Atraumatic osteonecrosis (ON) is an infarction of an area of bone secondary to a microvascular insult which leads to bone resorption and potentially structural collapse. The disease typically affects epiphyseal bone on the convex side of a joint, likely due to the lack of collateral circulation, and most commonly affects the femoral head, followed by the distal femur, proximal humerus and talus. In regards to osteonecrosis of the femoral head, the mean age at presentation is 38 years. Twenty-five percent of patients with the condition are less than 25 years of age. It may be bilateral in up to 72% of the cases. Approximately 20,000 new cases are diagnosed in the U.S. each year and about 10% of total hip arthroplasties are performed to treat degenerative disease secondary to osteonecrosis.

Osteonecrosis of the Femoral Head with Subchondral Collapse


Insert vascular anatomy of hip


The majority of osteonecrosis cases are idiopathic, related to high dose corticosteroid use or related to alcohol abuse. Additionally, there is increasing evidence for the role of clotting disorders in the development of this disease.

Up to 38% of patients with ON have a history of high dose corticosteroid use (~2 grams of prednisone within 2-3 months), however only ~5% of patients with a history of high dose corticosteroid use develop osteonecrosis

. Similarly, up to 29% of patients with ON have a history of alcohol abuse, however only ~5% of patients with a history of alcohol abuse develop ON

ON is a multifactorial process that is not clearly defined as of yet. The pathophysiology of corticosteroid and alcohol-related ON has been studied extensively and strong evidence is in support of increased adipogenesis as a major underlying mechanism. It has been reported that corticosteroids and ethanol can induce a pluripotent bone marrow cell line to preferentially differentiate into adipocytes in vitro. Increased adipogenesis causes venous sinusoidal compression, which leads to venous congestion, intraosseous hypertension, impaired arterial inflow, and ultimately, infarction

However, as stated earlier, only 5% of patients with a history of high-dose corticosteroid use or alcohol abuse develop ON. Many groups are studying genetic traits that may predispose to the development of ON. One are of interest is clotting disorders. One series reported that  82% of patients with osteonecrosis had at least one clotting disorder vs. 30% of controls

. Some of the more common abnormalities seen included IgG anti-cardiolipin antibodies, stimulated plasminogen activator, and plasminogen activator inhibitor-1. Similar results have been reported in other studies

. Other predisposing factors may include mutations in a multidrug resistance gene, genetic variation in alcohol-metabolizing enzymes, and abnormal type II collagen.

The table below lists many of the conditions associated with osteonecrosis (Mont et al 2006).

Direct Causes

Indirect Causes

Less Common

Hematologic disorders
Gaucher’s disease
Sickle cell disease/trait

Alcohol abuse
Renal failure
Organ transplant

Carbontetrachloride poisoning
Cushing’s disease

Natural History and Classification

At least 17 classification systems have been used. The majority are branches of the Ficat and Arlet System outlined below. This classification system also outlines the natural progression of the disease


Imaging Findings


Asymptomatic patient with signs of osteonecrosis on MRI only


Symptomatic patient with signs seen on MRI only


Diffuse sclerosis, subchondral cysts. No femoral head collapse


Subchondral fracture (“Crescent sign”) with or without collapse


Collapse of the femoral head and secondary degenerative joint disease

The Kerboul Angle, also known as the “Combined Necrotic Angle” is a system used to quantify size of the lesion. To calculate, first the center of the femoral head is identified. Two lines are then drawn from this point to the borders of the lesion on both AP and Lateral radiographs. The sum of the angles on the AP and Lateral radiographs is the Kerboul angle. Lesions are classified as small, medium, or large:

  • Small: 160º or less
  • Medium: 161º – 199º
  • Large: 200º or more

Patient History and Physical Findings

The most common presenting complaint is groin pain. Pertinent patient history includes history of steroid use, alcohol abuse, and clotting abnormalities. Other relevant history may relate to the associated factors in the above table. Mechanical symptoms are may also be present once collapse of the femoral head has occurred. Physical exam findings include a painful, decreased range of motion, which worsen with degeneration of the joint.

Imaging and Diagnostic Studies

MRI is the modality of choice for diagnosing and monitoring Stage 0-I disease (99% sensitivity and specificity)

and is preffered over bone scans. Both hips should be imaged in idiopathic cases, as osteonecrosis is often found in the contralateral, asymptomatic hip. A well-demarcated, crescent-shaped lesion in the subchondral bone is typically seen. T1 images reveal a single-density line demarcating the normal-ischemic bone interface. T2 images reveal a double-density line representing an area of hypervascular granulation tissue. Bone scans are not used as frequently due to a high rate of false positives.

Once the disease has progressed to Stage II, plain films are typically the preferred imaging modality. Radiographic changes appear months after disease onset.

