The Ranawat Award. Histology of nine structural bone grafts used in total knee arthroplasty

The cross section radiographs and histology of nine bone grafts were examined to determine whether grafts are durable enough to support a total knee implant when the load is shared by host bone, graft bone, and a stemmed component. All cases had cemented total knee arthroplasties with stemmed components adjacent to bulk grafts. The cases included autografts and allografts, which had been in situ for an average of 41 months (range, 20-62 months). Seven of the grafts were retrieved postmortem from three patients (four knees), and two were retrieved at revision surgery from one patient. The allografts all were intact, but had not revascularized. The autografts were viable bone. New bone was being laid down on the dead graft bone at the periphery of the allografts. No change in the bone to cement interface, no graft collapse, no development of radiolucent lines, and no component loosening occurred in these cases. The promising clinical results of bone grafts in total knee arthroplasties were confirmed by the examination of these grafts at the cellular level. Using stemmed components in bone grafted knee reconstructions may have increased graft durability and protected the grafts from fatigue failure.     

Acetabular reconstruction with impacted morselized cancellous allografts in cemented hip arthroplasty: a histological and biomechanical study on the goat

Bone defects in total hip arthroplasty revision surgery can be restored with different types of bone graft. The use of impacted morselized allograft chips in combination with cement is the treatment of our choice. To establish the incorporation capacity of the grafts and mechanical stability of the implant, an animal model in the goat was developed. An acetabular defect was created and restored with morselized grafts and a cemented cup. Postoperative performance of the reconstruction was followed both histologically and biomechanically. Histology showed that consolidation of the graft with the host bone bed had occurred within 3 weeks. In the following period a front of vascular sprouts infiltrated the graft. Graft resorption, woven bone deposition, and subsequent remodeling resulted in a new trabecular structure. This structure contained only scarce remnants of the original dead graft material. At the graft-cement interface, graft resorption and new bone formation had resulted in areas of direct vital bone-cement contact. Locally, a soft tissue interface was present. After longer follow-up periods, progressive interface formation and loosening of the cups were found in most animals. Mechanical testing showed that the stability of the reconstruction increased during the first 12 postoperative weeks. Thereafter, the stability decreased, probably by soft-tissue interface formation at the graft cement interface. We conclude that cemented morselized allografts have a high capacity to incorporate. Initial cup stability is adequate to provoke graft incorporation with decreasing stability after the incorporation process has been completed.     

Structural allograft in two-stage revisions for failed septic hip arthroplasty

We report 11 patients having revision of total hip arthroplasty using massive structural allografts for failure due to sepsis and associated bone loss. All patients had a two-stage reconstruction and the mean follow-up was 47.8 months (24 to 72). Positive cultures were obtained at the first stage in nine of the 11 patients, with Staphylococcus epidermidis being the most common organism. The other two patients had draining sinuses with negative cultures. There was no recurrence of infection in any patient. The mean increase in the modified Harris hip score was 45 and all the grafts appeared to have united to host bone. Two patients required additional procedures, but only one was related to the allograft. Complications included an incomplete sciatic nerve palsy and one case of graft resorption. Our results support the use of massive allografts in failed septic hip arthroplasty in which there is associated bone loss.     

The Rotaglide total knee arthroplasty. Prosthesis design and early results

The Rotaglide knee (Cozim Medical, UK) is a three-part knee containing a mobile polyethylene meniscal platform, imparting reduced loosening forces to the tibia. The femoral component design provides a high degree of congruency throughout the range of motion from 0 degree to 110 degrees, and both femoral and tibial components ensure minimal bone removal. The system is versatile, including a large number of component options, and it may be used in both primary and revision arthroplasties. The first results in 170 cemented knees (161 patients) with 2- to 5-year follow-up periods (average, 3.1 years) were very encouraging (excellent or good in 95% of knees, based on the British Orthopaedic Association knee assessment system). Poor results were seen only in revision cases. There have been no mechanical implant failures and no platform bearing dislocations, and the platforms continue to move as documented by postoperative roentgenograms, which show the metal markers of the platforms moving anteriorly in flexion and posteriorly in extension.     

