Initially described in 1910 by three separate physicians (Arthur Thornton Legg, Jacques Calve, Georg Perthes) observing identifiable changes in the hip joint with the advent of X-rays, Legg-Calvé-Perthes disease (LCPD) is the term that is now used to describe idiopathic avascular necrosis of the femoral head in the pediatric population. Although there are many identifiable causes of avascular necrosis in the pediatric patient, such as sickle cell disease, systemic inflammatory disease, chemotherapy, radiation, or prolonged steroid use, the term LCPD is reserved for pediatric patients presenting with noticeable hip abnormality without an identifiable cause. Current theories regarding its etiology include trauma, vascular abnormalities, or increased mechanical loading.
Incidence of Legg-Calvé-Perthes disease (LCPD) typically peaks between ages 4 and 8 years, although it can be seen in patients as young as age 2 years and as old as age 12 years.1 This age range can be helpful in differentiating patients with LCPD from those with slipped capital femoral epiphysis, who tend to present at an average age of 12 to 13 years. LCPD also has a strong predilection for males, with a 4:1 ratio of male to female patients. It presents unilaterally 89% of the time.2 Several studies suggest that a positive family history for LCPD, low birth weight, passive smoke exposure during pregnancy, certain human leukocyte antigen subtypes, and coagulation abnormalities (eg, factor V Leiden, protein C deficiency) are all associated with increased prevalence of LCPD.1,2 Patients with LCPD can be small in stature for their chronologic age, with disproportionately small hands and feet compared to their head size.1
A 7-year-old boy presented to the pediatric clinic with a history of insidious onset, right-sided thigh and knee pain. He described diffuse pain about the knee with some radiation proximally into the thigh. The pain was not intense, but worse with bearing weight on the right side and it had steadily increased over the course of several months. He denied any recent trauma; however, he had slowly begun having difficulty participating in physical activities due to the pain. He had no significant medical or surgical history. Family history was noncontributory. He was developmentally normal and had accomplished all of his milestones at the appropriate age. On physical examination, the patient had a right-sided antalgic gait. He had full range of motion at the right knee with no tenderness to palpation or visible swelling, and it was stable to varus/valgus stress and anterior/posterior drawer tests. He had normal sensation distally in his foot, with a strong, palpable dorsalis pedis pulse. Plain radiographs of the knee were obtained (Figure 1). The knee radiographs were negative for any fracture or bony abnormality. There was no visible knee joint effusion. The patient was discharged from the clinic with conservative treatment with rest, nonsteroidal anti-inflammatory drugs, and a scheduled follow-up appointment.
(A) Anterior/posterior radiograph and (B) lateral radiograph of the right knee.
The patient returned to the clinic 6 weeks later with continued pain but now with a worsening limp on the right side. There was no improvement in his symptoms, and he described the pain as a little worse. During the physical examination at this visit, close observation of his gait cycle revealed a Trendelenburg gait on the right side, which was confirmed using the one-leg Trendelenburg test. He continued to have full range of motion and strength in his knee; however, examination of his right hip revealed guarding during the examination along with decreased internal rotation and abduction compared to the contralateral side. Due to the positive findings in his hip physical examination, anterior/posterior (AP) and lateral radiographs of the bilateral hips were taken (Figure 2). The AP view of the hip suggests some flattening of the femoral head. The frog lateral view of the hip clearly demonstrated a large lucency in the anterolateral aspect of the femoral head.
(A) Anterior/posterior (AP) radiograph of the pelvis of patient with right-sided Legg-Calvé-Perthes disease showing subtle sclerosis and collapse laterally in the femoral head. (B) Frog lateral view better depicts the subchondral lucency (arrow) that can easily be missed on the AP view, highlighting the importance of obtaining this view.
Despite the patient's presentation of thigh and knee pain, the combination of his physical examination and his imaging suggested that the patient's pain was originating from his hip joint. His knee radiographs showed no abnormalities during the first visit, and his knee examination was normal. The subsequent AP pelvis radiograph showed mild changes in the lateral aspect of the femoral head. Radiographic changes were much more clearly demonstrated on the frog lateral view. These hip radiograph findings are classic changes of avascular necrosis of the femoral head. Because there was no identifiable underlying pathology or risk factor for avascular necrosis in this patient, the leading diagnosis was LCPD.
