Cortical blood flow regulation during hypoxemia in experimental head injury

JOURNAL

OF SURGICAL

RESEARCH

43, 86-93 (1987)

Cortical Blood Flow Regulation during Hypoxemia in Experimental Head injury PAUL J. FEUSTEL,

PH.D.,

AND LOUIS R. NELSON,

M.D.

S. R. Powers, Jr. Trauma Research Center, Departments of Surgery and Physiology, Albany Medical College of Union University, Albany, New York 12208, and Department of Biomedical Engineering, Rensselaer Polytechnic Institute, Troy, New York 12181 Presented at the Annual Meeting of the Association for Academic Surgery, Washington, D.C., November 5-8, 1986 The mechanism of increased susceptibility of the traumatized brain to hypoxemia (HYP) was investigated by measuring the local cortical blood flow (LCBF) by hydrogen clearance. Following pentobarbital anesthesia and head injury (HI) using a repetitive acceleration/deceleration injury, five surviving cats were ventilated to maintain arterial pCOz (28.3 -t 1.8 Torr, mean + SEM). Three control (C) animals received no head injury. LCBF (ml/min/lOO g) measurements at two cortical locations were made between 30 and 60 min postinjury, at 10 min and 40 min of HYP (P,Oz = 29.5 f 2.6 Torr), and at 30-min intervals for 4.5 hr after HYP. Before HYP, LCBF was not different in cortical areas in C (84 f 12 ml/min/lOO g) and HI (8 1 + 17 ml/min/ 100 g) animals. In C animals LCBF increased to 14 1 + 10 after 10 min of HYP and remained elevated at 40 min (P < 0.05). Ten minutes post-HYP, LCBF fell to 56 + 8 ml/min/ 100 g. Hypoxemia did not increase LCBF significantly in HI animals (87 + 17 ml/mitt/ 100 g). Post-HYP, LCBF in HI animals remained unchanged. In 7890 of HI LCBF measurements, clearances following a brief Hz inhalation were faster than clearances following tissue equilibration with Hz. This may be due to nonhomogeneous tissue perfusion in HI animals. In conclusion, after head injury there may be an attenuated microvascular blood flow response to hypoxemia and flow inhomogeneities which will accentuate tissue hypoxia. o 1987 Academic PBS, IIIC.

flow from metabolism have also been observed after injury [ 51. These observations led us to investigate the microcirculatory response to hypoxemia immediately following brain injury. The absence or attenuation of an appropriate blood flow response to hypoxemia may exacerbate tissue hypoxia with associated increased morbidity and mortality. We used an animal model of brain injury which demonstrates synergistic effects of injury and hypoxemia on morbidity and mortality [6, 71. These studies were intended to determine whether an attenuated blood flow response occurs in this model after injury.

INTRODUCTION

Human head injury is a highly variable, multifaceted problem resulting from different injury mechanisms, variable pathophysiologic processes, variable secondary factors, and various combinations of associated injuries. The most uncontrolled period in human head injury is the unmonitored period between head injury and in-hospital resuscitation during which varied degrees of hypoxemia and/or hypotension may be experienced, but are rarely documented. There is evidence that the sequence of events following blunt head injury in humans frequently results in hypoxemia [ l-31. Histopathologic studies of brains of patients who suffered blunt head injury and were subsequently autopsied over a 2-year period revealed lightmicroscopic evidence of ischemic/hypoxic injury throughout the brain [4]. Circulatory regulation abnormalities such as wide variations in blood flow and uncoupling of blood 0022-4804/87 $1.50 Copyright 0 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

METHODS

Healthy conditioned adult cats of either sex, weighing 2 to 4 kg, were used in these experiments. The animals received no solid food the night before the experiment but had water ad lib. The experimental protocol was 86

