Search

Object: The pathogenesis of traumatic brain swelling remains unclear. The generally held view is that brain swelling is caused primarily by vascular engorgement and that edema plays a relatively minor role in the swelling process. The goal of this study was to examine the roles of cerebral blood volume (CBV) and edema in traumatic brain swelling., Methods: Both brain-tissue water and CBV were measured in 76 head-injured patients, and the relative contribution of edema and blood to total brain swelling was determined. Comparable measures of brain-tissue water were obtained in 30 healthy volunteers and CBV in seven volunteers. Brain edema was measured using magnetic resonance imaging, implementing a new technique for accurate measurement of total tissue water. Measurements of CBV in a subgroup of 31 head-injured patients were based on consecutive measures of cerebral blood flow (CBF) obtained using stable xenon and calculation of mean transit time by dynamic computerized tomography scanning after a rapid bolus injection of iodinated contrast material. The mean (+/- standard deviation) percentage of swelling due to water was 9.37+/-8.7%, whereas that due to blood was -0.8+/-1.32%., Conclusions: The results of this study showed that brain edema is the major fluid component contributing to traumatic brain swelling. Moreover, CBV is reduced in proportion to CBF reduction following severe brain injury.

Leiomyosarcomas (LMSs) of the central nervous system are extremely rare; however, they are becoming more prevalent in immunocompromised patients. The authors present the cases of two patients with acquired immunodeficiency syndrome: one with LMS of the thoracic vertebral body and the other with LMS originating from the region of the cavernous sinus. The epidemiological and histological characteristics of LMS and its association with latent Epstein-Barr virus are discussed, as well as the treatments for this neoplasm.

Object: Recent animal studies demonstrate that excitatory amino acids (EAAs) play a major role in neuronal damage after brain trauma and ischemia. However, the role of EAAs in patients who have suffered severe head injury is not understood. Excess quantities of glutamate in the extracellular space may lead to uncontrolled shifts of sodium, potassium, and calcium, disrupting ionic homeostasis, which may lead to severe cell swelling and cell death. The authors evaluated the role of EEAs in human traumatic brain injury., Methods: In 80 consecutive severely head injured patients, a microdialysis probe was placed into the gray matter along with a ventriculostomy catheter or an intracranial pressure (ICP) monitor for 4 days. Levels of EAAs and structural amino acids were analyzed using high-performance liquid chromatography. Multifactorial analysis of the amino acid pattern was performed and its correlations with clinical parameters and outcome were tested. The levels of EAAs were increased up to 50 times normal in 30% of the patients and were significantly correlated to levels of structural amino acids both in each patient and across the whole group (p < 0.01). Secondary ischemic brain injury and focal contusions were most strongly associated with high EAA levels (27+/-22 micromol/L). Sustained high ICP and poor outcome were significantly correlated to high levels of EAAs (glutamate > 20 micromol/L; p < 0.01)., Conclusions: The release of EAAs is closely linked to the release of structural amino acids and may thus reflect nonspecific development of membrane micropores, rather than presynaptic neuronal vesicular exocytosis. The magnitude of EAA release in patients with focal contusions and ischemic events may be sufficient to exacerbate neuronal damage, and these patients may be the best candidates for treatment with glutamate antagonists in the future.

Cerebrospinal fluid (CSF) dynamics were correlated to the changes in ventricular size during the first 3 months posttrauma in patients with severe head injury (Glasgow Coma Scale score < or = 8, 75 patients) to distinguish between atrophy and hydrocephalus as the two possible causes of posttraumatic ventriculomegaly. Using the bolus injection technique, the baseline intracranial pressure (ICP), pressure volume index, and resistance for CSF absorption (R0) provided a three-dimensional profile of CSF dynamics that was correlated with ventricular size and Glasgow Outcome Scale (GOS) score at 3, 6, and 12 months posttrauma. Patients were separated into five different groups based on changes in ventricular size, presence of atrophy, and CSF dynamics. Group 1 (normal group, 41.3%) demonstrated normal ventricular size and normal ICP. Group 2 (benign intracranial hypertension group, 14.7%) showed normal ventricular size and elevated ICP. Group 3 (atrophy group, 24%) displayed ventriculomegaly, normal ICP, and normal R0. Group 4 (normal-pressure hydrocephalus group, 9.3%) had ventriculomegaly, normal ICP, and high R0. Group 5 (high-pressure hydrocephalus group, 10.7%) showed ventriculomegaly and elevated ICP with or without high R0. The GOS score in the nonhydrocephalic groups (Groups 1, 2, and 3) was better than in the hydrocephalic groups (Groups 4 and 5). It is concluded from these results that 44% of head injury survivors may develop posttraumatic ventriculomegaly. Posttraumatic hydrocephalus, as identified by abnormal CSF dynamics, was diagnosed in 20% of survivors and their outcome was significantly worse. This study demonstrates the importance of using CSF dynamics as an aid in diagnosis of posttraumatic hydrocephalus and identifying those patients who may benefit from shunt placement.

