Postischemic canine cerebral blood flow appears to be determined by cerebral metabolic needs.

Following a period of complete global cerebral ischemia and reperfusion there ensues a low flow state referred to as the delayed postischemic hypoperfusion state. It is unknown whether this low flow state contributes to neuronal injury or whether the magnitude of hypoperfusion correlates with the duration of ischemia. The latter question was addressed in 20 dogs in which complete global ischemia was induced by cerebrospinal fluid (CSF) compression for periods of 3, 9, 12, or 18 min. Following reperfusion, CBF (by sagittal sinus outflow) and CMRO2 were determined for 90 min, and results were correlated with the duration of ischemia. At 90 min postischemia the magnitude of decrease in CBF correlated crudely with the duration of ischemia (r = -0.67, p less than 0.01). For CMRO2 correlation of the magnitude of decrease with the duration of ischemia was more evident (r = -0.74, p less than 0.001). Furthermore, the postischemic ratio of CBF to CMRO2 was virtually identical for all dog groups regardless of the ischemic time. The adequacy of the ratio of CBF to CMRO2 was reflected by adequate oxygen levels in the sagittal sinus blood of all dogs. The authors conclude that the delayed postischemic hypoperfusion state is probably not an important determinant of neuronal injury since its magnitude appears to be primarily determined by the metabolic needs of the brain.

Nimodipine does not affect cerebral lactate levels following complete ischemia in dogs.

Nimodipine is known to improve postischemic cerebral blood flow (CBF) and neurologic outcome in experimental animals. Whether or not the two observations are related is unknown. This study searched for a possible improved rate of brain metabolic recovery in animals treated with nimodipine postischemia. Complete cerebral ischemia was produced for 11 min in 16 dogs, followed by reperfusion for 70 min. Prior to ischemia, glucose was administered (0.75 g X kg-1) in 12 dogs. Half of the glucose-treated dogs were given i.v. nimodipine, beginning 5 min postischemia (10 micrograms X kg-1 bolus followed by 1 microgram X kg-1 X min-1). The other half were given only saline postischemia. The remaining four dogs were given no glucose and received saline only postischemia. In all dogs, serial brain biopsies were taken at 2, 20, 40, and 70 min postischemia. In 5 dogs, the integrity of the blood-brain barrier (BBB) was tested by injection of Evans blue dye and postmortem examination of the brains. Brain biopsies were assayed for concentrations of phosphocreatine, ATP, ADP, AMP, glucose, lactate, and pyruvate. In all dogs, there was rapid restoration of a normal brain energy state following reperfusion. Brain lactate had returned to near normal in all dogs by 70 min postischemia, and the rate of lactate depletion was not different between groups. The integrity of the BBB was only minimally affected. A portion of the brain lactate was converted to pyruvate rather than crossing the BBB.(ABSTRACT TRUNCATED AT 250 WORDS)

The effect of profound levels of hypothermia (below 14 degrees C) on canine cerebral metabolism.

The goal of this study was to determine the temperature coefficient (Q10) for canine CMRO2 at temperatures below 14 degrees C. Eight dogs were anesthetized with halothane for surgical preparation. The animals were placed on total cardiopulmonary bypass and CBF was measured by direct sagittal sinus outflow. Duplicate measurements were taken at 37, 13, and 7 degrees C. The EEG became isoelectric at a temperature of 12.0 +/- 0.8 degrees C. The Q10 between 13 and 7 degrees C was 2.19 +/- 0.59. With rewarming to 37 degrees C, cerebral metabolic variables returned to control levels. Brain biopsies taken at the end of the study yielded normal values for brain energy stores. We conclude that the Q10 for CMRO2 at temperatures between 7 and 37 degrees C can be profoundly affected by the state of cerebral function as reflected by the EEG. In the absence of EEG activity, an expected Q10 value of 2.2 reflects only the direct effect of temperature on the rates of biologic reactions.

Canine cerebral metabolism and blood flow during hypoxemia and normoxic recovery from hypoxemia.