FEUSTEL

AND NELSON: CORTICAL

reviewed and approved by the Animal Research Committee at Albany Medical College and met all state and federal regulations regarding animal use. Initial anesthesia was intravenous methohexital (13.25 mg/kg) with subsequent anesthesia (after 25 min) maintained with pentobarbital (30 mg/kg, iv). Atropine (0.8 mg/kg im) was administered and the animal was intubated with a 3.0-mm cuffed endotracheal tube. The femoral artery and vein were surgically cannulated and arterial blood pressure was monitored using a strain gauge (Statham P50) attached to the arterial cannula. Rectal body temperature was monitored and maintained between 35 and 37°C using infrared heating lamps. Arterial blood gases were measured before and after trauma and just prior to each LCBF measurement. A repetitive acceleration-deceleration injury was administered to the skull encased brain of the anesthetized cat by a technique described previously [6,7]. Briefly, the injury consists of a translational plus rotational mode of injury in which the cat’s head is moved through a repetitive flexion-extension motion with the pivot at the base of the neck. The center of the head moves a total vertical excursion of 5.2 cm. From a neutral position, the flexion excursion is 3.8 cm. The total rotational component is 48” with the distance from the center of the head to the pivot being 6.7 cm. With this system the animal receives a total of 1400 positive and negative acceleration-deceleration impulses over a 67-set period at a speed of 1250 oscillations per minute. With this system at this speed, the average peak tangential acceleration force at the center of the head is 8Og and the average peak centrifugal force is 36g. Within a few minutes following experimental brain injury, mechanical ventilation with room air was started using a small animal respirator. The animal was paralyzed with gallemine triethiodide (6.5 mg/kg) and the respirator rate and volume were adjusted so that arterial PO2 was greater than 70 mm Hg and arterial PC02 was between 25 and 35 mm Hg as determined by repeated blood gas

PERFUSION

AFTER

HEAD INJURY

87

analyses. Additional doses of gallemine triethiodide, when necessary, were accompanied by supplemental pentobarbital(7.5 mg) so that surgical anesthesia was maintained throughout the experiments. The head was then fixed in a stereotaxic frame and a 5-mm-diameter opening was made on the right side of the skull approximately 10 mm lateral to the midline. This exposed the dura over the suprasylvian gyrus. Local cerebral blood flow was measured by the H2 clearance technique [8, 91. Electrodes were cut from 25-pm-diameter Tefloncoated platinum (Pt) wire (Medwire Inc.). The reference electrode was silver with a silver chloride coating. Under microscopic guidance, the Pt electrodes were inserted through a small slit in the dura. Two electrodes were inserted into the cortex to a distance of 1 to 2 mm. Pial vessels were avoided. If electrode insertion resulted in bleeding then the electrode was not used. The Pt electrodes were polarized to -0.3 V (E(Pt - Ag) = 0.3 V) with respect to the reference electrode for the hydrogen (HZ) oxidation reaction. Clearances were not performed until after a IO-min stabilization period during which a steady baseline was obtained. Clearances were measured following addition of H2 to the inspired air. Local cerebral blood flows were determined from the H2 clearance curves using a one-compartment model as previously described [ IO]. Clearance rates were measured using points in the interval of time between 80 and 30% of the peak H2 current. The first 20% of each clearance was eliminated to ensure that the arterial H2 concentration had fallen to negligible levels before beginning analysis. If a curve were multiexponential, as described by some investigators using larger electrodes [I 1, 121, then this technique yielded weighted average of flows closer to the higher compartment flow because of the elimination of the later points (unless the faster compartment has a very small contribution to the total flow). In theory, however, it has been shown that the smaller the electrode and the smaller the sample tissue compart-

