The Effect of Femoral Component Rotation on the Extensor Retinaculum of the Knee

Transcription

1 The Effect of Femoral Component Rotation on the Extensor Retinaculum of the Knee Kanishka M. Ghosh, 1 Azhar M. Merican, 2 Farhad Iranpour, 3 David J. Deehan, 1 and Andrew A. Amis 4 1 Orthopaedic Surgery Department, Newcastle University Hospital, Newcastle upon Tyne NE2 4HH, United Kingdom, 2 University of Malaya Medical Centre, Kuala Lumpur, Malaysia, 3 Musculoskeletal Surgery Department, Imperial College London, Charing Cross Hospital, London W6 8RF, United Kingdom, 4 Departments of Mechanical Engineering and of Musculoskeletal Surgery, Imperial College London, London SW7 2AZ, United Kingdom Received 20 July 2009; accepted 11 January 2010 Published online 9 March 2010 in Wiley InterScience ( DOI /jor ABSTRACT: Malrotation of the femoral component may cause patellofemoral complications after total knee replacement (TKR). We hypothesized that femoral component malrotation would cause excessive lengthening of the retinacula. Retinacular length changes were measured by threading fine sutures along them and attaching these to the patella and to displacement transducers. The knee post-tkr was flexed-extended while the quadriceps were tensed, then the measurements repeated after rotating the femoral component 58 internally and then 58 externally. Internal rotation shortened the medial patellofemoral ligament (MPFL) significantly from 1008 to 08 extension. External rotation lengthened the MPFL significantly from 908 to 08 extension. The transverse fibers of the lateral retinaculum showed no significant differences. The MPFL attaches directly from bone to bone, so it was lengthened directly by movement of the trochlea and patella, whereas the deep transverse fibers of the lateral retinaculum attach to the mobile iliotibial tract, so they were not lengthened directly. ß 2010 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 28: , 2010 Keywords: extensor retinaculum; total knee arthroplasty; femoral component rotation; patellar tracking; lateral release Malalignment is the most frequently discussed complication in total knee replacement (TKR) 1 5 and occurs in about 10 30% of patients, depending on surgical technique and the anatomical landmarks used. 6 Rotational malalignment of the femoral component is a common cause of revision surgery, 1 with consequences including flexion gap asymmetry, patellar maltracking, and varus or valgus malalignment of the lower extremity in the flexed position, taking the tibia out of the sagittal plane. 1,2,7,8 Not all patients with a malrotated femoral component report unsatisfactory results. Some studies reported that a relative internal rotation of the femoral component between 38 and 68 is tolerable, 5,6,9 while external rotation as high as 88 may not always cause clinical problems. 10 However, why some patients become so strongly symptomatic that they require revision surgery and others do not are open to question. Retinacular tensions are important in patellofemoral stability, especially in early knee flexion. 11 Abnormal tensions as a result of femoral malrotation after TKR may contribute to patellar maltracking, component wear, and pain. Even in experienced hands clinical estimation of the epicondylar axis is inaccurate and cannot be relied upon as the sole determinant of femoral component rotation. 12 This is especially pertinent in patients with advanced degenerative disease where anatomical landmarks are less distinct. In this situation, knowledge of retinacular behavior may help in achieving optimal patellar tracking. Our aim was to test the hypothesis that external femoral component rotation would move the prosthetic trochlea laterally, shortening the lateral retinaculum, and lengthening the medial retinaculum, and vice versa for internal rotation. Correspondence to: A.A. Amis (T: þ ; F: þ ); a.amis@imperial.ac.uk ß 2010 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. MATERIALS AND METHODS The length change patterns of the medial and lateral retinacula were measured across the range from 08 to 1108 knee flexion-extension with the prosthetic femoral component implanted in neutral rotation and then at 58 internal/ external rotation. Eight right knees were obtained from a tissue bank following approval from the local Research Ethics Committee. A power analysis was based on prior data; 13 the experiment had 80% power to detect a smallest average difference between pairs of 2.1 mm for the medial patellofemoral ligament (MPFL) and 4.4 mm for the lateral retinaculum with an alpha level of The specimens were sealed in polyethylene bags and frozen