Fulminant hepatic failure (FHF) is a critical illness. Cerebral Glucose and Oxygen Metabolism in Patients With Fulminant Hepatic Failure

Transcription

1 RAPID COMMUNICATION Cerebral Glucose and Oxygen Metabolism in Patients With Fulminant Hepatic Failure Gitte Irene Strauss, * Kirsten Møller, Fin Stolze Larsen, * Jens Kondrup, * and Gitte Moos Knudsen Hyperammonemia and hyperventilation are consistent findings in patients with fulminant hepatic failure (FHF), which may interfere with cerebral glucose and oxygen metabolism. The aim of the present study is to evaluate whether cerebral oxidative metabolism is preserved early in the course of FHF and whether hyperventilation has an influence on this. We included 16 patients with FHF, 5 patients with cirrhosis of the liver, and 8 healthy subjects. Concomitant blood sampling from an arterial catheter and a catheter in the jugular bulb and measurement of cerebral blood flow by the xenon 133 wash-out technique allowed calculation of cerebral uptake of glucose (CMR gluc ) and oxygen (CMRO 2 ). Both CMR gluc and CMRO 2 were reduced in patients with FHF compared with those with cirrhosis and healthy subjects, i.e., v and mol/100 g/min (P <.05) and v and mol/100 g/min (P <.05). Arteriovenous difference in oxygen and oxygen-glucose index were normal in patients with FHF. Institution of mechanical hyperventilation did not affect glucose and oxygen uptake and hyperventilation did not affect lactatepyruvate ratio or lactate-oxygen index. In conclusion, we found that cerebral glucose and oxygen consumption are proportionally decreased in patients with FHF investigated before clinical signs of cerebral edema. Our data suggest that cerebral oxidative metabolism is retained at this stage of the disease without being compromised by hyperventilation. (Liver Transpl 2003;9: ) From the Departments of *Hepatology, Infectious Diseases, and Neurobiology, Rigshospitalet, University of Copenhagen, Denmark. Supported in part by The Laerdal Foundation for Acute Medicine; Savvaerksejer Jeppe Juhl and Wife Ovita Juhls Foundation; Rigshospitalet; University of Copenhagen; The Novo Nordisk Foundation; The King Christian the 10th Foundation; The AP-Møller Foundation; The Danish Hospital Foundation for Medical Research, Region of Copenhagen; The Faroe Islands and Greenland; The Health Research Council; and The Beckett Foundation. Address reprint requests to Gitte Irene Strauss, MD, Department of Hepatology A-2121, Rigshospitalet, University of Copenhagen, Blegdamsvej 9, DK-2100 Copenhagen, Denmark. Telephone: ; FAX: ; gstrauss@dadlnet.dk Copyright 2003 by the American Association for the Study of Liver Diseases /03/ $30.00/0 doi: /j.lts Fulminant hepatic failure (FHF) is a critical illness with multiorgan failure and hepatic encephalopathy (HE). Cerebral edema and intracranial hypertension are feared and often fatal complications. The pathophysiological background for these complications is not fully understood, but alterations in cerebral metabolism and cerebral blood flow (CBF) seem to be of importance. 1 Metabolism of the brain is different from most other tissues in several aspects. It uses glucose only as an energy source under normal physiological conditions, and in contrast to most other organs, brain cells are permeable to glucose and can use glucose without the intermediation of insulin. 2 However, when necessary, the brain can use other substrates as energy sources, such as -hydroxybutyrate ( -OHB) and acetoacetate. 3 Hyperammonemia is a consistent finding in FHF, and arterial ammonia concentration seems to correlate with the development of cerebral edema and intracranial hypertension in patients with FHF. 4 The exact biochemical mechanisms involved in cerebral ammonia toxicity are not fully established, but ammonia-induced alterations in cerebral glucose metabolism may be of importance. 