The relationships between bone marrow and osteonecrosis of the proximal femur are complex. Changes in the bone marrow signal 16 are observed on magnetic resonance imaging (MRI) scans of patients with osteonecrosis; the fatty marrow conversion 15 is increased in the intertrochanteric portion of osteonecrotic hips; a decrease of osteogenic stem cells 4,12 is present in the bone marrow of some patients with osteonecrosis.
Bone marrow may be considered as a system with two parts: one, the hematologic part and two, the stroma system containing the mesenchymal stem cell pool that provides osteoblasts. Despite this osteogenic characteristic, the clinical use of marrow as an osteogenic graft has remained limited to the treatment of nonunion. 3 Bone grafts have been used in the treatment of avascular osteonecrosis for many years, not only to provide structural support to the subchondral bone, but also to enhance bone formation in osteonecrosis. 6,20
After experiencing poor results with the technique of core decompression alone described by Ficat et al, 5 the current authors began to do a modified type of core decompression with autologous bone marrow grafting. To date, 330 patients (534 hips) have been operated on with this technique. 10,12 By December 1996, 116 patients (189 hips) had been operated on with this technique. These patients are the basis of the current study because they have had a followup of more than 5 years.
MATERIALS AND METHODS
Patients and Osteonecroses
Between 1990 and December 1996, 116 patients (189 hips) with avascular osteonecrosis at early stages were treated with decompression and autologous bone marrow grafting. No patient was treated by core decompression alone; other conservative techniques (osteotomies or total hip replacement) were used for patients with osteonecrosis at later stages. From 1990 to 1996, 658 hips with osteonecrosis were operated on; therefore, the 189 hips in the current series represent only 29% of the hips with osteonecrosis treated at the current authors’ institution. The percentage of hips affected by osteonecrosis in this series of 189 hips was 16% in patients taking corticosteroids, 30% in patients with excessive alcohol intake, and 34% in patients with sickle cell disease (Table 1). In 73 patients, the procedure was done bilaterally under the same anesthesia. Patients were allowed weightbearing using crutches for 10 days after surgery and were full weightbearing without crutches thereafter.
The diagnosis of femoral head osteonecrosis was made using anteroposterior (AP) and lateral plain radiographs or MRI scans. 19 All MRI scans were obtained with a 1.5-T supraconducting unit. T1-weighted spin-echo images were obtained with repetition times of 400 to 600 ms and echo times of 20 to 40 ms. Long images with TR (repetition time) of 1500 to 2240 ms, and TE (echo times) of 30 to 100 ms were used to obtain images of varying T2 weighting (transverse relaxation time). Images were obtained in coronal and sagittal sections. The diagnosis of osteonecrosis on MRI scans was based on bandlike abnormal signals, on bandlike hypointense zones on T1-weighted images, and on matching hyperintense zones on short tau inversion (STIR images).
Anteroposterior and frog lateral radiographs were taken when the patient entered the study and every year thereafter. All radiographs were graded using the method of Steinberg et al. 19 Stage 1 was defined as a hip with normal radiographs but an abnormal MRI scan. Stage II was defined as a hip with an abnormal MRI scan and an abnormal radiograph in a patient presenting with sclerotic or cystic changes in the femoral head but an absence of crescent line. Stage III was defined as a femoral head with a crescent sign (evaluated as a radiolucent line on the lateral radiograph) without flattening of more than 1 mm, the normal contour being established by the use of concentric circles. Stage IV was defined as a flattening of more than 1 mm without joint narrowing. Because crescent line and collapse are considered major radiographic signs by most of the classifications but sometimes are ambiguous, the term crescent line only was used for hips with Stage III osteonecrosis and the term collapse was used for hips with Stage IV osteonecrosis. Stage V was defined as a hip with secondary changes in the acetabulum including joint-space narrowing, sclerosis, and cyst formation. Stage VI was defined as a hip with advanced degenerative changes.