at 208C prior to use. The mean age of the donors was 69 years (57 to 87 years). The specimens included 20 cm each of femur and tibia. After thawing, they were kept moist using occasional water spraying. The skin, underlying fat, and muscles other than the distal quadriceps were removed. Care was taken to preserve the retinacula, fascia of the quadriceps muscles, and the iliotibial band (ITB). The quadriceps were then separated into 6 components: rectus femoris (RF), vastus intermedius (VI), vastus lateralis longus (VLL), vastus lateralis obliquus (VLO), vastus medialis longus (VML), and vastus medialis obliquus (VMO). The distal tendinous fibers of the muscles merged together and were left intact to ensure that the actions of the muscles were as physiological as possible. VI was separated from the femur. Cloth strips were sewn to the proximal end of each muscle component and the ITB by stitching through the whole muscle bulk, providing attachments for muscle loading cables (Fig. 1). RF and VI were looped together to form a central muscle group. The tibia was transected 15 cm below the joint line. The fibular head was fixed to the tibia by two bone screws to maintain its anatomic position, then the distal part was removed. A brass tube was cemented into the femoral medullary cavity using PMMA bone cement. An intramedullary rod was cemented into the tibia. The knee was mounted in the test rig by clamping the femoral intramedullary tube (Fig. 1). The muscles were loaded in their physiological directions relative to the femoral axis. 14,15 A load of 175 N was applied to the quadriceps muscle groups by 1136

2 FEMORAL ROTATION AND THE RETINACULA 1137 Figure 1. A right knee placed on the jig with the quadriceps muscles pulled in their physiological directions. Optical trackers were secured to the bones. LVDTs mounted on the base of the test rig measured length changes of the extensor retinaculum via sutures (published originally in Ref. 13). hanging fixed weights on a cable and pulley system. The ITB was loaded to 30 N, 08 lateral, and 68 posterior to the femoral axis. 16,17 The quadriceps tensions were according to the mean physiological cross-sectional areas of the muscles: RF þ VI 35%, VLL 33%, VLO 9%, VML 14%, and VMO 9%. 14,18 To preserve the retinacula and reduce confounding factors such as variable suture tensioning affecting retinacular length changes after closure of a standard medial parapatellar approach, a longitudinal transpatellar approach was used. The lengths of the medial and lateral retinacula were not significantly affected by this arthrotomy. 19 A cruciate retaining TKR (Genesis II, Smith & Nephew, Memphis, TN) was used. The procedure was performed by the same consultant surgeon on all eight knees. The patella was resurfaced using a dome shaped inlay button centered along the median ridge. A digital vernier caliper was used to adjust the resection so that the thickness of the assembly matched the intact patella within 0.5 mm. The femoral component had a thicker posterolateral than posteromedial condyle, enabling the implant to sit at 38 external rotation from the posterior condylar axis and maintain an equal flexion gap without the need to remove excess bone from the posteromedial condyle. The component was positioned ensuring the posterior condyles of the implant were horizontal; this was checked with a spirit level and defined as neutral rotation. A fixed marker screw in the anterior femur was used as a datum to ensure that the anterior trochlear height and the femoral length matched the natural knee within 0.5 mm. After establishing this position, the femoral component was cemented over a conversion module from the TKR revision system, which allowed it to be attached to an intramedullary rod that ran through the femoral tube and protruded proximally (Fig. 2a). This protruding rod allowed the femoral component to be rotated within the closed knee. Extra bone cuts on the distal femur allowed rotation of the femoral component while preserving the ligament attachments (Fig. 2b). The tibial cut was made using an extramedullary guide aligned with the rod. The trial tibial tray was allowed to float to a functional position relative to the femoral component in all cases this lay just lateral to the medial edge of the tibial tubercle. Flexion and extension gaps were assessed subjectively in full extension and 908 of flexion. The Figure 2. (a); Rotation of the femoral component was allowed by mounting it onto a conversion module with a long intramedullary rod, so that it could be rotated within the closed knee from the proximal end. (b); The femoral condyles were trimmed to allow the movement. transpatellar arthrotomy was closed using two cannulated cancellous screws placed across the patella. 