Many aspects of medial patellofemoral reconstruction are still controversial
Today, in some aspects medial patellofemoral ligament (MPFL) research can be compared to the ACL research, however, there is some substantial differences in the number of papers, attention and scientific achievements. ACL studies have focused on the biomechanics of ACL reconstruction, prevention and the anatomical factors, such as the notch configuration and the tibia slope. In respect to these reconstructions, the choice of graft, single bundle or double bundle, foot prints, bone tunnels, tension, fixation, ingrowths, postoperative regime, return to sports, learning curve, factors for failures, including additional underlying instabilities, have been studied.
Patellofemoral (PF) surgeons can learn that even though ACL and MPFL reconstructive surgery have similarities, they must be aware of the differences. It is difficult to outline all the differences, but the importance of bony stability in the PF joint and the check rein nature of the MPFL makes import differences. With respect to mechanism of trauma, the rupture of both the ACL and MPFL have similarities. Both often happen with the knee forced into slight knee flexion/valgus, with the femur in internal rotation and the foot fixed. Afterwards, in the more chronic phase, PF instability has a more episodic and variable nature compared to the ACL deficiency. This, as well as the lower incidence, makes the PF research challenging. Nevertheless, we can in many ways learn from the ACL research and at least get inspired.
Based on the past 10 years of success with the MPFL reconstructions (MPFLR), it is my impression the MPFLR is seen by some as a panacea for PF instabilities. However, an increasing number of failures and complication reports and also revision cases are now beginning to emerge. The importance of reporting and analysis of the failures meticulously, in order to identify the patients and methods with less good results, is well known and it can help us to optimize the indications for isolated MFPLR and to fine-tune our operative techniques.
Today, the Danish registry for ACL reconstructions has collected valuable data for 20,000 reconstructions. I believe this can be the future road for PF instabilities as well, and an initiative has been done for the establishment of a similar Danish register for PF surgery. Hopefully, the Danish health care system can provide support for this project.
Femoral attachment site
Many papers have focused on the anatomic femoral MPFL foot print, and the results have been a bit conflicting and no consensus exists. The anatomic studies have all been based on standard knees, and yet no study has involved knees with predisposing anatomical factors for patella instabilities, such as trochlear dysplasia or patella alta. For the graft attachment site, most authors advocate a so-called near isometric point The Shoettle point has been accepted for the MPFLR, but unfortunately is subject to error with imperfections of radiographic technique and resulting inaccurate graft placement. We know inappropriate femoral attachment can have devastating consequences with reduced flexion or extension, increased forces in the PF joint, arthrofibrosis and graft failure. The most common problem is the lack of knee flexion and from this perspective it has been advocated that a more distal femoral position might be safer than a more proximal site. Thaunat recommended a favorable anisometry with a progressive slackening of the graft as the knee flexes and then to fixate the graft in extension. Personally I believe in restoring the anatomy. However, when this is said, I must admit that I always tend to place my graft a little bit in the anisometric distal position simply because the opposite can lead to the above mentioned flexion problems. On the other hand, you cannot be too enthusiastic with the distalization either because there have been reports of patients with extension deficit and pain in extension, based on a too distal fixation.
Victor and colleagues demonstrated that the proximal part of the native MPFL is tight in extension and slack in flexion, and that is just opposite for the distal part of the MPFL. There is general agreement that the MFPL relatively slackens in flexion and stability in higher degrees is dependent on the trochlea. In cases with trochlea dysplasia, the bony stability in flexion often is less, meaning the forces in the MPFL graft are relatively more importance in flexion. So when performing MPFLR in dysplastic cases (personally I do a concomitant trochleoplasty), this has to be taken into account. Therefore, especially in these situations a distally placed graft could be a problem.
Several reports have found isolated MPFLR in higher degrees of trochlea dysplasia results in deteriorating results and sometimes leads to failure. Another thing is that we know little about the correct attachment site when a concomitant transposition of the tibial tuberosity or a trochleoplasty is performed. In my experience with trochleoplasty procedures, I have found the attachment site changes simply because a lot of bone is removed in the proximal part of the articulation. The result is the graft becomes at little loose in extension and to compensate for this, the graft must be placed more distally to avoid tightening of the graft in flexion.
Regarding the skeletal immature patient, the use of fluoroscopy and a distal from the physis attachment site seems to be the right thing to do. I previously used the adductor sling method, but the proximal attachment site makes the graft too tight in flexion and correspondingly to loose in extension. I am not that nervous about damaging the physis in patients who have been menstruating for at least 1 year. First of all, the risk for growth disturbance is low and if any harm should occur to the medial physis, then this will lead to a small varus change, likely to be beneficial for the patella stability.
