This article was updated on November 17, 2015. The title, which had previously read “Medial Patellofemoral Reconstruction in Children and Adolescents” now reads “Medial Patellofemoral Ligament Reconstruction in Children and Adolescents.”
An erratum has been published: JBJS Reviews. 2015 Dec 15;3(12):e4.
➢ The medial patellofemoral ligament (MPFL) is an important static stabilizer of the knee and resists lateral translation of the patella.
➢ Acute dislocation of the patella is associated with a high rate of MPFL tears.
➢ Conservative measures, such as physical therapy and activity modification, have a role in the treatment of primary patellar dislocation in patients at low risk for recurrent instability. However, younger patients, who are at higher risk, and patients with recurrent instability generally require operative intervention.
➢ While many surgical techniques can be used in an attempt to restore the native anatomy and function of the MPFL, the outcomes of MPFL reconstruction are superior to those of repair.
Patellar dislocations are a common disabling knee injury in children and adolescents. Approximately 90% of primary patellar dislocations result in injury to the medial patellofemoral ligament (MPFL)1,2. MPFL injuries in children and adolescents, and the management of those injuries, differ from those in adults in numerous ways. In the past decade, studies have demonstrated better results with MPFL reconstruction than with medial repair and/or lateral release3-5.
The purpose of this article is to review recent literature on the natural history of pediatric patellar dislocations, review advances in imaging following acute patellar dislocation in pediatric patients, discuss the biomechanical properties of the MPFL, and evaluate the surgical options as well as the outcomes of MPFL reconstruction.
This article will cover:
• The physical examination findings, classification, natural history, and anatomic risk factors for patellar dislocation in children and adolescents.
• Radiographic and magnetic resonance imaging (MRI) findings associated with patellar dislocation and those of patients at risk for patellar dislocation as well as anatomic indices such as patella alta, trochlear dysplasia, and tibial tubercle-to-trochlear groove (TT-TG) distance and tibial tubercle-to-posterior cruciate ligament (TT-PCL) distance.
• The anatomy of the MPFL with a focus on its relationship to the distal femoral physis as well as its biomechanical properties.
• Surgical options for MPFL reconstruction in children and adolescents.
• A systematic review of reported outcomes of MPFL reconstruction and an evidence-based treatment algorithm for addressing pediatric patellar instability.
Patellofemoral instability is a complex disorder that most commonly affects children and adolescents. The incidence of patellar dislocation has been reported to be between 5.8 and 43 per 100,000 individuals6,7. The peak incidence occurs between the ages of ten and seventeen years, and girls have a higher incidence than boys7.
Pediatric patellar dislocation can be grouped into four main types: syndromic, obligatory, fixed lateral, and traumatic (Table I)8. Patellar dislocations in patients with Ehlers-Danlos syndrome, Marfan disease, or another condition characterized by ligamentous laxity are in the syndromic category. Dislocations that occur with every episode of knee flexion and reduce with every episode of knee extension are classified in the obligatory group. If the patella remains dislocated lateral to the trochlea during flexion and extension, the patient is considered to have a fixed lateral dislocation. The final category, traumatic, accounts for the majority of patellar dislocations. In one study, 70% of 189 patellar dislocations occurred during sports activities or dancing7.
Natural History Following Patellar Dislocation
In order to create an algorithm for treatment of patients with patellofemoral instability, it is important to review the natural history following the first episode of patellar dislocation. The rate of redislocation has been reported to be between 17% and 71%7,9-13 (Table II). Risk factors for recurrent instability include a younger age at the time of the initial dislocation, trochlear dysplasia, patellofemoral malalignment, and a positive family history of patellofemoral dislocation7,9-11. Fithian et al.7 reported on the natural history of 189 patellar dislocations and found girls between ten and seventeen years old to be at the highest risk for recurrent instability. They reported a 17% rate of recurrent instability following the first dislocation and a 49% rate of recurrence in those with more than one episode of instability.
