Patellofemoral instability (PFI) is a common cause of knee pain and disability in the paediatric population. Patella alta, lateralised tibial tubercle, medial patellofemoral ligament (MPFL) deficiency, genu valgum and trochlear dysplasia are well known risk factors.

A prospective database was created including patients referred through our physiotherapy pathway following first-time patella dislocation. Patella alta and lateralisation of the tibial tuberosity was treated with a Fulkerson-type tibial tubercle osteotomy(TTO). Medial patellofemoral ligament was reconstructed using quads tendon autograft pull-down technique. A modified Sheffield protocol was used postoperatively allowing weightbearing in a hinged knee brace.

Forty patients were identified with 8 patients having bilateral presentations. Male to female ratio was 12:28 with an age range of 4–17 years. Eight patients had congenital PFI, five patients acquired PFI through traumatic patella dislocation and twenty-seven patients developed PFI from recurrent dislocations. Structural abnormalities were found in 38(95%) of patients. Patella alta (Caton-Deschamps index >1.2) was identified in 19(47%) patients, genu valgum in 12(30%) patients, increased tibial tubercle-trochlear groove distance(TT-TG>20mm) was present in 9(22.5%) patients and persistent femoral anteversion(> 20 deg) in 7(17%) patients.

Eight patients were treated with TTO and MPFL reconstruction, three patients with MPFL reconstruction alone and five patients had guided growth for genu valgum correction. Ten patients are awaiting surgery. No postoperative patients had recurrence of PFI at their latest follow up.

PFI is a common problem in the adolescent paediatric population with identifiable structural abnormalities. Correcting structural pathology with surgery leads to predictable and safe outcomes.

We aim to determine the differences in lower limb joint kinematics during the golf swing of patients who had undergone Total Knee Arthroplasty (TKA) and a control group of native knee golfers.

A case-control study was undertaken with ten golfers who had undergone TKA (cruciate retaining single radius implant) and five age and matched golfers with native knees. Each golfer performed five swings with a driver whilst being recorded at 200Hz by a ten-camera motion capture system. Knee and hip three-dimensional joint angles (JA) and joint angular velocities (JAV) were calculated and statistically compared between the groups at six swing events.

The only significant differences in knee joint kinematics between TKA and control groups was a lower external rotation JA in the left knee during the backswing (p=0.010). There was no significant difference in knee JAV between the groups. Both hips demonstrated significantly (p=0.023 for left and p=0.037 for right) lower flexion in the TKA group during the takeaway swing event, and there was lower internal rotation in the backswing and greater external rotation in the downswing of the right hip. There was also slower left hip extension JAV in the downswing.

Normal knee kinematics were observed during the golf swing following TKA, with the exception of reduced external rotation in the left knee during the back swing and the right during the down swing. The differences demonstrated in the hip motion indicate that they may make compensatory movements to adjust to the reduced external rotation demonstrated in the knee.

Management of highly displaced acromioclavicular joint (ACJ) injuries remain contentious. It is unclear if delayed versus acute reconstruction has an increased risk of fixation failure and complications. The primary aim of this was to compare complications of early versus delayed reconstruction. The secondary aim was to determine modes of failure of ACJ reconstruction requiring revision surgery.

A retrospective study was performed of all patients who underwent operative reconstruction of ACJ injuries over a 10-year period (Rockwood III-V). Reconstruction was classed as early (<12 weeks from injury) or delayed (≥12 weeks). Patient demographics, fixation method and post-operative complications were noted, with one-year follow-up a minimum requirement for inclusion. Fixation failure was defined as loss of reduction requiring revision surgery.

104 patients were analysed (n=60 early and n=44 delayed). Mean age was 42.0 (SD 11.2, 17–70 years), 84.6% male and 16/104 were smokers. No difference was observed between fixation failure (p=0.39) or deep infection (p=0.13) with regards to acute versus delayed reconstruction. No patient demographic or timing of surgery was predictive of fixation failure on regression modelling. Overall, eleven patients underwent revision surgery for loss of reduction and implant failure (n=5 suture fatigue, n=2 endo-button escape, n=2 coracoid stress fracture and n=2 deep infection).

This study suggests that delayed ACJ reconstruction does not have a higher incidence of fixation failure or major complications compared to acute reconstruction. For those patients with ongoing pain and instability following a trial of non-operative treatment, delayed reconstruction would appear to be a safe treatment approach.

