Knee Osteonecrosis 

Updated: Jan 23, 2018
Author: Amar Rajadhyaksha, MD; Chief Editor: Thomas M DeBerardino, MD 



Osteonecrosis is a disease characterized by a derangement of osseous circulation that leads to necrosis of osseous tissue. Osteonecrosis of the knee has been divided into two separate entities: spontaneous osteonecrosis of the knee (SPONK) and secondary osteonecrosis.[1, 2]  A rare form that is sometimes considered a separate entity is postarthroscopy osteonecrosis, which affects patients who have undergone arthroscopic knee surgery, most commonly meniscectomy.[3]

History of the Procedure

Ahlback et al first reported on osteonecrosis of the knee in 1968.[4] The osteonecrosis that Ahlback described now is referred to as spontaneous osteonecrosis of the knee (SPONK).


In osteonecrosis, the lesion can extend to the subchondral plate and result in collapse of the necrotic segment. This can lead to disruption of the joint line, resulting in painful secondary arthritis.



The knee is the second most common site for osteonecrosis, but it is affected much less often than the hip. The true incidence of the disease is unknown, but osteonecrosis of the knee is believed to account for approximately 10% of cases of osteonecrosis.


The etiologies of both spontaneous osteonecrosis of the knee (SPONK) and secondary osteonecrosis are poorly understood. A systematic review by Hussain et al concluded that the meniscus plays a central role in the development of the disorder. They hypothesize that meniscectomy and meniscal tears, particularly of the medial meniscus posterior root, increase contact pressures and create an environment conducive to insufficiency fractures. These authors suggest that SPONK would be more accurately described as subchondral insufficiency fractures of the knee.[5]

Trauma may be a causative factor in SPONK. SPONK commonly is seen in elderly women with osteoporosis and may be associated with insufficiency fracture secondary to low bone mineral density in women over 60.[6, 7] See the Fracture Index WITH known Bone Mineral Density (BMD) calculator. Osteoporotic bone is more susceptible to microfracture with minor trauma, which leads to fluid accumulation in the marrow space. The intraosseous edema causes increased pressure within the marrow cavity, which may, in turn, lead to ischemia and necrosis.

Another possible cause may be vascular compromise to subchondral bone, resulting in osseous ischemia and subsequent edema. Again, edema leads to a rise in intraosseous pressure that further compromises blood flow, thus worsening ischemia and necrosis.

Although the etiology of secondary osteonecrosis is unknown, several risk factors are associated with the disease. Corticosteroid use is the most significant risk factor; other risk factors include the following[8] :

The pathogenesis of this condition is poorly understood. One possible mechanism is microvascular disruption in the subchondral bone that causes infarction. This circulatory compromise leads to bone marrow edema, with resultant ischemia and necrosis.

The mechanism by which corticosteroids contribute to osteonecrosis also is unclear. One hypothesis is that an increase in the size of the marrow fat cells decreases circulation and leads to ischemia. Other possible contributors to the etiopathogenesis are coagulopathies, fat emboli, and thrombus formation.


Clinical presentation of spontaneous osteonecrosis of the knee (SPONK) and secondary osteonecrosis is summarized in the following table.

Table. Clinical Presentation of SPONK and Secondary Osteonecrosis (Open Table in a new window)

Physical Characteristic


Secondary Osteonecrosis


Typically >55 y

Typically < 55 y

Sex (male-to-female ratio)



Associated risk factors


Corticosteroids, alcohol, SLE, sickle cell disease, caisson disease, Gaucher disease, fat emboli, thrombus formation

Other joint involvement


Approximately 75%


99% unilateral

Approximately 80% bilateral

Condylar involvement

One (usually medial femoral condyle or either tibial plateau)



Epiphyseal to the subchondral surface

Diaphyseal, metaphyseal, epiphyseal


Commonly sudden onset of pain and increased pain with weightbearing, stair climbing, and at night

Usually long-standing insidious pain; patient may have symptoms and signs of an underlying disorder, such as SLE


Pain localized to affected area; small synovitis or effusion may occur; ligaments are stable; range of motion may be limited by pain or effusion

Pain is difficult to localize; ligaments are stable; range of motion is grossly intact but may be limited by pain

Differential diagnosis

Osteonecrosis of the knee is commonly mistaken for osteochondritis dissecans, primary osteoarthritis, meniscal tears, bone bruises, transient osteopenia of the knee, and pes anserine bursitis.[9, 10] Therefore, it is important to identify osteonecrosis correctly and to differentiate between SPONK and secondary osteonecrosis, so as to treat each patient appropriately. Spatiotemporal gait analysis has been advocated as a means of distinguishing osteonecrosis of the knee from other knee disorders and determining its functional severity.[11, 12]


Indications for intervention are guided by the clinical signs and symptoms of osteonecrosis combined with radiographic or magnetic resonance imaging findings. A thorough history and physical examination are necessary to recognize associated risk factors and to differentiate spontaneous osteonecrosis of the knee (SPONK) from secondary osteonecrosis. 

