Osteonecrosis of the Femoral Head,
at the Pre-prosthetic Stage
P. Hernigou
Hôpital Henri Mondor - Créteil - France
Osteonecrosis of the femoral head (ONFH) is a disabling condition that affects mainly younger subjects, in the midst of their working lives; to this day, it has remained a devastating disease. Its treatment stands at the borderline between Medicine and Surgery, and requires a thorough understanding of the pathogenesis, the natural history, and the treatment options as a function of the different disease stages, which are best assessed with further investigations.
Osteonecrosis (ON) may be defined as the death of the cell components of bone - both osteocytes and bone marrow cells. ONFH is not a specific entity, but the final common pathway of various conditions that impair the blood supply to the femoral head - hence the frequently used term avascular necrosis.
The chief causes of nontraumatic ONFH are treatment with corticosteroids, sickle cell disease (SCD), and chronic alcohol abuse.
In patients on long-term steroids, there is a notoriously high incidence of avascular necrosis. Even patients who are given high-dose steroids for short periods of time (e.g. for the management of cerebral oedema) are at increased risk of ON. Also, since the introduction, some years ago, of steroid therapy after the majority of organ transplants, ONFH has frequently been observed in kidney transplant recipients. Of course, the mechanism underlying the ON in patients treated with steroids is complex: it involves changes in the walls of capillaries within the bone, and an accumulation of fat in the reticular cells and the fat cells of the bone marrow.
In SCD, the main change in the blood which gives rise to ON is the diminished deformability of the red blood cells (RBCs) as the haemoglobin S (HbS) molecules polymerize into rigid aggregates. When the oxygen saturation of the haemoglobin falls, HbS polymers will be formed, and the originally disk-shaped RBCs will become less deformable. These stiff cells will obliterate the microcirculation. This mechanism of altered RBC deformability accounts for the epiphyseal ON in SCD; in other words, the ON observed is the manifestation of rheological disturbances.
Initially, all nontraumatic forms of ON were considered to be idiopathic; nowadays, many different causes are recognized: caisson disease, Gaucher’s disease, radiation, SLE, collagen disease, pregnancy, blood disorders, chemotherapy, organ transplants, pancreatitis, etc. In addition to these established causes, there are other risk factors that have not been universally accepted. Thus, chronic alcohol abuse is so common that some wonder whether it is a predisposing factor for ON. However, a daily intake of 150 mL of ethanol (the equivalent of 1.5 L of red wine with an alcohol content of 10%) appears to be the exposure threshold for alcohol-associated ON. Obviously, the amount of alcohol should be seen in relation to the subject’s body weight. Lipid disorders, which often occur in a context of occlusive vascular disease, diabetes, or atheromatosis, have frequently been suggested as causative factors or co-factors.
While, as a rule, the mechanisms leading to ON have two processes in common (bone ischaemia, and bone marrow disturbance), there are still several unanswered questions.
What is the mechanism of cell death in ONFH? Which cell population dies first - the osteocytes or the bone marrow cells?
There is no conclusive evidence that the death of the osteocytes or the bone marrow cells is due to ischaemia caused by alterations of the vessel walls, embolism, or thrombosis. Of course, the high rate of ON in some disorders (such as caisson disease), and the physiological and biochemical disturbances involved, point to a causal relationship between the disease and the ON, and suggest that local ischaemia is, at least, an important factor in this process of cell death. However, even in these cases, the actual mechanism responsible for the presumed reduction in blood flow is not known with any degree of certainty. There are no observations that would support the view that (fat or other) embolisms are the main and causative abnormality.
The bone cells may be affected by a metabolic disorder, and their nutrition may be compromised by a purely local reduction in blood supply to a level that is not, in itself, incompatible with cell survival. Under such circumstances, any additional adverse factors may be fatal to the bone cells. These factors may be cytotoxic agents such as ethanol; or substances such as cortisone, regardless of whether the hypercortisonism is exogenous or endogenous. The agents concerned may directly affect the cells or their precursors; equally, they may act through capillary endothelial lesions, to produce vascular insufficiency.
The roles played by these factors (and many others, which may also cause cell death) probably vary with the different disorders (alcohol abuse, diabetes, gout, kidney transplants, etc.), and in different patients.
What is the role of vascular pathology in the process of osteonecrosis?
The question is whether the first disturbances are localized in a wedge-shaped arterial micro-territory, or whether, even at this early stage, they are distributed throughout a larger part of the femoral head. If the underlying disorder is a vascular one, the question is whether the disorder is intraosseous, extraosseous, or both; whether it is in the arteries, in the capillaries, in the endothelium, or in the veins. The most recent investigations suggest that the mechanism is one of distal emboli which retrogradely affect the arterial circulation, or that the underlying disorder is on the venous side.