5 In experimental studies of hyperammonemia, both reduced 6,7 and increased 8 cerebral glucose metabolism have been reported. Also, studies of patients with FHF reported both reduced, 9,10 normal, 10 and increased 11 cerebral glucose metabolism, whereas cerebral oxygen metabolism was reduced in all studies The aim of the present study is to evaluate cerebral oxidative substrate consumption in patients with FHF within the first 24 hours after the development of HE (i.e., well before the development of intracranial hypertension) 13 to evaluate whether changes in cerebral oxidative metabolism are an early event in patients with FHF. For comparison, a group of patients with cirrhosis of the liver without HE and a group of healthy subjects also were investigated. Furthermore, because hyperventilation is observed in many spontaneously ventilating patients with FHF and may be used in the treatment of intracranial hypertension, we evaluated the effect of acute mechanical hyperventilation on cerebral oxidative metabolism Liver Transplantation, Vol 9, No 12 (December), 2003: pp

2 Cerebral Metabolism in FHF 1245 Table 1. Clinical and Laboratory Data for 8 Healthy Subjects, 5 Patients With Cirrhosis, and 16 Patients With Fulminant Hepatic Failure HE Mean Arterial Pressure (mm Hg) Ammonia ( mol/l, reference interval 11-32) Coagulation Factors (U/L) II, VII, X 70% Creatinine 130 mol/l L-P Ratio Hematocrit (Volume fraction) Control Cirrhosis * FHF 4 (3-4) 78 11* * * * * NOTE. Data presented as mean SD, except HE, which is presented as median (range). Abbreviations: HE, hepatic encephalopathy; L-P, lactate-pyruvate. *P 0.05 versus control (ANOVA). Methods The study included three groups (Table 1), i.e., 16 patients with FHF, 5 patients with cirrhosis of the liver without HE, and 8 healthy subjects. The regional Scientific Ethical Committee (KF /98) approved the study, and informed consent was obtained from patients and healthy subjects or the patient s next of kin. Cirrhotic patients and healthy subjects were studied only during spontaneous ventilation, whereas patients with FHF were studied during mechanical normoventilation and hyperventilation. FHF Causes of FHF were acetaminophen intoxication (n 12), unknown (n 3), and acute hepatitis B (n 1). All patients were in stage 3 to 4 HE and on mechanical ventilation (Servo 900 C; Siemens, Solna, Sweden). The study was conducted an average of 10 hours (range, 4 to 24 hours) after the institution of mechanical ventilation. Fourteen patients were sedated with midazolam (6 mg/h; range, 2 to 10 mg/h). All patients were administered N-acetylcysteine (6.25 mg/kg/h) and 20% glucose (85 ml/h) intravenously and lactulose, 15 ml, three times daily through a gastric tube. All medications remained unchanged before and during the entire study period. No patient had clinical signs of intracranial hypertension before or during the study (pronation seizures, sluggish pupillary reaction to light, pupillary asymmetries, or systemic hypertension). Eleven of 16 patients fulfilled the King s College criteria for liver transplantation (Table 1). Patients With Liver Cirrhosis and Healthy Subjects Patients with cirrhosis of the liver were recruited from the outpatient clinic. Cause of cirrhosis was alcoholic in all patients (n 5), and they were all classified as Child s class A. No patient was treated with ammonia-lowering therapy, and they all had abstained from alcohol for at least 6 months. Healthy subjects underwent a physical examination, electrocardiogram, and blood test screening before inclusion in the study. All subjects were investigated in a supine position after an overnight fast, and glucose 5% (17 g/h) was administered intravenously for approximately 1.5 to 2 hours before the infusion of xenon 133 was started. Catheters Catheters were placed in the radial artery and retrogradely in the internal jugular vein for monitoring of mean arterial pressure and blood sampling, respectively. The cubital vein on the opposite arm of the arterial catheter also was cannulated and used for xenon 133 infusion. CBF Data for CBF have been published previously. 14 In the present study, only data for cerebral oxidative substrate metabolism are reported. A detailed description of the two methods has been published previously. 