Injuries to the tarsometatarsal (Lisfranc) joint are not common, and the results of treatment are often unsatisfactory. Since no individual is likely to see many such injuries, we decided to make a retrospective study of patients from five different centres. In this way 119 patients with injuries of the Lisfranc joint have been collected. This paper classifies these injuries and describes their incidence, mechanism of production, methods of treatment, results and complications. Sixty-nine of the patients attended for review: 35 of these had been treated by closed methods, 27 had had an open reduction and seven patients had had no treatment. On the basis of our study we suggest that these injuries should be classified according to the type of injury rather than the nature of the deforming force and that their treatment be based upon this classification. It seems that, whatever the severity of the initial injury, prognosis depends on accurate reduction and its maintenance.

A study was undertaken to assess the degree of inter-observer error when a panel of observers classified the radiographs of patients with early Perthes' disease, using Catterall grouping and "at risk" signs. The anteroposterior and lateral radiographs, taken within three months of diagnosis of Perthes' disease, were available for 69 hips and were shown in turn to 10 observers. The radiological end-results were assessed at least four years from diagnosis. The results showed a poor ability of the observers to delineate Groups 1, 2 and 3, with a more satisfactory performance in Group 4 and when Groups 2 and 3 were combined. Interpretation of "at risk" signs was unsatisfactory except when there was an increase in medial joint space greater than two millimetres. The end-results correlated well with early Catterall grouping and "at risk" signs when these were correctly interpreted.