Various etiologies of AVN include trauma (30-50% femoral neck fractures lead to AVN), steroids, alcohol, pancreatitis, hemoglobinopathies (i.e. sickle cell disease, polycythemia, etc.), dysbarism, and Gaucher's disease. Although the disease mechanism is not worked out, some proponents suggest that fat embolism is a common denominator. Fat emboli from fatty liver, destabilization and coalescence of plasma lipoproteins, and/or disruption of fatty bone marrow and other fat deposits are three mechanisms by which AVN can occur. In the case of dysbaric osteonecrosis formation of microthrombi from red cell aggregated and platelets on the gas bubble surface and not the actual bubble is the cause of mechanical vascular occlusion. With steroids abrupt cessation or dose changes can lead to hyperlipidemia and emboli from sudden release of stored fat globules in hepatocytes. Steroids prevent release of fat into the system by the hepatocytes (4). Other possible mechanisms for steroid osteopathy are related to steroid-induced hypophosphatemia (5).
The central factor in AVN is vascular insufficiency from occlusion leading to cellular anoxia and death of hematopoietic cells at 6-12 hours, osteocytes and other bone cells at 12-48 hours, and marrow fat cells at 2-5 days (1). Usually the femoral head cartilage is not affected because it is nourished by synovial fluid. The femoral head consists of a cortical shell of bone surrounding a mass of cancellous bone consisting of myeloid elements, fat, and sinusoids. The femoral head has an end-arterial vascular supply with poor collaterals. Similar vasculature exists in other bones which are also prone to AVN; these include: knees, humeri, tali, scaphoid, and lunate bones. Vascular insufficiency may be due to blocked arterial inflow or blocked venous outflow, leading to a rise in intramedullary pressure and compromising perfusion further (4).
Five stages are described in AVN. Stage I, consists of an asymptomatic hip, bone infarction demonstrated histologically, normal to subtle mottled lucency in the femoral head, photopenia on bone scan (due to no bone repair or formation of hydroxyapatite matrix). In Stage II, the hip is asymptomatic to mildly symptomatic and the femoral head contour is preserved, The infarct appears as an area of increased density at the periphery of the femoral head (due to laying down of new bone between necrotic trabeculae), MDP activity is increased (due to hyperemia and formation of hydroxyapatite crystal). In Stage III, the patient is symptomatic, the femoral head is flattened subtly with development of the crescent caused by fracture through dead subchondral bone, histology shows necrotic trabecula and marrow on both sides of the fracture line (fracture occurs because rate of bone resorption exceeds rate of repair) and there is failure of differentiation of mesenchymal cells into osteoblasts, MDP activity remains increased due to continued cellular activity. In Stage IV, pain is greater than in Stage III, collapse of the necrotic segment with step deformity of the femoral head, bone repair is ineffective, MDP activity remains increased due to cellular activity. In Stage V the pain is persistent, cystic changes are seen radiographically in both the femoral head and acetabulum, MDP activity remains increased due to continued ineffective repair (4).
Mitchell, et al. observed greater sensitivity for MRI compared to scintigraphy. In a series of 56 patients, MR was abnormal in all patients. In 80% of patients, a high signal inner border was present inside a high intensity peripheral rim (the double line sign). In 41 patients with AVN, the sensitivity of bone scintigraphy was 90%. In advanced stages, MRI failed to detect head flattening in six cases which was demonstrable on plain films.
Other investigators have also shown MRI to have greater sensitivity compared to radionuclide imagery. Markisz, et al. found a sensitivity of 100% for MRI compared to 81% for scintigraphy, while specificity for both was 100% (2). Thickman, et al. found a greater sensitivity for MR (98%) than for scintigraphy 86% although scintigraphy was more specific (79% vs. 71%) (6).
Overall, MRI is the most sensitive modality for early stages of osteonecrosis, but should be correlated with plain films in advanced stages. Scintigraphy is useful in Stage I disease over plain films, but is superseded by MRI in sensitivity in general. MR may also be useful in following progress with and without therapy for AVN.
2) Markisz JA, et al. Segmental patterns of avascular necrosis of femoral heads: early detection with MR imagery. Radiology 1987; 162:717-720.
3) Mitchell DG, et al. Femoral head avascular necrosis: correlation of MR imaging, radiographic staging, radionuclide imaging, and clinical findings. Radiology 1987; 162:709-715.
4) Kenzora JE (Editor), et al. Idiopathic osteonecrosis orthopedic clinics of North America. October 1985; p. 595-710.
5) Tumeh SS, McNeil BJ. Disease of the hip in children. ACR Nuclear Radiology, Syllabus IV.
6) Thickman D, et al. Magnetic resonance imaging of avascular necrosis of femoral head. Skeletal Radiology 1986;15:133-140.
J. Anthony Parker, MD PhD, email@example.com