Spontaneous Osteonecrosis of the Knee (SPONK/SONK)

Spontaneous osteonecrosis of the knee (SPONK/SONK) is the most common form of osteonecrosis around the knee. It must be distinguished from secondary osteonecrosis, which carries systemic risk factors and often multifocal disease. The current understanding is that SONK begins as a subchondral insufficiency fracture in osteopenic bone rather than as primary vascular occlusion.

Why early diagnosis matters. Recognition on MRI before radiographic collapse allows a trial of protected weight bearing that can prevent progression in small lesions. Larger lesions or established collapse usually require unloading osteotomy or arthroplasty, which makes accurate lesion sizing and staging essential for surgical planning.

Who. Peak age is 60-70 years, with a female predominance of 2-3:1; the classic patient is a woman over 60. The incidence is underestimated, because many cases are labelled as osteoarthritis or a meniscal tear.

Where. The medial femoral condyle is involved in greater than 90 percent of cases, at its weight-bearing surface, and the disease is bilateral in less than 5 percent, unlike secondary osteonecrosis. Lateral condyle or tibial plateau involvement suggests an alternative diagnosis or secondary osteonecrosis. Lateral-condyle SONK does occur, though less commonly, so inspect both condyles rather than assuming every lesion is medial.

How fast. The sudden severe pain is often mistaken for a meniscal tear or a flare of osteoarthritis. Collapse can occur within weeks to months.

The fracture comes first. Current evidence supports that SONK is initiated by a subchondral insufficiency fracture in osteopenic bone rather than by primary avascular necrosis. Reduced trabecular thickness and connectivity in elderly patients lower the threshold for microfracture, so osteopenic subchondral bone fails under normal physiological load. Repetitive daily weight bearing then propagates the fracture line.

The necrosis follows. The fracture disrupts the local blood supply, isolating a segment of subchondral bone and producing focal osteonecrosis. This explains the typical single-site medial condyle lesion in elderly patients without systemic risk factors.

The meniscus concentrates the load. Medial meniscal extrusion increases focal load on the condyle. A posterior root tear destroys hoop stress, the meniscus extrudes, and load concentrates on exactly the medial condyle that fails, so root tears both predispose to and accelerate the overload.

From fracture to collapse. Loss of structural support lets the weight-bearing surface flatten under continued loading. Joint incongruity then accelerates cartilage wear and secondary osteoarthritis; the medial joint line narrows and varus malalignment develops, which further increases medial compartment load in a vicious cycle. Fibrocartilage or fibrovascular tissue fills the defect as a repair attempt, and if collapse persists the end stage is secondary degenerative change.

Secondary osteonecrosis is a different disease. It involves true vascular occlusion from fat emboli, coagulopathy or steroid-induced adipocyte hypertrophy, and is often multifocal.

View larger

View larger

Examiners classically divide osteonecrosis around the knee into three types: spontaneous and secondary, contrasted above, and post-arthroscopic osteonecrosis (PAON), which is equally testable.

What it is. New osteonecrosis appears in the operated compartment, not present pre-operatively, in an older patient weeks to months after knee arthroscopy, classically after partial medial meniscectomy. The leading explanation is again a subchondral insufficiency fracture in the now mechanically altered, and often chondromalacic, compartment, rather than thermal or laser injury alone. Management mirrors SONK and is size- and collapse-based: protected weight bearing for small lesions, UKA or osteotomy for large or collapsed lesions.

The medicolegal trap. Because it appears after surgery, it can be mistaken for a surgical complication, so a pre-operative MRI documenting normal subchondral bone is invaluable.

Two measures run side by side: the stage, which describes how far the structure has failed, and the size of the lesion.

View larger

Lesion size. Size drives the treatment decision, and the teaching bands combine width and area.

Collapse depth. Depth of collapse greater than 2 mm on sagittal images is an independent predictor of failure of conservative treatment.

