Introduction

Femorotibial malalignment exceeding ±3° is a recognised contributor of early mechanical failure after total knee replacement (TKR). The angle between the mechanical and anatomical axes of the femur remains the best guide to restore alignment. We investigated where the femoral head lies relative to the pelvis and how its position varies with respect to recognised demographic and anatomic parameters. We have tested the hypothesis of the senior author that the position of the centre of the femoral head varies very little, and if its location can be identified, it could serve to outline the mechanical axis of the femur without the need for sophisticated imaging.

Purpose

To examine the clinical characteristics of patients undergoing knee arthroplasty with a pre-operative Oxford Knee Score >34 (‘good’/‘excellent’), and assess the appropriateness of surgical intervention for this group.

Purpose

The Oxford Knee Score (OKS) is a validated and widely used PROM that has been successfully used in assessing the outcome of knee arthroplasty (KA). It has been adopted as the nationally agreed outcome measure for this procedure and is now routinely collected. Increasingly, it is being used on an individual patient basis as a pre-operative measure of osteoarthritis and the need for joint replacement, despite not being validated for this use. The aim of this paper is to present evidence that challenges this new role for the OKS.

The Oxford Knee Score (OKS) is a valid and reliable self-administered patient questionnaire that enables assessment of the outcome following total knee replacement (TKR). There is as yet no literature on the behavioral trends of the OKS over time. Our aim is to present a retrospective audit of the OKS for patients who have undergone TKR during the past ten years.

We retrospectively analysed 3276 OKS of patients who had a primary TKR and had been registered as part of a multi-surgeon, outcome-monitoring program at St. Helier hospital. The OKS was gathered pre-operatively and post-operatively by means of postal questionnaires at annual intervals. Patients were grouped as per their age at operation into four groups: 60, 61-70, 71- 80 and >80. A cross-sectional analysis of OKS at different time points was performed.

The numbers of OKS available for analysis were 504 pre-operatively, 589 at one-year, 512 at two-year and gradually decreasing numbers with 87 knees ten-year post-operatively. There was as expected a significant decrease (improvement) of the OKS between pre-operative and one-year post-operative period and then reached a plateau. Beyond eight years, there is a gradual rise in the score (deterioration). The younger patients (60) showed a significant increase in their average OKS between one and five-years post-operatively. However beyond five years, they followed the trend of their older counterparts. When the twelve questions in the OKS were analysed, certain components revealed greater improvement (e.g. description of knee pain and limping) than others (e.g. night pain).

The OKS is seen to plateau a year after TKR. According to the OKS the outcome of the TKR is not as good in the younger age group as compared to the older age group. Further investigation is required to ascertain the cause of this observed difference.

The performance of total knee arthroplasty in deeply flexed postures is of increasing concern as the procedure is performed on younger, more physically active and more culturally diverse populations. Several implant design factors, including tibiofemoral conformity, tibial slope and posterior condylar geometry have been shown directly to affect deep flexion performance.

The goal of this study was to evaluate the performance of a fixed-bearing, asymmetric, medial rotation arthroplasty design during kneeling activities.

Thirteen study participants (15 knees) with primary total knee arthroplasty (Medial Rotation Knee, Finsbury, Surrey, UK) were observed while doing a step activity and kneeling on a padded bench from 90° to maximum comfortable flexion using lateral fluoroscopy. Subjects averaged 74 years of age and nine were female. Subjects were an average of 17 months post-operative, and scored 94 points on the International Knee Score and 99 on the Functional Score. Digitised fluoroscopic images were corrected for geometric distortion and 3D models of the implant components were registered to determine the 3D position and orientation of the implants in each image.

During the step activity, the medial and the lateral femoral contact point stayed fairly constant with no axial rotation from 0 to 100° of flexion. At maximum kneeling flexion, the knees exhibited 119° of implant flexion (101°-139°), 7° (-7° to 17°) tibial internal rotation, and the lateral condyle translated backwards by 11 mm.

