Surgeons often protect Tendon-bone reconstructions such as rotator cuff repairs by off loading them. We investigated the effect of limb position and boundary conditions in an in-vitro rabbit patella tendon-bone repair model. Patella tendons were repaired back to the tibia in eight hindlimb cadavers with 2 mitek anchors(Mitek, Westwood, MA) and 3-0 Ethibond (Ethicon, Sommerville, NJ) using two techniques, one involving simple sutures and the other involving crossing over between the sutures. A loading mechanism through the patella tendon was constructed using static weights over a pulley mechanism. The contact area and force at the PT-bone interface were measured using a TekScan pressure sensor (6911, TekScan, South Boston, MA). The contact footprint (area and normal force) was acquired under four configurations: (1) knee full extension with interface unloaded, (2) knee 45° flexion with interface unloaded, (3) knee full flexion with interface loaded by limb weight alone, (4) tendon loaded with limb weight and 20N force applied through tendon loading mechanism. The contact area force footprint changed substantially between the different suture techniques and loading configurations. Crossing over of sutures appears to provide an increased and more evenly distributed force across the tendon-bone interface. Repair off-loading was accompanied by a decrease in the contact footprint force and pressure. The force in both suture techniques increased with increasing flexion angle and was substantially increased by both bearing the weight of the dependent limb and by an axial load in the patellar tendon. Off loading a repair may not provide optimal environment for healing.

Aim: To measure contact pressures in vivo in patients with unicompartmental arthritis fitted with osteoarthrosis (OA) braces to see if the arthritic side of the joint is unloaded.

Method: A thin flexible sensor (TekScan) was manoeuvered arthroscopically into the medial compartment of the knee joint under local anaesthesia in patients with unicompartmental OA of the knee undergoing either therapeutic or diagnostic arthroscopy. All 15 patients had been fitted with a brace before the arthroscopy. Measurements were made within the compartment of double leg stance and single leg stance. Ground reaction force using a load cell was measured for 14 patients and the knee sensor data were normalised relative to this. Recordings were then repeated with the patients with different commercially available braces.

Results: The first two groups of patients showed significant reductions in pressures. Normalised knee sensor forces were reduced to 68%(Sd 22%) and 61%(Sd31%). In the last group of patients, reductions in pressure recordings were less between no-brace and brace. Three patients produced low signals suggesting incorrect sensor replacement.

Conclusions:

Significant unloading of the osteoarthritic compartment could be observed by applying manually a valgus force to the knee.

Introduction: The biomechanical properties of tendon and ligament have long been the subject of intense research. The understanding of the ultrastructure as it relates to the biomechanical function and clinical demands have often considered the ultimate properties at failure alone. Tendons and ligaments are predominately loaded in-vivo at subfailure loads and often in the initial toe region. To date, little work has focussed on the viscoelastic properties of the tendon in the initial toe region. The biomechanical behaviour at these low loads may reflect the unique mechanical interactions between the fasciles and collagen fibrils. This study examined stress relaxation of ligaments in the initial non-linear portion of the load vs. displacement curve.

Methods: Six flexor tendons (2.5 mm wide x 1mm thick) were harvested from 18 month cross bred whethers and stored in 0.145 M NaCl until testing. Tensile testing was performed on a MACH 1 Micromechanical Testing Machine (BIOSYNTECH, Laval, Quebec, Canada) in 0.145M phosphate buffered saline at room temperature. Tendons (gauge length 30mm) were displaced to 0.5, 1 and 5% strain at a loading rate of 50 microns/sec and stress relaxation measured over a period of 300 seconds and repeated for 4 sequences. Data was analysed using an analysis of variance.

Results and Discussion: Peak loads at 0.5 % strain ranged from 50 g (sequence 1) to 130 g (sequence 4) while at 5 % strain peak loads reached upwards to 1600 g. These loads are well within the initial toe region of the load-displacement behaviour of the ligament. The MACH 1 testing system provide a reliable and highly accurate system to control micron level displacements and mg load resolution. Recently, Yamamoto and coworkers reported the stress relaxation behavior and strain rate effects of collagen fascicles differed greatly from those of bulk tendons. The differences in tensile and viscoelastic properties between fascicles and bulk tendons may be attributable, in part, to ground substances, mechanical interaction between fascicles, and the difference of crimp structure of collagen fibrils. The present study supports an important role of tissue ultra-structure at low loads with regard to stress relaxation. Subtle changes in ground susbtance, water content or biochemical consituents not evident in testin

Introduction: Various plating devices and screw systems are available for single and multi-level cervical fusions. Recent reports regarding screw migration under torsional load and a “windshield wiper effect” has brought to light the importance of plate and screw design as well as the choice of graft.