88

JOURNAL

OF SURGICAL

RESEARCH:

ment the less likely it is that multicompartment clearances will be observed [ 131. At 60 min posttrauma, hypoxemia was initiated by mechanically ventilating the paralyzed animal at the same rate and volume with a mixture of 7.5% O2 in NZ. This continued for 60 min, at which time mechanical ventilation with 2 1% O2 was resumed. Clearances were performed by two techniques (see below) prior to hypoxemia, after 10 min of hypoxemia, at 40 min of hypoxemia, 10 min after hypoxemia, and subsequently at 30min intervals for up to 5 hr. Control animals were treated identically except that there was no head injury. Ordinarily the clearance of inert gas from normal brain tissue gives reliable data regarding the blood flow because the clearance of inert gases is not diffusion dependent [ 141 and because intercompartmental exchange of inert gas is insignificant for the usually large volume over which the gas is measured. This may not be the case with smaller electrodes in inhomogeneously perfused brain [ 131 or in volumes adjacent to unperfused tissue [ 15, 161. A measurable flow may be attributed to a section of cortex, which from the standpoint of oxygen delivery has no effective flow, because H2 gas can diffuse into and out of that tissue from a remote location. In order to test for the presence of intercompartmental diffusion (and presumably focal areas of low perfusion), flows were determined by using two methods of H2 gas loading. The first was the several breaths technique in which H2 is added to inspired air for two to four breaths. In the second method clearances began from a steady state of tissue H2 concentration. A steady state was reached by adding small flows of H2 to the inspired air until steady currents were achieved in both electrodes. At this point all tissue, including poorly perfused ischemic cortex, was equilibrated with HZ. HZ was then removed from the inspired air. A significantly different time constant between the several breath clearance and the clearance from the steady state was taken as evidence of intercompartmental diffusion-dependent exchange of HZ,

VOL. 43, NO. I, JULY 1987

Since clearances following a several breath HZ inhalation can be regarded as an upper limit on actual flow and clearances following a steady state can be regarded as a lower limit, mean flows reported below include diffusion-dependent electrodes where the flow for that electrode is the average of the two determinations. Statistical analysis was by analysis of variance with significance assessed at the 0.05 level. Subsequent application of the Student-Newman-Keuls multiple range test was used to determine whether blood flows differed with respect to time and hypoxemia [ 171. RESULTS

Of seven animals receiving head injury one died immediately. A second died at 10 min of hypoxia before clearances could be measured (prehypoxemia blood flows were 9 1 and 40 ml/min/ 100 g). Blood pressure and blood gas values for the remaining five animals and three controls are reported in Table 1. In six cortical areas of three control animals blood flow prior to hypoxemia was 83.8 f 11.7 ml/min/lOO g (mean -t SEM). With hypoxemia, flow increased to 141.3 f 9.9 ml/min/ 100 g at 10 min and remained elevated at 139.6 f 14.4 ml/min/lOO g after 40 min (P < 0.05; Fig. 1). After hypoxemia, flow fell to 56.3 + 7.7 ml/min/lOO g and remained at levels not significantly different from control. Four and a half hours after the onset of hypoxia flow was 66.4 + 14.2 ml/ min/ 100 g. Prehypoxemic blood flow, calculated using the average of the two clearance techniques, was 8 1.O f 17.1 ml/min/ 100 g in 10 cortical areas of head injured animals and was not different from control animals. Blood flow did not change during hypoxemia (Fig. 1). Over the 3.5 hr following hypoxemia, flows in four cortical areas fell to levels which could not be measured in 15 min (i.e., < 10 ml/min/ 100 g) and were assigned to zero flow (Fig. 2). Two electrodes failed for technical reasons. Three and a half hours