There are conflicting reports regarding the effects of hypoxemia on the cerebral metabolic rate for oxygen (CMRO2). Accordingly, we examined the changes in CMRO2 during normoxia, progressive hypoxia (PaO2 of 37, 27, and 23 mm Hg), and normoxic recovery from hypoxia. Measurements were made in dogs anesthetized with nitrous oxide (60-70%) and halothane (less than 0.1%) in oxygen. Arterial-cerebral venous blood oxygen content differences and cerebral blood flow (CBF) were measured simultaneously, the latter by a technique (collection of sagittal sinus outflow) previously validated for conditions of near-maximal CBF. The duration of each of the three hypoxic exposures was approximately 10 min. CMRO2 was significantly decreased (14%) only when the arterial blood oxygen tension was reduced to 23 mm Hg. CBF increased progressively to a maximum of 153% of control. Posthypoxemic brain biopsy values for cerebral metabolites obtained 40 min after normoxemia had been restored were normal. These results, in conjunction with an unchanged CMRO2 at 40 min normoxic recovery, suggest that no gross irreversible brain cell damage occurred. We conclude that with progressive hypoxemia, CMRO2 remains stable until oxygen demand exceeds oxygen delivery, resulting thereafter in a progressive reduction in CMRO2.

Postischemic canine cerebral blood flow is coupled to cerebral metabolic rate.

Following complete global cerebral ischemia and reperfusion, a brief period of reactive hyperemia is followed by a prolonged period of low flow commonly referred to as the delayed postischemic hypoperfusion state. It is generally assumed that this low-flow state may be injurious because of inadequate substrate delivery, thus implying that flow is no longer coupled to metabolic needs. This relationship of CBF to CMRO2 was examined in six anesthetized dogs that were subjected to 12 min of complete ischemia induced either by CSF compression or aortic occlusion. Following reperfusion and onset of the low-flow state, which stabilized at 45 min postischemia, control normothermic (37 degrees C) measurements of CBF and CMRO2 were determined. Thereafter, femoral arterial blood was circulated through a heat exchanger (42.5 degrees C), and brain temperature was increased to 40 degrees C and measurements were repeated. The brain was then cooled back to 37 degrees C for a final set of normothermic measurements. Thereafter, brain biopsies were taken to determine the energy state of the brain. CMRO2 changed approximately 6%/degrees C. CBF paralleled the change in CMRO2. Accordingly, the ratio of CBF to CMRO2 remained constant throughout at a value of 8 to 9, demonstrating maintained coupling. The brain energy state was normal at the end of the study. The authors conclude that postischemic CBF is modulated by the brain's metabolic needs.

Delayed treatment with nimodipine improves cerebral blood flow after complete cerebral ischemia in the dog.

Ten minutes of complete cerebral ischemia was produced in 18 dogs by temporary ligation of the aorta and venae cavae. Dogs were randomly assigned to one of three groups. A bolus dose of 10 micrograms kg-1 nimodipine, a dihydropyridine calcium entry blocker, followed by a constant infusion of 1 microgram kg-1 min-1 was given at 15, 30, or 60 min post ischemia. Cerebral blood flow and metabolism were measured for 2 h postischemia. Delayed treatment with nimodipine ameliorated or reversed the cerebral hypoperfusion that routinely occurs after complete ischemia. In the groups treated at 15 and 30 min, CBF remained above 60 ml min-1 100 g-1. In the group treated at 60 min, there was a progressive decline in CBF to 37 ml min-1 100 g-1. Following treatment with nimodipine, CBF immediately increased and was maintained above 50 ml min-1 100 g-1 for the remainder of the study. Once treatment with nimodipine was begun, CBF was approximately double that of an untreated group. Changes in CBF reflected changes in cerebrovascular resistance. Nimodipine had no effect on cerebral metabolism. Since the postischemic hypoperfusion state is believed to contribute to the ultimate neurologic damage following complete ischemia, treatment with nimodipine, even if delayed up to 60 min, may improve the outcome.

Cerebral vascular autoregulation and CO2 reactivity following onset of the delayed postischemic hypoperfusion state in dogs.

A small number of animal studies have suggested that during the delayed postischemic hypoperfusion state, CO2 reactivity of the cerebral vasculature is lost whereas autoregulation is retained. These findings, however, are inconsistent with the bulk of experimental evidence which demonstrates that CO2 reactivity is more robust and may be retained in pathologic circumstances which abolish autoregulation. These opposing viewpoints were therefore further evaluated in 18 dogs in which complete global ischemia was induced by cerebrospinal fluid (CSF) compression for periods of 12 (n = 12) and 18 (n = 6) min. Following 45 min of reperfusion and with onset of the delayed postischemic hypoperfusion state, autoregulation and CO2 reactivity were evaluated using a continuous measurement of CBF (by sagittal sinus outflow). CO2 reactivity was tested over a PaCO2 range of 20 to 60 mm Hg; autoregulation was tested over a blood pressure range of 60 to 140 mm Hg. Results demonstrated that after both 12 and 18 min of complete global ischemia, autoregulation and CO2 reactivity of the cerebral vasculature were both present, but attenuated. In the case of CO2 reactivity, the slope of the CBF response was decreased approximately 75%. In the case of autoregulation, the response in some dogs was incomplete as compared with their preischemic response.