The investigators undertook a retrospective analysis of ventriculostomy infections to evaluate their relationship to monitoring duration and prophylactic catheter exchange. In 1984, the results of an epidemiological study of ventriculostomy-related infection were published. One of the conclusions of the paper was that the incidence of ventriculostomy-related infections rose after 5 days of monitoring. This led to the recommendation that catheters be prophylactically changed at 5-day intervals if prolonged monitoring was required. A recent randomized prospective study on central venous catheters showed no reduction in infection with prophylactic catheter exchanges. This has led the authors to reexamine their experience with ventriculostomy infections. Data on 584 severely head injured patients with ventriculostomies were prospectively collected in two data banks, The Traumatic Coma Data Bank and The Medical College of Virginia Neurocore Data Bank. These data were retrospectively analyzed for factors associated with ventriculostomy related infections. It was found that there is a relationship of ventriculitis to monitoring duration but it is not simple or linear. There is a rising risk of infection over the first 10 days, but infection then becomes very unlikely despite a population that continues to be at risk. Patients in whom catheters were replaced prior to 5 days did not have a lower infection rate than those whose catheters were exchanged at more than 5-day intervals. Based on these data, it is recommended that ventriculostomy catheters for intracranial pressure monitoring be removed as quickly as possible, and in circumstances in which prolonged monitoring is required, there appears to be no benefit from catheter exchange.

Although mortality and morbidity rates from head injury have been reduced substantially by improved prehospital interventions, intensive care, and aggressive management of intracranial pressure (ICP), successful treatment of the primary brain injury has been elusive. In experimental models, tromethamine (THAM) has been effective in treating head injury; this drug acts by entering the cerebrospinal fluid compartment, reducing cerebral acidosis and ICP, and reversing the adverse effects of prophylactic hyperventilation on early recovery. In this randomized prospective clinical trial, THAM was studied to determine if it had beneficial effects in the early management of severe head injuries and if the adverse effects of hyperventilation could be prevented. A total of 149 patients with severe head injury (Glasgow Coma Scale scores of < or = 8) were randomly assigned to either a control or a THAM group. Both groups of patients matched in terms of clinical parameters, including age, sex, number of surgical mass lesions, number in each Glasgow Coma Scale stratum, and first ICP measurement. All patients were treated by a standard management protocol, intubated, mechanically ventilated, and maintained in the pCO2 range of 32 to 35 mm Hg for 5 days. Tromethamine was administered as a 0.3-M solution in an initial loading dose (body weight x blood acidity deficit, average 4.27 cc/kg/hr) given over 2 hours, followed by a constant infusion of 1 ml/kg/hr for 5 days. Outcome was measured at 3, 6, and 12 months postinjury. Although analysis indicated no significant difference in outcome between these two groups at 3 months, 6 months, and 1 year, there was a difference regarding ICP. The time that ICP was above 20 mm Hg in the first 48 hours postinjury was less in patients treated with THAM (p < 0.05). Also, the number of patients requiring barbiturate coma was significantly less in the THAM group (5.48% vs. 18.4%, p < 0.05). The authors conclude that THAM ameliorates the deleterious effect of prolonged hyperventilation, may be beneficial in ICP control, and warrants further study as to the dosage and timing of administration.