Aim: This study examined the relative stability of cervical plating systems under pure bending and axial-torsional fatigue using the Cloward type graft.

Methods: Five fresh-frozen human cervical and 10 porcine spines assessed by dual-energy x-ray absorptiometry (DEXA) scanning and then reconstructed at the C_2–3 and levels using the anterior Cloward technique. C_4–5 Two different plating systems (a solid plate and a hollow plate) were used and alternated between the C2–3 and C4–5 levels. Strain gauges placed on the plates themselves. The systems were subjected to pure bending and torsional loading.. Five kilogram loads were used to apply bending moments to the spine and did not differ between the two systems evaluated. Bending moments and displacement angles were recorded for the pure bending loading regime and torque versus time was recorded for the torsional fatigue loading.

Results: Strain gauge analysis revealed minimal strains on the plates under the loading conditions. Torque versus time was measured, and the decay constant was calculated from the decay curves. The hollow plating system decayed quicker than the solid plating system. Angular displacement under pure bending was minimal. The hollow system plate system resisted greater torque compared with the solid system. The decay curves eventually reached an asymptote for the both systems. This implied that the systems become stable under fatigue loading. The X-rays illustrated no failure at the screw/ bone interface (i.e. No “wiper” effect) after torsional fatigue.

Aim: To define the contact force and contact area at the glenoid labrum-bone interface between suture sites in an open transosseous Bankart repair, and to assess how these contact parameters are altered by tying adjacent sutures to each other.

Methods: Twelve capsulolabral avulsion lesions were created in fresh-frozen human shoulder specimens and were repaired using a standard transosseous suture technique. The contact forces and contact areas were measured at the labrum-bone interface between sutures before and after repair. Using the free suture ends, either a single or double strand knot was then tied between adjacent suture sites and the contact parameters were measured again.

Results: The contact forces and contact areas under the soft tissue bridges between transosseous sutures were mildly increased during repair (before repair: average force=5.53g, area=2.25mm²; after repair: force=11.7g, area=3.13mm²). However, both the contact forces and areas increased significantly when a single or double strand of suture was tied over the soft tissue bridge. The double strand technique resulted in a significantly greater increase in contact forces and areas than the single strand technique (single strand average force=70.1g, area=6.75mm²; double strand average force=95.15g, area=8.0mm² p< 0.05).

Conclusions: The contact parameters between labrum and bone in a Bankart repair were increased when the suture strands from adjacent transosseous repair sites were linked. Increasing contact force or contact area may improve healing at the bone-soft tissue interface, and may reduce the risk of “spot welding” repairs. This, in turn, may reduce the failure rate of Bankart repairs.

Introduction: Whether or not to resurface the patella in total knee replacement (TKR) is controversial. One concern is the possibility of progression of the arthritis in the patellofemoral joint that has not been resurfaced when exposed to the stress of articulating with the femoral component.

Methods: The cohort comprised six knees for Trac TKA (Biomet). The assessment involved the use of an electronic sensor system(Iscan, Tekscan). The readings were taken on an anaesthetised patient, during surgery. A tourniquet was not used. A subvastus operative approach was used.

Results: The contact area and contact stress increased with flexion with and without the femoral component in place. We measured no increase in patellar stress when the patella that had not been resurfaced articulated with the femoral component.

Introduction: Unicomparmental knee replacements have a long clinical history of success as well as failure. Recently, in Australia some 40% of knee surgery performed consists of unicompartmental knees for the treatment of medial compartment OA. This increased use of unicompartmental knees is in part due to advances in surgical technique through a minimally invasive approach. Loading conditions at the tibia-implant interface will play an important role in the stress/strain distributions at the proximal tibia. The use of an all PE tibial insert versus a metal backed component may provide a different strain disribution to the proximal tibia. This study examined the influence of metal backed and polyethylene tibial components in unicompartmental knee replacements with and without cement fixation on the initial strain distributions under various loading conditions.

