Between 1971 and 1991 we treated 98 patients with giant-cell tumours, 15 of whom presented with a pathological fracture. They were most common around the knee (12). Nine fractures were intra-articular. The tumours were treated by curettage and acrylic cementing (10), excision and endoprosthesis (1), excision and allograft (1), curettage and autologous graft (2) or by resection of the fibular head (1). Four patients had local recurrence, three of whom were cured by repeat curettage and cementing. Pathological fracture through a giant-cell tumour is not a contraindication to treatment by curettage and acrylic cementing.

From 1974 to 1989, we treated 50 patients with a simple dislocation of the hip: 38 were posterior dislocations and 12 were anterior. All dislocations primarily treated at our hospital were reduced by closed methods within three hours (mean 85 minutes (10 to 180)) and 43 were reviewed after an average follow-up of 8 years (2 to 17). It is widely held that isolated hip dislocation reduced within six hours gives an excellent outcome, but we found a significant number of complications. There were radiological signs of partial avascular necrosis in two, mild osteoarthritis in seven, and moderate degeneration in two. Heterotopic ossification was seen in four patients, but 29 of 33 MRI examinations were normal. Objective evaluation according to the Thompson and Epstein (1951) criteria showed fair and poor results in 3 of 12 anterior dislocations, but in 16 of 30 posterior dislocations. In six of the seven patients with no other severe injury, the hip had an excellent or good result; in only three of the eight patients with severe multiple injuries was this the case. The important factors in the long-term prognosis appear to be the direction of the dislocation and the overall severity of injuries.

We analysed metastasis-free survival after local recurrence in a population-based series of 432 patients treated for soft-tissue sarcoma of the extremities or trunk wall. Local recurrence was found in 124 patients; of these, 73 had no detectable metastases when the last local recurrence was diagnosed. No patient received chemotherapy for local recurrence. The 73 patients had a 5-year metastasis-free survival rate of 0.73 after the last local recurrence, compared with 0.47 in all 124 patients with local recurrence. The 5-year metastasis-free survival rate was 0.76 in the 308 patients who had never developed local recurrence, and 0.68 in the whole series. The selection of patients with local recurrence but no concurrent metastasis creates a subset of the population with a good prognosis; this makes it difficult to evaluate the effect of treatment for local recurrence on survival rate. In studies of prognosis this group of patients should be analysed separately from patients with primary tumours.