Deep hypothermia is currently used to protect the

Transcription

1 Optimal Blood Flow for Cooled Brain at 20 C Takao Watanabe, MD, Naoshi Oshikiri, MD, Kiyoshige Inui, MD, Setsuo Kuraoka, MD, Takashi Minowa, MD, Jun Hosaka, MD, Toshiki Takahashi, MD, and Yasuhisa Shimazaki, MD Second Department of Surgery, Yamagata University School of Medicine, Yamagata, Japan Background. Optimal conditions for deep hypothermic perfusion and protective brain blood flow remain unclear. Methods. Dogs (n 52) underwent 120 minutes of cardiopulmonary bypass at 20 C with perfusion flow rates of 2.5, 5, 10, 20, 40, and 100 ml kg 1 min 1.We examined the effect of the various flow rates and different perfusion pressures on brain blood flow, metabolism, and intracellular ph. Results. The brain was ischemic and acidotic when the perfusion flow rate was less than 5 ml kg 1 min 1 and pressure was less than 10 mm Hg. When perfusion pressure was higher than 10 mm Hg, cerebral cortex blood flow was more than 9 ml 100 g 1 min 1 and intracellular ph, higher than The cerebral metabolic rate for oxygen decreased at a flow rate of 2.5 ml kg 1 min 1. The cerebral metabolic ratio of glucose to oxygen and the cerebral vascular resistance were lowest when perfusion pressure was 10 to 30 mm Hg. Full-flow (100 ml kg 1 min 1 ) perfusion caused paradoxical brain acidosis; a flow of 40 ml kg 1 min 1 provided the best results. Conclusions. Both extremely low-flow perfusion and excessive perfusion cause brain acidosis. Low-flow perfusion at a pressure of 20 mm Hg provides cerebral vasorelaxation and aerobic metabolism during operations at 20 C. (Ann Thorac Surg 1999;68:864 9) 1999 by The Society of Thoracic Surgeons Deep hypothermia is currently used to protect the brain during operations for aortic arch aneurysms and congenital heart diseases. Cerebral perfusion instead of total circulatory arrest may further enhance brain protection even during deep hypothermia. Several groups have suggested a reduction in cerebral metabolism [1 5] and a minimal flow rate during cardiopulmonary bypass (CPB) [3 8] to protect the brain during deep hypothermia. However, in contrast to the warm brain, where the threshold pressure and blood flow are known, the protective ranges in cerebral blood flow and perfusion pressure to maintain aerobic metabolism in the cooled brain have not been established [9]. We examined the effects of deep hypothermic perfusion at 20 C on brain blood flow and metabolism to clarify the optimal perfusion flow and pressure for brain protection during surgical procedures. Material and Methods Animal Preparation and CPB All animals received humane care in compliance with the Guide for Animal Experimentation, Yamagata University School of Medicine and the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH publication 85-23, revised 1985). Fifty-two adult mongrel dogs weighing kg Accepted for publication March 2, Address reprint requests to Dr Watanabe, Second Department of Surgery, Yamagata University School of Medicine, Iida-Nishi 2-2-2, Yamagata , Japan. were anesthetized with sodium pentobarbital (25 mg/kg for induction and 4 mg kg 1 h 1 for maintenance) and ventilated mechanically. The brain was stained with neutral red (1.8 g in 200 ml of saline solution) administered intraperitoneally. Catheters for pressure monitoring and blood sampling were placed in the thoracic aorta and inferior vena cava through the femoral artery and vein. The sagittal sinus was exposed and cannulated with a 14-gauge venous catheter through a