Patella/quadriceps attachment site
The patella/quadriceps foot print of the MPFL hasn’t been given nearly the attention as the femoral foot print. Most of the anatomical studies have localized it on the proximal two-thirds of the patella. Lately, after having done dissections both inside and outside of the knee joint, new discoveries and breakthroughs have made it clear the MPFL has a broad, fan-shaped configuration and that most proximal fibers are the most substantial fibers. Fulkerson and Edgar, as well as Mochizuki and colleagues, described that the most proximal fibers attaches to the vastus intermedius tendon (the distal quadriceps tendon), not the patella itself. Mochizuki also found that the distal fibers of MPFL interdigitate with the deep layer of the medial retinaculum and they also attach to the medial margin of the patella tendon. As far as I know, no one has yet focused the attachment site in AP level of the patella and how this may influence especially the tilting of the patella.
Based on anatomical studies, Fulkerson and Edgar developed a new technique for reconstruction of the MPFL and they named it the medial quadriceps tendon-femoral ligament (MQTFL) to be anatomically accurate. The advantage is attachment to the patella by drilling bone tunnels or using screws and anchors is avoided and the abovementioned tilting forces on the patella is avoided. Our few anatomical studies have confirmed the above mentioned anatomic findings and we have now done 20 MQTFL reconstructions, giving promising but short-term follow-up.
Besides the attachments sites, the most critical step in the MPFLR is the tensioning of the graft and several methods have been advocated. Thaunat fixates the graft in extension, while others fixate the graft with the knee in 30° to 45° of flexion. In this situation, the patella is tracked by the deeper trochlea. To prevent a detrimental over constrained graft, some fixate the graft in the longest length. Beck and colleagues showed that 2 N of graft tension restored normal patellar translation.
When I fixate the graft, I try to obtain close to the isometric point. As mentioned above, with a slight distal placement, giving a slight anisometric, but less likely to be over tight in flexion I don’t hesitate to replace my Beath pin to obtain the right spot. My technique is always a femoral through tunnel in a 30° cranial direction to avoid any central structures and eventually implants. The Beath pin with the graft sutures goes out laterally. I tension the graft firmly without tension and this is followed by cycling of the knee, from full extension to 90° of flexion, causing the graft to adapt to the appropriate length. The graft is normally, alone by the friction forces in the bone tunnel, kept in the obtained length, so it can be fixated in 30° to 45° of flexion, using an interference screw. I always test if the correct stability is achieved and that the MPFL function as a checkrein. I don’t hesitate to replace my interference screw if I am not satisfied.
Using the “no tension” fixation, the risk of over constraining the graft is reduced. I must admit I haven’t used intraoperative fluoroscopy to confirm the dissection and Beath pin placement as I recognize the relevant anatomy. This is imperative. I have good results, but since I haven´t done postoperative radiograph routinely and since I haven´t done ten years follow-up, the situation is that I could be unaware of complications. Rarely, I do concomitant lateral release in isolated MPFLR in selected cases with trochleoplasty or tibial tuberosity transposition. I do the release after MPFLR in order to avoid over constraining of the graft on the lateral side.
Type of graft
Several type of graft for MPFLR has been found useful. Perhaps the hamstrings are the most popular, followed by the superficial quadriceps graft, introduced by Steensen and colleagues. The adductor tendon, patella tendon, allografts as well as artificial graft have been found useful. Schoettle has advocated to use the gracilis tendon instead of the semitendinosus tendon simply because it is strong enough. The strength of the native MPFL approximate 208 N with a mean stiffness of 24 N.
Compared to the gracilis tendon, the semitendinosus tendon has more important function as a dynamic valgus stabilizator and ACL agonist and it is likely that it is important for the prevention of ACL ruptures. The superficial quadriceps tendon seems to be attractive since it reduces the patella site fixation problems and also it has a more flat structure and to a higher degree it mimics the native MPFL. The down side is the long scar in front of the knee. Fink and Storz have suggested a quadriceps tendon harvesting system to reduce this problem. The adductor magnus tendon as graft has also been advocated, however, I find that the proximal insertion is unfavorable.
My personal preferences for about 10 years have been the gracilis tendon with two anchors in the patella, double bundle and an interference screw in the femur. Lately, I have been encouraged by the study from Fulkerson and colleagues about MQTFL reconstruction, and after a few anatomical dissections, I must agree with the existence of the MQTFL structure. Therefore, we have now performed about twenty MQTFL reconstructions and the procedure is faster, cheaper and elegant, compared to the MPFL procedure with drilling or anchors into the patella and associated risk of complication such as fracture being avoided. I have observed two cases of patella stress fracture with anchors into the patella.