Materials and Methods: Three cadaveric tibias (mean age 47 years old) were cleaned of all soft tissue and strain gauged. Rosette strain gauges (TML Ltd., Tokyo, Japan) were placed at 2 levels on the tibial cortex. The intact tibia were embedded in a low melting point alloy at a standard height and tested using an MTS 858 Bionix testing machine (MTS Systems, Min., MI). The tibia were tested in nuetral, varus and valgus positions at zero and sixty degrees of flexion. A 1500N was applied for 15 seconds and the strains measured. A K-Scan sensor (Tekscan, Boston, MA) was used to confirm the varus and valgus loading positions and to obtain a contact footprint and pressure for the intact and reconstructed tibias under the loading conditions (Fig. 1). Following intact testing, the tibias were templated and reconstructed by a surgeon familiar with the technigue. The implants were investigated with and without cement fixation and compared to their respective all polyethylene component if it was available using the same loading regime as the intact tibias. Principal strains were calculated.

Results: Tibial cortical strain distributions were significantly different at the proximal and distal sites under the loading conditions examined. The strain distribution for metal backed components was greater than the all PE design. Increasing flexion angle shifted the peak strains posteriorly. Metal backing and all PE tibial inserts presented different strain distributions on the medial side under nuetral and varus loading. Lateral compartment strains did not differ between designs, were higher proximal and decreased dramatically at the distal gauges. Cementless fixation tended to overload compared to the intact condition. Figure 2 presents the strain distribution for a typical metal backed and all poly unicompartmental knee in the nuetral position.

Discussion: Metal backed unicompartmental components overloaded the proximal cortex of the tibia. All polyethylene tibial inserts did not overload the proximal cortex and had similar strain distribution to the intact tibia. Cemented fixation allows the transfer of load to the distal tibial cortex via the proximal cortex and subchondral bone, provided that the bone cement has inter-digitised the subchondral bone.

Introduction: The stress and strain in the proximal femur after total hip arthroplasty are influenced by the geometry of the implant in addition to its materials properties and applied loading. The addition of a third taper in the medio-lateral plane may provide additional stability and improved load transmission.

Aim: To examine the relative stability of double and triple tapered stems in two finite element (FE) models.

Methods: The geometry of a polished, double-tapered and a triple-tapered stem were scanned using a three dimensional technique. Two FE models of the stems were created using PATRAN. The models were analysed using the ABAQUS. Tied and sliding contact conditions were allowed between the implants and the cement mantle. The interface at the distal tip of the stem was removed to represent the scenario with a distal centraliser present.

Results: When tied contact was assumed, both stems displayed similar von Mises’ stress distributions. The peak stresses remained constant in the double tapered stem, with a marked translation of regions of high stress towards the distal tip with the introduction of sliding contact conditions. Peak stresses in the triple tapered stem decreased, but displayed a more continuous distribution along the implant with sliding contact. Torsional loading of the stems reduced the magnitude of the distal tip stresses.

Conclusions: The triple-tapered geometry displayed a more even distribution of stresses along the length of the implant. The double-tapered geometry displayed a high stress state at the distal tip of the implant.

Introduction: The transition from fixed bearing to mobile bearing total replacements represents a recent trend in Australia with the introduction of many new designs. The complex kinematics of mobile bearing designs coupled with the importance of proper surgical techniques including soft tissue balancing presents a number of factors that may influence the short and long-term success of these implants. There have been few reports in the literature on the performance of many of the new mobile bearing design with regards to initial wear of the polyethylene (PE) insert.

Aim: To investigate the patterns of wear on three mobile bearing knee PE inserts that had been retrieved soon after implantation.

Methods: Three mobile bearing knee PE inserts were retrieved at the time of revision surgery and submitted to our laboratory for examination. The proximal and distal articulating surfaces were examined by measuring surface roughness (Ra) using a Surfanalyzer (5400 (Federal Products Co., Providence, RI, USA) following ISO 97. Optical microscopy and scanning electron microscope (SEM) analyses were used to locate and identify patterns of wear.

Results: The average time in service for the PE inserts was 18.6 months. The maximum Ra values were noted on the anterior-lateral side for all implants. Optical and SEM analysis revealed wear mechanisms that included burnishing, scratches, pitting and cold flow. Damage to the distal surface was noted in all samples with extensive wear tracks noted in the LCS and TRAC knees.

