Legg Calve Perthes Disease

Damian Mc Cormack,
International Fellow in Paediatric Orthopaedic Surgery,
Atlanta Scottish Rite Hospital,
Atlanta, Georgia.

Perthes disease remains enigmatic. Its cause is unknown. Its natural history is unpredictable. Its treatments are controversial. This article hopes to clarify what is know and what is thought about Perthes disease for those in training who seek a compromise between doing nothing for some good reason and doing something for no good reason. The article is again written in question and answer format for clarity and to maintain a logical approach and to perhaps encourage the reader to seek more questions before applying more answers.

1. What is Perthes disease?

Perthes disease is a condition of unknown etiology, which involves the immature hip joint and results in gradual collapse of part or all of the proximal femoral epiphysis with subsequent healing and remodelling of the femoral head and neck. In the child the initial disease process may cause synovitis with a limp and stiffness. The initial process is painless but pain may intervene later. In the adult symptoms may return as a result of degenerative arthritic changes in the hip. The final shape of the femoral head and neck dictates the clinical outcome. The better the head fits the acetabulum the less the risk is for later arthritic change. A spherical head in a spherical acetabulum is optimal but an egg shaped head in an egg shaped congruous acetabulum is also favourable. An incongruous fit or a subluxed or flattened head will have a poorer outcome. Because the capital epiphysis contributes to longitudinal growth of the femoral neck, the neck may ultimately be short but broad. The greater trochanter is unaffected and continues to grow. The adult deformity may include as a flat head, coxa plana, with a short neck, coxa brevia, and relative trochanteric overgrowth, further contributing to a Trendelenburg gait.

Therefore Perthes disease is a disease of childhood which has significant implications for the adult hip joint. The goal of treatment in the child is to minimise the risks of arthritic change in the adult.

2. Who first described Perthes disease?

Five names are associated with the description of Perthes. Author Legg from Boston described a case as " atypical tuberculous coxitis" in 1909. Jacques Calve encouraged his student, Paul Sourdat, to study tuberculous hips in Berck sur Mer, France, using his new X ray machine. Sourdat described several atypical cases in 1909 and Calve wrote further on 10 cases in 1910. Georg Perthes, working in Tubingen, Germany, described "arthritis deformans juvenilis" in 1910. Henning Waldenstrom described the radiographic features first, in 1908, in Stockholm but believed the process to be tuberculosis. The condition could therefore be called Waldenstrom Sourdat Calve Legg Perthes disease (fig 1).

What causes Perthes disease?

The cause is unknown. Before the use of radiographs Perthes disease was hidden under the gambit of tuberculous hip disease. Early suggestions for a cause included trauma, infection, and rickets. More recently it has become apparent that there are many factors which contribute to the ultimate degeneration and subsequent regeneration of the femoral head. Most recently Glueck has implicated a venous thrombolytic mechanism. His group have demonstrated deficiencies in protein C and protein S in patients with Perthes disease, as well as elevated lipoprotein A and hypofibrinolysis. These clotting mechanism disorders have not been confirmed. However similar femoral head collapse and fragmentation is seen in sickle cell disease and thalassemia as well as leukaemia, lymphoma and idiopathic thrombocytopenia.

Arterial occlusion is another potential cause. Inoue demonstrated that not one but two sequential infarcts were necessary to produce head necrosis in dogs. Studies with angiography and technetium scans have added some evidence to the arterial occlusion theory.

Many children developing Perthes disease are noted to be hyperactive. How this is linked with the disease is uncertain but the association seems real. Many patients with Perthes disease have delayed bone age compared with chronological age although this seems to accelerate towards normal by puberty. Full discussions on etiology are found elsewhere.

Who gets Perthes Disease?

Perthes disease can present at any time from infancy to maturity but most cases present between ages 4 and 8. Boys are affected 4 to 5 times more frequently than girls. It is not an inherited condition but there may be a family history in 2 to 20 percent. This may reflect environmental or social predisposition. It is reported to be commoner in urban areas and in certain geographic locations such as Japan, the Arctic( Eskimos) and central Europe. The classical candidate is the small hyperactive inner city boy!

How does Perthes disease present?

