8.10 The patellofemoral joint

Disorders of the patellofemoral joint present in one of two ways:

Anterior knee pain may arise from the patellofemoral joint, or may arise remotely—typically the hip. Surgical management of anterior knee pain has uncertain outcomes. On the other hand instability may be mechanical or functional. If mechanical then successful surgical correction should be considered.

Around 40% of elective referrals to an orthopaedic department are for knee pain, of which about a third are for pain felt anteriorly. Accurate assessment and appropriate management is therefore essential. The algorithm for elucidating the likely problem starts with the age of the patient (see Figure 8.10.1).

The patellofemoral joint has evolved in man in response to bipedal gait. To effect hip abductor power the greater trochanter of the upper femur has become more lateral. This has resulted in the quadriceps line of action becoming more valgus with respect to the insertion of the patellar ligament into the tibial tubercle, creating an angle (the quadriceps or Q-angle). To overcome the resulting lateral pull to the patella, the lateral trochlear facet of the femoral sulcus has become more flared, to act as a buttress against lateral patella displacement.

The patellofemoral joint is made up of the quadriceps tendon, patella and patellar ligament, and tibial tubercle along with the medial and lateral retinaculum within which lie the medial patellofemoral ligament and lateral patelloiliotibial band ligament. Other condensations within the medial and lateral retinaculum (e.g. the medial and lateral patellotibial bands) are also described, but are of less clinical relevance. The soft tissue envelope (including the patella) is known collectively as the ‘extensor mechanism of the knee’. Recently it was realized that the distal femoral sulcus may develop abnormally resulting in trochlear dysplasia.

The patella is a sesamoid bone that acts as a marker for the alignment of the extensor mechanism. The trochlear groove and an arch of articular cartilage around the intercondylar notch make up the femoral side of the joint. Except in deep flexion, the tibial articular surface comes into contact with a different part of the femur to the patella and the majority of intercondylar notch osteophytes result from patellofemoral disease. The femoral sulcus acts like a pulley for the extensor mechanism. The forces exerted through the normal patellofemoral joint depend on activity such that at level walking it is half body weight, on going upstairs is three to four times body weight, and on squatting it is seven to eight times body weight.

The movements of the patellofemoral joint are complex. In full extension only the distal part of the patella articular surface is in contact with the femoral groove. It mainly rests on the synovium overlying the supracondylar fat pad. As flexion increases, the contact area on the patella sweeps proximally as a broad strip until 90-degree flexion, when the proximal part is in contact with the distal groove. From 90-degree flexion the odd (or extreme medial) facet articulates with the lateral edge of the medial femoral condyle, and the lateral facet articulates with the medial edge of the lateral femoral condyle. The medial facet lies in contact with the synovium overlying the anterior cruciate ligament. In deep flexion the patella effectively bridges the intercondylar notch. At 135 degrees of flexion the patella articulates with parts of both the medial and lateral femoral articular surfaces that also come into contact with the anterior meniscal horns (Figure 8.10.2).

Wherever there is uncovered articular cartilage, a synovial fold, usually enveloping fat, covers the articular surface. These folds include the peripatellar fringe of synovium, the supracondylar fat pad, and the infrapatellar fat pad with its alar folds. In extension, the supracondylar fat pad has a distal leading edge (abolished at arthroscopy by the introduction of fluid) which descends into the femoral sulcus as the knee flexes. It moves proximally as the knee extends, aided by the intra-articular negative pressure, and the action of the articularis genu muscle. As the patella articular surface contact sweeps superiorly during flexion, the synovium of the infrapatellar fat pad covers the exposed part. Specifically the alar folds point proximally in the midline. Beyond 90-degree flexion the alar folds move away from one another. It should be noted that the articular surfaces are swept by synovium during knee movement. This is probably important for cartilage nutrition.

Standard plain radiographs are adequate as a screen for patellofemoral problems. The anteroposterior (AP) standing view gives the least information. The lateral view should have the posterior femoral condyles strictly overlapping. The trochlear groove can then be properly assessed. Figure 8.10.3 shows the normal groove. The skyline view (tangential patella) may be taken in various degrees of knee flexion. The sulcus angle is useful as a screening tool. If greater than 140 degrees, this is associated with patellar instability. The angle changes according to knee flexion (Figure 8.10.4) and therefore is not useful for research purposes, as this is difficult to control.

Computed tomography (CT) is helpful, and shows the shape of the subchondral bone. It can measure the rotational align-ment of the lower limb as well as the patellar tilt angle, and tibial tubercle–trochlear groove (TTTG) distance (Figure 8.10.5). The TTTG is more useful than the Q-angle as it measures where the patella should be, rather than where it is. The normal TTTG is 10mm.