Total elbow arthroplasty (TEA) usage is increasing owing to expanded surgical indications, better implant designs, and improved long-term survival. Correct humeral implant positioning has been shown to diminish stem loading in vitro, and radiographic loosening in in the long-term. Replication of the native elbow centre of rotation is thought to restore normal muscle moment arms and has been suggested to improve elbow strength and function. While much of the focus has been on humeral component positioning, little is known about the effect of positioning of the ulnar stem on post-operative range of motion and clinical outcomes. The purpose of this study is to determine the effect of the sagittal alignment and positioning of the humeral and ulnar components on the functional outcomes after TEA.

Between 2003 and 2016, 173 semi-constrained TEAs (Wright-Tornier Latitude/Latitude EV, Memphis, TN, USA) were performed at our institution, and our preliminary analysis includes 46 elbows in 41 patients (39 female, 7 male). Patients were excluded if they had severe elbow deformity precluding reliable measurement, experienced a major complication related to an ipsilateral upper limb procedure, or underwent revision TEA. For each elbow, saggital alignment was compared pre- and post-operatively. A best fit circle of the trochlea and capitellum was drawn, with its centre representing the rotation axis. Ninety degree tangent lines from the intramedullary axes of the ulna and humerus, and from the olecranon tip to the centre of rotation were drawn and measured relative to the rotation axis, representing the ulna posterior offset, humerus offset, and ulna proximal offset, respectively. In addition, we measured the ulna stem angle (angle subtended by the implant and the intramedullary axis of the ulna), as well as radial neck offset (the length of a 90o tangent line from the intramedullary axis of the radial neck and the centre of rotation) in patients with retained or replaced radial heads. Our primary outcome measure was the quickDASH score recorded at the latest follow-up for each patient. Our secondary outcome measures were postoperative flexion, extension, pronation and supination measured at the same timepoints. Each variable was tested for linear correlation with the primary and secondary outcome measures using the Pearson two-tailed test.

At an average follow-up of 6.8 years (range 2–14 years), there was a strong positive correlation between anterior radial neck offset and the quickDASH (r=0.60, p=0.001). There was also a weak negative correlation between the posterior offset of the ulnar component and the qDASH (r=0.39, p=0.031), and a moderate positive correlation between the change in humeral offset and elbow supination (r=0.41, p=0.044). The ulna proximal offset and ulna stem angle were not correlated with either the primary, or secondary outcome measures.

When performing primary TEA with radial head retention, or replacement, care should be taken to ensure that the ulnar component is correctly positioned such that intramedullary axis of the radial neck lines up with the centre of elbow rotation, as this strongly correlates with better function and less pain after surgery.

This prospective, randomized, controlled trial compares patient outcome after non-operative care versus open reduction and tunneled suspension device fixation (ORTSD) for grade III or IV acromioclavicular joint disruptions.

Sixty patients aged between sixteen and thirty-five years with an acute grade III or IV AC joint disruption were randomized to receive ORTSD fixation or non-operative treatment. Functional assessment was conducted at six weeks, three months, six months, and one year using the Disabilities of the Arm, Shoulder and Hands (DASH), Oxford Shoulder Scores (OSS) and Short Form (SF-12). Reduction was evaluated using radiographs. Complications were recorded, and an economic evaluation performed.

There was no significant difference in DASH or OSS at one year between non-operative and ORTSD groups (DASH score, 4.67 versus 5.63; OSS, 45.72 versus 45.63). Patients undergoing surgery had inferior DASH scores at 6 weeks (p<0.01). Five patients who failed non-operative management subsequently received surgery. Overall cost of treatment was significantly greater after ORTSD fixation (£796.22 vs £3359.73 (p<0.01)).

ORTSD fixation confers no functional benefit over non-operative treatment at one year. While patients managed non-operatively generally recover faster, a significant group remain dissatisfied following non-operative treatment requiring delayed surgical reconstruction.

Introduction

Flexor Hallucis Longus (FHL) tendon transfer is a well-recognised salvage operation for irreparable tendon Achilles (TA) ruptures and intractable Achilles tenonopathy. Several case series describes the technique and results of arthroscopic FHL tendon transfers. We present a comparative case series of open and arthroscopic FHL tendon transfers from Southmead Hospital, Bristol, UK.

