The aetiology of Legg-Perthe's disease, or idiopathic avascular necrosis of the femoral head, is unclear. Less than 10 per cent of the cases are bilateral, and very few occur in multiple family members; thus, it probably is not an inherited condition. There have been a few twin studies, in which one member has the disease, and the other may or may not be affected.

Legg-Perthes' disease tends to afflict boys in a ratio of about 4:1 over girls, primarily occurring between ages 4 and 8. It may, however, appear in someone as young as 2 years old or as old as 12 years.

Part of the difficulties in understanding the aetiology and mastering the treatment stems from the lack of a good experimental model. It is possible to make the femoral head avascular in an immature animal. Blood vessels do not cross the physis after about the first year of life; the head is nourished only by the fragile vessels that line the femoral neck, the retinaculum of Weitbrecht. By destroying these external capsular-type vessels, the head dies, but the characteristic picture of Perthes' disease is not reproduced. Another approach has been to fill the joint with a non-permeable fluid that can exceed the capillary pressure and also lead to bone death.

The difficulty is that there is a characteristic pattern of Perthes' disease indicated by four stages. The first stage is presumably an inflammatory stage, when there are no bony changes, but widening of the distance between the femoral head and the centre of the acetabulum occurs, a widened ‘tear drop-head’ distance (Fig. 1) 2575. Normally, even with some rotation, the differences in this distance between the two sides, is only 1 mm. In toxic synovitis, which is an inflammation of the hip, often confused with early Perthes' disease, the difference rarely exceeds 2 mm. In Perthes' disease that distance is increased. Some experimental work suggests that this increase represents actual overgrowth of cartilage. Other work suggests it represents relative overgrowth, i.e., the bone is not ossifying, so that the distance to the bony epiphysis becomes greater. The disease may also represent some synovitis, since the children with widening usually have the most acute loss of motion; the widening is certainly more than relative, since the distance to the metaphysis is also increased.

The absence of a good experimental model that reproduces both the pathology and radiography of the disease hampers our understanding. In experimental studies, physiological fluids have not been able to generate enough pressure to occlude the circulation. Toxic synovitis, which is presumably a viral affliction, has not been found to be a precursor of Perthes' disease. There have been associations of between 1 and 10 per cent of synovitis preceding Perthes' disease but most workers feel that the last number represents early Perthes' disease that was not well documented. Trauma has been implicated; haemorrhage following an intracapsular hip fracture in a child could lead to avascular necrosis. The higher incidence in boys tends to favour a traumatic aetiology. However, those children who have serious intracapsular haemorrhage invariably have excruciating pain, and this circumstance is not seen in Perthes' disease.

The second finding in early Perthes' disease, i.e. stage 1, is usually the subchondral crescent sign, best seen on the frog lateral film (Fig. 2) 2576. While this finding is an early radiographic feature, we now recognize that it represents a fracture induced by revascularization. Presumably therefore, even though this is one of the earliest radiographic features, it represents a later stage in the disease, suggesting (and supported by core biopsies on femoral heads), that this disease represents a series of insults or infarcts. For this reason technetium bone scans may not be helpful at this stage. Since revascularization is already established, the scan may appear normal. Bone changes occur before the scan shows impaired vascularization.

The second stage of the disease, radiographically, was originally thought to be the avascular stage. The bone becomes dense, and may collapse. There are various theories for the increased density. All may be partially true. One holds that the avascular bone is not resorbed, but the surrounding bone becomes osteoporotic, and hence the density is relative. Others suggest the density is secondary to calcification of marrow fat, or to collapse and impaction of the dead bone. The most commonly held theory today is that it represents creeping substitution, with new bone laid down on trabeculae of old dead bone. This mechanism is consistent with the suggestion that revascularization occurs much earlier than suspected in this condition.

The third stage of the disease clearly represents an extensive revascularization, whereby a substantial part of the head is resorbed, and disappears. The fourth stage involves reconstitution of the femoral head.