Clinical use of bone allografts

Modern techniques of bone allograft surgery provide a treatment modality for management of difficult skeletal defects. In oncological limb-salvage surgery, allograft reconstructions permit re-establishment of skeletal continuity and function after a wide resection of bone tumour. Bone allografts are increasingly used in salvage of difficult bone stock deficiencies following failed total joint replacements. Union between the allograft and the host bone takes place slowly and the use of autogenous bone graft at the graft-host junction is recommended for induction of repair. Internal repair (revascularization and substitution of the original graft bone with new host bone) also progresses slowly and seems to be confined only to the superficial surface and the ends of the graft. Biomechanically, a massive allograft may serve a structural function in the absence of advanced revascularization and creeping substitution processes. Infection of an allograft is a disastrous complication, whereas non-union of the graft-host junction and fracture of the graft are amenable to surgical treatment. Osteochondral allografts tend to show gradual deterioration of the articular cartilage with time, necessitating occasionally late resurfacing arthroplasty. It is evident that there is more active immune response to osteochondral grafts than was thought previously. Bone allografts induce cell-mediated and antibody-mediated cytotoxicity specific for donor antigens similar to that seen after organ transplantations. Not only the basic mechanisms of bone allograft rejection but also the clinical features of bone allograft rejection are poorly characterized. Clinically, new non-invasive imaging techniques should be applied in determining the metabolic activity of bone in order to find the optimal loading of healing allografts. Although the clinical results of massive bone allografts are still not completely predictable, the method has proved to be a technically and biologically feasible alternative for non-biological skeletal reconstructions.     

Spatial differences in gap junction gating in the lens are a consequence of connexin cleavage

Thirty patients had 32 cementless total hip arthroplasty revisions and were evaluated postoperatively for clinical function (Harris Hip Score) and radiographic evidence of implant stability. Of the 26 femoral components revised, 16 were revised with anatomic long-stem femoral prostheses, and 10 were revised with straight mid-stem-length components. All components were collared and had circumferential proximal fiber-mesh porous coating. Seven of 16 patients had radiographic subsidence after revision with long-stem components (2 to 30 mm); 6 of 10 patients had subsidence after revision with mid-stem femoral components (2 to 25 mm). Of the 13 patients with femoral subsidence, 8 had calcar reconstruction with allograft bone; of the 13 patients without radiographic subsidence, 8 did not require calcar reconstruction. One of 27 fiber-mesh, porous-coated acetabular components migrated (30 mm). No components have been removed or revised. Even with circumferential proximal porous coating, femoral implant stability remains unpredictable in total hip arthroplasty revision.     

Meniscal allograft transplantation for unicompartmental arthritis of the knee

Sixty-seven meniscal allografts were transplanted in the knees of 63 patients between 1988 and 1994. Before surgery, all patients experienced refractory disabling knee pain secondary to a prior total meniscectomy with advanced unicompartmental osteoarthritic changes as verified by arthroscopy. At a mean followup of 31 months (range, 1.0-5.5 years), 58 knees (86.6%) attained a good to excellent results-Twenty-one knees received isolated meniscal allografts, with 19 achieving good to excellent results (90.5%). Five knees received a medial or lateral meniscal allograft with an anterior cruciate ligament reconstruction, and 4 (80.0%) obtained good to excellent results. Thirty-four knees received a meniscal allograft in combination with either a valgus high tibial osteotomy, varus high tibial osteotomy, or varus distal femoral osteotomy to correct for preoperative varus or valgus deformities, with 29 (85.3%) attaining good to excellent results. The remaining 7 knees underwent a combined medial meniscal allograft, valgus high tibial osteotomy, and anterior cruciate ligament reconstruction with 6 (85.7%) attaining good to excellent results. The most frequent complication was a traumatic posterior horn tear in 6 knees at a mean of 21 months after surgery (range, 9-43 months), most likely the consequence of unsuccessful healing of the posterior horn of the graft.     

Atractylodes lancea autotetraploids induced by colchicine treatment of shoot cultures

We compared Boneloc bone cement with conventional cement (Palacos) in fixating the tibial component during 2-5 years in 19 patients with gonarthrosis undergoing total knee arthroplasty in a prospective randomized study. Boneloc displayed significantly larger migration, subsidence and lift-off than Palacos. The difference was identifiable already within 3 months postoperatively, but became significant at 12 months. In the Boneloc group, all components showed subsidence of the posterior part and lift-off of the anterior part of the tibial component, whereas in the Palacos group, the locations of subsidence and lift-off were evenly distributed about the edge of the implant. At 5 years, both Boneloc knees so far investigated were clinical failures with high migration rates. We conclude that, even in total knee arthroplasty, there is a substantial risk that Boneloc leads to inferior clinical results, but later than in hip replacements.     

