Purpose of the study: Follow-up of patients with a total hip arthroplasty with an Atlas^® cup revealed unique acetabular osteolytic defects which remained asymptomatic for long periods. We thus conducted a systematic review.

Material and method: Our retrospective analysis included 217 Atlas^® elastic impactable cups implanted consecutively from January 1993 to June 1995 and reviewed clinically and radiographically at mean 13.1 years.

Results: The incidence of acetabular osteolysis was 16%, the leading cause of replacements which occurred on average 8.8 years after the initial implantation. The overall actuarial cup survival was 76% at 13.5 years and 81% taking revision for periprosthetic osteolysis as the endpoint. Univariate analysis found a significant link between osteolysis defects and significant wear (p< 0.0001), Devane activity 4 or 5 (p=0.0005), low thickness polyethylene (p=0.006), and use of Zircone or alumina heads versus metal heads (diameter 22). There was no statistical link between the presence or not of a hydroxyapatite coating, despite a trend for less osteolysis with hydroxyapatite coating. At multivariate analysis, the only factors significantly linked with the presence of osteolysis were significant wear and thin polyethylene insert in the metal back.

Discussion: Mid-term follow-up appeared sufficient to assess the development of osteolysis considering the delays describe by others. The incidence of osteolysis, despite the high incidence and early appearance in our series, was very probably underestimated by the radiographic analysis. These results suggest we should search for defects systematically with this type of implant, especially if there is measurable wear. Although it is difficult to set a cut off, a polyethylene thickness less than 10mm should be avoided to limit the high risk of osteolysis defects. A more powerful statistical analysis and examination of explants is advisable in order to ascertain the reasons for this abnormally high level of failure by osteolysis at 13 years follow-up.

Conclusion: Systematic radiological and clinical review of all patients demonstrated an insufficient overall actuarial survival and an important rate of periprosthetic osteolysis. The association between periprosthetic osteolysis and polyethylene were was confirmed. Thin inserts appear to play an important role in this osteolysis via an abnormal wear and poor tolerance to wear debris from these implants.

Purpose of the study: Spontaneous repair of lost deep chondral tissue is minimal in the knee joint. A clinical trial of chondrocyte autografts as described by Brittberg and Peterson was undertaken by the Nantes University Hospital and the French Society of Arthroscopy in 1999.

Material and methods: Twenty-eight patients, mean age 28 years, underwent surgery in eight centers. Etiologies were: osteochondritis (n=14), isolated posttraumatic chondorpathy (n=7), chondropathy and full-thickness ACL tear (n=7). All lesions involved the condyles and were deep (ICRS grades 3 and 4). Mean surface area involved after debridement was 490 mm2 (range 150–1000 mm2). Patients were followed three years after the autologous grafting to assess functional outcome. An MRI was obtained at 2–3 years. Thirteen control arthroscopy procedures were performed including eight with biopsy specimens for histology and immunohisto-chemistry studies.

Results: Twenty-six patients were reviewed at more than two years. There were no general complications, three patients presented a partial avulsion of the graft treated by arthroscopy and one underwent arthrolysis at six months. Function improved in all patients except three and pain improved in all. The ICRS score improved from 43 points (range 19–70) to 77 points (range 39–84). Sixteen control MRIs were available and showed that the graft was hypertrophic in eleven cases, on level in four, and insufficient in one. Marginal integration was good in 11 cases and partial in five. Subchondral integration was complete in ten cases and mediocre in six. The arthroscopic score was nearly normal (score 8–11) in eight cases and abnormal in five (score 4–7). The histological class according to Knutsen (hyaline richness) was: one in group 1 (> 60%), three in group 2 (> 40%), four in group 3 (< 40%) and one in group 4 (bony or fibrous tissue). Function score (r=0.78 and MRI score (r=0.76) were correlated with arthroscopic sores. There was no correlation with the histological results.

Discussion: Clinical outcome was improved in more than 80% of cases, similar to results reported for histological series. The arthroscopic and histological results were equivalent to those reported by Knutsen but inferior to those reported by Bentley or Peterson.

Purpose of the study: We conducted a prospective randomized study to compare use of macroporous biphasic calcium phosphate ceramic and bone autografts for filling medial open wedge osteotomies of the proximal tibia.