15,16 Substrate Analysis Blood samples were drawn simultaneously from the radial artery and internal jugular bulb. Samples for blood gas analysis were drawn into heparinized tubes, immediately placed on ice, and analyzed (ABL 625; Radiometer, Copenhagen, Denmark) within 5 to 10 minutes. Samples for glucose and lactate analysis were drawn into tubes containing dried fluoride oxalate, placed on ice, and centrifuged within 5 to 10 minutes at 3,000 rpm for 15 minutes at 4 C. Plasma was stored at 20 C until analysis using a YSI 2700 Select Biochemistry Analyzer (Yellow Springs Instruments Co Inc, Yellow Springs, OH). Analysis was performed in duplicate, and the mean was used for calculations. Because analysis of glucose and lactate was performed on plasma, these values subsequently were converted to whole-blood values. 17 Whole-blood samples for determination of acetoacetate, -OHB, and pyruvate were immediately deproteinized with perchloric acid (1 mol/l) and centrifuged at 3,000 rpm for 15 minutes at 4 C within 5 to 10 minutes of withdrawal. According to Williamson et al, 18 3 ml of supernatant was neutralized with potassium phosphate (3 mol/l), centrifuged, and stored

3 1246 Strauss et al at 80 C until analysis for acetoacetate and -OHB. The remaining supernatant was frozen at 80 C until analysis of pyruvate. A double-beam spectrophotometer (Cary 1E; Varian Australia Pty Ltd., Mulgrave, Australia) was used for analysis of all three substrates. Amino Acids and Ammonia Data for arterial concentrations and metabolic rates of the individual amino acids have been published previously. 14 In the present report, only correlations with arterial ammonia level are included. Calculations Cerebral arteriovenous difference in oxygen (avdo 2 ) was calculated by using hemoglobin concentration and oxygen saturation in arterial and internal jugular venous blood according to the following formula: avdo 2 hb(mmol / L) (SaO 2 SvO 2 ) 0.01 (PaO 2 PvO 2 ) where SaO 2 is arterial oxygen saturation, SvO 2 is venous oxygen saturation, 0.01 is the solubility constant of oxygen in plasma in mmol/l/kpa, and PaO 2 and PvO 2 are arterial and venous partial pressure of oxygen in kpa, respectively. The cerebral metabolic rate (CMR) of a given substrate x was calculated according to Fick s principle, as follows: CMR x CBF avd x where avd x is arteriovenous difference of the substrate. Hence, avdo 2 expresses the relationship between cerebral oxygen metabolism and CBF. In case of luxury perfusion, i.e., CBF greater than metabolic needs, avdo 2 will decrease. The oxygen-glucose index (OGI) was calculated as follows: OGI avdo 2 / avd gluc where avd gluc is arteriovenous difference of glucose. During aerobic conditions, approximately six molecules of oxygen are used to oxidate one molecule of glucose. If an excess of glucose is metabolized relative to oxygen, the OGI will be less than 6. Hence, an OGI less than 6 indicates that some glucose is being metabolized anaerobically. Conversely, an OGI exceeding 6 indicates that substrates other than glucose are being metabolized aerobically. To take other substrates (carbohydrates) into account, the oxygen-carbohydrate index was calculated as follows: Oxygen-carbohydrate index avdo 2 / (avd gluc [(avd lac avd pyr ) / 2]) where avd lac is the arteriovenous difference of lactate and avd pyr is the arteriovenous difference of pyruvate. The lactate-glucose index (LGI) was calculated as follows: LGI avd lac / avd gluc The LGI is the fraction of cerebral glucose uptake that is metabolized and excreted from brain as lactate. The denominator is always positive; thus, if lactate is excreted, the numerator will be negative, and the resulting LGI will be negative, i.e., a negative LGI indicates that some glucose is being converted to lactate anaerobically. Conversely, a positive LGI indicates cerebral lactate uptake. Because one molecule of glucose is converted to two molecules of lactate, an LGI of, e.g., 0.1 indicates that 5% of cerebral glucose uptake (CMR gluc ) may be accounted for by lactate excretion. The lactate-oxygen index (LOI) was calculated as follows: LOI avd lac / avdo 2 Because the denominator is always positive, a positive LOI indicates cerebral lactate uptake, whereas a negative LOI indicates cerebral lactate efflux. The LOI gives an approximate indication of the magnitude of cerebral anaerobic metabolism relative to oxidative metabolism. Normally, a slight cerebral lactate efflux is present. 