Magnetic resonance imaging (MRI) is very useful. It gives information on the chondral shape, as well as bony and soft tissue abnormalities, not only within the patellofemoral joint, but also the tibiofemoral joint.

Isotope bone scanning can be useful in unexplained anterior knee pain, and can confirm abnormal metabolism within the bone.

Anterior knee pain is the term used to describe pain in and around the patella. Until the 1970s, the term ‘chondromalacia patellae’ was synonymous with patellofemoral pain, because softening was noted on the undersurface of the patella. This term has now been replaced by ‘anterior knee pain’, which is a symptom and not a diagnosis. Other terms used are: patellofemoral syndrome, patellofemoral pain syndrome, patellofemoral joint syndrome, and extensor mechanism disorder.

There are two principal views as to the aetiology of anterior knee pain. The first is patellar malalignment with respect to the femoral sulcus. This causes localized peak loading of the articular cartilage which results in chondral damage and pain. Correction of the malalignment should therefore decrease the patient’s symptoms by offloading the affected area. However, there is poor correlation between articular cartilage lesions and pain. Patellar malalignment may only explain some patellofemoral pain symptoms.

Until recently it has been suggested that anterior knee pain may be due to a wide range of physiopathological processes such as an increase in intraosseous pressure and increased bone remodelling. However, the most recent view is that the tissues of the patellofemoral joint are in balance at a molecular level (tissue homeostasis) and supraphysiological mechanical loading and chemical irritation of the nerve endings has a direct effect on unbalancing this. An inflammatory cascade then occurs resulting in peripatellar synovitis. The peripatellar synovium is richly innervated and has been shown to be extremely sensitive to light touch. Once the synovium is inflamed, it will be continually aggravated by activities of daily living, resulting in prolonged symptoms.

However deciding what causes the pain and which structure is involved is still controversial. Many patients with marked patellar malalignment never experience pain, whilst others, with apparently no malalignment, or changes to their activities, experience problems. Disorders that can present as anterior knee pain are outlined in Box 8.10.2.

Pain felt at the front of the knee is common in adolescence, affecting both a third of boys and girls. Of these, a third of the girls and a tenth of the boys present to primary care physicians. Adolescent anterior knee pain should not be considered a disease. In this group it is important to confirm that there is no abnormality in the knee. Despite symptoms, examination of the knee is normal. Specifically there is no effusion, the vastus medialis obliquus is present and functioning, there is no patellar apprehension, the patellar tracks normally, there is no evidence of meniscal pathology, the ligaments are intact, and the patient can squat. Variations and abnormalities that may be found are: hypermobility syndrome, spine or hip pathology, tight musculature, and rotational or postural abnormalities of the lower limb. The long-term prognosis is good.

The commonest cause of instability of the patella is poor muscle power and/or coordination that results in a functional instability. Various anatomical abnormalities may be present that increase the risk of patellar dislocation, but if functional stability can be achieved, the surgical correction of the abnormalities is unnecessary. This is similar to the situation found in anterior cruciate ligament rupture.

Patellar dislocation is the extreme of a continuum of malalignments of the extensor mechanism which include a tight lateral retinaculum with a lateral tilt of the patella, and subluxation of the patella. The displacement is laterally. Medial displacement of the patella is typically iatrogenic. It is important to note that although the patella may be subluxated or dislocated, the problem is not the kneecap. Various factors may lead to patellar instability:

The single biggest risk factor for recurrent patellar dislocation is a positive family history.

The incidence of patellofemoral dislocation is reported to be 5–43 per 100 000. The factors associated with dislocation are listed in Box 8.10.4. Of those patients who present with a primary dislocation, about 17% will go on to have a recurrence. Of those who present with a dislocation, 50% have a history of previous dislocation or subluxation. However, in a recent randomized controlled trial the only factor that was associated significantly with recurrent dislocation was a positive family history.

There are many ways to classify joint dislocation/subluxation. In the patellofemoral joint it is important to appreciate the difference between a dislocation in a joint with pre-existing normal anatomy, from one where the anatomy is abnormal. To dislocate a normal patellofemoral joint requires significant force to be applied to the medial side of the patella (typically in a tackle in football). Because the femoral sulcus is normal, the risk of an osteochondral fracture is high. If an osteochondral fracture occurs, prompt treatment is needed if long-term disability is to be avoided. These can be considered as a true ‘traumatic’ dislocation. Those with abnormal anatomy can be considered to sustain an ‘atraumatic’ dislocation. This is not to say that there has not been an injury, but usually there is no direct blow to the patella, or it has occurred with minimal force. Atraumatic dislocators have a low risk of osteochondral fractures. In the acute setting, examining the opposite knee will show whether there are anatomical abnormalities, as these are bilateral.