Three distal femoral axes have been described to aid in alignment of the femoral component; the Trans Epicondylar Axis (TEA), the Posterior Condylar Axis (PCA) and the Antero Posterior (AP) axis. Our aim was to identify if there was a reproducible relationship between the axes which would aid alignment of the femoral component. This is the first study compare all three distal femoral axes with each other using magnetic resonance imaging (MRI) in a Caucasian population. Our sample group represents real life patients awaiting total knee arthroplasty (TKA), as opposed non-arthritic or cadaveric knees.

We identified the relationship between these rotational axes by performing MRI scans on 89 patients awaiting TKA with patient-specific instrumentation. Measurements were taken by two observers.

Patients had a mean age of 62.5 years (range 32–91). 51 patients were female. The mean angle between the TEA and the AP axis was 92.78° with a standard deviation of 2.51° (range 88° – 99°). The mean angle between the AP axis and the PCA was 95.43° with a standard deviation of 2.75° (range 85° – 105°). The mean angle between the TEA and the PCA was 2.78° with a standard deviation of 1.91° (range 0° – 10°).

We conclude that while there is a reproducible relationship between the differing femoral axes, there is a significant range in the relationship between the femoral axes. This range may lead to greater inaccuracy than has previously been appreciated when defining the rotation of the femoral component. There is most variation between the PCA and the AP axis. The TEA's relationship with the PCA and AP appears important in defining rotation. Due to the well accepted difficulty in defining the TEA intra-operatively, there may be a role for patient-specific instrumentation in TKA surgery with pre-operative MRI.

We present the 2 year results for the first 54 patients after Selective Dorsal Rhizotomy (SDR) delivered in Bristol, concentrating on change in Reimers' migration index.

Eligible patients are selected at the SDR multidisciplinary meeting. Physiotherapy assessment is performed pre-operatively and at 6, 12 and 24 months post-surgery. Data collected includes GMFCS, Gross Motor Function Measure (GMFM) 88 and Modified Ashworth score for spasticity for major lower limb muscle groups, amongst other data. Pelvic radiographs are taken pre-operatively and at 2 years post-operatively. Reimers' migration index was measured using the hospital picture archiving and communication system (PACS).

The mean age was 7.2 (3.9–17.5) at the time of surgery. Pre- and post-operative pelvic radiographs were available for 30 patients (60 hips). 57% (n=34) hips showed an increase in migration percentage (mean 5.4%, range 0.1–17.5%) and 43% (n=26) hips showed a decrease (mean 4.0%, range 0–15.5%). Overall no significant difference was found in Reimers' migration index at 2 year follow up (mean increase 1.3% (95% CI −0.3–3.0), p=0.12).

There was an improvement in GMFCS category (by 1 grade) for 9 patients and a worsening for 1 patient at 2 year follow up. The Modified Ashworth score for spasticity improved in all patients. There was a mean improvement of 1.7 in the hip adductors and 2.4 in the ankle plantar flexors.

There was a statistically significant improvement in the GMFM 88 D and E domains of 14.7 (95% CI 11.3–18.1), p<0.0001 and 11.4 (95% CI 7.4–15.7), p<0.0001 respectively.

We found no evidence that SDR leads to worsening hip subluxation at 2 year follow up. All patients had improvement in lower limb spasticity. Overall there was a statistically significant improvement in function, as shown by GMFM 88 domains for standing, walking, running and jumping.

The National Joint Registry (NJR) for England & Wales provides a useful reference for hospitals to assess and compare their current practice with national figures. We aimed to identify patient and surgical factors at time of primary total knee arthroplasty (TKA) responsible for the revision.

A retrospective case note review was performed of all revision TKAs performed at Musgrove Park Hospital 2005–2010. Exclusion criteria included primary TKA performed elsewhere.

38 TKAs and 13 unicompartmental knee arthroplasties (UKAs) were revised. Mean time to revision was 1.8 years and 2.1 years respectively. Reason for revision was pain or disease progression in 54% revised UKAs and infection in 53% revised TKAs. 35% infected cases were therapeutically anticoagulated, 75% ASA 3 and 50% had a BMI >30. No problems were identified with surgeon grade or level of supervision. 5 cases of revision were of a trial primary prosthesis.