Patients with medial rotation knee arthroplasty exhibited medial pivot action with no paradoxical translation. The knees exhibited excellent kneeling flexion and posterior translation of the femur with respect to the tibia. The axial rotation in MRK was within the range of normal knee kinematics from -10 to 120 (perhaps 140).

Aims

To establish the cost of primary hip (THR) and knee (TKR) arthroplasty in an elective orthopaedic centre in the UK and to compare it with current government reimbursement to NHS hospitals and the costs in North America.

Introduction: Increased cell senescence has been reported in the human intervertebral disc (IVD) and was associated with degenerative pathology, particularly herniation. Increased IVD innervation and blood vessel ingrowth is associated with disc degeneration and the development of back pain. This preliminary study examines whether there is a relationship between the prevalence of senescent IVD cells and the extent to which the tissue is innervated and/or vascularised.

Methods: Specimens of herniated IVD (n=16 patients: aged 36–71) were stained for senescence associated β-galactosidase activity (SA β-gal), then snap frozen and cryosectioned prior to immunolocalisation procedures to detect nerves (NF200) or blood vessels (CD34). Stained sections were counterstained with DAPI to reveal cell nuclei. The proportion of SA β-gal +ve cells was scored and the extent of neural and blood vessel ingrowth semi-quantitated.

Results: The proportion of SA β gal +ve IVD cells ranged from 6% – 91% (median=16%) and was significantly correlated with age. The degree of neural or blood vessel ingrowth ranged from tissue which contained numerous (i.e. ≥10) positive cells/cell processes to tissue which was completely aneural or avascular. However, there was no clear relationship between the presence of SA β-gal +ve IVD cells and IVD innervation or vascularisation.

Conclusions: Cell senescence has been associated with up-regulated expression of catabolic enzymes, e.g. MMPs and increased synthesis of trophic cytokines, e.g. VEGF. Such cellular activity might by thought to contribute to the pathological ingrowth of nerves or blood vessels into the IVD. The data presented here, however, does not support such a hypothesis.

Conflicts of Interest: None

Source of Funding: Institute of Orthopaedics, RJAH Orthopaedic Hospital

Introduction: Intervertebral disc (IVD) cell transplantation is used to treat back pain. However, IVD cell activity may also contribute to pathology, e.g. IVD cells can undergo senescence or promote nerve growth, which in the IVD is associated with discogenic back pain. Serum deprivation of bovine IVD cells results in cell senescence. We have examined the influence of oxygen supply combined with serum deprivation on human IVD cells.

Methods: Cells from herniated IVD (n=3 patients) were subjected to serum deprivation and then cultured under hypoxic (1%) or atmospheric (21%) conditions for 10 days. IVD cell growth, viability and cell senescence (via Senescence Associated β-galactosidase activity; SA β-gal) were examined. The growth and migration of HMEC-1 (endothelial) and SH-SY5Y (neuronal) cells treated with conditioned medium from the IVD cell cultures (1% versus 21% oxygen) were subsequently monitored.

Results: Hypoxia significantly decreased IVD cell proliferation, but was also found to reduce cell senescence. Hence, the proportions of SA β-gal positive IVD cells in 1% and 21% oxygen at day 10 were 18±6% and 56±10%, respectively. There was no marked difference in cell viability (> 95%). Conditioned medium from IVD cells cultured under hypoxia stimulated endothelial and neural cell growth (determined via the MTS assay) and endothelial cell migration and neurite outgrowth to an extent that was significantly greater than conditioned medium from IVD cells cultured at 21% oxygen.

Conclusions: The trophic activity of human IVD cells is responsive to oxygen supply. However, hypoxia may influence the capacity of IVD cells to reduce back pain for better or worse.

Conflicts of Interest: None

Source of Funding: Institute of Orthopaedics, RJAH Orthopaedic Hospital.

Background: Tranexamic acid has been shown to be effective in reducing blood loss and transfusion requirement in cardiac surgery and total knee replacement surgery. The most effective dose of tranexamic acid in hip arthroplasty surgery is not yet known. We investigated the effect of a pre-operative bolus 1g intravenous tranexamic acid on intra- and post-operative blood loss, transfusion requirement, and risk of venous thromboembolism following total hip arthroplasty.