Material and methods: This phase III pragmatic clinical trial was designed for direct patient benefit. Randomization was performed in the operating room after completing the osteotomy. Twenty-six men and 14 women, mean age 51 years (range 19–75 years) were included. A biomaterial implant was used for 22 patients and an autograft for 18. Mean correction was 10 mm (range 6–15). One patient was excluded from the analysis, no patient was lost to follow-up. All patients were reviewed at minimum two years follow-up.

Results: At three months, knees were less painful with less subjective functional impact after filling with an autograft (pain VAS 3.1 versus 2.1 and function VAS 3.4 versus 2.5). These results were more balanced at six months (pain 1.6 versus 1.8 and function 1.8 versus 2.1) and remained stable at one and two years. The IKS knee scores were symmetrical at one at two years for both groups (IKS1 93 versus 86 and IKS2 90 versus 90). Bone healing was achieved within the usual delay. Axial alignment was stable at two years in both groups. There were eleven complications (28%), nine requiring revision: infected hematoma (n=1), intraoperative vascular injury without serious consequences (n=1), loss of correction (n=2), nonunion after filling with biomaterial (n=1), iliac abscess after filling with autograft (n=2) and painful calcification of the iliac region requiring resection (n=1).

Discussion: We observed three factors which can favor mechanical failure after filling with a ceramic material: intraoperative rupture of the lateral hinge, obesity, and excessively early unprepared weight bearing.

Conclusion: Although the difference did not reach significance, the risk of mechanical complications appears greater with macroporous en bloc ceramic filling. This material is less tolerant to comorbid conditions (obesity) and requires very precise technique as well as careful observance of postoperative care (no early weight bearing). Nevertheless, this method does have the advantage of avoiding painful sequelae and complications related to harvesting the iliac graft.

Purpose of the study: Revision acetabular surgery with bone stock deficiency is a difficult problem. The use of cementless component and bioactive ceramics seemed to be a promising alternative.

Since 1996, we have been filling bone defect at the time of revision with macroporous calcium phosphate ceramic. We reported our first experience between 1996 and 1999.

Material and methods: The procedure was carried out in 35 hip reconstructions ( 35 patients ) at a mean follow-up of 6 years ( range 5 to 7,4). The average age of the patients was 56 years( range 28 to 83).

2 patients died of a cause unrelated to the procedure and 2 patients were lost of follow-up. Bone defect were classified into type I ( 4 hips), type IIA ( 8 hips ), type IIB ( 5 hips), type IIC ( 9 hips), IIIA ( 4 hips ), type IV ( 5 hips ) according to Paprosky classification.

The functional status of the patients was evaluated according to the Merle d’Aubign ip rating.. The interfaces bioactive ceramics/bone base and bioactive ceramics/cementless component, as well as the homogeneity and the density of the graft were examined radiologically.

Results: Functionally, the Merle d’Aubigné hip rating improved, increasing from11,3 to 15,9. Failure of fixation of the acetabular component occurred in 11,4 % of the acetabular reconstructions ( 4 hips in 35 patients). One was diagnosed as loose on the basis of radiographic criteria alone and the other three hips had a loose acetabular component at reoperation. We observed no failed reconstruction when the acetabular component is in direct contact with host bone on 50 % or more of its surface area.

We saw no radiolucent lines or spaces at the interface between bioactive ceramics and the host bone. Morphological changes or a decreased in graft volume were not seen, except for the patient with the loose cup.

Discussion-Conclusion: The findings of the present study support the use of bioactive ceramics and cement-less acetabular component in the presence of loss bone in order to achieve the goals of a revision hip replacement, provided that at least 50 % support of the cup can be obtained with host-bone.

Purpose: In osteosarcoma, tumour progression leads to osteolysis via direct proteolytic mechanisms and/or osteoclast activation. Nitrogen biphosphonates (N-BP) like zolebronate inhibit osteoclast function and apoptosis of osteoclasts and other tumour cells. In animal models, N-BP decrease bony progression of myeloma, bone metastasis, and breast and prostatic tumours. In vitro studies have demonstrated a synergetic action with classical anti-cancer drugs on apoptosis for myeloma and breast cancer cell lines. The purpose of the present study was to investigate the effect of zoebronic acid on osteosarcoma growth, alone or in combination with ifosfamide.

Material and methods: A rat model accepting osteosarcoma transplant was used for the study. Four series of seven rats were treated with zoledronate (100 mg/kg on day 7, 14, 21 and 28 after implantation) in combination or not with ifosfamide (30 mg/kg on day 1”, 14 and 15). Thirty-five days after implantation, the rats were sacrificed to evaluate tumour volume, presence of metastasis, radiography, and pathological examination of the tumour. Zoledronate was also studied in vitro on an OSRGA osteosarcoma cell line isolated from the same tumour.