Relevant Anatomy

The blood supply to the knee joint comes from two major sources: the descending genicular artery (from the femoral artery) and the popliteal artery. Major branches of the descending genicular artery include the saphenous, deep oblique, and an articular branches.

The popliteal artery gives off numerous muscular branches and five major articular branches. These articular arteries anastomose to form extensive collateral circulation around the knee joint.


There are few contraindications to surgical intervention. Cardiovascular or respiratory disease that would compromise the patient's ability to cope with anesthesia must be recognized. Obvious disorders aside (ie, severe systemic disease, sepsis), patients with osteonecrosis of the knee (especially secondary) often are young and have few surgical contraindications.



Laboratory Studies

Lab tests are limited for osteonecrosis. Some tests that may be helpful include the following:

  • Sickle-cell screening, especially in African Americans

  • Lipid profile

  • Screening for coagulopathies (eg, protein S and protein C deficiencies, factor V Leiden disease)

Imaging Studies

Plain radiographs

  • Always obtain anteroposterior (AP), lateral, and tunnel view plain radiographs when entertaining the diagnosis of osteonecrosis. See image below.

    Plain radiograph of a knee affected by osteonecros Plain radiograph of a knee affected by osteonecrosis; note the cystic and sclerotic lesions in a serpiginous pattern.
  • Initially, no abnormalities may be seen, but as the disease progresses, flattening of the weightbearing portion may occur.

  • A radiolucent area forms in the subchondral bone, surrounded by a rim of sclerosis. Later in the disease course, the subchondral bone collapses, leading to secondary arthritic change and, possibly, a valgus or varus deformity.

  • In spontaneous osteonecrosis of the knee (SPONK), these lesions usually are seen in the medial femoral condyle, whereas in secondary osteonecrosis, both the medial and lateral sides may be affected.[13]


  • MRI can depict osteonecrosis before it is visible on plain radiographs. See image below.

    MRI confirming diagnosis of osteonecrosis; bilater MRI confirming diagnosis of osteonecrosis; bilaterality suggests secondary osteonecrosis.
  • MRI can also reveal the extent of disease more precisely than plain radiographs.[14]

  • In SPONK, lesions are isolated to a single condyle (usually medial) or plateau.

  • In SPONK, on T1-weighted images, osteonecrosis is seen as a discrete area of low signal intensity, replacing the high-intensity signal normally produced by marrow fat.

  • The T2-weighted image shows an area of low signal intensity surrounded by a high-intensity signal caused by edema.

  • The T1- and T2-weighted images in secondary osteonecrosis are similar to those in SPONK but are larger, are more serpiginous, and may be multifocal. Lesions usually are seen in the epiphyseal region of the distal femur or proximal tibia.

Technetium-99m scans

  • Bone scans have been used to diagnose SPONK and may show osteonecrotic lesions before plain radiographs do.

  • In SPONK, bone scans usually show a localized area of radioisotope uptake in the medial femoral condyle.

  • Bone scans are less effective for diagnosing secondary osteonecrosis than for diagnosing SPONK.

  • Studies have reported that bone scans provide a correct diagnosis of secondary osteonecrosis in only 40-70% of cases.[15]

  • Because secondary osteonecrosis commonly presents bilaterally, bilateral symmetric uptake may be read incorrectly as degenerative changes or as a negative study.

  • Bone scans generally are unreliable. Therefore, plain radiographs remain the initial imaging study of choice, with MRI for confirmation.

Because specificity and sensitivity of MRI are 98% in osteonecrosis, MRI is the diagnostic study of choice.

Other Tests

The following methods are outdated and are mentioned for the sake of completeness. Core biopsy, however, can be useful for pathologic diagnosis of osteonecrosis of the knee.