The long-term sequelae of patellofemoral instability are yet to be clearly delineated. The consensus from the historical and recent literature is that there is a role for physical therapy and symptomatic management for patients with a first-time patellar dislocation who are at low risk for recurrent instability (i.e., who have no anatomic risk factors)14,15. However, operative treatment should be considered for patients at high risk for recurrent instability—namely, younger, skeletally immature patients—as the risk of cartilage damage from recurrent instability is high.
History and Physical Examination
A thorough history is essential when assessing a patient with patellofemoral instability. Patients should be asked about prior patellar dislocations and their frequency and duration. In addition to understanding the events associated with the patellar dislocation, clinicians should question the patient regarding other joint instability and ligamentous laxity. Most patients with acute patellar dislocation are found to have a moderate to large knee effusion on physical examination, and when the patient is too young to convey an accurate history the knee effusion may be the only sign of the patellar dislocation. Concentric reduction of the patella within the trochlea should be confirmed with clinical examination and radiographic evaluation.
A patient with patellofemoral instability should be assessed for lateral translation of the patella with manual manipulation, during which the clinician should look for an apprehension sign. We also perform a moving apprehension test by applying a lateral force on the patella and assessing for apprehension throughout the knee range of motion16. This test is then repeated with a medially directed force on the patella, again while assessing the patient for a lack of apprehension throughout the knee range of motion. Clinicians should assess for a J sign by having the patient sit with the knee flexed 90° and then actively extend the knee while the examiner observes the patella for a lateral shift during terminal extension17. The patient should also be examined while he or she is in the standing position to assess overall alignment and to look for genu valgum. A torsional profile of the femur and tibia should be performed as well as an assessment of the Q angle (the quadriceps femoris muscle angle, which is the angle created between a line from the anterior superior iliac spine to the center of the patella and a line from the center of the patella to the tibial tubercle).
While most patellar dislocations are the result of trauma, anatomic abnormalities including patella alta, trochlear dysplasia, genu valgum, or laxity of the soft-tissue restraints of the knee may predispose certain individuals to recurrent patellar instability (Table III). In females, femoral anteversion is associated with a higher prevalence of patellar dislocation7. Anatomic variations that increase the Q angle can predispose a patient to patellofemoral instability. For example, increased femoral anteversion, genu valgum, and external tibial torsion can all increase the Q angle and affect patellofemoral alignment.
The tibial tubercle-to-trochlear groove (TT-TG) distance measures how lateral the tibial tubercle is relative to the center of the trochlea in the axial plane. An increased TT-TG distance is a risk factor for patellofemoral instability. While this distance was originally described as being measured on axial radiographs, Dejour et al.18 later described measuring the TT-TG distance using computed tomography (CT) with the knee in 30° of flexion, and Schoettle et al.19 confirmed that magnetic resonance imaging (MRI) can be used reliably to make the same measurement. Given that the position of the tibia relative to the femur varies throughout flexion and extension, Seitlinger et al.20 proposed using the distance between the tibial tubercle and medial border of the posterior cruciate ligament (TT-PCL distance). The TT-PCL distance is independent of knee rotation or trochlear morphology and may be used as an alternative to the TT-TG distance for assessing the need for a distal realignment procedure. The Q angle is falsely reduced in patients with a laterally subluxated patella, and for this reason the TT-TG distance is preferred over that measurement.
The patella is stabilized against lateral dislocation by both osseous and soft-tissue structures. The soft tissues in the medial aspect of the knee are the major contributors to stabilization against lateral dislocation from 0° to 30° of knee flexion. However, beyond 30° of knee flexion, the trochlea creates an osseous confine for the patella, preventing lateral dislocation in a normal knee. Patients with a dysplastic trochlea lack this osseous barrier to patellar dislocation and are more likely to have patellofemoral instability. Similarly, hypoplasia of the patella or lateral femoral condyle can increase patellofemoral instability.
