Rotational positioning of the femoral component during the realisation of a total knee arthroplasty is an important part of the surgical technique and remains a topic of discussion in the literature. The challenge of this positioning is important because it determines the anatomical result and its effect on the flexion gap and clinical outcome mainly through its impact on patellofemoral alignment. The intraoperative identification of the axis transepicondylar visually or by navigation is not reliable or reproducible. The empirical setting to 3 ° of external rotation, the procedure used to cut or dependent or independent is not adapted to the individual variability of knee surgery. Indeed, the angle formed by the posterior condylar axis and trans-epicondylar axis is subject to large individual variations.

The authors propose a novel technique, using the navigation of the trochlea to determine the rotation of the femoral component. The principle is to consider the rotation of the femoral implant as “ideal” when it makes a perfect superposition of the prosthetic trochlea with the native bony trochlea on patellofemoral view at 60° when planning the femur. The bottom of the prosthetic trochlea is well aligned with the trochlea groove, identified during the trochlear morphing, itself perpendicular to the trans-epicondylar axis. The authors hope to encourage centering patellofemoral joint prosthesis, thus favoring the original kinematics of the extensor apparatus.

The purpose of this study is to demonstrate firstly, that the navigation of the trochlea is a reliable and reproducible method to adjust the rotation of the femoral component relative to the trans-epicondylar axis taken as reference and the other, the rotation control by this method is not done at the expense of the balance gap in flexion.

It is a bi-centric study prospective, nonrandomised, including continuously recruited 145 patients in two French centers. All patients were included in the year 2010 and have all been revised three months and one year of surgery. The average age of patients was 71 years [53, 88]. It was made no selection of patients who have all been included consecutively in the study and in the two centres. In all cases, the rotation of the femoral component was determined by intraoperative navigation of the trochlea. The authors compared the alpha angle (angular divergence between the plane and the posterior bicondylar plane and trans-epicondylar axis) obtained by this method and that calculated on a pre-or postoperative scan. The authors also measured the space between femur and tibia internal and external side in flexion (90°) to assess the impact on the balance in flexion.

There is excellent agreement between the results obtained by the method of CT scan and the trochlear navigation technique. In addition, this technique allows us to achieve a quadrilateral space gap in flexion. The authors found large individual variation in the distal femoral epiphyseal torsion (angle alpha). They demonstrate that the navigation of the trochlea is a reliable and reproducible method to adjust the rotation of the femoral component relative to the trans-epicondylar axis taken as reference and provides, concomitantly, a quadrilateral space gap in flexion.

Rotational positioning of the femoral component during the realisation of a total knee arthroplasty is an important part of the surgical technique and remains a topic of discussion in the literature. The challenge of this positioning is important because it determines the anatomical result and its effect on the flexion gap and clinical outcome mainly through its impact on patellofemoral alignment. The intraoperative identification of the axis transepicondylar visually or by navigation is not reliable or reproducible. The empirical setting to 3 ° of external rotation, the procedure used to cut or dependent or independent is not adapted to the individual variability of knee surgery. Indeed, the angle formed by the posterior condylar axis and trans-epicondylar axis is subject to large individual variations.

The authors propose a novel technique, using the navigation of the trochlea to determine the rotation of the femoral component. The principle is to consider the rotation of the femoral implant as “ideal” when it makes a perfect superposition of the prosthetic trochlea with the native bony trochlea on patellofemoral view at 60 ° when planning the femur. The bottom of the prosthetic trochlea is well aligned with the trochlea groove, identified during the trochlear morphing, itself perpendicular to the trans-epicondylar axis. The authors hope to encourage centering patellofemoral joint prosthesis, thus favouring the original kinematics of the extensor apparatus.

The purpose of this study is to demonstrate firstly, that the navigation of the trochlea is a reliable and reproducible method to adjust the rotation of the femoral component relative to the trans-epicondylar axis taken as reference and the other, the rotation control by this method is not done at the expense of the balance gap in flexion.