There is still controversy over whether or not patients should be hyperventilated after traumatic brain injury, and a randomized trial has never been conducted. The theoretical advantages of hyperventilation are cerebral vasoconstriction for intracranial pressure (ICP) control and reversal of brain and cerebrospinal fluid (CSF) acidosis. Possible disadvantages include cerebral vasoconstriction to such an extent that cerebral ischemia ensues, and only a short-lived effect on CSF pH with a loss of HCO3-buffer from CSF. The latter disadvantage might be overcome by the addition of the buffer tromethamine (THAM), which has shown some promise in experimental and clinical use. Accordingly, a trial was performed with patients randomly assigned to receive normal ventilation (PaCO2 35 +/- 2 mm Hg (mean +/- standard deviation): control group), hyperventilation (PaCO2 25 +/- 2 mm Hg: HV group), or hyperventilation plus THAM (PaCO2 25 +/- 2 mm Hg: HV + THAM group). Stratification into subgroups of patients with motor scores of 1-3 and 4-5 took place. Outcome was assessed according to the Glasgow Outcome Scale at 3, 6, and 12 months. There were 41 patients in the control group, 36 in the HV group, and 36 in the HV + THAM group. The mean Glasgow Coma Scale score for each group was 5.7 +/- 1.7, 5.6 +/- 1.7, and 5.9 +/- 1.7, respectively; this score and other indicators of severity of injury were not significantly different. A 100% follow-up review was obtained. At 3 and 6 months after injury the number of patients with a favorable outcome (good or moderately disabled) was significantly (p less than 0.05) lower in the hyperventilated patients than in the control and HV + THAM groups. This occurred only in patients with a motor score of 4-5. At 12 months posttrauma this difference was not significant (p = 0.13). Biochemical data indicated that hyperventilation could not sustain alkalinization in the CSF, although THAM could. Accordingly, cerebral blood flow (CBF) was lower in the HV + THAM group than in the control and HV groups, but neither CBF nor arteriovenous difference of oxygen data indicated the occurrence of cerebral ischemia in any of the three groups. Although mean ICP could be kept well below 25 mm Hg in all three groups, the course of ICP was most stable in the HV + THAM group. It is concluded that prophylactic hyperventilation is deleterious in head-injured patients with motor scores of 4-5.(ABSTRACT TRUNCATED AT 400 WORDS)

In the belief that secondary cerebral compression, hypoxia, and ischemia materially influence the outcome from severe head injury, a standardized protocol was followed in 160 patients, with emphasis on early diagnosis and evacuation of intracranial mass lesions by craniotomy, artificial ventilation, control of increased intracranial pressure, and aggressive medical therapy. Of these patients, 36% made a good recovery, 24% were moderately disabled, 8% were severely disabled, 2% were vegetative, and 30% died. The mortality rate compares favorably with outcomes in similar patients reported from other centers and there has been no increase in the numbers of severely disabled or vegetative patients. It is proposed that vigorous surgical and medical therapy, by preventing or reversing secondary cerebral insults, enables some patients who would have died to make a good recovery without increasing the proportion of severely disabled patients.

Measurements of intracranial pressure (ICP) were begun within hours of injury in 160 patients with severe brain trauma, and continued in the intensive care unit. Some degree of increased ICP (greater than 10 mm Hg) was present on admission in most cases (82%), and in all but two of the 62 patients with intracranial mass lesions requiring surgical decompression; ICP was over 20 mm Hg on admission in 44% of cases, and over 40 mm Hg in 10%. In patients with mass lesions only very high ICP (greater than 40 mm Hg) on admission was significantly associated with a poor neurological picture and outcome from injury, while in patients with diffuse brain injury any increase in ICP above 10 mm Hg was associated with a poorer neurological status and a worse outcome. Despite intensive measures aimed at prevention of intracranial hypertension, ICP rose over 20 mm Hg during the monitoring period in 64 of the 160 patients (40%). Postoperative increases in ICP over 20 mm Hg (mean) were seen in 52% of the patients who had had intracranial masses evacuated, and could not be controlled by therapy in half of these cases. Even in patients without mass lesions, ICP rose above 20 mm Hg in a third of the cases, despite artificial ventilation and steroid therapy. Of the 48 patients who died, severe intracranial hypertension was the primary cause of death in nearly half and even moderately increased ICP (greater than 20 mm Hg) was associated with higher morbidity in patients with mass lesions and those with diffuse brain injury. Measurement of ICP should be included in management of patients with severe head injury.

There has been a rapid expansion of knowledge in the field of nutrition and metabolism with regard to the general surgical patient. However, only recently has there been greater appreciation of the benefits of adequate nutrition and appropriate metabolic care of the neurosurgical patient. In this review, the authors attempt to outline 1) the metabolic response to stress in general, and how it applies to the neurosurgical patient; 2) how best to provide adequate nutritional support for the neurosurgical patient; 3) the effects of nutrition on neurotransmitters; and 4) the effect of diet and nutrition on patients with malignant brain tumors.

