Introduction

Limb deformity is usually assessed clinically assisted by long leg alignment radiographs and further imaging modalities (MRI and CT). Often decisions are made based on static imaging and simple gait interpretation in clinic. We have assessed the value of gait lab analysis in surgical decision making comparing surgical planning pre and post gait lab assessment.

The aim of this study was to define the long-term outcome following adolescent clavicle fracture.

We retrospectively reviewed all adolescent fractures presenting to our region (13–17years) over a 10-year period. Patient reported outcomes were undertaken at a minimum of 4 years post-injury (QuickDASH and EQ-5D) in completely displaced midshaft fractures (Edinburgh 2B, >2cm displacement, n=50) and angulated midshaft fractures (Edinburgh 2A2, >30 degrees angulation, n=32).

677 clavicle fractures were analysed. The median age was 14.8 (IQR 14.0–15.7) and 89% were male. The majority were midshaft (n=606, 89.5%) with either angulation (39.8%) or simple fully displaced (39.1%). Only 3% of midshaft fractures underwent acute fixation (n=18/606), all of which were fully displaced. The incidence of refracture following non-operative management of midshaft fractures was 3.2% (n=19/588), all united with non-operative management. Fracture type, severity of angulation or displacement was not associated with risk of refracture. There was one case of non-union encountered following non-operative management of all displaced midshaft fractures (0.4%, n=1/245). At a mean of 7.6 years following injury, non-operative management of both displaced and angulated fractures had a median QuickDASH was 0.0 (IQR 0.0–2.3), EQ-5D was 1.0 (IQR 1.0–1.0). 97% of angulated fractures and 94% of displaced fractures were satisfied with their final shoulder function.

We conclude that Non-operative management of adolescent midshaft clavicle fractures result in excellent functional outcomes with a low rate of complications at long-term follow up. The relative indications for surgical intervention for clavicle fractures in adults do not appear to be applicable to adolescents.

Purpose

This study re-examined the influence of the head shaft angle (HSA) on hip dislocation in a large cohort of children with cerebral palsy (CP).

To evaluate the outcome of combined tibialis anterior tendon shortening (TATS) and calf muscle-tendon lengthening (CMTL) in spastic equinus.

Prospectively collected data was analysed in 26 patients with hemiplegic (n=13) and diplegic (n=13) cerebral palsy (CP) (GMFCS level I or II, 14 males, 12 females, age range 10–35 years; mean 16.8 years). None had received botulinum toxin A injections or surgery in the preceding six and 12 months respectively. All patients had pre-operative 3D gait analysis and a further analysis at a mean of 17.1 months (± 5.6months) after surgery. None was lost to follow-up. Twenty-eight combined TATS and CMTL were undertaken and 19 nineteen patients had additional synchronous multilevel surgery. At follow-up 79% of patients had improved foot positioning at initial contact. Statistically significant improvements were seen in the Movement Analysis Profile for ankle dorsi-/plantarflexion (4.25, p=0.032), maximum ankle dorsiflexion during swing phase (11.68°, p<0.001), and Edinburgh Visual Gait Score (EVGS) (4.85, p=0.014). Diplegic patients had a greater improvement in the EVGS than hemiplegics (6.27 -vs- 2.21, p = 0.024).

The originators of combined TATS and CMTL showed that it improved foot positioning during gait. The present study has independently confirmed favourable outcomes in a similar patient population and added additional outcome measures, the EVGS, foot positioning at initial contact, and maximum ankle dorsiflexion during swing phase. Study limitations include short term follow-up in a heterogeneous population and that 19 patients had additional surgery. However, distinguishing between the natural history of CP and interventions and isolating the effects of one intervention from others in multilevel surgery are well recognised difficulties in cohort studies in CP. TATS combined with CMTL is a recommended option for spastic equinus in ambulatory patients with CP.

A goal attainment scale (GAS) was used to evaluate outcomes of surgical and non-surgical interventions to improve gait in children with diplegic cerebral palsy (CP).

