The meniscofemoral ligaments: secondary restraints to the posterior drawer

The meniscofemoral ligaments: secondary restraints to the posterior drawer ANALYSIS OF ANTEROPOSTERIOR AND ROTARY LAXITY IN THE INTACT AND POSTERIOR-CRUCIATE-DEFICIENT KNEE Chinmay M. Gupte, Anthony M. J. Bull, Rhidian D. Thomas, Andrew A. Amis From Imperial College, London, England W e have tested the hypothesis that the meniscofemoral ligaments make a significant contribution to resisting anteroposterior and rotatory laxity of the posterior-cruciate-ligament-deficient knee. Eight cadaver human knees were tested for anteroposterior and rotatory laxity in a materialstesting machine. The posterior cruciate ligament () was then divided, followed by division of the meniscofemoral ligaments (MFLs). Laxity results were obtained for intact, -deficient, and -MFLdeficient knees. Division of the MFLs in the -deficient knee increased posterior laxity between 15 and 90 of flexion. Force-displacement measurements showed that the MFLs contributed 28% to the total force resisting posterior drawer at 90 of flexion in the intact knee, and 70.1% in the -deficient knee. There was no effect on rotatory laxity. This is the first study which shows a function for the MFLs as secondary restraints to posterior tibial translation. The integrity of these structures should be assessed during both imaging and arthroscopic studies of -injured knees since this may affect the diagnosis and management of such injuries. J Bone Joint Surg [Br] 2003;85-B:765-73. Received 19 August 2002; Accepted after revision 20 February 2003 In addition to their primary attachments to the tibia, the menisci may be connected to the femur by the meniscofem- C. M. Gupte, MRCS, Research Fellow Departments of Mechanical Engineering and Bioengineering and Musculoskeletal Surgery A. M. J. Bull, PhD, Lecturer Department of Bioengineering R. D. Thomas, FRCS Ed (Orth), Consultant Orthopaedic Surgeon Department of Musculoskeletal Surgery A. A. Amis, DSc, Professor of Orthopaedic Biomechanics Departments of Mechanical Engineering and Musculoskeletal Surgery, Imperial College, Exhibition Road, London SW7 2BX, UK. Correspondence should be sent to Dr C. M. Gupte. 2003 British Editorial Society of Bone and Joint Surgery doi:10.1302/0301-620x.85b5.13771 $2.00 oral ligaments. Most commonly, the posterior horn of the lateral meniscus is connected to the intercondylar area of the femur by a meniscofemoral ligament. In the human knee (Fig. 1), two such ligaments may be present. The anterior meniscofemoral ligament (amfl) of Humphry passes anterior to the posterior cruciate ligament (), attaching to the lateral aspect of the medial femoral condyle close to the distal articular cartilage. 1 Similarly, the posterior meniscofemoral ligament (pmfl) of Wrisberg passes posterior to the and attaches proximally, close to the roof of the intercondylar notch. Less commonly, other connections between the menisci and the femur have been described. 2,3 Although the MFLs were described more than 100 years ago, 4 their function remains undetermined. Radoïévitch, 5 in 1932, suggested that they acted as a third cruciate ligament. This implied a role in controlling anteroposterior (AP) laxity. Clancy et al 6 have provided the only evidence which suggests a clinical function of the MFL. In patients with an acute rupture of the, they noted that the posterior drawer with the tibia in internal rotation was reduced if the MFLs were intact. These observations, however, lacked objective assessment of laxity of the knee. Ritchie et al 7 performed a sequential cutting study showing that sectioning of the MFLs in the intact knee increased posterior translation from a mean of 3.11 mm to 3.46 mm in response to a posterior drawer force of 88 N. However, the assessment was performed at only one angle of flexion (90 ) and it is unclear whether this increase was statistically significant since no post-hoc tests were used after analysis of variance between the different testing conditions. Several cadaver studies 8-10 have examined the effect of division of the on posterior laxity of the knee. 11 All have involved division of both the MFLs and the, almost as if the MFLs were part of the itself. Thus far there have been no studies involving testing laxity after selective division of the MFLs in the -deficient knee. The management of injuries of the remains controversial. While some authors 12,13 agree that an isolated rupture of the is best treated conservatively, it is unclear which associated ligamentous ruptures require operative management. 14 It is likely that additional structures in the knee, which may include the MFLs, provide a synergistic reinforcement to the function of the. 15 VOL. 85-B, No. 5, JULY 2003 765