Is there an underlying abnormality?
Regardless of the underlying cause, most forms of ON (even those that are apparently idiopathic), have a number of pathogenetic features in common. Disorders of lipid metabolism are frequently seen. This would suggest that avascular necrosis stems, first and foremost, from an increase in the volume of the fat cells in the bone marrow. Since the marrow is confined in the inextensible space of the femoral head, this lipid overload will produce increased intraosseous pressure, which will lead to a bone compartment syndrome, and, eventually, to the death of the bone cells.
(SIGMA) The quality of the bone tissue: ON tends to be more severe, and to progress more rapidly, in the combined osteoporosis and osteomalacia that is seen in patients treated with steroids; in the osteoporosis associated with chronic alcohol abuse; in the renal osteodystrophy of kidney transplant recipients; and in other bone tissue pathologies.
What is the role of mechanical factors?
(SIGMA) Independently of the initiating biological factor that causes the necrotic phenomenon in a single event, there are mechanical factors involved in the production of ONFH. This is obvious from the fact that the necrotic zone is situated around the upper pole of the femoral head, and that the collapse of the head always occurs in the main weight-bearing region. The importance of mechanical factors is highlighted by a comparison of the strength of a trabecula of dead bone, and that of a living trabecula. Necrotic bone has a modulus of elasticity that is ca. 70% less than that of normal living cancellous bone; and the ultimate breaking load of necrotic bone is about half that of viable bone. Trabecular fractures have been described even in healthy femoral heads; their numbers are increased in porotic heads. If the bone is necrotic, there should be even more fractured trabeculae; worse, these fractures in dead bone cannot heal. Thus, the fracture burden will gradually increase, and in the maximally stressed zones of the femoral head, a subchondral fracture will occur. This fracture will show up on radiographs as a curved line, the so-called crescent sign. The influence of mechanical factors is also shown by the fact that the way in which the necrotic segment becomes demarcated matches the isobar pattern in a sphere that has been compressed at one site. The zone around the necrotic segment will be exposed to abnormally high loads, which will further worsen the pressure pattern and the ischaemic phenomena, causing yet more microfractures, and, thus, contributing to the spread of the necrotic lesions.
The secondary cellular and tissular response pattern is marked by the apposition of new viable bone on the surface of the dead trabeculae. This creeping substitution may be such as to fill in all the spaces in the cancellous bone, to produce what is, to all intents and purposes, cortical bone tissue.
The extent and magnitude of this response will vary with the extent of the necrosis and with the zone affected by ON.
What happens at the junction between viable and dead bone
The repair of necrotic cancellous bone consists in two different processes, which appear to occur independently of each other: cell proliferation and invasion of the femoral head by reparative tissue; and the differentiation of mesenchymal cells into osteoblasts, which lay down new bone on the surface of the dead trabeculae. Later on, osteoclasts appear. These cells are derived either from mesenchymal precursors or from blood monocytes.
In the initial phase, there is intensive proliferation, and the osteoblasts rapidly lay down large amounts of new bone. However, after this reparative front has progressed several millimetres, the process soon comes to a halt. Osteoblast formation ceases or slows down considerably; as a result, the reparative front is made up of fibrous tissue and clusters of capillaries and mesenchymal cells.
What happens at the subchondral end plate
What happens in the repair of the cortical end plate below the cartilage differs markedly from the reparative processes in the cancellous bone of the femoral head. At the subchondral level, bone resorption far exceeds new bone formation, and the net result is loss of subchondral bone. The reparative process does not stop at the subchondral end plate: it continues into the articular cartilage, which may become ossified. This invasion by reparative tissue, and the local reactions set up by this invasion, pave the way for the osteoarthritic changes that will ultimately occur. The changes in the femoral head as an organ are not directly brought about by cell death: a dead femoral head may function for years without any apparent structural compromise. What causes a change in the mechanical properties of the femoral head as an organ, to make it deform and collapse, is the action of living cells involved in the reparative process. The starting point of the fracture is the zone of least resistance created on the lateral side of the femoral head, by the resorption of the subchondral bone and its overlying cartilage; this resorption is the result of the reparative process. In idiopathic ON, the fracture extends underneath the cartilage into the necrotic bone, as a result of stresses at the interface between the cancellous bone and the dense subchondral end plate. This process will lead to subchondral separation, producing the typical crescent sign seen on radiographs.
