Medial collateral ligament reconstruction graft isometry is effected by femoral position more than tibial position

The three principal structural elements of the “medial ligament complex (MCL)” have been described as the superficial MCL (sMCL), the deep MCL (dMCL) and the posterior oblique ligament (POL) [16, 37, 45].

There is a wide consensus that the majority of isolated grade I and II MCL injuries heal with rehabilitation alone [12, 18, 20]. Most patients return to sports at 3 months and there are excellent long-term patient-reported outcomes [29]. Although grade III injuries also may heal without surgery, some patients remain symptomatic following conservative treatment, necessitating reconstruction [20]. Combined sMCL and posterior oblique ligament (POL) injuries may be associated with an increased incidence of failure to respond to non-operative treatment and may result in persistent MCL instability [31].

Residual peripheral ligament laxity is an important cause of ACL graft laxity and failure [41]. The ACL functions in conjunction with the medial ligament complex to prevent anteromedial instability and is a secondary restraint to valgus rotation [48]. Thus, if the medial structures remain compromised, the ACL graft may be exposed to increased loads, potentially leading to graft failure [4, 5]. Similarly, although combined high-grade MCL and PCL injuries are rarer [19], because, particularly the POL has a role in restraining both posterior translation and internal rotation in extension [45], the need for combined medial and PCL reconstruction has been suggested [6, 18, 26, 35, 40].

Medial reconstructions are challenging as it is difficult to reproduce the biomechanical behaviour of flat, sheet-like structures [30, 39]. It is a widely accepted principle that isometric ligament reconstruction may be advantageous in reducing the likelihood of unwanted graft behaviour [1]. Inappropriate sMCL graft positions could result in abnormal graft tensioning patterns leading to either persistent laxity or over-constraint [2]. However, graft positioning is complicated by controversy about the location of the sMCL femoral attachment. Historically, this has been described as being on the medial epicondyle (ME) [8, 17, 24, 45], whereas more recently, LaPrade et al. [23] reported that the sMCL attaches to a depression 5 mm posterior and 3 mm proximal to the ME. In contrast, Liu et al. [28] have described the fibres enveloping the ME. These different interpretations of the anatomy may relate to the fact that there is a confluence of fibres in the region of the ME that makes it difficult to identify a precise attachment site.

Conversely, the native POL is known to be anisometric [41] restraining valgus, internal tibial rotation and posterior tibial translation near extension and slackening as the knee flexes. Little is known about the precise length change pattern of its fibres. It has a linear attachment extending from just posterior to the attachment of the longitudinal parallel fibres of the sMCL around the base of the adductor tubercle in a posterior and then slightly proximal direction. Reconstruction of the central arm of the POL has been suggested [9] but the length changes of a graft in this position are ill defined.

The goal of this study was to (1) examine the length change patterns of the native medial structures of the knee and (2) determine the effect on graft length change patterns for different tibial and femoral attachment points for previously described medial reconstructions. The aim was to recommend optimal femoral and tibial attachment positions for medial reconstructions that closely reproduce native ligament length change behaviour.

Eight fresh–frozen cadaveric knee specimens were obtained from the local tissue bank. The knee specimens were dissected and tested with the necessary permissions from the “Gesetz über das Leichen-, Bestattungs und Friedhofswesen (Bestattungsgesetz) des Landes Schleswig–Holstein vom 04.02.2005, Abschnitt II, § 9 (Leichenöffnung, anatomisch)”.

The most important findings of this study were that small changes in the femoral attachment site of medial reconstructions resulted in significant changes in fibre length change behaviour and that utilising the tibial insertion of the semitendinosus tendon (as described for a Bosworth/Lind reconstruction) had no significant effect on fibre length change pattern.

The length change patterns of the native MCL strongly depended on its fibre region. The anterior part of the MCL was tight in flexion, whereas the posterior part of the MCL was tight in extension (p < 0.001), implying a reciprocal tensioning pattern of these fibre regions. This reciprocal behaviour of the MCL was first observed by Brantigan and Voshell [7] and later confirmed by several other works [3, 13, 21, 33, 46]. Recently, Willinger et al. [50] looked at the length change pattern of the anterior and posterior borders of the sMCL of 10 cadaveric knee specimen using a kinematics rig. The anterior region of the MCL showed a lengthening of 6% at 0–100° knee flexion and the posterior portion presented a 6% decrease of length at 100° knee flexion. This is similar to the present study’s lengthening of 4.1 ± 3.7% until 100° of flexion and a decrease of length for the posterior border of 5.4 ± 1.9% at 100° knee flexion. However, their increase and decrease of length was constant, whereas the present study showed an initial slackening of 2–3% (10–20° knee flexion) for all measured tibiofemoral MCL combinations. This initial slackening may be due to the external rotation of the tibia (screw home mechanism) in full extension, which will increase the length of the medial structures.