Personal goals were recorded pre-intervention from children and/or their carers attending the Edinburgh Gait Laboratory since 2012. Twenty children underwent orthopaedic surgery (Group 1) and 25 children underwent a non-orthopaedic intervention (Group 2). Patients were excluded if the intervention was <9 months before the study period. Post-operatively children and/or their carers were contacted by telephone to complete the mGAS questionnaire, rating the achievement of goals on a 5-point scale. The majority of goals related to structure and function and were similar between groups, with goals relating to stability and lower limb structure most frequently recorded. Attaining an improvement in pain was stated more frequently by Group 1 children. The GAS formula was used to transform the composite GAS into a standardised measure (T-score) for each patient. A t-test was used to determine if the change in T-score was significantly different from 0, i.e. no change. Both groups on average achieved their goals (mean change in T-score for Group 2 11.1, vs 21.1 for Group 1). The difference between these two means was significant (p = 0.012). Additionally 16 children had undergone a follow-up gait analysis, but the relationship between the change in Gait Profile Score and GAS, assessed by Pearson's correlation coefficient was statistically insignificant.

Both surgical and non-surgical interventions enabled children to achieve their goals, although Group 1 reported higher achievement. GAS reflect patients’ aspirations and may be as relevant as post-intervention kinematic or kinetic outcomes.

Controversy remains as to whether the contralateral hip should be fixed in patients presenting with unilateral slipped capital femoral epiphysis (SCFE). This study compares the outcomes of those patients who had prophylactic fixation with those who did not.

We identified 90 consecutive patients with a mean age of 12.3 years presenting to the study centre with SCFE from a prospective operative database. The patient's notes and radiographs were retrospectively analysed for post-operative complications, re-presentation with a contralateral slip, and the presence of a cam lesion.

The mean length of follow-up was 8 years (range 3 to 13). Fifty patients (56%) underwent unilateral fixation and 40 patients underwent bilateral fixation, of which 4 (4%) patients had simultaneous bilateral SCFE and 36 (40%) had prophylactic fixation of the contralateral hip. Twenty-three patients (46%) that underwent unilateral fixation, went onto have contralateral fixation for a further SCFE. Two patients from this group had symptomatic femoracetabular impingement from cam lesions and one patient required a Southwick osteotomy for a severe slip. Five patients (10%) that had unilateral fixation only demonstrated cam lesions on radiographic analysis, being suggestive of an asymptomatic slip. No post-operative complications were observed for the contralateral hip in patients that had prophylactic screw fixation and no cam lesions were identified on radiographic assessment.

This study suggests that the contralateral hip in patients presenting with unilateral SCFE should be routinely offered prophylactic fixation to avoid a further slip, which may be severe, and the morbidity associated with a secondary cam lesion.

Fractures repair by two mechanisms; direct fracture healing and indirect fracture healing via callus formation. Research concerning the effects of bisphosphonate on fracture repair has solely assessed indirect fracture healing. Patients with osteoporosis on bisphosphonates continue to sustain fragility fractures. A proportion of osteoporotic fractures require plate fixation. Bisphosphonates impair osteoclast activity and therefore, may adversely affect direct fracture healing that predominates with plate fixation.

Five skeletally mature Sprague-Dawley rats received daily subcutaneous injections of 1mg/kg Ibandronate (IBAN). Similarly, five control rats received saline (CONTROL). Three weeks following commencement of injections a tibial osteotomy was rigidly fixed with compression plating similar to that seen in routine clinical practice. Fracture healing was monitored with radiographs. Six weeks post plate fixation, animals were sacrificed. Radiographs were performed of the extricated tibiae following plate removal. The visibility of the osteotomy site was scored as totally visible, partially visible or absent as previously described. Mechanical testing was conducted on the healing osteotomies via 4-point bending.

Fractures healed without visible external callus. In the IBAN group three animals had totally visible osteotomy lines and two had partially visible osteotomy lines. The CONTROL group had three animals with absent osteotomy lines and two with partially visible osteotomy lines. The mean (±SD) stress at failure for the healing tibial osteotomies at 6 weeks was 28.8 (±23.97)MPa in the IBAN group and 37.4(±29.20) MPa in the CONTROL group (p=0.62)

Our results indicate that Ibandronate adversely affected direct fracture repair as demonstrated by the radiographic density of the fracture line. The strength of the repair was reduced but this did not reach statistical significance. Our results suggest that a sample size of 220 animals is required to detect a 15% difference (alpha 0.05, beta 0.2) which suggests the effect of bisphosphonates on direct fracture repair may be small.

Aim

To compare the results of total knee replacement in a consecutive series of morbidly obese patients (body mass index (BMI) > 40 kg/m²) with a matched group of non-obese (BMI< 30 kg/m²) patients.

