One hundred and two consecutive patients with cerebral haematoma were prospectively allocated to one of two groups according to their CT on admission. Group 1 patients were thought to have a high probability of an underlying structural lesion and underwent cerebral angiography acutely. If normal, this was repeated at three months. Group 2 patients were thought not to have such a lesion and underwent angiography at three months. Patients older than the mean age of the study population, and hypertensive patients were much less likely to harbour an aneurysm or arteriovenous malformation than younger or normotensive patients (p < 0-01; sensitivity 87-9%, specificity 88.6%). An aneurysm or arteriovenous malformation was the cause of haemorrhage in 12-8% of hypertensive patients, in 309% of patients with haematoma involving the basal ganglia, and 18-2% of those with posterior fossa haemorrhage. Features of CT in isolation give a sensitivity of 77-2% and a specificity of 84-2%. Follow up angiography in group 1 showed an arteriovenous malformation in one of seven patients in whom the original study was normal. These results contrast sharply with data from previous retrospective studies. The decision to investigate a patient with cerebral haematoma should be primarily based on the patient's clinical condition, rather than on the site of haemorrhage. (7 Neurol Neurosurg Psychiatry 1994;57: 1180-1186 Stroke is caused by spontaneous cerebral haemorrhage in 4% to 119% of patients"2 although rates of up to 40% have been quoted.3 Acute mortality ranges from 18°% to 75%, depending on the size and position of the haematoma and the age and clinical condition of the patient.'45 To prevent recurrence, it is important to identify patients in whom haemorrhage is secondary to an underlying structural lesion, such as an aneurysm, vascular malformation, or tumour. Pathological studies have indicated that primary haemorrhage is most likely to occur in certain sites-in particular the basal ganglia and posterior fossa.6 Factors such as -age, pre-existing systemic hypertension, and haematoma location are commonly employed in selecting patients for angiography. The influence of clinical features and CT on final diagnosis has been described retrospectively,7-" but to our knowledge there has been no prospective angiographic assessment in life of the relative frequency of primary and secondary haemorrhage in patients with cerebral haematoma.
MethodsAll patients presenting consecutively to this hospital between 1989 and 1992 with spontaneous cerebral haematoma were entered into the study. They were assigned to one of two groups after evaluation of their CT on admission by two or three consultant neuroradiologists. If the haematoma was thought likely to be due to an underlying structural cause, angiography was carried out as soon as the patient's condition allowed, and was repeated three months later if negative (group 1). If, however, such a lesion was not suspected, angiography was delayed by three months (group 2).The radiologists were...