Ischemia-reperfusion does not affect reactivity of isolated canine basilar artery.

The effect of ischemia-reperfusion on endothelium-dependent relaxations and reactivity of vascular smooth-muscle cells was studied in rings of basilar arteries obtained from six dogs exposed to 12 min of complete global cerebral ischemia followed by 100 min of reperfusion. Three sham-operated control dogs served as controls. Ischemia was induced either by an increase in intracranial pressure or by aortic occlusion. The rings were suspended for isometric tension recording in physiological salt solution. Ischemia-reperfusion did not affect endothelium-dependent relaxations to vasopressin and bradykinin. In rings without endothelium relaxations to sodium nitroprusside, molsidomine (SIN-1), and papaverine as well as contractions to 5-hydroxytryptamine and KCl were preserved. These results demonstrate that in large canine cerebral arteries, ischemia-reperfusion of these durations does not affect relaxations mediated by activation of endothelium or direct relaxations and contractions of vascular smooth-muscle cells.

Lidoflazine does not improve neurologic outcome when administered after complete cerebral ischemia in primates.

In order to investigate the effects of the calcium entry blocker lidoflazine on neurologic outcome in primates following an episode of global brain ischemia, 12 pigtail monkeys (Macaca nemestrina) were subjected to 17 min of complete cerebral ischemia, followed by 48 h of intensive care treatment and daily neurologic evaluations for 96 h. The monkeys were randomly assigned to receive, in a blind fashion, either lidoflazine 1.0 mg/kg (n = 6) or inactive lidoflazine solvent (n = 6) at 5 min, 8 h, and 16 h postischemia. One monkey in the lidoflazine group did not meet preestablished protocol criteria and was excluded from data analysis. The remaining monkeys were well matched for age, sex, and other physiologic variables. Neurologic outcome was not significantly different between the lidoflazine- and placebo-treated groups (p greater than 0.5). No monkey in either group achieved a normal neurologic exam by 96 h postischemia. Three lidoflazine-treated monkeys and two placebo-treated monkeys died prior to the 96-h neurologic evaluation. These deaths were judged to be neurologic in origin. The authors concluded that lidoflazine does not improve neurologic outcome in primates when administered after 17 min of complete cerebral ischemia.

The effects of nimodipine on cerebral blood flow and metabolism.

Nimodipine, a calcium entry blocker, was administered in increasing doses of 0.1-3.0 micrograms kg-1 min-1 to six dogs after they had recovered consciousness from a surgical preparation that was conducted under general anesthesia and while they were under the influence of total spinal anesthesia. CBF was measured with a sagittal sinus outflow technique and CMRO2 was calculated as the product of CBF and the arteriovenous O2 difference. Nimodipine did not influence either CBF or CMRO2. There was a decrease in the cortical pyruvate level at the end of the study, but no significant change in phosphocreatine, ATP, lactate, or energy charge when compared with six control dogs. It has previously been reported that nimodipine increases the CBF in global ischemia with a potentially beneficial effect on the neurological outcome. With no effect on normal CBF or metabolism, this suggests that nimodipine may be useful in a variety of ischemic situations without fear of either a steal phenomenon or untoward effects on intracranial pressure.

Nimodipine improves cerebral blood flow and neurologic recovery after complete cerebral ischemia in the dog.

Ten minutes of complete ischemia was produced in 11 dogs by temporary ligation of the aorta. Immediately before the ischemic episode, the dogs received nimodipine, a new calcium entry blocker, 10 micrograms kg-1, i.v., followed by an infusion of 1 microgram kg-1 min-1 for 2 h. Post-ischemic cerebral blood flow and metabolism were measured for 120 min in six dogs. Neurologic recovery was evaluated 48 h post-ischemia in five dogs. The results were compared to previously determined controls. Nimodipine nearly doubled cerebral blood flow in the delayed post-ischemic hypoperfusion period, compared to untreated dogs (approximately 45% versus 25% of pre-ischemic control values), but had no significant effect on metabolism. Nimodipine also improved neurologic recovery. Four of five treated dogs were normal and one was moderately damaged, whereas six of seven controls were either severely damaged or dead. This suggests that the delayed hypoperfusion state occurring after complete cerebral ischemia probably does contribute to the ultimate neurologic damage, and that nimodipine offers a potential protective effect.