Pre-operative co-morbidities such as known coronary artery disease have commonly deemed a patient at ‘high risk’ for primary elective Total Hip Arthroplasty (THA).

We prospectively collected data on 1744 patients who underwent primary elective THA between 1998 and 2004. 273 had a history of cardiac disease defined as a previous hospital admission with a diagnosis of angina pectoris or myocardial infarction. 594 patients had hypertension defined as that requiring treatment with antihypertensives. We also had data on pre-operative age, sex and body mass index (BMI).

There was no statistically significant increase in early mortality at 3 months with a history of cardiac disease or hypertension and this remained so when adjusting for the other factors in a multivariate analysis. Sex or BMI also did not have a statistically significant effect on the risk of death within 3 months. Increasing age was the only significant risk factor for early mortality (P<0.001).

Longer term mortality at 2 and 5 years in relation to these factors was also examined. Statistical analysis revealed that coronary history now showed a highly significant association (P<0.001) with long term mortality, in patients who survived more than 3 months. 95% confidence intervals for percentage mortality at 5 years were 9.7 - 21.7 with a cardiac history compared to 4.8 - 8.8 without a cardiac history. This remained significant (P=0.002) when adjusted for the other factors. Hypertension continued to have no effect, nor did BMI. Age remained a significant risk factor. Females had a slightly lower long term death rate than males, following THA.

The overall long term mortality following THA was less than expected from the normal population, even in the subgroup with a coronary history.

This study will assist clinicians when advising patients seeking primary elective THA, who have one of these common risk factors.

The dynamic association between the immune system and the skeletal system has recently been appreciated. It has been suggested that cells involved in the inflammatory cascade might modulate the bone fracture repair process. Interestingly a number of studies have demonstrated that ability of the T lymphocyte to affect bone remodelling and health profoundly. For example the presence of T lymphocytes has been shown to increase bone resorption during experimental induced arthritis. We wanted to investigate the role of specific T lymphocytes in fracture repair and required an in vivo model to deplete CD4 and CD8 T lymphocytes selectively.

Method The cell lines of Anti-L3T4 (CD4), Anti-Lyt-2 (CD8) and Anti-phytochrome were obtained from ECACC and produced by Edinburgh University group, Immunosolv. Anti-phytochrome (AFRC MAC 51) antibody was used as the control. To each group of 5 murine models 50ìL of CD4 or CD8 or control antibody was injected ip on days 0, 1, 2, 7, 8, and 9. The body weight and behaviour were measured. On day 20 the spleens were sampled and a single cell suspension was created for each murine model. PE CD8 antibody and FITC CD4 antibody were then added to each sample. Each spleen sample was then cell sorted using the FACS machine.

Results Compared to the control group the murine model injected with CD4 antibody had only 1.14 % of CD4 T cells remaining (mean 2.462 % +/− 0.270). Similarly, the murine models injected with CD8 antibody had only 1.8% of CD8 T cells remaining (mean 1.723 % +/− 1.036).

Conclusion Our results suggest that to investigate the role of T lymphocytes in fracture repair, specific T lymphocytes can be successfully depleted with the repeated use of antibodies.

Preoperative co-morbidities such as known coronary artery disease have commonly deemed a patient at ‘high risk’ for primary elective Total Hip Arthroplasty (THA). We prospectively collected data on 1744 patients who underwent primary elective THA between 1998 and 2004. 273 had a history of cardiac disease defined as a previous hospital admission with a diagnosis of angina pectoris or myocardial infarction. 594 patients had hypertension defined as that requiring treatment with antihypertensives.

We also had data on preoperative age, sex and body mass index (BMI).

There was no statistically significant increase in early mortality at 3 months with a history of cardiac disease or hypertension and this remained so when adjusting for the other factors in a multivariate analysis. Sex or BMI also did not have a statistically significant effect on the risk of death within 3 months. Increasing age was the only significant risk factor for early mortality (P< 0.001).

Longer term mortality at 2 and 5 years in relation to these factors was also examined. Statistical analysis revealed that coronary history now showed a highly significant association (P< 0.001) with long term mortality, in patients who survived more than 3 months. 95% confidence intervals for percentage mortality at 5 years were 9.7 – 21.7 with a cardiac history compared to 4.8 – 8.8 without a cardiac history. This remained significant (P=0.002) when adjusted for the other factors. Hypertension continued to have no effect, as did BMI. Age remained a significant risk factor. Females had a slightly lower long term death rate than males, following THA.