The extent of the necrosis was measured on AP and lateral radiographs. The lesion size was assessed quantitatively by digitizing the area of the femoral head occupied by bone with abnormal texture on radiographs or on MRI scans for Stage I and expressing it as a percentage of the entire femoral head. Cuts 5 mm posterior to the midcoronal MRI scan were used to estimate the size of the osteonecrosis in Stage I. The degree of collapse was measured using a template overlay of circles in 1-mm diameter increments.
Patients were seen in followup every 3 months for the first year, every 6 months for the second year, and yearly thereafter. Immediately before surgery and at each followup, patients were evaluated by the use of the Harris hip score 8 and by AP and lateral radiographs. Results were determined by the change in Harris hip scores 8 from preoperative evaluation to the most recent followup, by the change in the radiographic progression determined by the classification of Steinberg et al, 19 and by the need for subsequent total hip arthroplasty. For the patients who required additional surgery, total hip replacement was the procedure of choice for almost all the patients.
Marrow was aspirated from the anterior iliac crests with the patient under general anesthesia. After deep insertion of a beveled needle 6 to 8 cm long and 1.5 mm in internal diameter into spongy bone, the marrow was aspirated into a 10-mL plastic syringe. The needle was moved toward the surface through the same insertion site and successive aspirations were begun again, always turning the needle 45° after each aspiration. The marrow was aspirated in small fractions, to reduce the degree of dilution by peripheral blood. Using the same skin opening, several perforations were made into the iliac crest. All aspirates were pooled in plastic bags containing cell culture medium and anticoagulant solution (citric acid, sodium citrate, dextrose). Pooled aspirates then were filtered to separate cellular aggregates and fat.
Bone Marrow Concentration
The aspirated material was reduced in volume to increase its stem cell content. This was done by removing some of the red blood cells (the nonnucleated cells) and the plasma, in such a way as to retain only the nucleated cells, that is the mononuclear stem cells and the monocytes, and the lymphocytes.
The marrow was concentrated in a cell separator (Cobe 2991, Gambro BCT, Paris, France). A 5-minute centrifugation at 400 g (g = gravitation) forced the polynuclear cell layer, which was heavier because of the volume of its nuclei, to the periphery, where it was collected and separated from the remainder of the marrow. The leukocyte layer was removed at a flow rate of 100 mL per minute for 40 to 50 seconds. The lighter, red cells without nuclei were found in the center and were recovered with the plasma. All that remained was the mononuclear layer containing the stem cells. This centrifugation method reduced a 150-mL bone marrow aspirate to a concentrated myeloid suspension of approximately 30 mL of stem cells; it is this suspension that was poured into a syringe for reinjection.
Core Decompression and Bone Marrow Grafting Osteonecrosis
Patients were placed on a table with an image intensifier with a C arm. The decompression was done with a percutaneous approach using a 3-mm diameter trephine (trocar of Mazabraud, Collin, France). The bone marrow was injected into the femoral head using a small (Mazabraud) trocar. The instrument was introduced through the greater trochanter, as in conventional core decompression. Its position in the femoral head and in the necrotic segment was monitored with biplane fluoroscopy. Usually only one hole was made. In four cases, two holes were made because the first hole did not seem to be exactly in the center of the osteonecrosis. Because at the time of treatment the plain radiographs showed little if any evidence of necrosis, the preoperative MRI scans and the image intensifier views were used together to determine the site of the lesion. Before the injection of the bone marrow, a few milliliters of contrast was injected to check the area in the femoral head through which the injected bone marrow spread. It was established that the contrast medium does 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 showed that even a small hole relieved the intraosseous pressure. 9 Ficat et al 5 theorized that avascular necrosis was a compartment syndrome caused by increased intramedullary pressure 21,22 and venous stasis. During bone marrow injection, the pressure in the femoral head increased, but a normal pressure pattern was restored once the injection was finished, exactly as in intraosseous pressure measurements. In the patients in the current study, no complications were observed during anesthesia; in particular, there was no reduction in oxygen saturation and no change in the pulse rate or blood pressure. The authors are confident that all the injected bone marrow probably did not remain in the osteonecrotic fragment or in the femoral head. Leakage of marrow may have occurred in some patients through the trephine site although the entry point of the trephine usually was closed by a small piece of bone. Some leakage of marrow also may have occurred through the circulation of the proximal femur. However, according to the authors’ experience with radionuclide imaging in five patients at the beginning of their experience, the greater part of the bone marrow remains in the osteonecrosis or in the femoral head.