19 The incisions in the quadriceps and patellar tendon were approximated using Vicryl 2/0 (Ethicon Co., Somerville, NJ). Tibiofemoral flexion was measured dynamically using a Polaris Optical Tracking System (NDI Intl., Waterloo, ON, Canada). One optical tracker was placed on each of the femur and tibia. The specimen orientation was then defined by digitizing landmarks including the most posterior points of the femoral condyles to define the medial-lateral direction. Knee flexion was calculated as the angle between the femoral and tibial axes in the sagittal plane perpendicular to the medial lateral axis using Visual 3D (C-Motion Inc., Rockville, MD). 20 Retinacular length changes were measured using a previously described technique, 13 using monofilament sutures (Ethilon 2/0, Ethicon Co.) attached to Linear Variable Displacement Transducers (LVDTs Solartron Metrology, Bognor Regis, UK). Length changes were recorded using device specific software (Solartron Orbit Excel). To measure length changes in the MPFL, a suture was passed from the LVDT through a screw eye secured at the epicondylar origin of the MPFL, then threaded along the ligament s fibers and secured at the superomedial corner of the patella. Because the deep transverse band in the lateral retinaculum originates and inserts into two mobile structures (ITB and lateral patella), we measured their

3 1138 GHOSH ET AL. movements separately to calculate the length change of the fibers in between. The transverse band linked the superolateral corner of the patella to the deep aspect of the ITB. One suture passed from the LVDT through an eyelet secured on the ITB to the patella along the transverse band of fibers. A second LVDT had a suture passing to the ITB eyelet only (Fig. 1). Thus, length change in the deep transverse band was calculated from the difference between the two values. Ten preconditioning cycles of passive flexion-extension were performed with the quadriceps loaded before testing. The knee was flexed by pushing posteriorly against the tibial intramedullary rod, using a transverse nylon rod, to avoid imposing secondary moments. Data were sampled at 12 Hz while the quadriceps extended the knee slowly from 1108 to 08. Testing was first conducted with the standard TKA with the femoral component in neutral rotation, then repeated with the femoral component internally rotated 58 and then externally rotated 58. Two-way repeated-measures ANOVA with post-hoc Student s paired t-tests were used to examine length changes between neutral and internal rotation or neutral and external rotation and knee flexion at 108 intervals, for each of the retinacula. Significance was set at p < RESULTS The mean pattern of length change in the MPFL after TKR in neutral rotation was close to isometric, with a lengthening of mm as the knee extended from 1108 to 308 (p ¼ 0.044) followed by a shortening of mm as the knee extended fully (Fig. 3). With the femoral component in neutral rotation, the deep transverse band of the lateral retinaculum shortened by a mean of mm as the knee extended from 808 to 08 (p < , Fig. 4). Internal rotation of the femoral component caused an overall shortening of the MPFL when compared to neutral rotation (p < ). Significant changes occurred from 1008 to 08 of extension; as the knee extended this became more significant: 1008: mm (p < 0.05); 908 to 808: mm (p < 0.01); Figure 3. Length changes in the MPFL (mm) versus knee flexion (8), with the femoral component in neutral rotation, 58 external rotation (ER), and 58 internal rotation (IR) (mean SD, n ¼ 8). The neutral rotation data are offset 28 for clarity. Zero length change ¼ length of MPFL in intact knee at 08 flexion. Figure 4. Length changes in the lateral retinaculum (mm) versus knee flexion (8), with the femoral component in neutral rotation, 58 external rotation (ER) and 58 internal rotation (IR) (mean SD, n ¼ 8) The neutral rotation data are offset 28 for clarity. Zero length change ¼ length of lateral retinaculum in intact knee at 08 flexion. 