Cerebral Blood Flow, Brain ph, and Oxidative Metabolism in the Cat During Severe Insulin-Induced Hypoglycemia

Transcription

1 Journal of Cerebral Blood Flow and Metabolism 2: Raven Press, New York Cerebral Blood Flow, Brain ph, and Oxidative Metabolism in the Cat During Severe Insulin-Induced Hypoglycemia *John M. Cilluffo, :j:robert E. Anderson, tjohn D. Michenfelder, and *Thoralf M. Sundt, Jr. *Departments of Neurologic Surgery and tallesthesiology, Mayo Clinic; and tcerebral Vascular Research Center, St. Mary's Hospital, Rochester, Minnesota Summary: 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 that the P ac02 - CBF response curve is not dependent upon the metabolic integrity of cerebral tissue and is mediated by pathways separate from those of autoregulation. Key Words: Barbiturates-Brain ph-cerebral metabolism-electroencephalogram Halothane-Insulin hypoglycemia. The effect of severe hypoglycemia on brain ph, cerebral blood flow (CBF), and metabolism have been studied by Norberg and Siesjo (1976) in rats. Others have studied CBF in humans subjected to insulin shock therapy (Kety et ai., 1948; Eisenberg and Seltzer, 1962; Della Porta et ai., 1964; Gottstein and Held, 1967). The results of these investigations have been controversial, since some have found a marked increase in CBF in profound hypoglycemia (Norberg and Siesjo, 1976), while others found Address correspondence and reprint requests to Dr. Sundt at Cerebral Vascular Research Center, Alfred Building, Room 4-437, St. Mary's Hospital, Rochester, Minnesota Abbreviations used: CVR, Cerebrovascular resistance; PCr, phosphocreatine. either no change (Kety et ai., 1948; Eisenberg and Seltzer, 1962; Gottstein and Held, 1967) or a slight increase (Della Porta et ai., 1964). Our data from a previous series studying hypoglycemia in cats over a 3-h period of insulin administration showed that there was no change in CBF or brain ph at serum glucose levels of 1.0 j.tmol/ml (18 mg%) (Cilluffo et ai., 1981). Although we achieved significant hypoglycemia during this study, the electroencephalogram (EEG) was altered in only 6 of 14 of our animals resuscitated with glucose. The current study was designed to correlate the effects of severe hypoglycemia, determined by serum glucose levels and EEG activity, with cerebral metabolic function measured by CBF, brain 337

2 M. CILLUFFO ET AL. ph, cerebrovascular resistance (CVR), and brain tissue levels of phosphocreatine (PCr), adenosine triphosphate (ATP), adenosine diphosphate (ADP), adenosine monophosphate (AMP), glucose, lactate, and pyruvate. MATERIALS AND METHODS General Experimental Design Twenty-six cats were divided into two groups. Twelve cats were anesthetized with pentobarbital and 14 cats with halothane. In each group, five animals were used as controls and were not given insulin. The experimental subgroups consisted of seven cats anesthetized with pentobarbital and nine cats anesthetized with halothane. In the experimental groups, measurements were made at varying levels of arterial carbon dioxide tensions (P ac02) prior to and after the creation of a hypoglycemic state induced by multiple, regularly spaced injections of both intravenous and intramuscular insulin to the point of severe EEG depression, or to 6 h if this was not achieved. Baseline measurements were made at a P ac02 of 40 mm Hg. Insulin was given and the Paco2-CBF response was verified in each animal. In all animals the EEG and mean arterial blood pressure (MABP) were monitored continuously. Animal Preparation Adult cats weighing 1.8 to 3.5 kg were fasted for 24 h, although water was available. The animals were anesthetized with either intraperitoneal pentobarbital (Nembutal, Abbott Laboratories, North Chicago, IL) in a dosage of 32 mg/kg or halothane (Fluothane, Ayert Laboratories, New York, NY) using 4.0% for induction, 2.0% during the surgical preparation, and 0.5%- 1.0% during the experiment. The animals were anesthetized, underwent tracheostomy, placement of femoral arterial and venous catheters, ventilated on a model 670 ventilator (Harvard Apparatus, Millis, MA) and given 0.15 mg pancuronium bromide to abolish respiratory efforts. Body temperature was held at approximately 36 C with a warming pad (K-pad, Gorman-Rupp, Belleville, OH). The skin, subcutaneous tissue, and muscle were