This study identifies variations in presentation and demographics between structural and non-structural (muscle patterning) shoulder instability.

We analysed 1020 unstable shoulders (855 patients) from our institution database. Demographic details, direction and aetiology were obtained from medical records. Anterior dislocations comprised 67%, posterior 31% and inferior 2% of all directions of instability and 75 shoulders had multidirectional instability.

Structural causes were dominant in anterior instability (traumatic 39% and atraumatic 38%) and muscle patterning in posterior (81%) and inferior (90%) instability. Males accounted for 64% of all patients (73% of all structural patients and 53% of muscle patterning patients. Mean age at presentation was 25 years old (structural patients 28 years and muscle patterning patients 21 years old). There were 690 unilaterally unstable shoulders (57% right- and 43% left-sided); the dominant arm was affected in 58% overall, in 42% of all left-handers and only 33% of left-handers with muscle patterning. Bilateral shoulder instability occurred in 19% of all patients (12% of patients with structural instability and 28% of those with muscle patterning instability). For muscle patterning, the mean age at onset of symptoms was 14 years, and mean length of symptoms before presentation was 8 years. There was a trimodal distribution of age at onset of symptoms corresponding to peaks at 6, 14 and 20 years. In the group with onset of muscle patterning under 10 years old, there was a higher proportion of females (71% vs 47%), laxity (63% vs 29%) and bilaterality (54% vs 42%), and fewer presenting with pain (17% vs 50%).

Muscle patterning instability is associated with a demographic and presentation profile which may help distinguish it from structural forms of instability. As age at presentation increased, pain increased and joint laxity decreased. Bilaterality did not appear to be associated with gender, the presence of laxity or pain.

We report 5 cases of linked shoulder and elbow replacement (LSER) following failure of single-joint arthroplasty. Whilst total humeral replacement has been reported for treatment following resection for tumour we know of no reports of linked shoulder and elbow prostheses for arthropathy alone. Between May and December 2005, 2 patients with total elbow arthroplasty and 3 patients with total shoulder arthroplasty were revised to LSER for loosening of the long humeral stems or periprosthetic fracture. Custom-made prostheses were produced using computer-aided design and manufacture technology. There were no early complications including infection. All 5 patients reported early improvement of symptoms, with the ability to bear weight axially through the limb, restored. This technique avoids the problem of a stress riser between the stems of separate shoulder and elbow replacements and solves the problem of salvage of long-stemmed implants where no further humeral fixation is possible.

This study identifies variations in presentation and demographics for different forms of shoulder instability. We analysed 1020 unstable shoulders (855 patients) from a previously presented database. Demographic details, direction and aetiology were obtained from medical records. Anterior dislocations comprised 67%, posterior 31% and inferior 2% of all directions of instability and 75 shoulders had multidirectional instability. Structural causes were dominant in anterior instability (traumatic 39% and atraumatic 38%) and muscle patterning in posterior (81%) and inferior (90%) instability. Males accounted for 64% of all patients (73% of all structural patients and 53% of muscle patterning patients. Mean age at presentation was 25 years old (structural patients 28 years and muscle patterning patients 21 years old). There were 690 unilaterally unstable shoulders (57% right- and 43% left-sided); the dominant arm was affected in 58% overall, in 42% of all left-handers and only 33% of left-handers with muscle patterning. Bilateral shoulder instability occurred in 19% of all patients (12% of patients with structural instability and 28% of those with muscle patterning instability). For muscle patterning, the mean age at onset of symptoms was 14 years, and mean length of symptoms before presentation was 8 years. There was a trimodal distribution of age at onset of symptoms corresponding to peaks at 6, 14 and 20 years. In the group with onset of muscle patterning under 10 years old, there was a higher proportion of females (71% vs 47%), laxity (63% vs 29%) and bilaterality (54% vs 42%), and fewer presenting with pain (17% vs 50%). As age at presentation increased, pain increased and joint laxity decreased. Bilaterality did not appear to be associated with gender, the presence of laxity or pain. Muscle patterning instability is associated with a demographic and presentation profile which may help distinguish it from structural forms of instability.