Results: Zoledronate demonstrated efficacy by reducing the osteolysis induced by the sarcoma, but also on local tumour progression (75%) in comparison with untreated animals. In vitro, zoledronate inhibited cell proliferation by 60%. The ifosfsamide-zoledrnoate combination produced greater reduction in tumour progression than ifosfamide alone.

Conclusion: This work demonstrates for the first time that zoledronate has an effect on osteosarcoma tumour progression, either by a direct effect or by an antiosteoclastic effect and that the effet increases the efficacy of classical antitumour drugs such as ifosfamide.

Background: Certain pelvic fractures with posterior shearing force lesions raise a difficult problem for stabilisation. Assemblies bridging the two sacroiliac joints and anchored in the vertebral pedicles have been proposed to overcome this difficult osteosynthesis problem. But this type of fixation bridges intact joints. We used an original computer-assisted osteosynthesis technique in a patient with an unstable non-union of a displaced fracture passing through the S1 pedicle and the sacral foramina associated with major pubic disjunction.

Case report: The instability of the half-pelvis led to pubalgia and movement of the non-union focus which in turn led to S1 sciatalgia preventing the sitting position. The fracture and the displacement could not be treated initially because of open visceral lesions which required definitive colostomy and a long period of intensive care. The technical problem was to stabilise the pelvis with a posterior osteosynthesis which could not use the S1 pedicle. The fixation had to be sufficiently rigid to compensate for the impossible anterior fixation of the pubic symphesis.

Operative technique: The original solution was to bridge the non-union transversally using two spinal bars (CD instrumentation) applied on the healthy side with two polyaxial screws in S1 (in the pedicle and the wing) and on the non-union side with two polyaxial screws inserted in the iliac wing passing between the internal and external corticals. The assembly was completed with two transiliosacral screws passing through S1 and S2 after demounting and avivement of the non-union.

Discussion: The computer superposed a virtual image of the instruments on the CT images allowing precise insertion of the two iliac screws which had to pass between the bone tables over a long trajectory. The two transiliosacral screws could thus be inserted into S1 and S2 percutaneously with minimal neurological risk. Monitoring the progression allowed safe avivement of the nonunion to the anterior border of the sacrum. The final assembly thus associated four screws and two bars on the posterior part of the sacrum and two anterior screws, enabling perpendicular compression of the nonunion while maintaining the spine and contralateral sacroiliac joint. At five years, the patient remains free of dysesthesia and can sit and walk with little limitation.

Conclusion: This particular case illustrates perspective computer-assisted osteosynthesis methods.

Purpose: Circumferential arthrodesis of the lumbar spine is necessary in certain selected situations (lumbar stenosis with instability and preserved disc height or spondylolisthesis). Posterior lumbar interbody fusion (PLIF) raises the risk of significant bleeding and fibrosis around the roots as well as neurological complications. Transforaminal lumbar interbody fusion (TLIF) can avoid excessive bleeding and root displacement. The cages are inserted via a unilateral approach.

Material and methods: This prospective single-centre study included twenty patients (nine men and eleven women), mean age 49 years. Indications for lumbar surgery were degenerative spondylolisthesis in nine patients and discal lumbar pain with foraminal stenosis in five. The clinical status was assessed with the Oswestry score, SF-36 and a visual analogue scale (VAS). Radiological assessment was based on inter-body fusion, segmentary lordosis, and lumbopelvic parameters. TLIF was associated with a posterior approach for insertion of titanium pedicular screws (CDH, Medtronic Sofamor Danek). Temporary unilateral distraction opened the foramen. Unilateral arthrectomy enabled a lateral approach to the disc without involving the roots and avoiding any movement of the dural sac. The disc was resected and the body endplates were prepared before introducing two cages (pyramesh) filled with macroporous ceramic granules (BCP) mixed with autologous bone marrow. Installation to two contourned rods enabled segmentary compression to stabilise the cages in association with posterolateral fusion.