Patella alta is a common anatomic variant that is associated with patellofemoral instability, although no causative link has been proven largely because of a lack of temporally related evidence. While many believe that patella alta may be a cause of patellofemoral instability18,21-30, it may in fact be a result of patellofemoral instability. One longitudinal study showed that patella alta improved to within normal ranges after MPFL reconstruction31. It has also been theorized that a high-riding patella enters the confines of the trochlea at a higher degree of flexion than in a normal knee, which increases the range of motion of the knee in which the patella is restrained mostly by soft tissues and may also contribute to instability events31.
As mentioned, from full knee extension to 30° of flexion, the soft tissues of the knee, rather than osseous structures, stabilize the patella. The MPFL is the primary restraint to lateral patellar dislocation1,32. The MPFL originates in the sulcus on the medial femoral condyle between the medial femoral epicondyle and the adductor tubercle and inserts on the superior half of the patella. Acute patellofemoral dislocation results in a tear in the MPFL in >90% of cases33; a torn or incompetent MPFL is a risk factor for recurrent patellar instability34.
Anteroposterior, lateral, and Merchant radiographs of the knee should be obtained to assess patients with patellofemoral instability. Merchant views, made with the knee in 30° of flexion and the x-ray beam angled to match the degree of flexion, are preferable to sunrise views, in which the knee is in hyperflexion and the patella is forced into the trochlear groove35. Thirty degrees of knee flexion allows better depiction of the congruence angle between the trochlea and the patellar facets. Additionally, a standing full-length hip-to-ankle anteroposterior radiograph is useful for assessing alignment preoperatively.
The degree of patella alta can be measured on the lateral radiograph. Multiple indices have been described for evaluating patella alta, including the Caton-Deschamps index36, Insall-Salvati ratio37, modified Insall-Salvati ratio38, and Blackburne-Peel index39 (Fig. 1).
MRI can be used to identify osteochondral lesions or loose bodies that need to be addressed operatively. One study of 195 patients with patellar dislocation identified osteochondral lesions in 49% and loose bodies in 13%40. Another MRI study identified osteochondral injury in 70% of eighty-two patients following an acute episode of instability41. Additionally, MRI is useful to confirm a recent history of patellar dislocation, particularly in patients who are too young to provide an accurate history. A patellar dislocation causes cartilaginous trauma and bone marrow edema in the lateral femoral condyle and medial patellar facet during spontaneous reduction, which is represented by increased signal intensity on fluid-sensitive, fat-suppressed imaging, typically short tau inversion recovery (STIR) sequencing.
MRI can be used to accurately measure the TT-TG distance. In a study of the TT-TG distances on 608 MRIs in patients between the ages of nine and sixteen years, Dickens et al.42 reported that the TT-TG distance varied according to the chronological age of the patient. Specifically, the TT-TG distance was positively associated with the natural logarithm of age. In their pediatric population, the average TT-TG distance was 8.5 mm in patients without patellar instability compared with 12.1 mm in patients with patellofemoral instability. Therefore, age-related normal values, as opposed to the fixed adult parameters, should be considered when measuring the TT-TG distance in skeletally immature patients.