Several other studies [10, 11, 27, 43, 44, 47] found a constant decrease in length [11, 27] or a near-isometric behaviour [43] for the MCL, which was different to the present study. This is most likely due to differences in the testing setup, muscle loading, and dissection methods. Feeley et al. [11] assessed length changes of a modified Bosworth MCL reconstruction technique similar to the Lind reconstruction. They found a small, but significant difference in overall length change (2.7 ± 1.2 mm vs. 4.1 ± 2.3 mm) when the tibial attachment of the reconstruction was changed from the position of the proximal part of the sMCL tibial attachment site to the insertion to the semitendinosus The present study, however, did not find a significant difference, neither in TSR (5.4 ± 2.1 vs. 5.6 ± 1.5; n.s.), nor in overall length change pattern (n.s.).

It is known from previous studies that there is no perfect isometric tibiofemoral combination [38]. The MCL is an approximately 10-mm wide, flat structures with different fibre regions tensioning throughout flexion. Reconstructions of the medial side of the knee, however, cannot precisely imitate this complex tensioning pattern [49] and small changes in femoral graft position can result into different graft length change patterns. Moving the attachment point from the posterior edge of the ME to anterior will result in graft tightening rather than slackening in flexion. When performing medial reconstructions, surgeons should be aware of this effect when positioning the femoral tunnel to avoid unwanted graft behaviour as inappropriate graft tension can also induce over-constraint or ‘stretching out’ of the graft. If in doubt, an intra-operative check of the isometry should be performed.

The present study showed that there is an increase in the distance between femoral and tibial attachments of the POL of almost 30% as the knee flexes. The findings confirm that a POL graft should be tensioned and fixed with then knee in extension. Tensioning and fixation at higher knee flexion angles would lead to loss of motion and possibly graft failure due to rapid graft elongation likely to overcome the strain required for ligament failure [34].

This study had some limitations in addition those inherent in cadaveric knee biomechanical testing, including the age of specimens and number of knees tested. There are inconsistencies in the literature regarding the exact femoral attachment site of the sMCL. Some authors describe it posterior and proximal to the ME [11, 23], while others describe the sMCL as enveloping the ME [28, 36]. In this study, the knee was carefully dissected and the proximal attachment of the sMCL was found to envelope the ME. The anterior-most sMCL fibres attached to the anterior border of the ME, the posterior most fibres attached to the posterior border and the middle fibres to the superior border. However, it was also noted that in flexion the fibres of the sMCL appeared to bend posteriorly towards their femoral attachment as they coursed proximally from the tibia. This complex fibre pattern has been noted by the other authors [13, 32] and cannot be reproduced with sutures, which have a completely linear course, possibly leading to subtle length change differences for the native sMCL. However, the linear fibre course is representative of sMCL reconstruction.

In addition, fibre length change pattern in response to tibia-femoral rotatory loads (e.g. intern/external rotation, valgus rotation) were not investigated. Knees were also tested in an open chain extension rig against gravity, which only involved one muscle loading state. Different muscle loading states or involvement of the hamstring muscles were not considered. Finally, the present study setup only allowed for the measurement of length changes and not the measurement of actual tensile strain. This could be important further information for helping to determine the optimum angle of knee flexion angle for graft fixation and the amount of pre-tensioning required [25].

The complex behaviour of the native MCL could not be imitated by a single point-to-point combination and surgeons should be aware that small changes in the femoral MCL graft attachment position will significantly effect graft length change patterns. Reconstructing the sMCL with a semitendinosus autograft, left attached distally to its tibial insertion, would appear to have a minimal effect on length change compared to detaching it and using the native tibial attachment site, perhaps obviating the need for tibial fixation. A POL graft should always be tensioned near extension to avoid capturing the knee or graft failure.