There were large numbers of high risk patients (increased BMI, ASA grade or on anticoagulants), which may contribute to increased infection rates and subsequent revision. We suggest avoiding trial prostheses and recommend delaying anticoagulation reintroduction until the wound is fully healed. This information may be useful for aiding with patient selection and consent.

To describe complications and reoperation rates associated with degenerative spinal deformity surgery

A retrospective review of prospectively collected data from a single spinal surgeon in the United Kingdom. A total of 107 patients who underwent surgery, of 5 or more levels, for primary degenerative kyphoscoliosis between 2006 and 2012 were identified. Clinical notes were reviewed and post-operative complications, reoperation rates, length of follow up and mortality were analysed.

A total of 107 patients, average aged 66.5 years (range 52 – 85), with 80% women. 105 patients underwent posterior surgery, two patients required both anterior and posterior surgery. The average number of instrumented levels was 8.3; 10% 5 levels, 15% 6 levels, 11% 7 levels, 14% 8 levels, 15% 9 levels and 35% had fusions of 10 levels and above. 58% included fixation to sacrum or pelvis. 93% had a decompression performed and 30% had an osteotomy. There were 40 complications recorded within the follow-up period. Infection occurred in 7 patients (6.5%). All were successfully managed with debridement, antibiotic therapy and retention of implants. There were 4 dural tears (3.7%). One patient developed a post-operative DVT (0.9%). No patients sustained cord level deficits. Prevalence of mechanical complications requiring re-operation was 26% (28 patients). 5 patients (4.7) required revision surgery for symptomatic pseudarthrosis, 7 patients (6.5%) underwent revision fixation for metal work failure (broken rods/screw pull-out) and 16 patients (14.9%) underwent revision surgery to extend fixation proximally or distally due to adjacent segment disease (symptomatic proximal junction kyphosis 4.7%; osteoporotic fracture 3.7% and junctional/nerve root pain 6.5%). Overall reoperation rate was 32.5% at an average of 1.9 years following primary surgery (range 1 week–6 years). 37% patients remain on regular outpatient review (average 3.8 years following first surgery; range 2–6 years). 52% have been discharged after a mean follow-up of 2.3 years. 11 patients had died since their surgery (10.2%) at an average 4.1 years following their spinal surgery (range 1 –5.9 years).

Overall complication rate was 37.3%. 32.5% of patients were re-operated for infective or mechanical complications. 52% of patients had been discharged at an average of 2.3 years following their surgery. 10.2% of patients had died within 6 years of surgery.

The prevalence of degenerative spondylolisthesis (DS) increases with age. With an ageing population there will be increasing demands for adult deformity surgery, with associated significant financial and complication costs. The prevalence of lytic spondylolisthesis (LS) is 6–8%. Our clinical observation is that it is extremely rare to see LS in the presence of DS and therefore the objective was to formally describe the prevalence of LS in combination with DS to assess the hypothesis that 6–8% of patients with DS should also have LS. If this is not the case it may further our understanding of the pathogenesis of DS, which could aid in future prevention or treatment.

A retrospective review of erect lateral lumbar radiographs demonstrating lumbar spondylolisthesis was performed. Radiographs were identified and analysed on the hospital Synapse Picture Archiving and Communication System (PACS). Search criteria were radiographs requested by primary care and patients aged over 60.

101 patients with spondylolisthesis were identified. 89% were requested for back or leg pain. There were 89 patients with DS (69.7% women) and 12 with LS (83% men). The average age of DS and LS patient was 75 and 70 respectively (not significant). There were no cases found with both DS and LS. This was significantly different to the expected amount of 6% (p=0.03). The level of DS was at L3/4 in 11.2% (n=10), L4/5 in 79.8% (n=71) and L5/S1 in 16.9% (n=15). 2 levels were involved in 7 patients. 94.4% (n=84) demonstrated a Meyerding grade 1 slip and 5.6% (n=5) grade 2 slip. For LS the slipped level was at L5/S1 in all cases. 66.7% (n=8) had a grade 1 slip and 33.3% (n=4) a grade 2 slip.

We found no cases of lytic spondylolisthesis in the presence of degenerative spondylolisthesis. We hypothesis that the presence of spondylolysis or lytic spondylolisthesis may be protective against development of degenerative spondylolisthesis. More work is needed to explore this further.