reflected over the midsagittal plane and the temporalis muscle was reflected bilaterally to expose the skull. The bone overlying a 3 x 2 cm rectangular area on the parietal region was removed bilaterally using a high-speed air drill (Hall, Inc., Dallas, TX). The underlying dura was removed with the aid of the operating microscope. The dural margins were elevated and the bleeding points coagulated with the bipolar forceps. Three screws were placed in the skull at the bone margins on the right side for EEG monitoring. The dura was replaced bilaterally with a thin plastic sheet (Saran Wrap, Dow Chemical Co., Midland, MI), which kept the brain moist and prevented surface oxygenation. The Paco2 and Pao2 were adjusted to obtain the optimal values of 40 mm Hg and mm Hg, respectively, prior to any experimental procedure. The arterial serum glucose was measured at the initial time of arterial blood gas measurement in all animals and just prior to biopsy. In the control group, biopsy was performed immediately following the establishment of a satisfactory Paco2 and Pao2 measurement and no insulin was administered. In the experimental group, a PE-50 cannula was inserted into the right lingual artery so its tip lay within the carotid artery. This was used in the CBF - 133Xe measurement technique described below. Methods of Measurement Vital Sign Measurements Arterial blood pressures were monitored using strain-gauge transducers attached to the femoral artery catheter. Core body temperature was measured with a rectal thermometer. The femoral artery catheter was the source of samples used to measure arterial Pao2, Paco2, and ph. These measurements were performed on an Instrumentation Laboratory Model 113 Ultra Micro blood gas analyzer (Instrumentation Laboratory, Inc., Lexington, MA). Cerebral Blood Flow Measurements Cerebral blood flow was measured by the clearance of a O.I-ml bolus of normal saline containing 200 /LCi of 133Xe (total volume of each injection was 0.3 ml and included 0.2 ml of the solution of umbelliferone). The bolus was delivered into the right lingual artery catheter. Details of this technique have been previously described (Sundt et ai., 1978). The initial slope technique was used to estimate blood flow. Studies of CBF were made at P ac02s of 40, 20, J Cereb Blood Flow Metabo{, Vol. 2, No

3 CBF, BRAIN ph AND METABOLISM IN INSULIN HYPOGLYCEMIA , and 40 mm Hg, and again just prior to cerebral biopsy. Instrumentation Electroencephalographic and microspectrofluorometric instrumentation have been described previously in detail and will not be reviewed here (Sundt et ai., 1978; Anderson et ai., 1980; Sundt and Anderson, 1980b). Brain ph Intracellular brain tissue ph was measured from an area of cortex free of large pial vessels using a ph-sensitive fluorescent indicator (umbelliferone) injected simultaneously with the xenon. The details of this technique have also been reported (Sundt et ai., 1978). Tissue Metabolite Levels Brain biopsies were taken from both exposed areas of parietal cortex. This was performed using a technique that deposits a sample of brain ( mg) into liquid nitrogen within 1 s (Kramer et ai., 1968). This tissue was stored at -76 C and prepared for analysis in a refrigerated box (- 25 C) as described by Folbergrova et ai. (1972). Tissue extracts were analyzed with enzymatic fluorometric methods for PCr; ATP, ADP, and AMP; glucose (Lowry et ai., 1964); lactate and pyruvate (Lowry and Passonneau, 1972). Data Analysis The Student's t test was used for statistical comparison of the groups. A value for p of less than 0.05 was regarded as significant. All mean values are reported with the standard error of the mean (SEM). The tables and figures were constructed to permit rapid comparison between the controls and experimental animals, the halothane versus barbiturate animals, and those that achieved a significant EEG change and those that did not. Insulin Administration Insulin administration was begun after initial data were obtained in all the experimental animals. Regular insulin (Ileitin, Eli Lilly and Co., Indianapolis, IN) was administered in a dosage of 30 U/kg/h divided into four equal doses and given intravenously at 15-min intervals. An additional 125 U of insulin was given each hour intramuscularly. Insulin was continued until either