL sharing the restraint [1,4]. Cadaveric findings and the clinical counterpart: Isolated ACL sectioning often produces no significant increase in pure internal rotation in flexion [1,2]; the angle increases after the anterolateral structures are sectioned as well [1,2,4]
Pivot-shift test. Principal load and what is observed: Internal rotation plus valgus near extension produces lateral subluxation; after the force is released during flexion, reduction is posterior translation with external rotation [3] Principal passive restraints: Near extension, the ACL limits anterior translation and internal rotation at 0° [1,2,4,5], and the LMPR resists valgus [1]; in flexion, the KFs/ITB and the C/ALL restrain the motion [1,4]. Cadaveric findings and the clinical counterpart: An isolated ACL tear can be positive; in the awake patient, specificity is high and sensitivity is moderate [11]; sectioning the lateral structures and the LMPR as well increases instability in the model [1,3,4]

Positive and negative awake pivot-shift tests do not carry the same diagnostic weight

The pivot-shift test is close to the giving-way patients describe when they stop suddenly and cut. Producing it by hand requires the muscles of the limb to relax, so that the lateral plateau can subluxate anteriorly near extension and then reduce, during continued flexion, by posterior translation with external rotation [3]. Sokal et al., in a 2022 systematic review and bivariate random-effects meta-analysis, examined clinical diagnosis of partial or complete ACL tears in awake, unanesthetized patients. They included 24 studies that excluded combined injury of other knee ligaments. Arthroscopy or open exploration was the gold standard in 17, and MRI was the reference standard in 7. On QUADAS-2, 2 of the 24 studies were rated as having no risk of bias, and 12 entered the bivariate model. For the pivot-shift test in the awake state, pooled specificity was 94% (95% CI, 88%–97%) and pooled sensitivity was 55% (95% CI, 47%–62%). The positive likelihood ratio was 10.70 and the negative likelihood ratio was 0.48 [11].

In awake outpatient examination, a clearly positive pivot-shift test supports the diagnosis of an ACL tear. A negative pivot-shift test cannot be used to exclude an ACL tear [11]. In the acute phase, pain and effusion provoke protective co-contraction of the hamstrings and quadriceps, and that co-contraction actively limits tibial motion. If the examiner’s hands leave the rhythm of applying and then releasing multiplanar stress, the reduction can be masked as well [3,11]. A pivot-shift result in an awake outpatient cannot be set directly against a pivot-shift examination in the operating room under muscle relaxation from general or spinal anesthesia. When the pivot shift is negative in an awake patient, but the history is a typical cutting sprain with joint swelling or giving way, the Lachman test and MRI should be used to check the knee further [11].

Marked rotational instability and injuries that may coexist

In sequential cadaveric sectioning, isolated ACL sectioning under quasi-static valgus plus internal rotation mainly increases anterior translation near extension and at low and moderate flexion. Sectioning the anterolateral soft tissues and the LMPR as well significantly increases anterior translation and internal rotation over a wider flexion–extension range [1,4]. An increment in cadaveric laxity does not match clinical pivot-shift grade one to one. It is not a rule for assigning the laxity to one tissue, and it cannot be used to derive an indication for surgery or for arthroscopy. When marked rotational laxity or a pronounced pivot shift is found, the injury history, the rest of the physical examination, and MRI or other imaging can be used to look for two categories of potentially coexisting injury.

  • Anterolateral complex injury: edema or loss of continuity in the posterolateral Kaplan fibers at the distal femur, the deep capsulo-osseous ITB, and the anterolateral capsule/ALL. Anatomy and imaging of these structures are described in Chapter 1.7 [1,4].
  • Injury of the lateral meniscus posterior root and the posterior meniscal body: tear or substantial loss at the posterior-root attachment, at the meniscofemoral ligament connection, and in the posterior portion of the meniscal body. If the patient already has an independent indication for surgery and arthroscopy is already under way, the integrity of these structures can be inspected during the same procedure [1,3].