Extensor mechanism maltracking may also be classified according to whether the patella dislocates in extension or flexion. The former is more common. The latter is always due to a tight lateral retinaculum, and usually tight quadriceps. Both have to be addressed if surgical correction is anticipated. The terms ‘habitual’ and ‘obligatory’ are also used. They can be defined as voluntary dislocation in the former and involuntary in the latter although they are frequently used interchangeably.

Dejour classified extensor mechanism malalignment into major, objective, and potential patellar instability. Major is when dislocation has been documented and is associated with significant abnormalities, objective is when instability symptoms are associated with abnormalities, and potential is in the asymptomatic patient with anatomical abnormalities. He has a further group called ‘painful patellar syndrome’ where there is pain, but normal anatomy. This is the group where operative intervention is unlikely to be helpful. Abnormal anatomy can be stretched medial structures, especially the medial patellofemoral ligament (hypermobility syndrome), patella alta, or trochlear dysplasia. A combination of these is typical in major and objective patellar instability.

In the last 10 years the importance of the shape of the femoral sulcus in patellofemoral pathology has been realized. Abnormalities in shape can lead to pain, and also increase the risk of an unstable extensor mechanism. The normal femoral sulcus has an angle of under 140 degrees, when measured from the subchondral bone. It is present at birth as cartilage, and continues into adulthood. Dysplasia may occur with overgrowth of the distal physes, which may abolish or, in extreme cases, reverse the angle. It is most important in the first 20 degrees of knee flexion, and may not be obvious on standard skyline films because they are taken at more than 30 degrees flexion. Dejour classified trochlear dysplasia on the plain lateral x-ray into types I, II, and III (Figure 8.10.6). The normal groove runs smoothly into the anterior cortical line of the femur. If it passes anterior to this, crossing the medial condylar line to reach the lateral condylar line anterior to the anterior condylar line, then this shows trochlear dysplasia. If the patella tracks normally, it raises the joint reaction force.

It should be noted that the groove in the dysplastic trochlea lies medial to the normal groove position. CT scans can also measure the patellar tilt angle, and the TTTG distance. If the TTTG distance is greater than 20mm then either the tibial tubercle may be moved medially 10mm or the trochlear groove laterally to correct it to normal. The former is achieved by a tibial tubercle osteotomy, and the latter by trochleoplasty. A trochlear boss of greater than 6mm and a Dejour type III dysplasia would indicate that a trochleoplasty may be beneficial.

Rarer is a lateral condylar hypoplasia. The groove line does not pass anterior to the anterior femoral cortical line, and neither does the lateral condylar line. There is therefore no lateral buttress to resist lateral translation of the patella secondary to the Q-angle. This can be corrected by the Albee trochleoplasty, which elevates the lateral femoral condyle. If performed for hyperplasia, it precipitates patellofemoral osteoarthritis (PFOA).

It can therefore be seen that patellar instability is associated with anatomical abnormalities. Figure 8.10.7 shows the differences in trochlear shape, patellar position, length and thickness of the medial structures, most notably the medial patellofemoral ligament (MPFL), and also indicates laxity or tightness in the lateral retinaculum. Note that the lateral retinaculum is lax in the figure, and therefore lateral release is not appropriate.

This is a genetically-based disorder of collagen covering a wide variety of genetic abnormalities from the extreme of Marfan’s and Ehlers–Danlos syndromes to benign joint hypermobility syndrome. Screening involves assessing the Beighton score. This tests joint flexibility. True hypermobility syndrome is assessed using the Brighton criteria (Box 8.10.5).

Hypermobility syndrome is diagnosed in the presence of two major criteria/one major and two minor criteria/four minor criteria. Two minor criteria will suffice where there is an unequivocally an affected first-degree relative.

Patients who are hypermobile may experience pain that will not be improved by an operation. Given that they have very elastic tissues, the folly of performing a lateral release becomes evident in this group. Operations for instability have a poor outcome, mainly because the operative repair stretches over time. However a medial patellofemoral ligament reconstruction with a hamstring graft shortens the lever arm enough to stabilize the patella. Lateral release should not be performed. Because of the tissue elasticity, it is not usually necessary to perform a trochleoplasty for a shallow or flat dysplasia, since there does not seem to be a significant increase in the joint reaction force after stabilization.

