Perthes’ disease is an idiopathic avascular necrosis within a developing femoral head.

It was first described in 1910 by 3 independent observers - Arthur T Legg (USA), Jacques Calvé (France) and Georg Perthes (Germany). An extensive description was written in 1913 by Perthes, hence the condition widely bears his name alone.

Perthes’ disease has an annual incidence in the UK of 7.5 per 100,000 0-14 year olds per year. This equates to a childhood risk of 1 in 1000. It is 4 times more common in boys, with 80% of cases occurring between the age of 4 and 8 years old.

Right sided Perthes’ Disease. The radiograph demonstrates coxa plana and sclerosis of the right hip.

The incidence of Perthes disease varies widely across the UK. There is twice the amount of Perthes’ disease in the Northern parts of the UK, compared to the Southern parts. The disease appears most common in Northern Europe, with a low incidence of disease in India, and yet lower in equatorial regions. These differences exist even after adjusting for the differences in race.  

UK Incidence of Disease. Reproduced with permission from Perry et al. Legg-Calve-Perthes Disease in the UK: Geographic and temporal trends in incidence reflecting differences in degree of deprivation in childhood. Arth Rheum. 2012; 64: 1673 – 79.

Deprivation Graph, demonstrating three times the incidence of disease in the most deprived quintile of deprivation, compared to the most affluent. Reproduced with permission from Perry DC et al. Perthes Disease: Deprivation and Decline. Arch Dis Child. 2011; 96(12): 1124 – 28.

There are many risk factors that have been suggested, but the only one that is consistent and very strong is socioeconomic deprivation. At least in Europe, Perthes’ disease is intertwined with socioeconomic deprivation, with the risk of disease being three times greater in the most deprived groups, compared to the least deprived. This is strong evidence of an environmental disease determinant, with very little to suggest that there may be any genetic disease component.

Other risk factors that have been suggested include:

• Tobacco Smoke/ Wood Smoke Exposure (Good Evidence of an association, but this may simply be reflecting the socioeconomic deprivation)
• Congenital genitourinary abnormalities (Good Evidence of an association)
• Abnormal growth and small stature (Good evidence of an association).
• Hyperactivity and behavioral abnormalities (Good evidence of an association, though this may be a consequence of disease treatment).
• Manganese deficiency (Equivocal evidence of an association)
• Prematurity (Good evidence of no strong association)
• Low birth weight (Good evidence of no strong association)
• Genetics (Good evidence of no strong association)
• Heart Disease in adulthood (Early evidence of an association).

Several theories have been suggested to explain the mechanism of disease. The most widely favoured is an interruption to the lateral ascending vessel, which is the main vascular supply to the hip in children. Several suggestions have been offered to how this may occur:

Intravascular – hypercoagulability may result in thrombosis of vessels. There is however fair evidence to suggest that there is no association (i.e. Protein S deficiency, Protein C deficiency or Factor V Leiden deficiency)

Vascular – abnormalities of structure of the lateral ascending branch. It has also been suggested that there may be a generalised vascular dysplasia in affected children underpinning both disease, and the resultant growth failure.

Extravascular – repeated episodes of intracapsular pressure. A link with transient synovitis has therefore been proposed, though the evidence to support an association is very weak.

This can occur in around 15% of patients. It should not be at the same stage simultaneously. Patients who present in this way should have alternative diagnoses sought, which include multiple epiphyseal dysplasia, spondyloepiphyseal dysplasia, gauchers disease, sickle cell disease or steroid use.

Bilateral Perthes’ disease – this should prompt consideration of an alternative cause. Bilateral disease is typically asynchronous, such that the disease exists at different stages in either hip. Perthes disease of the right hip was treated at 6 years old. 7 years later the patient presented with left hip pain.

The disease stages were described by Waldenstrom in 1909 (although he believed that the underlying cause was tuberculosis). The stages are still recognised and used in modern practice. The ossific nucleus experiences a temporary insult, leading to cell death. This then resolves, with the necrotic bone being resorbed and replaced with new bone.

The stages of disease, as described by Waldenstrom, are widely misquoted. The true classification, according to the original description, is detailed below:

Initial – The ischaemic insult has occurred, causing bone to become necrotic. There is joint space widening (Waldenstrom’s sign) but little else to see on plain radiographs. The joint space widening occurs owing to cessation of growth of the bony epiphysis, yet continued growth of the cartilaginous tissues. Radiographs demonstrate a widened joint space, but little else.

Picture of Waldenstrom’s stages 

Sclerosis - The ossific nucleus becomes sclerotic. The altered bony architecture also causes flattening/ collapse of the femoral head. The sclerosis is in part to collapse, but largely due to calcification of the necrotic marrow.  Radiographs demonstrate flattening, and a small sclerotic nucleus.

Fragmentation – The necrotic bone undergoes resorption to be replaced by fibrocartilage. only occurs after reossification has firmly ensued. Radiographs demonstrate fragmented appearance of the femoral head.