FEIJSTEL AND NELSON: CORTICAL

PERFUSION

89

AFTER HEAD INJURY

TABLE 1 MEAN ARTERIAL BLOOD

PRESSURE (MABP) AND BLEND GAS DATA (MEAN AFTER HYP~XEMIA IN CONTROL AND HEAD INJURED

+ SEM) BEFORE, DURING, ANIMAL.S

Hypoxia Prehypoxia Control MABP (Torr) fQ2 (Ton) pC02 (Torr) PH

116 73.9 31.9 7.318

Head injury MABP (Torr) PO2 (TOM pC02 (Torr) PH

123 + 84.1 f 28.2 f 7.409 +

10 min

f 13 +- 4.9 f 0.5 + .030

Posthypoxia 40 min

10 min

11 2.6* 0.7 ,027

118+ 14 27.1 + 5.0* 28.4 + 1.9 7.300 _+ .036

114+ 15 93.3 Ik 7.1 25.8 + 1.8 7.358 -t .050

124 84.0 28.9 7.436

133*9 32.6 f 3.7* 23.3 f 3.2 7.411 + .035

116k8 32.1 f 3.8* 27.1 2 0.7 7.363 f .047

110 f 86.3 + 25.3 f 7.368 +

113+6 87.7 f 6.4 25.2 f 3.0 7.390 _+ .044

128 25.5 24.7 7.367

8 4.7 1.6 .030

AND

+ f f +

7 6.4 2.1 .053

4.3 hr + f f +

12 7.2 0.5 .039

* Different from prehypoxic levels; P < 0.05.

after hypoxemia blood flow averaged 15.1 + 6.4 ml/min/lOO g in the remaining electrodes. Figure 2 shows the flow responses in CORTICAL

BLOOD

FLOW

(ml/m~n/lOOgl 200

-----

POST

HYPOXIA 0 -1

’

I

!

0

1 TIME

(hours

HEAD

I

/

INJURY

‘j---j

’

I 2 post

3

4

hypoxia)

FIG. 1. The local cortical blood flow versus time in control and head injured animals. Values are means + SEM. Animals were hypoxemic between time 0 and 60 min. Head injury was induced 30 to 60 min prior to hypoxemia except in control animals. In control animals, average blood flow during hypoxia (at 10 and 40 min) was significantly greater than the blood flow during all normoxic measurements. In head injured animals there was no significant change in blood tlow.

individual cortical areas of control and head injured animals. Areas of cortex in head injured animals showed a less uniform response when compared with cortical areas in control animals. Clearances by the two techniques were not different in control cortex (Fig. 3). However, in cortex following injury, some cortical areas consistently showed a faster clearance following the several breaths or “pulse” administration of Hz than that following the steady-state administration (Fig. 4). In normal cortex, 5 1% of all measurement pairs had faster clearances following the brief inhalation. In injured cortex, 78% of measurements demonstrated faster clearances after the several breath technique, with some electrodes showing “flows” twice those measured following a steady-state equilibration. When considering only the flows determined from the steady-state clearances, the prehypoxemic blood flow in the injured animals was 69.3 +- 19.1 ml/min/lOO g, which was lower, but still not significantly different, than that of the control animals (83.4 + 10.6 ml/min/ 100 g). Flows determined from clearances from the steady state alone were 70.9 + 11.2 ml/min/ 100 g in injured animals during hypoxemia (control = 140.0 + 7.6 ml/min/ 100 g). At four and a half hours after

90

JOURNAL CORTICAL

BLOOD

OF SURGICAL

RESEARCH:

FLOW

~ml/minI100gl

04

:

:

0

:

1 TIME

:

2 (hours

post

(

3

(

4

hyooxlal

Ftc. 2. Local cortical blood flow in individual electrodes in head injured and control animals.

hypoxia, flows averaged 8.1 + 5.3 ml/min/ 100 g (control = 69.0 f 19.9 ml/min/lOO g). DISCUSSION

VOL. 43, NO. 1, JULY

1987

In control animals the blood flow response to hypoxemia was comparable with measurements made in cortical gray matter with microspheres [ 191. Although electrode insertion must cause some localized tissue damage, the reversible vasodilation which occurs in control animals must be taken as evidence that the local circulation in the volume of tissue over which H2 is measured is largely intact. The slightly higher flows observed before hypoxia compared with after hypoxia may reflect some persistent hyperemia due to electrode insertion. The technique-dependent difference in clearances seen in injured animals suggests that circulatory disturbances, such as flow inhomogeneities, may exist following injury. Clerances following pulse inhalations of H2 were faster (higher flow) than clearances following the equilibration method in head injured animals. This may indicate that there is an intercompartmental diffusive exchange of H2 in injured cortex. Specifically, during the clearance following the pulse administration, PULSE