766 CHINMAY M. GUPTE, ANTHONY M. J. BULL, RHIDIAN D. THOMAS, ANDREW A. AMIS amfl Fig. 1 Photograph of a left tibial plateau viewed from its posterolateral aspect. The anterior and posterior meniscofemoral ligaments (amfl and pmfl, respectively) attach to the posterior horn of the lateral meniscus (LM) and pass either side of the posterior cruciate ligament (). LM pmfl We performed a cadaver study to test the hypothesis that the MFLs make a significant contribution to limiting the AP and rotatory laxity of the -deficient knee. A technique was devised to identify accurately the MFLs and to allow selective division of the without compromising the integrity of the MFLs. Evidence of a stabilising role for the MFLs would have a bearing on the diagnosis and management of injuries to the. Materials and Methods Specimens. Ethical permission for the study was obtained and knees were harvested from donors within 48 hours of death, the femur and tibia being divided approximately 8 cm from the joint line. These were fresh-frozen in sealed polyethylene bags at -20 C and defrosted before testing. During testing, the joints were covered with wet tissue to prevent dehydration. A total of 34 knees was used for the various studies described. Of these, 20 were used for a validation of identification study, six in a control study and eight in a study on the testing of laxity. Preparation of specimens. The shafts of the femur and tibia were denuded of their muscle attachments, while ensuring that the joint capsule and ligaments remained intact. The fibula was resected distal to its neck, and the proximal tibiofibular joint was stabilised using two 4.5 mm transcortical bone screws passed through the fibular head into the tibia as described by previous authors. 9,16 These were adjusted such that the distance between the fibula and tibia was within 1 mm of that allowed by the interosseous membrane. Potting. The femur and tibia were positioned in stainlesssteel cylinders such that their long axes were along the central axis of the cylinder. This was achieved using the method of Amis and Scammell. 17 The bones were secured in these positions using polymethylmethacrylate cement. Identification of the anterior MFL. A medial parapatellar incision allowed access to the anterior aspect of the knee. amfl ACL LM Fig. 2 Photograph showing identification of the amfl. Oblique fibres anterior to the (which lies behind the marker screw S) were followed distally by retracting the anterior cruciate ligament (ACL) to ensure that they attached to the posterior horn of the lateral meniscus (LM). The final test of meniscal attachment was the tug test (see text). The anterior cruciate ligament (ACL) was retracted inferolaterally to reveal the anterior surface of the in the flexed knee. The amfl was identified in this region by the obliquity of its fibres, which contrasts with the vertical orientation of the fibres of the in this situation (Fig. 2), and its proximal attachment between the and the articular cartilage. The amfl was followed distally to ensure that it attached to the lateral meniscus. This was confirmed using a pair of forceps to pull on the suspected amfl (the tug test). The posterior horn of the lateral meniscus moved anteriorly and medially if it was an amfl which was attached to this structure. If this movement was not detected, the fibres were not connected to the lateral meniscus and were oblique fibres of the. After identification of the amfl, the parapatellar incision was closed. Identification of the pmfl. A posterior arthrotomy 2 cm long was performed in order to gain access to the posterior S THE JOURNAL OF BONE AND JOINT SURGERY

THE MENISCOFEMORAL LIGAMENTS: SECONDARY RESTRAINTS TO THE POSTERIOR DRAWER 767 Lateral femoral condyle Medial femoral condyle LM pmfl divided pmfl Lateral tibial condyle amfl Fig. 3 Photograph showing identification of the pmfl. The posterior arthrotomy was made just below the fibres of the oblique popliteal ligament. Again, oblique fibres posterior to the were followed distally to confirm their attachment to the lateral meniscus (LM). Fig. 4 Photograph showing the divided with both MFLs intact. aspect of the knee. The line of this incision was parallel to and just distal to the oblique popliteal ligament (Fig. 3). A fibre-splitting approach was used in order to minimise the possibility of accidentally dividing the MFL as described by previous authors. 