Results: We report a cohort comparison study of 73 patients who underwent primary hip arthroplasty. Thirty-six patients received tranexamic acid (TA group), and thirty-seven received no tranexamic acid (control group). Blood loss was measured directly intra-operatively, and indirectly post-operatively by haemoglobin and haematocrit measurement. Deep vein thrombosis (DVT) was investigated by venography.

Patient demographics were similar between both groups. There was no significant difference in intra-operative blood loss between both groups. The early post-operative blood loss and total blood loss were significantly less in the tranexamic acid group. This effect of tranexamic acid was more significant in females who showed a dose-related relationship between tranexamic acid dose and blood loss. Fewer patients in the tranexamic acid group required blood transfusion. There was no increased incidence of DVT in the tranexamic acid group.

The use of a single pre-operative 1g bolus of tranexamic acid administered before surgery is a safe, cost-effective method of reducing post-operative blood loss following total hip arthroplasty. The effect is more significant in females at this dose.

Aim: To assess the functional outcome and longevity of patients who are mentally competent when they sustain an intra-capsular fracture of the femoral neck.

Methods: Prospective cohort study of fifty female patients over the age of seventy years of age with a displaced intra-capsular fracture of the femoral neck. Patients with known, cognitive impairment terminal illness or active infection were excluded from the study. Patients were treated with a large head total hip replacement. Functional outcome was assessed using the the Barthel index score and the Charnley-modified Merle d’Aubigne score yearly for five years. The Oxford hip score was also recorded annually from three to ten years post op. The date and cause of death were recorded from hospital notes and death certificates.

Results: The mean age at time of femoral neck fracture was 81 years. Three patients died before discharge from hospital. The mortality rates were 18% at twelve months, 28% at two years, 44% at five years and 92% at ten years. Those patients surviving had mean Oxford scores of 24 at 3 years, 26 at 5 years, 23 at 7 years and 25 at ten years. The cause of death was ischaemic heart disease in 22%, cancer in 10%, bronchopneumonia in 12% and simply ‘old age’ in 14%.

Conclusion: Elderly female patients who are not cognitively impaired at the time of displaced intra-capsular fracture of the femoral neck have a 56% chance of living five years. Function scores in the survivors are maintained, and are better than would be expected if the patients had undergone hip hemiarthroplasty. Total hip replacement in this population may be justified.

Aims: Minimum ten year clinical, radiological and postmortem results of the flexible Cambridge Cup. This acetabular prosthesis was designed to replace the horseshoe shaped articular cartilage and provide physiological loading with minimal resection of healthy bone.

Method: Fifty female patients aged over 70 years with a displaced intra-capsular fracture of the femoral neck were recruited to the ethically-approved prospective study. They underwent implantation of the Cambridge Cup, which has an outer polybutyleneterephthalate shell and an inner UHMWPE bearing, with a Thompson-type hemiarthroplasty. The cups were manufactured with a 60μm plasma-sprayed coating of hydroxyapatite. This was removed from half the cups to simulate the effects of long-term HA resorption. Implants were sterilised by gamma irradiation in air. Independent clinical and radiological assessments were undertaken before discharge, at 6 weeks, one year, 18 months, two -, five-, seven- and ten years. Patients were scored using the Barthel Index, the Charnley-modified Merle d’Aubigne scores and latterly the Oxford hip score. The date and cause of death were obtained from hospital records and death certificates. Fifteen Cambridge Cups were retrieved post-mortem for histological and wear analysis

Results: The mean functional scores recovered to levels before fracture. These scores decreased with advancing age at five years. The mortality rates were 16%, 28%, 46% and 92% at 1, 2, 5 and 10 years. The Oxford hip scores in patients surviving between five and ten years were maintained.

The HA coated implants remained asymptomatic. Three uncoated components required revision for migration. No evidence of accelerated UHMWPE wear was seen on retrievals or radiographs. Histological analysis of the retrieved HA coated specimens showed excellent bony fixation, uncoated cups showed predominantly fibrous tissue.

