We have reviewed 66 consecutive Souter-Strathclyde arthroplasties of the elbow implanted in 59 patients between 1982 and 1993. Thirteen patients (15 elbows) (19.6%) died. Sixteen elbows (24.2%) were revised, six for aseptic loosening (9%), four (6%) because of fracture or loosening after a fracture, three (4.5%) for infection and three (4.5%) for dislocation. Four patients refused to attend for review.

In 33 elbows with a follow-up of 93 months (60 to 167) complete relief of pain was achieved in 22 (67%) when seen at one year. After ten years or more 36% of the elbows were painfree and 64% had occasional slight pain especially under loading or stress. The mean gain in the arc of movement was 16°, but a mean flexion contracture of 33° remained. The main early complications were intraoperative fractures of the epicondyles (9%), postoperative dislocation (4.5%) and ulnar neuropathy. The incidence of ulnar neuropathy before operation was 19%. After operation 20 patients (33%) had an ulnar neuropathy, in seven of which it had been present before operation, and of these weakness of the hypothenar muscle occurred in two. The probability of survival of the Souter-Strathclyde elbow prosthesis based on the Kaplan-Meier calculation is 69% at ten years.

We report the long-term results of the Matti-Russe operation for pseudarthrosis of the scaphoid in 100 cases, reported previously by Mulder in 1968. Clinical results for 77 patients and radiographic data for 74 were reviewed at 22 to 34.8 years after surgery. In general, there was satisfactory relief of pain and stiffness but some patients had limitation of motion and reduced grip-strength, with usually slight osteoarthritic changes. There was poor correlation between subjective, objective, and radiographic results but 88% of the patients were satisfied with their results.

Fourteen cases of hemiarthroplasty for four-part fractures of the proximal humerus were reviewed. Pain relief was satisfactory, but function was limited, mainly due to loss of glenohumeral abduction despite electromyographic proof of actively contracting abductors in all cases. Analysis of special radiographs of nine cases showed a direct relationship between the clinical results and the "humeral offset", or distance between the geometric centre of the humeral head and the lateral aspect of the greater tuberosity. This offset affects the lever arms of the glenohumeral abductor muscles. The implications for surgical technique and for the design of shoulder prostheses are discussed.

A study of structures which obstruct reduction of hip dislocation was performed on 15 hips by magnetic resonance imaging (MRI). Before treatment started, MRI studies were performed on 10 patients, six of whom were treated conservatively, after which further MRI studies helped to establish a concentric reduction. In the other four conservative treatment failed and they were operated on; in them the MRI studies were compared with arthrographic and surgical findings. In all but one of these 10 patients, MRI enabled us to differentiate between an everted and an inverted limbus. In five other patients with unsatisfactory development of the hip following closed reduction, MRI was compared with earlier arthrographic studies. MRI provided accurate anatomical information which would not have been obtained by arthrography. It clearly has great potential in assisting the surgeon to select the appropriate form of treatment.