What happens in the articular cartilage
The articular cartilage remains viable for a long time (Fig. 1); it will go on functioning normally, despite the necrosis and subsequent reparative processes. The cartilage cells are nourished from the synovial fluid, which allows them to survive. In fact, later on the surviving cartilage cells appear to produce collagen faster than do healthy adult cartilage cells. Histochemical analysis has shown normal collagen and glycosaminoglycan levels early on; much later, with the onset of arthritic changes, there is first a relative, and eventually an absolute, loss of proteoglycans.
 |
| Figure 1 Gross photograph of a femoral head, with normal-looking cartilage over the necrotic segment |
Development of OA
OA of the hip is the result of excessive pressure acting on the healthy parts of the femoral head; of the incongruency between the deformed head and the acetabulum; of the invasion of cartilage by reparative capillaries; and, later on, by pannus formation, which completes the destruction of the cartilage.
ONFH most commonly involves the anterosuperior portion of the femoral head. An MRI scan of a normal femoral head will show a uniformly high signal intensity on T1- and T2-weighting, throughout the femoral head. In the overwhelming majority of cases, the basic pattern of ON consists in a zone of decreased signal intensity on T1- and T2-weighted images. This low-intensity zone is at the site shown as a necrotic region on plain radiographs, in typical osteonecrosis with femoral head collapse. The low-intensity zone may be homogeneous or heterogeneous, with a speckling of high intensity against the low-intensity background, both on T1- and on T2-weighted images.
Early on, the most typical image is a thin low-intensity band on T1- and T2-weighting, which goes to the subchondral bone and is more or less concave towards the top of the femoral head.
With gadolinium, the contrast of the marrow spaces is enhanced, and the sensitivity of the technique for the detection of ON is improved. Gadolinium contrast enhancement may also be useful in screening for femoral head perfusion problems after hip fractures.
A variety of imaging protocols may be used for the visualization of the different features. MRI provides excellent soft tissue contrast, and allows images to be produced in virtually any plane. The standard protocol usually starts with a T1-weighted axial localizer (short TR, short TE). So-called T2-weighted spin-echo sequences enhance the specificity of the technique. In certain cases, these images may be replaced by fast T2 gradient-echo sequences, which save imaging time. So-called STIR (short T1 inversion recovery) sequences suppress the marrow fat signal. These sequences give a strong signal of tissues with long T1 and T2 relaxation times (granulation tissue and joint fluid), which improves the contrast between the bone marrow and the abnormal tissues. Also, chemical shift MR imaging to produce specific water or fat images may be used, to give better characterization of the different constituents.
On the T1-weighted scans, MRI shows a curved low-intensity band with its concave side towards the top of the femoral head; this band defines an upper polar sector of varying size; the chief feature is the lack of homogeneity of the pattern in that sector.
The examination of surgical specimens, which is made easier by the use of surface coils, shows a marked decrease in signal intensity in the area of the subchondral end plate and in the zone between the dead and the viable bone that demarcates the necrotic segment. The in-between zone of the actual segment is heterogeneous, both in the surgical specimen and when imaged in vivo. Microradiographs of the bone sections show the well-known three zones of necrosis: the cartilage, with the subchondral end plate still attached to it, is separated from the necrotic segment; the necrotic segment consists of a delicate and regular openwork of bone, whose uniform appearance is at odds with the heterogeneous MRI signal pattern. The necrotic segment is separated from the living bone by a remodelling zone, which perfectly matches the low-intensity band seen on MRI scans. Since the changes in MRI signal intensity are fat-related, the histological workup also shows patterns related to the presence or absence of fat cells in the zones studied. In the zone of fibrosis and osteolysis that demarcates the necrotic segment (Fig. 2),
 |
| Figure 2 Histological pattern in ONFH. The peripheral portion of the necrotic segment is bordered below by a fibrovascular zone devoid of fat cells, which does not produce a signal on MRI. |
the tissue newly formed by bone remodelling is chiefly fibrovascular, without any fat cells. This accounts for the peripheral low-intensity band seen in the femoral head. Within the necrotic segment, the mottled MRI pattern correlates well with what is shown by histology: in some areas, the fat cells are “mummified”, with intact cell walls. These zones give a normal-intensity signal on MR imaging, because the triglycerides are still intact inside the cells, and have not been broken down. By contrast, in areas where the fat cells have been destroyed, there will be decreased signal intensity. This is accounted for by the fact that once the fat cell wall has been disrupted, the triglycerides are released. In the environment containing sodium salts from the joint fluid (transudate), the triglycerides undergo saponification (breakdown), and produce a weak MR signal. Similarly, in the subchondral zone, with its trabecular fractures and resorptive processes, histology will show rupture of the fat cells, which will, once again, produce a low-intensity signal. Thus, a “histologically dead zone” may produce different MR signals, with the necrotic segment retaining a normal bone pattern for several months. This is why the subject of investigations at an early stage of ON, and the staging of ON, is somewhat more complex than might be assumed: in theory, ischaemic necrosis follows cell anoxia; the fat cells can survive this anoxia for between two and five days. Even after that time, it is not certain that MRI will be diagnostic at this very early stage, since the fat cells will remain “mummified”, with the triglycerides still inside the walls of the cells. A low signal intensity on the MRI scans is not due to cell death as such, but to the release and the breakdown of the triglycerides. However, this process can probably occur in the subchondral zone only if there has been trabecular fracture and resorption. Thus, the question is whether, early in the course of the disease, a weak MR signal in the subchondral zone is not, in fact, evidence of subchondral separation (Fig. 3), which, at that stage, would otherwise be revealed only by histology, since the lesion would not yet show up on radiographs. Subchondral separation is the radiological cut-off between Stages II and III in the Arlet & Ficat classification of ON; histologically, this separation may be present at an earlier stage.