Cerebral blood flow, brain pH, and oxidative metabolism in the cat during severe insulin-induced hypoglycemia.

The effects of severe hypoglycemia on brain pH, cerebral blood flow (CBF), and other physiologic and metabolic parameters were studied in 26 cats subjected to insulin hypoglycemia. Two groups were utilized to compare the effects of anesthesia. The halothane group was composed of 14 animals and the barbiturate group contained 12 animals. Insulin was administered by both the intravenous and intramuscular routes until there was a severe electroencephalographic (EEG) change or until 6 h had elapsed. The cerebral responses to hypoglycemia demonstrated the following: CBF was unaffected by severe hypoglycemia in normotensive animals with or without EEG changes; brain pH was essentially constant in all groups regardless of glucose levels, CBF, or EEG; and the EEG abnormalities corresponded closely to brain glucose levels. Cerebral adenosine triphosphate and phosphocreatine levels were lowest in the animals with isoelectric EEGs. We conclude that CBF and brain pH in the normotensive cat under general anesthesia are relatively unaffected by insulin hypoglycemia despite the presence of severe EEG changes and diminished cerebral energy reserves. The study suggests tha the PaCO2-CBF response curve is ot dependent upon the metabolic integrity of cerebral tissue and is mediated by pathways separate from those of autoregulation.

Cerebral blood flow and neurologic outcome when nimodipine is given after complete cerebral ischemia in the dog.

Ten minutes of complete cerebral ischemia was produced in 26 dogs by temporary ligation of the aorta and the venae cavae. Twenty dogs received nimodipine, a calcium entry blocker, 10 micrograms kg-1 i.v. 2 min after the ischemic period, followed by 1 microgram kg-1 min-1 for 2-3 h. Six dogs received only the solvent used for nimodipine. Fourteen dogs received nimodipine for 3 h and were subsequently evaluated neurologically up to 48 h postischemia. In the 12 other dogs, CBF and metabolism were followed for 2 h postischemia while either nimodipine or the solvent only was infused. The results were compared to previously published results for untreated dogs and dogs given nimodipine before the ischemic event. Nimodipine had the same effect on postischemic CBF whether started before or after the ischemic event, nearly doubling the flow when compared with untreated controls, whereas the solvent alone caused only a slight increase in CBF over control. By contrast, nimodipine initiated in the preischemic period significantly improved the neurologic outcome, but when initiated in the postischemic period the results were equivocal, such that the outcome was not significantly different from either the untreated group or the group in which nimodipine was initiated preischemia. Metabolic measurements did not give any indication of a specific effect of nimodipine, nor could the metabolic results be used as an indicator of neurologic outcome. The results are consistent with a beneficial effect of nimodipine following complete cerebral ischemia; however, evaluation of neurologic functional effects will require a more sensitive model.

The effects of dizocilpine maleate (MK-801), an antagonist of the N-methyl-D-aspartate receptor, on neurologic recovery and histopathology following complete cerebral ischemia in primates.

The present study was designed to determine if the noncompetitive excitatory amino acid antagonist, dizocilpine maleate, when administered after a 17 min period of complete cerebral ischemia in primates, would improve postischemic neurologic function and hippocampal histopathologic outcome when compared to placebo-treated animals. Ten pigtail monkeys were anesthetized and subjected to complete cerebral ischemia using an established neck tourniquet model. Five minutes postischemia, five monkeys received dizocilpine 300 micrograms/kg i.v. over 5 min, followed by an infusion of 150 micrograms/kg/h for 10 h. This produced plasma levels of the drug in excess of 30 ng/ml for the duration of the infusion. An additional five monkeys were treated with an identical volume of saline placebo. All monkeys received intensive care for the initial 24 to 48 h postischemia. At 96 h postischemia, there was no significant difference in neurologic function between the two groups (p = 0.53, with the placebo group having the numerically better outcome). There also was no significant difference between hippocampal histopathology scores between dizocilpine and placebo-treated monkeys. The authors conclude that dizocilpine is not an efficacious therapy in the treatment of neurologic injury that occurs following complete cerebral ischemia in this primate model.

Cerebral protection by barbiturate anesthesia. Use after middle cerebral artery occlusion in Java monkeys.