We present the use of dynamic electromyographic analysis (DEMG) in the diagnosis of muscle patterning instability. DEMG’s were requested in 168 of 562 muscle patterning shoulders with suspected subclinical or clinically complex muscle patterning instability. An experienced neurophysiologist (blinded to the clinical findings and direction of instability) inserted dual-wire tungsten electrodes into pectoralis major, latissimus dorsi, infraspinatus and anterior deltoid. Muscle activity was recorded during rest, flexion, abduction, extension, and cross-body adduction. 5 investigations were abandoned. The timing and magnitude of muscle activity was noted and compared to the clinical diagnosis and direction of instability. DEMG identified a total of 204 abnormal muscle patterns in 163 shoulders. The examination was normal in 13 patients (8%). A single muscle was abnormal in 63 shoulders, 2 muscles in 55, 3 muscles in 9, and all 4 muscles in one shoulder. Over-activation of pectoralis major was identified in 58%, and latissimus dorsi in 70%, of shoulders with anterior instability. In posterior instability, latissimus dorsi was overactive in 76%, anterior deltoid in 14% and infraspinatus was under-active in 24%. Pectoralis major and Latissimus dorsi were both overactive in 38% of anterior, 29% of posterior and 38% of multidirectional instability. Abnormal muscle patterns were identified in 52 shoulders with subclinical muscle patterning. A further 98 shoulders had 134 clinically abnormal muscle patterns. These were confirmed by DEMG in 57 cases (sensitivity 43%), and DEMG’s were normal in 77 (specificity 43%). DEMG also identified 65 additional muscles as abnormal in the 98 clinically abnormal shoulders. DEMG performed by an experienced neurophysiologist provides additional information regarding abnormal muscle activation in selected complex or subtle cases of muscle patterning instability in which clinical examination has a low sensitivity and specificity.

The restoration of pain-free stable function in gleno-humeral arthritic cases in various situations such as rotator cuff deficiency, old trauma and failed total shoulder arthroplasty is a challenging clinical dilemma. The Bayley-Walker shoulder has been designed specifically for very difficult cases where surface replacement devices do not provide sufficient stability. This device is a fixed-fulcrum reversed anatomy prosthesis consisting of a titanium glenoid component with a CoCrMo alloy head that articulates with an UHMWPE liner encased in a titanium alloy humeral component that has a long tapered grooved stem. The centre of rotation of the Bayley-Walker shoulder is placed medially and distally with respect to the normal shoulder in order to improve the efficiency of the abductor muscles. An important problem in devices of this type is obtaining secure and long-lasting fixation of the glenoid component. The glenoid component relies on fixation through the cortical bone by using threads, which protrude through the anterior surface of the scapula at the vault of the glenoid. It is HA coated for subsequent osseointegration. The purpose of this study was to investigate fixation of the glenoid component.

A 3D finite element model of the glenoid component implanted in a scapula was analysed using Abaqus. The implant was placed in position in the scapula, with the final 2–3 screw threads cutting through the cortical bone on the anterior side at the vault of the glenoid due to the anatomy in this region. The analysis was performed for two load cases at 60° and 90° abduction. A histological study of a retrieval case, obtained 121 days after implantation, was also conducted.

The FEA results showed that most of the forces were transmitted from the component to the cortical bone of the scapula, the remaining load being transmitted through cancellous bone. In particular the area where the threads of the glenoid component penetrated the scapula showed high strain energy densities. Histology from the retrieved case showed evidence of bone remodelling whereby new bone growth resulting in cortical remodelling had occurred around the threads.

Both the FEA and histological study show that fixing the component at multiple locations in cortical bone may overcome the problems of glenoid loosening associated with constrained devices. The Bayley-Walker device has been used on a custom basis since 1994; 81 Bayley–Walker shoulders for non-tumour conditions and 43 Bayley-Walker glenoid components have been used in association with a bone tumour implant, with good early results. Radiographically, radiolucencies have not been observed and overall the comparisons with the original Kessel design are positive.

Over a six-year period, 166 shoulder arthroscopies were performed in 153 patients with symptoms of subluxation, but no firm evidence of dislocation and no clear-cut clinical diagnosis. Arthroscopy confirmed the working diagnosis in 133 shoulders (80%), but changed it in 33 (20%). In 26 shoulders the presumed direction of instability was changed and in seven a different diagnosis was made. Arthroscopy was particularly helpful for patients with shoulder instability of uncertain type or direction.

Double-contrast shoulder arthrograms were performed in 20 patients at an average of 30 months after operative repair of a torn rotator cuff. In 18 out of 20 shoulders the contrast medium leaked into the subacromial bursa indicating a defect in the rotator cuff. Despite this, 17 patients had complete relief of pain and 15 had a full range of shoulder elevation. The results suggest that a completely watertight closure is not essential for a good functional result, and that arthrography may not be helpful in the investigation of failure of repair.

The results of 63 operative repairs of chronic tears of the rotator cuff in 61 patients are reviewed retrospectively; the mean follow-up was 32.7 months. Fifty-four patients presented with symptoms of persistent pain and seven patients with gross loss of movement. All the patients had failed to respond to conservative treatment. Results were assessed in terms of relief of pain, restoration of movement, the patients' ability to return to work and whether they were satisfied with the results. Overall, a good result in terms of relief of pain was achieved in 40 shoulders. In 31 shoulders (30 with pain and one without pain) the operation included particular measures to decompress the subacromial space; 26 of the patients achieved relief of pain which was significantly better than in those patients whose operation did not include a decompression. The complications and failures are discussed. It is suggested that operative repair of the chronically torn rotator cuff of the shoulder is a worthwhile operation and that the operation should include an adequate decompression of the subacromial space.