 |
| Figure 3 Histological pattern in ONFH, with clearly visible subchondral fracture |
Pathohistologically, bone marrow necrosis is the feature most readily detected. The only sign of osteocyte necrosis is the disappearance of these cells, leaving empty and often widened osteocyte lacunae. The actual pattern seen will depend upon the stage of the necrosis. At an advanced stage (in a femoral head that has collapsed and is being examined after its removal for THR), the conical sequestrum is made up of dead trabeculae, with empty osteocyte lacunae. The bone marrow is usually a magma, with loss of haematopoietic cells, and fat cells with absent nuclei and ruptured cell walls. Sometimes, the marrow is replaced by eosinophil debris, without any remaining identifiable cell structures. This necrotic zone, of varying size, is situated under the subchondral end plate. In the concave reparative zone, there is fibrovascular proliferation; dead trabeculae are completely or partially resorbed by osteoclasts, and replaced or covered with viable appositional bone, by osteoblasts.
The prognosis of ONFH is governed by the collapse of the femoral head as a result of subchondral fracture; this lesion first produces a radiolucent crescent, and subsequently a loss of sphericity of the femoral head. The treatment to be given will often be a function of this loss of sphericity and the subchondral separation. MRI is an excellent modality that correlates well with the histological findings; it may also be used to establish the extent of the ON; however, it is not very suitable when it comes to detecting the collapse. The best imaging technique to provide evidence of subchondral separation or loss of sphericity is still conventional radiography.
The earliest radiographic sign is an increased density inside the femoral head; however, the main contribution of conventional radiography is the demonstration of subchondral separation. In fact, it is the most reliable modality for the detection of the crescent sign and the loss of femoral head sphericity. The crescent sign should be looked for on lateral views, which open up the hip joint and, consequently, show the crescent. Loss of sphericity is also seen earlier on the lateral films, because of the anterosuperior site of the lesion. The most reliable lateral incidence to show loss of sphericity is the frog lateral view.
These advantages of conventional radiography notwithstanding, MRI can, of course, furnish information to help establish the age of the osteonecrotic lesion. If the necrotic zone is still of high intensity on the MR images (Fig. 4), the patient will usually have an asymptomatic hip or one that has been causing pain for less than six months. If, on the other hand, the zone is of low signal intensity (Fig. 5), then three quarters of the patients will have hip pain, and the pain history will be longer than six months. Thus, low-intensity ON is, as a rule, of longer standing than is ON that is still giving a high-intensity signal on MRI scans.
 |
| Figure 4 MRI pattern of a small incipient ONFH lesion. A.p. and lateral views. The zone inside the necrotic segment is still of high intensity. |
 |
| Figure 5 Long-standing necrosis. The necrotic zone is of low signal intensity. |
Joint effusion shows up well on MR images, giving a high-intensity signal on T2-weighting; a large effusion may be taken as indirect evidence of ON with subchondral collapse: prior to collapse, there may be a small effusion; however, in general, massive effusion occurs only in cases of subchondral collapse.
Many surgical treatments have been devised in ONFH. Of course, the large number of treatment modalities shows that the condition is difficult to treat, and that there are limits to what can be done.
The most important aspect of the management of ONFH is an early diagnosis of the condition.
Pre-collapse
Customarily, ONFH patients have been treated nonsurgically. However, purely medical treatment will allow only few patients to live with their condition for one or two years. Usually, the patients are kept off weight-bearing on the affected side; however, this treatment principle has several limitations: it does not abolish the muscle tone around the joint, even when the patient is lying down. The use of two elbow crutches (as a substitute for axillary crutches) does not provide the required complete weight relief; and even if crutches afford some protection, there is still the question of how long the patients should be kept off weight-bearing. Between the diagnosis of ONFH and the loss of femoral head sphericity, up to 4 or 5 years may elapse. Strictly speaking, the patients should be kept off weight-bearing for this entire period of time. Seeing that the patients are usually young and in the midst of their working lives, it would be futile to attempt such prolonged weight relief. It should also be borne in mind that if patients are kept on nonsurgical treatment for a long period of time, eventual surgery may come too late, since the lesions will, by then, have progressed to a stage where femoral head preserving surgery can no longer help.