‘CBF’

CONTROL

(mllmtn/iOOgI

The results support the hypothesis that following head injury there is an attenuated microvascular blood flow response to hypoxemia. This is consistent with the results of Lewelt et al. [ 181 who used similar techniques and found a decreased flow response following injury in a fluid percussion model in cats. In the present model of head injury previous results indicate that there is a small (5%) immediate mortality due to the head injury alone [6]. Of those that survive the injury and are subsequently subjected to hypoxemia for 1 hr there is a mortality of 50%. There is no significant delayed (>lO min postinjury) mortality associated with either the injury without hypoxemia or the hypoxemia without injury. This absence of a significant blood flow response to hypoxemia may exacerbate tissue hypoxia to such an extent that irreversible tissue damage occurs, resulting in the higher mortality rates.

240

-

180..

120..

STEADY

STATE [ml/min/

0

60

120

160

‘CBF’ 1 OOg) 240

FIG. 3. Flows measured by clearances following muhiple breath inhalations of Hz (pulse) versus flows measured by clearances after steady-state H2 inhalations in control animals. Determinations are distributed around the line of identity, indicating the lack of significant intercompartmental diffusion. (+) Normoxia, (X) hypoxia.

FEUSTEL PULSE

‘CBF’

~ml/min/

1 OOgl

AND NELSON: CORTICAL TRAUMA

FIG. 4. Flows measured from clearances following multiple breath inhalations of Hz (pulse) versus flows measured by clearances after steady-state Hz inhalations in head injured animals. Different symbols represent different electrodes. Several electrodes show faster clearances following multiple breath determinations.

H2 gas may be removed from the neighborhood of the electrode not only by blood flow but also by diffusion into neighboring poorly perfused areas. Thus the clearance is accelerated and the measured flow is higher than the actual flow. Following an inhalation to a steady state, the area surrounding the electrode may be supplied with HZ from adjacent areas of lower perfusion. This would result in a slower clearance not due to lower actual blood flow. In 75% of LCBF measurements in head injured animals (compared with 5 1% in controls), clearances following brief H2 inhalation were faster than clearances following tissue equilibration. Intercompartmental diffusion may be attributed to inhomogeneous tissue perfusion. Using the H2 clearance technique, Halsey et al. [ 151 have shown significant intercompartmental diffusion in ischemic areas following middle cerebral artery occlusion in cats. Also, significant intercompartmental HZ diffusion was found to occur in the superficial area of the medulla, where there is a large adjacent unperfused

PERFUSION

AFTER

HEAD INJURY

91

(CSF) tissue compartment [16]. Using 14C antipyrine autoradiography, heterogeneities of blood flow have also been shown to occur in models of incomplete cerebral ischemia [20]. Flow inhomogeneities may be a result of regional tissue edema, local metabolic and humoral factors, or small foci of ischemia. Tissue edema itself would tend to decrease measured perfusion due to a decrease in capillary density if the tissue volume over which the measurement is made is assumed to remain constant. These inhomogeneities would further accentuate tissue hypoxia. There is some evidence that similar blood flow unresponsiveness and maldistribution occur in head injured patients. Obrist et al. [5] found that metabolism and flow are uncoupled in those head injured patients who have abnormally high blood flows relative to the cerebral metabolic rate. In this more acute animal model the average blood flows were not found to be high but there was an apparent lack of coupling of blood flow to cerebral oxygen demand. These findings indicate that periods of even relatively mild hypoxemia may be especially dangerous following head injury. In the present study P,C02 fell slightly with hypoxemia, most likely due to increased buffering by unsaturated blood. This P,C02 decrease may have attenuated the hypoxic blood flow response slightly. The blood flow response to hypocapnia appears to be more resilient and is often retained following head injury [5, 211. Severe hypocapnia can also lead to flow inhomogeneity. Grote et al. [22] found that in normal animals severe hypocapnia (12 Torr) leads to inhomogeneous blood flow together with shifts of tissue oxygen tensions into the hypoxic range. Although moderate CO;! reductions (to 20 Torr) did not lead to tissue hypoxia in Grote et al. ‘.s studies the threshold for tissue hypoxia by hypocapnia following brain injury may be lower if inhomogeneities already exist. Autoregulation may be impaired following injury [ 181. Although we did not specifically test for the presence or absence of au-

