The clinical features, treatment, and outcome were reviewed for 48 patients with a haematoma and 71 patients with an infarct in the posterior fossa in order to develop a rational plan of management. Clinical features alone were insufficient to make a diagnosis in about half of the series. Patients with a haematoma were referred more quickly to the neurosurgical unit, were more often in coma, and more often had CT evidence of brain stem compression and acute hydrocephalus. Ultimately, 75% of the patients with a haematoma required an operation. By contrast, most patients with an infarct were managed successfully conservatively. Early surgical management in both cerebellar haemorrhage and infarct (either external ventricular drainage or evacuation of the lesion), associated with early presentation and CT signs of brain stem compression and acute hydrocephalus, led to a good outcome in most patients. Of the patients with cerebellar haematoma initially treated by external drainage, over half subsequently required craniectomy and evacuation of the lesion; but, in some cases, this failed to reverse the deterioration. In patients with a cerebellar infarct, external drainage was more often successful. The guidelines, findings, and recommendations for future management of patients with posterior fossa stroke are discussed. (J Neurol Neurosurg Psychiatry 1995;59:287-292)

We have retrospectively reviewed 23 conscious patients, in whom a CT scan diagnosis of acute subdural haematoma was made, and in whom craniotomy for evacuation was not initially performed. These highly selected patients represent 3% of 837 patients with acute subdural haematoma, presenting over a five year, eight month period to the Institute of Neurological Sciences, in Glasgow (1986-1991). Patients with any other associated intracranial abnormalities, such as cerebral contusions, as shown on CT, were excluded from this report. All patients were followed by serial CT scanning, and neurological assessments. Cerebral atrophy was present in over half of the sample. In 17 of our patients, the acute subdural haematoma resolved spontaneously, without evidence of damage to the underlying brain, as shown by CT or neurological findings. Six subsequently required burr hole drainage of a hypodense liquid subdural haematoma. In each of these patients, haematoma thickness was greater than 10 mm. Haematoma volume was significantly larger (53 +/- 6 ml versus 32 +/- 2 ml) in the group who came to operation. The mean delay between injury and operation in this group was 15 days. We conclude that certain conscious patients with small acute subdural haematomas, without mass effect on CT, may be safely managed conservatively.

In an attempt to examine in vivo the early metabolic consequences of severe acute head injury, 1H MRS was performed in four patients from 8 to 25 h (mean 15 h) following trauma. In three of these patients, decompressive surgery was performed 4-5 h prior to the MRS. High levels of lactate (area of lactate peak >50% of the mean areas of the NAA, choline-containing, and creatine-containing compound peaks) were found at 8 h posttrauma in the one patient who was not operated on and at 10 h posttrauma in one of the patients who underwent surgery. In the other two postoperative patients, at 18 and 25 h after trauma, lactate levels were found to be low (lactate peak <20% of the mean area of the other three peaks). In the one patient who had a follow-up at 6 days and who had the largest initial lactate levels, these remained high. These findings suggest that high levels of lactate may not be an inevitable consequence of severe head injury and that similar MRS studies should be performed on each individual patient before therapies to reduce lactate are considered. There appeared to be no correlation between the relative amounts of lactate and outcome.

Aims-To study the mechanism of action of steroids in patients with peritumorous oedema. Methods-To investigate early cerebral metabolic changes proton magnetic resonance spectroscopy (1H-MRS) was used before and 11 to 14 hours after treatment with dexamethasone (12 mg oral loading and 4 mg four times daily maintenance). Nine patients (two men, seven women, mean age 54) with pronounced oedema associated with various intracranial tumours (two astrocytomas, three meningiomas, two glioblastoma, and two metastases) were examined using MRI and MRS. SE1500/135 volume selected MRS (mean volume 21 ml) were performed on an oedematous region and a contralateral region. All spectra were acquired with and without water suppression. Metabolite peak area ratios were determined. Results-Regions of oedema had significantly (P < 0-01) higher unsuppressed water than the contralateral regions, as expected. There was no change at this early time point after dexamethasone. The ratio of the area of choline containing compounds to that creatine and phosphocreatine compounds was determined after which the serial ratios of these before and after were calculated (a serial ratio of 1-0 would indicate no change in the choline to creatine ratios after steroid administration). The mean serial ratios for the area of oedema were 1'02 (SEM 0.08) and 1-10 (0.08) for the contralateral volume of interest, indicating no significant changes. However, significant changes (P < 0.02) were found in the N-acetyl-aspar-