Very early in the course of the disease, core decompression remains the most logical treatment modality, if one accepts that the condition is a compartment syndrome, with increased pressure inside the femoral head. Core decompression has been in use for a considerable time; the results reported by different authors vary, no doubt because of different patient populations and different ON stages treated with decompression. However, while fundamental research and clinical studies have shown that dead bone may be revascularized by living bone, the reparative osteogenic potential is slight in ONFH: the number of bone progenitor cells in the uninvolved part of the femoral head and in the trochanteric region is less than in healthy subjects. It would, therefore, make sense not only to core, but to introduce new cells. This can be done by placing a tibial graft, a vascularized graft or cancellous bone into the coring tract. The same result may be produced more readily by harvesting bone marrow from the anterior iliac crests, concentrating the marrow thus obtained, and reinjecting it into the necrotic zones.
Post-collapse
Once the femoral head has lost its sphericity, core decompression will still afford pain relief, but will not be able efficaciously and lastingly to halt the gradual collapse of the weight-bearing zone. The important goal to achieve at this stage is the immobilization of the necrotic segment. A mobile segment will produce increased pressure in the femoral head, and may contribute to the progression of the necrosis. To this end, attempts have been made to use early reconstruction, with debridement of the necrotic zone and replacement of the dead bone with autologous bone reinforced with a vascularized fibular graft, to support the subchondral bone at risk of collapse. This is an attractive approach, which does, however, require the patient to be kept off weight-bearing for a long period of time (3 - 6 months). The surgery involved is technically demanding; the postoperative management is quite cumbersome (3 - 6 months off weight-bearing); often, both hips are affected; and the treatment may not work. This is why only few patients have been managed in this way.
Since the survival of the cartilage cells is ensured by their nutrition from the synovial fluid, the femoral head may be restored to sphericity by elevating the necrotic segment and keeping it in this restored position, by the injection of cement. In this approach, acrylic cement is used to give the femoral head its correct shape, and to delay the progression to OA which follows the deformation of the head. Cement injection has several advantages: it allows immediate weight-bearing; it provides immediate pain relief; and it does not interfere with the patient’s work for an unduly long period of time. Also, conversion to a hip replacement is straightforward, should the pain recur. In such cases, femoral head resurfacing would undoubtedly be a suitable treatment option.
For femoral head preserving surgery, osteotomies of various kinds (valgus, varus, flexion, deflexion, rotational) have been used for a long time. These osteotomies are still difficult to perform. Also, they may require prolonged weight relief after surgery, and conversion to hip replacement in case of failure may be difficult.
Following the onset of arthritic changes
Once OA has developed, or the femoral head has suffered major collapse, joint replacement is the only viable option (Fig. 6). Surface replacement arthroplasties have given good results in some cases, but have come to be less frequently used, because of the rapid deterioration seen in some cases, which contrasts with the good results habitually seen following THR.
 |
|
| Figure 6 Gross and radiographic appearance of ONFH, showing the subchondral fracture | |
Very early in the course of the disease, core decompression remains the most logical treatment modality, if one accepts that the condition is a compartment syndrome, with increased pressure inside the femoral head. Core decompression has been in use for a considerable time; the results reported by different authors vary, no doubt because of different patient populations and different ON stages treated with decompression. However, while fundamental research and clinical studies have shown that dead bone may be revascularized by living bone, the reparative osteogenic potential is slight in ONFH: the number of bone progenitor cells in the uninvolved part of the femoral head and in the trochanteric region is less than in healthy subjects. It would, therefore, make sense not only to core, but to introduce new cells.
The treatment of ONFH with bone marrow autografts is based upon the view, now commonly held, that the osteogenic cells derive from a stem cell in the bone marrow stroma. It is thought that this stem cell gives rise to osteoblasts, chondroblasts, fibroblasts, etc.; and that the actual cell type obtained from that one precurseur will depend upon the culturing conditions used. The clonogenic potential of the different cells has been demonstrated; each cell colony is derived from a single stem cell is known as a CFU-F (fibroblast colony-forming unit).
