Polyethylene wear remains an important cause of failure in knee replacements. Retrieval studies, simulators and simple X-ray methods produce wear data that may be inaccurate or unrealistic. We have developed an accurate RSA system for measuring wear in-vivo. Using this system we have found wear rate in a fixed bearing TKR to be about 0.1mm/yr[ Four Oxford uni-compartmental knee replacements, with excellent clinical results were studied ten years after implantation. RSA X-rays were taken in double leg stance with the knee in full extension and 15 and 30 degrees flexion. Following RSA calibration, silhouettes of the components on the stereo X-rays were extracted using a Canny edge detector and were matched to silhouettes projected using CAD models to determine the 3D component position. The average minimum thickness of the bearing was determined and was compared with the measured minimum thickness of 14 unused bearings to calculate linear penetration. The average linear penetration after average 10 years implantation (range 8.5 to 10.25 years) was 0.16 mm (SD 0.13 mm). The average penetration rate was 0.017 mm/year (SD 0.011 mm/year). The maximum linear penetration rate was 0.027 mm/year. The penetration rate is similar to that obtained in a retrieval study [
The kinematic profile of single axis design TKR was closer to normal especially near extension. During mid-flexion, abnormal anterior femoral translation was noticed with the polyradial design. No significant difference was noted between CR and CS designs.
Increasing the width of the proximal section of a polished tapered stem enhances its rotational stability.
Unicompartmental knee replacement (UKR) is an established and effective treatment for early unicondylar osteoarthritis of the knee. However good results will only be achieved with a UKR if appropriate implants, indications and surgical techniques are used. There are now many UKR available. The majority have been introduced recently and have no published clinical results, as a result it is not clear how well they will function. Wear is a potential problem with UKR because of thin polyethylene and small contact areas. To minimise wear we use a device with a fully congruous unconstrained mobile bearing, the Oxford UKR. The indications for UKR are confusing. The Oxford UKR is recommended for medial compartment osteoarthritis with full thickness cartilage loss and a functionally intact Anterior Cruciate Ligament. The Varus deformity should be correctable and there should be full thickness cartilage in the lateral compartment. It is appropriate for about one in four osteoarthritic knees needing replacement. With fixed bearing devices, because of problems with wear, the indications are narrower and contraindica–tions include young patients and damage to the Patello-femoral joint. These devices can however be used in the lateral compartment. There is currently a vogue to consider UKR as a pre-TKR. Under these circumstances the indications are relaxed and worse results are achieved. UKR are now routinely implanted through a minimally invasive approach, which decreases morbidity and aids recovery. There is concern that the small incision will compromise implantation. The techniques used range from free hand with a burr to sophisticated instrumentation. We use a mill to precisely restore ligament balance and function to normal and have shown that, with this instrumentation, the device can be implanted as precisely through a short incision as through a standard one. If appropriate implants, indications and surgical techniques are used then UKR achieve better short term functional results than both HTO and TKR, and they can achieve a long term survival that is similar to TKR and better than HTO. Under these circumstances we believe that UKR is the treatment of choice.
Radiolucencies are generally thought to be a manifestation of loosening. They are commonly seen under the tibial component of the Oxford Unicompartmental knee replacement. However, they are not associated with long term failure [ Eight medial Oxford Unicompartmental knee prostheses were examined post-operatively, at 12 months and at 24 months using Roentgen Stereophotogrammetric Analysis (RSA). The serial radiographs were calibrated and three dimensional CAD models were matched to the extracted component shapes. Implanted bone marker balls were then selected on all radiographs. The components were superimposed. Migration calculation was achieved by comparing relative positions of components to bone marker balls in the serial radiographs. Fluoroscopically screened X-rays were taken to determine whether there were any radiolu-cencies under the tibial component. By 12 months, the femoral component had migrated significantly anteriorly (0.35 mm) and proximally (0.61 mm). The tibial component had migrated significantly distally (0.42 mm) and anteriorly (0.47 mm). Although on average there appeared to be a small further migration between 12 and 24 months, this was not statistically significant in any direction. Six of the eight patients had radiolucencies at 24 months. None of the radiolucencies were was complete. The maximum radiolucency was 1 mm thick and the average was 0.57 mm. There was no correlation between radiolucency and migration. The proximal femoral migration and distal tibial migration can be explained by the load across the knee. We cannot yet explain the anterior tibial migration. Radiolucencies almost always occur and are not related to migration.This confirms they are not associated with loosening.