Conclusion: The surface roughness analyses showed asymmetrical wear on the distal PE interfaces as well as wear on the proximal PE interfaces. The presence of embedded particles and debris suggests a third-body mechanism. Dislocation and general instability may have exacerbated the early signs of wear in these components.

Introduction: The classic teaching in flexor tendon repair suggests that a 10mm bite is important for the integrity of the repair regardless of the other features of the technique. Although this has been widely accepted since Bunnell’s first descriptions of accurate flexor tendon repair there appear to be little data to support it. An extensive review of the literature showed no biomechanical data relating specifically to size of bite in flexor tendon repair. We hypothesised that decreased bite may cause less damage to the tendon during repair while still offering adequate mechanical strength.

Aim: To investigate the effect of different bite sizes on the mechanical properties of flexor tendon repairs.

Methods: Twenty fresh-frozen cadaveric flexor tendons were divided at their centres. One side of a modified Kessler repair was used on each side taking a 6mm bite on one side and a 10mm bite on the other. The tendons underwent tensile testing on a mechanical testing frame by pulling on the ends of the suture with the tendon secured in pneumatic grips. Data for stiffness and ultimate load to failure were recorded.

Results: An increased bite size made no significant difference to stiffness of the repairs. There was a difference in load to failure noted but this was not significant. The ultimate load to failure was noted after the specimens had been distracted over 2mm, which would result in clinical failure.

Conclusions: These results suggested that a 10mm bite may be excessive in flexor tendon repair and could cause more tissue damage than lesser bites. Further study of in vivo effects of decreased bite size is required.

Introduction: Acetabular osteolysis is common behind cups with holes (the reported incidence is 9% to 36%). Fluid pressure has been implicated in the pathogenesis of osteolysis.

Aim: To test the hypothesis that a polyethylene liner in a metal cup can act as a pump in vivo.

Methods: This study was performed during revision surgery in six cases. The components were from several manufacturers. All were ingrown uncemented cups that had osteolytic lesions associated with holes in the cup.

A cannula was placed through the capsule into the hip joint and another was placed through the periosteum and bone of the ilium into the osteolytic lesion above the ingrown cup. The continuity of these two spaces through the holes in the cup was confirmed by the injection of methylene blue. Pressure transducers were then connected to both cannulae. Measurements were taken while applying compression and distraction forces across the artificial hip joint.

Results: Compression and distraction loads produced a rise (48mmHg) and a fall (35mmHg) respectively in the pressure in the osteolytic lesion but no change in the hip joint pressure, thereby, demonstrating a pumping action. After exposing the prosthesis we were able to demonstrate 1mm to 2mm of in-out excursion of the polyethylene liner in the metal cup, which may explain the mechanism of the pumping effect.

Discussion: The polyethylene liner in the metal shell can act as a pump. Compression and distraction forces, such as occur in normal gait, produce changes in fluid pressure, which are transmitted through the holes and may cause osteolysis behind the cup.

Methods: Sixty pseudo-capsular tissues from loose shoulder, hip and knee (20 each) arthroplasties and 30 capsular tissues from primary total joint arthroplasty (TJA) patients (10 each; 12 rheumatoid arthritis [RA], 18 osteo-arthrosis [OA]) were investigated for mRNA and protein expressions of IL-1ß (interleukin-1 b, IL-1Ra (interleukin-1 receptor antagonist), MMP 1 (matrix matalloproteinase-1)-, TIMP 2 (tissue inhibitor of MMPs-2) using in situ hybridisation and immunohistochemistry. Polyethylene and metal debris in the same sections were semi-quantified simultaneously.

Results: IL-1ß mRNA and proteins were expressed in most RA primary and revision tissues and were less expressed in OA primary tissues. In contrast, IL-1Ra mRNA was found in most primary OA tissues and less in RA primary and the revision tissues. The ratio of staining intensities of IL-1ß/IL-1Ra mRNA was higher in revision and primary RA tissues compared with the primary OA tissues. MMP-1 protein expression was correlated with the IL-1ß/IL-1Ra ratio. Polyethylene (PE) debris was found in 56 out of 60 of the revision tissues. Their sizes were different in the hip (mainly small, < 30 mm in diameter), the knee (mainly large, > 300 mm) and the shoulder (all sizes). The expressions of the detected factors were highly correlated with the concentration of the PE debris but with not their sizes.