When red bone marrow is transplanted, the graft will contain osteogenic precursors, which will repopulate the osteonecrotic bone. In the early stages of the disease, the femoral head will still be round (Fig. 7). By definition, the necrotic zone will be acellular, at least as far as osteocytes and bone marrow cells are concerned. However, before Stage III, the bone framework is still intact; in particular, it will have retained its strength, even though the cell population in the upper end of the femur is abnormally small. This is why it was thought that conventional core decompression should be supplemented by an autograft of cells harvested by bone marrow aspiration from the ipsilateral iliac crest. We have used this approach in the treatment of Stage I and Stage II ONFH.
 |
| Figure 7 Gross appearance of a femoral head with pre-Stage III ONFH. The affected zone is abnormally “white”, and bordered by a small red revascularization zone. |
 |
 |
| Figure 8 Technique of bone marrow harvesting and reinjection, in the treatment of ONFH by bone marrow autograft | |
The bone marrow is harvested under general anaesthesia (Fig. 8). The usual sites are the anterior iliac crests; the posterior crests are less frequently used. A bevelled metal trocar of 6 - 8 cm length and a bore of 1.5 mm is pushed deep into the cancellous bone. A 10 mL syringe that has been flushed with heparin is used to aspirate the marrow. Once the needle has been inserted to the desired depth, the tip is swept around a full circle in 45° steps, with the bevel pointing in different directions at each step. Bone marrow is withdrawn at each of these points. Once this 360° aspiration has been performed at one site, the needle is brought out and reinserted at a different site, where the 360° sweep in 45° steps is repeated. This procedure is continued until a sufficient quantity of bone marrow has been harvested. The cell content of the marrow thus obtained will be greater if the marrow has been aspirated in small (2 mL) fractions, since, under these conditions, the proportion of contaminating peripheral blood will be less. The same percutaneous tract may be used for multiple punctures of the iliac crest. All the marrow aspirated is discharged into a plastic collection bag containing ACD (acid citrate dextrose) anticoagulant solution. It is then filtered, to remove fat aggregates and clots.
Bone marrow harvesting is most conveniently done by two operators, each working on one iliac crest. An assistant places the aspirated material in the collection bag and flushes the syringes with heparin.
The aspirated material needs to be reduced in volume in order to increase its stem cell content. This is done by removing some of the RBCs (the non-nucleated cells) and the plasma, in such a way as to retain only the nucleated cells, i.e. the mononuclear stem cells as well as the monocytes, the lymphocytes, and some PMNs. This involves a considerable amount of bone marrow handling. Although the technique as such is suitable for large volumes of marrow, great care must be taken to ensure that everything is done with full sterile precautions, so as to obtain a product that is safe for reinjection. Two techniques are available: bone marrow may be harvested, concentrated, and reinjected under the same anaesthetic, in which case concentration must be performed sufficiently speedily so as not to exceed a time of ca. 30 minutes; or the harvested material may be frozen and reinjected at a later stage; in that case, there is no constraint on the time taken to concentrate the marrow.
For reinjection under the same anaesthetic, a COBE 2991 blood cell washer was used. With this technique, the bone marrow is centrifuged for 5 minutes at 400 g (g = gravity). Leucapheresis is performed during 40 to 50 seconds at a collection rate of 100 mL/min. This centrifuging technique yields a “concentrated myeloid suspension” of ca. 50 mL stem cells, from a 300 mL volume of aspirated bone marrow; the stem cell concentrate is placed in a syringe for reinjection.
The bone marrow is injected into the femoral head using a small (Mazabraud) trocar. The instrument is introduced through the greater trochanter, as in conventional core decompression. Its position in the femoral head and in the necrotic segment is monitored with biplane fluoroscopy. Since, at the time of treatment, the plain radiographs will show little if any evidence of necrosis, the preoperative MRI scans should be used together with the image intensifier views, to determine the site of the lesion. After the injection of the bone marrow, a few millilitres of contrast should be injected, in order to check the area in the femoral head through which the injected bone marrow will spread. It has been established that the contrast medium will not damage the bone progenitor cells.
Although the bore of the trocar is small compared with the trephines normally used for core decompression, femoral head and trochanteric region pressure measurements have shown that even a small hole will relieve the intraosseous pressure. If, during the bone marrow injection, the pressure in the femoral head is found to rise, a normal pressure pattern will be restored once the injection is finished, exactly as in intraosseous pressure measurements. In our patients, no complications were observed during anaesthesia; in particular, there was no reduction in oxygen saturation, and no change in the pulse rate or blood pressure.
There are still several theoretical and practical questions that remain unresolved. Thus, it is not known how many cells need to be injected, or what the optimum concentration is. The technique’s mechanism of action is, undoubtedly, complex. During the injection, the femoral head is flushed, and some of the fat is removed, as the 25 mL volume of bone marrow is injected. Also, the bone marrow contains not only stem cells but also bone morphogenetic proteins such as BMP-2, which are introduced into the femoral head and into the necrotic segment. The procedure, which may be used in other conditions as well, holds great promise because of its inherent advantages: it is straightforward; it involves autologous transplantation; and it is well tolerated, since both the marrow harvesting and the injection into the femoral head are done percutaneously. Obviously, its usefulness has to be weighed against that of other treatment modalities. The decision will be primarily a function of the stage of ON: once the femoral head has lost its sphericity, bone marrow autografts can no longer be used.