severe slowing or isoelectricity was obtained in the EEG, or until 6 h had passed. Because of technical problems we could not go beyond 6 h of observation prior to biopsy. RESULTS Control Preparations There were no significant differences between the control values of PCr, ATP, ADP, AMP, lactate, and pyruvate in the halothane animals, compared with the pentobarbital group. The only significant difference between these two groups was the level of serum and brain tissue glucose, with the halothane group being higher in all instances. These data are summarized in Table 1. Experimental Preparations Systemic Response to Hypoglycemia These values are graphically depicted in Fig. and summarized in Tables 2 and 3. The lines of each graph demonstrate the animals with severe EEG changes (dashed lines) and those with no EEG change despite prolonged high-dose insulin administration (solid lines). The mean baseline serum glucose levels were higher for the halothane group than for the barbiturate group. These values are comparable to our previously reported data in seven animals in each group, with serum glucose values of 4.67 p,mollml and 3.41 p,mollml for the halothane and barbiturate groups, respectively (Cilluffo et al., 1981). The mean serum glucose levels at the time of biopsy in the animals with an EEG change were 0.62 p,mollml (11 mg%) for the halothane group and 0.45 p,mollml (8 mg%) for the barbiturate group, a difference that is not statistically significant. The MABP in the halothane group was statistically lower at the time of biopsy in the animals showing an EEG change; in the barbiturate group it was not. There was a statistically significant fall in heart rate in the halothane animals with hypoglycemic slowing of the EEG. This effect represents the increased vagal tone seen in the terminal stages of hypoglycemic coma. The arterial ph was unaffected by severe hypoglycemia whether or not the EEG was slowed. J Cereb Blood Flow Me/abol, Vol. 2, No.3, 1982

4 M. CILLUFFO ET AL. TABLE 1. Metabolic measurements according to electroencephalographic (EEG) activity Serum Number of PCr ATP ADP AMP Glucose Lactate Pyruvate glucose Condition EEG cats (/Lmo]Jg) (/Lmo]Jg) (/Lmollg) (/Lmollg) (/Lmo]Jg) (/Lmo]Jg) (/Lmo]Jg) (/Lmo]Jg) Halothane Control Normal ± ± ± ± ± ± ± ± 0.65 Hypoglycemia No change ± ± ± ± ± ± ± ± 0.08 in EEG EEG change ± ± ± ± 0.04 O. to ± ± ± ± 0.37 Pentobarbital Control Normal 3.57 ± ± ± ± ± ± ± ± 0.55 Hypoglycemia No change 2.87 ± ± ± ± ± ± ± ± 0.49 in EEG EEG change ± ± ± ± ± ± ± ± O. to EEG and CBF Response to Hypoglycemia These data are summarized in Tables 2 and 3 and graphed in Fig. 2. The CBF was significantly higher in the halothane group, compared to the barbiturate animals. Following confirmation of the cerebral Paco2-CBF response, CBF returned to baseline. Cerebral blood flow remained constant during severe hypoglycemia in all halothane-anesthetized animals. In the barbiturate group, CBF was unal- tered in the three animals with no EEG change despite severe hypoglycemia. In the four barbiturate animals that did demonstrate a severe EEG change from hypoglycemia, the CBF remained constant in two and increased moderately in two, for an overall increase that was not statistically significant. The initial CBF for the groups that showed an EEG change were 0.84 ml/g/min and 0.40 ml/g/min for the halothane and barbiturate groups, respectively. Cerebral blood flow at the time of biopsy was 0.90 Pa CO2 (torr) ph arterial MABP (mmhg) Heart rate (beats/min) CBF (ml/g/min) ph brain Cerebrovascular resistance (mm Hg/100 g/min) Tissue indicator perfusion (ml/g/min) I -- === "" ='-:.;-=---t. 66 ' ' : tl '-- -'- --'_ : :.. -: ==== =--- =-: ; : Brain Measurements Systemic Measurements Halothane Pentobarbital _ EEGCh.nQ. Arterial glucose 6. 0 ::=";"EEG (,umol/ ml) =..::: =====-= =---J L.._,,_,,_/_ --_ _ -_ -_-_-_- _--_ _-_- -, _ --' 74. _ :o_ ======-"I '0 =? i P:c=:--= C77:=- o x o x Time (min) Time (min) J 3 j FIG. 1. This series of graphs is derived from the experimental cats in each anesthetic group subjected to insulin hypoglycemia. Insulin in an intravenous dose of 30 U/kg/h was given in equal aliquots every 15 min and an additional 125 U intramuscularly was given until we achieved either an isoelectric EEG or 6 h had passed. The Paco2 was varied as shown at the beginning of each experiment to verify intact autoregulation. The dashed lines represent those animals who demonstrated severe EEG changes and the solid lines demonstrate those who did not. Period X designates the time of brain biopsy, which varied from 4-6 h depending on the EEG reponse. Cerebrovascular resistance was determined as the mean arterial blood pressure (MABP) divided by the cerebral blood flow (CBF). This expresses in mm Hg the pressure needed for 1 ml of blood to pass through 100 g of brain in 1 min. J Cereb Blood Flo)\' Metabol, Vol. 2, No.3, 1982