Hypoxemia is a nearly constant accompaniment of head injury. Diverse theories have been proposed to explain this relationship. The authors report the case of a patient who suffered an episode of severe, transient, arterial oxygen desaturation during "controlled" brain trauma: an otherwise uneventful stereotaxic biopsy of a small germinoma of the hypothalamus. Evidence is provided that pure ventilation-perfusion mismatching, without pulmonary edema, underlay the hypoxemia. The hypothalamus is intimately involved in matching pulmonary ventilation to perfusion; the hypoxemia of various brain injuries may be mediated by perturbation of this structure.

Brain-tissue acidosis inferred by cerebrospinal fluid (CSF) lactic acidosis is considered to play an important role in the clinical course of severe head injury. Ventricular CSF lactate concentration was studied in 19 patients during the first 5 days after severe head injury. All patients were intubated, paralyzed, and artificially ventilated so that PaCO2 was kept at 33.2 +/- 5.0 mm Hg and PaO2 at 122 +/- 18 mm Hg (mean +/- standard deviation). The mean Glasgow Coma Scale score on admission was 5.73 +/- 2.42. The first CSF sample was drawn within 18 hours after head injury. Over the first 4 days postinjury, patients with a poor outcome had significantly higher ventricular CSF lactate levels than did those with moderate disabilities or a good outcome. Patients showing favorable outcome had a significant decrease in ventricular CSF lactate levels 48 hours after injury. This decrease was not observed in patients with a poor outcome. Increased ventricular CSF lactate concentration was also reliably associated with increased intracranial pressure (ICP). Ventricular CSF lactate levels did not correlate with the magnitude of intraventricular bleeding. Arterial and jugular venous blood lactate levels, although high after head injury, were usually lower than the levels in the ventricular CSF and reached a normal range by the 3rd day following head trauma. At that time, the ventricular CSF lactate concentration was still above normal in patients with a poor outcome but had decreased to normal in patients with moderate disabilities or a good outcome. Ventricular CSF pH did not generally correlate with the ventricular CSF lactate concentration in patients under controlled ventilation; however, in a few patients close to death or with ventricular infection, a correlation was noted. Ventricular CSF lactate levels were not related to cerebral blood flow. In this study, profiles of ventricular CSF lactate concentration are defined in relation to the patients' clinical course and outcome. High ventricular CSF lactate concentration is present within 18 hours after severe head injury. Its decrease to normal in the following 48 hours is a reliable sign of clinical improvement; however, ventricular CSF lactate levels that are persistently high or that increase over time indicate the patient's deterioration. Serial assessment of ventricular CSF for acid-base status and metabolites in head-injured patients with a ventricular catheter already placed for ICP monitoring is useful in the evaluation of prognosis and clinical course.

The authors studied intracranial pressure (ICP) and intracranial compliance as defined by the pressure-volume index (PVI) in 34 severely head-injured patients with a Glasgow Coma Scale score of 8 or less. The objective of the research was to determine if there was a correlation between the pressure-volume status and subsequent increase in ICP. The PVI and ICP measurements were obtained serially, and the temporal course of the pressure-volume status and ICP was determined during the 5-day period following injury. Aggressiveness of ICP was quantified by a therapy intensity level scale. A clear relationship between the PVI measured soon after injury and subsequent development of ICP emerged. Following mechanical trauma the PVI is reduced, and the degree of reduction and extent of biomechanical recovery are closely related to outcome and development of raised ICP.

The posttraumatic creatine kinase-BB isoenzyme (CKBB) activity and lactate concentration in ventricular cerebrospinal fluid (CSF) have been studied in 29 patients with severe head injuries. The CKBB activity reaches its maximum a few hours after trauma, and has a monoexponential drop with a half-time of approximately 10 hours. Ventricular CSF lactate concentration continues to rise in patients with a poor outcome, and decreases only slowly and inconsistently in most of the other patients. Thus, increase of lactate in the ventricular CSF is not, like CKBB, a direct one-stage consequence of the trauma but is due to continuous production from a derangement of metabolism caused by the trauma. Since even higher ventricular CSF lactate levels can be survived when not caused by head injury, and since no significant pH changes were related to the ventricular CSF lactic acidosis in these artificially ventilated patients, it is concluded that ventricular CSF lactic acidosis is indicative of a severe, although not necessarily intractable, disturbance of brain function associated with intracellular lactate production and acidosis.