Differential Diagnosis

Transient osteoporosis of the hip
Femoroacetabular impingement


Nonoperative therapy

Observation of Asymptomatic Lesions:

When a patient develops ON for the first time they will usually present with pain in just one hip. As these patients are being worked up or treated, ON is often incidentally found in the contralateral hip on x-ray or MRI. Min et al published a study in 2008 that examined the natural history of asymptomatic hips. At an average follow-up of 8 years, only 23% of all asymptomatic hips had collapsed. The size of the lesion was highly predictive of outcome. Survival was 100% for lesions occupying less than 15% of the femoral head, 79% for medium-size lesions (15-30% of the femoral head), and 9% for lesions larger than 30% of the femoral head

. This study suggests that small lesions could potentially be observed.However, a similar study published in 2004 examined only the fate of very small lesions and found that 73%  went on to collapse within 12 years. This group concluded that detecting collapse in very small Stage I lesions was difficult and that the patients should be followed very closely for many years

In summary, observation of small, asymptomatic lesions is controversial. Collapse rates of 23% – 73% have been reported. Medium-sized, asymptomatic lesions should be treated. Large lesions have a poor prognosis and treatment is largely directed by age and overall health.

Observation of Symptomatic Lesions:

Stulberg et al performed a randomized control trial comparing nonoperative management to core decompression.Operative management was more successful at all stages. Their conclusion was that core decompression produced better results than conservative management {ref: 2060201}

Mont et alperformed a meta analysis that included 819 hips treated nonoperatively and 1206 hips treated by core decompression. Patients in all stages did better with operative management. This study concluded that there is currently no role for nonoperative management in symptomatic hips (including protected weight bearing) {ref: 8595753}


Encouraging results have been reported for numerous medical and biophysical therapies, including statins, bisphosphonates, low molecular weight heparin, stanozolol, iloprost, extracorporeal shock-wave therapy, hyperbaric oxygen, puerarin (herbal), and electromagnetic therapy.


Pritchett et al published a retrospective review in 2001 that reported the incidence of osteonecrosis in 284 patients who had received high dose corticosteroids while concurrently on a statin. At 7.5 years follow-up MRI’s were performed on all patients who complained of hip pain. Only 3 patients (1%) developed osteonecrosis which is significantly lower than the historical value of 5% for patients receiving high-dose steroids

. A similar effect has been demonstrated in a rabbit study


A prospective, randomized, controlled trial has reported the results of 54 patients with large pre-collapse or early post-collapse lesions randomized either into a group that received observation only or a group treated medically with alendronate for 25 weeks. Seven percent (2/29) of the lesions treated with alendronate had progressed at a minimum 2-year follow-up versus 76% (19/25) of the controls

Low Molecular Weight Heparin

Enoxaparin has been reported to decrease disease progression in patients with an underlying thrombophilia and/or hypofibrinolysis and Stage I or II disease

Operative Treatment – Femoral Head-Sparing Techniques

Core Decompression

Ficat and Arlet first described core decompression in 1964while investigating painful hipsin patients who had normal radiographic findings.The rationale behind the procedure is to reduce intraosseous pressure in the femoral head, thereby improving arterial inflow and potentially decreasing pain.The technique entails directing a-wire into the necrotic area underfluoroscopicguidance, which is then over-drilled with a drill and trephine. Typically the patient is kept on protective weight bearing for approximately six weeks to reduce the incidence of proximal femoral fracture, which is the main complication of this technique. In general, core decompression is indicated for precollapse lesions with less than 30% involvement of the femoral head. A success rate of 84% for Stage I lesions and 65% for Stage II lesions has been reported {ref:  8595753}.Efficacy of the procedure drops significantlyonce the femoral head develops a subchondral fracture

A modification of this technique has been described and entails involves multiple percutaneous small drillings rather than one larger core

. Lower morbidity and the fact that it can be done on an outpatient bases are among the rationale for the variation. Similar results to core decompression and zero complications have been reported

Another variation includes placement of a tantalum rod in the core tract. This technique theoretically provides subchondral support and decreases the risk of proximal femoral fracture. Accurate placement of the device is key. The one or two studies published on this technique have shown comparable or slightly better results than those of standard core decompressions {ref: 17079367}. Indications for this technique remain the same as for standard core decompression.One issue to keep in mind is that if these patients eventually require a total hip arthroplasty, the tantalum rod will require removal. 