In ONFH, subchondral separation and the loss of sphericity of the femoral head constitute a point of no return in the course of the disease. It is generally accepted that once this stage has been reached, the condition will inexorably progress to OA, in the shorter or the longer term. This is why attempts have been made to restore the sphericity of the femoral head by elevating the necrotic segment and keeping it in its correct position, by the injection of cement. In this approach, the acrylic cement is used to restore the rounded femoral head pattern, relying on the fact that the cartilage cells will survive because the articular cartilage is nourished by the synovial fluid.
The technique has been used in the treatment of osteonecrotic femoral heads that had lost their sphericity but had not yet progressed to arthritic changes.
The surgical technique involves exposure of the hip by a Smith-Petersen approach and a T-shaped capsular incision to expose the anterosuperior aspect of the femoral head as well as the anterior part of the neck. The incision of the capsule must be limited inferiorly, so as not to damage the lateral circumflex femoral artery.
The necrotic segment is usually obvious: while the cartilage over the lesion is grossly indistinguishable from the cartilage over the healthy part of the femoral head, the necrotic zone is commonly surrounded by a groove or low ridge. Pressure with a dissector may also be used to detect the necrotic zone, which will spring back after having been depressed, rather like an indented table tennis ball (Fig. 9). This phenomenon is no doubt due to the elasticity of the overlying cartilage and the mobility of the necrotic segment, which is detached from the living bone and from the subchondral end plate.
 |
|
| Figure 9 Diagrammatic representation of loss of femoral head sphericity in ONFH. The cartilage is still grossly normal, but the head may be depressed like a table tennis ball. |
Elevating the articular cartilage and restoring the sphericity of the femoral head can be readily achieved by the use of a pin inserted between the viable bone and the necrotic segment. Levering on the pin will, as a rule, allow the cartilage to be raised sufficiently to restore the rounded pattern of the femoral head; this restoration of the contour can be appreciated by direct vision. If need be, two or three pins may be used; these pins may be left in situ to steady the segment during the cement injection.
Low-viscosity cement is injected using a cement gun (Fig. 10); the cement is placed under the detached segment and, if necessary, in the subchondral zone. The object is to finish up with a spherical head and a necrotic segment that will neither move nor allow itself to be depressed (Fig. 11). After surgery, weight-bearing is allowed from Day 3, and a standard programme of hip physiotherapy is initiated.
 |
|
|
| Figure 10 Cement injection technique: The collapsed segment is lifted, and cement is injected into the subchondral zone and below the segment.
Reproduced with permission of the Journal of Bone and Joint Surgery (1993, 75B, 875-880) |
 |
|
|
| Figure 11 Intraoperative view. Starting at the top: The necrotic segment is readily identified, and may be depressed with a dissector. - The segment is elevated with a pin; through this tract, cement is subsequently injected. - Postoperative view.
Reproduced with permission of the Journal of Bone and Joint Surgery (1993, 75B, 875-880) |
This technique shows that it is possible, early in Stage III ONFH, to restore the sphericity of the femoral head (Fig. 12), and that the immobilization of the necrotic segment affords immediate pain relief. The technique is beneficial because of the straightforward postoperative management, and the fact that the patient is allowed to weight-bear immediately. Also (unlike certain rotational osteotomies), it does not jeopardize later conversion to THR. However, despite all these benefits, the technique is not a miracle treatment of ON that preserves the femoral head. On the other hand, considering the high rate of osteotomy failure necessitating conversion to THR, cement injection appears to be justified. It should probably be used only in cases where the ON is not very far advanced and not progressing very rapidly. In our study, the best results were obtained in adult patients with idiopathic ONFH, and in SCD cases. While the procedure cannot alter the natural history of ON, it probably allows the patient a more comfortable and painfree life until, eventually, he or she will have to undergo THR, between five and ten years after the collapse of the femoral head.