It is a bi-centric study prospective, nonrandomised, including continuously recruited 145 patients in two French centres. All patients were included in the year 2010 and have all been revised three months and one year of surgery. The average age of patients was 71 years [53, 88]. It was made no selection of patients who have all been included consecutively in the study and in the two centres. In all cases, the rotation of the femoral component was determined by intraoperative navigation of the trochlea. The authors compared the alpha angle (angular divergence between the plane and the posterior bicondylar plane and trans-epicondylar axis) obtained by this method and that calculated on a pre-or postoperative scan. The authors also measured the space between femur and tibia internal and external side in flexion (90°) to assess the impact on the balance in flexion.

There is excellent agreement between the results obtained by the method of CT scan and the trochlear navigation technique. In addition, this technique allows to achieve a quadrilateral space gap in flexion. The authors found large individual variation in the distal femoral epiphyseal torsion (angle alpha). They demonstrate that the navigation of the trochlea is a reliable and reproducible method to adjust the rotation of the femoral component relative to the trans-epicondylar axis taken as reference and provides, concomitantly, a quadrilateral space gap in flexion.

Aim

This controlled study uses gait analysis to evaluate patients' pre- and post-ankle arthroplasty, post-ankle arthrodesis and compares the results with a healthy control group to assess whether these theoretical benefits are borne out in clinical practice.

Purpose of the study: The purpose of this prospective controlled trial was to evaluate the influence of the tibial stem design on the primary fixation of the tibial base of a total ankle arthroplasty (TAA) and on radiographic remodelling of adjacent bone.

Material and methods: Sixty Salto™ TAA with a short stem were studied with one to three years follow-up. This series was mated with a control series of 60 TAA using conventional ‘cannon’ implants. All patients were assessed clinically with the AOFAS score and radiographically on the loaded ankle.

Results: The two consecutive series were reviewed a mean 23 months. The mean AOFAS score at review was 83.6 points (mean gain 45 points). The functional analysis did not disclose any difference between the two series. There was no evidence of tibial piece migration. The incidence of lucent lines, all partial (32.5%) and bone remodelling was higher in the cannon TAA series than in our short-stem TAA series, but the difference did not reach statistical significance. Considering all prostheses together, the incidence of bone condensation was 39.3%; and bone defects 16%. Young age appeared to be the only factor correlated with the development of defects (p=0.01). One progressive defect was grafted. Two prostheses were removed (one talar necrosis and one malposition).

Discussion: Primary fixation of the tibial base of the Salto™ prosthesis is excellent, irrespective of the design of the anchor stem. Suppression of the cannon did not reduce statistically the phenomena of osteolysis. The mechanism of the periprosthetic osteolysis is probably plurifactorial, associating biochemical and mechanical factors.

Introduction: The goal of this study is to relate our experience about a third generation model of a Total Ankle Arthroplasty (TAA) Salto^® (Tornier™) by evaluating a prospective, homogenous and continuous serie of 130 TAA.

Materials and Methods: From 1997 to 2002, 130 TAA had been implanted to 125 patients, mean age 57 year (19 to 84). This procedure was performed by two experimented surgeons. Indications for arthoplasty were post-traumatic osteoarthritis in 57%, osteoarthritis in chronic laxity in 15%, rheumatoïd arthritis in 12%, primitive osteoarthritis in 10%, and 6% other. Collection of preop, postop and follow up datas was prospective. Radiographs were numerised and treated by a specific software. We considered as a failure when the implant was removed.

Results: After a mean follow up period of 44 months [12 – 108], 10 patients were lost to follow up, without complication, 4 patients died, without relationship to the TAA, but they had sufficient follow up, 9 arthroplasties were converted to arthrodesis, leaving a 92,5% success rate. Postoperative main complications were cutaneous problems (18/130). In one case, a skin necrosis led to a secondary infection that requested implant removal.

Long folllow up showed that main complications were pain and bone cysts, which required arthrodesis in 8 cases. Survivorship analysis at 108 months were 83% [IC5%: 72–95]. At follow up, clinical AOFAS ankle score was significantly raised (31% preoperative to 84% at follow up). A SF36 quality of life score was available for 85 patients. Physical score was 60, mental score 66 and total score was 64.

Discussion: This study has the avantages of being prospective and continuous. Clinical results and failure rate were encouraging. Infection rate less than 1% may be in relation with the low rate of patient with rheumatoïd arthritis. No significant difference of the result could be find according to the initial indication.

Conclusion: These mid-term results are concordant with orther series of third generation TAA. We remain concerned because of bone trabeculation modification and pain without obvious anatomical abnormality: long term follow up is necessary. Mid term results confirm TAA as a therapeutic option for ankle pathology.

Purpose of the study: The aim of this study was to analyze the mechanical function of the ankle after implantation of a total ankle arthroplasty. Gait analysis included kinematic and dynamic parameters of the lower limbs before and after prosthesis implantation in comparison with ankle fusion.

Material and methods: This prospective non-randomized study included three cohorts of patients. The first cohort included 12 patients presenting osteoarthritic lesions of the ankle requiring total ankle arthroplasty; these 12 patients were reviewed six months postoperatively and for six of them twelve months postoperatively. The second cohort was composed of 12 patients reviewed twelve months after tibiotalar arthrodesis. The third cohort was composed of 12 healthy volunteers who participated in the same study protocol. The gait analysis was conducted with the Motion Analysis optoelectronic system. Parameters recorded were: self-selected speed, fastest speed, stride rate, step length, stride symmetry and length, symmetry of floor contact, and symmetry of toe lift-off. In addition, patients participated in specific tests to step over an obstacle and go up and down stairs.

Results: Gail was slower an asymmetrical in patients with ankle fusion. Fusion enabled good recovery of gait speed but at the cost of imbalanced weight-bearing times and asymmetrical toe lift-off. Total ankle arthrodesis provided little improvement in gait speed but enabled progressive and persistent recovery of symmetrical gait.

Discussion: It is well established that an ankle prosthesis improved joint force and motion in comparison with the osteoarthritic ankle. Fusion provides good clinical and kinetic results but at the const of compensation by the joints above and below the ankle. Few studies have examined gait symmetry which in our opinion would be a good criterion for evaluating the quality of gait.

Conclusion: The raw data obtained in this study demonstrate that patients with an osteoarthritic or fused ankle can recover gait speed but that they retain a limp. Total ankle arthroplasty enables a more balanced fluid and symmetrical gait which is much more comfortable for the patient.

Introduction: The emergence of multidrug resistant Gram negative bacilli susceptible to hardly any beta lactam compound has led to infections close to a therapeutic dead end. In such circumstances, Imipenem-cilastatin (I-C) is often the only remaining therapeutic option. We report our experience with the prolonged administration of high-doses of I-C in the treatment of osteoarticular infections with bacteria resistant to other beta-lactam agents (or 4^l generation cephalosporins in 14 cases).

Materials and methods: Our retrospective study over 7 years included 29 patients with septic arthritis (n=3) continuous osteitis (n=6), septic non-union(n=12) and prosthetic joint infections (n=8). Treatment included an extensive surgical debridement and post-operative combination antibiotherapy with intravenous I-C and aminoside (54%) and/or fluoroquinolones (46%) and/or fosfomycin (29%). Associated microorganisms requiring yet additional antimicrobial agents were associated in 17 (59%) cases. I-C was maintained for an average of 46 days (extremes 21–90), at an average dose of 3,8g/day (extremes 2–6). The bacteria warranting I-C were cephalosporinase hyperproducing Enterobacter cloacae (38%), extended spectrum beta-lactamases producing enterobacteria (31%), Pseudomonas aeruginosa (21%) and/or Acinetobacter baumanii (21%).

Results: Early outcome was favorable in 24 patients (82%). Two patients relapsed with the bacteria requiring I-C. Three failed to negate succion fluid cultures : one was discharged with no change in his condition, one agreed to a leg amputation and the third died of candidemic septic shock in SICU with drainage fluid still bactériologie ally positive. Repeated secondary colonization and systemic infection with yeasts led to a monitoring of yeast load. Per os amphotencin B and immediate treatment of urinary colonization prevented further systemic dissemination of candical infections. No other tolerance incidents were noted. Acquired resistance occurred only once in a P. aeruginosa isolate while Imipenem-cilastatin was chosen to cover an ESBL producing Escherichia coli. Secondary treatment with ceftazidime was then successful in eradicating P. aeruginosa.

Conclusion: I-C has been widely used for the treatment of mixed flora infections as a wide spectrum antibiotic.

We report good tolerance of high posology long term administration in documented osteoarticular indications if yeast colonization is properly monitored, and eradication rates are comparable to those reported in infections with susceptible bacteria.

Purpose: Prosthetic replacement of the ankle joint is a controversial issue. Minimally invasive noncemented third-generation implants with a third component have enabled improved clinical results and prolonged mid-term implant survival. These results authorise prosthetic implantation as an alternative to arthrodesis in selected patients. New developments in prosthesis concepts and design, aimed at overcoming the insufficient results of earlier implants, require early validation with a prospective clinical and radiological follow-up. The objective of this work was to present the mid-term results with the SALTO prosthesis implanted in 42 patients.

Material and methods: Forty-two SALTO prostheses implanted between February 1997 and December 2000 were followed prospectively for two to six years. The implant design, which mimics the anatomic asymmetry of the talar dome, uses a mobile polyethylene insert and optional fibular resurfacing for optimal primary and long-term stabilisation. Data were collected prospectively using a computer database which provides the AOFAS score. Metrological analysis of the digitalized x-rays (AP, lateral and stress) were used to study the precision of the insertion, implant stability, and prosthesis kinematics. Posttraumatic osteoarthritis predominated (n=29) in this series. Mean age was 54 years (30–79).

Results: None of the patients were lost to follow-up. Three patients had a revision procedure for arthrodesis (persistent pain in two and sepsis in one). The clinical score was excellent or good in 88% of patients. The mean clinical score was 20.5 points preoperatively and 70 points at last follow-up. The radiological analysis demonstrated the precision of the insertion technique and did not disclose any evidence of significant implant mobilisation with time. There were no failures related to the lateral malleolar implant (n=12). Radiographically, mean ankle movement was 15.2–23°. Changes in the periprosthetic bone were noted.

Discussion: The quality of the results with the prosthesis under study and the short- and mid-term stability enable envisaging implantation when the local or regional anatomic conditions suggest arthrodesis would produce unsatisfactory results. The improvement in function (exceptionally total recovery) remains difficult to predict. Further follow-up of these patients is needed.

Purpose: We reviewed retrospectively 349 cases of infected total hip arthroplasty treated by prosthesis replacement. The surgical strategy, 127 single-stage procedures and 222 two-stage procedures, was determined by the surgeon on a case by case basis.

Material and methods: At least one positive sample during the clinical history was required for inclusion in the series. Results of all bacteriological samples collected pre- and intra-operatively were noted. Samples were considered reliable if obtained from a deep site (puncture, biopsy, intraoperative specimen) and non-reliable if obtained from any other site. We studied the agreement between preoperative and intraoperative samples, taking the intraoperative samples as the reference, in order to determine the effect of complete preoperative knowledge of the causal germ on the outcome of infection treatment at last follow-up.

Results: For single-stage replacement procedures, preoperative samples were reliable in 74 cases (58%) and non reliable in seven (6%); they were sterile or absent in 46 cases (36%). Intra-operative samples were positive in 103 cases (81%). Agreement between the preoperative and intraoperative samples was observed in 48 cases (38%). The rate of success was not different if the surgeon had or did not have reliable knowledge of the causal germ(s) preoperatively: successful treatment in 66 cases (89%) with knowledge and successful in 46 cases (87%) without knowledge. For two-stage procedures, preoperative samples were reliable in 155 cases (70%) and non-reliable in 15 (7%); they were sterile or absent in 52 cases (23%). Intraoperative samples were positive in 178 cases (80%). Agreement between preoperative and intraoperative samples was observed in 107 cases (48%). The rate of success was not different if the surgeon had or did not have reliable knowledge of the causal germ(s) preoperatively: successful treatment in 133 cases (86%) with knowledge and successful treatment in 56 cases (84%) without knowledge.

Conclusion: Reliable preoperative knowledge of the causal germ(s) did not affect the rate of success for single-stage or two-stage total hip arthroplasty replacement procedures. These findings do not corroborate the notion that it is absolutely necessary to recognise the germ(s) causing the infection before undertaking a single-stage replacement procedure.

Purpose: The purpose of this work was to evaluate the mid-term clinical and radiological outcome of a homogeneous and continuous series of third-generation total ankle prostheses (resurfacing, cylindric, noncemented, triple-compartment).

Material and methods: From March 1990 to June 1996, 26 patients aged 57 years (32–73) were treated with a New Jersey LCS (n=5) or Buechel-Pappas (n=21) prosthesis. Most of the patients (n=21) had a posttraumatic ankle. Preoperatively, mean ankle motion was 17°. The AOFAS score was used for the clinical assessment. The position of the prosthesis and its stability over time were assessed on the x-rays together with the insert, the bone-prosthesis interface, the bone structure and periprosthetic ossifications.

Results: Intra and postoperative complications were malleolar fracture (n=5), haematoma (n=1), late wound healing (n=1), insert instability (n=2), and medial malleolar conflict (n=1) requiring reoperation with preservation of the implant. At mean follow-up of seven years, two patients had died and one was lost to follow-up. Three had an arthrodesis: failed fixation at two years, secondary talar mobilisation at seven years, and secondary infection at eight years. For the other twenty patients: the AOFAS score was poor for two patients (including one with patent wear), fair in two, and good in 16. Mean joint motion at last follow-up was 24°. Radiographically, there were no significant changes in the position of the tibial and talar elements. Anchorage of the tibial element was fibrous in half of the patients and ossesous in the other half. A macro defect aspect was observed below the talar element in four patients. There were active periprosthetic ossifications in the majority of the patients.

Discussion: The insufficient ancillary for this prosthesis and its old concept explain the frequency of malleolar fractures and the level of the functional outcomes. Improved prostheses should be used. The stability of the results observed in this series of patients followed up to 12 years is an argument favouring indicating prosthesis insertion as an alternative to arthrodesis, particularly since revision of an arthrodesis is not particularly difficult, even with an iliac graft. The radiographic evidence of periprosthetic ossifications or bone resorption, particularly under the implant, emphasises the need for prolonged surveillance.

Purpose: We report results of management of infected total knee arthroplasty (TKA). Our aim was to analyse the different therapeutic options and identify factors predictive of cure.

Material and methods: This retrospective multicentric analysis included a consecutive series of 179 cases of infected TKA. Revision TKA was performed for 77 knees in two operative times, 30 in one operative time. Synovecetomy was performed for 26 knees, arthrodesis for 36 and amputation for nine. Minimum follow-up was two years. For each case, we assessed cure of infection and functional outcome. Non-parametric statistical tests were used to compare outcome.

Results: Mean follow-up was 41.2 months. There was a 17% death rate during the first two years after management of infected TKA. Cure was achieved in two-thirds of the revised cases (in one or two operative times) and in 54% of the synovectomy cases. Arthrodesis yielded cure in 50%. Staphylococci was identified in 65%. The functional outcomes of revised prostheses (two times) was less satisfactory if the anterior tibial tubercle was removed, if the extensor system was involved, or if a flap cover was necessary (p< 0.05). There was no significant difference in functional outcome for one-time and two-time operations. Cure was achieved in 95% of the synovectomies performed before 16 days. Relapse occurred in 95% of those performed after 56 days. Arthrodesis performed in patients with major bone loss failed. Among the arthrodesis failures, 50% were related to mechanical failure and the other 50% to recurrent infection.

Discussion: We discuss these results and indications for different treatments of infected total knee arthroplasty. For each therapeutic option, we analysed factors allowing hope for cure and good functional outcome.

Conclusion: The objective of treatment of an infected TKA is to achieve cure and maintain satisfactory function, often a difficult goal to reach. Major mutilating surgery can be avoided by early adapted care.

Purpose: We analysed prospectively our experience with the zincrone-polyethylene bearing for total hip arthroplasty for implantations performed between 1987 and 1997. This bearing is designed to reduce in vivo wear due to the excellent biological qualities of this ceramic.

Material and methods: Total hip arthroplasty was performed on 867 hips over the ten year period (1987–1997) in patients followed for mean seven years. All had titanium or stainless steel stems with a cemented or press fit cup (inner diameter 26-mm or 28-mm). We used Y-TZP zircone, a polycrystalline tetragonal-phase zircone stabilised by adjunction of ytterium (Prozyr). Mean patient age was 58 years (17–87). Mean body weight was 70 kg (40–125) and mean height was 167 cm (140–196). The sex-ratio was 1.3.

Results: Overall survival determined with the Kaplan-Meier method was 60% at 13 years (95%CI=55–65%). Revisions (all causes including infection) were performed on 118 hips. To our knowledge, for the entire cohort, one patient presented a zircone head fracture. Radiographic analysis showed a particular type of fracture related to isolated acetabular loosening. Comparative analysis of this type of loosening for 785 heads (26-mm) and 92 heads (28-mm) was not significant (log rank test). Taking surgical revision as the endpoint for isolated acetabular failure, implant survival at 13.5 years was 74% (95%CI=68–80), confirming the particular mode of failure with this bearing.

Discussion: This prospective analysis showed the absence of superiority of the zircone head in vivo for reduction of wear in comparison with metal heads in young subjects. The purely epidemiological data reveal formal evidence of the deleterious nature of zircone in contradiction with the expected beneficial tribiological effect. We have no physicochemical explanation for these failures. The pathology findings were non-specific, but the failure was real.

Conclusion: Considering these findings, we have abandoned use of this ceramic since 1997.

We report the results of a prospective study of 140 consecutive cases of acetabular revision using large frozen femoral head allografts and cemented all polyethylene acetabular components. The mean follow-up time was 10 years (5 Ð 16).

Thirty patients died, seven were lost to follow-up and 26 had failed and undergone further surgery. Nineteen failures were due to aseptic failure and collapse of the graft. Kaplan-Meier survival analysis calculated a mean survival at 10 years of 88.5% for revision for any reason.

We compare all reported techniques of acetabular reconstruction for similar defects and recommend a surgical strategy based on the available evidence, but weighted towards a preference to reconstitute bone stock rather than removing further bone in the revision situation.

Aims: To review the early results of shoulder arthroplasty in the weight bearing shoulder of long term pareplegic patients. We have been unable to find previously published results of this subgroup of shoulder arthroplasty patients in the litterature. Method: The case notes and X-rays of 5 female patients who had undergone shoulder arthroplasty were analysed. These patients all had longstanding paraplegia and were wheelchair bound. All patients has been prospectively scored with the American Shoulder and Elbow Surgeon’s (ASES) function score and the Constant score. Results: The patients had a mean age of 70 yrs (61–88) at the time of surgery. Three of the 5 shoulders and rotator cuff tears at the time of surgery, 2 of which were repaired. The results are summarised in the table. There was no radiological evidence of failure apart from the migration of the single cemented glenoid component. Conclusion: In view of the satisfactory improvement in pain in particular following these procedures, we feel that it is reasonable to continue to offer this procedure in this subgroup of patients. We will however remain vigilant with regard to any further complications arising because of the increased loading in these prostheses at the medium and long term.

Purpose: Fractures of the anterior acetabular wall with preservation of the pelvic inlet are rare. These lesions were not noted or classiþed by Judet and Letournel in their classiþcation system Ð Ç In fractures of the anterior wall, the anterior part of the articular horseshoe breaks off with a major portion of the middle segment of the anterior column È. The ilio-inguinal approach was recommended for the surgical treatment of these fractures. Method : We have encountered two cases involving purely the anterior wall with preservation of the pelvic inlet, rather than the anterior wall fracture described by Judet and Letournel. We have identiþed only two other cases in the international literature. The recognition that these fractures were not as that described by Judet and Letournel was essential, as an alternative surgical approach was necessary for reconstruction. The ilioinguinal approach of Judet and Letournel is the technique of choice in anterior fractures, but provides only very limited potential for intra-articular manipulation through the line of fracture. The Smith-Petersen approach was practically the only feasible choice as intra-articular exposure is difþcult or impossible to obtain with the classical approaches used for anterior acetabular fractures. Results: Both of our cases were reconstructed via the Smith-Petersen approach. Postoperative review at 18 months demonstrated painless, fully mobile hips with evidence of radiological union of the fractures. Conclusion: We propose designating as ñfracture of the anterior columnñ all the anterior fractures described by Judet and Letournel (column and wall) that justify the ilioinguinal approach. We would reserve the term ñanterior wallñ for large anterior acetabular rim fractures, with the extension exclusively lateral to and below the pelvic brim. We recommend the use of a Smith-Petersen approach for reconstruction once this fracture pattern is identiþed.

Purpose: We have used monoblock cryopreserved femoral heads for acetabular reconstruction without supporting material since 1985 for cemented total hip arthroplasty in patients with major acetabular defects. From 1985 to 1995, 140 reconstructions were performed. We present a prospective analysis of the clinical and radiological outcome at 10 years follow-up.

Material and methods: Mean age of the population was 61 years. Most of the patients had had two prior interventions. According to the SOFCOT classification, the 140 defects were: grade II 50%, grade III 35% and grade IV 15%.

The cryopreserved graft (femoral head bone bank) was used to reconstruct the acetabular defect. The graft was adapted to the size of the defect to fashion a congruent construct aimed at achieving primary stability. We did not use any supporting material in addition the primary osteosynthesis with one or two screws. A poly-ethylene cup was cemented in the graft. Most of the cement was applied onto the graft which was reamed to the size of the acetabulum. We retained a theoretical 6-year follow-up for review. All patients were seen for follow-up assessment using the Postel-Merle-d’Aubigné (PMA) clinical score and standard x-rays analysed according to the Oakeshott method. Kaplan-Meier survival curves were plotted taking change in status, revision for clinical failure as the endpoint.

Results: Mean overall follow-up was 8.5 years; it was 10 years for patients with an implant still in situ. Eight patients (5.7%) lost to follow-up were included in the series retained for analysis at mean 5-year follow-up. Thirty-five patients died during the follow-up period (25%). These patients had been followed for a mean four years before their death. Radiologically, cup tilt was not significantly altered over time. Conversely, the centre of the cup, measured from the U line, was not modified in the patients who died or in the group of living patients without revision at last follow-up. It was modified in the group of failure group: mean 28 mm postoperatively in the failure group reaching 39 mm at time of failure (ANOVA < 10-3). We had 26 failures (18%) which occurred at six years (mean); there was a peak at two years and another at nine years. Mean Kaplan-Meier survival was 13.5 years (95CI 12.5–14). The PMA clinical score improved from 3/5/3 (11) pre-operatively to 5.3/5.6/4.3 (15.2) at last follow-up.

Discussion: The overall results at 10 years in this series were globally satisfactory with a success rate above 70%. Failures were related to radiologically demonstrated graft compression with ascension of the centre of the cup measured from the U line. Comparing these results with data in the literature shows an improvement over the Harris series (7-year follow-up in a small group of 48 patients).

Conclusion: This method of acetabular reconstruction reserved for major bony defects has provided a 73% rate of success at ten years.

We evaluated the use of a hemipelvic acetabular transplant in twenty revision hip arthroplasties with massive acetabular bone defects (Paprosky IIIB) at a mean follow-up of 5-years (4–10 years). These defects were initially trimmed to as geometric a shape as possible by the surgeon. The hemipelvic allografts were then cut to a geometric shape to match the acetabular defects and to allow tight stable positioning of the graft between the host ilium ischium and pubis. The graft was further stabilised with screw fixation. A cemented cup (without a reinforcement ring) was entirely supported by the allograft in all procedures.

We report 65% good intermediate-term results.

There were seven failures (five aseptic loosening and two deep infections). Radiographic bone bridging between the graft and host was evident in only one of these cases. Aseptic graft osteolysis began radiographically at a mean of 14 months and revision occurred at a mean of 2 years in the 5 aseptic failure cases. All 5 cases could be reconstructed again due to the restoration of bone stock provided by the hemipelvic graft. One infected case was able to be reconstructed using impaction allografting and the other was converted to a Girdlestone hip.

Thirteen of twenty acetabular reconstructions did not require revision. Radiographic bone bridging between the graft and host was evident in 12 cases. In 2 cases, ace-tabular migration began early (at 5 and 27 months) but stopped (at 35 and 55 months). These 2 cases have been followed for 6 and 9 years respectively, with no further migration. Two dislocations occurred but did not require acetabular revision.

The function of these hips is good with a mean Postel Merle D’Aubigne score of 16.5.

We feel that these are satisfactory intermediate term results for massive acetabular defects too large for reconstruction with other standard techniques.

Purpose: We report our experience with acetabular reconstruction using cyropreserved bone bank hemipevli without a scaffold and total hip arthroplasty for major acetabular defects. Between 1985 and 1999, among 262 acetabular reconstructions requiring massive allografts using cryopre-served bone, 20 cases were performed with hemipelvi.

Material and methods: Mean age of the population was 56 years. The acetabulum had been operated on a mean three times. The 20 defects corresponded to Paprosky grade IIIB or SOFCOT grade IV bone loss. Clinical and radiological review of the 20 hips was made at a mean five years after treatment. None of the patients was lost to follow-up. The overall Postel Merle d’Aubigné (PMA) score at last follow-up was 17 for preoperatively scores at D2, M4 and S3 respectively. The acetabular defects were major and poorly described by the conventional systems. For example, the mean height of the bony defects was about 10 cm measured from the base of the radiographic U and the superolateral rim of the remaining roof.

Results: Globally, 13 patients had not required a reoperation at last follow-up. We had one postoperative death and two early displacements as well as two infections including one haematogenous infection. The Oakeschott criteria were used to analyse the review radiographs. Aseptic lysis of the graft was observed in five cases (generally around the 13th postoperative month) that required revision; a bone graft and a supporting ring were used in all cases because more bone stock was available than for the first revision. Among the 13 cases that did not require a new procedure, there were two with an ascended graft displacing the centre of rotation about 10 mm, followed by radiographic stability. The overall functional score for these 13 hemipelvi at last follow-up was 17 demonstrating the superior functional result compared with arthroplastic resection, the only alternative for such important loss of bone stock. It is not possible to implant a large non-cemented socket in these cases. Radiographic fusion was achieved, documented in 13 cases by the development of bony bridges or disappearance of the interface with oriented lines of force. Early graft resorption does not appear to occur when a metallic scaffold is associated (Garbuz).

Discussion: In all, 19 hips still had their total arthroplasty at last follow-up (one patient with failure preferred trocahntero-iliac coaptation.

Conclusion: Due to the inefficacy of alternative methods, this mode of restoration for major bone loss of the acetabular region (which facilitates secondary revision) appears to provide satisfactory results since the probability of preserving the prosthesis at a mean five years was slightly greater than 3/5. A stronger metallic scaffold may be the solution for the future.