Residual rotational instability after reconstruction

Some patients, after intra-articular single-bundle ACL reconstruction, have less passive anterior laxity but still have a positive pivot shift or rotational instability during activity. The cadaveric work that tested a reconstructed state, Noyes et al. [2], and the sequential-sectioning studies [1,4,6,7] have to be kept apart in two senses: a result observed directly at time zero, and a mechanism inferred from the geometry of loading.

In group 1 of Noyes et al., a 10 mm single-bundle BPTB graft in the native central footprint carried a high in situ force under the simulated pivot shift sampled at 25°, and it reduced the mean lateral-compartment anterior-translation difference after ACL sectioning. It did not meet the TOST equivalence criterion set in that paper. Further division of the ALL and the capsulo-osseous ITB changed lateral anterior translation only slightly in the 25° simulated pivot shift and produced a small increase in isolated internal tibial rotation at moderate and deep flexion. Manual-tension tenodesis of the distal ITB, added after that, further decreased anterior translation and shared the load on the graft, but from 25° to 90° it markedly overconstrained isolated internal tibial rotation [2]. The study tested only a static distal ITB attachment. It did not simulate dynamic function of the proximal ITB, and it did not use the time course of the clinical clunk as an outcome. The time-zero result cannot be extended to “every intra-articular single-bundle reconstruction is sufficient to eliminate the pivot shift in living patients,” or to “modern low-tension combined anterolateral procedures have no clinical value” [2].

Residual rotational instability after surgery in a living patient may come from two mechanisms.

The first is a change in the position and tension of the intra-articular graft. This is an inference from loading geometry and from the evidence that the two bundles share load. Noyes et al. did not include a nonanatomic-tunnel control. In their model, the BPTB graft limited anterior translation of the lateral plateau at 25° under three conditions at once: the graft was in the native central footprint, the initial tension was still present at the moment of testing, and the graft carried a high in situ force [2]. Gabriel et al. [6] and Wu et al. [7] found that the AMB and the PLB jointly resist anterior translation under combined rotation. If the femoral or tibial tunnel is off the anatomic attachment, for example a femoral tunnel placed high and anterior so that the graft runs more vertically, or if the graft elongates and loses tension while it remodels, the intra-articular restraint to anterior subluxation near extension may decrease.

The second is extra-articular soft-tissue injury or lateral-meniscus injury that is present and has not been treated. The Noyes model kept an intact lateral meniscus. It did not simulate active tension in the proximal ITB or repeated shear-cycle loading in vivo [2]. The sectioning results of Willinger et al. [1] and Kittl et al. [4] in ACL-deficient knees are the reference. If a patient also has a severe injury of the Kaplan fibers or the deep ITB, an anterolateral capsular injury, or an avulsion of the LMPR, and that injury is left untreated, the lateral compartment during flexion lacks parallel restraint against internal rotation and lacks hoop support against valgus. The combined load on the intra-articular graft may increase. Internal-rotation laxity or a residual pivot shift may remain at moderate and deep flexion [1,2,4]. These cadaveric studies show how the structures share load under the loads that were applied. They do not include biological remodeling of a graft, neuromuscular control, or later cartilage outcome. Comparison of clinical results between anatomic single-bundle and double-bundle reconstruction, comparison among different combined anterolateral procedures, the specific indications for surgery, and the parameters of postoperative rehabilitation still have to be decided by controlled clinical studies and by follow-up [1,2,5-7].

Research funding and conflicts of interest. Among the main primary biomechanical studies cited in this chapter, Willinger et al. [1] received support from the German Society for Arthroscopy and Joint Surgery, the DFG, and Smith & Nephew funding paid to a research account at Imperial College London. Noyes et al. [2] received institutional research support from Arthrex. Naendrup et al. [3] received support from the Wallace H. Coulter Foundation, and the corresponding author disclosed a consultancy with Smith & Nephew. Kittl et al. [4] received a Smith & Nephew research fellowship and support from the German Society for Arthroscopy and Joint Surgery. Markolf et al. [5] received support from NIH R01 AR40330, the OREF, and other sources. Gabriel et al. [6] received support from NIH AR39683 and from the Musculoskeletal Research Center at the University of Pittsburgh. Wu et al. [7] received support from NIH AR055612 and AR051078 and declared no conflict of interest.