  • Venography

  • Interosseous pressure measurements

Diagnostic Procedures

Core biopsy may be useful for pathologic diagnosis of osteonecrosis of the knee.

Histologic Findings

Macroscopic pathology

Early in osteonecrosis of the knee, there may be slight discoloration and flattening of the articular cartilage. As the disease progresses, a line of demarcation becomes evident, and an osteochondral flap overlies the area of osteonecrosis. Late in the disease course, secondary arthritic changes occur, leading to a cartilage defect filled with necrotic debris and to signs of osteoarthritis (eg, osteophyte formation, eburnated bone).[16]

Microscopic pathology

On microscopic examination, the osteonecrotic bone shows empty lacunae and fatty degeneration within the center of the lesion. The surrounding area shows evidence of osseous healing, including osteoblastic activity, fibrovascular granulation tissue, and cartilage formation.


Aglietti devised the following classification system for spontaneous osteonecrosis of the knee (SPONK), which was a modification of an earlier classification by Koshino[17] :

  • Stage I: Plain radiograph findings are normal. Diagnosis must be made from MRI or bone scan.

  • Stage II: Radiographs show flattening of the weightbearing portion of the condyle

  • Stage III: Radiographs show a radiolucent area surrounded by sclerosis

  • Stage IV: Radiographs show a more defined ring of sclerosis and subchondral bone collapse forming a calcified plate, sequestrum, or fragment

  • Stage V: Narrowing of the joint space, osteophyte formation, and/or femoral and tibial subchondral sclerosis is shown

For secondary osteonecrosis, Mont and Hungerford developed the following staging system, which is a modification of the Ficat and Arlet staging of osteonecrosis of the hip[18, 19, 20, 21, 22, 23, 24] :

  • Stage I - Plain radiographs reveal no change, but MRI scan findings are positive

  • Stage II - Radiographs reveal cystic and sclerotic changes in the distal femur and/or proximal tibia

  • Stage III - Subchondral collapse is seen as the crescent sign

  • Stage IV - Evidence of degenerative changes is present on both sides of the joint (eg, joint space narrowing, osteophytes); at this stage, it may be difficult to distinguish osteonecrosis from osteoarthritis of the knee on plain radiographs



Medical Therapy

Nonoperative and operative treatment options are available for osteonecrosis of the knee. Spontaneous osteonecrosis of the knee (SPONK) and secondary osteonecrosis can be treated nonoperatively when the patient is asymptomatic. However, once a patient becomes symptomatic, treatment options for the two entities differ.

Spontaneous osteonecrosis of the knee

Nonoperative treatment has been shown to produce good results in symptomatic patients with SPONK. Treatment encompasses a conservative regimen of protected weightbearing with crutches, analgesics, nonsteroidal anti-inflammatory medications, and physical therapy focused on strengthening the quadriceps and hamstring muscles. Lotke et al reported on 87 knees with SPONK: 36 of these knees were treated nonoperatively, with only one progressing to arthroplasty; the 35 remaining knees did well.[25, 26, 27]

Observational studies have suggested beneficial effects of bisphosphonates in SPONK. However, a double-blind, placebo-controlled study in 30 patients with SPONK found that the bisphosphonate ibandronate has no beneficial effect on clinical and radiological outcome over and above anti-inflammatory medication.[28]

Secondary osteonecrosis

Outcomes of nonoperative treatment in symptomatic secondary osteonecrosis are relatively poor. In one study of secondary osteonecrosis treated nonoperatively, 31 of 51 knees eventually required arthroplasty. Therefore, operative therapy usually is recommended once the patient is symptomatic.

Pharmacotherapy aimed at the putative pathophysiology of the disease has had mixed results. Medications that have been used include nifedipine and lipid-lowering agents such as gemfibrozil (Lopid).

Surgical Therapy

A systematic review of the literature on procedures to preserve the joint in patients with osteonecrosis of the knee found that core decompression prevented additional surgical treatment in pre-collapse knees with a failure rate of 10.4%. Autogenous and osteochondral grafts decreased the need for additional surgery in both pre-collapse (0%) and post-collapse knees (10.5%).[29]


Arthroscopic debridement for the treatment of osteonecrosis of the knee has had mixed results. Arthroscopy may not alter the natural course of the disease. Patients with SPONK may have degenerative tears of the menisci. Debridement of these tears does not improve osteonecrosis of the bone. In addition, partial meniscectomy has been hypothesized to cause further degeneration of the knee joint. Thus, arthroscopy is controversial, with questions arising on the possibility of increased interosseous pressure.