Regional cerebral ischemia was produced in 18 Java monkeys by permanent middle cerebral artery (MCA) occlusion. All monkeys were thereafter paralyzed (pancuronium bromide, 0.05 mg/kg/hr) and sedated (diazepam, 0.1 mg/kg/hr) for a 48-hour period. Thirty minutes after MCA occlusion, pentobarbital sodium anesthesia was induced in nine of the monkeys (14 mg/kg) and maintained for 48 hours (7 mg/kg every two hours), with continuous supportive care. After 48 hours, all drugs were discontinued; the monkeys were observed for five days, and then killed. Seven of the control monkeys developed a cerebral infarction, three did not survive past the 48 hours of intensive care, and the other four had a notable neurologic deficit. All pentobarbital monkeys survived the seven days, but four had a cerebral infarction and two of these had a notable neurologic deficit. These differences were statistically significant.

Cerebral blood flow during cardiac operations: comparison of Kety-Schmidt and xenon-133 clearance methods.

This study simultaneously compared the standard Kety-Schmidt and the modified xenon-133 (133Xe) clearance techniques for measuring cerebral blood flow (CBF) and cerebral metabolic rate for oxygen (CMRO2) during cardiac operations. The validity of the CBF method is important because our management of the patient during cardiopulmonary bypass (CPB) is based, in part, on our understanding of the cerebral hemodynamics during CPB. In 20 patients undergoing coronary artery bypass grafting, CBF and CMRO2 were determined by both methods. Measurements were made before onset of CPB and once during CPB. Ten patients underwent CPB with systemic normothermia (37 degrees C) and 10 with systemic hypothermia (27 degrees C). Anesthesia consisted of fentanyl and midazolam. CPB pump flows were kept at 2.2 to 2.4 L.min-1.m-2 and alpha-stat pH management was used. Xenon-133 clearance significantly underestimated CBF and CMRO2 relative to the Kety-Schmidt technique before CPB and at both bypass temperatures. Values obtained by 133Xe clearance were approximately 50% of that measured by the Kety-Schmidt method. The modified 133Xe technique as typically used during cardiac operations does not appear to measure CBF accurately; this leads to corresponding errors in CMRO2 calculations. Determination of CMRO2 and cerebral autoregulatory function during cardiac operations appears to be more appropriate if based on the more direct Kety-Schmidt technique. Accordingly, our management of CPB with respect to cerebral perfusion as it has been determined by the modified 133Xe clearance method may require reassessment.

The effect of the excitatory amino acid receptor antagonist dizocilipine maleate (MK-801) on hemispheric cerebral blood flow and metabolism in dogs: modification by prior complete cerebral ischemia.

The effect of the N-methyl-D-aspartate (NMDA) receptor antagonist dizociplipine maleate (MK-801) on cerebral blood flow (CBF), cerebral metabolic rate for oxygen (CMRO2), intracranial pressure and systemic variables was examined in 6 normal dogs (Group I). In 6 additional dogs (Group II), the effects of a prior 11 min episode of complete cerebral ischemia on the response to dizocilipine was studied. CBF was measured with a sagittal sinus outflow technique and CMRO2 was calculated as the product of CBF and the arterial to sagittal sinus O2 content difference. Dizocilipine was administered as a 150 micrograms/kg i.v. bolus followed by a 75 micrograms.kg-1.h-1 infusion for 90 min. Plasma dizocilipine levels were greater than 25 ng/ml for the duration of the infusion. The CSF levels were approximately half the plasma levels. Five minutes after initiation of dizocilipine treatment, Group I dogs experienced a 63% increase in heart rate (P less than 0.01) and an 8% decrease in the mean arterial blood pressure (P less than 0.05). Over the same time interval. CBF increased by 85% (P less than 0.01) and intracranial pressure nearly doubled (P less than 0.05). In addition, dizocilipine treatment in all Group I animals resulted in EEG quasiperiodic bursts of delta-waves and polyspikes on a background of beta-activity. With the exception of the intracranial pressure, the above changes in systemic and cerebral variables persisted for the duration of the drug infusion. Intracranial pressure was no longer significantly elevated after 80 min of drug infusion. Hemispheric CMRO2 was unchanged by dizocilipine in Group I dogs. There was a decrease in the cortical glucose level at the end of the study, but no significant change in phosphocreatine, ATP, lactate, or energy charge when compared with 6 laboratory normals. An identical dose of dizocilipine administered after an 11 min episode of complete cerebral ischemia resulted in no significant changes in either cerebral or systemic variables. The absence of systemic effects in Group II dogs suggests that dizocilipine administration in normal dogs results in a centrally mediated activation of the peripheral sympathetic nervous system. The uncoupling of CBF and CMRO2 observed following dizocilipine treatment is similar to that reported for two other known NMDA antagonists, ketamine and phencyclidine. If administration of dizocilipine results in improved histopathological and neurological outcome following an episode of complete cerebral ischemia, this improvement is unrelated to changes in postischemic CBF or hemispheric CMRO2.