Conclusions: The high ratio of IL-1ß/IL-1Ra in primary RA and revision tissues and its correlation with MMP-1 expression and PE debris concentration indicated that an over-expression of IL-1ß and/or regulation downwards of IL-1Ra is an important event in inflammatory disorders and the foreign body reaction in TJA. A therapeutic strategy with IL-1Ra, that has been considered in RA treatment may thus contribute to the longevity of prosthesis of a TJA.

Introduction: Proximal bone resorption is a common problem after total hip arthroplasty. This has been attributed to stress shielding and has been reported to be more pronounced for cemented than for uncemented implants.

Aim: To investigate the cortical strain distribution of a new proximal “fit and fill” cementless, titanium, femoral, hip prosthesis based on the SROM design.

Methods: Strain gauges were mounted on five fresh-frozen cadaveric and five saw-bone femora and checked against a template for the prosthesis. The strain gauges were placed at four levels on the anterior, posterior, medial and lateral cortices corresponding to the Gruen zones. Two extra strain gauges were placed on the proximal posteromedial cortex. Loading was applied to the intact and reconstructed femora in the ISO 7206–4 orientation and single legged stance in an MTS servo-hydraulic testing machine. Data were analysed using analysis of variance.

Results: The strain distributions following reconstruction and multi-axis loading (ISO 7206–4 orientation) approximated the strains in an intact femur in the diaphysis. The proximal posteromedial cortical strains were approximately 50% of those of the intact femur.

Conclusions: The strains observed in the proximal femur following reconstruction in the present study are considerably higher than most others reported in the literature. A number of factors may contribute to the high proximal strains observed. This study has illustrated that geometric design and material selection along with surgical technique may allow for greater loading to proximal bone and enhance the long term integrity of this type of implant.

Introduction: A tibial tubercle osteotomy can be used in the exposure of severe articular deformity and the tight knee in total knee arthroplasty, especially revision surgery. This osteotomy has been popularised by Dr. L. Whiteside [1] who described transosseous wiring to secure the osteotomy following joint reconstruction. Other fixation techniques including the use of cables and screws may provide options for this technique. The current study exmained 3 different fixation methods for tibial tubercle osteotomy using an in-vitro sheep model.

Materials and Methods: Tibial tubercle osteotomies (5 cm in length) were performed in ten adult sheep tibias. The osteotomies were fixed sequentially using circumferential Dall-Miles cables (Howmedica, Ritherford, NJ) (Fig. 1), transosseous wires and lastly 2 AO screws. Testing of each fixation configuration was performed using an MTS 858 Mini Bionix servo hydraulic testing machine (MTS Systems Corporation, USA). The loading regime used a cyclic 200 N load applied along the line of the patellar tendon with micromotion measured at the osteotomy and adjacent bone using optical sensors (MEL, Bahnhofstr, Germany). Data was analyzed using ANOVA. Micromotion at the end of the osteotomy fragment was recorded for 150 cycles for each tibia following reconstruction with cables, wires, and screws in succession.

Results: Cable fixation provided the most stable construct followed by screws and wires respectively. Wire fixation had the greatest variation in micromotion (370 microns). The ovine tibia model provides a reproducible bone bed to evaluate different fixation strategies for tibial tubercle osteotomy. Clinically, differences may even be marked considering anatomic and bone quality issues as well as magnitude of the load that have not been addressed in this in-vitro study.

Discussion: Fixation of the osteotmoy is an important surgical technique. Wolff et al. found that major complications related to the surgical technique occurred in 23% of the knees performed in 26 cases [2]. Reis et al. [3] observed fixation with 3 or 4 titanium screws was sufficient after a follow-up period of 18 months. Twenty-nine of the osteotomies healed primarily. One patient developed postoperative displacement of the tibial tubercle requiring additional screw and suture fixation. This study has shown that micromotion of a tibial tubercle osteotomy fixation in sheep is dependent upon fixation technique. Cables provided the most stable fixation compared to screws and wires in an ovine tibial model.