Conclusion: The uncemented Cambridge Cup was implanted in a challenging environment of osteoporotic bone. Clinical, radiographic and post-mortem results up to ten years are excellent.

Publication of normal and expected outcome scores is necessary to provide a benchmark for auditing purposes following arthroplasty surgery. We have used the Oxford knee score to monitor the progress of knee replacements undertaken since 1995, the start of our review programme. 4847 Oxford assessments were analysed over an 8 year follow-up period.

The mean pre-operative Oxford knee score was 39.2, all post-operative reviews showed a significant improvement. Patients with a BMI > 40, and the under 50 age group showed early deterioration in outcome scores, returning to pre-operative levels by 5 and 7 years respectively. There was no significant difference in outcome between surgeons performing < 20 knee replacements a year and those performing > 100 / year.

The age of the patient at the time of surgery and the pre-operative body habitus have been identified as factors affecting long term outcome of total knee replacement surgery. Awareness of these factors may assist surgeons in advising patients of their expected outcomes following surgery.

Sectioned femoral components retrieved from failed hip resurfacing arthroplasties show resorption of proximal femoral bone or formation of a fibrous membrane at the bone cement interface. Our study uses Finite Element Analysis (FEA) to examine the effects of the implant orientation on bone remodelling following hip resurfacing arthroplasty. A radiographic analysis of the proximal femur following hip resurfacing was conducted in order to draw a comparison to the FEA findings.

A 3D FEA model of the Birmingham Hip Resurfacing (BHR) was created based on the geometry and material properties of a 45 year-old female donor hip. Hip joint and muscle loads were applied. Bone remodelling stimuli was determined using changes in strain energy. A range of implant orientations were compared to study the affect on bone remodelling. A retrospective radiological analysis was undertaken on 100 hips with a minimum of 5 years follow up. Femoral neck diameter was measured at post-op, 2 and 5 years, as well as neck and stem shaft angles.

FEA showed that valgus orientation was associated with increased resorption underneath the shell. Varus orientation showed increased bone formation at the stem tip. The radiological analysis identified 2 distinct patterns of neck thinning. Slow thinners (76%) had < 5% reduction in neck diameter at 2 years and < 10% at 5 years. Rapid thinner (24%) had > 5% thinning at 2 years and > 10% at 5 years. The mean percentage reduction in neck diameter was significantly different between the two groups at the two time points (p< 0.01). The rapid group had a higher proportion of valgus aligned implants (88%) and a significant decrease in reconstructed offset (p=0.0023).

The FEA results have shown that stem alignment can affect bone resorption resurfacing. FEA results were consistent with the radiological findings. Additional retrieval studies are necessary to help understand aetiology of implant failures.

Adverse bone remodeling in the proximal femur may be detrimental to the long term survival of resurfacing prosthesis. Bone resorption beneath the femoral shell and thinning of the femoral neck have been observed. We present a radiological analysis of the incidence, rate, site of neck thinning and changes observed within the femoral neck, in 100 cases, with a minimum of five years follow-up.

Femoral neck diameter was measured at zero, two and five years post-operatively, at the head neck junction and five mm distally. Pre and post-operative head to neck ratios, natural and reconstructed offset, femoral neck-shaft and stem-shaft angles and cup inclination angle were measured.

Two distinct patterns of neck thinning were observed. In 76 cases (slow thinning group), we observed a reduction of < 5% of original neck diameter at two years and < 10 % at five years (mean 1.5%, sd+/− 1.5). In 24 cases (rapid thinning group), a reduction of > 5% of original neck diameter at two years and > 10% at five years (mean 10.4%, sd+/− 4.8) was observed. The difference in the percentage reduction in neck diameter was significantly different between the two groups at both time points (p< 0.01). Larger head-neck ratios were observed in the rapid thinning group, both pre and post operatively (p< 0.01).