Study of Aspirated and Reinjected Marrow
The study of aspirated and reinjected marrow was important because it allowed the quality of each operation to be controlled and in particular it allowed verification that the concentration and separation processes had not destroyed the stem cells. Furthermore, it was the only means to determine the quantity of stem cells aspirated and reinjected for each patient.
Nuclear cells in the aspirate were counted as follows: the quantity of medullary nuclear cells per kilogram of marrow was calculated using a conservative formula that takes into account a high blood dilution. In each milliliter of aspirate, it was estimated that medullary cells were represented only by the difference between the nuclear cell count in the aspirate and the count in peripheral blood (which is done during general anesthesia); the number of nuclear cells of presumed medullary origin per kilogram was expressed as follows:EQUATION
with V = total volume of aspirate in milliliters, including harvesting medium; NP = nuclear cell count per milliliters in the collection bag that leaves the operating room, including the harvesting medium; V-100 = exact volume of aspirate, after subtraction of the 100 mL of harvesting medium; NS = nuclear cell count per milliliter of peripheral blood drawn during the general anesthesia; P = patient’s weight in kilograms. As an example, for a total final volume of 300 mL containing 14 ×106 nuclear cells/mL, obtained from a 70-kg adult with a leukocyte count of 4 × 106/mL as determined under general anesthesia, it is estimated that the medullary nuclear cell count is 5 × 107 per kg, for a total of 0.35 × 1010 nuclear cells.
The number of stem cells which were aspirated and reinjected were evaluated as follows: although the nuclear cell count is essential, the information it gives regarding progenitor cells in the marrow is not complete. In fact, there is no constant ratio between average marrow cellularity and number of precursors. The best biologic criterion for quantitative and functional evaluation of stroma cells aspirated for use as an autograft therefore is the study of the precursors by the in vitro cloning of nuclear cells of the marrow. To measure bone marrow activity, the fibroblast colony-forming unit was used as an indicator of stromal cell activity. The fibroblast is not an osteogenic cell but according to the theory of pluripotential cell lines, 3,4 osteocytes develop from colony-forming units of progenitor cells in the marrow. There is little doubt that these colonies are clonal (originate from one cell); so, in vitro cultures of the fibroblast progenitor cells were obtained. Marrow was cultured in vitro before and after concentration to ensure that the concentration process does not alter the number of stem cells in the sample. After collection, heparin was added to the marrow samples. The number of nucleated cells was counted using a standard Malassez hemocytometer (Zintle, Germany) and buffy coats were collected after centrifugation of the aspirates at 1300 g for 10 minutes. Cells were washed once and resuspended in Hanks balanced salt solution without Ca++ and Mg++. For the fibroblast colony-forming units, quadruplicate aliquots of 2 × 106 cells were inoculated in 25-mL tissue culture flasks containing 10 mL of culture medium supplemented with 20% fetal calf serum, penicillin (100 U/mL) and streptomycin (100 mg/mL). Culture flasks were placed in a humidified incubator with 5% CO2 and maintained at 37° C. The growth medium was renewed completely every 3 to 4 days. Cultures were read on the tenth day. Fibroblastic colonies were stained with Giemsa and counted under an inverted microscope at 25 magnification. An aggregate of cells containing more than 50 fibroblasts was scored as a colony. Results were expressed as the mean number of fibroblast colony-forming units per 106 bone marrow cells. The fibroblastic nature of the colonies were shown by immunofluorescence staining with antibodies against fibronectin and Type I and Type III collagen. Results are expressed as the number of colony-forming units per 106 medullary cells or the number of colony-forming units per milliliter.