JOURNAL

92

OF SURGICAL

RESEARCH:

toregulation, the lack of significant blood pressure changes during the measurement periods indicates that the lack of a flow response to hypoxemia could not be due to a decrease in blood pressure. This does not preclude the possibility of brain circulatory damage arising from transient hypertension which occurs during and immediately after injury [23]. Pentobarbital anesthesia is a potential complicating factor in these experiments. Pentobarbital is known to reduce cortical blood flow in proportion to the reduction in metabolism [24]. In focal ischemia models of brain injury, pentobarbital, in large doses with additional drugs for blood pressure support, has been shown to decrease local cerebral blood flow in well-perfused cortex so that local oxygen tensions remain unchanged, while improving the flow to metabolism ratio in poorly perfused cortex [lo]. The effect of pentobarbital in the present experiments would therefore most likely be to protect flow distribution regulation. The lack of a blood flow response to hypoxemia and the flow heterogeneity in head injured animals occurred despite this possible protective effect. We conclude that following head injury there are both an attenuated microvascular blood flow response to hypoxemia and heterogeneities in microvascular blood flow distribution. Because of these two effects, hypoxemia may lead to more severe tissue hypoxia following head injury than would otherwise by expected. ACKNOWLEDGMENTS The authors gratefully acknowledge the expert technical assistance of Donald Szarowski. This work was sup ported by NIH Research Center Grant GM 15426 from the National Institute of General Medical Sciences.

REFERENCES 1. Cook, A. W., Browder, E. J., and Lyons, H. A. Alterations in acid-base equilibrium in craniocerebral trauma: A determinant of survival. J. Neurosurg. 18:366-370,1961.

VOL. 43, NO. 1, JULY

1987

2. Froman, C. Alterations of respiratory function in patients with severe head injuries. Brit. J. Anaesth. 40: 354, 1968. 3. Sinha, R. P., Ducker, T. B., and Perot, P. L. Arterial oxygenation: Findings and its significance in central nervous system trauma patients. JAMA 224: 1258-1260, 1973. 4. Graham, D. I., Adams, J. H., and Doyle, D. Ischemit brain damage in fatal non-missile head injuries. J. Neural. Sci. 39: 213-234, 1978. 5. Obrist, W. D., Langfitt, T. W., Jag& J. L., Cruz, J., and Gennarelli, T. A. Cerebral blood flow and metabolism in comatose patients with acute head injury: Relationship to intracranial hypertension. J. Neurosurg.

61:241-253,

1984.

6. Nelson, L. R., Auen, E. L., Bourke, R. S., Barron, K. D., Malik, A. B., Cragoe, E. J., Popp, A. J., Waldman, J. B., Kimelberg, H. K., Foster, V. V., Creel, W., and Schuster, L. G. A comparison of animal head injury models developed for treatment modality evaluation. In R. G. Grossman and P. L. Gildenberg (Eds.), Seminars In Neurological Surgery, Head

Injury:

Basic and Clinical

Aspects.

New

York Raven Press, 1982. Pp. 117-127. 7. Nelson, L. R., Bourke, R. S., Popp, A. J., Cragoe, E. J., Signorelli, A., Foster, V., and Creel, W. Evaluation of treatment modalities in severe head injuries using an animal model. In A. J. Popp, R. S. Bourke, L. R., Nelson, and H. K. Kimelberg (Eds.), Neural Trauma. New York: Raven Press, 1979. Pp. 297-311. 8. Aukland, K., Bower, B. F., and Berliner, B. W. Measurement of local blood flow with hydrogen gas. Circ. Rex

14: 164-187,

1964.