The classical presentation is an intermittent painless exercise related limp in a 4 to 8 year old. A traumatic incident often precipitates the visit to the doctor. If pain is present it may be episodic and centred in the knee or thigh as commonly as in the groin. Examination reveals limitation in abduction and internal rotation in an otherwise well and sometimes hyperactive child. The limp is more apparent after exercise or late in the day and is often minimized by the child in the examination room. The limp may be a Trendelenburg limp with a positive Trendelenburg test or may more resemble an antalgic limp. Atrophy of thigh muscles is often noted and spasm may restrict hip movement during the examination. More severe cases may loose so much abduction that they present with an adduction contracture and occasionally a flexion contracture at the hip joint. Rotation is lost before flexion and extension in these cases.

What else might cause a painful limp and stiffness in a child?

Infection ( hip joint, femur or S.I. joint), transient synovitis ( in the younger child), trauma (apophyseal avulsions from the pelvis, traumatic hip dislocations, fractures), tumour (especially osteoid osteoma), slipped capital epiphysis ( in teenagers), other causes of osteonecrosis ( Sickle cell etc).

What are the plain radiological features of Perthes disease?

The child should be assessed with an a p and frog lateral radiograph of the hips and pelvis (without lead shielding). The findings depend on the stage of the disease process and can range from an almost normal appearance to gradual increase in density of the femoral head and lucency in the metaphysis to fragmentation of the ossific nucleus and partial or complete collapse of the head to eventual healing. These evolutionary changes have been documented and staged by Waldenstrom.

What are the Waldenstrom stages of disease progression in Perthes disease?

Stage 1: Initial phase of increased density:

During the first 6 to 12 months following the causative event, the femoral head appears slightly smaller and gradually increases in density. There may appear to be loss of density in the adjacent metaphysis. The first sign may be lateralization of the head in the acetabulum due to soft tissue swelling within the joint. The physeal plate is blurred. A subchondral fracture may be visible during this phase. The phase of increased density ends when lucencies begin to appear in the otherwise dense ossific nucleus. (Fig 5)

Stage 2: Fragmentation phase:

Thereafter the ossific nucleus fragments and collapses and this process continues over a period ranging from 2 months to 3 years ( mean 8 months). The extent of the head which fragments is critical to the final outcome. Commonly it is the central portion which collapses while the periphery maintains its integrity. If the lateral rim of the head is intact it will become the weight bearing portion and will protect the central portion from further collapse. If the lateral rim, or pillar, loses its integrity, then the entire head flattens. This point has been observed by many authors and is integral to all of the prognostic classifications but was specifically addressed by Herring. It will be further discussed later. In very mild cases there may be no fragmentation but resolution of the increased density. In mild cases the fragmentation may be limited to the anterior central aspect of the head and be visible only on the frog lateral view. The fragmentation phase ends with the appearance of new bone in previously lucent areas. (Fig 3 and Fig 4)

Stage 3: Reossification phase:

During the next 1 to 10 years (average 4 years) the head reossifies. Radiolucent areas in the head begin to ossify and become dense. Reossification begins centrally and works outwards. Thus bone begins to appear at the periphery and the altered shape of the head emerges. During this healing phase the shape of the femoral head may improve gradually. (Fig 6)

Stage 4: Healed phase:

This refers to the final shape of the head and neck. In fact there may be considerable time between the end of reossification and the end of growth. Therefore following reossification there is no further new bone deposited but the head shape may continue to evolve until maturity.

What is Waldenstrom's sign?

In about 1/3 of patients a linear fracture is visible in the subchondral region of the femoral head, usually best seen in the frog lateral view. This occurs early in the initial phase just as the density of the head is increasing. It does not therefore mark the onset of the fragmentation phase.

Are there any radiographic signs outside the femoral head in Perthes disease?

In the metaphysis lucencies can appear early on and may represent true metaphyseal cysts in some but are probably physeal cysts in the majority. The "Sagging Rope Sign" in the metaphysis is a dense curved line which represents the projected posterior rim of the head. Premature physeal arrest may occur although the growth disturbance is not due to bridging of the plate. Trochanteric overgrowth and head /neck redirection may result. "Bicompartmentalization" of the acetabulum refers to the often transient adaptive change which may occur in the acetabulum to accommodate the deformed head.

So what are the possible final radiographic appearances of a hip following Perthes disease?