5 CBF, BRAIN ph AND METABOLISM IN INSULIN HYPOGLYCEMIA 341 TABLE 2. Halothane: Systemic responses to hypoglycemia Flat EEG (n = 7) Time of measurement 0 15 min 30 min 45 min Flat EEG Arterial ph 7.23 ± ± ± ± ± 0.04 Paco2 (mm Hg) 40.0 ± ± ± ± ± 1.30 Pao2 (mm Hg) ± ± ± ± ± 8.04 MABP (mm Hg) 92.9 ± ± ± ± ± 8.8 CBF (mllg/min) 85.4 ± 9.6" 51.3 ± ± ± ± 14.1 Brain ph 7.01 ± ± ± ± ± 0.02 Indicator perfusion (mlll00 g/min) 0.62 ± ± ± ± ± 0.06 Glucose (!Lmollml) 5.48 ± ± 0.08a CVR (mm Hg/100 g/min) 1.17 ± ± ± ± ± 0.19 Heart rate (bpm) 183 ± ± ± ± ± 13a Time of measurement 0 15 min TABLE 2-Continued No change in EEG (n = 2) 30 min 45 min 6h Arterial ph 7.37 ± ± 0.05 p.co2 (mm Hg) 40.1 ± ± 0.3 p.o2 (mm Hg) ± ± 7.3 MABP (mm Hg) 92.5 ± ± 8.5 CBF (mllg/min) ± ± 12.5 Brain ph 7.07 ± ±0.12 Indicator perfusion (mllioo g/min) 0.63 ± ± 0.21 Glucose (!Lmollml) 7.69 ± 3.08 CVR (mm Hg/100 g/min) 0.63 ± ± 0.25 Heart rate (bpm) 186 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 0.37a 0.25 ± ± ± ± ± ± 12.0 a Statistically significant (p < 0.05). Halothane CBF > barbiturate CBF (p < 0.05). mug/min and 0.79 mug/min, respectively, for the two groups. The CVR was unaffected by the severe decrease in serum and brain tissue glucose levels. Tissue clearance of the ph indicator was also unaffected by the decreased glucose availability. As in our previous experiments (Cilluffo et al., 1981), we found no significant change in brain ph in severe hypoglycemia in either anesthetic group, re- gardless of whether the EEG was normal, slowed, or isoelectric. Cerebral Metabolite Response to Hypoglycemia Table 1 shows the values of the various metabolites measured. There were significant changes induced by hypoglycemia, which will be discussed below. J Cereb Blood Flow Metabot. Vat. 2, No

6 M. CILLUFFO ET AL. TABLE 3. Pentobarbital: Systemic responses to hypoglycemia Flat EEG (n = 7) Time of measurement 0 15 min 30 min 45 min Flat EEG Arterial ph 7.28 ± ± ± ± ± 0.03 Paco2 (mm Hg) 41.8 ± ± ± ± ± 1.44 PaD2 (mm Hg) ± ± ± ± ± 8.47 MABP (mm Hg) 84.9 ± ± ± ± ± 10.4 CBF (ml/glmin) 48.5 ± ± ± ± ± 12.1 Brain ph 6.99 ± ± ± ± ± 0.04 Indicator perfusion (mllioo glmin) 0.44 ± ± ± ± ± 0.13 Glucose (flmollml) 4.77 ± ± 0.1" CVR (mm HgilOO glmin) 1.91 ± ± ± ± ± 0.32 Heart rate (bpm) 168 ± ± ± ± ± 27.0 Time of measurement 0 15 min TABLE 3-Continued No change in EEG (n = 2) 30 min 45 min 6h Arterial ph 7.31 ± ± 0.02 PaC02 (mm Hg) 34.7 ± ± 0.88 PaD2 (mm Hg) ± ± 30.2 MABP (mm Hg) 82.7 ± ± 25.4 CBF (mllglmin) 40.8 ± ± 1.20 Brain ph 7.00 ± ± 0.03 Indicator perfusion (milioo glmin) 0.22 ± ± 0.03 Glucose (flmollml) 3.22 ± 0.4 CVR (mm HgilOO glmin) 2.13 ± ± 0.97 Heart rate (bpm) 160 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 0.49" 0.99 ± ± ± ± ± ± 13.9 " Statistically significant (p < 0.05). Halothane CBF > barbiturate CBF (p < 0.05). DISCUSSION Fluorescent Indicator Analysis The use of a ph-sensitive fluorescent indicator is a relatively new technique. Umbelliferone is a fatsoluble, ph-sensitive, fluorescent indicator that is freely diffusible across the blood -brain barrier and is nontoxic (Sundt et ai., 1978). A nomogram can be constructed based on fluorescence ratios, and com- parison of the relative intensity of simultaneous clearance curves of the indicator to this nomogram permits the accurate determination of brain ph. Details have been reviewed elsewhere (Sundt and Anderson, 1980b). Measurements using this method are uniquely free of the tissue damage associated with any type of microelectrode and no diffusion barriers interfere with measurements. In previous studies using this method, we have found that brain ph (1) is relatively more acid at J Cereb Blood Flow Metaho/, Vol. 2, No.3, 1982