The mean age of the patients at the time of decompression and autologous bone marrow grafting was 31 years (range, 16–61 years). Seventy-five patients (65%) were males and 41 patients were females (35%). The disease was unilateral in 43 patients and bilateral in 73 patients. The followup for these patients was between 5 and 11 years with a mean of 7 years.
Preoperatively, hips were placed in the following stages according to the classification of Steinberg et al 19: Stage I, 59 hips (31%); Stage II, 86 hips (46%); Stage III, 12 hips (6%); and Stage IV, 32 hips (17%). No hips had surgery later than Stage IV with this technique, and only symptomatic hips had surgery in this study.
Three complications were observed in the 116 patients (189 hips) who had decompression and autologous bone marrow grafting. These complications only were observed in patients with sickle cell disease: one patient had pneumonia, one patient had pain after surgery, and one patient had secondary alloimmunization after transfusion. No complications were recorded during the anesthesia; in particular, oxygen saturation, heart rate, and blood pressure remained unaltered. No pulmonary embolism, thrombophlebitis, or intertrochanteric fracture was observed.
At the most recent followup (average, 7 years; range, 5–11 years) 34 hips (22 patients) among the 189 hips (116 patients), that is, 18% of the hips and 20% of the patients, required total hip replacement at a mean of 26 months after decompression and autologous bone marrow grafting (range, 8–76 months). Most of the patients were treated at early stages, because only 44 of the 189 hips had Stage III or Stage IV disease. Seventy-eight percent of the total hip replacements were done during the first 3 years postoperatively and 6% were done after 5 years. For the entire group, the mean preoperative Harris hip score 8 was 69 points (range, 49–92 points). Patients who required total hip replacement had 57 points before autologous bone marrow grafting and 41 points before total hip replacement, for a mean loss of 16 points. Patients who did not require total hip replacement at the latest followup had 72 points preoperatively and 89 points postoperatively, for a gain of 17 points. In these patients who did not require total hip replacement, the preoperative and the postoperative Harris hip scores were 77 and 93 points for Stage I, 74 and 90 points for Stage II, 70 and 81 points for Stage III, and 68 and 74 points for Stage IV, respectively.
The preoperative stage of the 34 hips requiring total hip replacement (Table 2) was Stage I for two hips, Stage II for seven hips, Stage III for five hips, and Stage IV for 20 hips. In the 155 hips not requiring total hip replacement (Table 3), preoperatively, the hips were Stage I (57 hips), Stage II (79 hips), Stage III (seven hips), Stage IV (12 hips), and at the most recent followup, they were Stage 0 (12 hips), Stage I (32 hips), Stage II (72 hips), Stage IV (21 hips), Stage V (11 hips), and Stage VI (seven hips). In patients who did well clinically (without total hip replacement), 59% of the hips (32 Stage I and 59 Stage II) were radiographically stable (Table 3), 33% of the hips (13 Stage I, 20 Stage II, seven Stage III, 12 Stage IV) had radiographic progression, and 12 Stage I (8% of the hips) seemed to have decreased to Stage 0 on MRI scans (no signal of osteonecrosis could be seen on MRI scans taken at the most recent followup). When the hips were in Stage III or in Stage IV preoperatively, the patients had a poor result and all the hips progressed in stage (Table 3).
There was a significantly higher number of patients who had total hip replacement when the preoperative radiograph showed Stage IV disease. Even if patients with Stage IV osteonecrosis improved after decompression and bone marrow grafting, this technique does not avoid progression to Stage V and Stage VI, after collapse, even for hips that had relatively small or intermediate amounts of femoral head flattening (Stage IV-A and Stage IV-B) preoperatively. To determine the effect of the lesion size on outcome of hips at earlier stages, the 145 hips that were in Stage I and in Stage II before autologous bone marrow grafting were evaluated together. The size of the lesion on average was 25% of the femoral head (range, 10%–45%). Using the chi square test, patients with smaller lesions (< 25%) had a significantly better outcome (absence of collapse) than patients with larger lesions (20 hips with collapse).