Radiotherapy for elderly patients

Fifty-seven revision total hip arthroplasties in fifty-six patients were performed with a Harris-Galante porous-coated acetabular component by one surgeon, and the patients were followed prospectively for a mean of seven years (range, five to twelve years). A trochanteric osteotomy was performed in forty hips, and a posterior approach with an extended anterior capsulectomy was used in the other seventeen. The acetabular defect was classified as segmental in seven hips, cavitary in twenty-three, and combined in twenty-one; six hips had no notable defect. A bulk allograft was used in eleven hips, and morseled cancellous-bone allograft or autogenous graft was used in thirty-four hips; twelve hips did not have bone-grafting. Both the femoral and the acetabular component were revised in forty-five hips, and only the acetabular component was revised in twelve. Thirty-nine hips (68 per cent) had a good or excellent clinical result according to the Harris hip score. The acetabular component was well fixed in the fourteen hips that had a fair result and the four hips that had a poor result. The acetabular component was considered to have migrated if there was a change in the angle of the cup of 5 degrees or more or a change in the horizontal or vertical position of the cup of more than three millimeters. Despite varying degrees of bone loss, no acetabular component had radiographic evidence of loosening at the latest follow-up examination. No component was revised and no revisions were scheduled. One hip was debrided for a late metastatic infection, but the component was well fixed and was not revised. There were no complications related to the use of screws for fixation. These mid-term results confirm the early success of acetabular revisions performed with fixation of a titanium fiber-metal-coated hemispherical component with multiple screws and no cement.     

Bone allografts are immunogenic and may preclude subsequent organ transplants

The authors report a case of a 41-year-old woman with diabetes and chronic renal failure in whom antihuman leukocyte antigen antibodies developed after she received a frozen bone allograft that limited her access to organ donors. The patient had a chondrosarcoma of the right distal femur. A wide resection with segmental total knee arthroplasty was followed by a revision using a composite bone allograft prosthesis. After revision, broadly reactive lymphocytotoxic antibodies developed in the patient. The patient's panel reactive antibody level rose from 28% to a peak of 70%. Panel reactive antibody expresses the percentage of a panel of human leukocyte antigen type T lymphocytes from 40 individuals (representative of all human leukocyte antigen Class I histocompatibility antigens) to which antihuman leukocyte antigen Class I lymphocytotoxic antibodies have developed in the recipient as measured by the antiglobulin crossmatch method. The specificity of the patient's primary antibody is found in 45% of donors available in Illinois since 1988 (N = 1606). Because a positive crossmatch precludes kidney and pancreas transplantation, at least 45% of cadaver organ donors were excluded from use for this patient. This is an unusual case that focuses on the potential impact of bone allografts in patients who may need subsequent organ transplantation.     

Characteristics of impaired awareness after traumatic brain injury

The condylar constrained total knee arthroplasty was performed on 29 patients undergoing 33 procedures and were reviewed clinically and radiographically at an average follow-up of 5 years (range, 2-10 years). There were 21 women and 8 men. The average age at the time of surgery was 70 years (range, 32-84). Of the 16 knees that were revision total knee arthroplasties, 8 had a previous infected total knee arthroplasty, and 17 knees had severe deformities requiring the use of the condylar constrained prosthesis. The patients were rated according to the Knee Society clinical and radiological evaluation protocol. Measurements of femoral and tibial component position were obtained as well as femoral tibial angle, patella position, and cement bone radiolucencies. All clinical measurements were made by an independent physical therapist. Clinical results revealed an improvement from an average preoperative knee score of 38 points to an average postoperative score of 86 points. The clinical results for 19 (58%) knees were excellent, 8 (24%) had a good result, 1 (3%) was fair, 2 (6%) were poor, and 3 (9%) were failures. The patients' average functional levels increased from 24 to 58. The final average flexion was 96 degrees. Three knees have been revised (9%). One was revised for recurrent infection, one for periprosthetic fracture, and one for mechanical loosening of the tibial component. There were no other knees with evidence of radiologic loosening. We conclude that the condylar constrained total knee prosthesis provides an acceptable solution for revision and complex primary total knee replacements at an intermediate follow-up term of 5 years.     