midline craniotomy after heparinization (4 mg/kg). The heart was exposed through a midline sternotomy, and animals were connected to a CPB circuit with cannulation of the ascending aorta and right atrium. Animals were cooled to 20 1 C (measured nasopharyngeally and rectally) by CPB with a perfusion flow rate of 100 ml kg 1 min 1. Thereafter, they underwent deep hypothermic perfusion for 120 minutes at 20 C with a perfusion flow rate randomly set at 2.5,5(n 10), 10, 20, 40, or 100 ml kg 1 min 1. The aortic root was cross-clamped throughout perfusion. The pressure in the inferior vena cava was maintained near 0 by setting the venous reservoir at an adequate height. Ten million colored microspheres (15 m) (E-Z Trac, Los Angeles, CA) were injected into the arterial line of the CPB circuit at the end of core cooling and at 60 minutes of deep hypothermic perfusion, both at 20 C. After 120 minutes of perfusion at 20 C with a flow rate of 2.5, 5, 10, 20, 40, or 100 ml kg 1 min 1, the brain was excised and sliced for measurements of brain blood flow and regional intracellular ph (phi) by The Society of Thoracic Surgeons /99/$20.00 Published by Elsevier Science Inc PII S (99)

2 Ann Thorac Surg WATANABE ET AL 1999;68:864 9 BRAIN BLOOD FLOW AND METABOLISM AT 20 C 865 The CPB circuit was primed with a crystalloid solution that resulted in hemodilution to 20.6% 3.6% of hematocrit during deep hypothermic perfusion. Returned blood was oxygenated by a membrane oxygenator (HPO20HC; Senko Medical Instrument Mfg Co, Ltd, Tokyo, Japan) in which arterial carbon dioxide tension was kept relatively high ( mm Hg at 37 C for 60 minutes of deep hypothermic CPB) to avoid cerebral vasoconstriction. Acid-base balance was normally regulated with sodium bicarbonate (base excess, for 60 minutes of deep hypothermic perfusion). A calibrated roller pump was used for perfusion. Arterial, systemic venous, and sagittal sinus blood was sampled and analyzed for ph, oxygen and carbon dioxide tensions, and oxygen content with a blood gas analyzer (ABL-300; Radiometer, Copenhagen, Denmark) and glucose content with a glucose and lactate analyzer (YSI 2300 STAT; Yellow Springs Instrument Co Inc, Yellow Springs, OH). The cerebral and whole body metabolic rates (CMRO 2, CMRGlu, VO 2 ) rates for oxygen and glucose were calculated using the Fick principle. Brain Regional Blood Flow and phi Brain slices were divided into 11 parts: anterior, middle, and posterior cerebral cortex, thalamus, caudate nucleus, hippocampus, pons, cerebellum, and anterior, middle, and posterior cerebrum. These slices and samples (2 to 3 g) of the kidney and jejunum were chemically digested to count trapped microspheres. Regional tissue blood flow was calculated from the sample weight, microsphere count, and perfusion flow rate as described previously [10]. Whole-brain blood flow (WBBF) per 100 g of brain tissue was calculated using whole-brain samples as follows: WBBF (rcbf sample weight)/whole-brain weight, where rcbf is regional cerebral blood flow in each sample. Cerebral vascular resistance (units 100 g) was calculated as perfusion pressure divided by wholebrain blood flow. Brain phi was measured with the modified photometric method of LaManna and coworkers [11, 12]. The photoabsorption spectrum of the dye changes linearly with phi between 5.5 and 7.8. Color changes in brain slices were recorded on color slides at 20 C immediately after deep hypothermic perfusion. Using a dualwavelength color spot scanner (CS-9000; Shimadzu, Kyoto, Japan), we measured