The relative motion between a prosthesis, the cement mantle and its’ host bone during weight bearing is not well understood. Using Radiostereophotogrammetric Analysis (RSA), we examined the dynamically inducible micromotion that exists at these interfaces when an increased load is placed through the prosthesis. Dynamically inducible micromotion was measured in the femoral components of 21 subjects undergoing total hip replacement with polished Exeter stems. Two standing RSA studies were performed, at 3 and 12 months postoperatively. Firstly in double-leg stance, and secondly fully weight bearing through the operated hip. Subjects had no signs of clinical or radiological signs of loosening at 1 year. Significant micromotion was detected at the prosthesis-cement interface at 3 months. Similar patterns of micromotion were observed at 12 months. The prosthesis appeared to bend during single-leg stance weight bearing, however this accounted for less than half of the total observed movement. Conventional RSA studies were conducted at 3 months, 6 months and 1 year to confirm that the implants showed normal migration patterns. This study demonstrates that movement exists between the prosthesis and bone during cyclical weight bearing. This dynamically inducible micromotion probably occurs at the prosthesis-cement interface. It could account for the wear that is observed on the surface of retrieved secure prostheses. This may be a mechanism by which failure eventually occurs.
The aim of the study was to measure in-vivo the 10-year linear and volumetric polyethylene wear of a fully congruent mobile bearing unicompartmental knee arthroplasty (OUKA). We studied six OUKA’s that had all been implanted 10 years previously. Each patient was examined in even double leg stance at a range of knee flexion angles, in a calibration cage. A stereo pair of X-ray films was acquired for each patient at 0°, 15° and 30° of flexion. The films were analysed using an RSA style calibration and a CAD model silhouette-fitting technique. The position and orientation of each femoral and tibial component was found relative to each other and the bearing position inferred. Penetration of the femoral component into the original volume of the bearing was our estimate of linear wear. The volumetric wear is calculated from the measured linear wear and the known surface area of the bearing. In addition eight control patients were examined less than 3 weeks post-operation where no wear would be expected.and 30We studied seven OUKA’s that had all been implanted at least 10 years previously. A stereo pair of X-ray films was acquired for each patient at 0 Results: The control group showed no measured wear. The seven OUKA’s had an average maximum depth of linear penetration of 0.40 mm at a mean follow-up of 10.9 years. The linear wear rate was 0.033 mm/year. The volumetric material loss was 79.8 mm3. If a steady gradual material loss is assumed, 8 mm3 of UHMWPE was lost per year. Polyethylene particulate wear debris continues to be implicated in the aetiology of component loosening and implant failure knee following arthroplasty. The OUKA employs a spherical femoral component and a fully congruous meniscal bearing to increase contact area and theoretically reduce the potential for polyethylene wear. The results from this in-vivo study confirm that the device has low 10-year linear and volumetric wear in clinical practice. This may protect the device from component loosening in the long term.
When the Oxford unicompartmental meniscal bearing arthroplasty (UCA) is used in the lateral compartment of the knee 10% of the bearings dislocate. An in-vitro cadaveric study was performed to investigate if the anatomy and joint geometry of the lateral compartment was a contributory factor in bearing dislocation. More specifically, the study investigated if the soft tissue tension of the lateral compartment, as determined by the length of the lateral collateral ligament (LCL), was related to bearing dislocation. A change in length of greater than 2 mm is sufficient to allow the bearing to dislocate. The Vicon Motion Analysis System (Oxford Metrics, Oxford, UK) was used to assess length changes in the LCL of seven cadaveric knees. Measurements were made of the LCL length through knee flexion and of the change in LCL length when a varus force was applied at a fixed flexion angle. Measurements were made in the normal knee and with the knee implanted with the Oxford prosthesis. In the intact knee the mean LCL change was 5.5mm (8%) over the flexion range. After implantation with the Oxford UCA the mean change in length was only 1 mm (1%). There was a significant difference in the LCL length at 90° (p=0.03) and 135° (p=0.01) of knee flexion compared to the intact knee. When a varus force was applied the LCL length change of the intact knee (5.4 mm) was significantly different (p=0.02) to that of the knee with the prosthesis implanted (2.7 mm) This study used a new method to dynamically measure LCL length. It found that after implantation of the Oxford lateral UCA the LCL remains isometric over the flexion range and does not slacken in flexion as it in the normal knee. This would suggest that the soft tissue tension was adequate to contain the bearing within the joint. However, when a varus force was applied the LCL did not sufficiently resist a displacing force producing an LCL length change greater than 2 mm. The evidence provided by this study is contradictory. The “lack of change in LCL length through flexion” suggests that the ligament remains tight through range and is unlikely to allow dislocation. However, the amount of distraction possible when an adducting moment is applied is sufficient to allow bearing dislocation. The length tension properties of the lateral structures are therefore implicated in the mechanism of dislocation.