Duo-condylar knee arthroplasty: hospital for special surgery design

This is a report of 94 knees in 88 patients with the duo-condylar type of knee arthroplasty. The follow-up period of time was between 2 to 4 years with an average of 3 years. The rheumatoid to osteoarthritic patient ratio was 3 to 1. The overall results were excellent in 37.5 per cent, good in 37.5 per cent, fair in 16 per cent, and poor in 9 per cent. The main causes of failure and poor results were: (1) under or over correction of deformity leading to subluxation and/or instability of the knee; (2) loosening of the tibial component, and (3) symptoms arising from the patellofemoral joint. The revision rate is 5.5 per cent. The progressive radiolucency at the cement bone bond is 26 per cent of which 16 per cent is up to 1 mm and 10 per cent is between 1.5 to 3 mm. To further improve the results of arthroplasty, one should take into consideration (1) replacement of the patellofemoral joint, (2) insertion of the prosthesis in the proper anatomical location under correct tension of the ligaments and capsule with the help of proper instrumentation and (3) improvement in fixation of the tibial component.     

Diagnosis and treatment of osteogenic sarcomas, techniques of reconstructions. Future prospects

PATIENTS AND METHODS: We have performed 34 massive bone-cartilage grafts with a follow-up of 2 to 7 years (1988-1993) including 5 complete joint grafts of the knee. Between 1988 and 1993, 8 massive diaphyso-metaphyseal bone grafts were performed. Joint reconstructions using massive bone-cartilage allografts are increasingly used in routine oncology surgery. Long-term rehabilitation and possibilities of immediate anatomic reconstruction of the articular surface, together with mid-term results suggest that the functional results are promising compared with major reconstruction prostheses. Indications for operations are being increasingly widened to younger subjects who have undergone partial or total joint exeresis for tumour. Sleeved prostheses were used for 12 reconstructions (1988-1993) for sarcoma of the knee. RESULTS The risk of sepsis are comparable for the different groups and are mainly related to the quality of the skin repair during chemotherapy. Fractures of the graft occur when the fixation is insufficient or when rehabilitation exercises were too aggressive. Non-consolidation was exceptional when the junction between the allograft and the receiver bone is not surrounded with autologous spongious autografts. Joint instability and arthrosis depend on the stability of the ligament reconstruction. To this day, no Charcot type joint disease has been demonstrated, periarticular innervation has maintained joint trophism. DISCUSSION: There are still some incompletely resolved problems concerning the revascularization of the graft, its integration into the skeleton, the outcome of the grafted cartilage and that of the ligament formations attached to the graft or used as allografts. These massive grafts must be studied over a longer period of time but the early results are encouraging. Sleeved grafts using bone-bank specimens could be an intermediary solution which appears to be indicated in cases where the tumoural resection was particularly large removing bone, cartilage, ligaments and muscles. With these sleeved prostheses, the muscles can be refixed onto the graft thus reducing the risk of shank fracture and loosening. The use of a tibial graft with the patellar tendon is helpful in reconstructing the extensor apparatus. However, if rehabilitation is not undertaken rapidly and followed regularly for several months, the graft favours the development of muscular adherances which can be a major limitation to joint mobility.     

Implant stability in revision total hip arthroplasty: allograft bone packing following extended proximal femoral osteotomy

One method of revising the femoral component in revision total hip arthroplasty in the presence of compromised femoral bone stock is to pack the upper femur with particulate allograft and then to cement the femoral component into the allograft bed. This technique is being used clinically with encouraging results. Additionally, surgical exposure of the femoral canal during revision total hip arthroplasty can be greatly improved with an extended trochanteric osteotomy, which is subsequently repaired with wires or cables. To assess the feasibility of performing the allograft bone packing technique following an extended trochanteric osteotomy, the stability of this construct in a cadaver model was measured, using micromotion sensing instruments and loads applied on a materials testing machine. The stability of the cemented allograft impaction construct following extended trochanteric osteotomy was comparable to the stability of the control construct, which consisted of a similar impacted allograft construct without osteotomy. The stability of the osteotomized side was comparable to that of the control side. It is concluded that the initial in vitro stability of the allograft impaction technique following extended proximal femoral osteotomy is adequate to justify experimental in vivo use.     

Contents of total and protein-bound carbohydrates are low in leukemic leukocytes from patients with acute myelogenous leukemia