photoabsorption at two peaks, 440 and 535 nm, in eight regions of the brain. A standard curve for brain ph was obtained as a function of the photoabsorption ratio at 535 and 440 nm using diluted and acid-base titrated brain homogenates with neutral red at 20 C. Methods for blank ph standard curve and equation for phi have been detailed previously [12], where phi dye absorption 535/440 (r 0.992). Statistical Analysis Results are expressed as the mean the standard deviation or individually in scattergrams. Analysis of variance and Student s t tests (paired and unpaired) were used for statistical analyses; a p value of less than 0.05 was considered significant. Hemodynamics, cerebral metabolic ratio of glucose to oxygen, cerebral cortex blood flow, and phi were further analyzed against perfusion pressure because it may have a dominant effect on cerebral blood flow [13]. The correlation between cerebral cortex blood flow and phi was examined using the same samples. Logistic regression lines were divided into two plots when a breakdown threshold was observed. Lotus Improv R2J (Lotus Development Co, Cambridge, MA) and Microsoft Excel 5.0 (Microsoft Co, Tokyo, Japan) were used. Results Organ Blood Flow and Metabolism Whole-brain blood flow remained in a very low range across the wide variety of perfusion flow rates and pressures (Table 1; Fig 1). It decreased significantly at 60 minutes of perfusion at 20 C with flow rates of 20, 10, 5, and 2.5 ml kg 1 min 1 compared with baseline (end of cooling perfusion) ( p 0.05). It was lower at 10 and 20 ml kg 1 min 1 than at 40 and 100 ml kg 1 min 1 ( p 0.026) but higher than at 2.5 and 5 ml kg 1 min 1 ( p 0.027). Tissue blood flow in the kidney and jejunum was linearly correlated with perfusion flow rate (r 0.73 and 0.72, respectively) and logarithmically correlated with perfusion pressure (r 0.64 and 0.66, respectively). On the other hand, brain blood flow was preferentially maintained when perfused at low flow and low pressure. Blood flow delivery to the brain, weighing 0.61% 0.10% of body weight, was 4.5% 3.3% and 0.91% 0.40% of total perfusion flow at perfusion pressures of less than and greater than 20 mm Hg, respectively ( p ) (Fig 2). Cerebral vascular resistance at 60 minutes of deep hypothermic perfusion had a linear positive correlation with perfusion pressure higher than 10 mm Hg (r 0.85) (Fig 3), where vasorelaxation was observed in a low pressure range but one greater than 10 mm Hg. Whole-body oxygen consumption was logarithmically correlated with perfusion flow rate (r 0.81). The cerebral metabolic rate for oxygen decreased only with perfusion of 2.5 ml kg 1 min 1 ( p 0.004). It was lower at 60 minutes of deep hypothermic perfusion in the 2.5 and 100 ml kg 1 min 1 groups than in the 40 ml kg 1 min 1 group ( p and p 0.01, respectively). The cerebral metabolic rate for glucose did not decrease but increased at a perfusion flow rate of 10 ml kg 1 min 1 ( p 0.013) (see Table 1). Brain Regional Blood Flow and phi Brain blood flow changed with perfusion flow rate in all regions ( p , one-way analysis of variance), but distribution of regional blood flow was homogeneous even during very low flow perfusion (Table 2). Brain regional phi was higher after 40 ml kg 1 min 1 perfusion than after perfusion at lower flow rates (Table 3). It was lower after perfusion at 100 ml kg 1 min 1 than after 40 ml kg 1 min 1 perfusion in the middle and posterior cortex and the thalamus. After perfusion at flow rates of 2.5 and 5 ml kg 1 min 1, the brain was acidotic.