The indications for unicompartmental knee arthroplasty (UKA) can be contentious. Concerns exist about implanting a UKA in younger individuals. It has been reported that it is contra-indicated in patients under the age of 60. The suggestion is that younger patients may compromise their outcome after surgery by their increased activity levels. However, the number of publications with data on UKA in young patients is minimal and the age/activity related criteria for UKA remains unknown. The aim of the study was 1) to confirm that younger UKA patients have higher activity levels than older patients, and 2) to test the hypothesis that younger, more active patients have inferior outcome when compared to older less active patients. Fifty consecutive patients undergoing UKA who were under 60 years (Group Y) and 50 patients over 60 years (Group O) were recruited. Indications were anteromedial arthritis with full thickness lateral compartment cartilage, a functioning ACL and a correctible varus. All patients underwent Oxford UKA using the minimally invasive technique. This device employs an unconstrained fully congruous meniscal bearing. Outcome was evaluated using Knee Society Scores. Activity level was documented using Tegner Activity Score. Average age of patients for group Y and O was 55 and 68 years respectively. Minimum follw-up was 2 years. pre-op. knee scores were comparable for each group. (y: 38.8, O: 35.8) and patients in both groups siginificantly improved after surgery. A significant difference in Tegner score was found between the two groups after surgery (y:4.1, O:2.6). More than 50% of younger patients regularly participated in high demand activities like skiing, tennis and manual labour. No statistically significant differences in any other post op. knee scores or complication rate were found despite adequate study power. Younger patients have increased activity levels after UKA when compared to an older age group. There is no evidence that the outcome of UKA in younger more active patients will be inferior to those who are older and more sedentary.
The Birmingham reSurfacing Arthroplasty (BSA) is a metal on metal prosthesis with no published independent clinical studies. Despite this, it is increasing in popularity, especially as an alternative to stemmed prostheses in younger patients. This study presents the 1year migration results of the BSA femoral component using Roentgen Stereophotogrammteric Analysis (RSA). Twenty six subjects underwent a BSA, through the postero-lateral approach using CMW3G cement, with RSA marker balls placed intra-operatively. The femoral component migration was measured at intervals of 3, 6 and 12 months using the Oxford RSA system. Geometric algorhythms were used to identify the femoral component. The data was examined for distribution prior to analysis. All statistical analysis was performed using the t-test. The data was normally distributed. The 1 year migration results of the BSA femoral component are displayed below. All cemented implants migrate in vivo. The majority of cemented stemmed implant migration occurs within the first post-operative year. High rates of migration within the first post-operative year correlate with premature component failure in some instances. The BSA is a fundamentally different design to most cemented prostheses, despite this we know that very large migrations, those in excess of 2mm/year in any direction are generally regarded as poor indicators of long term outcome. These results suggest that the BSA femoral component is an inherently stable device as it does not migrate significantly within the first post-operative year. Only long-term independent clinical studies and continued RSA follow-up will enable a comprehensive evaluation of the device.
This study examined the effect of surgical approach on the 1year migration of the Exeter stem, using Roentgen Stereophotogrammetric Analysis (RSA). There is evidence that implants with increased early migration, particularly those moving into internal rotation, are likely to have a higher failure rate. A total of 46 patients awaiting THR were allocated into 2 groups. Both groups underwent an Exeter Total Hip replacement, one through the Posterior Approach (PA), and the other through the Hardinge Approach (HA). RSA was used to assess post-operative stem position and migration at 1 year. Post-operative gait analysis and clinical evaluation was used to assess a subgroup of patients. There was no difference in the initial stem position of each group within the femoral canal. Significant differences in migration were found for migration of the head distally and posteriorly and for the tip medially. During gait, the legs of the PA group were relatively internally rotated when compared to those of the HA group. Relative to the supporting bone, the PA implants internally rotate and also rotate into valgus further than the HA implants. The starting position of the prostheses was the same and the cement and prosthesis characteristics were similar. From this, we can infer that surgical approach and therefore muscle function, have a large influence on component migration. During gait, the PA group had internally rotated legs compared to the HA group. This is probably due to short external rotator weakness in the PA group. As a result the internally rotating component of the posteriorly directed joint reaction force on the femoral stem will be larger. This probably explains the rapid internal rotation seen in the implants of the PA group. With the HA group, because of abductor damage there is likely to be a smaller component of the joint reaction force acting on the stem. This would explain the smaller rates of observed coronal plain rotation (rotation into valgus). Profound differences exist in early migration, between the HA and PA. Different muscle function may account for this and probably influences long-term outcome.