Patients may have an ache in the thigh, groin, or knee. Most commonly, they will come to attention because of a limp, rather than discomfort. The limp is a lurch to the involved side, and there may be a positive Trendelenburg test, i.e., when the patient stands alone on the involved leg, the opposite pelvis droops, because the abductor muscle cannot support the hip. Motion of the hip may be almost absent, it may present as a loss of abduction and internal rotation, or it may have completely normal motion. The earliest radiographic findings of widening of the tear-drop neck distance, and the crescent sign on the frog lateral film have already been mentioned. About 10 per cent of patients have bilateral disease. It is, therefore, less common to have a bilateral presentation, in which case hypothyroidism, which gives irregular ossification, or one of the epiphyseal dysplasias should be considered. As mentioned above the bone scan may not be helpful, because of the presence of a revascularization stage. MRI may be more helpful in determining the extent of the disease, but its value has not yet been fully demonstrated.

The difficulty in agreeing on treatment regimens for these children stems from the fact that there is such great variability in the severity of the condition. Further, it is difficult to decide what is a good or bad result. Radiological grading systems, such as the Mose concentric circle technique, can show an abnormal femoral head although the clinical result is often quite good. Even with quite deformed femoral heads, there are few symptoms. Long-term studies have failed to show much development of osteoarthritis. The hips do not deteriorate severely until age 50 to 60 years.

Younger children, especially below the age of 4, do extremely well (Figs. 9 and 10) 2583,2584. Also, the less the femoral head is involved, the better the prognosis. Therefore a grading system has been established: Grade I represents slight involvement of the superolateral head, and Grade II represents about one-half of the head involved. Grades III and IV represent almost all of the head, and Grade IV, total involvement. With lesser degrees of femoral head involvement, there is sufficient of a ‘strut’ remaining and the femoral head does not flatten badly or extrude. Recent work suggests that the presence of an identifiable strut gives the best prognosis. There are many more refinements of this procedure, but the problem for the practitioner is that when a patient is first seen, despite bone scans and MRI, it is impossible to assess how much of the femoral head is involved; therefore, assigning no treatment to Group I and II patients, or considering that Group III and IV patients always have bad results if untreated is unacceptable.

Another approach has been to designate what features led to a poor prognostic output; ‘head at risk’ features were identified. Probably the only two features that stand out are calcification lateral to the epiphysis, and lateral displacement of the head. If the head is laterally displaced either the head is already large and deformed, or there is synovitis, pushing the head out of the socket. Similarly, lateral calcification suggests that there will be a very large head. Therefore, large femoral heads, with short necks that are flat, and not completely covered by the acetabulum give poorer results. Our knowledge of the factors that lead to arthritis suggests that if part of the femoral head, or of any joint, does not make contact with an opposing cartilage surface, the nutrition of the cartilage is affected and degenerative changes ensue. Therefore treatment should ideally be geared towards developing a normal round femoral head; if this is impossible, achieving a head that is moulded to the acetabulum, i.e. congruous, and making contact with the entire articular surface should be the aim. Failing these, having the head completely included in the acetabulum should be the aim.

Clearly, in determining treatment, it is necessary to remember that the femoral head cartilage contains a semicircular growth plate that is different from the physis, and is the plate responsible for longitudinal growth. The acetabulum also has a growth plate, that also responds to pressure, or can be moulded (Fig. 12) 2586. The cartilage of the head generally remains alive (probably nourished by diffusion from the synovial fluid), even if the ossific nucleus dies. Like all growth plates, this responds to pressure; increases in pressure slow growth. The younger the patient, the more growth cartilage there is to remodel. Even if a young patient has complete involvement of the femoral head, and the femoral head flattens, because there is much more growth to take place, the head and acetabulum will remodel to each other and produce a fairly good result by the end of growth.

Another important consideration is that of motion. A round femoral head means that there has been motion in many planes. An oval head suggests motion can take place in only one direction. If a child presents himself with Grades I or II involvement, and motion is good, and an arthogram or MRI shows that the head is round and well covered, probably no treatment is indicated. Regardless of the grade, if there is major restriction of motion, the child should be treated with 24-h traction with appropriate periods of relief, until the motion is restored. Some children have more spasms than others; some retain their motion throughout the disease, while others require night traction and may need prolonged periods of traction or crutch walking.