9. Young, W. H2 clearance measurement of blood flow: A review of techniques and polarographic principles. Stroke 11: 552-564, 1980. 10. Feustel, P. J., Ingvar, M. C., and Severinghaus, J. W. Cerebral oxygen availability and blood flow during middle cerebral artery occlusion: Effects of pentobarbital. Stroke 12: 858-863, 198 1. 11. Jamieson, A. D., and Halsey, J. H. Regional cerebral blood flow determination by the hydrogen clearance technique and comparison with oxygen availability in the rabbit. Stroke 4: 904-9 11, 1973. 12. Symon, L., Branston, N. M., and Chikovani, 0. Ischemic brain edema following middle cerebral artery occlusion in baboons: Relationship between regional water content and blood flow at I and 2 hours. Stroke 10: 184-191, 1979. 13. Pearce, P. A., and Adams, J. M. Measurement of RCBF by Hz clearance: Theoretical analysis of diffusion effects. Stroke 13: 347-355, 1982. 14. Ohta, Y., and Farhi, L. E. Cerebral gas exchange: Perfusion and diffusion limitations. J. Appl. Physiol. 46(6): 1164-l 168, 1979. 15. Halsey, J. H., Capra, N. F., and McFarland, R. S. Use of hydrogen for measurement of regional cere-

FEUSTEL

AND NELSON: CORTICAL

PERFUSION

bra1 blood flow. Problem of intercompartmental diffusion. Stroke 8: 351-357, 1977. 16. Feustel, P. J., Stafford, M. J., Allen, J. S., and Severinghaus, J. W. Ventrolateral medullary surface blood flow determined by hydrogen clearance. J. A&. Physiol. 56( 1): I50- 154, 1984. 17 Dixon, W. J., and Brown, M. B. Biomedical Computer Programs, P Series Berkeley: Univ. of California Press, 1979. 18. Lewelt, W., Jenkins, L. W., and Miller, J. D. Effects on experimental fluid-percussion injury of the brain on cerebrovascular reactivity to hypoxia and to hypercapnia. J. Neurosurg. 56: 332-338, 1982. 19. Marcus, M. L., Heistad, D. D., Ehrhardt, J. C., and Abboud, F. M. Total and regional blood flow measurements with 7-, IO-, 15, 25-, and 50-pm microspheres. J. Appl. Physiol. 40: 501-507, 1976. 20. Ginsberg, M. D., Budd, W. W., and Welsh, F. A. Diffuse cerebral ischemia in the cat: I. Local blood

21.

22.

23.

24.

AFTER HEAD INJURY

93

flow during severe &hernia and recirculation. Ann. Neural. 3: 482-492, 1978. Cold, G. E. Cerebral blood flow in the acute phase after head injury. Part 2: Correlation to intraventricular pressure (IVP), cerebral perfusion pressure (CPP), P,C02, ventricular fluid lactate, lactate/pyruvate ratio and pH. Acta Anaesthesiol. &and. 25(4): 332-335, 198 1. Grote, J., Zimmer, K., and Schubert, R. Effects of severe arterial hypocapnia on regional blood flow regulation, tissue PO1 and metabolism in the brain cortex ofcats. Pfrugers Arch. 391(3): 195-199, 1981. Wei, E. P., Dietrich, W. D., Povlishock, J. T., Navari, R. M., and Kontos, H. A. Functional, morphological, and metabolic abnormalities of the cerebral circulation after concussive brain injury in cats. Circ. Res. 46: 37-47, 1980. Nilsson, L., and Siesjo, B. K. The effect of phenobarbitone anesthesia on blood flow and oxygen consumption in the rat brain. Acta Anaesthesiol. Stand. (Suppl.) 57: 18-24, 1975.