MPFL Anatomy, Biomechanics, and Relationship to the Distal Femoral Physis
The MPFL lies in the second layer of the medial soft-tissue complex of the knee as described by Warren and Marshall43. The MPFL is made up of two major bundles, a short oblique bundle that inserts on the superior patellar pole and a second, inferior straight bundle44. The patellar attachment of the MPFL is a fan-like structure that inserts at the junction between the proximal and middle thirds of the superomedial border of the patella44. The femoral origin of the MPFL is on the medial femoral condyle in a sulcus distal to the adductor tubercle and proximal to the attachment of the medial collateral ligament (MCL)44. Schöttle et al.45 dissected eight adult cadaver knees to further delineate the anatomic relationships of the MPFL and thereby determine the radiographic landmarks for MPFL reconstruction. They found that the femoral origin lies approximately 1.3 mm anterior to the posterior femoral diaphyseal cortex and described a reproducible point 2.5 mm distal to the posterior origin of the medial femoral condyle and proximal to the level of the posterior point of the Blumensaat line. Another cadaveric study, of sixteen skeletally immature specimens, identified the MPFL femoral attachment distal to the medial aspect of the distal femoral physis46. Recently, Shea et al.47 identified the center of the MPFL footprint at or below the physis in six cadaveric knees, but in two specimens from older donors the proximal extent of the MPFL origin extended above the medial aspect of the physis. The complex three-dimensional concave shape of the distal femoral physis must be taken into consideration in pediatric patients48. A lateral radiograph used in isolation to identify the MPFL femoral origin may result in false localization to a point proximal to the physis49.
Mochizuki et al.50 found that the proximal fibers of the MPFL were mainly attached to the vastus medialis obliquus fibers while distally the ligament invests mostly in the medial retinaculum. This anatomic relationship allows the MPFL to act as a checkrein to keep the patella and patellar tendon medial, which maintains the tracking of the patella within the trochlea throughout flexion and extension.
Studies assessing the zone of injury in MPFL rupture have produced varying results (Table IV)1,33,41,51-55. In one recent study in which STIR sequences were used to identify the precise insertion site of the MPFL as well as the zone of injury in acute tears, the MPFL was found to insert an average of 5 mm distal to the medial aspect of the physis and the injury was isolated to the patellar attachment in 61% of forty-four patients53. Another recent study evaluated fifty patients following a primary patellar dislocation and showed that the skeletally immature patients were more likely to have an injury of the patellar attachment site of the MPFL than the skeletally mature patients as determined by preoperative ultrasound56.
The main function of the MPFL is to provide restraint against lateral translation of the patella from 0° to 30° of knee flexion. The MPFL has relatively low tension throughout flexion and extension, maintaining between 2 and 10 N of force57. Overall, the MPFL remains isometric between 0° and approximately 90° of knee flexion and becomes slack beyond 90° of flexion; however, there is variation in the length relationships between the different bundles of the MPFL throughout flexion and extension58-60. Unlike the anterior cruciate ligament (ACL) or posterior cruciate ligament (PCL), which both have substantially higher resting tension, the MPFL functions as a checkrein for the patella, guiding the patella into the trochlea and preventing dislocation until osseous constraints take over. When a lateral force is applied to the patella, as in the case of traumatic patellar dislocation, the native MPFL can withstand a force of approximately 200 N before tearing61. This load to failure is much lower than that of the ACL, which fails at loads exceeding 1725 N62.
Patients with no anatomic risk factors (Table III) or associated pathologic findings following a primary traumatic patellar dislocation are counseled about activity modification and offered a course of physical therapy for quadriceps and core strengthening to improve patellar tracking. While we are not aware of any randomized controlled trials comparing nonoperative with operative management of first-time patellar dislocation, the authors of a review of five cohort studies including 339 patients concluded that there is insufficient evidence to identify a significant difference between the results of operative and nonoperative treatment63. If a patient sustains a second patellar dislocation event, the risk of a recurrent episode of instability increases from 13% to nearly 50%, and operative treatment is indicated for most such patients7.
The main options for operative treatment of patellofemoral instability include guided growth if there is pathologic genu valgum, distal realignment, and MPFL reconstruction. Figure 2 demonstrates our treatment algorithm. Clinicians use anatomic and radiographic parameters to determine which patients would benefit from each option. Patients with pathologic genu valgum may benefit from temporary medial hemiepiphysiodesis64. This technique uses a plate to tether the medial growth plate while allowing lateral-sided growth with the effect of gradual correction of valgus malalignment. In general, this technique is considered for patients with genu valgum of >10° and more than two years of growth remaining.