The successful approach to the failed knee with bone deficiency is dependent upon thorough planning prior to surgery in order to have the resources available in terms of adequate bone allograft and suitable revision implants. The approximate size of bone stock deficiency can be calculated from preoperative radiographs and similarly ligamentous incompetence can often be diagnosed clinically prior to surgery. Smaller defects of up to 1 to 1.5 cm in depth and localized in the main to a single side of the tibial plateau or to a single femoral condyle can be dealt with using smaller grafts that may be local autograft or allograft, or modular wedges. Larger tibial defects can be compensated for using conventional revision systems by thicker polyethylene and augmented baseplates, but once the flexion-extension gap reaches approximately 40 mm this is no longer possible and structural graft or customized componentry becomes necessary. Femoral defects larger than about 1 cm that cannot be made up by augments necessitate grafting. The need to use a large proximal tibial allograft also may dictate the operative approach used to expose the joint, especially in the situation of a multiply-operated tight knee. In such cases the use of a quadriceps turndown may be more advisable than the use of a tibial tubercle osteotomy as the osteotomy may well not have an adequate bed to heal to following the reconstruction. Several series have reported cases of patellar tendon avulsion and the clinical results following this complication usually are not satisfactory. Preoperatively it is important to identify, if possible, the case that is likely to require a more extended approach because of a tight soft tissue envelope. The reports of results of series of revision total knee arthroplasty in the setting of significant bone loss are at present confined to short-term followup. The clinical results of these series are satisfactory at this early point in time, but decision regarding the durability of reconstructions requiring major structural allografting awaits longer-term study. Of concern is the devastating complication of infection following such revision surgery, the risk of which is amplified in the setting of prior infection. In addition, the long-term viability of major structural grafts in the setting of loading is uncertain as the risk of graft collapse in the process of incorporation is not known. Notwithstanding these concerns, major grafting is sometimes the only recourse to achieve satisfactory revision of a failed arthroplasty. The use of such major grafts is therefore cautiously supported and because of the risks inherent in such surgery we believe that such surgery should be carried out in the setting of specialist interest units.     

Urinary bladder marsupialization for treatment of obstructive urolithiasis in male goats

In this study, 18 patients (17 men and 1 woman; mean age 61 years) with a previously infected vascular graft underwent vascular reconstruction with cryopreserved arterial allografts. Treatment consisted of first total (n = 11) or partial removal (n = 7) of infected prosthetic grafts. Revascularizations were aortoaortic (n = 2), aortobifemoral (n = 8), aortounifemoral (n = 3), femorofemoral (n = 2), iliofemoral (n = 1), or femoropopliteal (n = 2) bypasses. Four patients died postoperatively (22%)-one of septic necrosis of the allograft, one of septic rupture of the aortic anastomosis of a previous bypass, one of multiorgan failure, and one of mesenteric infarction. One allograft occluded within 30 days (5.5%), leading to an above-knee amputation. In the remaining patients, routine arteriography or duplex scan showed patent allografts. For the 14 survivors, the mean follow-up period was 20 months (range: 1-45 months). Two patients died-one of septicemia not related to the allograft, and one of multiple organ failure. Among the 12 survivors, 3 patients with non-ABO-compatible allografts developed different types of long-term alterations. One patient had a hemorrhage due to femoral allograft rupture at 45 days, and two patients had aortic allografts dilatation with mural thrombus, necessitating a prosthetic replacement in one patient. Cryopreserved allografts used for the treatment of infected vascular graft are useful in selected cases, although they are not totally resistant to infection. Patients should be followed closely to detect significant long-term alterations of the allografts.     

Migration and wear of hydroxyapatite-coated press-fit cups in revision hip arthroplasty: a radiostereometric study

Screw-fixated and hydroxyapatite-coated press-fit cups were studied using radiostereometry in 29 revision and 14 primary arthroplasties. The acetabular defects in the revision cases varied from none to type 3 (wall defects) according to Gustilo-Pasternak. Morsellized allograft was used in 25 revisions. Nine of these cups rested on less than 50% living bone. After 2 years, the mean migration in the revised group reached 0.36, 0.21, and 0.49 mm in the horizontal, longitudinal, and anteroposterior (AP) directions. The mean rotations varied between 0.5 degrees and 0.7 degrees depending on direction. The primary implants displayed smaller mediolateral migration and AP tilt. The mean proximal wear rate for the whole group was 0.11 mm/y. A central gap on the postoperative AP view implied less migration. The size of the preoperative bone defects or amount of bone-graft used had no influence on the migration. Despite extensive use of morsellized allograft, this implant displayed the smallest migration so far reported in revision hip arthroplasty.     