7 CBF, BRAIN ph AND METABOLISM IN INSULIN HYPOGLYCEMIA 343 e c: 0 0 "It) C") >( e GI GI """""" ! A OL- L- -L L-_ rA O ---- o Time (sec) FIG. 2. Photograph of a typical fluorescent clearance curve of umbelliferone. The equation for the ratio of 370 and 340 nm excited 450 nm fluorescence in the vascular space (point A) and in the tissue (points B, C, and 0) is total fluorescence at 450 nm with 370 nm excitation nm background fluorescence from NAOH-NAOPH -0- total fluorescence at 450 nm with 340 nm excitation nm background fluorescence from NAOH-NAOPH = 450 nm fluorescence at 370 nm excitation attributable to indicator nm fluorescence at 340 nm excitation attributable to indicator. Thus, at point A, the arterial spike, the equation is = = (ph 7.175; from nomogramsimultaneous arterial sample, ph = 7.165); at point S, equation equals = = (ph 6.980, from nomogram); similar calculations at points C and o produce brain ph measurements of and 6.980, respectively. Points B, C, and 0 are separated from each other by 24-s intervals. Analysis of further points on the clearance curve of umbelliferone is less reliable since a greater portion of fluorescence in this portion of the curve is attributable to background due to NAOH and NAOPH, and, therefore, values here are less reliable. Nevertheless, calculations on these points of the curve appear unchanged from points S, C, and O. normocapnia under light halothane anesthesia (ph, 7.04) than under light barbiturate anesthesia (ph 7.11) (Anderson et al., 1980); (2) varies with the depth of anesthesia (Anderson et ai., 1980); (3) changes with Paco2 but not with acute changes in arterial ph (Sundt and Anderson, I980a); and (4) agrees quite closely with the calculated equivalent intracellular ph reported by Siesjo (1978) and Davidian et al. (1978). Important information can also be acquired from analysis of the indicator's clearance curve. Immediately upon leaving the intravascular space, umbelliferone enters a compartment too acidic to be the extracellular compartment (Sundt and Anderson, 1980a). Parenthetically, this is evidence for an intracellular transport route across the blood -brain barrier (namely, capillary to glial cell to neuron) (Sundt and Anderson, 1980a). The correlation of these brain ph measurements with the calculated intracellular ph and the data from clearance curves both suggest that these measurements primarily affect intracellular ph. However, in that umbelliferone is fat-soluble, it is concentrated in the membranes. Thus, we are, in fact, measuring the ph at the interface of the various membranes and the cytosol, extracellular fluid, etc. (Sundt and Anderson, I980b). It represents an "equivalent brain ph," somewhat analogous to the terms "equivalent intracellular ph" used by Siesjo to recognize the fact that the intracellular space has areas of differing ph with multiple gradients and, therefore, a calculated ph is a representation of the whole (Siesjo, 1978). Thus, a change in extracellular ph would be expected to influence our measurements. Metabolic Changes with Hypoglycemia The primary purpose of the work reported here was the correlation of brain blood flow with brain ph to determine if these two measurements remained "coupled" in the setting of hypoglycemia. Although we were not primarily interested in the metabolic changes occurring in areas of hypoglycemia, other than as they might be reflected by changes in brain ph, a number of metabolic determinations were made to be certain that we were, in fact, measuring blood flow from a brain that was suffering from a reduction in glucose below its critical levels. This study again demonstrates the relative resistance of the cat brain to hypoglycemia. We were able to achieve a severely slowed or isoelectric EEG in seven of nine halothane cats, but in only four of seven barbiturate cats. The levels achieved, i.e., 0.62 /-tmouml in the halothane group and 0.45 J Cereb Blood Flow Metabol. Vol. 2. No