Late Phase/ Healing – Creeping substitution occurs, with fibrocartilage being replaced with new bone. The malleable femoral head once again develops a firm bony-structure. Radiographs demonstrate coalescence of the femoral head, with the central aspect of the femoral head being the last to reform.

Containment is concerned with trying to prevent deformity of the hip, by maintaining the shape within the acetabulum. The femoral head loses is its architecture during the sclerosis stage and ‘softens’. The architecture is not regained until the late phase. The principle of containment is to ensure that the femoral head remains firmly within the confines of the acetabulum throughout the ‘soft phases’ – like an ice cream kept in a scoop. Once the late phase occurs the shape has been determined, and no significant remodeling occurs (despite the widespread misconception that there is a late remodeling phase).

1. Early stage-disease (pre-collapse);

The internal bony architecture of the head is deficient and is subject to plastic change. Early changes in head shape may be modified during this period. The principle is to recognise and treat cases in which the femoral head is not contained, to promote the development of a spherical, congruent femoral head.

2. Intermediate stage disease (post collapse, but pre-healing);

It is believed that little can be done surgically to alter the disease process at this stage.

3. Healed disease;

Irreversible changes have now occurred. The principle is to treat the consequences of suboptimal femoral head shapes;

1. Painful hinge abduction.
2. Femoro-acetabular impingement.
3. Degenerate change.

As always, classifications are limited by their applicability to practice. There are several described for Perthes’ disease.

Herring – Describes the degree of collapse of the lateral column;

1. No collapse
2. Collapsed by less than 50%
3. Collapsed by more than 50%

In theory, the lateral column is a key structure, acting as a strut to withstand load bearing. When it is involved, the remaining areas of the femoral head are exposed and also collapse. This relates well to prognosis and is reproducible, but has limited clinical application, because it is only defined at the end of the fragmentation phase (i.e. post-collapse). It has been demonstrated that the optimal time to intervene is pre-collapse.

Herring’s Classification

Catterall – This relates to the appearance of the hip on the lateral radiograph. This was described before Herring’s classification and also relates well to prognosis. This is also unable to be determined until post-collapse. A somewhat complex classification that can broadly be simplified by splitting the femoral head into quarters, and knowing that the anterior quarter is always involved.

Catterall’s classification

Salter Thompson – Describes the presence of a subchondral lucency during the initial phase. This can be likened to the ‘crescent sign’ seen as a marker of impending subchondral collapse in adult avascular necrosis. If the lucency involves more than 50% of the femoral head, there is an ultimately greater resorption of the head and a worse prognosis. Unfortunately, only 1/3 of patients have a lucency of any size and thus this classification is also limited in its applicability, though is the only one that can be used to inform treatment (as it occurs pre-collapse).

Picture of subchondral lucency of Salter Thompson. This is greater than 50% of the femoral head, suggesting a poor outcome.

Stulberg – Has described the shape of the femoral head at completion of disease. This classification is useful for long-term prognosis, with respect to the risk of osteoarthritis, and can be modified into a more simple guide for routine practice;

1. Spherical, congruent head – carries the same risk for osteoarthritis as the general population
2. Aspherical, but congruent head – Likely to develop osteoarthritis above 50 years old.
3. Aspherical and incongruent head – likely to develop premature osteoarthritis below 50 years old.

Picture of Modified Stulberg Classification

Early Perthes’ disease commonly presents with a painful limp. Pain is localised to the groin, thigh or knee. Painful internal rotation is the earliest sign. The patient may subsequently develop decreased abduction. Restricted range of motion is an important guide for treatment, and should therefore be examined closely.

A child presenting with acute groin, thigh or knee pain should be assumed to have significant hip pathology and promptly investigated.

Plain radiography – Radiographs will demonstrate changes consistent with the Waldenstrom stage.

MRI – Can be used to show marrow signal changes, and is therefore useful in early diagnosis if radiographs are not diagnostic. More recent work is centered on MR angiography and blood flow measurements, although this is experimental and currently offers no benefit with regards to planning treatment or prognosticating. Young children (<6 years) generally do not tolerate an MRI scan without a general anaesthetic.

Isotope bone scan – Does not require general anaesthesia and can be useful in initial investigation in the face of normal plain radiographs. The classic finding is a ‘cold spot’, in contrast to the increased uptake or normal images seen with many of the other differential diagnoses of hip pain.

An MRI of a child with early Perthes’ disease. The T2W image demonstrates decreased signal within the femoral head, and oedema around the physis.

Catterall identified radiographic signs of the head at risk for collapse. When viewed in the context of Herring’s classification, it is apparent that these signs are describing impending failure of the lateral column and are therefore of value. In some centres they are used as the primary prompts for surgical interventions (as the Herring or Catterall classifications predict the prognosis the end of fragmentation, when the optimal time to intervene has passed).

1. Calcification lateral to the epiphysis
2. Gage’s sign (A lucent area in the lateral epiphysis and metaphysis)
3. Lateral subluxation of the head
4. A horizontal physis

Catterall also described ‘the child who is not doing well’, represented by progressive loss of movement; this is the basis for the ‘containment theory’.