Primary radiotherapy of carcinoma of the supraglottic larynx--a multivariate analysis of prognostic factors

Managing the patella and balancing the patellofemoral joint space is one of the most difficult aspects of performing a primary total knee replacement (TKR). The situation is compounded in the revision situation. Unfortunately, an otherwise well-performed TKR will fail because of problems with the patella or the extensor mechanism. This article discusses various aspects of the management of the patella and extensor mechanism during revision TKR with an objective of minimizing complications and maximizing functional outcomes. The topics covered include exposure of the patella and extensor mechanism during revision surgery, whether or not to remove all prior patellar implants, the technique for removal of a prior implant, the management of bone loss or fractures of the patella during revision TKR, the insertion (or noninsertion) and fixation of a new implant, and the balance of the patellofemoral joint space, including avoidance of patella baja or patella alta. A compilation of scientific and "no-so" scientific data and experience gleaned over the past 26 years of total knee replacement surgery will be presented.     

Revision of failed prosthetic anterior cruciate ligament reconstruction

Revision of failed prosthetic anterior cruciate ligament reconstructions will continue to be necessary as previously placed prosthetic devices fail with time. These patients often present with recurrent instability, pain, swelling, or effusions. Graft rupture and the generation of particulate debris are common causes of these symptoms. To effectively treat these patients requires careful preoperative evaluation and planning. Operative treatment includes removing the prosthesis and metal fixation devices, evaluating femoral and tibial bone stock, and determining adequacy of previous tunnel position. Staging of the operation may be a necessity if significant bone loss or poor tunnel position on either the femoral or tibial side requires bone grafting. The use of an autogenous bone-patellar tendon-bone graft is suggested and has proven to be effective in restoring knee stability in these revision cases. The ultimate outcome following revision of failed prosthetic ligaments may be limited by associated intraarticular pathology often seen in this patient population.     

Intracellular distribution of phospholipase C gamma 1 in cell lines with different levels of transformation

PURPOSE OF THE STUDY: To assess after 83 months of follow-up, the results of 19 femoral revisions carried out according to an original method combining a cemented stem and bone reconstruction by means of impacted-morcelized bone allograft protected by a titanium mesh. MATERIALS: Twenty hips (18 patients mean aged 58 at surgery) were included between 1986 and 1991. Five hips had a least one previous prosthetic revision, one hip was revised because of septic loosening. No patient was lost for follow-up, but two had died during the follow-up period: one patient died one month after surgery was excluded, one other died 7 years after the index procedure and was included with his last hip rating. Loss of femoral bone stock was severe according to the SOFCOT four stage rating system: 2 femurs were grade II, 14 grade III, and 3 grade IV. Femoral stem migration was assessed with landmarks recommended by Walker. All the measurements were performed with a digitizer (OrthoGraphics). METHODS: All the procedures were carried out through a posterolateral approach, augmented by 4 trochanteric osteotomies and 5 distal femoral windows. After prosthesis and cement removal, a bone plug was placed into the medullary canal. Then, cancelous bone morcelized allografts were impacted in the femoral defects through the medullary canal. A titanium mesh cylinder was placed into the femur to separate the graft from the cement introduced later to obtain fixation of the revision stem. The stem was extended about 5 centimeters over the distal edge of the grafts in order to bridge the femoral defects. The mesh was extended only in front of the grafts and was used to protect them from excessive cement penetration. RESULTS: Functional improvement was noticeable since the Merle d'Aubigné Hip score improved from 9.8 to 16.3 at follow-up. The pain score improved from 2.1 to 5.5 and walking score from 2.3 to 5. Adverse effects occurred during the first cases and were related to cement removal: 3 greater trochanter fractures, 5 distal femoral perforations and 2 non displaced femoral shaft fractures. The septic revision had recurrence of infection associated with radiolucent lines > 2 millimeters and the only one graft resorption. One trochanteric non-union was observed but no prosthetic dislocation. Only one femoral stem migration (4.4 millimeters) was detected without any other radiographic features of loosening after 9 years of follow-up. This stem was considered as loosed, but was not revised because of few clinical symptoms. Only 2 radiolucent lines less than 2 millimeters at the bone cement interface in Gruen's zones 3 to 5. Likewise, no radiographic feature of stress-shielding was observed. On follow-up X-rays, 3 hips had corticalisation of the grafts, and 12 hips demonstrated normal cancelous trabeculations in the grafts. CONCLUSION: Satisfactory functional and radiographic results were obtained with this method after 5 to 10 years of follow-up instead of severe preoperative femoral bone stock impairement. Likewise, we observed only one recurrence of loosening diagnosed with the help of digitized X-ray examination. Only one significant (> 3.5 mm) femoral stem migration was detected. Radiographic features of femoral reconstruction were observed but without histologic proof of graft integration. This method uses a longer stem than the "Exeter", but avoids a high rate of femoral stem migration and appears compatible with femoral bone reconstruction.