8 M. CILLUFFO ET AL. JLmoVml in the barbiturate group, we believe, are about as low as can be achieved in the cat. Certainly the levels are lower than most previously published series in humans (Kety et ai., 1948; Della Porta et ai., 1964) and are comparable to those achieved in rats (Norberg and Siesj6, 1976; Agardh et ai., 1981a). Although we made measurements from only one single area of cat brain, the parietal-temporal cortex, there are reportedly only minimal differences among various regions in the cerebral cortex during insulin hypoglycemia (Ferrindelli and Chang, 1973; Agardh et ai., 1981a). To our knowledge, the only differences in regional cerebral metabolism during hypoglycemia that have been reported are between the cerebellum and cerebral cortex (Agardh et ai., 1981a). Buchweitz et al. (1980) showed no regional differences in oxygen consumption in the anesthetized cat brain. Thus, we believe our sampling technique can be assumed to be representative of cortical metabolism in the cat brain during hypoglycemia. Brain tissue glucose levels correlated most closely with changes in the EEG secondary to hypoglycemia. The serum glucose levels in some animals were significantly depressed, and yet some of these animals did not demonstrate an altered EEG. All animals with severe slowing or isoelectricity of the EEG had very low brain tissue glucose levels. These findings agree with those of Ferrindelli and Chang (1973), who found that brain tissue glucose levels rather than serum glucose levels were most reflective of behavioral alterations in hypoglycemic mice. Although Lewis et al. (1947a,b) and Feise et al. (1976) have shown that EEG and clinical behavior correlate well, this relationship cannot be observed in a paralyzed anesthetized cat. Levels of other energy metabolites were also depressed in hypoglycemic brain samples. The changes in the determinants summarized in Table 1 compare favorably with the changes found by other workers in the various parameters investigated, with appropriate allowances for differences in species and effects of anesthetic agents. Specifically, the brain glucose and lactate levels during profound hypoglycemia were similar to those reported by Pelligrino and Siesj6 (1981) and Agardh et al. (1981b). It is well known that various anesthetic agents alter the ratio of brain glucose to serum glucose (Biebuyck, 1973; Mayman et ai., 1964). This, plus the effects of halothane on serum glucose levels (Biebuyck and Alberti, 1971, 1972), explains the difference between the control brain glucose levels in the two groups. All hypoglycemic animals showed decreased PCr and ATP levels when compared with controls. Furthermore, the level of decline was more pronounced in those animals with EEG changes. Brain tissue lactate levels were also diminished, but here we did not find changes that were highly significant, although there was an unexplained difference between the halothane and barbiturate groups. The low level of lactate in these animals may have some relevance in comparing the hypoglycemic brain with the anoxic brain. Cerebral Blood Flow and Brain ph As indicated above, our primary goal in this study was to compare the changes in CBF and brain ph in cats during hypoglycemia. Correlating these data with previous investigations, we hoped to be able to make some observations concerning the mechanism of the CBF- metabolic blood flow couple. In that we found no major changes in CBF during profound hypoglycemia, we are unable to add any positive data on this subject. The preservation of a Pacoz- CBF response in these animals was interesting and suggested to us that this response, although lost in ischemia and commonly used as a test of cerebrovascular reactivity, is not dependent on cerebral metabolic activity and has a different mechanism from autoregulation and the coupling of CBF and metabolism (Lassen, 1974; Sokoloff, 1978). However, it should be noted that this response was monitored early in the experiment before profound hypoglycemia had been achieved. The differences in CBF in the current study from those found by Norberg and Siesj6 (1976), Agardh et al. (l981a), and Abdul-Rahman et al. (1980) must be explained. These investigators, using a variety of techniques including 14C-iodoantipyrine autoradiology (Abdul-Rahman et ai., 1980), found an increase in CBF in the rat during profound hypoglycemia. Although this might be a species difference, the difference in the data can best be explained by examining the work of Nilsson et al. (1981). They found that during severe hypoglycemia, CO2 responsiveness to the brain was preserved while autoregulation was abolished. Ventilated normoglycemic rats under 70% nitrous oxide (N20) had a higher than normal blood pressure, but CBF remained normal in these animals with preserved autoregulation. However, during hypoglycemia, au- J Cereb Blood Flow Metabol, Vol. 2, No.3, 1982