Aim: Radio-frequency treatment is used clinically in unstable joints to reduce the length of the supporting soft tissues to help provide stability. The mechanical properties after treatment have not been adequately studied. Since there is a change in the tissues’ ultra-structure with treatment we hypothesised that different collagenous tissues may have varying responses to radio-frequency treatment.

Methods: Ovine extensor tendons and cadaveric gleno-humeral capsules were tested on a MTS machine to investigate the dynamic and failure properties before and after radio-frequency treatment. Three radio- frequency treatments of different power (5, 10 and 20W) were used and two different treatment times (10s and 30s) to investigate the effects of treatment power and time on changes in the mechanical properties.

Results: The tissue shortening that was produced in the tendons and capsules was progressive with increases in treatment wattage and time. The tendon failure-force and stiffness were significantly reduced by the radio-frequency treatment but no significant changes were found in the capsules. Considering the dynamic properties only, the tendons showed significant changes with treatment. The mechanical properties were significantly different between control and treated groups but not between the treatment settings.

Conclusions: The tissue type altered the effect of radio-frequency treatment on the mechanical properties. Varying the treatment wattage and time did not significantly alter the changes observed with the largest difference being between control and treated tissue at any treatment setting. Therefore, radio frequency was proven to shorten collagenous tissues in a predictable manner but changes to the mechanical properties depend on the tissue type.

Introduction: Uncemented humeral components rely heavily on initial stability and fixation as a function of the design of the implant. Concerns over initial torsional stability of humeral components have motivated the development of a variety of design concepts.

Aim: To investigate the torsional stability of two types of cementless humeral shoulder prostheses.

Methods: Twelve fresh-frozen cadaveric humeri were cleaned of all soft tissues and prepared for reconstruction with the two types of cementless humeral shoulder prostheses. The humeri were embedded in a low melting point alloy and tested in a servohydraulic-testing machine. The loading applied to the humeri was a controlled angle loading regime at ± 1.5 degrees for 150 cycles. Torque versus time was measured, and the exponential time constant was calculated.

Results: The Z implant displayed overall a tightening effect, and a positive time constant. Whereas the G implant displayed a negative time constant, i.e. a loosening of the implant.

Discussion: These differences reflect the initial stability achieved immediately following surgery and may have important implications for bone in-growth and long-term stability.

Introduction: Abnormal fracture healing in aged, post-menopausal or ovariectomised patients remains a clinical problem. Understanding the distribution and regulation of biomolecular factors in fracture healing in oestrogen deficient rats may have clinical implications for developing novel therapeutic strategies for enhancing osteoporotic fracture healing. Our previous work demonstrated that bone morphogenetic proteins (BMPs), transforming growth factor beta (TGF-ß) and their signal transducers, Smads, played important roles in normal fracture healing. Insulin-like growth factor I (IGF-I) has been indicated playing a role in the maintenance of bone mass. Matrix metalloproteinases (MMPs) has been indicated to play a role in bone matrix degradation. Those factors in ovariectomised fracture healing have not yet been reported.

Aim: To investigate the expression of BMP-2, 7, TGF-ß, Smads1–7, IGF-I, IGF-I receptor 1a (IGF-IR1), MMPs and TIMPs by a quantitative immunohistochemistry in a fracture model in an ovariectomised rodent (OVX).

Methods: Age-matched, normal, female rats served as controls. The animals were sacrificed in groups of six at one, two, three, four and six weeks after the fracture.

Results: The highlights of our results were the lack of IGF-I in the early stage of fracture healing (up to two weeks) in OVX rats and the greater expression of MMP-1 in OVX rats at all groups when compared with the normal rats.

Conclusions: Our data suggested that the regulation downward of IGF-I in the OVX fractures resulted from estrogen deficiency and may have the function to stimulate MMP-1 activity. Over-expressed MMP-1 degraded collagen matrix in the cortex and inhibited the woven bone matrix formation during OVX fracture healing.

Introduction: Hip instability is a complex and challenging problem. In experienced units, up to 4% of patients undergoing total hip arthroplasty will require revision surgery to treat hip instability, with only 60% of these treatments being successful. Many authors reporting results with various constrained systems available have described dislocation rates post implantation of the constrained component of 4% to 29%.

Method: The thirteen patients who underwent placement of a constrained component as a revision procedure in our unit from 1989 to 2000 were reviewed.