Polyethylene wear can be an important cause of knee replacement failure. Six TKRs in young, active patients with excellent Oxford Knee Scores and Knee Society Scores, mean 76 months post knee replacement and 5 control patients, 2 weeks post TKR, were selected. Each patient had weight bearing stereo radiographs of at 0, 15, 30, 45 and 60 degrees of flexion while standing in a calibration grid. These x-rays were analysed using our Radio Stereometric Analysis (RSA) system. The three-dimensional shape of the TKR (manufacturer’s computer aided design model) was matched to the TKR silhouette on the calibrated stereo radiographs for each angle of flexion. The relative positions of the femoral and tibial components in space were then determined and the linear and volumetric penetration was calculated using Matlab software. The accuracy of the system was found to be 0.3mm (CAD model tolerance 0.25mm). The mean linear wear in the control patients was 0.02mm (range −0.19 to +0.23mm). Average linear penetration in the study group was found to be 0.6 mm at 6 years, giving an overall linear wear rate of 0.1mm/year. Average penetration volume at 76 months was 399mm3. The average volumetric wear rate was 63mm3/year. It is possible to measure volumetric wear in vivo using RSA. Volumetric wear rate was found to be 63mm3 per year. Studies on retrieved normally functioning hip replacements have shown volumetric wear rates of 35mm3 per year. However, clinical outcomes of knee replacements are comparable to those of hip replacements, suggesting that the knee has a more effective mechanism for dealing with polyethylene wear particles.
Unicompartmental knee replacements (UKR) have many advantages over total replacements (TKR), including better function and less morbidity. However, in general, they have a higher failure rate. To minimise the UKR failure rate it is essential that the implant should not wear out, and that the appropriate indications and surgical techniques are used. The Oxford UKR has a fully congruent mobile bearing and has been shown in a retrieval study to have minimal wear. The indications for the use of Oxford UKR are clearly defined. It is recommended for medial compartment osteoarthritis, with a functionally intact Anterior Cruciate Ligament. The Varus deformity should be correctable and there should be full thickness cartilage in the lateral compartment, which is best demonstrated on a valgus stress radiograph. It is appropriate for about one in four osteoarthritic knees needing replacement. The designer, Mr Goodfellow, achieved a 98% (CI 93% to 100%) survival at 10 years, using the appropriate indications. However, data from the designer is open to bias. An independent series of 420 Oxford UKR from Dr Svard achieved a 94% (CI 86% to 100%) survival at 15 years, with no loss to follow-up. In the Swedish Knee Arthroplasty Register, in centres implanting at least 2 UKR per month, the survival rate of the Oxford UKR was 93% at 8 years. In centres doing very few UKR the failure rate was higher. These poor results were probably because of inappropriate indications or technique. To address the problem of inconsistent results new simplified instrumentation (Phase 3) has recently been introduced. This instrumentation has been specifically designed for a minimally invasive approach. Patients recover three times quicker after minimally invasive UKR than after TKR. A Radiographic comparison demonstrated the Oxford UKR can be implanted as precisely through a short incision as through a standard incision. A fluoroscopic study demonstrated that knee kinematics after minimally invasive UKR are virtually normal and are substantially better than after TKR. At one year the first 58 minimally invasive Oxford (Phase 3) UKR implanted by a single surgeon had an average flexion of 135°, Knee Society knee score of 97 and function score of 92. We conclude that UKR is the treatment of choice for medial compartment osteoarthritis provided appropriate implants, indications and surgical techniques are used.
Results at one year (TMK first): AKSS(Knee) 91.6 / 84.1 (p=0.003), OKS 39.8 / 37.6 (p=0.006), ROM 104 / 104 (p=0.364), Pain (AKSS) 47.3 / 41.7 (p=0.01), Pain (OKS) 3.5 / 2.9 (p=0.006).