Osteochondral grafts

Studies of osteochondral allografts have yielded mixed results. Bayne et al used fresh allografts in six knees with spontaneous osteonecrosis of the knee (SPONK), with only one good result.[30] The authors suspect that these results were due to the poor compliance of elderly patients, resulting in allograft fragmentation. Three knees with steroid-induced secondary osteonecrosis also failed the grafting procedure. This may be due to continued use of corticosteroids, which may lead to poor vascularization of the graft and subsequent subsidence.

In contrast, Tirico et al reported excellent results with fresh osteochondral allografts in seven patients with SPONK lesions on the medial femoral condyle. On median follow-up of 7.1 years (range, 4.5-14.1 years), none of the patients had required additional surgery and none of the grafts had failed. All the patients expressed extreme satisfaction with the results.[31]

Some surgeons have focused on using osteochondral autografts. This procedure, commonly referred to as OATS (osteochondral autologous transfer system), was first introduced by Matsusue in 1993.[32] Using this procedure, Hangody et al reported a 2-5 year follow-up with good or excellent results in up to 90% of cases.[33] Other grafting methods may improve results, but further studies are required.[34]

High tibial osteotomy

High tibial osteotomy (HTO) has been used in patients with SPONK, with encouraging results.[35, 36] Aglietti et al described 31 patients treated with high tibial osteotomy, with 21 of these knees having ancillary bone grafting.[17] Of the 31 knees, 87% had excellent to good results at a mean follow-up of 6.2 years, and only two knees progressed to arthroplasty. Use of high tibial osteotomy in secondary osteonecrosis is limited because most of these patients have bicondylar femoral involvement and also may have tibial involvement.

Core decompression

The principle behind core decompression is reduction of interosseous pressure, thereby restoring adequate circulation. This procedure has been used with some success in the earlier stages of osteonecrosis. Core decompression is a lesser procedure than total knee arthroplasty and has been shown to delay the need for joint replacement.

Core decompression has been used with some success in SPONK. Forst et al reported successful outcomes in their study of 16 knees, in which core decompression of the femoral condyle was performed in precollapsed lesions.[37] Results have not been as encouraging in SPONK as in secondary osteonecrosis. Therefore, it should be reserved for refractory cases.

Mont et al reported on 79 knees that were treated for secondary osteonecrosis.[38] Forty-seven knees were treated with core decompression, and 32 were treated nonoperatively. For core decompression, clinical success was achieved in 73% (34 of 47) of knees (good to excellent Knee Society Scores) at a mean follow-up of 11 years (range of 4-16 y). Radiographically, 17 of the 47 (36%) knees progressed to Ficat and Arlet stage III or IV, as opposed to 24 of 32 knees (75%) treated nonoperatively.

Unicondylar (unicompartmental) knee arthroplasty

Unicondylar arthroplasty has been used with success in SPONK, as the disease usually is confined to one condyle, as in the image below.[39] This procedure is not, however, recommended for secondary osteonecrosis, as the disease can affect both condyles.[40] Marmor reported an 89% success rate in a study of 34 knees with medial femoral condyle osteonecrosis treated with unicondylar replacement.[41, 42, 43]

Two separate studies found that long-term survival after unicompartmental knee arthroplasty (UKA) was greater than 94% after 10 years.[44, 45]

Knee osteonecrosis. Plain radiograph of a unicompa Knee osteonecrosis. Plain radiograph of a unicompartmental knee arthroplasty.

Total knee arthroplasty

Knee arthroplasty is indicated in the late stages of the disease, when patients have severe pain that does not respond to other treatments, as in the image below. Total knee arthroplasty is an appropriate intervention for late-stage secondary osteonecrosis with degenerative changes, for patients with severe pain, or for those with functional disability. It has had varying success with SPONK. Bergman and Rand reported that 87% of 38 knees treated with total knee arthroplasties had excellent or good results.[46] Of those knees, 27 had SPONK and nine had secondary osteonecrosis. These results were inferior to those of total knee replacements performed for other diagnoses.

Plain radiograph of a total knee arthroplasty perf Plain radiograph of a total knee arthroplasty performed for osteonecrosis of the knee.