Data from 523 patients admitted to the Medical College of Virginia with severe head injury and known 6-month outcomes were analyzed in order to determine the optimal combination of early-available prognostic factors. Twenty-one prognostic indicators noted in the emergency room at admission were used to predict outcomes into four categories: good, moderately disabled, severely disabled, or vegetative/dead. A combination of the patient's age (in years), the best motor response (graded in the usual six-point scale), and pupillary response (in both eyes) was found to be the most accurate indicator. The model correctly predicted outcome into one of the four outcome categories in 78% of cases ("specifically accurate predictions"). If predictions into an outcome category adjacent to the actual outcome were accepted, this model was accurate in 90% of cases ("grossly accurate predictions"). A set of three simple graphs based on this model can be used for rapid early estimation of probable outcome in a severely head-injured patient at admission.

A consecutive series of 330 severely head-injured patients was studied prospectively. All of the patients were treated with the same protocols by the same physicians and staff in the same intensive care unit. All of the patients had intracranial pressure (ICP) monitoring. Of the 330 patients, 100 were in the pediatric age group (0 to 19 years of age) and 230 were in the adult group (20 to 80 years of age). Statistical analyses were performed with regard to outcome, Glasgow Coma Scale (GCS) score, ICP course, and incidence of surgical lesions. The average emergency room GCS score as well as the 24-hour GCS score for each group was the same. The percentage of patients having ICP that was normal, increased but reducible, and increased but not reducible in each group was the same. The pediatric patients had a significantly higher percentage of good outcomes (43%) than the adult patients (28%) (p less than 0.01). They also had a significantly lower mortality rate (24%) than the adult patients (45%) (p less than 0.01). At 1 year following injury, 55% of pediatric patients had a good outcome compared to 21% of adults (p less than 0.001); this trend was evident at 3 months, with the same p value. Pediatric patients with normal ICP had a higher percentage of good outcomes (70%) than the adult patients with normal ICP (48%) (p less than 0.05). There was no significant difference in outcome in pediatric and adult patients with mass lesions or with increased ICP, regardless of whether or not the pressure was reducible. There was a much higher incidence of surgical mass lesions in adult patients (46%) than in pediatric patients (24%) (p less than 0.001).

Autoregulation of cerebral blood flow ("CBF15") was tested in a series of 26 pediatric patients (mean age 13.2 years) with severe head injury (average Glasgow Coma Scale (GCS) score 5.5) in the acute stage. A baseline 133Xe CBF measurement was performed and then repeated, after blood pressure was increased by 29% with intravenous phenylephrine or decreased by 26% with intravenous trimethaphan camsylate. Correlations were made between CBF and clinical condition, outcome, time after injury, intracranial pressure (ICP), and pressure-volume index (PVI) changes, and the site of injury (hemispheres, diencephalon, or brain stem). The site of injury was determined with multimodality evoked potential measurements. Autoregulation was intact in 22 (59%) of 37 measurements. There was no correlation with GCS score, outcome, time after injury, site of injury, or way of testing (decreasing or increasing blood pressure). Autoregulation was statistically significantly more often impaired when CBF was either below normal -2 standard deviations (SD) (reduced flow) or above normal +2 SD (absolute hyperemia). In cases with intact autoregulation, mean ICP decreased from 17.5 to 15.0 mm Hg with higher blood pressure and increased from 19.0 to 21.3 mm Hg with lower blood pressure. When PVI was measured during the blood pressure manipulations, it was found to change in a direction opposite to the ICP change. The consequences of these findings in the management of ICP problems with blood pressure control are discussed.

An analysis of clinical and demographic data of 264 patients with severe head injury showed that a combination of the Glasgow Coma Scale (GCS) score, oculocephalic responses, and age can provide a simple but reliable prediction of outcome in severe head injury. Addition of other clinical data, excluding intracranial pressure and evoked potentials, improved the predictability only negligibly. A simple chart, which is constructed from the application of the logistic regression model, can be used to determine the odds of a good outcome from the combination of the three factors. A method is given by which the GCS score of a patient with a missing verbal response score can be accurately approximated in order to complete the chart. Among other values, the odds of a good outcome provide the clinician with a reliable measure of the relative severity of a patient's injury. The accuracy of the chart in prediction is expected to be 80% or above.

