Preliminary experience with the occasional good survival of patients in Hunt and Hess Grade IV or V with aneurysmal subarachnoid hemorrhage (SAH) led to a prospective management protocol employed during a 2 1/2-year period. The protocol utilized computerized tomography (CT) scanning to diagnose SAH and to obtain evidence for irreversible brain destruction, consisting of massive cerebral infarction with midline shift or dominant basal ganglia or brain-stem hematoma. These patients, along with those who exhibited poor or absent intracranial filling on CT or angiography, were excluded from active treatment and given supportive care only. All other patients had immediate ventriculostomy placement and, if intracranial pressure (ICP) was controllable (less than or equal to 30 cm H2O without an intracranial clot or less than or equal to 50 cm H2O in the presence of a clot), went on to have craniotomy for aneurysm clipping. Aggressive postoperative hypertensive, hypervolemic, hemodilutional therapy was subsequently employed. Of 54 patients with poor-grade aneurysms, ventriculostomy was placed in 47 (87.0%) and yielded high ICP's in the overwhelming majority, with the mean ICP being 40.2 cm H2O. Nineteen poor-grade aneurysm patients received no surgical treatment and survived a mean of 31.8 hours with 100% mortality. Thirty-five patients underwent placement of a ventriculostomy, craniotomy for aneurysm clipping and intracranial clot evacuation, and postoperative hypertensive, hypervolemic, hemodilutional therapy. The outcome at 3 months of the 35 patients who were selected for active treatment was good in 19 (54.3%), fair in four (11.4%), poor in four (11.4%), and death in eight (22.9%). It is concluded that poor-grade aneurysm patients usually present with intracranial hypertension, even those without an intracranial clot. Based on radiographic rather than neurological criteria, a portion of these patients can be selected for active and successful treatment. Increased ICP can be present without ventriculomegaly, and immediate ventriculostomy should be performed. As long as ICP is controllable, craniotomy and postoperative intensive care can effect a favorable outcome in a significant percentage of these patients.

Four patients underwent a far lateral-combined craniotomy procedure for extensive tumors of the clivus and craniocervical junction. Their presentation, operative, and clinical course are discussed. All patients had improved at their follow-up examination (mean follow-up, 10.7 months). This approach can now be applied to extensive tumors of the petroclival region and craniocervical junction to decrease morbidity.

Our experience with the use of external lumbar subarachnoid drainage in 5 children with severe diffuse head injuries is presented. All patients had Glasgow Coma Scale scores of 8 or less at 24 h after injury and were initially treated with ventriculostomies. Two children required surgical evacuation of focal mass lesions. Within 72 h of admission, all children manifested high intracranial pressures (ICP) refractory to maximal therapy, including hyperventi-lation, furosemide, mannitol, and barbiturate coma. After the institution of lumbar drainage, 3 of the 5 children had an abrupt and lasting decrease in ICP, obviating the need for continued barbiturates and hyperventilation. Three children survived, 2 of whom made good recoveries; 1 child is functional with disability. ICP varied passively with the height of the drainage bag in these surviving patients. Two patients died, most likely from uncontrolled ICP before the lumbar drain was placed. We conclude that controlled external lumbar subarachnoid drainage is a potentially useful treatment for severe diffuse pediatric head injury when maximal medical therapy and ventricular cerebrospinal fluid (CSF) evacuation have failed to control high ICP. Posttraumatic CSF circulation disruption, white matter cerebral edema, and intracranial venous hypertension can be treated with this modality in the absence of focal mass lesions.