18 The posterior aspect of the was closely inspected in order to identify the pmfl. Again, the obliquity of fibres and the proximal and distal attachments were assessed to ensure that this structure was correctly identified. In particular, attachment to the lateral meniscus was confirmed to ensure against misidentification of oblique fibres of the. 19 Validation study of identification. In 20 knees, the capsule of the joint was opened entirely and the collateral ligaments were divided in order to give a clear view and thus validate the identification procedures described above. Testing sequence. The knee with intact ligaments and sutured parapatellar incision was regarded as the intact specimen. The AP and rotatory laxity of the intact specimen was tested as described below. Following these tests, the parapatellar incision was reopened, and the was selectively divided between any MFL(s) which were present, leaving them intact (Fig. 4). The incision was resutured and the -deficient knee was tested. This procedure was repeated a second time after dividing the MFL(s) in the -deficient knee. Thus laxity was determined for the intact, -deficient and -MFL-deficient knees. Testing AP laxity. In the intact knee this was done after identification of the MFLs and closure of the parapatellar incision. The tibia, in its cylinder, was incorporated into a four-degrees-of-freedom rig, which was attached to the crosshead of an Instron 1122 materials testing machine (Instron, High Wycombe, UK) (Fig. 5). This allowed the tibia to move in all possible degrees of freedom relative to the femur in response to an AP drawer test at a fixed angle of flexion and included internal/external and varus/valgus Rig Tibia Femur Baseplate ML VV Fig. 5 PD Photograph of the testing apparatus. After positioning and fixing in the stainless-steel cylinders, the femur and tibia were incorporated in a fourdegrees-of-freedom rig. This was attached to the crosshead of the materials testing machine. The tibia could be positioned at a fixed angle of flexion relative to the femur by adjusting the femoral baseplate. Each degree of freedom is shown by the double-headed arrows indicating the motion of the bearings in the rig (VV, varus/valgus; PD, proximal/distal; ML, medial/lateral; and IE, internal/external). IE VOL. 85-B, No. 5, JULY 2003

768 CHINMAY M. GUPTE, ANTHONY M. J. BULL, RHIDIAN D. THOMAS, ANDREW A. AMIS Posterior translation 20 16 12 8 4 0 NS /MFLs cut cut 0 30 60 90 120 Flexion angle (degrees) Fig. 6 Graph showing the effect of division of the on posterior laxity in the intact knee, and the effect of division of the MFLs in the deficient knee. Division of the resulted in an increase in posterior translation at all angles of flexion in the intact knee (paired t-test and ANOVA, p < 0.005). Division of the MFLs resulted in an increase in posterior translation in the -deficient knee between 15 and 90 of flexion (*, p < 0.005; NS, not significant; SE bars are shown). rotations, and medial/lateral and proximal/distal translations. Thus, during application of the anterior and posterior drawer, the tibia was free to move relative to the femur in a normal passive physiological manner. 17 The force applied during the drawer tests passed through the anatomical centre of the joint as the femur and tibia were aligned along the longitudinal central axis of the cylinders. The femur, in its cylinder, was fixed to the base of the machine by an adjustable feature. Movement of the crosshead imposed AP translations on the tibia at fixed angles of flexion. The angle of flexion of the knee was set by an adjustment on the femoral baseplate. The AP laxity was found for a force of 100 N applied cyclically in the anterior and posterior directions at 0, 15, 30, 60, 90 and 120 of flexion, after five cycles for preconditioning of the viscoelastic structures. The rate of displacement was controlled at 25 mm/min. The machine crosshead translation was used as a measure of AP translation. The stiffness of the rig was found to be 1.6 kn/mm and the results were presented within this resolution. The neutral position of the joint (0 N force at 0 mm displacement) was found at each angle of flexion and defined as the bisect of the points of inflexion in the AP drawer hysteresis curve for the intact joint. Data were analysed using the graphical plotter on the testing machine. The same neutral position was determined for each testing condition by the superposition of the load-displacement curve for the anterior drawer, because it was found that the anterior portion of the hysteresis curve did not change for the testing conditions. AP laxity was measured for each condition of the joint. The knee was removed from the rig to perform the cutting sequences described. However, the precise positions of the cylinder in each fixture were marked. The knee was reinserted in the testing apparatus in order to maintain its previous position. This ensured that the zero-load neutral position was constant. Testing rotatory laxity. The neutral