A spectrum of deformity may occur ranging from a normal looking round head and long neck to slight flattening of the head with a congruent acetabulum to a very flat head ( coxa plana) with an incongruent acetabulum. There may be growth disturbance and premature arrest of the physis with relative over growth of the greater trochanter. This spectrum of deformity has been classified by Stulberg and will be discussed later.

What other conditions can give these radiological appearances in the proximal femur?

Epiphyseal dysplasias (Spondylo epiphyseal dysplasias "SED" dwarfism, Multiple epiphyseal dysplasia or "MED" dwarfism, haemoglobinopathies, endocrine disorders (hypothyroidism), several syndromes(Schwartz Jampel, Tricho Rhino Phalangeal syndrome), post traumatic osteonecrosis and several other rarities. Bilateral Perthes disease is asymmetrical, one side beginning months or years before the other. Therefore symmetrical changes suggest another diagnosis. (Fig 2)

Are there any other ways to assess Perthes disease?

In certain circumstances further imaging modalities are useful. Arthrography will demonstrate "hinge abduction", due to lateral extrusion of the head with impingement against the acetabular rim with abduction. MRI will allow for early diagnosis of Perthes but this is of no real therapeutic value. It may allow better 3 dimensional assessment of the head in the future. CT scans will be of use in the rare patient with mechanical locking of the joint. Occasionally loose fragments occur in the joint. Ultrasound will demonstrate the early joint effusion. Scintigraphy and SPECT scans will establish an early diagnosis.

What is the natural history of Perthes Disease?

The natural history is unknown because there have been no studies of untreated patients with Perthes disease. Many patients have been treated conservatively but with very variable results. Some children have minimal radiographic changes and symptoms and very good outcomes. Others go through all the radiographic Waldenstrom stages described above and have symptoms for a year or more with gradual resolution. Some remain symptomatic into the late teens.

Does anything help predict the prognosis for the individual?

For an individual patient the most important factor is age of onset. Onset before age 6 produces a good result in the majority although a small subgroup will fare badly. Onset between 6 and 9 years fare moderately well. Onset after age 9 suggests a poor outcome. The other important factor is the radiographic appearance. This may be classified in a number of ways.

What are the plain radiographic classifications of Perthes disease?

Catterall's 4 part classification; types I, II, III, IV.
Salter and Thomson's classification; types A and B.
(Herrings) Lateral pillar classification; Types A, B and C.

What is the Catterall 4 part classification of Perthes disease?

Catterall proposed 4 groups of patients with Perthes disease in 1971. They are designated by Roman capitals. Groups I and II had a good prognosis (in 90%) and required no intervention. Groups III and IV had a poor prognosis (in 90 %) and required treatment. The classification is applied to the frog lateral and AP film during the fragmentation phase (i.e. when lucencies begin to appear in the previously dense head)

Catterall Group I: Involvement only of the anterior epiphysis (therefore seen only on the frog lateral film)

Catterall Group II: Central segment fragmentation and collapse. However the lateral rim is intact and thus protects the central involved area.

Catterall Group III: The lateral head is also involved or fragmented and only the medial portion is spared. The loss of lateral support worsens the prognosis.

Catterall Group IV: The entire head is involved.

Catterall's classification has a significant inter and intra observer error.

What is the Salter and Thompson Classification?

Salter and Thompson recognized that Catterall's first two groups and second two groups were distinct and therefore proposed a two part classification.

Salter & Thompson Group A: Less than 1/2 head involved.
Salter & Thompson Group B: More than 1/2 head involved.

Again the main difference between these two groups is the integrity of the lateral pillar.

What is the (Herring) Lateral Pillar Classification?

In 1992 Herring proposed a classification based specifically on the integrity of the lateral pillar. It is determined from the a p film only, at the start of the fragmentation phase.

Lateral Pillar Group A: There is no loss in height of the lateral 1/3 of the head and minimal density change. Fragmentation occurs in the central segment of the head.

Lateral Pillar Group B: There is lucency and loss of height in the lateral pillar but not more that 50% of the original (contralateral) pillar height. there may be some lateral extrusion of the head.

Lateral Pillar Group C: There is greater than 50% loss in the height of the lateral pillar. The lateral pillar is lower than the central segment early on.

Catterall also described a series of radiographic "head at risk of collapse" signs. What are they?