The overall long term mortality following THA was less than expected from the normal population, even in the subgroup with a coronary history.

This study will assist clinicians when advising patients seeking primary elective THA, who have one of these common risk factors.

Fracture repair is a wound healing process that in young healthy patients usually proceeds to uncomplicated union. However, the healing cascade is delayed with increasing age, medication and certain diseases such as rheumatoid arthritis.

Recently the important role of the immune system in fracture repair has become apparent within the emerging subject of Osteoimmunology. Patients with rheumatoid arthritis have an altered immune system and therefore we have investigated the hypothesis that patients with rheumatoid arthritis have a higher incidence of non-union after a fracture compared to patients without rheumatoid arthritis.

Method: The Edinburgh Royal Infirmary computer database was searched over a 10 year period (May 1996- May 2006) to identify all patients with non-union out of the total number of patients presenting with fractures. These patients groups were then subdivided into patients that had and did not have rheumatoid arthritis. Patients were excluded if they were lost to follow up, or if the fracture either occurred before the May 1996 or management continued passed May 2006. In this study non-union was defined as failure to heal within expected timescale and lack of progression at serial x-rays (all non-union were diagnosed at least 3 months from fracture).

Results: From May 1996 through to May 2006, 8,456 patients with fracture were defined. 71 of these patients with fractures had rheumatoid arthritis. Of these patients 63 had union of their fractures whilst 8 patients developed non-union of their fracture (11.3%). In a total of 8385 non rheumatoid arthritis patients 164 developed non-union of their fracture compared to 8221 patients who had union of their fractures (2%). Comparison between these two patient groups suggests rheumatoid arthritis patients are more likely to develop non-union of traumatic fractures (Chi squared test, p value < 0.001).

Patients with rheumatoid arthritis who progressed to non-union were on the following medication, Gold (1), Indomethacin (1), Non steroidal anti-inflammatories (4), Combination analgesia (2), Antihypertensives (2), Omeprazole (1) and Thyroxine (1).

Discussion: The results from this study suggest that patients with rheumatoid arthritis have a greater incidence of non-union after a fracture compared to patients without rheumatoid arthritis. This maybe due to the abnormal immune system in rheumatoid arthritis patients. However rheumatoid arthritis patients are often on a number of medications and these drugs rather than the innate alteration of the immune system may be responsible for the altered healthy response. However whether as a result of the rheumatoid arthritis itself or the medication, our study demonstrates a higher non-union rate in the rheumatoid arthritis patients and this needs to be taken into account when treating rheumatoid arthritis patients with fractures.

An estimated 10% of patients have problems with fracture healing. Initial studies have revealed that it is likely that both the innate and specific immune systems play a role in fracture repair, but this has not been attributed to particular components, cells or their products. It is known that the functionality of the immune system is impaired with age and this may account for the higher rate of delayed union in elderly patients.

We used a validated mouse model of a reproducible closed tibial fracture. In order to prevent any foreign body inflammatory/immune response no artificial internal fracture fixation was used and instead external support was provided using a Plaster of Paris cast. The role of the specific immune system was studied using an immunodeficient Balb/c SCID (Severe Combined Immuno Deficient) mutant mouse. The SCID mice were matched for age, sex (all males) and weight to the control, wild type Balb/c mice. Mechanical (4 point bending) and radiographic (Radiographs scanned and calculations of callus area, index and density made with image analysis software) measures were used to assess fracture repair at 21 days.

Mechanical measurements revealed an enhancement of fracture healing in the SCID mouse strain compared to the control strain, with stress at yield and Young’s modulus higher in SCID mice than controls. (Stress at yield: 4.2 +/− 0.23MPa in Controls, 7.1 =/− 0.6MPa in SCIDs, P< 0.01; Young’ Modulus: 22.1 +/− 2.99MPa in Controls, 60+/− 9.9MPa in SCIDs P< 0.01). There were no significant differences seen in mechanical properties of unfractured bone between the two strains. Radiographic analysis revealed no significant differences in callus area or index (both measurements of callus size) but callus density was significantly higher (P< 0.01) in the SCID subjects compared to controls (2.6 +/− 0.06E5 Greyscale in SCIDs vs. 2.2 +/− 0.09E5 in controls).

We conclude that an abnormality of the immune system due to either lack of the specific immune system (T and B cells) or an enhancement of the innate system results in increased mineralization, stiffness and strength of fracture healing, and that further investigation might result in novel therapies directed toward avoidance of non/delayed-union.

The results of 41 consecutive total knee replacements performed on morbidly obese patients with a body mass index > 40 kg/m², were compared with a matched group of 41 similar procedures carried out in non-obese patients (body mass index < 30 kg/m²). The groups were matched for age, gender, diagnosis, type of prosthesis, laterality and pre-operative Knee Society Score. We prospectively followed up the patients for a mean of 38.5 months (6 to 66). No patients were lost to follow-up. At less than four years after operation, the results were worse in the morbidly obese group compared with the non-obese, as demonstrated by inferior Knee Society Scores (mean knee score 85.7 and 90.5 respectively, p = 0.08; mean function score 75.6 and 83.4, p = 0.01), a higher incidence of radiolucent lines on post-operative radiographs (29% and 7%, respectively, p = 0.02), a higher rate of complications (32% and 0%, respectively, p = 0.001) and inferior survivorship using revision and pain as end-points (72.3% and 97.6%, respectively, p = 0.02).

Patients with a body mass index > 40 kg/m² should be advised to lose weight prior to total knee replacement and to maintain weight reduction. They should also be counselled regarding the inferior results which may occur if they do not lose weight before surgery.

Introduction: The Depuy Pressed Fit Condylar (PFC) total knee arthroplasty (TKA) is well established with reported 10-year survival rates of 93–97%. The PFC was modified, leading to the introduction of the Sigma TKA in 1997. The theoretical advantages of the Sigma system include increased contact area between the femoral component and the tibial insert. We report the first 5-year clinical and radiographic follow-up data for the Sigma TKA.

Methods: Over a ten-month period, 211 Sigma TKAs were performed in 179 patients. Patients were seen at a specialist nurse-led clinic at admission and at 6 months, 18 months, 3 years and 5 years after surgery. Data were recorded prospectively at each visit. Radiographs were obtained at the 5-year follow-up appointment.

Results: Of 211 knees, 187 (150 patients) were alive at 5 years. 5 were lost to follow up. 5 knees (2.4%) were revised: 4 for infection and 1 underwent change of polyethylene insert at 4.9 years. 5-year survival with an endpoint of revision for any reason was 97.4%; with an endpoint of revision for aseptic loosening it was 99.5%. The median American Knee Society score was 93/100 at 5 years compared with 25/100 at admission. Of 145 radiographs, 17 (11.7%) showed radiolucent lines. None showed radiographic loosening of either component. 28 (19.3%) had alignment outside the range of 7±3° valgus.

Discussion: These results suggest that the Sigma TKA gives excellent clinical results after five years. Further follow-up studies are required to see if this performance is maintained in the long term.

Introduction To assess five-year survivorship and clinical outcome following unicompartmental knee replacement (UKR) and total knee replacement (TKR) in pre-operatively matched patient groups.

Methods From a prospective database of over 600 patients undergoing knee replacement for osteoarthritis, 70 primary Oxford UKRs (62 patients) were matched pre-operatively with 70 primary PFC TKRs (68 patients) for age, sex, body mass index, knee range of movement and Knee Society Score (KSS). The two groups were assessed at six, eighteen, thirty-six and sixty months following knee replacement and compared for survivorship of the prosthesis (with failure defined as revision for any reason or ‘worst-case’ assuming all patients lost to follow up are revised), post-operative ROM achieved and clinical outcome based on the KSS (reported as separate knee and function score).

Results Three (5 percent) patients in the UKR group and two (3 percent) in the TKR group were unavailable for follow-up at five years. Five-year survivorship based on revision for any reason was 90 percent (95 percent Confidence Interval (CI): 82 to 98) for UKR and 100 percent (95 percent CI: 100 to 100) for TKR (log rank test: p=0.009). The ‘worst-case’ five-year survivorship (assuming all patients lost to follow up are revised) was 85 percent (95 percent CI: 76 to 94) for UKR and 97 percent (95 percent CI: 93 to 100) for TKR (log rank test: p=0.02). The mean post-operative ROM achieved was 105.3 degrees following UKR and 98.3 degrees following TKR (difference 7.0 (95 percent CI 3.3 to 10.6), p< 0.001). There was no difference in the KSS between the two groups for the knee score (difference 0.1 (95 percent CI: −4.3 to 4.5), p=0.9) and function score (difference 2.6 (95 percent CI: −1.7 to 6.9), p=0.2).

Conclusion In comparable patients with osteoarthritis of the knee, survivorship remains superior for TKR at five years. The ROM achieved is greater for UKR but there is no difference in the overall clinical outcome following UKR or TKR. We believe that UKR should only be performed in carefully selected cases because mid-term clinical outcome is similar following UKR or TKR and the complication rate may be higher for UKR.

Introduction and Aims: Several non-comparative studies for uni-compartmental knee replacement (UKR) and total knee replacement (TKR) have been published, but few provide a direct comparison in well-matched patients. The aim of this study was to assess five-year survivorship and the quality of outcome following UKR and TKR in pre-operatively matched patient groups.

Method: From a prospective database of over 600 patients undergoing knee replacement for osteoarthritis, 70 primary medial Oxford UKRs (62 patients) were matched pre-operatively with 70 primary Kinematic TKRs (68 patients) for age, sex, body mass index, active range of movement (ROM) and Knee Society Score (KSS). The two groups were assessed at six, 18, 36 and 60 months following knee replacement and compared for survivorship of the prosthesis (with failure defined as revision for any reason or ‘worst-case’ assuming all patients lost to follow-up are revised), quality of outcome based on the KSS (reported as separate knee and function score) and the ROM achieved.

Results: Three (4%) patients in the UKR group and two (3%) in the TKR group were unavailable for follow-up at five years. Five-year survivorship (Kaplan-Meier methodology) was significantly different for the two groups using revision for any reason as an endpoint – 90% for UKR, 100% for TKR (log-rank test, p=0.01). Assuming all patients lost to follow-up were revised, the ‘worst-case’ five-year survivorship was also significantly different for the two groups – 85% for UKR, 97% for TKR (log-rank test, p=0.02). Analysis of the KSS at six, 18, 36 and 60 months showed no significant difference between the two groups for the knee score (p=0.48) and the function score (p=0.1). The ROM achieved at five years was greater for the UKR group and the difference was statistically significant (UKR: Least square mean 105 degrees, standard error 1.2; TKR: Least square mean 98 degrees, standard error 1.2, p< 0.001).

Conclusion: In comparable patients with osteoarthritis of the knee, survivorship remains superior for TKR at five years. The ROM achieved is greater for UKR, but there is no difference in the overall quality of outcome following UKR or TKR. We believe that UKR should only be performed in carefully selected cases because mid-term functional outcome is similar following UKR or TKR and the complication rate may be higher for UKR.

Introduction: Previous studies have suggested that bilateral total hip replacement (THR) is a safe and economically advantageous procedure. It has not become routine practice in many centres due to persisting concerns about complications and the patients’ ability to rehabilitate. In this study we compare one group of patients who underwent bilateral THR with another group who had only a unilateral procedure despite osteoarthritis in both hips. They were assessed for complications and functional outcome.

Methods: All data collected prospectively between 1998 and 2002. All patients had osteoarthritis and were recruited during the same time period . Group A comprised 49 patients who were listed for bilateral THR. Group B consisted of 215 patients listed for unilateral THR, with both hips symptomatic. Pain and function were assessed using the Harris Hip Score preoperatively and at 6 months post operatively. All intra and postoperative complications were recorded.

Results: There was no statistical difference between the groups for sex , pathology, body mass index or American Society of Anaesthetists (ASA) scores.

There is a difference in age between the two groups with the bilateral group being younger (p< 0.01).

Although group A had a higher death rate, it was not statistically significant. The deep infection and revision rate was similar for both groups. The average length of stay for group A was 13 days compared to 8 days for group B.There was a significant difference in the likelihood of blood transfusion with a transfusion rate of 78% in group A and 22% in group B (p< 0.001).

Conclusion: This study has demonstrated no significant increased risk of death, deep infection or need for revision with bilateral hip replacements. The increased likelihood of blood transfusion should be offset against the definite economic advantages and the good early outcome from performing bilateral total hip replacements.