708 to 08: mm (p < 0.001). External rotation caused an overall lengthening (p < ). Significant changes were found from 808 to 08 of knee extension; 808: mm (p < 0.01) and 708 to 08: mm (p < 0.001, Fig. 3). Internal or external rotation of the femoral component caused small but significant (p ¼ for internal and p < for external) overall differences in length of the lateral retinaculum (Fig. 4). However, no significant differences occurred at any flexion angle with internal or external rotation. DISCUSSION We tested the hypothesis that malrotation of the femoral component following TKR would stretch the retinacula. This was quantified by measuring length changes in the principal structures of the medial and lateral retinaculum. We speculated that, because the patella is guided by the trochlear groove, femoral malrotation would cause maltracking. We hypothesized that the resultant tilt and shift of the patella would cause significant lengthening of the retinaculum on one side along with shortening of the other, and vice versa. Five degree of internal or external rotation caused significant length changes in the MPFL. External rotation resulted in lengthening and internal rotation caused shortening. Significant differences in length change patterns occurred over a wider arc of motion when the femoral component was internally rotated ( ). In contrast, femoral component rotation did not cause significant length changes in the deep transverse fibers of the lateral retinaculum. We have not reported the length changes of the retinacula as stretching or slackening because our interobserver comparisons found such large differences of identification of the transition point between slack and taut fibers

4 FEMORAL ROTATION AND THE RETINACULA 1139 that to report what was stretching or slackening would be unreliable. Because the patella is guided by the trochlear groove, it moves with the prosthesis when the femoral component is malrotated (Fig. 5). Malrotation when the knee is near extension causes more marked changes in trochlear groove position than in flexion, where the groove is close to the axis of rotation of the femoral component. External component rotation shifts the trochlear groove laterally and decreases the effectiveness of the lateral ridge in resisting lateral translation of the patella. This is partly counteracted by the reduction of lateral forces from the reduced Q angle. The lengthening of the MPFL in this experiment indicates that a resultant lateral shift and/or tilt of the patella occurred within the lateralized trochlear groove. 7,21 Conversely, internal component rotation causes the lateral flange to rise, shifts the trochlear groove medially, and only the shallow medial trochlear facet resists medial translation of the patellar component. The reduction in the length of the MPFL meant that a medial translation and/or tilt of the patellar component had occurred. The lateral retinaculum behaved in an unexpected manner: femoral component rotation of 58 did not cause significant changes in the length of the deep transverse fibers. Therefore, the ITB moved backwards and forwards not only with flexion and extension, but also with the patellar tilt and shift that accompanied the movement of the prosthetic trochlea (Fig. 5). Thus, femoral component external rotation allowed the deep transverse fibers of the lateral retinaculum to move posteriorly and the ITB with them, so length changes in the transverse band were nonsignificant. In degenerative disease, capsular structures are stiffer, 22 possibly resulting in abnormal tensioning of the lateral retinaculum, but no data exist to allow us to simulate that situation. Since doing this experiment, we found 23 that the deep transverse fibers of the lateral retinaculum are an order of magnitude stiffer than the MPFL, 24 so the MPFL is more likely to demonstrate length changes. Figure 5. Schematic representation showing how femoral component rotation slackens/tightens the MPFL (y) directly due to its bone-to-bone attachments. However, the lateral retinaculum is attached to the mobile ITB, so it is not stretched directly. Patellofemoral problems cause a considerable proportion of complications after TKR, including loosening, wear, instability, and patellar fracture. 1 In vitro studies showed that femoral component malposition affects patellofemoral kinematics and abnormal tensioning of the soft tissues may also occur. 1,21,25,26 In general, the position of the patella followed the movement of the trochlear groove, shifting and tilting laterally with femoral external rotation. Over-stuffing the patellofemoral joint causes significant lengthening of the retinacula 27 and insertion of a TKR may increase retinacular tension. 28 Our data could be useful for computer modeling of the patellofemoral joint. Most studies of femoral component rotation have rotated the component from the posterior condylar axis to the transepicondylar axis and then noted more physiological tracking. Because the Genesis II was designed to sit in 38 external rotation from the posterior condylar axis, direct comparisons with other published work required that the implants were taken beyond the transepicondylar axis in both directions. Miller et al. 29 found that if the femoral component was aligned parallel to the transepicondylar axis, more normal patellar tracking occurred, minimizing patellofemoral shear forces early in flexion. Clinical studies have shown that external rotation of the femoral component results in a reduction in lateral release rates. 