Know where the area figures come from, because the classic threshold is lower than the bands above. Rozing, Insall and Bohne studied 90 knees and found the prognosis depends on the size of the radiolucent lesion, with osteoarthritis likely to develop once the lesion exceeds 2.3 square centimetres, a figure still quoted as the original cut-point. The 40 percent width band is the one used in the later literature to separate lesions. The bands in the table are a practical composite rather than a single validated rule, so do not reassure a patient with a 3 square centimetre lesion on the strength of the "small" row: by Rozing's threshold that lesion is already in the range that tends toward arthritis. Treat area and width as concordant risk signals and act on the worse of the two.

View larger

View larger

Not every painful older knee with bone marrow oedema is SONK. Transient bone marrow oedema syndrome (transient osteoporosis) of the knee is a self-limiting, reversible cause of pain and oedema that must not be over-treated as if it were a collapsing insufficiency fracture.

On MRI. Transient bone marrow oedema syndrome shows diffuse oedema without a discrete subchondral fracture line, without a focal subchondral low-signal collapse zone and without articular flattening. SONK has the focal subchondral fracture line and, later, collapse; a subchondral band of marked hypointensity adjacent to the endplate favours an insufficiency fracture, not transient oedema.

Natural history. Transient bone marrow oedema syndrome resolves spontaneously over weeks to months with protected weight bearing and analgesia, and does not progress to collapse. It can be regional or migratory, and the hip is the classic site.

Why it matters. Misreading transient oedema as SONK risks unnecessary surgery; misreading SONK as transient oedema risks allowing a preventable collapse. Look specifically for the subchondral fracture line.

View larger

History. The pain is sudden and severe, medial, often after minor activity or at night, and the patient frequently denies any injury. Night pain that wakes the patient distinguishes it from mechanical meniscal symptoms. The pain may improve slightly and then worsen with collapse, and function suffers: antalgic gait, reduced walking distance, difficulty with stairs.

Risk factors. Ask specifically about steroids, alcohol, haemoglobinopathy such as sickle cell, and Gaucher disease. Their presence points to secondary osteonecrosis.

Examination.

• Gait: antalgic; a varus thrust may be present if collapse is advanced
• Swelling: mild effusion, with medial joint-line tenderness
• Range of motion: limited by pain, with a possible flexion contracture if chronic
• Stability: usually stable, unless there is secondary ligamentous laxity from collapse
• Alignment: measure the mechanical axis; varus deformity is common in advanced cases

Telling it apart. Sudden onset without trauma in an elderly patient with night pain should raise suspicion for SONK. Mechanical locking or catching suggests meniscal pathology, and a gradual increase in pain over months is more typical of primary osteoarthritis. Always obtain MRI in this demographic with unexplained medial knee pain.

Radiographs first. They are the first-line test, but a normal radiograph does not exclude early SONK.

• Views: weight-bearing AP, Rosenberg (45-degree flexion PA), lateral and skyline
• Look for: subtle flattening of the medial femoral condyle, subchondral sclerosis, early joint-space narrowing
• Compare with the contralateral knee; varus alignment greater than 5 degrees increases suspicion

View larger

View larger

View larger

MRI is the gold standard. It makes the diagnosis before radiographic change. Obtain coronal and sagittal T1, T2, STIR and proton-density sequences; staging is performed on the coronal and sagittal T2-weighted and STIR images.

The diagnostic finding is the line, not the oedema. Look for a subchondral T2 hypointense line in the affected condyle with extensive surrounding bone marrow oedema-like signal. The oedema is the eye-catching part and the part that is not specific: avascular necrosis, osteochondritis dissecans and osteoarthritis all produce it. The subchondral line, its length and its distance from the articular surface are what make the diagnosis.

View larger

View larger

What to measure. Measure on the coronal slice of maximum width, and size the low-signal subchondral component rather than the surrounding oedema field.