The viability of bone underneath the femoral head may be compromised as a consequence of a non-physiological bone loading mechanism. FEA has predicted stress shielding underneath the femoral head and loading of the mini stem. Compromised blood supply of the retained epiphyseal remnant may play a part in femoral head resorption.

Femoral neck thinning is a phenomenon of unproven aetiology which affects almost 25% of our resurfacing cases.

Introduction: Sectioned femoral components retrieved from failed hip resurfacing arthroplasties show resorption of proximal femoral bone or formation of a fibrous membrane at the bone cement interface, in a proportion of cases. We hypothesize that both scenarios create a functional discontinuity zone (FDZ), which exacerbates offloading the proximal bone and promoting resorption. Our study uses finite element modeling to examine the effects of the presence of an FDZ on bone remodeling following hip resurfacing arthroplasty. A radiographic analysis of the proximal femur following hip resurfacing was also conducted in order to draw a comparison to clinical findings.

Methods: The hip resurfacing FE models were oriented in 3 distinct stem-shaft angles: 136 ‘neutral’, 120 ‘varus’ and 150 ‘valgus’. A low-modulus (2 MPa) FDZ (approx. 2 mm thick) was simulated beneath the implant head. Femoral joint and muscle loads were applied to simulate peak joint loading during gait. Interface stress was compared for the normal and simulated FDZ resurfaced femurs. Bone remodeling stimuli was determined using changes in strain energy. A retrospective radiological analysis was undertaken on 98 hips (74 males and 24 females) with a minimum of 5 years follow up. Measurements of the prosthesis-shaft angle, pre–and post-operative femoral head offset and femoral neck diameter at 2 and 5 years were undertaken.

Results: The presence of the simulated FDZ in the FE analysis resulted in increased proximal-medial bone resorption and slightly greater bone formation surrounding the stem. Correspondingly, device-bone interface stresses were found to decrease proximally under the loading platform and increase at the stem, particularly adjacent to the stem-head junction. The valgus BHR femur led to increased resorption, especially around the periphery of the neck and on the medial side. The radiological analysis identified 2 groups; 22 hips (Group 1) had a mean 5.61mm (sd 2.07) reduction in neck diameter over 5 years and 76 hips (Group 2) demonstrated slow reduction in neck diameter, mean 1.13mm (sd 0.97). Neck thinning at 2 and 5 years was significantly greater for Group 1 (p< 0.0001). Group 1 hips had significantly greater reduction in femoral offset (p=0.041), with greater valgus angle oriented components (p=0.09). Reduction in femoral offset was significantly associated with greater valgus orientations (p< 0.0001). The Group 1 revision rate was 36.4% compared to 2.6% in Group 2 (p< 0.0001).

Discussion: The FE results support the hypothesis that the presence of a FDZ decreases load transfer to the proximal bone, resulting in increased medial stress shielding and resorption. These results are consistent with the Group 1 clinical findings. In order to better understand the cause of implant failures in hip resurfacing arthroplasty, additional retrieval studies are necessary.

Aim: The aim of this study was to determine whether increasing obesity has an influence on the age at which joint replacements are undertaken at our centre.

Materials and Methods: The database was analyzed for age, oxford score and body mass index (BMI) at the time of surgery.

The patients were divided into 5 groups based on their BMI, BMI< 25(normal), 25–29.9(overweight), 30–34.9(obese), 35–39.9(moderately obese), 40 or more(morbidly obese).

BMI < 25 was treated as a control group for comparison.

Statistical analysis was done using t test.

Results: In total, 1369 patients were analyzed, 1025 with hip replacement and 344 with knee replacement The difference in mean oxford score at surgery was not statistically significant between the groups (p> .05). For those undergoing hip replacement, the mean age of morbidly obese was 10 years less as compared to those with BMI < 25. For those with knee replacement, the same difference was 13 years. The age at surgery fell as the BMI > 35 for both hip and knee replacement (p< . 05). This association was found to be stronger for patients with knee replacement than with hip replacement.

Conclusion: This study shows that there is a positive association between obesity and the age at which hip and replacements are required. Obese people with BMI > 35 are likely to require joint replacement at an earlier age as compared to people with BMI < 25. The age at which joint replacement in required falls as the BMI increases over 35. This association is strong for patients requiring total knee replacement, and moderate for patients with hip replacement.

The potential for bone remodeling in the proximal femur may be detrimental to the long term survival of resurfacing prosthesis.

A retrospective analysis of radiological changes in the femoral neck was undertaken for 96 patients (100 hips, 76 males and 24 females), with a minimum of 5 years following hip resurfacing. The mean age at surgery was 53.8 years. Femoral neck diameter was measured post-operatively, at 2 and 5 years. Pre and post-operative head to neck ratios, femoral head-shaft offset, femoral neck and implant stem-shaft angles were also measured.

Two groups of patients were identified with differing rates of reduction in their femoral neck diameter. Over the first 2 years, Group A (24%) mean reduction was 2.02mm/year while Group B (76%) mean reduction was 0.33 mm/year. At 5 years, the Group A mean reduction was 5.64mm (sd±2.03mm) while Group B reduction was 1.16mm, (sd±0.97mm). The difference was significant at both time points (p< 0.01). Larger head-neck ratios were observed in the group A, both pre and post operatively (p< 0.01).

Finite Element Analysis has predicted stress shielding underneath the femoral head and loading of the mini stem. This may explain bone resorption underneath the shell and remodeling around the mini stem. Compromised blood supply of the retained epiphyseal remnant may also play a part in femoral head resorption. Group A with a larger proportion of females and femoral heads will potentially have a larger proportion of epiphyseal remnant retained. A further mechanism that could be influential in the development of neck thinning and bone resorption may be due to fluid pumping mechanism causing osteolytic erosion at the bone cement interface.

In conclusion, femoral neck thinning is a phenomenon of unproven aetiology which is affecting almost 25% of our resurfacing cases. Further investigations are needed to determine its aetiology and remedy.

INTRODUCTION: The American Knee Society score (AKSS) and the Oxford Knee score (OKS) are widely used health outcome measures for total knee replacements. The AKSS is a surgeon-assessed, variable weighted knee score. The OKS is a patient assessed equally weighted score. Our aim was to evaluate whether patient self assessment is a viable alternative to clinical review and whether it can provide enough information to identify which patient would require a clinic visit.

As there had been no previous studies correlating the two scoring systems, we investigated whether a correlation exists between the two scores at 2, 5 and 10 year periods. A correlation would allow us to determine what OKS value would achieve 90% sensitivity in identifying patients requiring clinical review at the above time points. This strategy would reduce the number of clinical visits required and its associated cost.

METHODS: We reviewed the data gathered prospectively from January 2000 to April 2006. All patients were part of an ongoing multi-surgeon single institution Knee Arthroplasty Outcome Programme. Preoperative, 2, 5 and 10 year post-operative OKS and AKSS were gathered from different cohorts. This method of comparison has been validated by previous publications. The scores were then analyzed using the Pearson correlation and linear regression. Different OKS values were analyzed for sensitivity and specificity.

RESULTS: 175 patients completed both the OKS and AKSS questionnaires preoperatively. 312 completed both scores at 2 years; 124 at 5 years and 57 patients at 10 years. The mean OKS, and the two AKSS components, the Knee score and Functional score improved significantly 2 years postoperatively when compared to their preoperative values. The Functional score deteriorated significantly from 5 to 10 years (p< 0.0001). There was good correlation between the OKS and the Knee score and Functional score at 2 years and a moderate correlation at 5 to 10 years. OKS > 24 showed more than 90 % sensitivity in identifying poor Knee scores in the 2, 5 and 10 year periods.

CONCLUSION: In this study, the good correlation of OKS and AKSS at 2-years suggests that postal Oxford questionnaire is sufficient in following up patients in the short term after total knee replacement. However, the moderate correlation at 5 and 10 years suggests that clinical evaluation is necessary.

We recommend that at 2 years, all patients complete an OKS questionnaire and if this is above 24, a clinical evaluation maybe required. Using this OKS value as a screening technique would allow a reduction of up to 50% in clinic visits and outpatient costs at the 2 year follow-up. This reduction is not as great at the 5 and 10 year periods. At these time periods, we recommend a clinical follow-up.

Aim: Our aim was to determine the in-hospital costs of total hip replacement (THR) and total knee replacement (TKR), and compare them to their reimbursement under the national tariff system and to our North American colleagues.

Methods: In 2004 an elective orthopaedic centre was set up in South West London which performs mainly primary lower limb arthroplasty. We used a retrospective analysis of financial statements from September 2004- December 2005 inclusive to establish operative costs (including implant), perioperative costs and post operative costs until discharge.

Results: A consecutive series of 1538 primary TKR patients (mean age = 72.2) and 1118 primary THR patients (mean age = 70.8) was studied from September 2004 to December 2005 accounting for 88% of the workload. Average post operative stay was 6.5 days for TKRs and 5.7 for THRs and 66% of the patients were ASA grade III or over. The cost including implant of a THR was £6054, and for a TKR it was £6499. After adjustment to allow direct comparison to tariff, our THR cost is 2.3% less than and 1.0% more than the 2004/5 and 2005/6 tariffs respectively. Our adjusted TKR cost is 5.0 % and 4.2% less than the 2004/5 and 2005/6 tariffs respectively. Our cemented and uncemented THR costs are 8.0% less than and 6.6% more than their respective tariffs introduced for the first time in 2005/6. Overall there is no large discrepancy between our THR/TKR costs and tariff reimbursements.

Discussion: Although in our unit our costs are similar to the re-imbursements used in the UK, we question the ability of general NHS hospitals in the UK to perform at these levels and prices for three reasons. Firstly, our high volume of joint replacement activity has enabled us to negotiate the most favourable implant prices in the UK. Secondly, length of stay in our unit is approximately 60% of national average. Thirdly, our unit is run without many of the infrastructure costs of a general hospital as well as the cost incurred by training junior staff and research and development. Costs were also found to be favourable compared with our American colleagues and similar to our Canadian colleagues.

Conclusion: Our elective only orthopaedic centre provides a cost effect way of performing primary arthroplasty surgery while maintaining high standards of care and twenty-four hour intensive care cover. We believe this cost effectiveness may be unachievable in general NHS hospitals in the UK.

Computer navigation assistance in total knee arthroplasty (TKA) results in more consistently accurate postoperative alignment of the knee prostheses. However the medium and long term clinical outcomes of computer-navigated TKA are not widely published. Our aim was to compare patient perceived outcomes between computer navigation assisted and conventional TKA using the Oxford knee score (OKS).

We retrospectively collected data on 441 primary TKA carried out by a single surgeon in a dedicated arthroplasty centre over a period of four years. These were divided according to use of computer navigation (group A) or standard instrumentation (group B). There were no statistical differences in baseline Oxford knee score (OKS) and demographic data between the groups. 238 of these had at least a one-year follow-up with 109 in group A and 129 in group B. Two year follow-up data was available for 105 knees with 48 in group A and 57 in group B and a three year follow-up for 45 with 21 and 24 in groups A and B respectively. 12 patients had completed four year follow-up with seven and five knees in groups A and B respectively.

The mean OKS at 1-year follow up was 24.98 (range 12– 54, SD 9.34) for group A and 26.54 (range 12– 51, SD 10.18) for group B (p = 0.25). Similarly at 2-years the mean OKS was 25.40 (range 12– 53, SD 9.51) for group A and 25.56 (range 12– 46, SD 9.67) for group B (p = 0.94). The results were similar for three and four-year follow ups with p values not significant. This study thus revealed that computer assisted TKA does not appear to result in better patient satisfaction when compared to standard instrumentation at midterm follow up.

It is known from long term analysis of conventional TKA that mal-aligned implants have significantly higher failure rates beyond eight to ten years. As use of computer navigation assistance results in a less number of mal-aligned knee prostheses, we believe that these knees will have improved survivorship. The differences in OKS between the two groups should therefore be evident after eight to ten years.