rotatory position of the tibia was defined as the position of the tibia at rest in the neutral AP position of the intact joint. A 5 Nm torque was then applied to the tibia, first in internal rotation, then in external rotation. This was achieved by hanging a mass of 2 kg on a moment arm 0.25 m from the tibia in both directions. The rotation of the tibia relative to the femur was measured by a protractor attached to the rig. This measurement was performed at each increment of flexion in the intact knee, and after the sectioning protocols described. Coupled rotation with flexion. As the tibia flexed relative to the femur, it was hypothesised that there was internal rotation of the tibia on the femur and conversely external rotation as the tibia extended. This phenomenon of coupled rotation was assessed by analysing the rotatory position of the tibia at the different angles of flexion, before and after division of the and MFLs. Control studies. The above parameters were tested before and after the posterior arthrotomy and suturing of the parapatellar incision in six knees. Thus, any effects of these interventions per se were assessed. Statistical analysis. The data consisted of measurements of laxity taken under different loading conditions and at various angles of flexion. Since there were three independent variables in the study (the angle of flexion, the ligament deficiency and the knee itself), a three-way repeated measures analysis of variance with Tukey s post-hoc test was used to analyse the laxity results between populations of the intact, -deficient and -MFL-deficient knees. The anterior and posterior translations relative to the neutral position were compared for each condition of the joint by a one-tailed paired Student s t-test with p = 0.05 for a 95% alpha level after applying Bonferroni s correction factor for multiple comparisons. A one-tailed t-test was used because cutting any restraining structure in the knee can only increase joint laxity and therefore any change in laxity can only be in one direction. The loads required to produce posterior tibial translations of 5 mm and 7 mm were calculated at 15 and 90 of flexion from the load-displacement curves for each sample. From this, the reduction in load required to produce the same translation after division of the and MFL was calculated. This allowed assessment of the percentage contribution of the, MFLs and combined other structures to resisting the posterior drawer force. THE JOURNAL OF BONE AND JOINT SURGERY

THE MENISCOFEMORAL LIGAMENTS: SECONDARY RESTRAINTS TO THE POSTERIOR DRAWER 769 Tibial translation 7 6 5 4 3 2 1 0 0 20 40 60 80 100 120 Flexion angle (degrees) Results Fig. 7 cut /MFLs cut Graph showing the effect of division of the and MFLs on a tibial translation from 100 N anterior drawer. The SE bars have been omitted for clarity; (t-test and ANOVA, p > 0.05 for all comparisons). Presence of MFLs. Eight knees were tested for laxity. They all possessed at least one MFL. In five knees both MFLs were present; in two there was only a pmfl and in one only an amfl. Validation of identification techniques. The amfl and pmfl were both correctly identified in all 20 knees in which the capsule of the joint was subsequently opened. Control studies. There was no significant effect of the posterior arthrotomy and suturing of the parapatellar incision by themselves on any of the testing conditions assessed. For example, posterior laxity at 90 was 7.6 ± 0.6 mm in the intact knee and 7.9 ± 0.8 mm after posterior arthrotomy and suturing of the parapatellar incision. The mean difference in posterior laxity therefore was 0.3 mm (p = 0.2, 95% CI -1.3 mm to 1.9 mm). Table I. Effect of division of the and MFLs on anterior translation from 100 N anterior drawer (95% confidence intervals (CI) shown in brackets) Angle (degrees) /MFLs 0 3.8 (2.9 to 4.7) 3.7 (2.8 to 4.5) 3.8 (3.0 to 4.6) t-test 0.25 0.2 15 5.7 (5.0 to 6.4) 5.8 (5.1 to 6.5) 5.7 (5.1 to 6.3) t-test 0.5 0.20 30 6.3 (5.7 to 7.0) 6.5 (5.8 to 7.2) 6.4 (5.8 to 7.0) t-test 0.2 0.20 60 6.0 (5.0 to 7.0) 6.0 (5.1 to 7.0) 6.1 (5.9 to 6.3) t-test 0.4 0.3 90 5.5 (4.6 to 6.4) 5.7 (4.7 to 6.6) 6.1 (5.3 to 6.9) t-test 0.08 0.08 110 5.6 (4.5 to 6.6) 5.8 (4.8 to 6.8) 6.1 (4.8 to 7.3) t-test 0.05 0.12 AP laxity. Division of the in the intact knee significantly increased posterior drawer (p < 0.01, Fig. 6) at all angles of flexion. The effect was greatest between 60 and 90 of flexion (75% to 77%). There was no significant increase in anterior translation (p > 0.05 all angles, Table I and Fig. 7). Division of the MFLs in the -deficient knee significantly increased posterior translation between 15 and 90 of flexion. This effect was greatest at 90 (34%). There was no significant