Ritter et al compared 32 knees with SPONK to 63 osteoarthritic knees.[47] The success rate was 82% in SPONK knees, with no statistical significance in success rates between the two groups. For secondary osteonecrosis, Mont et al reported on 31 knees treated with total knee arthroplasty; all of these patients had a history of corticosteroid use.[48] After 8.2 years of mean follow-up, results excellent to good in 55% of knees.

In a study of 3182 TKAs that were performed between 2005 and 2011, patients with knee osteonecrosis were associated with increased frequency of manipulation under anesthesia which is associated with poor postoperative ROM. Diabetes, high cholesterol levels, and tobacco smoking and preoperative knee ROM of less than 100° were also associated with increased frequency of MUA. The researchers concluded that patients with multiple risk factors may benefit from preoperative counseling to set realistic ROM expectations.[49]  

Preoperative Details

Core decompression

Determining areas of pain on the tibia and/or femur is essential. The lesion should be delineated carefully on MRI examination to determine which areas need to be cored and to determine extent of involvement.

Intraoperative Details

Core decompression

Intraoperatively, when coring the knee, a tourniquet is applied but rarely is inflated. For femoral coring, a small incision is made on the lateral or medial side under fluoroscopic control, just above the flare of the condyle, as in the image below. Then, a 3- to 6-mm Michelle trephine is inserted through the metaphyseal flare into the lateral and medial condyles up to within a few millimeters of the subchondral plate.

Knee osteonecrosis. Postoperative radiograph of a Knee osteonecrosis. Postoperative radiograph of a core decompression performed from the lateral side; the point of entry of the trephine, seen as a break in the cortex, is demarcated by the white arrow.

For tibial coring, the trephine is introduced through a small incision just medial to the tibial tubercle, being sure to avoid the medial saphenous nerve. The trephine then is guided into the medial and lateral plateaus. It is helpful to turn the trocar while advancing it within the bone to clear the teeth and obtain the best biopsy possible.

Postoperative Details

Core decompression

Fifty percent weightbearing with a cane or crutch for 6 weeks is advised, after which the patient may return to full weightbearing.


For excellent patient education resources, visit eMedicineHealth's First Aid and Injuries Center and Osteoporosis Center. Also, see eMedicineHealth's patient education articles Knee Pain, Knee Injury, and Knee Joint Replacement.


Core decompression

Complications are minimal when the surgeon is experienced. The operation causes minimal blood loss and is a quick procedure. Possible complications include infection, fracture, and failure of the procedure to alleviate symptoms.

Total knee arthroplasty

Although more extensive than core decompression, this is a relatively safe and effective procedure. Possible problems include, but are not limited to, reactions to anesthesia, deep venous thrombosis (DVT), injury to a nerve or blood vessel (peroneal nerve palsy), fracture, infection, swelling, and dislocation of the patella.

Outcome and Prognosis

Studies have shown the prognosis of spontaneous osteonecrosis of the knee (SPONK) to be related directly to the size of the lesion.[50] Aglietti et al reported that lesions greater than 5 cm2 had a worse prognosis than lesions with areas less than 3.5 cm2. Juréus et al reported that six of seven patients with lesions greater than 40% of the anteroposterior radiographic view of the condyle at the time of diagnosis underwent major knee surgery, whereas none of 10 patients with a lesion of less than 20% ever underwent surgery.[51]

Prognosis of secondary osteonecrosis is dependent on 2 factors, stage and location. In 1 study, stage I lesions had successful outcomes in 83% of lesions, whereas only 33% of stage IV lesions at presentation had successful outcomes. Lesions that involved the epiphysis of the distal femur fared better than lesions that affected the metaphyseal and/or diaphyseal region. As opposed to SPONK, size of the lesion has not been a valuable prognosticator. Knees with osteonecrosis in the distal femur and proximal tibia do not do worse than knees with lesions isolated to the distal femur or proximal tibia.

Future and Controversies

Atraumatic osteonecrosis is characterized by a poorly understood derangement of osseous circulation. There are no proven causes, only associated risk factors. Systemic corticosteroids and alcohol abuse continue to be the most commonly associated factors. Perhaps future treatment modalities will reverse the pathophysiology of the disease, rather than merely stopping progression or palliating symptoms.

The role of core decompression in atraumatic osteonecrosis has been questioned. However, many patients feel immediate relief after the procedure, and it also may halt the progression of disease. Core decompression is also a less invasive operative procedure than total knee arthroplasty.[52, 53]