Evaluation of the glutamate antagonist dizocilipine maleate (MK-801) on neurologic outcome in a canine model of complete cerebral ischemia: correlation with hippocampal histopathology.

This study was designed to determine if dizocilipine maleate (MK-801), administered following 11 min of complete ischemia in dogs, could favorably alter neurologic outcome and hippocampal damage. Eighteen dogs were anesthetized and subjected to complete cerebral ischemia by temporary occlusion of the ascending aorta and the venae cavae via a thoracotomy. Five min postischemia, 9 dogs were given dizocilipine 150 micrograms/kg, followed by an infusion of 1.25 microgram/kg/min for 8 h. Control dogs were given equal volumes of placebo. Dogs were evaluated neurologically at 24, 48, and 72 h; thereafter, the brains were perfused, fixed and harvested. There was no significant difference in outcome between dizocilipine- and placebo-treated dogs: 5 of 9 given dizocilipine were normal, 1 was mildly injured and 3 were severely injured or dead. In the control animals given placebo, 3 of 9 were normal, 2 were mildly injured and 4 were moderately to severely injured. Histopathologic examination was limited to the hippocampus. CA1 and CA2,3,4 pyramidal neurons were graded according to degree of injury on a 5-point scale. There were no differences in histopathologic grades between the two groups. However, in both groups combined there was a significant correlation between neurologic outcome grade and histopathologic grade. The only notable systemic effect of dizocilipine appeared to be prolonged sedation which extended beyond 24 h postischemia but was not evident at 48 h postischemia. The authors conclude that more outcome studies in more sensitive models are needed.

Closed recirculatory spinal subarachnoid perfusion for determining CSF dynamics.

A new method for determining the rates of cerebrospinal fluid (CSF) production under nonsteady-state conditions, namely, closed recirculatory spinal subarachnoid perfusion, was used to determine the effect of enflurane on the rate of CSF production in dogs. Considerable variability in results was observed such that there was no statistical difference in rates of production among animals that received enflurane 2.2%, enflurane 2.2% and nitrous oxide 60% to 70%, enflurane 3.2% and nitrous oxide 60% to 70%, or nitrous oxide 60% to 70% (controls). Possible sources of variability were sought in additional studies using a modification of the new method, and in an in vitro model. The results were compared to those obtained using an established method for determining rates of CSF production, namely, open ventriculocisternal perfusion. It was concluded that the sources of variability in the closed recirculatory method relate in part to adherence of the fluorescein-conjugated albumin tracer to glass and other surfaces, and to uneven flow and distribution of the tracer in the recirculatory system. When the open ventriculocisternal perfusion method was used, consistent results were obtained, demonstrating that CSF production rate increased significantly in animals that received enflurane. The authors conclude that the new closed recirculatory method is less reliable than the classical open perfusion method.

Influence of intraoperative antibiotic choice on the incidence of early postcraniotomy seizures.

Because of a suspicion that intraoperative penicillin antibiotics might be causing early postoperative seizures in craniotomy patients, a deliberate effort was initiated in 1987 to avoid these agents in favor of nonpenicillin antibiotics. This permitted a retrospective comparison of the incidence of early postoperative seizures in craniotomy patients who did and who did not receive intraoperative penicillins. Records of patients treated between July 1, 1984, and July 1, 1985, and between July 1, 1987, and July 1, 1988, were reviewed, for a total of 1316 procedures. There were no seizures in the 323 patients who underwent suboccipital craniectomy. However, among the 993 patients receiving supratentorial procedures there were 30 with seizures within the first 6 hours postoperatively, 19 of which were generalized seizures. The incidence of early seizures was 4.7% (20 cases) of the 427 patients given penicillins and only 1.8% (10 cases) of the 566 not given penicillins (p less than 0.01). Since patients undergoing surgery for intractable seizures have a high incidence of early postoperative seizures (11 of 92 in this series, or 12%), creation of a subgroup eliminated these from consideration but did not alter the relationship of penicillins to early seizures (p less than 0.02). The authors conclude that intraoperative (and early postoperative) penicillin antibiotic administration should be avoided, if possible, in patients undergoing craniotomy for supratentorial pathology.