effect of division of the MFLs on the anterior translation of the tibia relative to the femur (Table I and Fig. 7). Load sharing and posterior drawer force. The mean APdrawer hysteresis curves are shown in Figure 8. The mean posterior drawer force required to produce a translation of the tibia of 7 mm at 90 of flexion was reduced from 41.5 N (SE 15.0) in the -deficient knee to 4.4 N (SE 2.2) after division of the MFLs (paired t-test, p = 0.02; Table II). Load-sharing calculations showed that, in the intact knee, the MFLs contributed 28% of the total force resisting posterior drawer at 90 of flexion (Table III) for posterior dis- intact cut /MFLs cut Anterior displacement 2mm ACL intact 50 N Posterior force (N) Anterior force (N) 68% 28% MFLs 4% other 7mm Posterior displacement Fig. 8 Graph showing mean load-displacement curves for intact, -deficient, and -MFL-deficient knees tested at 90 flexion. Note the hysteresis effect of the anterior/posterior drawer cycle. Note also the relative percentage contributions of the and MFLs to resisting posterior drawer at displacement of 7 mm. The crossed hairlines mark the assumed AP neutral and zero load. VOL. 85-B, No. 5, JULY 2003

770 CHINMAY M. GUPTE, ANTHONY M. J. BULL, RHIDIAN D. THOMAS, ANDREW A. AMIS Table II. Mean values (SE) of force (N) generated at posterior translation of 5 mm and 7 mm of the tibia at 15 and 90 of flexion. The results of the paired t-tests (with correction factor) for intact versus cut and cut versus /MFLs cut are shown as p values 5 mm 7 mm cut /MFLs cut cut /MFLs cut 15 54.5 29.4 (10.9) 11.5 (2.5) 106.5 (20.4) 48.6 (13.4) 22.8 (4.8) (15.6) p value 0.002 0.079 0.001 0.052 90 57.7 17.3 (5.5) 3 (1.4) 102.5 (25.9) 41.5 (15.0) 4.4 (2.2) (16.8) p value 0.018 0.012 0.003 0.018 Table III. Percentage contributions (± SE) to resisting posterior drawer of the and MFLs in intact knee, and the MFLs in the -deficient knee Displacement MFLs Other MFL in deficient knee 15 of flexion 5 mm 48 (8) 21 (6) 31 (5) 42 (9) 7 mm 53 (10) 18 (5) 29 (6) 41 (8) 90 of flexion 5 mm 65 (9) 28 (8) 8 (2) 70 (8) 7 mm 68 (9) 27 (8) 5 (1) 65 (12) Discussion The main finding of our study was that the MFLs made a substantial contribution to resisting posterior tibial drawer in both the intact and -deficient knee. The load-share calculations suggested that the MFLs contributed 28% to the force resisting posterior drawer in the intact knee, and 71% in the -deficient knee at 90 of flexion. We believe that this is the first time that a mechanical role has been shown for these structures in the human knee. placement of 5 mm and 26.5% for a posterior displacement of 7 mm. In the -deficient knee the MFLs contributed 70.1% of resistance at displacement of 5 mm. At 15 of flexion the reduction in force after division of the MFLs was not statistically significant (p = 0.09). Tibial internal/external rotation. Division of the in the intact knee caused no significant change in internal rotation laxity at any angle of flexion (Table IV, Fig. 9), but increased external rotational laxity between 60 and 120 of flexion (Table V and Fig. 10) in response to an applied torque of 5 Nm. Division of the MFLs in the -deficient knee did not significantly affect rotational laxity in either direction (p > 0.1; Table IV). Coupled rotation. In the intact knee the tibia underwent a mean of 6.5 (SE 2.2 ) of internal rotation as the angle of flexion changed from 0 to 110. This coupled rotation with flexion did not change significantly in knees which were either - (6.9 ± 2.3; p = 0.8) or /MFL-deficient (7.3 ± 2.7; p = 0.8). Internal rotation (degrees) 25 20 15 10 5 0 0 30 60 90 120 Flexion angle (degrees) Fig. 9 /MFLs cut Graph showing the effect of division of the and MFLs on internal rotation of the tibia (in degrees) from internal rotation torque of 5 Nm (t-test and ANOVA, p > 0.05 for all comparisons). Table IV. Effect of division of the and MFLs on internal rotation as a result of a 5 Nm of internal rotation torque (95% CI shown in brackets) Angle (degrees) /MFLs 0 7.6 (5.9 to 9.3) 8.7 (7.0 to 10.4) 9.1 (7.0 to 11.2) t-test 0.12 0.36 15 12.8 (10.7 to 14.9) 14 (12.1 to 15.9) 14.5 (12.1 to 16.9) t-test 0.1 0.11 30 15.6 (13.2 to 18.0) 16.6 (13.9 to 19.3) 17.1 (14.6 to 19.6) t-test 0.2 0.20 60 18.5 (15.8 to 21.2) 17.9 (15.5 to 20.3) 18.1 (16.1 to 20.1) t-test 0.14 0.38 90 20.7 (16.9 to 24.5) 18.8 (15.9 to 21.7) 19.8 (16.6 to 23.0) t-test 0.4 0.08 0.08 120 21.3 (17.0 to 25.6) 19.3 (15.6 to 23.0) 20.4 (16.5 to 24.3) t-test 0.07 0.1 THE JOURNAL OF BONE AND JOINT SURGERY