Results: The indications for revision surgery included recurrent dislocation in eight and intraoperative instability in five revision hip arthroplasties. No patients were lost to follow up. The average follow-up was 43 months(range 14 to 121). The average age at time of surgery was 73 years(range: 52 to 84 years). No component has been revised. The average hip score after revision surgery was 72(range: 52 to 89). There have been no episodes of dislocation of the constrained arthroplasty. In seven cases the constrained arthroplasty was implanted into a previously placed well fixed shell.

Conclusion: Constrained acetabular components were a highly effective tool in the treatment of hip instability.

Introduction: The lordosis of the lumbar spine, flexion angle and body weight result in significant shear forces through the lumbar and lumbosacral disc spaces. These shear forces result in translational motion across the disc space, which is resisted but not completely abolished by pedicle screw stabilisation. Failure of lumbar interbody fusions through non-union may be related to translational micromotion at the vertebral endplate / bone graft interface. A porcine in vitro model was established to test whether variations in the design of inter-body implants and in particular, the presence of surface serrations would assist in resisting shear forces – especially those causing anterior translation.

Methods: Measurements of anterior vertebral translation were recorded on porcine cervical spine segments, subjected to 25 N antero-posterior shear load while under a 300 N compressive pre-load. Baseline testing was firstly performed on the intact specimens and following removal of the facet joints. The annulus, disc nucleus and cartilaginous endplates were then removed and the specimens were divided into two groups for testing using interbody implants. Four stainless steel blocks measuring 15 mm (length) × 5 mm (height) × 4 mm (width) were manufactured to act as intervertebral disc spacers. Two were made with smooth surfaces and two were made with 1 mm deep serrations on the upper and lower surfaces. One group was tested with two smooth and one with two serrated implants.

Results: Under 25 N shear load, the specimens tested with the serrated implants showed anterior vertebral translation of 0.046 ± 0.013 mm while those tested with the smooth surfaced implants measured 0.152 ± 0.075 mm (p < 0.01). A significant difference was also found between the stiffness of the specimens implanted with smooth surfaced (432.8 N/mm) and serrated (1088.4 N/mm) implants (p < 0.01). The value for peak load at failure for the specimens with smooth surfaced implants (150.43N) was less than those implanted with serrated implants (175.48 N), but not significantly different.

Discussion: The presence of surface serrations on the interbody implants significantly increased the resistance to shear forces in this model. In the clinical setting, we postulate that the degree of micromotion generated by anterior shear forces at interbody fusion sites should be substantially less when serrated implants are used and reduce the incidence of non-union. This may explain the improved fusion rates reported by contemporary authors when using some interbody implants. Further research is needed to clarify the combined effects of pedicle screw stabilisation and interbody implants upon shear displacement and variations in implant design.

Introduction: Arthrodesis of the shoulder joint is appropriate for several conditions, including paralysis, degenerative disease, infection, and salvage of failed arthroplasty. Two common complications of shoulder fusion, non-union and unacceptable arm position, may reflect failure to achieve rigid fixation during the surgical procedure. Numerous fixation techniques have been described, including plate fixation, external fixation, and screw fixation.

Aim: To compare the biomechanics of five fixation techniques of shoulder fusion in a human cadaveric model.

Methods: Twenty-five shoulder fusions were carried out in fresh-frozen human cadaveric specimens with the following five techniques: screw fixation alone (n=5), external fixation alone (n=5), external fixation supplemented with screw fixation (n=5), single plate fixation (n=5), and double plate fixation (n=5). Each specimen was tested on a servo-hydraulic machine under repeated physiologic loads to determine the bending and torsional stiffness.

Results: There was a statistically significant difference in bending and torsional stiffness between all five fixation techniques (ANOVA, p< 0.05). Normalised bending (B) and torsional (T) stiffness, in descending order, were: double plate (B=1.0, T=1.0), single plate (B=0.77, T=0.89), external-fixation with screws (B=0.68, T=0.74), external-fixation alone (B=0.40, T=0.53), and screws alone (B=0.13, T=0.26).

Discussion & Conclusion: Statistically significant differences in bending and torsional stiffness have been identified using five different techniques of shoulder fusion. The risk of the most common complications of this surgical procedure, non-union and unacceptable arm position, may be minimised if these biomechanical findings are applied to surgical decision-making.