The literature suggests that in children with severe head injury, cerebral hyperemia is common and related to high intracranial pressure (ICP). However, there are very few data on cerebral blood flow (CBF) after severe head injury in children. This paper presents 72 measurements of cerebral blood flow ("CBF15"), using the 133Xe inhalation method, with multiple detectors over both hemispheres in 32 children aged 3 to 18 years (mean 13.6 years) with severe closed head injury (average Glasgow Coma Scale (GCS) score 5.4). In 25 of the children, these were combined with measurements of arteriojugular venous oxygen difference (AVDO2) and of cerebral metabolic rate of oxygen (CMRO2). In 30 patients, the first measurement was taken approximately 12 hours postinjury. In 18 patients, an indication of brain stiffness was obtained by withdrawal and injection of ventricular cerebrospinal fluid and calculation of the pressure-volume index (PVI) of Marmarou. The CBF and CMRO2 data were correlated with the GCS score, outcome, ICP, and PVI. Early after injury, CBF tended to be lower with lower GCS scores, but this was not statistically significant. This trend was reversed 24 hours postinjury, as significantly more hyperemic values were recorded the lower the GCS score, with the exception of the most severely injured patients (GCS score 3). In contrast, mean CMRO2 correlated positively with the GCS score and outcome throughout the course, but large standard deviations preclude making predictions based on CMRO2 measurements in individual patients. Early after injury, there was mild uncoupling between CBF and CMRO2 (CBF above metabolic demands, low AVDO2) and, after 24 hours, flow and metabolism were completely uncoupled with an extremely low AVDO2. Consistently reduced flow as found in only four patients; 28 patients (88%) showed hyperemia at some point in their course. This very high percentage of patients with hyperemia, combined with the lowest values of AVDO2 found in the literature, indicates that hyperemia or luxury perfusion is more prevalent in this group of patients. The three patients with consistently the highest CBF had consistently the lowest PVI: thus, the patients with the most severe hyperemia also had the stiffest brains. Nevertheless, and in contrast to previous reports, no correlation could be established between the course of ICP or PVI and the occurrence of hyperemia, nor was there a correlation between the levels of CBF and ICP at the time of the measurements. The authors argue that this lack of correlation is due to: 1) a definition of hyperemia that is too generous, and 2) the lack of a systematic relationship between CBF and cerebral blood volume

The authors studied the relative contribution of cerebrospinal fluid (CSF) and vascular parameters to the level of intracranial pressure (ICP) in 34 severely head-injured patients with a Glasgow Coma Scale score of less than 8. This was accomplished by first characterizing the temporal course of CSF formation and outflow resistance during the 5-day period postinjury. The CSF formation and outflow resistance were obtained from pressure responses to bolus addition and removal of fluid from an indwelling ventricular catheter. The vascular contribution to the level of ICP was assessed by withdrawing fluid at its rate of formation and observing the resultant change in equilibrium ICP level. It was found that, with the exception of patients with subarachnoid hemorrhage, CSF parameters accounted for approximately one-third of the ICP rise after severe head injury, and that a vascular mechanism may be the predominant factor in elevation of ICP.

In certain subgroups of severely head-injured patients, the mortality rate remains unacceptably high. The authors describe a randomized, controlled trial of prophylactic pentobarbital therapy in a group of these patients. Pentobarbital was started as soon as possible after the head injury, regardless of the intracranial pressure (ICP), and was continued for a prescribed period of time. The study included 53 consecutive head-injured patients over the age of 12 years, who had either an acute intradural hematoma (subdural and/or intracerebral, large enough to warrant surgical decompression), or no mass lesion but whose best motor response was abnormal flexion or extension. All patients in the study were randomly assigned to a control group (26 cases) or a pentobarbital-treated group (27 cases) once the diagnosis had been made and informed consent obtained. All patients were treated with the same protocol of aggressive resuscitation, prompt diagnosis and treatment of mass lesions, and intensive care, with close follow-up monitoring. The randomization was effective in producing a close match between the control and treated groups with respect to age, sex distribution, cause of injury, neurological status, intracranial lesions, prevalence of early systemic insults, midline shift, and initial ICP. Outcome was essentially the same in each group. There was no difference between groups in the incidence of elevated ICP, the duration of ICP elevation, or the response of ICP elevations to treatment. Arterial hypotension occurred in 14 patients (54%) in the treated group and only two patients (7%) in the untreated group. Based on these data the authors cannot recommend the prophylactic use of pentobarbital coma in the treatment of patients with severe head injury. They also believe that its use is accompanied by significant side effects which can potentially worsen the condition of a patient with severe head injury.