3 866 WATANABE ET AL Ann Thorac Surg BRAIN BLOOD FLOW AND METABOLISM AT 20 C 1999;68:864 9 Table 1. Organ Blood Flow and Metabolism During Perfusion at 20 C a Perfusion Flow Rate (ml kg 1 min 1 ) Variable Baseline (100) (n 52) p Value (ANOVA) Perfusion pressure (mm Hg) b b 12 1 b 16 4 c 22 6 b d Whole-brain blood flow b c b e (ml 100 g 1 min 1 ) Brain vascular resistance (U 100 g) b e e Kidney blood flow c c c e e (ml 100 g 1 min 1 ) Jejunal blood flow (ml 100 g 1 min 1 ) b c e b b Whole-body oxygen consumption (ml kg 1 min 1 ) b b b b b CMRo 2 (ml 100 g 1 min 1 ) c f d CMRGlu (ml 100 g 1 min 1 ) d a Data are shown as the mean the standard deviation. b Significance: p 0.002, intragroup decrease from baseline at 60 minutes of perfusion. c Significance: p 0.01, intragroup decrease from baseline at 60 minutes of perfusion. d Significance: p 0.05, intragroup increase from baseline at 60 minutes of perfusion. e Significance: p 0.05, intragroup decrease from baseline at 60 minutes of perfusion. f Significance: p 0.01, intragroup increase from baseline at 60 minutes of perfusion. ANOVA one-way analysis of variance among groups; Baseline (100) end of cooling perfusion (100 ml kg 1 min 1, 20 C), data from all animals shown; CMRGlu cerebral metabolic rate for glucose; CMRo 2 cerebral metabolic rate for oxygen. Relationships Between Perfusion Pressure, Brain Blood Flow, and Metabolism Cerebral cortex blood flow was maintained within a narrow range logarithmically correlated with perfusion pressure (r 0.71). Critical closing pressure for the brain was between 5 and 10 mm Hg during perfusion at 20 C. Autoregulation [14 17] of cerebral blood flow was maintained during deep hypothermia across a wide range of perfusion pressures higher than 10 mm Hg (Fig 4). Cerebral cortex phi had a logarithmic correlation with perfusion pressures of less than 15 mm Hg (r 0.71) but not greater than 15 mm Hg (r 0.1) (Fig 5). Perfusion pressures higher than 10 mm Hg guaranteed a cerebral cortex blood flow of more than 9 ml 100 g 1 min 1 and a cerebral cortex phi of higher than 6.95 (see Figs 4, 5). When the perfusion pressure was less than 10 mm Hg and the cerebral cortex blood flow was less than 9 ml 100 g 1 min 1, the brain was acidotic (see Figs 5 and 6), and the glucose to oxygen metabolic ratio increased (Fig 7). As evidence of aerobic metabolism, the cerebral metabolic ratio of glucose to oxygen was lowest when perfusion flow rate was 40 ml kg 1 min 1 and pressure around 20 mm Hg. Perfusion at 100 ml kg 1 min 1 or very low flow elevated the value, findings suggesting anaerobic metabolism (see Fig 7). Comment Our data demonstrate that 120 minutes of perfusion at a flow rate of less than 5 ml kg 1 min 1 and a pressure Fig 1. Organ blood flow versus perfusion pressure at 20 C. Organ blood flows were logarithmically correlated with perfusion pressure: brain blood flow 5.3 ln (mm Hg) 5.1; kidney blood flow 104 ln (mm Hg) 212; jejunal blood flow 42 ln (mm Hg) 89, where r was 0.63, 0.64, and 0.66, respectively. Blood flow was lower in the brain than in the kidney and jejunum (p and p , respectively, two-way analysis of variance.) Fig 2. Percent blood flow delivery to the brain versus perfusion pressure at 20 C. Blood flow in the whole brain was divided by total perfusion flow.

4 Ann Thorac Surg WATANABE ET AL 1999;68:864 9 BRAIN BLOOD FLOW AND METABOLISM AT 20 C 867 Fig 3. Cerebral vascular resistance (CVR) versus perfusion pressure at 60 minutes of deep hypothermic perfusion (2.5 to 100 ml kg 1 min 1 ). Two plots, represented by the two linear regression lines, are shown: CVR 1.2 mm Hg 12 (r 0.52) and CVR mm Hg 0.62 (r 0.85) at a perfusion pressure of less than and greater than 10 mm Hg, respectively. of less than 10 mm Hg causes brain acidosis with persistent glycolysis even at 20 C. The results agree with those in a study by Swain and associates [6] and appear to lower the protective perfusion flow range previously estimated by Miyamoto and colleagues [4] to be more than 30 ml