 |
|
| Figure 12 Femoral head with subchondral separation, before and after cement injection | |
|
AUTHOR’S PUBLICATIONS
Ph. HERNIGOU, F. GALACTEROS, D. GOUTALLIER : Nécrose de hanche drépanocytaire (fréquence, aspect morphologique, évolutif et thérapeutique). Revue de Chirurgie Orthopédique 1989, Suppl. 1, 75, 120-121. Ph. HERNIGOU, M.C. VOISIN, E. DESPRES, D. GOUTALLIER : Confrontation de l’imagerie par résonnance magnétique nucléaire et de l’histologie dans les nécroses des têtes fémorales. Revue du Rhumatisme 1989, 56, (11), 741 -744. Ph. HERNIGOU, F. GALACTEROS, D. BACHIR, D. GOUTALLIER : Etude de 164 nécroses épiphysaires (hanches, épaules, genoux) chez 55 patients drépanocytaires, caractéristiques, aspect épidémiologique et étio-pathogénique. Revue du Rhumatisme 1989, 56 (12), 869-875. Ph. HERNIGOU, F. GALACTEROS, D. BACHIR, D. GOUTALLIER : Histoire naturelle de la nécrose de hanche dans le malade drépanocytaire. A propos de 104 nécroses. Revue de Chirurgie Orthopédique 1989, 75, 542-557. Ph. HERNIGOU, D. GOUTALLIER : Reconstruction in avascular necrosis of the spheric geometrical shape of the femoral head with acrylic cement. Bone Circulation and Bone necrosis, Ed. J. Arlet, B. Mazieres, Springer Verlag 1990, 353-355. Ph. HERNIGOU, F. GALACTEROS, D. BACHIR, D. GOUTALLIER : Deformities of the hip in adults who have Sickle-Cell disease and had avascular necrosis in childhood. : Journal of Bone and Joint Surgery 1991, 73 A, 1, 81-92. Ph. HERNIGOU, F. GALACTEROS, D. BACHIR, D. GOUTALLIER : Séquelles des nécroses de hanche de l’enfant drépanocytaire. Revue du Rhumatisme 1991, 58 (10), 643. Ph. HERNIGOU, B. COPIN, K. EZZAOUIA, D. GOUTALLIER : Dissection sous-chondrale dans les nécroses de hanche de l’adulte (confrontation entre l’aspect per-opératoire, les radiographies, le scanner, la résonnance magnétique nucléaire et l’histologie). Revue du Rhumatisme 1991, 58 (10), 695. Ph. HERNIGOU, D. GOUTALLIER : Reconstruction de la sphéricité de la tête fémorale des nécroses par relèvement du séquestre et comblement par du ciment. Revue du Rhumatisme 1991, 58 (10), 696. Ph. HERNIGOU, D. GOUTALLIER : Reconstruction de la sphéricité de la tête fémorale des nécroses par relèvement du séquestre et comblement par du ciment. Revue de Chirurgie Orthopédique 1992, Suppl. I, 213. Ph. HERNIGOU, D. BACHIR, F. GALACTEROS : Avascular necrosis of the femoral head in sickle-cell disease. J. Bone and Joint Surg. (BR) 1993, 75 B, 875-880. Ph. HERNIGOU, F. BEAUJEAN : Bone marrow activity in the upper femoral extremity in avascular osteonecrosis. Rhum. (Eng. Ed.) 1993, 60 (1), 610. Ph. HERNIGOU : La nécrose de hanche. Revue Synoviale 1993, 17, 11-16. Ph. HERNIGOU, F. BEAUJEAN : La moëlle osseuse, une clé dans la compréhension des nécroses de hanche idiopathiques. Revue du Rhumatisme et des Maladies Ostéoarticulaires 1993, 60, n° 10, 722. Ph. HERNIGOU : Nécrose de hanche et cicatrice du cartilage de croissance. Revue du Rhumatisme 1994, 61 (n° 10), 704. Ph. HERNIGOU, A. de LADOUCETTE : Ostéonécrose aseptique de la tête fémorale. Thérapeutique Rhumatologique Médecine-Sciences, Edit. Flammarion 1995, 357-359. Ph. HERNIGOU : Ostéonécrose des épiphyses de l’adulte. Edition technique, Encyclopédie Médico-chirurgicale Appareil Locomoteur 1995, 14-028-A-10 Ph. HERNIGOU : Traitement des nécroses de hanche (aux stades I et II) par autogreffe de moelle osseuse. Revue du Rhumatisme 1995, n° 10, 194. M.C. VOISIN, I. BROCHURIOU, M.H. SY, Ph. HERNIGOU : Aspect anatomo-pathologique de la tête fémorale dans les nécroses drépanocytaires au stade III. Revue du Rhumatisme 1995, 62 (n° 10), 742. Ph. HERNIGOU : Autologous bone marrow grafting of avascular osteonecrosis before collapse. Rev. Rhum. (Engl. Ed.) 1995, 62, 10, 650. Ph. HERNIGOU, F. BEAUJEAN : Abnormalities in bone marrow of the iliac crest in patients with osteonecrosis. Journal of Bone and Joint Surgery, 1997, 79A, 7, 1047-1053. Ph. HERNIGOU, F. BEAUJEAN : Autologues bone morrow grafting of avascular necrosis before collapsus. Journal of Bone and Joint Surgery, 1997, 79B, Supp. II, 171. |