Surgical interventions to correct patellar maltracking are traditionally divided into proximal realignment (lateral release and medial reefing) and distal realignment (Roux–Goldthwait and tibial tubercle transfers). The traditional management is a lateral release, combined with a double-breasting medial reefing proximally, and Roux–Goldthwait or medialization of the tibial tubercle (Elmslie) distally. This works and normally controls the dislocation. However, in the presence of a significant trochlear dysplasia, where there is no groove for the patella to track in, then the risk is that either the patella redislocates, or the increase in the joint reaction force caused by the patella tracking over the trochlear boss leads to pain and later PFOA. Few patients partake in sports after this treatment, and the risk of later PFOA is higher than non-operated patients. It should be noted that operated patients, however, have had more symptomatic knees.

More popular in the USA is proximal realignment with a Fulkerson osteotomy. This moves the tibial tubercle anteromedially and therefore reduces the joint reaction force. This helps if there is a trochlear dysplasia.

Over the last 10 years newer surgical procedures have been introduced. The importance of the MPFL has led to a number of different variations to reconstruct or repair it. One method is a free hamstring graft anchored to the upper medial border of the patella, and inserted into the origin of the MPFL between the medial epicondyle of the femur and the adductor tubercle.

This is a proximal realignment and avoids the need to do other procedures, especially if the patient is hypermobile. MPFL reconstruction has the advantage over traditional proximal realignment procedures in that it can be performed as a day case, does not interfere with the quadriceps, and so allows earlier return to function, and does not need splintage.

In the last decade there has been increasing interest in addressing hyperplastic trochlear dysplasia by deepening the trochlea. A number have been described, but the consistent feature is that the excess subchondral bone is removed and the ‘boss’ lowered. The articular surface is not removed, unlike some older described excision or abrasion procedures. The trochleoplasty can be subdivided into those that create a flexible osteochondral flap and those that create a thick one (Figures 8.10.8 and 8.10.9). In the former the articular surface has to be normal. The flap or flaps created are then anchored to the subchondral bone by a variety of methods. The aim of the trochleoplasty is to remove the boss, recreate the lateral buttress, and lateralize the trochlear groove. It is hoped that it will avoid later-onset PFOA.

Acute patellar instability is managed conservatively, where the aim is to achieve functional stability with strong coordinated muscles. Traditional treatment with plaster casting for acute dislocation is falling out of favour to temporary orthosis and early physiotherapy. There is no evidence that this is improves or worsens the outcome but patients prefer it. Recurrent instability may not need surgical treatment as patients may develop coping strategies. If operative treatment is decided upon, then patellar tracking can be controlled by MPFL reconstruction alone. It is now felt that lateral release should be avoided, especially in hypermobility syndromes. This also may be unnecessary when a deepening trochleoplasty is performed as this slackens the lateral structures. Dysplasia may need to be addressed if the patella tracks over a significant trochlear boss.

PFOA affects an older age group than patellofemoral instability. Since the patellofemoral joint sustains the highest loads of any joint, it is no surprise that wear of the bearing surface is a common finding as age increases. However, symptomatic isolated PFOA is uncommon. It appears that about 5% of patients with radiographic changes of patellofemoral degeneration are actually symptomatic, which interestingly matches the anterior knee pain rate following unresurfaced patellae in total knee replacement. In a population of symptomatic osteoarthritic knees presenting over the age of 60 it was found that 19% of men and 17% of women had isolated patellofemoral disease as compared to 5% of men and 10% of women between the ages of 40–60 years. This emphasizes that failure of the bearing surfaces increases with age. It is useful to consider PFOA as present when both surfaces of the joint have exposed bone. Lesser articular cartilage loss or damage can therefore be considered as a chondropathy. However, it must be emphasized that radiological changes in the patellofemoral joint do not always equate with pain or loss of function. This means that a patient may have anterior knee pain and marked radiological changes, but the pain may arise from elsewhere, notably the hip.

It would be anticipated that trochlear dysplasia would lead to PFOA. The evidence for this is increasing and it is interesting that patellofemoral arthroplasty is most successful in this group. Pre-existing trochlear dysplasia is difficult to assess in patients with PFOA and bone loss.

The majority of patients with radiographic changes of PFOA (Figure 8.10.10) have no symptoms. If symptomatic, most will improve by building up quadriceps strength and coordination. However there is no evidence that any particular rehabilitation regimen is beneficial.