kg 1 min 1 at 20 C. The use of a membrane oxygenator in the present study may have extended the protective flow range because it reduces gaseous microembolism [17]. In studies of deep hypothermia with bubble oxygenators, the threshold perfusion flow range for brain protection had been estimated at approximately one fourth of full-flow range [3 5, 8]. Under these conditions, a higher perfusion flow rate than presumed by Swain and coworkers [6] and by us in the present study would be required to fully oxygenate the brain because Fig 4. Cerebral cortex blood flow at 20 C. Blood flow was logarithmically correlated with perfusion pressure at 60 minutes of deep hypothermic cardiopulmonary bypass (DHCPB) (E 2.5 to 100 ml kg 1 min 1 ): cerebral cortex blood flow 9.0 ln (mm Hg) 11.1 (r 0.71). There was no correlation at the end of cooling perfusion ( baseline, 100 ml kg 1 min 1 ). brain swelling would be more severe with a bubble oxygenator than a membrane oxygenator owing to more microemboli and the inflammatory response [17]. In the ischemic brain, lack of oxygen and eventual reduction in the tricarboxylic acid cycle cause a lack of energy for cell survival. The mature brain compensates for this exclusively by anaerobic glycolysis. Glycolysis metabolizes each glucose molecule into two adenosine triphosphate molecules, two lactate molecules, and two hydrogen ions; the rate of glucose consumption to adenosine triphosphate production is 19-fold that during aerobic metabolism. Therefore, the ischemic brain metabolizes glucose at a higher rate than the normal brain with a reduced oxygen supply. In the latter case, the metabolic ratio of glucose to oxygen increases during the process to intracellular lactic acidosis [18]. As reported previously [4 8], incomplete perfusion at very low flow Table 2. Regional Cerebral Blood Flow During Perfusion at 20 C a Perfusion Flow Rate (ml kg 1 min 1 ) Area Baseline (100) (n 52) p Value (ANOVA) Anterior cortex b c Middle cortex b d d c Posterior cortex b d b c c Thalamus b b b Caudate nucleus d d Hippocampus b d b e Pons b d d Cerebellum d d b a Data are shown as the mean the standard deviation. b Significance: p 0.002, intragroup decrease from baseline at 60 minutes of perfusion. c Significance: p 0.05, intragroup decrease from baseline at 60 minutes of perfusion. d Significance: p 0.01, intragroup decrease from baseline at 60 minutes of perfusion. e Significance: p 0.05, intragroup increase from baseline at 60 minutes of perfusion. ANOVA analysis of variance; Baseline (100) end of cooling perfusion (100 ml kg 1 min 1, 20 C), data from all animals shown.

5 868 WATANABE ET AL Ann Thorac Surg BRAIN BLOOD FLOW AND METABOLISM AT 20 C 1999;68:864 9 Table 3. Brain Regional Intracellular ph After 120 Minutes of Perfusion at 20 C a Area Perfusion Flow Rate (ml kg 1 min 1 ) p Value (ANOVA) Anterior cortex b b c d Middle cortex b b b b Posterior cortex b b d b d Thalamus b b c b c Caudate nucleus b b b d Hippocampus b b b d Pons b c Cerebellum b b b d a Data are shown as the mean the standard deviation. b Significance: p 0.002, unpaired t test, versus 40 ml kg 1 min 1. c Significance: p 0.01, unpaired t test, versus 40 ml kg 1 min 1. d Significance: p 0.05, unpaired t test, versus 40 ml kg 1 min 1. ANOVA analysis of variance. would be as detrimental as circulatory arrest even at deep hypothermia because it may increase the metabolic insult with a persistent glucose supply [5, 7, 19]. As a result, the threshold flow range for protecting the brain has long been studied. At normothermia, the threshold cerebral blood flow to maintain normal brain function is estimated to be approximately 30 ml 100 g 1 min 1 at 37 C, below which brain acidosis occurs. The brain loses electrical activity at a cerebral blood flow of less than 20 ml 100 g 1 min 1 and loses further cellular membrane integrity at flows lower than 10 ml 100 g 1 min 1, known as the threshold for infarction [9]. Miyamoto and associates [4] showed that global cerebral blood flow