Method: Six TKRs in young, active patients with excellent Oxford Knee Scores and Knee Society Scores, mean 76 months post knee replacement and 5 control patients, 2 weeks post TKR, were selected. Each patient had weight bearing stereo radiographs of at 0, 15, 30, 45 and 60 degrees of flexion while standing in a calibration grid. These x-rays were analysed using our Radio Stereometric Analysis (RSA) system. The three-dimensional shape of the TKR (manufacturer’s computer aided design model) was matched to the TKR silhouette on the calibrated stereo radiographs for each angle of flexion. The relative positions of the femoral and tibial components in space were then determined and the linear and volumetric penetration was calculated using Matlab software.
*Oxford Hip and Knee Group: P McLardy-Smith, C Dodd, D Murray &
R Gundle
The Oxford hip score (OHS) is a patient-based instrument for assessment of outcome which is often used after total hip replacement, and the EuroQol 5D (EQ5D) is a patient-based generic questionnaire for health assessment. In an analysis of the outcome at one year of 609 revision hip replacements (RHRs), we compared the OHS and EQ5D scores, postoperative patient satisfaction and change in pain. About 25% of the operations were repeat RHRs. At one year, 57% of patients were very pleased with their operation. The correlation between preoperative and postoperative scores and change scores for the OHS and EQ5D was high. For both instruments the effect sizes were large, but the greater effect size of the OHS suggests that it is particularly sensitive to improvements after RHR. The effect scores of the OHS declined with the number of previous RHRs, while those for the EQ5D seemed less sensitive. Our results confirm the value of the OHS in assessing outcome after RHR.
We have developed a 12-item questionnaire for patients having a total knee replacement (TKR). We made a prospective study of 117 patients before operation and at follow-up six months later, asking them to complete the new questionnaire and the form SF36. Some also filled in the Stanford Health Assessment Questionnaire (HAQ). An orthopaedic surgeon completed the American Knee Society (AKS) clinical score. The single score derived from the new questionnaire had high internal consistency, and its reproducibility, examined by test-retest reliability, was found to be satisfactory. Its validity was established by obtaining significant correlations in the expected direction with the AKS scores and the relevant parts of the SF36 and HAQ. Sensitivity to change was assessed by analysing the differences between the preoperative scores and those at follow-up. We also compared change in scores with the patients’ retrospective judgement of change in their condition. The effect size for the new questionnaire compared favourably with those for the relevant parts of the SF36. The change scores for the new knee questionnaire were significantly greater (p <
0.0001) for patients who reported the most improvement in their condition. The new questionnaire provides a measure of outcome for TKR that is short, practical, reliable, valid and sensitive to clinically important changes over time.
We developed a 12-item questionnaire for completion by patients having total hip replacement (THR). A prospective study of 220 patients was undertaken before operation and at follow-up six months later. Each completed the new questionnaire as well as the SF36, and some the Arthritis Impact Measurement Scales (AIMS). An orthopaedic surgeon assessed the Charnley hip score. The single score derived from the questionnaire had a high internal consistency. Reproducibility was examined by test-retest reliability and was found to be satisfactory. The validity of the questionnaire was established by obtaining significant correlation in the expected direction with the Charnley scores and relevant scales of the SF36 and the AIMS. Sensitivity to change was assessed by analysing the differences between the preoperative scores and those at the follow-up. The standardised effect size for the new questionnaire compared favourably with that for the SF36 and the AIMS. The new questionnaire provides a measure of outcome for THR which is short, practical, reliable, valid and sensitive to clinically important changes.
To assist surgeons to select a total hip replacement (THR) we present comparative information on all such implants on the market in the UK. We identified 62 different primary THRs, manufactured by 19 companies; half had been introduced in the last five years, and only 30% have any results published in peer-reviewed journals. The prices range from 250 pounds to 2000 pounds, and the two cheapest implants have the longest reported follow-up. The number of THR implants available in the UK, and presumably the rest of the world, is rapidly increasing, but there is little or no scientific evidence that the newer, more expensive, implants are better than established designs. Some will undoubtedly be worse. We believe that this situation is unsatisfactory and make recommendations for improvement, in particular that preference be given to implants with good results in published peer-reviewed long-term clinical trials.
Survival analysis is a powerful tool for analysing the results of total joint replacement, but it has major drawbacks when the failure rates are very low. We have reviewed 35 recent survival analyses of joint replacements to assess the magnitude of these problems and make recommendations as to how they may be avoided.