THE MENISCOFEMORAL LIGAMENTS: SECONDARY RESTRAINTS TO THE POSTERIOR DRAWER 771 Table V. Effect of division of the and MFLs on external rotation as a result of a 5 Nm external rotation torque (95% CI shown in brackets) Angle (degrees) /MFLs 0 11 (7.5 to 14.5) 12.5 (10.5 to 14.6) 11.5 (8.1 to 15.0) t-test 0.08 0.13 15 18 (16.3 to 19.7) 19 (17.0 to 21.0) 18.7 (16.7 to 20.8) t-test 0.06 0.30 30 21.3 (19.0 to 23.2) 21.5 (18.6 to 24.3) 22.1 (19.7 to 24.5) t-test 0.32 0.19 60 21.1 (19.0 to 23.2) 23.1 (21.0 to 25.2) 23.3 (21.1 to 25.6) t-test 0.0006 0.23 90 18.6 (16.4 to 20.8) 22.1 (19.5 to 24.7) 23.2 (20.9 to 25.6) t-test 0.01 0.08 110 17.4 (15.7 to 19.1) 20.8 (18.6 to 22.9) 21.9 (18.8 to 25.0) t-test 0.02 0.20 External rotation (degrees) 25 20 15 10 5 /MFLs cut 0 0 30 60 90 120 Flexion angle (degrees) Fig. 10 Graph showing the effect of division of the and MFLs on external rotation of the tibia from an external rotation torque of 5 Nm (*indicates p < 0.05 for the t-test and ANOVA for intact and -deficient knees; all other comparisons showed p > 0.05). Isolation of the MFLs. Previous authors have described difficulty in identifying the MFLs reliably 18,19 for a variety of reasons including deficiencies in the anatomical approaches 18 and inaccurate determination of the distal attachments of fibres thought to be MFLs. 1 Our identification technique, which relied on isolating oblique ligamentous fibres in the vicinity of the, followed by positive identification of their distal attachments to the lateral meniscus aided by the tug test, was a reliable method of isolating the MFLs without disturbing the other structures which support the knee. There were no false-positive or false-negative results in the validation study. AP laxity. Division of the resulted in an increase in posterior drawer in the intact knee, which was minimal in full extension and greatest between 60 and 90 of flexion. This confirmed the findings of several previous studies. 7-9 Load-sharing calculations (Tables II and III) showed that in the intact knee, the MFLs contributed 28% to the total resistance to posterior displacement of 5 mm of the tibia at 90 of flexion. The contribution to resistance is even higher in the -deficient knee (70.1% at 90 for displacement of 5 mm; Table III). This supports the laxity findings and reinforces the hypothesis that the MFLs provide a significant amount of the resistance to posterior displacement in the -deficient knee, particularly in the mid-flexion range. Division of the MFLs increased posterior translation from a 100 N posterior drawer force by a mean of 34% at 90 of flexion. Thus Radoïévitch s hypothesis postulating that the MFLs act as a third cruciate ligament has been confirmed to some extent. 5 Several previous studies have reported changes in posterior laxity after division of the. 9-11 In these studies the MFLs were divided together with the. Future studies should consider the MFLs as structures separate from the, which contribute to resisting posterior displacement in their own right. Some in vivo supporting evidence has been provided by the study of Clancy et al. 6 In their patients with acute rupture of the, the posterior drawer, when performed with the tibia in internal rotation, was reduced if the MFLs were intact. Our study suggests that the posterior drawer would also be reduced with the tibia in neutral rotation. If the MFLs do limit posterior displacement in the deficient knee, it would be appropriate to assess their integrity during diagnostic studies such as arthroscopy or MRI. The difficulties in identifying the MFLs at dissection have been well described previously. 19 We speculate that it is also possible to misidentify the MFLs as fibres of the during arthroscopy. If the results of the present study are borne out in clinical practice, it is likely that the posterior drawer sign will be reduced in patients with complete rupture of the in which the MFL(s) remain intact. In these cases, there is a possibility of misdiagnosing a complete rupture of the as a partial rupture because the intact MFLs have been erroneously identified as fibres of the. It is therefore of paramount importance to identify accurately the MFLs by assessing their distal attachments to the lateral meniscus during arthroscopy. There are no studies which examine the incidence of injuries to the when the MFLs remain intact. Although it is unclear whether the stabilising influence of the MFLs in cadaver specimens has any clinical consequences, it could VOL. 85-B, No. 5, JULY 2003