Results: Mean operative time was three hours. Mean blood loss was 400 ml. The patients were verticalised on day three without a corset. Mean follow-up was six months with retrospective evaluation of the Oswestry score, SF-36, and VAS. Postoperative pain resolved rapidly. Two patients developed transient incomplete L5 deficit. Bony bridges around the cases and posterolaterally were identified on the six-month x-rays. Spine view confirmed the quality of the fusion and lumbopelvic parameters revealed restoration of segmentary lordosis.

Conclusion: The unilateral approach for TLIF is a reliable technique which does not compromise the roots. It enables very reliable primary stability and recovery of local segmentary lordosis. We are developing a minimally invasive percutaneous technique for this procedure.

Purpose: Adding bone marrow cells to ceramic materials provides an osteoprogenic capacity favouring bony regrowth. Likewise, addition of platelets, which contain growth factors, might increase the rate of bone formation. The purpose of this work was to quantify in vitro the osteogenic potential obtained by adding platelet-rich plasma (PRP) to the bone marrow culture on ceramic materials.

Material and methods: PRP was obtained by centrifugation of blood and added to bone marrow cells harvested from the iliac crest and cultured on biphasic macroporous ceramic materials. Addition of PRP was repeated with platelet counts every two days. Differentiation of bone marrow cell into cells with osteogenic potential was evaluated by quantifying alkaline phosphatase activity after 15 days culture.

Results: Proliferation of mesenchymatous cells was clearly enhanced in cultures with PRP (+31%). Mean prevalence of phasphatase-alkaline-positive colonies was also improved after addition of PRP (+38%). Similarly, alkaline phosphatase activity was higher after addition of PRP (+31%).

Discussion: Adjunction of PRP to bone marrow cells cultured on ceramic materials stimulates proliferation of osteoblast-like cells. Increased cell proliferation and differentiation observed in vitro provides quantitative elements favouring the combination of platelets with bone grafts using bone substitutes.

Purpose: We studied a consecutive multicentric series retrospectively to assess outcome after one-stage revision total hip arthroplasty without cement for infection-related failed total hip arthroplasty in patients with chronic infection (> 30 days).

Material and method: Thirty-four patients (12 women and 22 men) underwent primary arthroplasty between 1992 and 1998. Mean age was 67.8 years (range 45 – 89). Indications were primary hip degeneration (n=17), neck fracture (n=5), secondary hip degeneration (n=12). Twenty-two patients had risk factors for infection. The same treatment protocol was used for all patients: single-phase replacement of the infected implant with insertion of a hydroxyapatite coated implant without cement and a three-month antibiotic regimen adapted to intraoperative samples. Seven femorotomies were performed. Clinical, laboratory and radiological findings were used to assess outcome. Patients were considered to be cured when laboratory tests were normal and x-rays showed no signs of suspected complications.

Results: There were three failures and 31 successful revisions. Intraoperative fracture of the femur required osteosynthesis in six patients. five patients developed a drug-related complication. Isolated germs were: meticillin-sensitive staphylococcus (n=18), meticillin-resistant staphylococcus (n=6), multiple germs (n=2), other germs (n=6), undetermined (n=2). The three failures occurred in patients with meticillin-resistant staphylococcal infections.

Discussion and conclusion: Our 88% success rate is in line with data in the literature. For us, one-phase revision is the best first intention option: other techniques may be required for patients with severe infections, after failure of one-stage revision, or massive loss of bone stock. Based on the present results, we have continued to prefer this therapeutic option but long-term follow-up is needed to further support these early findings.

Purpose: Healing may be problematic after lag screw osteosynthesis of pertrochanteric fractures in osteoporotic bone due to the greater risk of defective fixation. Acrylic cement has been proposed to reinforce the fixation of the lag screw in these patients, principally to avoid the risk of cutout, but the acrylic cement’s thermal toxicity, its poor biocompatibility, and the difficult manipulation are important drawbacks. Cortoss® is a new composite biomaterial composed of bisphenol-aglycidly (bis-GMA) which could be an attractive alternative to classical cements. Cortoss is an injectable material with mechanical properties similar to human cortical bone. The purpose of this clinical study was to describe the new method for injection the material and to assess the anchoring force and safety of Cortoss in osteoporotic patients undergoing surgical fixation of pertrochanteric fractures.

Material and methods: This prospective study was approved by the local ethics committee. Twenty consecutive patients aged 70 years or more with pertro-chanteric fracture were included. The lag screw was inserted under fluoroscopic guidance, and the maximum insertion torque was measured. The screw was then unscrewed seven turns (length of the threaded head), and 2.5 cm Cortoss was injected via a polyimide catheter measuring 2.5 mm in diameter. The screw was then rescrewed in place to a troque 30% above that measured without Cortoss

Results: Eighteen women and two men, age 70 – 96 years, gave their informed consent to participate in the study. Mean maximal insertion torque without Cortoss was 1.23 Nm (min 0, max 4.8 Nm), which was increased 30% with Cortoss. Posto-operative x-rays showed that the screw head was embedded in an envelope of Cortoss and that the Cortoss had diffused into the adjacent bone. There were no adverse effects.

Discussion and conclusion: Cortoss provided effective reinforcement of the fixation as demonstrated by the higher insertion torque. Cortoss can also increase the mechanical resistance of the screw-cancellous bone interface, limit fixation default, and reduce secondary displacement of the lag screw in patients with osteoporotic bone. Improved stability can also reduce immobilisation time and facilitate bone healing without increasing the risk of secondary displacement and subsequent morbidity. Due to its safety, easy use, and efficacy, Cortoss provides a better alternative than acrylic cement for reinforcement of lag screw fixation of porotic bone.

Purpose: Combining biomaterials with non-cemented cups is a new approach to acetabular construction in patients with bone stock loss after failure of conventional techniques. We evaluated our early results and attempted to ascertain limitations.

Material and methods: Between January 1, 1996 and December 31, 2000, we changed 229 cups. The reconstruction combined biomaterials with a non-cemented cup for 48 patients (mean age 57 years, age range 29–84). Mean retrospective follow-up was 37 months (7–67) with two patients lost to follow-up early.

Two types of cups coated with hydroxyapatite were used depending on the acetabular potential for retaining the implant: 26 ATLAS press-fit cups (four screwed) and 22 Cerafit cups with Surfix anchor screws. Bony defects were filled with grains of macroporous calcium phosphate ceramic, alone or in combination with an autol-ogous bone graft (five patients) and/or an iliac bone marrow graft (24 patients). The Harris and modified PMA clinical scores were used for assessment. According to the Antonio classification, preoperative bone loss was grade II in 18, grade III in 27, grade IV in 3. We attempted to determine the percent of bone support under the cup before reconstruction. The centre of the prosthesis rotation and the interfaces with the biomate-rials were checked regularly.

Results: At short-term, this technique provided a clear functional improvement (Harris improved from 53.7 to 81.3 points). The bone-biomaterial interface did not show any lucent lines and tended to become homogeneous (31 cases). Seven patients (15.2%) developed millimetric lucent lines around stable implants. We had nine failures (19.6%) and performed four surgical revisions (8.7%) for major inclination of the cup in three and recurrent dislocation in one. There were also five patients with an asymptomatic implant migration. Excepting one case, the cup inclinations occurred when the bony support was less than 50% of the acetabulum while only one migration was noted with less than 50% bony support (p=0.02).

Discussion: At short-term this technique, which is easy to perform and less costly and safer than allografting, provides similar results. We observed a continuous construct between the receiver bone and the macroporous biomaterial in all cases. Mechanically, the cup was stable when the bony support was healthy and involved more than 50% of the acetabulum. Other reconstruction strategies should be considered in other cases.

Purpose: We report a retrospective monocentric evaluation of surgical lavage for early infection(≤ 30 days) after total hip arthroplasty.

Material and methods: Thirty-four patients, mean age 67.8± 12.1 years) who developed grade II infection in the Gustillo and Tsukayama classification were treated between 1992 and 1995 by surgical debridement and adapted antibiotherapy for a limited duration. Arthroplasty was indicated for primary hip degeneration (n=17), secondary hip degeneration (n=11), or neck fracture (n=6). There were 25 first-intention arthroplasties and nine revision arthroplasties. Seventeen patients had risk factors for infection.

All patient were followed for at least four years. Therapeutic efficacy was assessed on the basis of clinical and laboratory findings and radiographic results. We searched for biological or bacteriological factors predictive of good outcome. The chi-square test and Fisher exact tet were used to compare population distributions and Student’s t test and Mann Whitney test to compare means.

Results: Mean follow-up was five year, 74% of the patients developed recurrent infection. A second treatment protocol was implemented for only 56% of these patients and was effective against the infection and satisfactory for functional outcome (28% of the patients underwent revision for extraction of the prosthesis and 16% were given continuous antibiotics). There was a significant difference in the efficacy of the surgical lavage for delay ≤ 21 days (p = 0.02). Statistical analysis suggested the type of germ (p = 0.006), and presence of risk factors (p= 0.0052) had a significant influence.

Discussion: Attractive because minimally invasive, surgical lavage appears to limit acute symptoms without eradicating infection. Furthermore, if the failure is recognised late after infection has become chronic, the efficacy of revision is compromised and may lead to poor mid-term function. Like other authors, we identified delay to treatment as the one significant factor. These poor results have incited us to limit indications for surgical lavage to cases of infection diagnosed very early, programming secondary revision for very debilitated patients. Close and prolonged follow-up is indispensable.

Purpose: Indications for anterior release associated with posterior release for stiff idiopathic scoliosis in adults has varied from institution to institution. The traction view is taken as a useful tool to determine whether anterior release is necessary. The purpose of this study was to validate this hypothesis in a homogeneous group of patients with specifically defined idiopathic scoliosis and to determine predictive value of the traction view. Based on this prediction, we then compared postoperative frontal correction in patients with and without anterior release, performed thoracoscopically.

Material and methods: All patients had idiopathic thoracic scoliosis with a Cobb angle greater than 60° and less than 35% reduction on the standard traction view. Cotrel-Dubousset instrumentation was used for release/posterior fusion procedures. A posterior approach was used alone in group 1 patients (n = 46). Group 2 patients (n = 10) underwent thoracoscopic first intention anterior release/fusion.

Results: The postoperative Cobb angle was strongly correlated with the preoperative angle on the traction view (r = 0.86, p < 0.001). The traction view predicted the postoperative Cobb angle actually achieved rather than the degree of correction of the Cobb angle. The severity of the curvature, 81.5° and 83.3° in groups 1 and 2 respectively, and reducibility on the traction view, 61.6° and 62.1° in groups 1 and 2 respectively, were not different. Likewise the postoperative angles were not significantly different between the two groups (47.4° and 45.4° respectively).

Discussion: The traction views were found to predict reduction of the thoracic curvature even for stiff scoliosis but with a wide error. The two groups were comparable since there was no difference in the mean degree of reducibility under traction. Consequently, anterior release did not appear to improve the postoperative correction in the frontal plane.

Conclusion: The traction view does not appear to be sufficiently discriminating to determine the usefulness of anterior release associated with posterior release for the treatment of stiff idiopathic thoracic scoliosis.

We have evaluated the use of a synthetic porous ceramic (Triosite) as a substitute for bone graft in posterior spinal fusion for idiopathic scoliosis. In a prospective, randomised study 341 patients at five hospitals in the UK and France were randomly allocated either to autograft from the iliac crest or rib segments (171) or to receive Triosite blocks (170). All patients were assessed after operation and at 3, 6, 12 and 18 months.

The two groups were similar with regard to all demographic and baseline variables, but the 184 treated in France (54%) had Cotrel-Dubouset instrumentation and the 157 treated in the UK usually had Harrington-Luque implants. In the Triosite group the average Cobb angle of the upper curve was 56°, corrected to 24° (57%). At 18 months, the average was 26° (3% loss). In the autograft group the average preoperative upper curve of 53° was corrected to 21° (60%). At 18 months the mean curve was 25° (8% loss). Pain levels after operation were similar in the two groups, being mild in most cases. In the Triosite group only three patients had problems of wound healing, but in the autograft group, 14 patients had delayed healing, infection or haematoma in the spinal wound. In addition, 15 autograft patients had pain at the donor site at three months. Seven had infections, two had haematoma and four had delayed healing.

The haematological and serum biochemistry results showed no abnormal trends and no significant differences between the groups. There were no adverse events related to the graft material and no evidence of allergenicity.

Our results suggest that Triosite synthetic porous ceramic is a safe and effective substitute for autograft in these patients. Histological findings on biopsy indicate that Triosite provides a favourable scaffolding for the formation of new bone and is gradually incorporated into the fusion mass.

We performed biopsies during reoperation for minor complications in two active young patients 9 and 19 months after massive bone allograft implantation for bone tumour. The grafts were dead and resorption-apposition activity, when present, was predominantly in subperiosteal areas. Inflammatory infiltration was very seldom found.

Features considered as ‘microfractures’ or ‘microcracks’ were noted in the cortical ring together with the formation of woven bone, in areas with remodelling. Such cracks are likely to be of mechanical origin and do not inevitably lead to complications.