Three distal femoral axes have been described to aid in alignment of the femoral component; the Trans Epicondylar Axis (TEA), the Posterior Condylar Axis (PCA) and the Antero Posterior (AP) axis. Our aim was to identify if there was a reproducible relationship between the axes. Hopefully this will aid the surgeon to more accurately judge the rotation of the femoral cutting block by using the axes with the least variation. This is the first study compare all three distal femoral axes with each other using magnetic resonance imaging (MRI) in a Caucasian population awaiting total knee arthroplasty (TKA).

We identified the relationship between these axes by performing MRI scans on 89 patients awaiting TKA with patient-specific instrumentation. Measurements were taken by two observers.

Patients had a mean age of 62.5 years (range 32–91). 51 patients were female. The mean angle between the TEA and AP axis was 92.78°, standard deviation (SD) 2.51° (range 88°–99°). The mean angle between the AP axis and PCA was 95.43°, SD 2.75° (range 85°–105°). The mean angle between the TEA and PCA was 2.78°, SD 1.91° (range 0°–10°).

We conclude that while there is a reproducible relationship between the differing femoral axes, there is a significant range in the relationship between the femoral axes. This range may lead to greater inaccuracy than has previously been appreciated when defining the rotation of the femoral component. There is most variation between the PCA and the AP axis. Most systems have a cutting block with 3° of external rotation from the PCA and this would be parallel to the TEA in the majority, but not all, cases in this series. This data suggests that if the surgeon is to pick two axes to reference from, one should include the TEA.

The PFNA is used routinely at the RUH for unstable peri-trochanteric and femoral fractures. Failure of operative treatment is associated with increased morbidity and financial burden. We analysed surgical and fracture factors, aiming to identify those associated with fixation failure.

Retrospective analysis of 76 consecutive patients treated with a PFNA between 2009–2012 was performed. Patient demographics were assessed, along with fracture classification, adequacy of reduction, tip apex distance (TAD) and grade of surgeon. Failure was defined as metal work failure, non-union or need for repeat procedure.

The mean age was 78.9 years (25.9–97.4). 21 were male and 49 female. There were 17 failures (24.3%) (7 required further surgery). 10 failures were per-trochanteric, 2 sub-trochanteric and 5 mid-shaft fractures. Complications included 4 broken and 6 backed-out distal locking screws, 2 blade cut-outs, 1 nail fracture and 4 non-unions. All per-trochanteric were adequately reduced with a TAD <25 mm. 11/17 had consultant supervision.

A high rate of backed-out distal locking screws was identified. We found no concerns with adequacy of reduction, TAD or consultant supervision.

This study assesses the incidence of noise in ceramic on ceramic (COC) bearings compared to metal on polyethylene (MOP) bearings. Noise after MOP implants has rarely been studied and they never been linked to squeaking.

We have developed a noise characterising hip questionnaire and sent it along with the Oxford Hip Score (OHS) to 1000 patients; 509 respondents, 282 COC and 227 MOP; median age 63.7 (range 45–92), median follow up 2.9 years (range 6–156 months).

47 (17%) of the COC patients reported noise compared to 19 (8%) of the MOP patients (P=0.048). 9 COC and 4 MOP patients reported their hip noise as squeaking. We found the incidence of squeaking in the COC hips to be 3.2% compared to 1.8% in the MOP hips. Overall, 27% patients with noise reported avoiding recreational activities because of it and patient's with noisy hips scored on average 4 points less in the OHS (COC: P=0.04 and MOP: P =0.007).

This is the first study to report squeaking from MOP hip replacements. We therefore believe the squeaking hip phenomenon is not exclusive to hard bearings. Surprisingly, only a small proportion of patients described nose from their as a ‘squeak’. Noisy hip implants may have social implications, and patients should be aware of this. We have shown a relationship between noise and a lower OHS. However, longer follow-up and further study is needed to link noise to a poorly functioning implant.

The link between squeaking and ceramic on ceramic (CoC) bearings has been widely reported in orthopaedic literature and is described as a hard bearing phenomenon. We aim to look at the incidence of noise in CoC bearings compared to Metal on Polyethylene (MoP) bearing, which have yet to be linked to squeaking.