772 CHINMAY M. GUPTE, ANTHONY M. J. BULL, RHIDIAN D. THOMAS, ANDREW A. AMIS be speculated that knees with injuries of the in which the MFLs are intact may be more stable than those in which the MFLs are also compromised. Further clinical studies are required to assess the functional effects and prognosis of this subgroup of injuries. At present the MFLs are often sacrificed during reconstruction of the even if they are intact. This occurs during the process of clearing the soft tissues out of the intercondylar notch and when drilling the femoral tunnel. It may be possible to preserve the MFLs during such drilling, particularly in the single-bundled reconstruction. This would be of value if the stabilising functions of the MFLs were confirmed. Rotational laxity. Our study showed a small effect of dividing the on the rotational laxity of the knee between 60 and 120 of flexion. This is in agreement with previous studies, 8-10 and is explained biomechanically by the central position of the and its attachments. There was also no effect of division of the MFLs on rotational laxity in the -deficient knee. This seems counterintuitive, because the MFLs lie obliquely and pass across the centre of the joint in the coronal plane. Thus it would be reasonable to expect that internal rotation causes tension in the MFLs, and hence that division of the MFLs increases internal rotation. However, because the MFLs lie close to the centre of rotation of the knee, their effect on rotation may be negligible because of a short moment arm. Previous authors 10 have noted that division of the medial collateral ligament and posteromedial corner increases rotational laxity of the knee. 7 These structures were not evaluated in our study, but their contribution to rotational stability may overshadow that of the MFLs. Thus damage to these structures may be required before the MFLs contribute to rotational stability. This may also explain the lack of effect of dividing the MFLs on coupled rotation. However, there are no studies which have examined the effect of the posteromedial capsule in stabilising rotatory laxity in the -deficient knee. Limitations. Our study has some limitations since it is based on cadaver specimens. In particular, the roles of muscle tensions and joint forces have not been included. However, the methods used have previously been shown 20,21 to give data on joint laxity in vitro which matches those obtained by instrumented tests in vivo. 22 Concerns may also be expressed as to the effects of tissue lysis and dehydration after removal of the skin, and freezing of specimens. In order to minimise these effects, the joints were kept moist during testing. The methods used in our study do not have significant effects on the behaviour of the collagenous tissues being studied. 23,24 Thus, although this work cannot represent the functioning, living joint, it does give valid data regarding the passive joint-stabilising function of the MFLs. The drawer-testing protocol using the four-degrees-offreedom rig, allows the joint to move in an unconstrained manner. However, this does not allow the reproduction of joint kinematics when structures are divided, and therefore limits conclusions about the function of the MFLs in the intact joint. This limitation of cadaver testing was addressed by Fujie et al, 25 who developed a solution by the application of robotic technology and a universal force sensor. It would be of value to perform a similar investigation using this application. The contribution of the MFLs to load sharing in the intact knee is deduced from load-displacement curves. It would be of value to confirm this contribution by reversing the cutting sequence, i.e. by dividing the MFLs before division of the. One further assumption in our study is that the MFLs tested represent a homogeneous population. When both MFLs were present, they were simultaneously divided. Further studies should investigate sequential cutting of the MFLs when both are present in a given sample. We also observed that the amfl was tense in flexion while the pmfl was tense near full extension. These qualitative observations agree with those of Friederich and O Brien, 26 who suggested a reciprocal tightening and slackening of the MFLs during flexion of the knee. It may be speculated that the amfl resists posterior drawer in the flexed knee, whilst the pmfl performs this function near full extension. The relatively small number of specimens precluded testing this hypothesis in the present study. In conclusion, our study demonstrates that there is a significant contribution of the MFLs to resisting the posterior drawer of the tibia in the -deficient knee. We believe that this is the first study which demonstrates a mechanical function for these structures. It is likely that the MFLs supplement the function of the in AP stability. They do therefore require consideration in the patient with a suspected injury. If further in vivo studies confirm the function of the MFLs, this may have considerable bearing on the prognosis and management of -related knee injuries. Mr C. M. Gupte is supported by the Royal College of Surgeons of Edinburgh Lindsay Stewart Sports Fellowship, the Smith and Nephew Foundation Orthopaedic Fellowship, The Aircast Foundation and by the Wishbone Trust. Dr A. M. J. Bull is supported by the Arthritis Research Campaign, who also donated the materials testing machine. We also thank Mr Paul Aichroth and Professor Fred Heatley for their invaluable advice. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. References 1. Gupte CM, Bull AM, Thomas RD, Amis AA. A review of the function and biomechanics of the meniscofemoral ligaments. Arthroscopy 2003;19:161-71. 2. McCormack D, McGrath J. Anterior menisco-femoral ligament. Clin Anat 1992;5:485-7. 3. Wan AC, Felle P. The menisco-femoral ligaments. Clin Anat 1995;8:232-6. 4. Humphry GM. A treatise on the human skeleton (including the joints). Cambridge: MacMillan and Co, 1858. 5. Radoïévitch S. Les ligaments des ménisques interarticulaires du genou. Annales d Anatomie et Pathologie 1932;8:4-13. 6. Clancy WG, Shelbourne KD, Zoellner GB, et al. Treatment of knee joint instability secondary to rupture of the posterior cruciate ligament: report of a new procedure. J Bone Joint Surg [Am] 1983;65-A:310-22. THE JOURNAL OF BONE AND JOINT SURGERY

THE MENISCOFEMORAL LIGAMENTS: SECONDARY RESTRAINTS TO THE POSTERIOR DRAWER 773 7. Ritchie JR, Bergfeld JA, Kambic H, Manning T. Isolated sectioning of the medial and posteromedial capsular ligaments in the posterior cruciate ligament-deficient knee: influence on posterior tibial translation. Am J Sports Med 1998;26:389-94. 8. Gollehon DL, Torzilli PA, Warren RF. The role of the posterolateral and cruciate ligaments in the stability of the human knee: a biomechanical study. J Bone Joint Surg [Am] 1987;69-A:233-42. 9. Race A, Amis AA. Loading of the two bundles of the posterior cruciate ligament: an analysis of bundle function in a P drawer. J Biomech 1996;29:873-9. 10. Markolf KL, Mensch JS, Amstutz HC. Stiffness and laxity of the knee: the contributions of the supporting structures: a quantitative in vitro study. J Bone Joint Surg [Am] 1976;58-A:583-94. 11. Bergfeld JA, McAllister DR, Parker RD, Valdevit AD, Kambic H. The effects of tibial rotation on posterior translation in knees in which the posterior cruciate ligament has been cut. J Bone Joint Surg [Am] 2001;83-A:1339-43. 12. Harner CD, Hoher J. Evaluation and treatment of posterior cruciate ligament injuries. Am J Sports Med 1998;26:471-82. 13. Dandy DJ, Pusey RJ. The long-term results of unrepaired tears of the posterior cruciate ligament. J Bone Joint Surg [Br] 1982;64-B:92-4. 14. Miller MD, Bergfeld JA, Fowler PJ, Harner CD, Noyes FR. The posterior cruciate ligament injured knee: principles of evaluation and treatment. Instr Course Lect 1999;48:199-207. 15. Barton TM, Torg JS, Das M. Posterior cruciate ligament insufficiency: a review of the literature. Sports Med 1984;1:419-30. 16. Bergfeld JA, McAllister DR, Parker RD, Valdevit AD, Kambic HE. A biomechanical comparison of posterior cruciate ligament reconstruction techniques. Am J Sports Med 2001;29:129-36. 17. Amis AA, Scammell BE. Biomechanics of intra-articular and extraarticular reconstruction of the anterior cruciate ligament. J Bone Joint Surg [Br] 1993;75-B:812-7. 18. Kaplan E. The lateral menisco-femoral ligament of the knee joint. Bull Hosp Joint Dis 1956;17:176-82. 19. Gupte CM, Smith A, McDermott ID, et al. Meniscofemoral ligaments revisited: anatomical study, age correlation and clinical implications. J Bone Joint Surg [Br] 2002;84-B:846-51. 20. Butler DL, Noyes FR, Grood ES. Ligamentous restraints to anteriorposterior drawer in the human knee: a biomechanical study. J Bone Joint Surg [Am] 1980;62-A:259-70. 21. Piziali RL, Rastegar J, Nagel DA, Schurman DJ. The contribution of the cruciate ligaments to the load-displacement characteristics of the human knee joint. J Biomech Eng 1980;102:277-83. 22. Daniel D, Stone M. KT-1000 antero-posterior displacement measurements. In: Daniel D, Akeson W, O Connor J, eds. Knee ligaments: structure, function, injury and repair. New York, Raven Press, 1990:427-47. 23. Woo SL, Orlando CA, Camp JF, Akeson WH. Effects of postmortem storage by freezing on ligament tensile behavior. J Biomech 1986;19:399-404. 24. Viidik A, Sandqvist L. Influence of post-mortem storage on the tensile strength characteristics and histology of ligaments. Acta Orthop Scand 1965;79A:1-35. 25. Fujie H, Mabuchi K, Woo SL, et al. The use of robotics technology to study human joint kinematics: a new methodology. J Biomech Eng 1993;115:211-7. 26. Friederich N, O Brien W. Functional anatomy of the meniscofemoral ligaments. June 1990. Fourth Congress of the European Society for Knee Surgery Sports Traumatology and Arthroscopy (ESSKA). VOL. 85-B, No. 5, JULY 2003