Distal realignment procedures are indicated in patients with an increased age-specific TT-TG distance and/or a positive J sign on physical examination. Tibial tubercle transfer is an osseous distal realignment procedure that involves an osteotomy of the tibial tubercle and medialization with or without anteriorization of the tubercle65. This procedure corrects the lateralization of the tibial tubercle as measured by the TT-TG or TT-PCL distance but should be considered only for patients with closing proximal tibial physes.
There are multiple options for soft-tissue distal realignment procedures that may be safely used in children and adolescents. The relative indications for these procedures are similar to those for tibial tubercle transfers and include a large TT-TG distance and large Q angle. The Roux-Goldthwait procedure was the first described, and it has been modified by Marsh et al.66. With this technique, the surgeon creates a longitudinal split in the patellar tendon, transfers the lateral half of the detached patellar tendon deep to the medial half, and sutures this transferred half of the tendon to the periosteum (Fig. 3, A). The technique described by Nietosvaara et al.6 uses a hamstring graft to reconstruct a new patellotibial ligament and MPFL (Fig. 3, B). Because of poor reported outcomes with the classic Galeazzi procedure, we no longer recommend that technique, which involves a semitendinosus tenodesis to the medial aspect of the patella with a concomitant lateral capsular release (Fig. 3, C)67,68. Garin et al.69 and Luhmann et al.70 both described a patellar tendon transfer in which the patellar tendon is completely detached from the underlying tibial apophysis and secured medially with sutures to the periosteum of the proximal part of the tibia, which medializes the tendon by approximately 50% of its width (Fig. 3, D).
An isolated MPFL reconstruction is a valid surgical option for patients who have sustained a patellar dislocation but do not have a large Q angle or TT-TG distance. In the past, surgeons simply attempted to repair the ruptured medial soft tissues following a patellar dislocation, but the patient outcomes of medial repair combined with lateral release of capsular and soft tissues to improve alignment have been no better than those of nonoperative management4. While these procedures may have failed because they were not tailored to repair the MPFL on the basis of the documented zone of injury, the lateral release does not significantly improve patellofemoral kinematics5,71. Similarly, medial retinacular tissue imbrication is a nonanatomic procedure that can result in excessive medialization and abnormal patellar tracking72.
The well-documented failure of medial repairs has stimulated the development of a number of techniques for MPFL reconstruction over the past twenty years. As reported in the literature, MPFL reconstruction techniques have been performed with a variety of anatomic structures as grafts, including gracilis and/or semitendinosus hamstring autografts, adductor magnus pedicled grafts, and medial quadriceps grafts73-76.
The goal of MPFL reconstruction is to recreate the anatomic checkrein for the patella, particularly between 0° and 30° of knee flexion, without disrupting the distal femoral physis. One technique is to harvest the gracilis and/or semitendinosus, depending on the patient’s size, and use a doubled hamstring autograft to reconstruct the MPFL. The graft can be fixed in osseous tunnels or sockets with tenodesis screws, suture anchors, or docking techniques77-79. Some surgeons are advocating use of either suture anchors or incomplete osseous sockets (leaving an intact lateral patellar cortex) to avoid patellar fractures, which have been reported following MPFL reconstruction with osseous tunnels80. There has been a multitude of variations of this technique, including utilization of a pedicled quadriceps tendon76,81 (Fig. 4, A), a pedicled patellar tendon (Fig. 4, B), or a pedicled adductor tendon75 (Fig. 4, C) to reconstruct the MPFL. Similarly, other authors have advocated using the MCL or adductor tendon as a pulley73,74 (Fig. 4, D and E). Fulkerson and Edgar82 advocated suturing the graft to the vastus medialis obliquus fibers to avoid penetrating the patellar cortical bone.