9 CBF, BRAIN ph AND METABOLISM IN INSULIN HYPOGLYCEMIA 345 to regulation was abolished and CBF increased in these animals if the blood pressure was held at the relatively hypertensive levels ordinarily observed in the animals anesthetized with 70% N20. In animals in which the blood pressure fell to those levels observed in the unparalyzed awake animals, these differences were no longer observed. In order to further evaluate this hypothesis, six additional animals were studied after profound hypoglycemia, reflected by a flat EEG, had been achieved (Cilluffo et ai., unpublished observations). In animals with a normal blood pressure, the Paco2-CBF response was preserved. In animals with a low blood pressure, this response was lost. Elevations of MABP resulted in marked elevations of CBF (over 200% in two animals). Thus, the results of the current study, when considered with those of Siesjo and his coworkers, suggest that the Paco2-CBF response curve, athough lost in ischemia, is preserved during hypoglycemia if the animal remains normotensive, and that the mechanisms for this response curve are different from metabolic autoregulation. U sing methods similar to those used in this study, we have found that brain ph and CBF changes parallel fluctuations in P ac02 (Anderson et ai., 1980; Sundt and Anderson, 1980a). If, however, the Paco2 is held constant, brain ph remains essentially unchanged and apparently unrelated to major differences in CBF. For example, there are very significant differences in CBF in animals anesthetized with halothane and barbiturate (Anderson et ai., 1980; Sundt and Anderson, 1980b). Yet, if brain CBF is compared at an identical ph in animals anesthetized with these two different agents, one finds that CBF is 150% higher in those animals anesthetized with halothane. Furthermore, varying levels of halothane anesthesia at a constant Paco2 of 41 mm Hg have shown a CBF of 76 mv100 g/min at 0.1% halothane versus a CBF of 47 at 3% concentration-anesthetic concentrations that produce the same intracellular brain ph (Anderson et ai., 1980). Studies during seizures (Tenny et ai., 1981) indicate a very major difference in brain ph between the primary and secondary seizure foci without a concomitant difference in CBF. We conclude that (1) CBF does not change in hypoglycemia during normotension; (2) hypoglycemia abolishes autoregulation and, therefore, elevations in CBF can be expected with elevations in MABP; (3) the Paco2-CBF response curve is not metabolically coupled; and (4) major changes in CBF related to the anesthetic agent, the depth of anesthesia, and the metabolic rate (seizure activity) are identified that cannot be explained by changes in brain ph. It appears that brain ph is not the primary modulator of metabolically induced changes in CBF (Astrup et ai., 1978). REFERENCES Abdul-Rahman A, Agardh C-D, Siesjo BK (1980) Local cerebral blood flow in the rat during severe hypoglycemia, and in the recovery period following glucose injection. Acta Physiol Scand 109: Agardh C-D, Kalimo H, Olsson Y, Siesjo BK (1981a) Hypoglycemic brain injury: Metabolic and structural findings in rat cerebellar cortex during profound insulin-induced hypoglycemia and in the recovery period following glucose administration. J Cereb Blood Flow Metaboll:71-84 Agardh C-D, Chapman AG. Nilsson B, Siesjo BK (1981b) Endogenous substrates utilized by rat brain in severe insulininduced hypoglycemia. J Neurochem 36: Anderson RE, Michenfelder JD, Sundt TM Jr (1980) Brain intracellular ph, blood flow, and blood-brain barrier differences with barbiturate and halothane anesthesia in the cat. Anesthesiology 52: Astrup J, Heuser D, Lassen AN, Nilsson B, Norberg K, Siesjo BK (1978) Evidence against H+ and K+ as main factors for the control of cerebral blood flow: A microelectrode study. In: Cerebral Vascular Smooth Muscle and Its Control, Amsterdam. Elsevier, pp Biebuyck JF (1973) Anesthesia and hepatic metabolism: Current concepts of carbohydrate homeostasis. Anesthesiology 39: Biebuyck JF, Alberti KGMM (1971) Effects of halothane on blood metabolites and serum insulin in the rat. Diabetologia 7:471 Biebuyck JF, Alberti KGMM (1972) The effects of anaesthetic agents on carbohydrate metabolism in the rat. Clin Sci 42:4-5 Buchweitz E, Sinha AK, Weiss HR (1980) Cerebral regional oxygen consumption and supply in anesthetized cat. Science 209: Cilluffo JM, Anderson RE, Sharbrough FW, Sundt TM Jr (1981) Correlation of brain blood flow, intracellular ph and metabolism in hypoglycemic cats under halothane and barbiturate anesthesia. Brain Res 216: Davidian NM, Butler TC, Poole DT (1978) The effect of ketosis induced by medium chain triglycerides on intracellular ph of mouse brain. Epilepsia 19: Della Porta P, Marolo AT, Negri VU, Rossella E (1964) Cerebral blood flow and metabolism in therapeutic insulin coma. Metabolism 13: Eisenberg S, Seltzer HS (1962) The cerebral metabolic effects of acutely induced hypoglycemia in human subjects. Metabolism 11: Feise G, Kogure K, Busto R, Scheinberg P, Reinmuth OM (1976) Effect of insulin hypoglycemia upon cerebral energy metabolism and EEG activity in the rat. Brain Res 126: Ferrindelli JA, Chang M-M (1973) Brain metabolism during hypoglycemia. Arch Neurol 28: Folbergrova J, MacMillan V, Siesjo BK (1972) The effect of moderate and marked hypercapnia upon the energy state and upon the cytoplasm NAD/NADH+ ratio of the rat brain. J Neurochem 19: Gottstein U, Held K (1967) Insulinwirkun auf den menschlichen Hirnmetabolisms von Sloffwechselfesunden und Diabetikern. Klin Wschs 45:18-23 Kety SS, Woodford RB, Harmel MH, Freyhan FA. Appel KE, J Cereb Bluud Flow Metabol, Vol. 2, No.3, 1982

10 M. CILLUFFO ET AL. Schmidt CF (1948) Cerebral blood flow and metabolism in schizophrenia. The effects of barbiturate semi-narcosis, insulin coma and electroshock. Am J Psychiatry 104: Kramer RS, Sanders AP, Lesage AM, Woodhall B, Sealy WC (1968) The effect of profound hypothermia on preservation of cerebral ATP content during circulatory arrest. J Thorac Cardiovasc Surg 56: Lassen NA (1974) Control of cerebral circulation in health and disease. Circ Res 34: Lewis LD, Ljunggren B, Norberg K, Siesjii BK (l974a) Changes in carbohydrate substrates, amino acids and ammonia in the brain during insulin-induced hypoglycemia. J Neurochem 23: Lewis LD, Ljunggren B, Ratcheson RA, Siesjii BK (1974b) Cerebral energy state in insulin-induced hypoglycemia, related to blood glucose and to EEG. J Neurochem 23: Lowry OH, Passonneau JV, Hasselberger FX, Schulz DW (1964) Effect of ischemia on known substrates and co-factors of the glycolytic pathway in brain. J Bioi Chem 239: Lowry OH, Passonneau JV (1972) A Flexible System of Enzymatic Analysis. New York, Academic Press Mayman CI, Gatfield PD, Breckenridge BM (1964) The glucose content of brain in anaesthesia. J Neurochem 11: Nilsson B, Agardh C-D, Ingvar M, Siesjii BK (1981) Cerebrovascular response during and following severe insulin-induced hypoglycemia: CO2-sensitivity, autoregulation, and influence of prostaglandin synthesis inhibition. Acta Physiol Scand 111: Norberg K, Siesjii BK (1976) Oxidative metabolism of the cerebral cortex of the rat in severe insulin induced hypoglycemia. J Neurochem 26: Pelligrino D, Almquist L-O, Siesjii BK (1981) Effects of insulin-induced hypoglycemia on intracellular ph and impedance in the cerebral cortex of the rat. Brain Res 221: Pelligrino D, Siesjii BK (1981) Regulation of extra- and intracellular ph in the brain in severe hypoglycemia. J Cereb Blood Flow Metabol 1: Siesjii BK (1978) Brain Energy Metabolism. New York, Wiley Sokoloff L (1978) Local cerebral energy metabolism: Its relationships to local functional activity and blood flow. In: Cerebral Vascular Smooth Muscle and Its Control, Amsterdam, Elsevier, pp Sundt TM Jr, Anderson RE, Van Dyke (1978) Brain ph measurements using a diffusible, lipid soluble ph-sensitive fluorescent indicator. J Neurochem 31: Sundt TM Jr, Anderson RE (l980a) Intracellular brain ph and the pathway of a fat soluble ph indicator across the bloodbrain barrier. Brain Res 186: Sundt TM Jr, Anderson RE (l980b) Umbelliferone as an intracellular ph-sensitive fluorescent indicator and bloodbrain barrier probe instrumentation, calibration, and analysis. J Neurophysiol 44: Tenny RT, Sharbrough RW, Anderson RE (1981) Correlation of intracellular redox states and ph with blood flow in primary and secondary seizure foci. Ann Neurol 8: J Cereh Blood Flow Metahol, Vol. 2. No.3, 1982