Nonvascularized bone grafting

Cortical Strut Grafting (typically with a fibular allograft) is not commonly used today. Long-term success rates have been poor (as low as 30%) and future conversion to total hip arthroplasty is complicated. However, some studies have reported promising results using cortical strut grafting combined with bioactive materials such as autogenous bone marrow cells, demineralized bone matrix, and bone morphogenic protein

Two other nonvascularized bone grafting procedures termed the “Lightbulb” and “Trapdoor” procedures may be used for post-collapse disease. These procedures essentially entail an open surgery in which a window is cut in the femoral neck or femoral head and the subchondral void is packed with various types of bone graft

. One study had reported a 83% success rate for patients with Stage III disease and 33% for those with Stage IV disease at a 4-5 year follow-up using the trapdoor procedure

Vascularized bone grafting

In general, free vascularized fibular grafting is considered for patients <35 years-old with pre-collapse lesions or those with collapse of <2mm.The procedure typically employs two teams, one prepares the proximal femur while the other harvests the vascularized fibula graft. The free fibulaand cancellous bone graftare then inserted into a core tract and the peroneal vessels are anastomosed to either the lateral circumflex artery or the ascending branch of the first perforator.Patients are typically kept on protected weight bearing for ~6 months postop.In addition to providing structural support, this procedure theoretically provides some amount of revascularization of the femoral head. Results with this technique have been highly variable {ref: 12690862}. Complications include donor site morbidity (~24% rate){ref:  9692940} and proximal femur fracture (~2.5% rate)

. Lastly, conversion to total hip arthroplasty in patients who have undergone this procedure requires burring laterally to prevent the prosthesis from being placed in varus.

Proximal femoral osteotomy

Proximal femoral osteotomy is not widely accepted as a standard treatment in the US, however it has had more consistent results and is more commonly used in Asia. The rationale is to remove the necrotic or collapsing segment from the weight-bearing zone of the femoral head.This procedure should be considered in young patients with small to medium-size lesions (or a combined necrotic angle of less than 200). Ongoing steroid use should be considered a contraindication. The procedure is technically difficult and nonunion rates can be high. Similar to vascularized fibular grafts, success rates in the U.S. have been highlyvariable(as low as 30%).Additionally, the osteotomy can significantly complicate future conversion to THA due to the fact that anatomy is often distorted and retained hardware can be problematic.

Operative Treatment – Femoral Head-Sacrificing Techniques


One of the biggest advantages of hemiresurfacing is that proximal bone stock is conserved and the femoral canal is not violated,making later conversion to a THA easier and more successful. This is particularly important in younger patients who will likely require additional procedures. Additionally, patients are able to maintain a higher level of activity compared to those with a total hip replacementand wear debris is not an issue as opposed to metal on metal arthroplasty. Disadvantages include unpredictable relief of groin pain, which is a complaint in up to 20% of these patients. Also, this is a temporizing measure as failure will eventually occur secondary to acetabular cartilage erosion.Short-mid term results have been relatively encouraging with 5-7 year success rates are approximately 70-90%. At 10-15 years there has been a dramatic decrease in survivorship.

Metal on metal resurfacing

Metal on metal resurfacing was first used in the 1960’s and had poor results. It was reintroduced in the 1990’s and much improved results were seen with the improved manufacturing technology. The rate of femoral neck fracture has been reduced to 0.8% with improved technique and patient selection. Metal on metal resurfacing has similar advantages to hemiresurfacing and eliminates groin pain much more reliably. Complications include the the development of local reactions to metal wear debris and high serum levels of metal ions, however complications from this have yet to be demonstrated. Early results have been excellent with up to 95% survivorship at 5 years.


Hemiarthroplasty has poor long-term results. The high activity level in this patient population results in increased polyethylene wear and osteolysis. This technique should not be used in these patients.

Total hip arthroplasty

Historical results of total hip arthroplasty for the treatment of osteonecrosis have been poor secondary to the high activity level of this population. However short to mid-term results with newer bearing surfaces and uncemented techniques have been encouraging. Success rates of up to 89% have been reported at 15 years follow-up making THA a reliable, long term option. Many argue that any surgery which may compromise future conversion to THA (such as osteotomies and vascularized bone grafts) should be used with discrimination. Some suggest that any patient over the age of 30 with collapse should be treated symptomatically until symptoms warrant arthroplasty.

Treatment Summary

When considering treatment options, there are four factors to take into consideration: 1) Patient age; 2) Activity level and general health; 3) Size and location of lesion; 4) Stage of the lesion. Small, pre-collapse lesions are typically treated with core decompression +/- bone grafting. Medium-sized precollapse lesions and postcollapse lesions with less than 2mm depression can be treated with more aggressive head-conserving approaches in younger patients. However, in light of recent improvements in materials and techniques, patients over the age of 40-50 may be better served with symptomatic treatment until symptoms warrant total hip arthroplasty. Medications such as bisphosphonates should also certainly be considered for all of the aforementioned patients. The femoral head is unlikely to be salvaged if the lesion is extensive (>30% of the femoral head) and involves the weight-bearing zone or if there is greater than 2mm of collapse. Additionally, if there is marked acetabular involvement, total hip arthroplasty is the only appropriate choice.

Pearls and Pitfalls

Postoperative Care

Include immediate postoperative care and rehabilitation


Include functional and prosthetic survivorship data as applicable


Include overview of complications

Selected References

Mont et al. OKU 3: Hip and Knee Reconstruction. AAOS. 2006. Ch. 44: Osteonecrosis of the hip. 511-19.


AVN Hip.jpg (image/jpeg)