The etiologic factors had an influence on the risk of failure with the need for total hip replacement after the index surgery (Table 4). For example, among the 31 hips with osteonecrosis related to corticosteroid use, the 11 hips (eight patients) that had failed results and had total hip arthroplasty received an estimated number of progenitors of average 0.6 colony-forming units. At the most recent followup, the other 20 hips without arthroplasty received an estimated number of 5.4 colony-forming units. Total hip replacement was required in 12% of the 21 hips in patients with organ transplantation, in 11% of the 31 hips in patients taking corticosteroids, in 8% of the 56 hips in patients with excessive alcohol intake, in one of the 10 hips in patients with idiopathic osteonecrosis, but only in two of the 64 hips in patients with sickle cell disease.
The number of nucleated cells (Table 5) obtained from bone marrow aspiration and presumed to be from medullary origin was estimated to be 16.4 millions cells per mL (± 11.5). The number of colony-forming units obtained in the samples incubated in vitro was on average 12.4 ± 3.4 per 106 bone marrow nucleated cells. The average volume of bone marrow aspiration was 147 mL ± 12 mL. The average total number of colony-forming units obtained by bone marrow aspiration therefore was estimated to be 29,000 ± 21,000. The variation between patients was attributed to the differences in the number of nucleated cells (26%), to the difference in the volume of bone marrow aspirated (11%), and to the difference in the number of colony-forming units per million nucleated cells (63%).
There was a significant difference in the number of colony-forming units obtained from the iliac crest according to the etiologic factors of the osteonecrosis (Table 5). The number of progenitor cell counts (in terms of fibroblast colony-forming units) was significantly (p < 0.05 with the Mann-Witney U test) lower in the patients who had osteonecroses attributable to corticosteroid therapy, alcohol abuse, or organ transplantation than in the patients who had osteonecroses attributable to sickle cell disease or other causes. This may be in relation with the toxicity of steroids and of alcohol on the progenitor cells. 12
The average total number of colony-forming units obtained from bone marrow aspiration was estimated to be 29 × 103 cells after culture of samples in vitro. After concentration, cell recovery after culture in vitro was 84.3% for the colony-forming units. The average total number of colony-forming units injected by hip was estimated to be 25 × 103 cells. In each group of patients with different etiologic factors, the number of transplanted progenitor cells was different (Table 6) and had an influence on the outcome of the hips (Table 4). Hips that received a low number of transplanted cells had a more significant risk of failure at the latest followup than hips that received a high number of transplanted cells (Table 4). This may explain in part the influence of the etiologic factors on the outcome of the hips (Table 4).
Core decompression was described by Arlet and Ficat 2 and Ficat et al. 5 The aim of the technique should be to improve repair in the osteonecrosis, at least at earlier stages (Stages I and II) before mechanical failure of the femoral head has occurred (Stage IV).
Reconstruction repair has been observed after core decompression, but usually this repair is incomplete. One of the reasons for bone remodeling leading to an insufficient creeping substitution after osteonecrosis in the femoral head may be the small number of progenitor cells in the femoral head 8,13 of patients with osteonecrosis. Although fundamental research and clinical studies have shown that dead bone may be repaired by living bone, the reparative osteogenic potential is low in osteonecrosis: 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. Whether this decrease in progenitor cells is a cause or a consequence of the necrosis is not known 8,11,13 and probably will require additional studies; however, because of this lack of progenitor cells, treatment modalities should stimulate and guide bone remodeling by creeping substitution to preserve the integrity of the femoral head. Currently using progenitor cells or growth factors 11 may be one of the solutions. Therefore, it would make sense not only to use core decompression but to introduce new cells.