of less than 7 ml 100 g 1 min 1 during deep hypothermia at 20 C resulted in reduction of cerebral oxygen metabolism. Our data demonstrate that cerebral cortex blood flow of 9 ml 100 g 1 min 1 is necessary to maintain aerobic metabolism at 20 C. Both sets of data indicate the same range. Therefore, deep hypothermic metabolic reduction in the brain at 20 C may reduce the cerebral requirement for blood flow to approximately 25% to 30% of normal. The brain, on the other hand, apparently protected itself from luxury pressure and blood flow by elevating cerebral vascular resistance (see Fig 3). This allowed shunting of excess blood through organs less susceptible to ischemia than the brain. In 1984, Fox and associates [3] reported a loss of autoregulation during deep hypothermia, where cerebral vascular resistance did not change over a wide range of perfusion flow rates. However, their data were fundamentally affected by their extreme hemodilution (hematocrit value of less than 10%). Our data, in contrast, reveal that autoregulation of cerebral blood flow is maintained even at 20 C as long as hemodilution is in the clinical range. When full-flow perfusion (100 ml kg 1 min 1 ) was continued without consideration of hypothermic metabolic reduction, the phi in the middle and posterior cerebral cortex and the thalamus and the metabolic rate for oxygen were lower than those seen with 40 ml kg 1 min 1 perfusion. The ratio of the glucose to oxygen metabolic rates also increased during this luxury perfusion, as Fig 5. Cerebral cortex intracellular ph (phi) versus averaged perfusion pressure (APP) during 120 minutes of deep hypothermic perfusion. The phi was logarithmically correlated with perfusion pressures of less than 15 mm Hg: cerebral cortex phi 0.40 ln (mm Hg) 6.1 (r 0.71). Fig 6. Cerebral cortex intracellular ph (phi) after perfusion versus cerebral cortex blood flow (CCBF) at 60 minutes of deep hypothermic perfusion. There are two plots, as there was a breakdown threshold. Correlation lines with cerebral cortex blood flow are represented as cerebral cortex phi ml 100 g 1 min (r 0.59) and ml 100 g 1 min (r 0.04) when the cortex blood flow was less than and greater than 9.0 ml 100 g 1 min 1, respectively.

6 Ann Thorac Surg WATANABE ET AL 1999;68:864 9 BRAIN BLOOD FLOW AND METABOLISM AT 20 C 869 References Fig 7. Cerebral metabolic ratio of glucose to oxygen at 60 minutes of deep hypothermic perfusion. Cerebral metabolic rate for glucose was divided by that for oxygen. High values at perfusion pressures of less than 10 mm Hg and greater than 40 mm Hg represent anaerobic metabolism in the brain. it did during very low flow perfusion (see Fig 7). These findings agree with each other and suggest the paradoxical harm of luxury perfusion. Full-flow perfusion, when continued against downregulation of cerebral blood flow during deep hypothermia, could cause cerebral vasoconstriction that would harm the microcirculation. As shown in our study, excessive perfusion through the arch arteries during surgical procedures on the aorta can cause severe vasoconstriction with active autoregulation. This can not only make the brain paradoxically acidotic but also increase the risk of brain embolism, as blood flow at extreme velocity can easily carry the atheromatous debris in the sclerotic and constricted arteries to the brain. Our observation at 20 C demonstrates that there are both upper and lower limits in perfusion pressure ranges for brain protection during deep hypothermia; 10 to 40 mm Hg may be this protective range. It resembles the normothermic condition when both hypotension and hypertension can harm the brain. The data also demonstrate that there is an optimal range for perfusion flow rate and pressure. Perfusion at a flow rate of 40 ml kg 1 min 1 and a pressure of approximately 20 mm Hg provided low cerebral vascular resistance (brain vasorelaxation) with sufficient blood flow, the lowest cerebral metabolic ratio of glucose to oxygen, and the highest brain phi. We