After late reduction of the hip in children with developmental dysplasia the acetabular contour slowly improves and it is difficult to know if and when acetabular reconstruction is required. We studied the radiographs of 19 patients with unilateral dislocated or subluxated hips which had been reduced between the ages of one and two years. Preoperatively, all the affected hips showed acetabular dysplasia. After reduction they steadily improved for three years by which time none was dysplastic as measured by the acetabular index. After the age of ten years, when assessed by the more sensitive centre-edge angle, two were found to be dysplastic. It was not possible to predict these from early radiographs. We conclude that the decision to reconstruct an acetabulum should not be taken until three years after reduction, and that a few hips which appear to be developing satisfactorily at that time will, nevertheless, become dysplastic.
One-stage femoral lengthening is thought to have an unacceptably high complication rate and is not widely practised. We reviewed 17 patients after one-stage lengthening for femoral shortening with associated angular or rotational deformities. Minimal dissection of the bone ends was undertaken. The mean length gain was 4 cm (2 to 7), and the average time to union was 6 months (3 to 10). There were no neurovascular complications. Four patients had delayed or nonunion, but union was achieved after bone grafting. We conclude that with minimal dissection, and with iliac crest cancellous bone grafting, one-stage leg lengthening for correction of deformity and leg-length inequality of up to 7 cm, in selected patients, can be effected safely with a relatively short rehabilitation.
The orientation of an acetabulum or an acetabular prosthesis may be described by its inclination and anteversion. Orientation can be assessed anatomically, radiographically, and by direct observation at operation. The angles of inclination and anteversion determined by these three methods differ because they have different spatial arrangements. There are therefore three distinct definitions of inclination and anteversion. This paper analyses the differences between the definitions and provides nomograms to convert from one to another. It is recommended that the operative definitions be used to describe the orientation of prostheses and that the anatomical definitions be used for dysplastic acetabula.
Long posterior wall (LPW) Charnley acetabular implants are widely used as it is believed that the LPW helps to prevent dislocation. This has, however, not been proven statistically. In a preliminary study of these implants removed at revision marked erosion of the LPW was frequently seen, indicating that repetitive impingement may occur. The influence of the long posterior wall was therefore investigated mathematically. LPW and standard sockets were found to be equally likely to dislocate provided that the standard socket was anteverted 5 degrees more than the LPW socket. With simulated external rotation, LPW sockets impinge 30% earlier than standard sockets. When impingement occurs a torque is applied to the components, which increases the shear stresses at the cement-bone interface. The torques, although not large enough to dislodge the socket immediately, are repetitive and so may contribute to loosening. The LPW socket can generate twice as much torque as the standard socket and therefore is more likely to loosen.
We investigated in vitro a mechanism by which particulate debris may induce bone resorption and cause implant loosening. We first studied two standard particles: latex, which is considered to be inert, and zymosan, which is inflammatory. Macrophages that phagocytosed either particle became activated, and stimulated 15 times as much bone resorption as did control macrophages. For activation to occur, 100 times more latex than zymosan had to be phagocytosed. We also found that bone cement and polyethylene particles activated macrophages in a similar manner, and that the necessary amounts of these were intermediate between those of latex and zymosan. None of the particles were toxic. It was concluded that implant loosening may result from bone resorption stimulated by mediators released by macrophages that have phagocytosed particles of bone cement or polyethylene.
We investigated the possibility that the macrophages which are seen around implants may stimulate bone resorption and cause loosening. We found that macrophages release mediators that stimulate bone resorption, and that the amount of resorption increased by between 2.5 and 10 times when the macrophages adhered to a foreign surface. This bone resorption depended on the surface energy and roughness of the foreign surface, varying with these physical properties rather than with the chemical nature of the material. It is concluded that loosening of orthopaedic implants is likely to be influenced by the surface energy and roughness of the implant.
The results of treatment of injuries of the thoracic and thoraco-lumbar spine with neurological involvement have been reviewed in a retrospective study of 115 patients, of whom eighty-nine received conservative and twenty-six surgical treatment. Operation was reserved, in general, for patients with irreducible dislocations and incomplete neurological lesions, open reduction and internal fixation being the commonest procedure. Only three patients required a delayed spinal fusion for suspected instability after a period of conservative treatment. On the other hand, ten patients, eight of whom had been treated surgically, were left with severe chronic spinal pain. Of the patients treated conservatively, 35% showed significant neurological improvement compared to 38% of those treated surgically, but the latter group contained a much higher proportion of incomplete lesions with a far better prognosis. It is concluded that the place for early operation might be still further restricted.