An example of our preferred hamstring MPFL reconstruction with patellar sockets rather than complete patellar tunnels is shown in Figure 4, F and Video 1. In this technique, the femoral socket is identified on both lateral and anteroposterior fluoroscopic images to verify that the guidewire is positioned distal to the distal femoral physis. Once the sockets are reamed, the graft is positioned with the knee in 20° to 30° of flexion. The graft length is adjusted so that it holds the patella in the center of the trochlear groove. With the knee in extension, the graft should be positioned as a checkrein to allow for patellar excursion of up to one-fourth the width of the patella, as described previously2. Given reports of poor results of medial repair and lateral release3-5, we rarely perform lateral release in isolation. However, when the lateral retinaculum is tight and prevents centralized patellar tracking or neutral patellar tilt following MPFL reconstruction, we perform lateral retinacular lengthening. Postoperatively, patients are allowed to bear weight as tolerated with a brace in extension until quadriceps strength has returned. Early range of motion is encouraged.
While there have been a number of studies on the outcomes of MPFL reconstruction in skeletally mature patients77,79,83-85, there have been fewer referring specifically to skeletally immature patients (Table V)68,80,86. Grannatt et al.68 reported the results of a Galeazzi semitendinosus tenodesis in twenty-eight pediatric patients (thirty-four knees) with an average age of 11.1 years; 35% of the knees required a subsequent operation for recurrent patellofemoral instability. The few studies pertaining to children have shown MPFL-based reconstructions to provide better outcomes than the Galeazzi procedure. Nelitz et al.86 reported the results of MPFL reconstruction in twenty-one consecutive patients with open growth plates. At an average of 2.8 years postoperatively, eighteen were satisfied with the result and three were partially satisfied. Only four patients were unable to return to their preoperative level of sports activity. In a report of 179 cases of MPFL reconstruction in children and adolescents with an average age of 14.5 years, there was a 4.5% prevalence of recurrent instability at an average of 16.2 months postoperatively80. Various authors have also noted a mild decrease in radiographic patellar height as an indirect consequence of MPFL reconstruction, which could contribute to decreased postoperative patellofemoral instability31,83,87.
While MPFL reconstruction reduces the prevalence of patellar instability, the procedure is not without complications. A meta-analysis of MPFL reconstruction studies, which included adult patients, revealed a 26% rate of complications, including patellar fracture, loss of knee flexion, wound complications, and pain, in 629 knees88. In a study of 179 MPFL reconstructions performed during a six-year study period in patients less than twenty-one years of age, Parikh et al.80 reported a 16.2% complication rate80. Eight of the 179 patients had recurrent instability, eight had knee stiffness, and six had patellar fracture. Tanaka et al.89 described technical complications and failures of MPFL reconstruction, particularly with proximal misplacement of the femoral socket resulting in tightness in knee flexion. Typically, the MPFL is reconstructed with a quadrupled hamstring graft, the diameter of which is much larger than the diameter of the native MPFL, so an error in overtensioning such a graft could have a large effect on knee range of motion. A graft that is shortened by >3 mm and positioned proximally will double the compressive forces across the medial aspect of the patella90.
Patellofemoral instability is a complex disorder that most frequently affects children and adolescents. Acute dislocation of the patella is associated with a high rate of MPFL tears, and surgical reconstruction of the MPFL has improved patient outcomes. Advances in imaging and biomechanics have further delineated the anatomy and kinematics of the MPFL. Various operative techniques have been used with the aim of restoring the native anatomy and function of the MPFL, with MPFL reconstruction yielding superior outcomes compared with those of MPFL repair.
The authors thank Cynthia Conklin for her assistance with the illustrations in this manuscript.
Source of Funding: There was no external funding source for this manuscript.
Investigation performed at the Hospital for Special Surgery, New York, NY
Disclosure: None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any aspect of this work. None of the authors, or their institution(s), have had any financial relationship, in the thirty-six months prior to submission of this work, with any entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. One or more of the authors has had another relationship, or has engaged in another activity, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.
- Copyright © 2015 by The Journal of Bone and Joint Surgery, Incorporated