If an arthogram, MRI, or radiograph shows that the femoral head is flattening out or is uncovered by the acetabulum, there is a need for more action apart from making sure that motion is retained. Bracing or surgery can be employed. If the top of the femoral head is flat, the arthogram will indicate that the corresponding face of the acetabulum is also flat. If the hip is abducted, the flat part of the femoral head is now facing a round portion of the acetabulum (Fig. 13) 2587. Therefore, pressure is lessened on the dividing epiphyseal cartilage cells, which will grow faster, and the head is more likely to round out. Abducting the hip can usually cover the femoral head, thereby limiting lateral growth. However, if the hip is subluxed laterally, there is no acetabulum limiting or shaping the femoral cartilage; cartilage overgrows and becomes flat.

Increased abduction can be accomplished by a bilateral abduction brace, the Scottish-Rite brace (Fig. 14) 2588 being one of those more commonly accepted, or by varus osteotomy. There is little preference for either technique, but it is vital to maintain the motion. In the brace, if there is muscle spasm, the patient abducts the good leg more, and abducts the involved limb very little. In that case the treatment is worthless.

For surgery, we usually approach the femur laterally, through the fascia lata, and split the vastus lateralis, also taking down some of its origin from the vastus ridge. A dynamic compression plate is used, with the upper screws in the neck, and the plate is contoured to the degree of varus. By making a dome osteotomy, too much shortening is avoided. Patients are kept in a spica cast for 2 or 3 months.

There have been mixed reports of the value of achieving the same end by an innominate osteotomy. Salter has a good series of these patients, but is only using osteotomy when the head and the acetabulum are round. Obviously osteotomy can give improved coverage, but leaves the problem of a flat head unsolved (Fig. 16) 2590.

The older the child, the less cartilage there is to remodel. Sometimes doing an osteotomy in which a mature flat head is tipped into the socket leads to severe incongruity, pain, and reduced motion. The results in children over 10 years of age tend to be much poorer. For later uncovering of the head, a Chiari procedure can reduce the risks of arthritis. Other late complications with Perthes' disease include shortening and overgrowth of the trochanter since in severe cases, the physis may lose its blood supply. Stapling the opposite femur at the appropriate time should be considered. If it is early in the course of the disease, and it appears that the trochanter is overgrowing, the trochanteris physis can also have an epiphyseodesis. The child nearing maturity may need a trochanteric advancement.

Bluemm RG, Falke THM, Ziedees de Plantes BGJ, Steiner RM. Early Legg-Perthes' disease (ischemic necrosis of the femoral head) demonstrated by magnetic resonance imaging. Skeletal Radiol, 1985; 14: 95.

Bobechko WP. The Toronto brace for Legg-Perthes' disease. J Bone Joint Surg, 1976; 58B: 115.

Brotherton BJ. The long-term results of the treatment of Perthes' disease by recumbency and femoral head containment. J Bone Joint Surg, 1976; 58B: 131.

Caffey J. The early roentgenographic changes in essential coxa plana: their significance in pathogenesis. Am J Roentgenol, 1968; 103: 620.

Canario AT, Williams L, Weintroub S, Catterall A, Lloyd-Roberts GC. A controlled study of the results of femoral osteotomy in severe Perthes' disease. J Bone Joint Surg, 1980; 62B: 438.

Catterall A, et al. A review of the morphology of Perthes' disease. J Bone Joint Surg, 1982; 64B: 269.

Clarke TE, Finnegan TL, Fisher RL, Bunch WH, Gossling HR. Legg-Perthes disease in children less than four years old. J Bone Joint Surg, 1978; 60A: 166.

Coleman CR, Slager RF, Smith WS. The effect of environmental influence on acetabular development. Surg Forum, 1958; 9: 775.

Cotler JM, Donahue J. Innominate osteotomy in the 1980; treatment of Legg-Calvé-Perthes disease. Clin Orthop, 1980; 150: 95.