1. Gage's sign. a V shaped lucency in the lateral epiphysis.(? significance)
2. lateral calcification (lateral to the epiphysis) (implies loss of lateral support)
3. lateral subluxation of the head.(implies loss of lateral support)
4. A horizontal growth plate.(implies a growth arrest phenomenon and deformity)

How can one compare and classify the final outcome in Perthes at the end of growth?

The final shape of the head may be compared to a perfect circle using the Mose template and both a p and lateral images.

When the head contour is within 1 mm of a given circle on both views the result is deemed good.
When the head contour is within 2 mm of a given circle on both views the result is deemed fair.
When the head contour is greater than 2 mm off a given circle on both views the result is deemed poor.

Given that a congruous but aspherical head can perform well suggests that the Mose criteria are too strict and impractical.

What is the Stulberg Classification?

Stulberg proposed a 5 part classification of final outcome in Perthes which is of more practical value than that of Mose.

Stulberg I
The head is essentially normal as is the acetabulum. (spherical congruency)

Stulberg II
The head is within 2 mm on Mose' template in both views but is large ( coxa magna) or the neck is short. (spherical congruency)

Stulberg III
The head is more than 2 mm from circular, i.e. oval shaped but not flat, but congruous within a deformed steep acetabulum. (aspherical congruency)

Stulberg IV
The head is significantly flattened (undefined by Stulberg) but again is congruent in an equally deformed acetabulum (aspherical congruency)

Stulberg V
The head is flattened but the acetabulum has not accommodated to it and thus there is "aspherical incongruity". This carries the poorest prognosis and is seen with late onset Perthes.

How may Perthes disease be treated?

  1. Symptomatic treatment.
  2. Non operative "containment" of the head within the acetabulum.
  3. Surgical containment" procedures.
  4. "No treatment" is sometimes offered as an option by those trying to escape further discussion! However the unpredictability in the natural history for the individual demands periodic observation at least. Many children have periods of pain and stiffness which deserve simple symtomatic treatment. Patients with Perthes disease and relative trochanteric overgrowth may benefit from trochanter transfer and those with symptomatic loose bodies in their joint may benefit from their removal.
What are the possible symptomatic treatments for Perthes disease?

If a child has pain and joint stiffness then rest and NSAIDs may afford relief. Gentle traction in bed will relieve muscle spasm and pain and help to return a reasonable abduction range. " Slings and springs" traction rests the joint in a flexed position and allows active movement and is very effective. The sudden pain and synovitis which seem to occur around the time of initial fragmentation respond well. Short term crutch use is indicated but the old concept of prolonged non weight bearing in bed or on crutches is no longer supported.

What is the rational behind "containment" of the femoral head in the acetabulum?

Salter has demonstrated the " biologic plasticity" of the femoral head in pigs following a vascular insult. Containment of the head within the acetabulum is reported to encourage spherical remodelling during the reossification and subsequent phases. However if there is total head involvement and the lateral pillar collapses then the effect of containment is probable less. Therefore it seems that the extent of involvement of the head is the critical factor and containment simply optimizes the situation.

How can the hip be contained non surgically ?

Abduction encourages the head to travel beneath the lateral acetabular rim. Abduction is restricted in hips with moderate to severe disease and therefore abduction traction and adductor tenotomy may regain some movement. A Petri abduction cast may be placed under general anaesthesia and will hold the position. If abduction comes easier to the child then an abduction brace, such as the Atlanta Scottish Rite abduction brace may be used. It's use remains controversial, reflecting the difficulty in performing controlled studies in Perthes disease. Its advantage is that it allows ambulation and thus maintains motion in the hip. Bracing is continued until new bone formation is seen ( reossification phase).

What are the surgical options for containment?

A femoral subtrochanteric varus osteotomy will encourage the head to sit deeper in the acetabulum. It should optimally be done before the reossification phase when "biological plasticity" is greater. Extension and derotation may be added. Trendelenberg gait may worsen following varus osteotomy.

A pelvic osteotomy, such as that of Salter, will increase antero lateral coverage of the head. The hip should have a good range of motion before the procedure and have minimal deformity. He suggested proceeding only in those with onset after age 6 years. Adductor and iliopsoas releases are necessary. Stiffness following early Salter's osteotomy seems to be a problem. Other procedures such as shelf osteotomies have been used in the past. Cheilectomy , or excision of the extruded lateral head segment has a poor reputation.