This patient's presentation emphasizes the importance of ruling out hip pathology in children presenting with knee pain as their primary symptom, as referred pain is common. In a pediatric patient with knee pain, a normal physical examination of the knee should clue the physician to examine the hip joint during the same visit. Along with LCPD, other causes of hip pathology in the pediatric patient are important to consider in the differential diagnosis. Some causes of acute hip pathology include fracture, muscular strains or tendinitis, avulsion injuries, transient synovitis, septic arthritis, and osteomyelitis. Hip pathology that is subacute or chronic in nature can include slipped capital femoral epiphysis, inflammatory conditions (juvenile idiopathic arthritis), benign and malignant bone tumors (eg, osteoid osteoma, Ewing's sarcoma), or avascular necrosis attributable to another cause such as sickle cell disease or chronic steroid use. Femoral head collapse that is seen in bilateral hips generally suggests a more systemic cause such as skeletal dysplasia (eg, multiple epiphyseal dysplasia) or hypothyroidism.
The patient presented in this case had a typical presentation for LCPD. The symptoms are often mild and nonspecific in the early stages. Because this problem often occurs in children as young as age 4 years, the patient may not be able to describe the pain well. Although the pathology is in the hip, the pain often localizes to the thigh or knee. A proposed mechanism for this referred pain is thought to be because a sensory branch of the obturator nerve that innervates the hip also innervates the distal thigh or knee.3 It is not uncommon for a patient to initially be evaluated for knee pathology, with extensive testing that includes radiographs, blood tests looking for rheumatologic causes, and even magnetic resonance imaging (MRI). A simple examination of the hip looking for diminished range of motion, especially decreased abduction and decreased internal rotation, will often uncover the true source of pain. It is important that an examination of the hip always be included when evaluating children with knee pain.
Physical examination findings in the hip may vary depending on the timing of presentation of the patient. During the early phases of presentation, the patient may only have an occasional antalgic gait on the affected side with only slight changes in range of motion. This may be easily missed by the examining physician. As the natural course of the disease progresses, the patient may develop further decreases in range of motion, with limitation in internal rotation and abduction of the hip. This limitation in range of motion comes from several sources. Initially, muscle spasm and guarding limit the patient's range. Subsequently, development of a joint effusion can cause further stiffness. As more collapse of the femoral head takes place, range of motion further decreases and the patient may begin to develop hip flexor and adductor contractures.
If significant femoral head collapse occurs, the physician may notice a leg length discrepancy, with the affected side being shorter. At this point, the patient may display a Trendelenburg gait, which is marked by the trunk leaning toward the affected side and the pelvis dropping to the contralateral side during stance phase. The cause of this limp is due to the shortening of the ipsilateral gluteal musculature, leading to hip abductor deficiency. Without functional abductor musculature, the pelvis drops to the contralateral side during the stance phase and the patient leans toward the affected side to compensate.
Imaging is then used to confirm what is suspected on physical examination. The AP radiograph of the pelvis is used to visualize bilateral hip joints. Depending on the stage of the disease, the affected hip can show a continuum of changes, from sclerosis, to subchondral lucency, to loss of height in the femoral head and collapse. At times, the AP radiograph does not show any significant changes and the diagnosis can be missed. Obtaining a frog lateral radiograph is important to pick up any subtle lucencies in the anterolateral aspect of the femoral head. These simple initial imaging modalities are often all that are needed to confirm the diagnosis. Some surgeons use MRI to plan treatment options, but MRI is rarely needed to make the diagnosis of LCPD.
The pathophysiology of this disease can be subdivided into four stages. The initial stage is thought to arise from temporary interruption in the blood supply to the femoral epiphysis, leading to bone and cartilage cell death. Subsequently, there is collapse of trabecular bone, a stage termed “fragmentation.”4 In the next stage, reossification then occurs with formation of new bone in the femoral head. In the healed stage, the femoral head continues to remodel until skeletal maturity and normal trabecular patterns and radiodensity gradually return.
Note that avascular necrosis of the hip in children is very different from of the hip in adults. In adults with avascular necrosis, once the bone in the hip has collapsed, no new bone growth can be expected. Figure 3A and Figure 3B show the remarkable amount of bone regeneration that can occur in LCPD.
(A) Radiograph of a 7-year-old boy with extensive collapse of the femoral head from Legg-Calvé-Perthes disease. (B) Radiograph from the same patient, 8 years later. Note the excellent remodeling of the femoral head.
Currently, the most common classification system used to predict prognosis and guide treatment is the Herring classification (Figure 4). This classification, also referred to as the lateral pillar classification, is based on the height of the lateral 15% to 30% of the epiphysis. It is subdivided into four groups: Group A (no loss of lateral pillar height), Group B (<50% loss), Group C (>50% loss), and a hybrid Group B/Group C border (narrowed lateral pillar with approximately 50% height). Typically, patients with less collapse of the lateral pillar and younger age at presentation go on to have better outcomes.
Herring classification of the femoral head. (A) Normal femoral head. (B) Group A: no loss of height. (C) Group B: between 0% and 50% loss of height. (D) More than 50% loss of height. Reprinted with permission from Farsetti et al.11
Once the diagnosis of LCPD is made, the patient should be referred to an orthopedic specialist surgeon for further care. Treatment options vary, depending on the patient's age and the amount of collapse of the femoral head (as defined by the Herring classification).
The disease process can be frustrating for the patient's family. There is no “cure” for LCPD. It often takes several years for the disease to run its course. It is difficult to tell the family with certainty what measures will be needed to treat the problem, especially in its early stages. Patients may present with a “mild” form of LCPD, only to go on to develop more severe collapse of the femoral head, necessitating major operative intervention.
The main goal of treatment in LCPD disease is to prevent deformation of the femoral head while the reossification process takes place. The time period during the course of the disease where the femoral head has the highest risk of deformation is in the late phase of fragmentation stage, when necrotic bone is being actively resorbed, and during the early phase of resossification, when woven bone is being laid down in random manner before being organized into trabeculae along the direction of weight bearing forces.5
Although the treatment for LCPD is constantly evolving, the keys to initial treatment remain the same. These include restoration of range of motion and maintaining coverage (containment) of the femoral head in the acetabulum. Range of motion can be restored by restricting activity and using nonsteroidal anti-inflammatory drugs to improve pain and decrease the hip effusion. Most patients also begin physical therapy to regain range of motion.
If these patients do not improve with conservative therapy and go on to develop adductor contracture and progressive radiographic changes, initial surgical treatment to maintain coverage of the femoral head can include lengthening of the adductor longus muscle followed by casting to improve abduction. An example is shown in Figure 5A, where the patient presented at age 8 years and was noted to have left-sided LCPD with decreased abduction but adequate coverage of the femoral head, and thus underwent adductor tenotomy and casting. Figure 5B shows the intraoperative arthrogram after adductor tenotomy, showing good congruity of the hip joint. A follow-up examination 3 years later showed maintenance of joint congruity (Figure 5C). Daytime bracing to maintain abduction was used commonly in the past but is less common now.
(A) Radiograph of initial presentation of an 8-year-old boy with left-sided Legg-Calvé-Perthes disease with femoral head collapse and decreased abduction on physical examination. (B) Intraoperative arthrogram showing congruent hip joint after the boy underwent adductor tenotomy and casting. (C) Long-term follow-up image shows good articular congruity and resossification.
The key in treatment is to minimize forces acting on the femoral head across the acetabulum and prevent extrusion. This principle, also known as containment, helps keep the affected anterolateral aspect of the femoral epiphysis positioned in the acetabulum and avoid weight-bearing stresses that cause further deformation and extrusion.6
Surgical treatment to obtain containment of the femoral head can be achieved by reorienting the femur or the pelvis. The former method is called a proximal femoral varus osteotomy and positions the anterolateral (uncovered) aspect of the head under the acetabulum. By taking out a wedge from the proximal femur, the hip can be tipped more deeply into the socket. The other major method of containment is a pelvic osteotomy. In that procedure, a cut is made around the acetabulum to allow the socket to “roll over” the deformed femoral head and give it better coverage. An example of a pelvic osteotomy can be seen in Figure 6, with radiographs 8 months after the surgery showing maintenance of the articular congruity and some reossification.
(A) Anterior/posterior and lateral (B) radiographs of right-sided Legg-Calvé-Perthes disease with significant collapse and femoral head extrusion in a 6-year-old boy with decreased abduction and poor response to physical therapy. (C) He underwent pelvic osteotomy to provide better coverage of the femoral head. (D) Radiograph taken 8 months postoperatively shows maintenance of containment and some reossification.
Prognostic factors associated with improved outcomes include age at presentation and radiographic classification of femoral head collapse. Recent studies, including two large, multicenter prospective studies7,8 along with recent meta-analyses propose that patients older than age 6 years and with 50% loss of height/necrosis have improved outcomes and better head sphericity with operative treatment.9,10 Overall, patients presenting at a younger age with minimal radiographic changes respond well to conservative treatment, thus emphasizing timely diagnosis and treatment. For patients with advanced radiographic changes (eg, Herring class C), long-term prognosis remains guarded despite operative treatment.
It is of utmost importance to perform a thorough clinical examination on a pediatric patient presenting with knee or thigh pain. As a routine part of the physical examination, the hip should be assessed to rule out any possible pathology; obtaining plain radiographs of the hip during the initial assessment will also help. If LCPD is diagnosed in a patient presenting before the age of 6 years with little to no femoral head collapse, then the prognosis is good with conservative treatment such as physical therapy and nonsteroidal anti-inflammatory drugs. For older patients with increased collapse, surgical treatment to increase containment may be needed to maintain a congruent hip joint and range of motion.
- Perry DC, Hall AJ. The epidemiology and etiology of Perthes disease. Orthop Clin North Am. 2011;42(3):279–283. doi:10.1016/j.ocl.2011.03.002 [CrossRef]
- Loder RT, Skopelja EN. The epidemiology and demographics of Legg-Calve-Perthes' disease. ISRN Orthop. 2011;504393.
- Sakamoto J, Morimoto Y, Ishii S, et al. Investigation and macroscopic anatomical study of referred pain in patients with hip disease. J Phys Ther Sci. 2014;26(2):203–208. doi:10.1589/jpts.26.203 [CrossRef]
- Kim HKW, Herring JA. Pathophysiology, classifications, and natural history of Perthes disease. Orthop Clin North Am. 2011;42(3):285–295. doi:10.1016/j.ocl.2011.04.007 [CrossRef]
- Joseph B, Price CT. Principles of containment treatment aimed at preventing femoral head deformation in Perthes disease. Orthop Clin North Am. 2011;42(3):317–327. doi:10.1016/j.ocl.2011.04.001 [CrossRef]
- Joseph B. Management of Perthes' disease. Indian J Orthop. 2015;49:10–16. doi:10.4103/0019-5413.143906 [CrossRef]
- Wiig O, Terjesen T, Svenningsen S. Prognostic factors and outcome of treatment in Perthes' disease: a prospective study of 368 patients with five-year followup. J Bone Joint Surg Br. 2008;90(10):1364–1371. doi:10.1302/0301-620X.90B10.20649 [CrossRef]
- Herring JA, Kim HT, Browne R. Legg-Calve-Perthes disease. Part II: prospective multicenter study on the effect of treatment on outcome. J Bone Joint Surg. 2004;86-A(10):2121–2134.
- Saran N, Varghese R, Mulpuri K. Do femoral or Salter innominate osteotmies improve femoral head sphericity in Legg-Calve-Perthes disease? A meta-analysis. Clin Orthop Related Res. 2012;470:2383–2393. doi:10.1007/s11999-012-2326-3 [CrossRef]
- Nguyen NT, Klein G, Dogbey G, McCourt JB, Mehlman CT. Operative vs nonoperative treatments for Legg-Calve-Perthes disease: a meta-analysis. J Pediatr Orthop. 2012;32(7):697–705. doi:10.1097/BPO.0b013e318269c55d [CrossRef]
- Farsetti P, Tudisco C, Caterini R, Potenza V, Ippolito E. The Herring lateral pillar classification for prognosis in Perthes disease. Late results in 49 patients treated conservatively. J Bone Joint Surg Br. 1995;77(5):739–742.