This demonstrates marked metaphyseal cysts, and lateral subluxation of the head.

Many children in the early stages of Perthes’ disease maintain good hip motion. The head displays plastic deformation at this stage. As movement is intact, the entire head spends time being molded by the acetabulum. This can be termed ‘passive containment.’ Preserved abduction of the hip (ensuring that pelvic movement is prevented) is generally used to guide whether a child is passively containing their hip.

In other children, the hip range of motion may be reduced, such that a portion of the femoral head (in particular the lateral column) does not enter within the acetabular dome. This is due to synovitis and pain, which results in adduction of the hip relatively ‘uncovering’ the lateral column.

This leads to abnormal loading forces and undesirable plastic changes. Many clinicians feel that children in this situation benefit from intervention to cover the head and allow more desirable molding. This can be achieved by surgically moving the femoral head or the acetabulum - this is termed ‘active containment.’

Historically, 2 years of bed-rest, spica casts, or one of various devices were used to brace the hip in abduction. These methods were cumbersome, and are less acceptable in modern times. Some units still advocate such treatment.

Active (surgical) containment – Younger patient

There are various treatment strategies. The authors feel that younger patients (typically under the age of 8) can be appropriately managed with either a varus femoral osteotomy or a salter osteotomy. The important pre-requisite for these procedures, which can be demonstrated by examination under anaesthesia and arthrography, is the absence of hinge abduction – If the femoral head has already undergone collapse, and a saddle has formed then either containment method would make this problem worse.

Varus osteotomy results in an immediate leg length discrepancy, with shortening of the involved limb. This shortening may improved the coverage of the hip as the pelvis dips over the femoral head. There is also shortening of the abductor moment arm and a resulting trendelenburg gait. In the younger child, there is usually sufficient remodeling potential for these issues to resolve, although they should be monitored and leg length inequality managed as appropriate. Some authors advocate trochanteric physeal arrest to prevent relative trochanteric overgrowth, which can lead to problems with impingement and the abductor moment.

A salter osteotomy is also a common means of containment. This operates on the ‘good side’ of the joint, rather than the damaged side. They provide anterior and lateral coverage, and therefore add containment to the areas most affected by Perthes’ disease. However, this osteotomy also lengthens the leg. This lengthening may have the effect to tilt the pelvis, which may conversely uncover the hip a little.

Radiographs demonstrating two of the common procedures used to treat early Perthes’ disease with surgical containment (Varus Osteotomy and Salter Osteotomy) 

Active containment – Older patient

In older patients (typically over 8), we feel that the potential for proximal femoral remodeling is much lower and as such, we prefer to perform active containment procedures on the acetabular side. There are 2 broad types of pelvic osteotomy useful in Perthes’ disease;

• Reorientation – A Salter osteotomy may be used in Perthes’ disease with good outcomes. This hinges on the pubic symphysis and provides anterior and lateral coverage.
• Salvage – The shelf osteotomy has been employed in Perthes’ disease with encouraging results.

Reshaping (volume altering) are the other common type of pelvic osteotomy, however this should be avoided in Perthes’ disease. Unlike other paediatric hip pathologies, the acetabulum is typically uninvolved. Altering its shape and reducing its volume may have detrimental effects on the plastic femoral head.

Active containment of an older child with Perthes, disease using a Shelf salvage procedure.

Bisphosphonate – The role of bisphosphonates is under evaluation

Arthrodiastasis – Joint distraction with an external fixator may unload the femoral head and alter internal pressures, promoting revascularisation and reossification/remodeling. There is little evidence thus far to guide the role of arthrodiastasis.

This is guided by the specific problems the patient presents with;

Hip pain – An examination under anaesthesia and arthrogram is useful to determine whether ‘hinge abduction’ is occurring, due to impingement of a deformed femoral head or overgrowth of the greater trochanter. The articular surface and adduction should be assessed and the ‘position of best fit’ determined, whereby the hip is held in a position where the articular surfaces are at their most congruent. It is then possible to perform;

• An osteotomy to hold in the position of best fit (often a valgus osteotomy, with the pre-requisite that adduction should be intact to avoid the ‘hip stuck in abduction’ and limited adduction).
• Arthroscopic or open procedures to remove impingement lesions or reshape the femoral head.

Radiograph of a hip with a saddle deformity causing marked impingement, and a leg clinically fixed in a few degrees of adduction. A valgus osteotomy was performed to change the functional arc of motion, enabling twenty degrees of abduction. 

There is a paucity of evidence that have sought to determine outcomes of surgery for Perthes’ disease.

There are two large-scale cohorts of disease, which have determined the outcomes in Perthes disease.

Herring (2003) in a multicentre US study suggested that surgery may offer improved results in children >8 years old, with a hip graded as Herring B.

Wiig (2007) in a mulitcentre Norwegian study, demonstrated that amongst those >6years old with >50% head involvement, the outcomes of surgery were significantly improved.

There are no randomised clinical trials offering an insight into the optimal treatments in Perthes’ disease.