think the brain can be protected from ischemic acidosis when it is cooled to 20 C during aortic operations, even with a lower perfusion pressure than currently recommended (higher than 40 mm Hg) in the studies carried out at 25 C [16, 20]. In summary, during deep hypothermia, perfusion at relatively low pressures and flow rates may provide optimal brain protection with full microcirculatory function. The data emphasize the importance of brain vasorelaxation in maintaining aerobic metabolism. We thank Dr Julie A. Swain and Dr Kenneth M. Taylor for critically reviewing the manuscript and advising us on the study design. 1. Greeley WJ, Kern FH, Ungerleider RM, et al. The effect of hypothermic cardiopulmonary bypass and total circulatory arrest on cerebral metabolism in neonates, infants, and children. J Thorac Cardiovasc Surg 1991;101: Croughwell N, Smith LR, Quill T, et al. The effect of temperature on cerebral metabolism and blood flow in adults during cardiopulmonary bypass. J Thorac Cardiovasc Surg 1992;103: Fox LS, Blackstone EH, Kirklin JW, Bishop SP, Bergdahl LA, Bradley EL. Relationship of brain blood flow and oxygen consumption to perfusion flow rate during profoundly hypothermic cardiopulmonary bypass. J Thorac Cardiovasc Surg 1984;87: Miyamoto K, Kawashima Y, Matsuda H, Okuda A, Maeda S, Hirose H. Optimal perfusion flow rate for the brain during deep hypothermic cardiopulmonary bypass at 20 C. J Thorac Cardiovasc Surg 1986;92: Watanabe T, Washio M. Cerebral cellular response to profound hypothermia. Cardiol Young 1993;3: Swain JA, McDonald TJ Jr, Griffith PK, Balaban RS, Clark RE, Ceckler T. Low-flow hypothermic cardiopulmonary bypass protects the brain. J Thorac Cardiovasc Surg 1991;102: Scheller MS, Branson PJ, Cornacchia LG, Alksne JF. A comparison of the effects on neuronal Golgi morphology, assessed with electron microscopy, of cardiopulmonary bypass, low-flow bypass, and circulatory arrest during profound hypothermia. J Thorac Cardiovasc Surg 1992;104: Rossi R, van der Linden J, Ekroth R, Scallan M, Thompson RJ, Lincoln C. No flow or low flow? A study of the ischemic marker creatine kinase BB after deep hypothermic procedures. J Thorac Cardiovasc Surg 1989;98: Paczinsky R, Hsu CY, Diringer MN. Pathophysiology of ischemic injury. In: Fisher M, ed. Stroke therapy. Boston: Butterworth-Heinemann, 1995: Saitou H, Watanabe T, Zhang JW, et al. Regional cerebral blood flow during retrograde cerebral perfusion. J Surg Res 1997;72: Griffith JK, Cordisco BR, Lin CW, LaManna JC. Distribution of intracellular ph in the rat brain cortex after global ischemia as measured by color film histophotometry of neutral red. Brain Res 1992;573: Watanabe T, Iijima Y, Abe K, et al. Retrograde brain perfusion beyond the venous valves: hemodynamics and intracellular ph mapping. J Thorac Cardiovasc Surg 1996;111: Schwartz AE, Sandhu AA, Kaplon RJ, et al. Cerebral blood flow is determined by arterial pressure and not cardiopulmonary bypass flow rate. Ann Thorac Surg 1995;60: Iwabuchi T, Kutsuzawa T, Ikeda K, Nakamura T. Effects of blood gases on the pressure-flow relationships in canine cerebral circulation. Stroke 1973;4: Murkin JM, Farrar JK, Tweed WA, McKenzie FN, Guiraudon G. Cerebral autoregulation and flow/metabolism coupling during cardiopulmonary bypass. The influence of PaCO 2. Anesth Analg 1987;66: Sadahiro M, Haneda K, Mohri H. Experimental study of cerebral autoregulation during cardiopulmonary bypass with or without pulsatile perfusion. J Thorac Cardiovasc Surg 1994;108: Taylor KM. Central nervous system effects of cardiopulmonary bypass. Ann Thorac Surg 1998;66:S Rehncrona S. Brain acidosis. Ann Emerg Med 1985;14: Anderson RV, Siegman MG, Balaban RS, Ceckler TL, Swain JA. Hyperglycemia increases cerebral intracellular acidosis during circulatory arrest. Ann Thorac Surg 1992;54: Tanaka H, Kazui T, Sato H, Inoue N, Yamada O, Komatsu S. Experimental study on the optimum flow rate and pressure for selective cerebral perfusion. Ann Thorac Surg 1995;59: