Limb length disparity is a frequent complication after hip surgery inducing many surgeon-patients conflicts. To date no study has been able to precisely quantify such limb length disparity. EOS® system, currently validated to measure lower limb parameters, allows from two bi-dimensional numerical orthogonal radiographies in standing position to obtain a tri-dimensional reconstruction of lower limbs. A computerized system achieves the parameters calculation.

The aim of this study is to precisely measure the limb length disparities and the other hip parameters following total hip arthroplasty surgical procedure, by using a standard X-rays and using EOS® three-dimensional reconstructions.

Twenty-eight patients programmed for total hip arthroplasty have been included (i.e. thirty lower limbs). Two independent performers have carried out twice the measures either on standard X-rays and using three-dimensional reconstructions of the lower limb disparities prior and after the surgical procedure.

The inter and intra-observer reproducibility for the measure of the lower limb disparities have been of the EOS® measures have been respectively of 0.854 and 0.865 and for the standard X-rays of 0.717 and 0.726.

Mean length disparity observed was before Total Hip Arthroplasty of −0.328 cm (0.705; −1.266/0.530) and was of 0.088 mm (1.326; −1.635/0.632) after. We are able to decrease the lower limbs disparity in 69.1% and for the average of 0,416cm.

Using EOS® system has allowed assessing with greater precision the possibility to restore equal lower limb length.

This assessment has permitted introducing a new planning procedure including EOS® imaging associated to the fusion of the prosthetic tri-dimensional image in order to achieve adequate lower limb length.

Introduction

Osteoporosis is a metabolic disease of the bone responsible for a loss of bone resistance and an increase in fracture risk. World Health Organization (WHO) estimations are about 6.3 millions of femoral neck fractures in the world by 2050. These estimations make osteoporosis a real problem in term of public health.

Knowledge in biological tissues mechanical behaviour and its evolution with age are important for the design of diagnosis and therapeutic tools. From the mechanical aspect, bone resistance is dependent on bone density, bone architecture and bone tissue quality. If the importance of bone density and bone architecture has been well explored, the bone tissue quality still remains unstudied because of the lack of biomechanical tools suitable for testing bone at this microscopic dimension.

Therefore the goal of this study is to estimate the osteoporotic cancellous bone tissue mechanical behaviour at its microscopic scale, using an approach coupling mechanical assays and digital reconstruction.

Hard-on-hard bearing surface have been accepted as a valuable alternative for young and active patients needing a hip replacement because these combinations are resistant to wear. Initial development of alumina-on-alumina bearings faced complications such as fractures, and socket loosening. But, with the increasing number of prostheses implanted, noise occurrence appeared as a new complication. The primary aim of the present survey was to quantify the prevalence of having noise in a population receiving alumina-on-alumina hip arthroplasty.

Two hundred and eighty-four ceramic-on-ceramic hips were performed in 238 patients (126 males and 112 females) from January 2003 to December 2004. The average age at the index operation was 52.4 ± 13.4 years (range, 13 to 74 years). We used the same type of prosthesis for all patients manufactured in all cases by Ceraver-Osteal^®. Clearance between femoral and insert was between 20 and 50 microns in order to achieve minimal wear. The survey was conducted by an independent surgeon who did not participated in patients care during the last 6 months of 2007. He interviewed the patients by phone with a standardized questionnaire (appendix) that aimed to assess if noise was present and the characteristics of this noise if present. No suggestion was done on how they could describe the noise and they felt free to use the word that they considered to be the most adapted. Satisfaction was evaluated asking if the patient was very satisfied, satisfied or dissatisfied with its prosthesis.

When the noise was present, the X-ray was independently evaluated to assess if sign of component fracture was present.

Four patients (six hips) died of unrelated cause during the follow-up. Three patients (three hips) lived outside France and could not be followed (1.3%). Nine patients (ten hips) could not be traced and are considered lost to follow-up (3.8%). Two hundred and twenty-two patients with 265 hips were therefore surveyed. Among these 265 hips, 28 experienced noise generation (10.6%). It was defined as a snap for 6 patients, as a cracking sound by 6, as rustling by 6 patients, as a squeaking by 7 patients (2.6%), a tinkling by 2 patients, one patient was unable to define the sound she felt. No factor related to the patient influenced the occurrence of noise. Twelve patients were dissatisfied with the result of the hip prosthesis, 5 of them experienced noise (41.7%); 210 were satisfied or very satisfied 23 of them experienced noise (11%); this difference was significant (p=0.002). No patients required revision for noise.

The origins of noise occurrence are unknown but several hypotheses can be suggested.

Squeaking may be due to absence of sufficient lubrication. Other types of noise can be due to microseparation, occult dislocation, impingement between the femoral neck and the acetabular rim but demonstration remain an issue.

Introduction: While an increasing number of authors have reported on the long-term results of primary alumina total hip arthroplasty (THA) [1], strategies for revising a ceramic-on-ceramic THA are debated in the literature. According to some authors [2], the reimplantation of a ceramic head on a well-fixed femoral stem is inadvisable, as it may lead to a fracture of the newly implanted head. The aim of the present study was to evaluate the incidence of this specific issue, and to report on the clinical and radiological results of the revised hips.

Methods: Between January 1977 and December 2005, 138 consecutive alumina-alumina revision hip arthroplasties were performed in 127 patients. There were 79 women (62.2%) and 48 men (37.8%), with an average age of 67 years (range, 32–91 years). Among these, an isolated acetabular revision was performed in 108 cases. The reason for revision was aseptic loosening of the acetabular component in 98 hips, pain in 7, fracture of an alumina liner in 2, and recurrent dislocation in 1. The revised socket was a cemented alumina in 56 hips, a threaded screw-in titanium with an alumina core in 34, a pressfit titanium with an alumina core in 11, and bulk alumina in 7. Acetabular bone stock losses were classified according to the AAOS system. Most of the hips had a contained type II defect (86%). In all cases, the femoral stem was left in place and the acetabular component alone was revised. At the time of revision surgery, an aluminaalumina combination was implanted in 27 hips, an aluminapolyethylene combination in 56, a metal-poly-ethylene combination in 15, and a zirconia-polyethylene in 10. Overall, a ceramic head was reimplanted on a used femoral taper in 59 cases. Acetabular reconstruction with allografts supported with the Kerboull acetabular reinforcement device was performed in 31% of the hips.

Results: The mean follow-up period was 78 ± 37 months. Thirteen patients (15 hips) died a mean 37 months after surgery. Sixteen patients were lost to follow-up. Postoperatively, five hips had a recurrent dislocation, 2 a deep infection, 9 a trochanteric nonunion (21,3%), 6 a transient nerve palsy. 18 hips required a re-revision surgery, 12 of which for aseptic loosening of the acetabular component. Among the 59 ceramic heads implanted on a well-fixed stem, no fracture of the head occurred at a mean 81 months follow-up. Of the original 108 hips, 77 were available for clinical evaluation and 75 for radiological evaluation at least 2 years after surgery. The mean Merle d’Aubigné score increased from 10.1 ± 2.1 to 16.7 ± 1.1 at the latest follow-up (p< 0.001). Forty-five hips were graded excellent or very good (60%), 26 good (34,6%), 3 fair (4%), and 1 poor (1.3%). When revision for aseptic loosening was considered as a failure, the overall survival rate at 8 years was 96.3 ± 1.8%.

Discussion & Conclusions: In the present study, aseptic loosening of the acetabular component was the main reason for revision surgery. Osteolysis around ceramic implants was moderate and was related to the migration of the socket. Among the ceramic heads implanted on a used titanium trunnion, no fracture was observed. This approach is possible, in so far as careful inspection does not show any major imperfection of the morse taper [3]. As for other bearing surfaces, the management of aseptic loosening of al-al prostheses is based on the amount of osteolysis around the loosened socket.

Purpose of the study: TKA on genu valgum raises serious problems for the ligament balance. Excessive release of the lateral retracted ligaments exposes the knee to potential instability in the frontal plane. To resolve this problem and avoid implantation of a constrained TKA, we opted for osteotomy of the lateral condyle removing the insertion of the lateral collateral ligament and the popliteal muscle after release of the fascia lata. The purpose of our study was to evaluate the functional and radiographic outcomes of these patients.

Material and methods: This was a retrospective study from 2002 to 2006. All patients with degenerative joint disease of the knee with severe and/or fixed genu valgum were included. These patients were implanted with a navigated posterostabilised Wallaby TKA (Navitrack) associated with osteotomy of the lateral condyle fixed with screws after acquisition of the ligament balance. The diagnosis and surgical history were noted. The preoperative alignement was determined on the full limb x-ray and from navigation data. The following variables were reviewed: polyethylene height, lowering of the lateral condyle, blood loss, operative time. The postoperative alignment was established at least one year after surgery. Intraoperative, postoperative and late complications were noted. The Knee Society function scores were used.

Results: Fifteen patients, mean age 70 years were reviewed at mean 35 months. The mean duration of the operative time was 136 min with mean blood loss of 620 ml. The mean PE height was 13 mm. All operated knees were corrected with mean alignment improving from 17.71 to 1.5 valgus postoperatively. The function score improved from 35 preoperatively to 79 at last follow-up. There were no cases of patellar instability or secondary laxity. Two patients developed late reflex dystrophy. The only case of revision concerned one non-union of the lateral condyle (screw removed at four months) but had a function score of 85 at last follow-up.

Discussion: Performing an osteotomy of the lateral condyle in complement with the navigated posterostabilised TKA for fixed genu valgum enabled good relaxation and satisfactory functional results so that totally constrained implants can be avoided.

Purpose of the study: Classically, bone grafts are harvested from the iliac crests which can provide a limited volume of graft material. Using the reaming product might help spare iliac bone.

Material and method: A variable head RIA device (reaming, irrigation, aspiration) was used over the last year for ten patients who presented partial or segmental bone loss. The bone graft was constructed exclusively with the reaming produce following membrane induction using a cement scaffold. The tibia was reconstructed in nine cases and the humerus in one. Bone loss was 6 cm on average.

Results: The reconstruction healed in six cases within a mean delay of 6 months; the 4 other cases are under assessment. Complementary bone was necessary to achieve healing in one case. There were no complications involving the donor site excepting transient pain at the point of insertion.

Discussion: In this series associating an induced membrane and reaming produce, the time to healing appeared to be shorter than with cancellous iliac bone. The smaller size of the fragment may be a determining factor.

Conclusion: Reaming products collected from medullary cavities of the long bones can be used for reconstruction of bone loss.

Purpose of the study: Leg length discrepancy after THA is a common complication and source of recurrent complaints from patients. To date, no reliable and reproducible technique has come forward to enable accurate quantification of all radiological parameters of the lower limb. Nevertheless, preoperative planning for hip arthroplasty requires knowledge of many limb parameters, in particularly leg length discrepancy, femoral offset, or the head-neck angle. The most widely used method is to use the 2D radiographs. The EOS system uses two digitalised 2D images taken orthogonally in a weight-bearing position to enable 3D reconstruction of the lower limb. The inter- and intraoperator reproducibility has been studied and validated. The purpose of our study was to compare the inter- and intra-operator reproducibilities of the measures taken on the standard full-length x-ray and those determined on the 3D EOS reconstructions.

Material and method: Twenty-five patients scheduled for THA were included in this study (50 lower limbs). Two independent operators determine the measures on the AP EOS view and on the 3D reconstructions obtained from two orthogonal EOS images. The following parameters were measured: femur length, tibia length, limb length, HKA, HKS, femoral offset, neck-shaft angle, head diameter, and length of the femoral neck. Each observer performed two series of measurements. Interobserver reproducibility was assessed with the intraclass correlation coefficient (CI: 95%). Student’s t test was used to compare the clinical parameters measured on the 2D and 3D images.

Results: Inter- and intraobserver reproducibility were 0.867 and 0.903 on the 2D x-rays and 0.911 and 0.940 on the 3D reconstructions. The better reproducibility of the EOS reconstruction was confirmed for all parameters tested in this study. Comparison of the 3D and 2D measurements revealed significant differences.

Discussion: Our study demonstrated that measurements made on EOS 3D reconstructions offer better inter- and intraobserver reproducibility than those made on the standard AP view. In addition, the 3D reconstruction takes into consideration of the projection of the anatomic structures in the plane of the AP radiograph. The EOS appears to be a pertinent tool giving reliable results for the pre- and postoperative work-up for arthroplasty of the lower limb.

To date, no technique has proved to be reliable and reproducible in order to precisely calculate radiological lower limb parameters.

EOS^® system allows from two bi-dimensional orthogonal radiographies in standing position to obtain a tridimensional reconstruction. A computerized system achieves the parameters calculation.

The aim of the study was first to evaluate the inter and intraobserver reproducibility of the EOS^® system, secondly to compare EOS^® measures with X-ray orthoroentgenograms.

Twenty-five patients about to receive total hip arthroplasty were included (fifty lower limbs). Two independent performers have carried out twice the measures either on standard X-rays and using three-dimensional reconstructions (femoral parameters (length, offset, collo-diaphy-seal angle, neck length, and head diameter), tibiae length, limb length, HKA, HKS). The reproducibility was estimated by intraclass correlation coefficients.

The inter and intraobserver reproducibility of the EOS^® measures have been respectively of 0.881 and 0.916 and more specifically of 0,997 and 0,997 for femoral length, of 0.996 and 0.997 for tibiae, of 0.999 and 0.999 for limb length, of 0.893 and 0.890 for HKS, of 0.993 and 0.994 for HKA, of 0.892 and 0.914 for femoral offset, of 0.765 and 0.850 for collo-diaphyseal angle.

The inter and intraobserver reproducibility using orthoroentgenograms reached 0.854 and 0.902.

Our results show the EOS^® is a tool allowing reproducible measures. Furthermore 3D EOS^® reconstructions offer better reproducible measures for all parameters that the orthoroentgenograms.

Its use prior to the decision of surgery and during surgery planning for lower limb arthroplasty is for us essential for adjusting surgical procedure accordingly.

Purpose of the study: The objective of this study was to establish an experimental sheep model for a surgical procedure which has been clinically successful for repairing major loss of bone stock: the Masquelet technique.

Material and methods: A 25 mm bone defect was created in a metatarsal bone then filled with a cement filler. After six weeks, the cement was removed after opening the neoformed pseudosynovial membrane. The cavity was left empty in group 1 (n=6) or filled with a morcelized cancellous autograft harvested from the iliac crests in group 1 (n=6).

Results: The surgery was well tolerated in all animals which were able to used the injured limb the day after the operation. Radiographic images and histological findings 24 weeks after surgery demonstrated that healing had not been achieved in all of the animals in group 1. Inversely, healing was achieved in all animals in group 2 at 24 weeks. Immunohistochemistry of the neoformed pseudosynovial demonstrated :

an abundant vascular network,

Conclusion: The sheep metatarsal model is a model of critical size with low morbidity. This model could be used to:

evaluate new therapeutic strategies for bone regeneration in conditions close to clinical situations,

The presence of a pseudosynovial membrane might:

be a barrier against the diffusion of bone morphogenetic proteins outside the lesion and

Purpose of the study: Complex fracture-dislocation of the elbow, and subsequent surgical treatment, is often the source of a stiff joint. The purpose of this study was to assess the efficacy of a therapeutic protocol combining systematic insertion of a dynamic external fixator allowing early mobilization of the elbow with restitution of the radial height and the coronoid process.

Material and methods: This consecutive series of ten patient, six men and four women, mean age 49 years, age range 27–67 years, underwent surgery from 2002 to 2004. Three patients presented a posterior Monteggia fracture (two type IIA, one type IId associated with comminutive fracture of the trochlea), four patients presented a dislocation associated with a Masson 4 fracture of the radial head and two presented inveterated dislocations diagnosed three weeks after the traumatic event. In all, seven patients presented a fracture of the radial head and six a fracture of the coronoid process. For all patients, the operation consisted in stabilization with a dynamic external fixator of the elbow associated or not with restoration of the radial height with a radial head prosthesis (n=4) and reconstitution of the coronoid process (n=6). The lateral ligaments had to be reinserted in four elbows. The comminutive fracture of the rochlea was treated with an iliac crest graft.

Results: One patient died early. One patient presented pin tract infection and four developed heterotopic ossifications. At mean follow-up of twelve months, outcome was excellent in our patients, good in four, fair in one (Mayo clinic classification). All patients had a stable elbow. The mean range of motion was 89° flexion-extension and 145° pronationsupination.

Conclusion: In this series, systematic use of external fixation for complex fracture-dislocation of the elbow joint yielded satisfactory results when the element stabilizing the joint were appropriately restored and when rehabilitation was undertaken early.

Purpose of the study: Computer assisted surgery for total knee arthroplasty is widely used in Europe. The reliability of these systems appears to be very good with bone cuts within 3° of the planned mechanical axis. Nevertheless, the relationship between intraoperative measurements provided by the navigation system and the postoperative gonometry can be used to assess the quality of realignment. The purpose of this work was to determine this relationship.

Material and methods: Thirty-three knees operated on with the Navitrack® system were assessed. Two senior operaters performed all procedures. An Omnia® cemented prosthesis with an ultracongruent fixed plateau was used in all cases. Intraoperative measurements wer made with the definitive prosthesis after cement solidification without stress on the knee. Postoperative gonometry was undertaken when the intraoperative flexion had disappeared six weeks to six months after the operation. Gonometric measurements were made by an independent operator using a computerized system operating on digitalized x-rays. The gonometric protocol had been standardized previously and only the gonometric measurements in compliance with this protocol were retained for analysis. The difference between measurements was analyzed with the t test for paired variables. Search for correlations was also performed.

Results: On average the intraoperative deviation was 0.8±0.8° (3° valgus to 2.4° varus). The postoperative gonometry showed 1.7±1.1° (3.4° valgus to 4.3° varus). The mean difference between the intraoperative axis and the measured postoperative axis was significant (p< 0.0001). There was no significant correlation between intraoperative and postoperative measurements.

Discussion: Computer-assisted navigation systems have their limitations which should be measured. The present findings would demonstrate a significant difference of minimal amplitude between the intraoperative measurement and the postoperative gonometry. Although the clinical pertinence of this difference remains to be demonstrated, it must be kept in mind for safe use of these navigation systems.

Purpose: Tissue engineering offers new therapeutic perspectives with the possibility of producing cartilage tissue for a large number of patients. These structures are composed of an absorbable synthetic support and competent cells. Two types of cells can be proposed: articular chondrocytes harvested from the peripheral part of the joint, or mesenchymatous stem cells (MSC) present in the bone marrow and possessing chondrogenic potential. The purpose of this study was to determine the optimal cell source and the best supporting material for in vitro production of cartilage.

Material and methods: Isolated rabbit MSC were harvested and amplified with cell culture for 21 days. After this period, 20–40 million cells/ml were combined with polyglycolic acid sponges (3 types of sponges 1x1x0.2 cm2) and cultured in TGFß-enriched medium under specific dynamic conditions allowing gas exchange. The tissue obtained was compared with structures of identical size obtained with differentiated chondrocytes harvested from the same animals. The study included a histological analysis and immunohistochemistry for type I, II, and X collagen and biochemistry for DNA content, glycosaminoglycanes (GAG) and type II collagen.

Results: After 3 weeks in culture, the composites obtained with MSC preserved their size and had the white pearly aspect of hyalin cartilage. The histological analysis and immunohistochemistry tests for type II collagen confirmed the presence of a cartilaginous matrix throughout the thickness of the fragments. The GAG and type II collagen contents were significantly higher with MSC compared with chondrocytes, irrespective of the supporting material.

Discussion: This study demonstrated that cartilaginous tissue fragments can be obtained with MSC cultured on PGA supporting material under very specific conditions. Use of these cells offers the advantage of easy harvesting followed by in vitro amplification, and thus less harvesting morbidity. Complementary studies are needed to evaluate the behaviour of these living materials after implantation in the articulation.

Purpose: Fracture of alumina implants is a rare cause of total hip arthroplasty (THA) failure. Over the last twenty years, improvement in the design of ceramic implants has helped reduce this risk. We observed five cases of alumina implant fracture over the last year. The purpose of this work was to analyse these recent events and to examine our experience to search for an explanation.

Material and methods: Between 1976 and 2002, eleven patients, five women and six men, mean age 57 years (range 32–87 years) required revision surgery for fracture of an alumina component. The time interval from primary surgery to revision was 36.5 months (range 7 – 106 months). An alumina-alumina couple was involved in nine cases and an alumina-PE couple in two. The primary THA used a press-fit metal backed alumina cup for five hips, a cemented alumina cup for three hips and a massive press-fit alumina cup for one. The alumina head measured 32 mm for ten hips, 22.2 mm for one. The femoral stems were cemented in seven cases, and hydroxyapatite surfaced in four. The acetabular component was involved in four cases and the femoral head in seven.

Results: Identified causes of ceramic implant fracture were: high-power trauma (insert) (n=1), insufficient head diameter (n=1), and defective head match (n=1). For the eight other fractures, involving three insets and five heads, no triggering factor could be identified. The three insert fractures involved two 50-mm cups and one 52-mm cup. Insert design had been changed before these fractures. Mean time from insertion to fracture was 12 months (range 8.5–15 months) in these cases.

Discussion: To our knowledge, all reported cases of ceramic implant fracture have concerned the prosthetic head. The present series demonstrates a recent development, fracture of acetabular implants. It is hypothesised that these fractures could be explained by an insufficient cup diameter associated with a recent change in the design of the metal-backed flush ceramic insert. These inserts have been modified and now have an “overhang” design associated with a thicker layer of ceramic for the smaller cup diameters. These accidents remain rare but justify careful surveillance, particularly during the first postoperative year.

Purpose: We searched to define a fracture index useful for predicting the risk of bone fracture in children with essential bone cysts.

Material and method: We reviewed 96 children with an essential bone cyst. The following clinical data were recorded: age, cyst localisation, circumstances of diagnosis.

The radiological analysis was based on 193 AP and lateral x-rays. We measured: 1) the distance separating the superior pole of the cyst from the suprajacent growth cartilage, 2) the largest cyst diameter, 3) the greatest cyst height, 4) the thinnest cortical width facing the cyst, 5) the cyst surface area calculated exactly using surface area software and expressed as a a ratio of shaft diameter (S/d2, Kaelin index). These different parameters were compared for cysts associated with fracture or not.

Results: Mean age at diagnosis was 10.4 years (range 2 – 12.8 years). Most of the cysts were located in the upper portion of the humerus (72%). Fracture was the inaugural sign in 68% of the cases.

Comparing the two cohorts of patients demonstrated that the following differences were significant (Student’s t test): 1) cyst width (p=0.0038): below 16 mm none of the cysts fractured. For wider cysts, there was no difference between the fracture and non-fracture cysts. 2) cortical thickness (p=0.0002); cortical thickness greater than 5 mm protected against fracture. If the cortical measured less than 3 mm, the risk of fracture was greater than 50%. 3) Kaelin index: (p< 0.0001) was directly correlated with fracture risk but no cutoff could be identified.

For an 80 – 100% risk of fracture, the cyst must have the following characteristics: width > 30 mm, height > 75 mm, cortical thickness < 2.4 mm, Kaelin index > 5.

For a 50% risk of fracture, the cyst must have the following characteristics: width > 24 mm, height > 55 mm, cortical thickness < 3 mm, Kaelin index > 3.

Conclusion: The Kaelin index is reliable but difficult to calculate in the consultation setting. Cortical thickness is a good indicator. Its predictive value can be improved by correlating the height of the cyst with its width. These measure can be obtained easily during consultation.

Purpose: During revision of alumina-on-alumina prostheses, reimplantation of a ceramic head on the former stem raises the risk of head fracture. Furthermore, serious deformation of the metal head has been reported after failure of ceramic-on-ceramic prostheses. The purpose of this retrospective study was to evaluate the incidence of these complications and analyse clinical and radiological results of revision for alumina-on-alumina total hip arthroplasties.

Material and methods: This consectutive series included 107 patients (74 women and 33 men, mean age 65 years, age range 32–91 years) who underwent 118 revision arthroplasties from 1976 to 1997. The cause in the majority of the cases was isolated non-infected cup loosening (80%) or ceramic implant fracture (4%). There were 72 cemented alumina cups, 38 screwed cups, six press-fit metal-backed cups, and two massive alumina impacted cups. Cavitary acetabular osteolysis was present in 67% of the cases. The cup was replaced alone in 94 cases (80%). The implanted cup was a polyethylene cup in 93 cases and an alumina cup in 25. When the femoral stem was left in place, a new ceramic head was implanted in 49 cases (39 alumina, 10 zircone). The alumina head was replaced by a metal head in 18 cases.

Results: Four early dislocations and two infections were noted. Mean follow-up after revision surgery was 67 months. Fourteen patients were lost to follow-up. Sixteen patients required revision surgery, ten of them for repeated cup loosening. At last follow-up (mean 61 months), there were no failures among the 49 ceramic heads implanted on a stem left in place. No revisions were required for polyethylene cup wear or metal head wear. In all, 72 hips were evaluated at a minimum two years follow-up. The functional score at last follow-up was 15.2±3.5. Seven-year actuarial survival was 95.5% taking revision of non-infected revision as the endpoint.

Discussion and conclusion: Analysis of the results shows that acetabular loosening is the main cause of revision of alumina-on-alumina total hip arthroplasties. The femoral stem was loosened in only nine cases and could be preserved in 80% of the cases. Replacement of an alumina-on-alumina couple with a metal-on-polyethylene couple does not appear to be a contraindication in this series.

Purpose: Post-traumatic instability of the elbow is defined as a subluxation of the humeroulnar joint at least three weeks after trauma. Treatment is based on restitution of the three elements essential for stability: the coronoid process, the height of the head of the radius, repair of the lateral ligaments. The purpose of this study was to analyse treatment modalities for these unstable elbows and assess mid-term results.

Material and methods: This consecutive series included all patients who underwent surgery between 1992 and 2000. There were 22 patients (twelve men and ten women, mean age, 46 years, age range 26–74 years). The left elbow was involved in 16 cases (two dominant) and the right elbow in six cases (six dominant). The initial trauma was isolated dislocation in six patients, dislocation with fracture of the coronoid process and the head of the radius in nine, trans-olecranon fracture-dislocation in seven. Mean time from trauma to revision for instability was four months. A dynamic external fixator was used for stabilisation in all cases, with or without restoration of the height of the radius with a radial head prosthesis (n=12), and reconstruction of the coronoid process (n=7). Reinsertion of the lateral ligaments was necessary in 17 patients.

Results: Mean follow-up was 33 months. Six patients required a second procedure for transposition of the ulnar nerve in three and arthrolysis of the elbow joint in three. There was one failure requiring total elbow arthroplasty less than one year after revision. At last follow-up, outcome according to the Mayo Clinic classification was excellent in ten patients, good in five, fair in one, and poor in five (four trans-olecranon fracture-dislocations). Twenty patients had a stable elbow. Mean flexion-extension was 113° with a 19° mean extension deficit. At last follow-up, six patients had radiographic signs of osteoarthritis.

Discussion and conclusion: The results were directly correlated with the nature of the initial trauma and the quality of the restoration of he stabilising elements. The poorest results were observed after trans-olecranon fracture-dislocation, which led to osteoarthritic degradation in three out of four cases. In our experience, treatment of sequelae of elbow dislocation, or the terrible triade, can give satisfactory results with an appropriate treatment strategy.