• Width as a percentage of the entire medial femoral condyle width on the same slice
• Area in square centimetres
• Depth of collapse in millimetres, on sagittal images
• Extent of bone marrow oedema on STIR, which predicts pain, healing time and healing potential, although oedema extent alone does not define prognosis

Look for a medial meniscus posterior root tear, do not merely exclude one. Root tears and chondral lesions are found in many patients with subchondral insufficiency fracture and are considered contributory. Finding one does not overturn the diagnosis, it often explains it, and in a suitable patient it is a separate problem that may need addressing in its own right. Also look for ligament injury and for the secondary osteonecrosis pattern: serpiginous, geographic lesions, often bilateral or multifocal.

Soft-tissue signs. The metaphyseal burst sign and oedema tracking beside the distal femur or vastus fascia can support the diagnosis. They do not replace identification of the subchondral lesion.

View larger

Long-leg alignment film. Obtained when planning high tibial osteotomy or UKA, to measure the mechanical axis deviation, medial proximal tibial angle and lateral distal femoral angle. It determines whether a corrective osteotomy can unload the lesion or whether arthroplasty is required.

The decision. Lesion size, collapse and the patient decide treatment, and the early aim is to unload the condyle, by protected weight bearing or osteotomy. Before choosing, assess the lateral compartment, the patellofemoral joint, and the patient's age and activity.

• Width less than 40 percent, collapse less than 2 mm and a patient compliant with protected weight bearing: conservative care. Small lesions with minimal oedema may be observed with protected weight bearing and repeat MRI at 8-12 weeks
• Width 40-60 percent: borderline; consider bisphosphonates and close monitoring
• Width greater than 50 percent or collapse greater than 2 mm: a high likelihood of progression, so counsel toward surgical options early

Conservative care. Protected weight bearing is the mainstay, with a bisphosphonate as an adjunct, and review is expectant: serial radiographs and clinical review until the pain settles or collapse occurs.

1. Weeks 0-6: touch weight bearing with crutches or a walker; avoid full weight bearing, to prevent propagation of the subchondral fracture
2. Weeks 6-12: partial (50 percent) weight bearing, increasing gradually if pain allows; repeat radiographs or MRI to assess healing
3. Months 3-6: full weight bearing as tolerated once pain subsides and imaging shows resolution of oedema or stabilisation of the lesion

Surgery. Large lesions and collapse usually come to surgery, as can failed conservative care.

Osteotomy or UKA. The choice depends on age, activity level, alignment and lesion size. Osteotomy is preferred in the younger, active patient with correctable varus and preserved lateral cartilage. UKA is favoured in the older patient with a larger lesion, collapse or failed conservative care, where joint preservation is less critical.

Size stops mattering once you replace the joint. In a series of 71 knees taken to arthroplasty for end-stage SONK, there was no correlation between the size of the necrotic lesion and the failure rate, and unicompartmental arthroplasty gave slightly better function than total knee arthroplasty (WOMAC 1.0 vs 1.6, p=0.04) at a mean of 6.6 years. Lesion size predicts whether conservative care will fail; it does not predict how the arthroplasty will do, so it should not be used to argue a patient out of a unicompartmental implant.

Untreated disease. Subchondral collapse of the weight-bearing condyle is the pivotal event, converting a potentially reversible lesion into an irreversible one. Secondary osteoarthritis of the medial compartment follows, often rapidly once collapse occurs, with progressive varus malalignment, chronic pain and functional decline.

Complications of treatment.

• High tibial osteotomy: delayed union or nonunion, loss of correction, over- or under-correction, intra-articular fracture, common peroneal nerve injury, hardware irritation
• Unicompartmental knee arthroplasty: aseptic loosening, polyethylene wear, bearing dislocation (mobile-bearing designs), progression of disease in the other compartments, persistent pain, periprosthetic fracture, conversion to TKA
• Bisphosphonate therapy (off-label adjunct): rare atypical femoral fracture and osteonecrosis of the jaw with prolonged use
• Prolonged protected weight bearing: deconditioning, falls and venous thromboembolism risk in the elderly