A little detail, conflicting with the observation from Fulkerson, is that in respect to the femoral insertion site, I have observed the isometric spot to be at same spot as the MPFL. With respect to the superficial quadriceps tendon, I have yet only used this in revision cases and found it useful. Based on the study by Mochizuki and colleagues, with a more fan shaped configuration and the study by Victor and colleagues with the proximal part of the MPFL tight in extension and the distal part more tight in flexion, I have tried to fix the distal part of the gracilis graft at the patella site and the proximal part in the quadriceps tendon, and this seems function well, and at least it gives a more fan-shaped configuration, compared to the superficial quadriceps graft.
To do or not do postoperative immobilization is the question. In one hand, we want to protect the graft with a brace. On the other hand, we want to avoid arthrofibrosis. Personally, I don’t use braces and this is based on the principle that the graft is a check rein to secure the patella and when placed isometric the tension is equal through range of motion. Since I don’t perform isolated reconstruction in the cases where more extensive forces can be expected, as in cases with increases the tibial tubercle–trochlear groove distance or trochlear dysplasia, I believe the disadvantages of the brace outweigh the benefits.
There are many ways to fixate the graft and just like other reconstructive surgery, it is a matter of preferences and economy. Anchors, interference screw, periosteal sutures, docking technique and bone tunnels have been described as well as direct suture into the distal quadriceps tendon. A study from Berard and colleagues have focused tunnel widening and it seems that patella alta and higher degrees of trochlear dysplasia predispose to this phenomenon, however, the clinical relevance is unknown.
As you maybe have realized after reading this, many aspects of medial PF reconstruction are still controversial. As Matthew Bollier stated in his Patellofemoral Update blog, “only precise patient outcome instruments and long-time follow-up can help us to distinguish between right and wrong.”
Remember, and this especially applies to MPFL surgery that “there is nothing so bad you can’t make it worse than it was before surgery.” Medial patellofemoral reconstruction surgery should be reserved for orthopedic surgeons with particular interest and understanding in the PF joint.
Baldwin JL. Am J Sports Med. 2009;doi:10.1177/0363546509339909.
Beck P. Patellofemoral contact pressures and lateral patellar translation after medial patellofemoral ligament reconstruction. Am J Sports Med. 2007;35:1557-1563.
Bollier M. Arthroscopy. 2011;doi:10.1016/j.arthro.2011.02.014.
Feller JA. The medial patellofemoral ligament revisited : an anatomical study. Knee Surg Sports Traumatol Arthrosc. 1993;1:184-186.
Fulkerson JP. Arthrosc Tech. 2013;doi:10.1016/j.eats.2013.01.002.
Hopper GP. Am J Sports Med. 2014;doi:10.1177/0363546513518413.
Mochizuki T. Knee Surg Sports Traumatol Arthrosc. 2013;doi:10.1007/s00167-012-1993-7.
Nelitz M. Knee Surg Sports Traumatol Arthrosc. 2011;doi:10.1007/s00167-011-1548-3.
Nomura E. Anatomical analysis of the medial patellofemoral ligament of the knee, especially the femoral attachment. Knee Surg Sports Traumatol Arthrosc. 2005;13:510-515.
Parikh SN. Am J Sports Med. 2013;doi:10.1177/0363546513482085.
Philippot R. Knee Surg Sports Traumatol Arthrosc. 2009;doi:10.1007/s00167-009-0722-3.
Schottle PB. Technical note: anatomical reconstruction of the medial patellofemoral ligament using a free gracilis autograft. Arch Orthop Trauma Surg. 2008;128:479-484.
Shah JN. Am J Sports Med. 2012;doi:10.1177/0363546512442330.
Smirk C. The anatomy and reconstruction of the medial patellofemoral ligament. Knee. 2003;10:221-227.
Steensen RN. A simple technique for reconstruction of the medial patellofemoral ligament using a quadriceps tendon graft. Arthroscopy. 2005;21:365-370.
Steensen RN. The anatomy and isometry of the medial patellofemoral ligament: implications for reconstruction. Am J Sports Med. 2004;32:1509-1513.
Thaunat M. The favourable anisometry: an original concept for medial patellofemoral ligament reconstruction. Knee. 2007;14:424-428.
Tuxøe JI. The medial patellofemoral ligament: a dissection study. Knee Surg Sports Traumatol Arthrosc. 2002;10:138-140.
Victor J. Am J Sports Med. 2009;doi:10.1177/0363546509337407.
Wagner D. Knee Surg Sports Traumatol Arthrosc. 2013;doi:10.1007/s00167-012-2015-5.
Lars Blønd, MD, is a consultant surgeon, Koege University Hospital and chief surgeon, Teres Medical Group, Copenhagen, and President of the Danish Society of Sports Medicine. He can be reached at Ortopaedkirurgisk afdeling, Koege Universitets Sygehus, Lykkebaekvej 1, DK-4600, Denmark; email: Larsfirstname.lastname@example.org or www.larsblond.com.
Disclosure: Blønd has not relevant financial disclosures.