Normal ossification is characterized by transformation of a cartilage template to bone attributable to colonization of the cartilage by bone cell progenitors, than by hematopoietic stem cells. 11,14 Red (blood forming) bone marrow first appears in the bones of the foot at approximately 1 year of age. In healthy children, red marrow fills all the bone cavities and the spaces in the cancellous bone of the epiphyses. The needs associated with the expansion of blood mass during childhood must be met, despite the limited volume of the medullary cavities. This limited volume is attributable to two characteristics of the skeleton in childhood, namely, an abundance of cartilage and very thick cancellous bone trabeculae. 11,14 When growth stops at the end of adolescence, needs in terms of hematopoiesis reach a plateau, and in many areas of the skeleton the red marrow converts to fatty (yellow) marrow devoid of hematopoietic activity. Red marrow persists in some areas such as the iliac crest. The characteristics of bone marrow are well-documented 3,17 and have been used therapeutically during the past few years, 3,10,14 although they have not yet received sufficient attention in the area of bone diseases. Red marrow contains not only a hematopoietic component but also contains a stromal component in which osteogenic bone cell precursors are found. The ability of bone marrow to induce bone formation was shown in rabbits as early as 1869 by Goujon 7 and has been used by contemporary investigators 3 to treat nonunions and to potentiate the osteogenic effect of bone grafts or bone graft substitutes. It currently is thought that osteogenic cells derive from stem cells present in the bone marrow stroma and, consequently, that red bone marrow transplants are a source of osteogenic precursor cells. 3,4,17
The treatment of osteonecrosis with bone marrow autografts is based on the view, now commonly held, 3,17 that the osteogenic cells derive from a stem cell in the bone marrow stroma. 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 still is round. 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 still is 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. The current authors have used this approach in the treatment of patients with Stage I and Stage II osteonecrosis.
The current study includes a large number of patients operated on by only one surgeon and followed up for as many as 10 years. It was shown that if done properly, the procedure had an extremely low incidence of complications and was effective in treating patients with earlier stages of avascular necrosis as compared with patients in previous studies 1,18 who were treated nonoperatively. Significant differences in outcome were observed in patients who had the greater number of transplanted cells, as opposed to patients with a decrease of the number of progenitor cells in the iliac crest and a decreased number of transplanted cells. This phenomenon may be related to an increased repair capacity in the region of borderline osteonecrosis when progenitor cells are transplanted in the osteonecroses and in the femoral head. Such a phenomenon (repair of osteonecrosis) has been reported previously with systemic allogenic bone transplantation in some patients. 14
In vivo osteogenesis occurs 3 only if the density of implanted cells at the treated site is sufficiently high. To achieve this, either large amounts of concentrated bone marrow or bone marrow in combination with a growth factor can be used. The volume of bone marrow to be collected should be determined 3,17 based on the number of progenitor cells needed for successful treatment. In a normal femoral head without osteonecrosis, 13 this number of nucleated cells per milliliter was estimated to be average 3 × 106 per mL and the number of colony-forming units was estimated 19 per million nucleated cells. According to the average volume of a femoral head (60 cm3) the number of colony-forming units present, in absence of osteonecrosis, may be estimated to be 3000. When osteonecrosis is present, the number of colony-forming units can decrease to less than 200. The number of cells transplanted was greater than the normal number expected in the femoral head in most patients. However, several variables influence the number of cells that are obtained and particularly the disease at the origin of the osteonecrosis. The number of nucleated cells may decrease because the marrow cavity is hypocellular and the number of colony-forming units in some patients also may decrease independently of local cellularity. This was observed particularly after organ transplantation. In these circumstances (low number of transplanted colony-forming units) the results were significantly lower. In some patients, a prior study of bone marrow cell density should be considered. Bone marrow injection in combination with growth factors could be one of the solutions in the future for these patients.
1. Aaron R, Lennow D, Stulberg B: The Natural History of Osteonecrosis of the Femoral Head. In Urbaniak J, Jones JP (eds). Osteonecrosis. Rosemont, IL, American Academy of Orthopaedic Surgeons 261–265, 1997.
2. Arlet J, Ficat P: Forage-biopsie de la tête fémorale dans l’ostéonécrose primitive: Observations histopathologiques portant sur huit forages. Rev Rhum Mal Osteoartic 31:257–258, 1964.
