The complications of prostheses for fractures of the proximal humerus are understudied because the experience of each shoulder surgeon is limited and a standardised registration protocol is not yet available. A prospective study on complications in shoulder arthroplasty for fracture is, therefore, essential to explore variables that influence outcome. The purpose of this study is to report our experience with complications following arthroplasty for proximal humeral fractures.

In a multicenter study, four hundred and six patients treated with arthroplasty for proximal humeral fracture were prospectively followed during a nine year period; three hundred patients with a minimum of two years follow-up, at an average of forty-five months (range, 24–117), were available for review. Objective results were graded with the Constant score and range of motion. Subjective results were reported according to patient satisfaction.

At follow-up, the average Constant score was fifty-four points (range, 14–95) and active forward elevation was 103° (range, 10°–180°). Eighty-one percent of patients were satisfied or very satisfied. We observed a 59% rate of late (after three months) complications, including a high rate of tuberosity-related complications (72% malunion or nonunion). Initial tuberosity malposition was present in 35% of the patients. Secondary migration despite initial good positioning was observed in 24%.

Tuberosity complications were associated with poor final Constant score, poor range of motion and shoulder pain (p=0.001 for all items). A re-operation was performed in 5.3% of the cases. Patients who were mobilised according to the ‘early passive motion’ concept had double the incidence of secondary tuberosity migration, compared to those that were initially immobilised (14% versus 27%, p=0.004). Tuberosity complications are the most frequent late complication and they are associated with poor functional results. It is, therefore, incumbent upon the surgeon to maximise healing with adequate fixation of the tuberosities, followed by sufficient immobilization.

Unidirectional instability with or without hyperlaxity can be adressed as such; mostly traumatic changes like Bankart or capsule lesions are seen and they can be treated with the standard arthroscopic suture techniques.

Multidirectional instability (MDI) with or without hyperlaxity is quite rare. In the literature MDI is often mentioned, but most of the time it refers to multidirectional laxity with unidirectional instability. During arthroscopy often capsular redundancy is seen without obvious pathological changes. Several series have been described where arthroscopic capsulorraphy has shown to be reasonably succesfull Since the introduction of thermal shrinkage several series have been published, with poorer results compared to capsular shift: failure rates vary between 11 to 36%. The possible reason of this high failure rate is that 1) many patients with multi-directional laxity are included, 2) as well as the fact that after time the effect of the shrinkage disappears due to regeneration of the capsule. Posterior instability. In posterior subluxation, often posttraumatic, with hyperlaxity frequently as accompanying phenomenon, arthroscopic capsulorraphy has been rather successful. My personal series of 10 patients showed a failure rate of 50%. Shrinkage has been applied for this indication as well, with varying results. Since 1998 13 patients with posterior subluxations were treated in our hospital with shrinkage after the failure of extensive physical therapy. After 1 yr follow-up most of them were stable; after 18–24 months follow-up all showed recurrence of the posterior instability. The pain, often accompanying the subluxation, was however still absent at the latest follow-up.

Multidirectional instability with or without hyperlaxity is a not well defined clinical entity; for this reason the results of several treatment modalities are often not comparable. Posterior instability, especially subluxations are often posttraumatic, with some accompanying hyperlaxity. In both pathological conditions arthroscopic capsulorraphy seems to be more effective than shrinkage

Purpose: The purpose of this study was to analyse the factors affecting outcome after prosthesis implantation for osteonecrosis of the humerus in order to ascertain details concerning the indications for humerus or total prostheses in this condition.

Material and methods: Forty-seven women and 27 men, mean age 57 years with osteonecrosis of the humerus underwent surgery for implantation of 80 shoulder prostheses. According to the Arlet and Ficat classification as modified by Cruess there were five grade II, fifteen grade III, 41 grade IV and fourteen grade V shoulders, with five unclassifiable shoulders. there were 14 supraspinatus tears and five infraspinatus tears. The 26 total shoulder prostheses and 54 humeral prostheses were reviewed at a mean 47 months follow-up. Pre-operative x-rays were available for 65 shoulders and postoperative x-rays for 58.

Results: The subjective outcome was considered good irrespective of the grade (very satisfied 60%, satisfied 30%, dissatisfied 4%). The mean postoperative Constant score was 70 ± 15. The preoperative score was lower for more advanced disease but the difference was not significant at postoperative assessment. The total prosthesis provided better results in grade V shoulders and the hemiprosthesis in grade II and III shoulders.

The result of the humeral prostheses depended on the grade of necrosis. The postoperative Constant score was lower for advanced necrosis with lower scores for motion, force and pain (which was not influenced by age). At last follow-up, 31% of the x-rays revealed a significant narrowing of the glenoid-head space. The clinical outcome was significantly less satisfactory for all Constant scores. Rotator cuff tears, preoperative stiffness favoured glenoid wear. Tears of both cuff tendons led to poor outcome. Bone collapse with impaction of the head into the glenoid and medialisation of the humerus was a significant factor predictive of less satisfactory outcome.

Discussion: The grade of necrosis is a determining factor for indications for humerus or total shoulder arthroplasty for the treatment of osteonecrosis. Our objective results lead us to recommend a humeral prosthesis for grade II and III patients and a total shoulder arthroplasty for grade V patients. For grade IV patients, the choice depends on patient age, shoulder stiffness, the status of the cuff, and most importantly, collapsus of the head of the humerus with impaction-medialisation of the humerus.