Clinical outcomes for the surgical management of isolated PFOA are sparse, and usually case series are not exclusive to isolated PFOA. Traditionally total patellectomy was the operative treatment of choice, but the outcomes were poor due to loss of extensor power, and continuing pain. More recently, excising any lateral osteophytes or part of the lateral facet (partial lateral facetectomy) with a lateral release has been reported with variable success (Figure 8.10.11). Tibial tubercle osteotomy moving the tubercle anteriorly (Maquet) has been done with generally poor results, and more recently the Fulkerson osteotomy has been used. However, as with osteotomies for tibiofemoral osteoarthritis, more certain outcomes seem to be had from joint arthroplasty. Recently figures for significant numbers of patients undergoing patellofemoral joint replacement with medium-term outcomes are encouraging (Figure 8.10.12).

The patella may present with pathology in the bone. Infections including tuberculosis, Paget’s disease, secondaries and primary bone tumours have been reported, but are rare. Management depends on the pathology and is not specific to the patella.

Osteochondritis dissecans (OCD) affecting the patellofemoral joint is described as ‘anterior’ OCD and is rare. It more commonly affects the patella than the trochlear groove. It is thought to occur secondary to trauma, and may lead to a loose body. Treatment is as for OCD in the tibiofemoral joint.

Considerable efforts have been expended on repairing localized damage to articular cartilage. The standard method is microfracture. Other techniques have been described includeing carbon-fibre patches (matrix-support prosthesis), autologous periosteal transplantation, osteochondral plugs (mosaicplasty), and autologous chondrocyte implantation, ACI). Observational case series give good results, but as the quality of the scientific methodology improves the results look less good when compared to microfracture, although the latter may be worse in larger lesions. As stated earlier there is poor correlation between pain and the site of any lesion. Many of the lesions are traumatic in origin and secondary to trochlear dysplasia. Attempts at repairing the articular surface without addressing the dysplasia are therefore doomed to failure. The poor results of ACI in the patella probably reflect this. There is currently interest in the use of stem cells to repair damaged articular cartilage, but this is in its infancy.

Pain arising from the soft tissue envelope is usually due to mechanical overload. In adolescence this may present as elongation of the tibial apophysis (Osgood–Schlatter syndrome) that may become symptomatic where the pathology is described as an osteochondrosis. The patients are usually sporty and in Osgood–Schlatter syndrome present with a prominent tibial tubercle that may be tender. Treatment is aimed at activity modification to reduce the symptoms to a level the patient can tolerate. Any pain should settle with rest and not keep the patient awake. If it does, and after excluding a tumour, a period resting in a cylinder cast may be tried. This enforces rest and discourages teachers from over-exercising the pupil.

Osteochondrosis of the distal pole of the patella (Sinding–Larsen syndrome) is rarer but behaves similarly. In adult life the residual effect is an elongated distal pole, and implies an enthusiastic athlete in adolescence

Older patients, typically over 30 years old, may develop pain in the patellar ligament, or less commonly in the quadriceps ligament, when it is described as a tendinitis. It is also called jumper’s knee and is similar to tennis elbow. The pathology is not inflammatory but a tendinosis. It is associated with tight musculature, typically the quadriceps. It reflects the loss of elasticity in the tissues with increasing age, thus causing stiffness. Typically it occurs in patients who are trying to maintain their athletic abilities and results in extensor mechanism overload. Treatment includes addressing the psychological aspects of the patient’s desire to stay young.

Tumours such as ganglia and osteochondromata, arising within the infrapatellar fat pad are rare and present as anterior knee pain. Clinically there is fullness present either side of the patellar ligament, compared to the opposite side, which is tender. More difficult to diagnose is a haemangioma as this does not cause a swelling. Diagnosis is usually made on MRI scanning, which is the ideal way to investigate fat pad pathology.

All these lesions are easily investigated using MRI, especially when there is a doubt in the diagnosis, and concern about infection or tumour.

Stress fracture of the patella can occur in osteoporotic bone or more rarely in normal bone. Transverse fractures may need surgical intervention. The risk seems to occur if there is a tight lateral retinaculum (with a tilted patella on skyline x-ray) and a significant force is applied as in jumping, or tripping over a dog. The result is an autolateral release. Excision of the fragment is advised. A childhood variant is a sleeve fracture with loss of the articular surface. This needs early surgical repair.

It may be difficult to distinguish this fracture radiographically from a bipartite patella. The bipartite patella may have more than one fragment. Typically the fragment lies superolaterally, and occurs in 1% of knees. On skyline view it may be non-fused part of the patella, or a separate fragment that overhangs the lateral condyle. If symptomatic, an MRI scan may show oedema in the fragment, or reveal the cause of the anterior knee pain.

The symptomatic patellofemoral joint is difficult to manage, principally because it is difficult to be certain that the pain felt by the patient is actually arising from the joint. The majority of symptomatic patients should be managed conservatively. Surgical treatment gives best results in patellofemoral instability with defined anatomical abnormalities that are then corrected.