1,25,30 Internal rotation can be expected to stretch the lateral retinaculum, increasing the likelihood of lateral retinacular release. 29 However, our study did not show significant length changes in the transverse fibers of the lateral retinaculum following femoral malrotation, suggesting that other retinacular structures may be affected. The main limitations of our study arose from the use of normal cadaveric knees, without obvious degenerative disease. We could not simulate pathological changes such as tight bands in the lateral retinaculum. The clinical aim of TKR is to restore normal knee behavior, and so it is necessary to start with normal knees to isolate and understand the effects of femoral component rotation. Development of the cadaveric model should include the effects of degenerative pathology after clinical research to define the tissue changes. Factors such as the chronicity of the degenerative process and height of the patella will contribute to the large variability expected clinically, making meaningful experimental findings elusive. Also, there is a lack of uniformity in clinical studies, regarding which anatomical structures are released and differing and ill-defined indications for doing a lateral release. The force applied to the quadriceps was limited by tearing of the muscles of the elderly specimens, so normal physiological magnitudes were not attained. The muscle tension was constant across the arc of knee flexion, whereas in life the tension increases as we squat down; this may have affected the length change patterns recorded. Knee flexion beyond 1108 would have been desirable, but aspects of the rig prevented this. Despite these limita-

5 1140 GHOSH ET AL. tions, we measured the effects of femoral prosthesis malrotation on length changes in the retinacula. Measurements were made dynamically to capture transient events. The muscles were loaded in physiological directions. The ITB was also loaded in accordance with earlier studies 16,17 that showed that it affects patellar tracking. 17,31 Alterations in patellar tracking may cause abnormal retinacular tensions, and the extensor retinaculum contributes significantly to patellar stability. 15,18,32 Malrotation of the femoral component following TKR causes significant differences in length change patterns within the principal structures of the retinaculum. This may be related to lateral retinacular release after TKR, which may have an incidence as high as 50% 33 and has been related to loosening of the patellar component. 34 This effect was larger in the MPFL than in the deep transverse fibers of the lateral retinaculum, contrary to popular understanding. These length changes may be related to the tensile behavior of the retinacula: the MPFL stretches 26 7 mm to failure, 24 the deep transverse fibers of the lateral retinaculum 11 4 mm. 23 A recent study of patellar stability found that the lateral capsule was a greater contributor than the transverse fibers near knee extension, 32 so this structure may affect patellar tracking if the femoral component is malpositioned in internal rotation. It has become accepted that patellar maltracking often relates to excessive tension in the lateral retinaculum. Our study and previous work 13 found that the transverse fiber band of the lateral retinaculum shortens when the knee extends, because the ITB tract to which it attaches moves anteriorly with extension. In contrast, the MPFL attaches directly from bone to bone, so movement of the femoral component, and with it the patella, directly affects its length. ACKNOWLEDGMENTS K.M. Ghosh was supported by Smith & Nephew (UK) Ltd, A.M. Merican by the University of Malaya Medical Centre and the Arthritis Research Campaign, and F. Iranpour by the Furlong Charitable Foundation for Research. The tissue specimens were supplied by the International Institute for the Advancement of Medicine, Jessup, PA. REFERENCES 1. Akagi M, Matsusue Y, Mata T Effect of rotational alignment on patellar tracking in total knee arthroplasty. Clin Orthop Relat Res 366: Berger RA, Crossett LS Malrotation causing patellofemoral complications after total knee Arthroplasty. Clin Orthop Relat Res 356: Hakki S, Coleman S, Saleh K, et al Navigational predictors