3. Connolly J, Grese R, Lipello L, Dehne R: Development of an osteogenic bone-marrow preparation. J Bone Joint Surg 71A:684–691, 1989.
4. Cui MD, Wang GJ, Dalian G: Steroid-induced adipogenesis in a pluripotential cell line from bone-marrow. J Bone Joint Surg 79A:1054–1063, 1997.
5. Ficat P, Arlet J, Vidal R, Ricci A, Fournial JC: Résultats thérapeutiques du forage-biopsie dans les ostéonécroses fémoro-capitales primitives (100 cas). Rev Rhum Mal Osteoartic 38:269–270, 1972.
6. Gilbert A, Judet H, Judet T, Ayatti A: Microvascular transfer of the fibular for the necrosis of the femoral head. Orthopedics 9:855–890, 1986.
7. Goujon E: Recherches expérimentales sur les propriétés du tissu osseux. J Anat Physiol 6:399–412, 1869.
8. Harris WH: Traumatic arthritis of the hip after dislocation and acetabular fractures: Treatment by mold arthroplasty: An end-result study using a new method of result evaluation. J Bone Joint Surg 51A:737–755, 1959.
9. Hauzeur JP: Pasteils JL, Orloff S: Bilateral non-traumatic aseptic osteonecrosis in the femoral head. J Bone Joint Surg 69A:1221–1225, 1987.
10. Hernigou P: Autologous bone marrow grafting of avascular osteonecrosis before collapse. Rev Rhum (Engl Ed) 62:650–651, 1995.
11. Hernigou P: Growth factors released from bone marrow are promising tools in orthopedic surgery. Rev Rhum (Engl Ed) 65:79–84, 1998.
12. Hernigou P, Beaujean F: Abnormalities in bone marrow of the iliac crest in patients with osteonecrosis. J Bone Joint Surg 79A:1047–1053, 1997.
13. Hernigou P, Beaujean F, Lambotte JC: Decrease of mesenchymal stem cell pool in the upper femoral extremity of patients with osteonecrosis related to corticosteroid therapy. J Bone Joint Surg 81B:349–355, 1999.
14. Hernigou P, Bernaudin F, Reinert P, Kuentz M, Vernant JP: Bone marrow transplantation in sickle cell disease: Effect on osteonecrosis. J Bone Joint Surg 79A:1726–1730, 1997.
15. Koo KK, Dussault RG, Kaplan PA, et al: Fatty marrow conversion of the proximal femoral metaphysis in the osteonecrotic hips. Clin Orthop 361:159–167, 1999.
16. Mitchell DC, Rao VM, Dalinka MK: Avascular necrosis of the femoral head: Magnetic resonance imaging appearance correlated with radiographic staging and radionuclide, histologic and clinical findings. Radiology 162:709–712, 1987.
17. Muschler G, Boehm C, Easley K: Aspiration to obtain osteoblast progenitors cells from human bone marrow: The influence of aspiration volume. J Bone Joint Surg 79A:1699–1709, 1997.
18. Musso ES, Mitchell SN, Schink-Aslani M, Basset CA: Results of conservative treatment of osteonecrosis of the femoral head: A retrospective review. Clin Orthop 207:209–215, 1986.
19. Steinberg ME, Hayen GD, Steinberg D: A quantitative system for staging avascular necrosis. J Bone Joint Surg 77B:34–41, 1995.
20. Urbaniak JR, Cougan PG, Gunneson EB, Nunley JA: Treatment of osteonecrosis of the femoral head with free vascularized fibular grafting. J Bone Joint Surg 77A:681–694, 1995.
21. Wang GJ, Lennox DW, Reger SL, Stamp RC, Hubbard S: Cortisone induced intrafemoral head pressure changes and its response to drilling decompression method. Clin Orthop 159:274–278, 1981.
22. Wang GJ, Sweet DE, Reger SL: Fat cell changes as a mechanism of avascular necrosis of the femoral head in cortisone treated rabbits. J Bone Joint Surg 59A:729–735, 1977.
Richard A. Brand, MD; and Joseph C. McCarthy, MD—Guest Editors