Curtis BH, Gunther SF, Gossling HR, Paul SWl. Treatment for Legg-Perthes disease with the Newington ambulation-abduction brace. J Bone Joint Surg, 1974; 56A: 1135.

Evardsen P, Slordahl J, Svenningsen S. Operative versus conservative treatment of Calvé-Legg-Perthes disease. Acta Orthop Scand, 1981; 52: 553.

Green NE, Beauchanp RD, Griffin PP. Epiphyseal extrusion as a prognostic index in Legg-Calvé-Perthes disease. J Bone Joint Surg, 1981; 63B: 900.

Herring JA, Lundeen M, Wenger DR. Minimal Perthes' disease. J Bone Joint Surg, 1980; 62B: 25.

Inoue A, Freeman MAR, Vernon-Roberts B, Mizuno S. The pathogenesis of Perthes' disease. J Bone Joint Surg, 1979; 58B: 453.

Kamhi E, MacEwen GD. Treatment of Legg-Calvé-Perthes disease. Prognostic value of Catterall's classification. J Bone Joint Surg, 1975; 57A: 651.

Klisic PJ. Treatment of Perthes' disease in older children. J Bone Joint Surg, 1983; 65B: 419.

Laurent LE, Poussa M. Intertrochanteric varus osteotomy in the treatment of Perthes' disease. Clin Orthop, 1980; 150: 73.

Legg AT. The classic: an obscure affection of the hip joint. Clin Orthop, 1971; 79: 4.

Lloyd-Roberts GC, Catteral A, Salamon PB. A controlled study of the indications and the results of femoral osteotomy in Perthes' disease. J Bone Joint Surg, 1976; 58B: 31.

McAndrew MP, Weinstein SL. A long-term follow-up of Legg-Calvé-Perthes disease. J Bone Joint Surg, 1984; 66A: 860.

Meyer J. Legg-Calvé-Perthes, radiological results of treatment and their late clinical consequences. Acta Orthop Scand Suppl, 1977; 167: 1.

Mose K. Methods of measuring in Legg-Calvé-Perthes disease with special regard to the prognosis. Clin Orthop, 1980; 150: 103.

Perthes GC. Uver Arthritis deformans jufenilis. Deutsch Z Chir, 1910; 107: 111.

Petrie JG , Bitenic I. The abduction weight-bearing treatment in Legg-Perthes disease. J Bone Joint Surg, 1971; 53B: 54.

Ratliff AHC. Perthes' disease: a study of 16 patients followed up for 40 years. J Bone Joint Surg, 1977; 40B: 248.

Salter RB. The present status of surgical treatment of Legg-Perthes disease. Current concept review. J Bone Joint Surg, 1984; 66A: 961.

Salter RB. Legg-Perthes disease. The scientific basis for the methods of treatment and their indications. Clin Orthop, 1980; 150: 8.

Salter RB, Thompson G. Legg-Calvé-Perthes disease. The prognostic significance of the subchondral fracture and a two-group classification of the femoral head involvement. J Bone Joint Surg, 1984; 66A: 479.

Stulbert SD, Salter RB. The natural course of Legg-Perthes disease and its relationship to degenerative arthritis of the hip: long-term follow-up study. Orthop Trans, 1977; 1: 105.

Tachdjian MO. Pediatric Orthopedics. 2nd edn. Philadelphia: WB Saunders, 1990: 943.

Van Dam BE, Crider RJ, Noyes JD, Larsen LF. Determination of the Catteral classification in Legg-Calvé-Perthes disease. J Bone Joint Surg, 1981; 63A: 906.

Van der Heyden AM, van Tongerloo RS. Shelf operation in Perthes' disease. J Bone Joint Surg, 1981; 63B: 282.

Weinstein SL. Long term follow-up of patients with Legg-Calvé-Perthes disease. Orthop Trans, 1985; 9: 204.

Wynne-Davies, R. Some etiologic factors in Perthes' disease. Clin Orthop, 1980; 150: 12.