346 BRIEF REPORTS THE JOURNAL OF BONE AND JOINT SURGERY

The patient made an uncomplicated recovery, her haemothorax resolved and she was discharged after seven days. Follow-up examination including an MRI scan of the spine, 13 months later, showed no evidence of recurrence ofthe tumour or of other pathological changes. Discussion. The unusual feature of our patient was the presentation with symptoms of dyspnoea and a haemo-thorax, making the initial diagnosis difficult. A CT scan helped, but the final diagnosis of ABC was established only after surgical exploration and histological examina-tion. There are two possible explanations of the haemo-thorax. First, penetration of a segmental vessel by the expanding tumour; however, exploration did not reveal such a bleeding vessel. Secondly, spontaneous rupture of the ABC following perforation of its outer shell; to the best of our knowledge this complication has never been reported. On the contrary it is well documented that ABC, although it may grow at an alarming rate, remains well contained by a rim of reactive bone (Enneking 1983); if this is absent, the thickened periosteum and the compressed soft tissues circumscribe the lesion (Spjut et al 1971) and therefore no spontaneous bleeding occurs. However, profuse bleeding following evacuation of the blood clot from within the cavity does occur. The source of this bleeding seems to be the highly vascular lining that forms the inside wallofthe reactive bone (Schajowicz 1981). In our case the thin shell of the tumour was ruptured at one site and we therefore assume that a spontaneous evacuation of the clot took place into the pleural cavity and caused the haemothorax. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article