What other surgical procedures are occasionally required?

A valgus femoral osteotomy is sometimes performed in a flattened head with hinge abduction if the hip is congruous in adduction. Osteochondritic loose bodies may cause mechanical locking in the joint and can be removed by arthrotomy or occasionally arthroscopy.

Should the relative greater trochanteric overgrowth be addressed in Perthes disease?

The trochanteric overgrowth can be dramatic on radiographs but several studies have shown that a Trendelenberg gait does not always occur. If it does occur, and is significant, then trochanteric advancement may improve the gait. An alternative is to perform a trochanteric arrest at an earlier date but this assumes that the first statement will not apply to the particular child.

In very general terms who is to be treated and what are the treatment options in Perthes disease?

Despite all the complexity of the foregoing classification systems, etc, the management of Perthes disease can be summarized as follows.

For patients less that 6 years old the prognosis is good for the majority. If they are stiff or painful they respond to bed rest, traction, pain relieving anti-inflammatory medication and rest. There is no evidence that abduction splints or surgical intervention is warranted in the majority of these younger patients.

For patients between 6 and 8 years but with a bone age less than 6 and an intact lateral pillar (Herring A and B) the prognosis is similar to that for the first group and observation is as good as surgical intervention for the majority. If they have bone ages greater than 6 years and Herring lateral pillar classification B then "containment" of the head within the acetabulum seems to be warranted. This may be done by abduction bracing, femoral varus osteotomy or a pelvic osteotomy.

If they are between 6 and 8 and are in lateral pillar group C then the result of intervention are equivocal. Children presenting with Perthes disease at age 9 or older often have lateral pillar B or C and a poor prognosis. The trend is towards early containment of these hips although stiffness can be a problem following early pelvic (Salter's) osteotomy.

What else?

The AAOS are currently conducting a multicentred study of Perthes disease in an effort to clarify optimal treatment protocols. The study is being coordinated by Dr Herring in Dallas. To date there are no indications as to the final conclusions. The trend is towards more aggressive surgical containment.

Figure legends.

Fig 1.

    Legg , Calve and Perthes.
    Waldenstrom and Calve's student Paul Sourdat also described the first cases of Perthes disease.

Fig 2.
    Bilateral Perthes disease.
    Note the asymmetry of the appearances in each hip indicating different stages of the disease. Epiphyseal dysplasias and haemoglobinopathies tend to have symmetrical changes.

    Note also lateral extrusion of the head with widening of the medial joint space and fragmentation of the physes. Reossification is advancing on one side. The lateral 1/3 of the head or " pillar" is intact and protects the central portion from further deformation.

Fig 3. (Perthes 1)
    Unilateral Perthes disease in the fragmentation phase, demonstrating medial joint widening, lucency of the middle 1/3 of head segment and loss of lateral pillar integrity.
Fig 4. (Perthes 2)
    Unilateral Perthes with entire head involvement and fragmentation. The reossification phase has not yet begun.
Fig 5 (Perthes 3)
    Unilateral Perthes disease with widening of the medial joint space, blurring of the physis, increased density of the head and lucency between the medial and central 1/3's of the head corresponding to early fragmentation phase.
Fig 6 (Perthes 4)
    Unilateral Perthes in the reossification phase with a visible subchondral line similar to Waldenstrom's sign. However Waldenstrom's subchondral fracture is seen very early in the disease process, before fragmentation. In this case the lateral pillar has maintained some integrity.

The following references are suggested as an introduction to the vast literature on this subject.

Herring J A. Current Concepts review. The treatment of Legg Calve Perthes disease. A critical review of the literature.
JBJS. 1994; Vol 76 A: No 3, 448 458.

Herring J A, Neustadt J, Williams J, Early J, Browne R.
The lateral pillar classification of Legg Calve Perthes disease.
JPO. 1992; 12: 143 150.

Herring J A. Management of Perthes disease ( editorial).
JPO 1996; 16(1): 1 2.
Weinstein S. Legg Calve Perthes syndrome.
In Paediatric Orthopaedics, 4th ed. 951 991.
Editors Morrissy and Weinstein. Lippincott Raven.