References

  1. [1]

    Willinger L, Athwal KK, Holthof S, Imhoff AB, Williams A, Amis AA. Role of the Anterior Cruciate Ligament, Anterolateral Complex, and Lateral Meniscus Posterior Root in Anterolateral Rotatory Knee Instability: A Biomechanical Study. Am J Sports Med. 2023;51(5):1136–1145. PMID: 36917838. DOI: https://doi.org/10.1177/03635465231161071

  2. [2]

    Noyes FR, Huser LE, West J, Jurgensmeier D, Walsh J, Levy MS. Two Different Knee Rotational Instabilities Occur With Anterior Cruciate Ligament and Anterolateral Ligament Injuries: A Robotic Study on Anterior Cruciate Ligament and Extra-articular Reconstructions in Restoring Rotational Stability. Arthroscopy. 2018;34(9):2683–2695. PMID: 30173809. DOI: https://doi.org/10.1016/j.arthro.2018.04.023

  3. [3]

    Naendrup JH, Zlotnicki JP, Murphy CI, Patel NK, Debski RE, Musahl V. Influence of knee position and examiner-induced motion on the kinematics of the pivot shift. J Exp Orthop. 2019;6(1):11. PMID: 30888526. DOI: https://doi.org/10.1186/s40634-019-0183-7

  4. [4]

    Kittl C, El-Daou H, Athwal KK, Gupte CM, Weiler A, Williams A, Amis AA. 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(2):345–354. PMID: 26657572. DOI: https://doi.org/10.1177/0363546515614312

  5. [5]

    Markolf KL, Burchfield DM, Shapiro MM, Shepard MF, Finerman GAM, Slauterbeck JL. Combined Knee Loading States that Generate High Anterior Cruciate Ligament Forces. J Orthop Res. 1995;13(6):930–935. PMID: 8544031. DOI: https://doi.org/10.1002/jor.1100130618

  6. [6]

    Gabriel MT, Wong EK, Woo SLY, Yagi M, Debski RE. Distribution of in situ forces in the anterior cruciate ligament in response to rotatory loads. J Orthop Res. 2004;22(1):85–89. PMID: 14656664. DOI: https://doi.org/10.1016/S0736-0266(03)00133-5

  7. [7]

    Wu JL, Seon JK, Gadikota HR, Hosseini A, Sutton KM, Gill TJ, Li G. In Situ Forces in the Anteromedial and Posterolateral Bundles of the Anterior Cruciate Ligament Under Simulated Functional Loading Conditions. Am J Sports Med. 2010;38(3):558–563. PMID: 20044500. DOI: https://doi.org/10.1177/0363546509350110

  8. [8]

    Parsons EM, Gee AO, Spiekerman C, Cavanagh PR. The Biomechanical Function of the Anterolateral Ligament of the Knee. Am J Sports Med. 2015;43(3):669–674. PMID: 25556221. DOI: https://doi.org/10.1177/0363546514562751

  9. [9]

    Parsons EM, Gee AO, Spiekerman C, Cavanagh PR. The Biomechanical Function of the Anterolateral Ligament of the Knee: Response. Am J Sports Med. 2015;43(8):NP22. PMID: 26232458. DOI: https://doi.org/10.1177/0363546515597218

  10. [10]

    Hohmann E, Keough N, Molepo M, Arciero R, Imhoff A. The Anterolateral Ligament Complex Has Limited Impact on Anterior Tibial Translation or Internal Rotational Stability in Anterior Cruciate Ligament Deficient and Anterior Cruciate Ligament Reconstructed Knees: A Systematic Review of Biomechanical Cadaver Studies. Arthroscopy. 2025;41(11):4882–4899.e15. PMID: 40306468. DOI: https://doi.org/10.1016/j.arthro.2025.04.043

  11. [11]

    Sokal PA, Norris R, Maddox TW, Oldershaw RA. The diagnostic accuracy of clinical tests for anterior cruciate ligament tears are comparable but the Lachman test has been previously overestimated: a systematic review and meta-analysis. Knee Surg Sports Traumatol Arthrosc. 2022;30(10):3287–3303. PMID: 35150292. DOI: https://doi.org/10.1007/s00167-022-06898-4