We developed a noise characterizing hip questionnaire and sent that along with the Oxford Hip Score (OHS) to 1000 patients; 3:2 ratio of CoC to MoP. 282 CoC patients and 227 MoP patients returned the questions: 509 patients in total. Our patient database provided details on femoral head size and the acetabular inclination angle, for each respondent

47 (17%) of the CoC hip patients reported noise compared to 19 (8%) of the MoP hip patients (P=0.054). 9 CoC patients and 4 MoP patients reported squeaking, while clicking was the most frequent answer in both groups. 27% patients with noise reported avoiding recreational activities because of it. Patient's with noisy hips scored on average, 5 points worse in the OHS (CoC: P = 0.04 and MoP: P = 0.007) and were on average 5 years younger (CoC: P<0.001 and MoP: P=0.007). No correlation was found between noisy hips and femoral head size or inclination angle.

The squeaking hip phenomenon is not exclusive to hard bearing THA. Noise from patient's hips may have social implications and this should be highlighted when consenting a patient for either of these hip procedures. In both implants, we showed there to be a correlation between noise production and a lower OHS. However, longer follow up studies are needed to link noise to a poorly functioning implant

Metal on metal (MoM) bearings in hip surgery may result in increased blood levels of metal ions. The nature of the relationship between ion levels and failure is still not fully understood.

We compared three cohorts of resurfacing patients, grouped for brand and diameter of femoral component. We measured the blood level of Cobalt and Chromium ions at an average of 4 years postoperatively. The results were grouped as follows: Birmingham Hip Resurfacing ≥50 mm diameter, Durom resurfacing ≥50 mm diameter and Durom resurfacing <50 mm diameter.

120 patients were included in each group. There were significant differences in Cobalt levels between the groups. The median cobalt level for the BHR group was 8 nmol/L higher than the Durom ≥50 mm group (P<0.005). The Durom <50 mm group recorded cobalt levels 8.5 nmol/L higher than the Durom ≥50 mm group (P=0.0004). Revision rates were equal in the Large BHR's and Large Durom HRA (both 3.3%) however the small Durom HRA had a revision rate of 8.3%.

Elevated blood ion levels can indicate a failing MoM bearing. When similar ion levels were reported for BHR and small Durom the latter had significantly higher revision rates. The large BHR and large Durom HRA have similar revision rates yet the large Durom HRA had significantly lower metal ion levels. Only one of the patients having revision surgery (n=18) had metal ion levels above the MHRA threshold. This suggests ion levels do not absolutely predict the rate of HRA failure. Given similar revision rates with different ion levels between the large BHR and large Durom hips, mechanisms of failure leading to revision must be isolated from the release of metal ions. Therefore clinical and image based follow up are recommended in addition to ion level monitoring.

Patello-femoral arthritis can result in a considerable thinning of the patella. The restoration of an adequate patella thickness is key to the successful outcome of knee arthroplasty.

The objectives were (1) to establish a reproducible patella width:thickness index including chondral surface and (2) to investigate whether there is a difference between bone alone and bone/chondral construct thickness as shown by MRI.

Forty three MRI scans of young adults, mean age 27 (range 17–38), 34 male and 9 female, were studied. Exclusion criteria included degenerative joint disease, patello-femoral pathology or age under 16/over 40 (102 patients). The bony and chondral thickness of the patella and its width were measured. Inter/intra observer variability was calculated and correlation analysis performed.

We found a strong correlation between patella plus cartilage thickness and width (Pearson 0.75, P < 0.001). The mean width:thickness ratio was 1.8 (SD 0.10, 95% CI 1.77–1.83). Without cartilage the ratio was 2.16 (SD 0.15, 95% CI 2.11–2.21), correlation was moderate (Pearson 0.68, P < 0.001). The average patella cartilage thickness was 4.1mm (SD 1.1, 95% CI 3.8–4.5).

The narrow confidence intervals for the ratio of patella width:thickness suggest that patella width can be used as a guide to accurate restoration of patella thickness during total knee or patella-femoral replacement. We would recommend a ratio of 1.8:1.

Aim

The present study aimed to assess the accuracy of preoperative departmental ultrasound scans in identifying rotator cuff tears at our institution.

Introduction

There has been little research into the effect of suffering a simultaneous hip and upper limb fragility fracture. The aim of this study is to describe the characteristics of this important group of patients and to define the effect on outcomes such as mortality and length of stay.

The indication for rotator cuff repair in elderly patients is controversial.