in determining the necessity for collateral ligament release in total knee replacement. J Bone Joint Surg Br 91: Verdonk PC, Pernin J, Pinaroli A, et al Soft tissue balancing in varus total knee arthroplasty: an algorithmic approach. Knee Surg Sports Traumatol Arthrosc 17: Crottet D, Kowal J, Sarfert SA, et al Ligament balancing in TKA: evaluation of a force-sensing device and the influence of patellar eversion and ligament release. J Biomech 40: Barrack R, Schrader T, Bertot A, et al Component rotation and anterior knee pain after total knee arthroplasty. Clin Orthop Relat Res 392: Anouchi Y, Whiteside L, Kaiser A The effects of axial rotational alignment of the femoral component on knee stability and patellar tracking in total knee arthroplasty demonstrated on autopsy specimens. Clin Orthop Relat Res 287: Jerosch J, Peuker E, Philipps B, et al Inter-individual reproducibility in perioperative rotational alignment of femoral components in knee prosthetic surgery using the transepicondylar axis. Knee Surg Sports Traumatol Arthrosc 10: Insall JN, Scuderi GR, Komistek RDA, et al Correlation between condylar lift-off and femoral component alignment. Clin Orthop Rel Res 403: Nagamine R, Whiteside LA Effect of medial displacement of the tibial tubercle on patellar position after rotational malposition of the femoral component in total knee arthroplasty. J Arthroplasty 11: Bicos J, Fulkerson JP, Amis AA Current concepts review: the medial patellofemoral ligament. Am J Sports Med 35: Kinzel V, Ledger M, Shakespeare D Can the epicondylar axis be defined accurately in total knee arthroplasty? Knee 12: Ghosh KM, Merican AM, Iranpour-Boroujeni F, et al Length change patterns of the extensor retinaculum and the effect of total knee replacement. J Orthop Res 27: Farahmand F, Senavongse W, Amis AA Quantitative study of the quadriceps muscles and trochlear groove geometry related to instability of the patellofemoral joint. J Orthop Res 16: Christoforakis J, Bull AM, Strachan RK, et al Effects of lateral retinacular release on lateral stability of the patella. Knee Surg Sports Traumatol Arthrosc 14: Bull AMJ, Andersen HN, Basso O, et al Incidence and mechanism of the pivot shift. An in-vitro study. Clin Orthop Relat Res 363: Kwak SD, Ahmad CS, Gardner TR Hamstrings and iliotibial band forces affect knee kinematics and contact pattern. J Orthop Res 10: Senavongse W, Amis AA The effects of articular, retinacular, or muscular deficiencies on patellofemoral joint stability. J Bone Joint Surg [Br] 87B: Merican AM, Ghosh KM, Deehan DJ, et al The transpatellar approach for the knee in the laboratory. J Orthop Res 27: Amis AA, Senavongse W, Bull AMJ Patellofemoral kinematics during knee flexion-extension: an in vitro study. J Orthop Res 24: Armstrong A, Brien HJC Patellar position after knee arthroplasty. J Arthrop 18: Fishkin Z, Miller M, Ritter C Changes in human knee ligament stiffness secondary to osteoarthritis. J Orthop Res 20: Merican AM, Sanghavi S, Iranpour F, et al The structural properties of the lateral retinaculum and capsular complex of the knee. J Biomech 42: Mountney J, Senavongse W, Amis AA, et al Tensile strength of the medial patellofemoral ligament before and after repair or reconstruction. J Bone Jt Surg [Br] 87:

6 FEMORAL ROTATION AND THE RETINACULA Kessler O, Patil S, Colwell CW, et al The effect of femoral component malrotation on patellar biomechanics. J Biomech 41: Anglin C, Brimacombe JM, Hodgson AJ, et al Determinants of patellar tracking in total knee arthroplasty. Clin Biomech 23: Ghosh KM, Merican A, Iranpour F, et al The effect of overstuffing the patellofemoral joint on the extensor retinaculum. Knee Surg Sports Traumatol, Arthrosc 17: Ogata K, Ishinishi T, Hara M. Evaluation of patellar retinacular tension during total knee arthroplasty. J Arthropl 12: Miller MC, Berger RA, Petrella AJ Optimizing femoral component rotation in total knee arthroplasty. Clin Orthop Relat Res 392: Sodha S, Kim J, Kevin B, et al Lateral retinacular release as a function of femoral component rotation in total knee arthroplasty. J Arthroplasty 19: Merican AM, Amis AA Iliotibial band tension affects patellofemoral and tibiofemoral kinematics. J Biomech 42: Merican AM, Kondo K, Amis AA The effect on patellofemoral joint stability of selective cutting of lateral retinacular and capsular structures. J Biomech 42: Weber AB, Worland RL, Jessup DE, et al The consequences of lateral release in total knee replacement: a review of over 1000 knees with follow up between 5 and 11 years. Knee 10: Meding JB, Fish MD, Berend ME, et al Predicting patellar failure after total knee arthroplasty. Clin Orthop 466: