Avascular necrosis (AVN) is defined as cellular death of bone components due to interruption of the blood supply; the bone structures then collapse, resulting in pain, loss of joint function and long-term joint damage.
AVN usually involves the epiphysis (end part of a long bone), such as the femoral and humeral heads and the femoral condyles, but small bones can also be affected. In clinical practice, AVN is most commonly encountered in the hip.[1, 2] Most available data regarding the natural history, pathology, pathogenesis, and treatment of AVN pertains to femoral head necrosis.
AVN is associated with numerous conditions. Patients taking corticosteroids and organ transplant recipients are particularly at risk of developing AVN. AVN of the jaw has been described in patients taking bisphosphonates and, more recently, denosumab.[3, 4] For full discussion of this entity, see Bisphosphonate-Related Osteonecrosis of the Jaw.
For diagnosis, plain radiography is the most appropriate initial imaging study, although findings are unremarkable in the early stages of AVN.MRI is the most sensitive and specific imaging modality. See Workup.
Early diagnosis and appropriate intervention can delay the need for joint replacement. However, most patients present late in the disease course. Without treatment, the process is almost always progressive, leading to joint destruction within 5 years. See Treatment.
Although the pathophysiology of AVN is not fully understood, the final common pathway is interruption of blood flow to the bone. AVN often affects bones with a single terminal blood supply, such as the femoral head, carpals, talus, and humerus. The earliest pathologic characteristics of osteonecrosis are necrosis of hematopoietic cells and adipocytes followed by interstitial marrow edema[5] .
Osteocyte necrosis occurs after approximately 3 hours of anoxia, but histological signs of osteocyte death do not appear until approximately 24 to 72 hours after oxygen deprivation[5] . Interruption of the vascular supply and resultant necrosis of marrow, medullary bone, and cortex are theorized to be caused by the mechanisms listed below. However, individual patients usually have more than one risk factor; this indicates that the pathogenesis of AVN is likely multifactorial.
United States
The frequency of AVN depends on the site involved. The most common site is the hip; other locations include the carpals, talus, femur, metatarsal, mandible, and humerus. In the United States, approximately 15,000 new cases of AVN are reported each year. AVN accounts for more than 10% of total hip replacement surgeries performed in the United States. Osteonecrosis of the jaw associated with bisphosphonate has also been well studied and reported. Most patients with osteonecrosis of the jaw also had an ongoing malignancy and/or had undergone a recent dental procedure.[8, 3, 9]
International
In most countries, the incidence and prevalence of AVN are not well reported. A Japanese survey estimated that 2500-3300 cases of AVN of the hip occur each year; of these, 34.7% were due to corticosteroid use, 21.8% to alcohol abuse, and 37.1% to idiopathic mechanisms.[10]
Data on mortality rates associated with AVN are not available. Most data involve AVN of the hip. Mortality rates are low and vary based on the operative procedure used to treat AVN.
Morbidity rates are high and depend on the underlying cause. Morbidity rates associated with AVN of the hip are high; the prevalence of long-term disability is significant. Despite advances in orthopedic procedures, most patients with advanced AVN require more than one hemiarthroplasty or total hip replacement during their lifetime.
In a study that included 1706 patients who underwent total hip replacement for AVN of the femoral head, Lovecchio et al concluded that AVN is an independent risk factor for transfusion up to 72 hours postoperatively and for readmission up to 30 days postoperatively. Bleeding requiring transfusion was the most common medical complication, occurring in 19.6% of patients with AVN compared with 13.9% of those without AVN (P< 0.001).[11]
AVN has no racial predilection except for cases associated with sickle cell disease and hemoglobin S and SC disease, which predominantly occur in people of African and Mediterranean descent.
With the exception of AVN associated with systemic lupus erythematosus, AVN is more common in men, with an overall male-to-female ratio of 8:1.
AVN is a disease of middle age that most often occurs during the fourth or fifth decade of life and is bilateral in more than half of cases.
Avascular necrosis (AVN) may be asymptomatic and is occasionally discovered incidentally on radiographs. Symptoms depend on the affected joint. Medullary infarcts are usually silent, and infarcts of the small bones of the hands and feet are often symptomatic.
Pain in the affected joint is typically the presenting symptom of AVN, regardless of the location. Patients with AVN of the femoral head often report groin or anterior thigh pain that is exacerbated by weight bearing. The pain may initially be mild but progressively worsens over time and subsequently may be present at rest or at night.
Large infarcts, such as those due to Gaucher disease and hemoglobinopathies, are associated with very severe pain.
AVN may be primary or idiopathic. For AVN that is secondary or associated with an underlying condition or exposure, the following factors have been identified:
Initially, the physical examination findings of AVN may be unrevealing. Abnormal physical findings depend on the location and severity of disease. With progression of AVN of the hip, joint function deteriorates and the patient may walk with a limp. AVN of smaller, non–weight-bearing joints typically does not cause significant disability. Findings may include the following:
No laboratory test findings specifically suggest or confirm the presence of avascular necrosis (AVN). Plain radiographic findings are unremarkable in early stages of AVN. Nevertheless, the American College of Radiology (ACR) considers x-ray of the pelvis and hips the most appropriate initial imaging study in patients at risk for AVN who present with hip pain.[18] If radiographs are normal or show femoral head lucencies suspicious for osteonecrosis, magnetic resonance imaging (MRI) of the hips without contrast is most appropriate.[18]
The ACR advises that MRI is the most sensitive and specific imaging modality for diagnosis and provides optimal evaluation of the likelihood of articular collapse. Involvement of greater than 30-50% of the femoral head, often in the sagittal plane, indicates significantly increased risk of articular collapse.[18]
Additional ACR recommendations include the following[18] :
For more information, see Imaging in Avascular Necrosis of the Femoral Head.
Histology is the criterion standard for diagnosis of AVN. However, bone biopsy is not routinely performed because of the availability of sensitive noninvasive tests such as MRI.
Histology is the criterion standard for diagnosis of AVN, although it is usually unnecessary. The histologic specimen is usually obtained during surgery, although it is occasionally obtained during diagnostic bone biopsy. Histologic changes are observed in both cortical bone and bone marrow.
In the early stage, interstitial edema and plasmostasis is confined to the bone marrow. As the severity progresses, osteocytes continue to disappear and necrotic tissue fills the medullary spaces. New living bone is laminated onto dead trabeculae with partial resorption of dead bone. In the subchondral trabeculae, bone resorption exceeds formation leading to the net removal of bone, loss of structural integrity of trabeculae, subchondral fracture, and joint incongruity[5] .
Several different staging systems have been developed and continue to be used. Ficat initially developed an AVN staging system based on radiologic findings. This staging system was revised after the widespread use of MRI in the workup of AVN. The staging system presented in the below table is based on the consensus of the Subcommittee of Nomenclature of the International Association on Bone Circulation and Bone Necrosis (ARCO: Association of Research Circulation Osseous). The most important consideration is collapse of the femoral head cortex. Repair and complete recovery may be possible prior to collapse. Afterward, the collapse is irreversible[19] .
Staging of Avascular Necrosis
View Table | See Table |
Plain radiographic findings are unremarkable in early stages of AVN. Nevertheless, the American College of Radiology considers x-ray of the pelvis the most appropriate initial imaging study in patients at risk for AVN who present with hip pain.Both an anteroposterior view of the pelvis and a frog-leg lateral view of the hip are necessary, as articular collapse or cortical depression may be seen on only one of those projections.[18]
In children, the earliest radiographic findings of AVN include the following[18] :
Subsequentl findings in pediatric patients include fragmentation, resorption, reossification, and remodeling of the femoral head and neck.[18]
In mild-to-moderate AVN, radiographs demonstrate sclerosis and changes in bone density. In advanced disease, bone deformities, such as flattening, subchondral radiolucent lines (crescent sign), and collapse of the femoral head, are evident (see images below).
View Image | Avascular necrosis in the femoral head resulting from corticosteroid therapy. |
View Image | Avascular necrosis of the shoulder showing subchondral radiolucent lines (crescent sign). |
MRI is the most sensitive and specific imaging procedure for AVN, of the hip with an overall sensitivity that exceeds 90%. The specificity of MRI is also very high. The use of gadolinium is particularly useful in early detection.
MRI findings of AVN include decreased signal intensity in the subchondral region on both T1- and T2-weighted images, suggesting edema (water signal) in early disease. This relatively nonspecific finding is often localized in the medial aspect of the femoral head. This abnormality is observed in 96% of cases.
The next stage is characterized by a reparative process (reactive zone) and shows low signal intensity on T1-weighted scans and high signal intensity on T2-weighted scans. This finding is diagnostic for AVN (see images below).
View Image | Avascular necrosis of both femoral heads. This T1-weighted image shows decreased signal intensity in both femoral heads. |
View Image | MRI of the distal femur and proximal tibia. This T2-weighted image shows increased signal intensity in the marrow. |
Advanced AVN is characterized by deformity of the articular surface and by calcification, which are also easily detected with radiography and CT scanning.
In early AVN, osteoblastic activity and blood flow are increased; thus, the sensitivity of radionuclide bone scan is better than that of plain films at this stage.
The central area of decreased uptake is surrounded by an area of increased uptake. This phenomenon is known as the doughnut sign and indicates the reactive zone surrounding the necrotic area.
Limitations of bone scan include the following:
Computed tomography (CT) is not commonly used for assessment of osteonecrosis in pediatric patients. In adults, CT is used principally to provide information for surgical planning, by determining the severity and location of articular collapse and providing evidence of early secondary degenerative joint disease.[18]
CT scans show sclerosis in the central part of femoral head as an asterisk sign. Changes in the anterior part of the femoral head are easily observed
Medical management of avascular necrosis (AVN) primarily depends on the location and severity of disease, as well as the patient's age and general health. Treatment outcomes correlate directly with the stage of the disease. No medical treatment has proven effective in preventing or arresting the disease process.
Conservative measures include limited weight bearing with crutches and pain medications. This may be beneficial and is a reasonable initial course of action if the involved segment is smaller than 15% and far from the weight-bearing region. Immobilization may be helpful in some cases (eg, AVN of the distal femur or tibia). In advanced AVN, the disease course is unaffected by activity and will eventually require surgery.
Treatment with bisphosphonates may be helpful. Although large-scale, randomized, double-blind studies have not been performed, a systematic review of eight articles involving 788 hips suggested that alendronate has short-term efficacy in reducing pain, improving articular function, slowing of bone collapse progression, and delaying the need for total hip arthroplasty in adults with AVN[20]
Iloprost, a vasoactive prostaglandin analog that is approved for inhalational treatment of pulmonary hypertension, has shown clinical and radiographic benefits in early-stage AVN when administered intravenously. Iloprost produces vasodilation and increases microcirculation, reducing bone marrow edema and relieving pain and other accompanying symptoms; it inhibits platelet aggregation and diminishes the concentration of oxygen free radicals and leukotrienes.[21, 22]
Statin therapy to prevent corticosteroid-induced AVN may be helpful. Pritchett reported a 1% incidence of AVN in 284 patients who were on statin therapy during the entire period of corticosteroid treatment (average, 7.5 y).[23] The use of high-dose corticosteroids carries a reported 3-20% incidence of AVN.
Extracorporeal shockwave therapy (ESWT) has shown beneficial effects in early AVN of the femoral head. ESWT may relieve pain and improve hip function of the hip, and induce regression of AVN.[24, 25] ESWT may be combined with alendronate therapy.[24]
Several surgical procedures have been used in an attempt to treat AVN, with variable success. No surgical procedure is the consensual best among surgeons in the treatment of AVN. In early stages of AVN (precollapse), core decompression with or without bone graft is typically considered the most appropriate treatment. In late stages, characterized by collapse, femoral head deformity, and secondary osteoarthritis, total hip arthroplasty is the most appropriate treatment.
Researchers postulate that core decompression improves circulation by decreasing intramedullary pressure and preventing further ischemia and progressive joint destruction. The best results vary from 34-95%, which is significantly better than results of conservative treatment.[26] The best results are obtained when treating patients with early AVN (precollapse). Core decompression is also effective for pain control.
Bone graft options include structural cortical or medullary bone graft and vascularized bone graft with either a muscle-pedicle bone graft or free vascularized fibular graft.[27]
Bone grafting is combined with the following:
The best results have been reported with free vascularized bone grafts. Success rates of 70% and 91% have been reported in two small series.[28, 29]
Advantages of free vascularized grafts compared with total hip arthroplasty include the following:
Disadvantages of free vascularized grafts include the following:
In a review of bone grafting through a window at the femoral head-neck junction, Zuo et al found that factors predicting good clinical success were absence of femoral head collapse or degree of collapse < 2 mm (Association Research Circulation Osseous [ARCO] stage IIIa). Surgical failure rates were high in patients with femoral head collapse >2 mm (ARCO stage IIIb and IIIc) and in those with necrotic lesions involving the lateral pillar (L2 and L3 type). Finally, patients 40 years of age and older had a worse prognosis than those younger than 40.[30]
Several osteotomy procedures have been tried with variable success.
Intertrochanteric osteotomies have been performed in patients with posttraumatic AVN.
Transtrochanteric rotational osteotomy involves rotation of the femoral head and neck on the longitudinal axis. The necrotic anterosuperior part of the femoral head becomes posterior, and the weight-bearing force is transmitted to what was previously the posterior articular surface, which is not involved in the ischemic process. In 1992, Sugano and colleagues reported excellent results in 56% of patients who underwent this procedure.[31] Transtrochanteric rotational osteotomy is technically demanding.
Most patients with advanced disease (stage III and above) require total hip arthroplasty. Total hip arthroplasty provides excellent pain relief for many years, although most young patients require repeat surgery. With high failure rates (10-50% after 5 years), patients with AVN will probably need a second total hip arthroplasty during their lifetime.
Cheung et al reported excellent long-term survival of cementless total hip arthroplasty for managing AVN of the hip. In their review of 182 total hip arthroplasties, 117 of them for AVN, 19.1- year survival was 97.1% for AVN patients and 96.2% for non-AVN patients.[32]
The use of stem cells implanted via core decompression has been studied for the treatment of early-stage (precollapse) AVN of the femoral head.[33, 34, 35] For example, Pilge et al reported benefit with the application of autologous bone marrow concentrate, derived from iliac crest aspirates. In this study, compared with core decompression alone, the addition of bone marrow stem cells seemed to decrease pain and other joint symptoms, improve range of motion, and prevent collapse of the femoral head.[21]
Hernigou et al reported superior long-term outcome with core decompression plus bone marrow injection compared with decompression alone in the same patient. Their study included 125 consecutive patients (78 males and 47 females) with bilateral corticosteroid-induced AVN.[36]
All patients had AVN at the same stage in both sides (stage I or II); the hip with lower volume of osteonecrosis, as measured with MRI, was treated with core decompression, and the contralateral hip was treated with decompression and percutaneous injection of mesenchymal cells (MSCs) obtained from bone marrow concentration. The average total number of MSCs (counted as number of colony-forming units–fibroblast) in each injection was 90,000 ± 25,000 cells (range, 45,000 to 180,000 cells).[36]
On follow-up conducted an average 25 years after the first surgery (range 20 to 30 years), 35 hips (28%) treated with cell therapy had collapsed, compared with 90 (72%) treated with decompression only. Total hip arthroplasty had been performed in 30 hips (24%) in the cell therapy group, versus 95 (76%) in the decompression-only group (P < 0.0001).[36]
Houdek et al reported beneficial results from injecting the combination of bone marrow–derived mesenchymal stem cells and platelet-rich plasma into the femoral head after standard hip decompression in patients with corticosteroid-induced AVN. In their preliminary study of 35 hips in 22 patients, use of this technique improved pain and function, and more than 90% of femoral heads had not collapsed at a minimum of 2 years.[37]
Systematic reviews of studies with patient-reported outcomes have demonstrated clinical benefit with the use of stem cells for hip AVN, with a low rate of complications, but have highlighted the lack of standardization with this technique. Comparisons across studies have been complicated by differences in etiology and severity of AVN, cell sources and doses, adjuvant therapies used, and outcome assessment methodology.[34, 35, 38, 39]
Although the assumption is that the injected stem cells are taken up and differentiate into osteoblasts, Im points out that the survival of stem cells implanted into the hip has not been studied. In other organs, tracking studies have shown that injected or implanted stem cells usually exert paracrine effects and then largely die off. Measures to improve stem cell survival by enhancing the vascularity and osteogenic potential of the treatment site are currently under study, and include use of adipose stem cells and the addition of angiogenic factors, such as vascular endothelial growth factor (VEGF).[39]
AVN of the femoral condyles (knees) may respond to more conservative intervention such as arthroscopic lavage and debridement. AVN of the femoral condyles has a better prognosis than hip AVN, although osteoarthritis eventually develops.
In a study of 25 patients with 29 shoulders treated with arthroplasty for humeral head AVN, Ristow et al reported that both total shoulder arthroplasty and hemiarthroplasty are safe and effective treatments, regardless of etiology and radiographic staging. At a mean follow-up of 3.9 years (range, 1-8.5 years), all patients who underwent operative intervention showed statistically significant improvement in measurements of functionality.[40]
Obtain consultation with an orthopedic surgeon. Early intervention can save affected joints and obviate the need for joint replacement.
In early AVN, patients should use crutches or other supports to avoid weight bearing. In advanced AVN, the disease course is unaffected by activity; surgery is the only option.
No pharmaceutical treatment has been proven to prevent progression of avascular necrosis (AVN). Analgesics are often used as needed for pain relief.
The following precautions can be taken to minimize the risk of developing avascular necrosis (AVN) and to improve outcomes in these patients:
For patients undergoing pelvic radiotherapy, a variety of pharmacological interventions for preventing AVN have been studied, including supplemental calcium and/or vitamin D, bisphosphonates, selective estrogen receptor modulators, hormone replacement therapy (estrogen or testosterone), denosumab, and calcitonin. However, a Cochrane review concluded that the evidence relating to those interventions is of very low certainty.[41]
See the list below:
The prognosis of AVN depends on the disease stage at the time of diagnosis and the presence of any underlying conditions. More than 50% of patients with AVN require surgical treatment within 3 years of diagnosis. Half of patients with subchondral collapse of the femoral head develop AVN in the contralateral hip.[42]
Poor prognostic factors include the following:
See the list below:
Stage Clinical and Laboratory Findings Stage 0
Patient is asymptomatic. Radiography findings are normal. Histology findings demonstrate osteonecrosis.Stage I
Patient may or may not be symptomatic. Radiography and CT scan findings are unremarkable. AVN is considered likely based on MRI and bone scan results (may be subclassified by extent of involvement [see below]). Histology findings are abnormal.Stage II
Patient is symptomatic. Plain radiography findings are abnormal and include osteopenia, osteosclerosis, or cysts. Subchondral radiolucency is absent. MRI findings are diagnostic.Stage III
Patient is symptomatic. Radiographic findings include subchondral lucency (crescent sign) and subchondral collapse. Shape of the femoral head is generally preserved on radiographs and CT scans. Subclassification depends on the extent of crescent, as follows:
- Stage IIIa: Crescent is less than 15% of the articular surface.
- Stage IIIb: Crescent is 15-30% of the articular surface.
- Stage IIIc: Crescent is more than 30% of the articular surface.
Stage IV
Flattening or collapse of femoral head is present. Joint space may be irregular. CT scanning is more sensitive than radiography. Subclassification depends on the extent of collapsed surface, as follows:
- Stage IVa: Less than 15% of surface is collapsed.
- Stage IVb: Approximately 15-30% of surface is collapsed.
- Stage IVc: More than 30% of surface is collapsed.
Stage V
Radiography findings include narrowing of the joint space, osteoarthritis with sclerosis of acetabulum, and marginal osteophytes.Stage VI
Findings include extensive destruction of the femoral head and joint.