3 In the case of severe cerebral hypoxia, the LOI may be expected to decrease, i.e., become more negative, as the oxygen supply becomes insufficient for aerobic metabolism. The lactate-pyruvate (L-P) ratio was calculated as L-P Ca lac /Ca pyr Where Ca lac is arterial concentration of lactate and Ca pyr is arterial concentration of pyruvate. When oxygen supply becomes insufficient for aerobic metabolism, pyruvate is converted to lactate because the reduced form of nicotinamide adenine dinucleotide cannot be oxidized to the oxidized form of nicotinamide adenine dinucleotide in mitochondria by the electron transport chain. Hence, in the case of hypoxia, the L-P ratio will increase. Statistics Results are presented as mean SD. Normality was tested by the Kolmogorov-Smirnov test. Kruskal-Wallis analysis of variance (ANOVA) on the ranks with Dunn s method for multiple comparisons was used for comparison between more than two groups. Mann-Whitney U test was used for comparison of unpaired data within groups, and Wilcoxon s test, for paired data. Bonferroni correction factor 5 was applied to P to adjust for multiple comparisons between two unpaired groups and paired data. Correlation was tested by Spearman s rankorder test. P less than.05 is considered significant. Results Eight of 16 patients with FHF recovered spontaneously (survivors), 2 patients underwent emergency liver transplantation, and 6 patients died of cerebral herniation (nonsurvivors) days after the study. The two patients who underwent transplantation are not

4 Cerebral Metabolism in FHF 1247 Figure 1. Box plot of arterial concentrations of (A) oxygen, (B) glucose, and (C) lactate (mmol/l) and (D) pyruvate, (E) acetoacetate, and (F) -OHB ( mol/l) in 15 patients with FHF, 8 healthy subjects (controls), and 5 patients with cirrhosis of the liver. Ends of the boxes define the 25th and 75th percentiles, with a line at the median and error bars defining the 10th and 90th percentiles. included in comparisons between survivors and nonsurvivors. In one patient, only data for cerebral oxygen uptake (CMRO 2 ) during normoventilation were obtained. Arterial concentrations of glucose, lactate, pyruvate, and -OHB were significantly greater, whereas arterial concentrations of oxygen were significantly lower, in patients with FHF compared with controls (Fig. 1; ANOVA). CMR gluc and CMRO 2 were significantly reduced in patients with FHF compared with CMR gluc and CMRO 2 in healthy subjects and patients with cirrhosis, i.e., CMR gluc was v and mol/(100 g min) (P.05; Fig. 2) and CMRO 2 was v and mol/(100 g min) (P.05; Fig. 3), respectively. In addition to glucose, net cerebral lactate uptake of mol/(100 g min) was found in patients with FHF, whereas there was no net cerebral lactate uptake in Figure 2. Cerebral metabolic ratios of glucose (white bars) and lactate (black bars) in healthy subjects (n 8), patients with cirrhosis (n 5), and patients with FHF (n 15). Values expressed as mean SD. * P <.05 versus controls.

5 1248 Strauss et al Figure 3. Substrate covered (white bars) versus measured (black bars) CMRO 2 in healthy subjects (n 8), patients with cirrhosis (n 5), and patients with FHF (n 15). Values expressed as mean SD. * P <.05 versus substrate covered (paired Wilcoxon s). healthy subjects and patients with cirrhosis, i.e., and , respectively (Fig. 2). There were no net cerebral fluxes of pyruvate, acetoacetate, or -OHB (data not shown). avdo 2 was similar for all groups, i.e., mol/ml in patients with FHF, mol/ml in patients with cirrhosis, and mol/ml in healthy subjects (P not significant [NS]; Table 2), whereas avd gluc was significantly lower in patients with FHF and patients with cirrhosis than healthy subjects, i.e., and v mol/ml (P.05; Table 2). Indices of cerebral oxidative metabolism are listed in Table 2. OGI was similar in healthy subjects and patients with FHF, whereas it was significantly increased in patients with cirrhosis (Table 2). When other carbohydrates were taken into account, i.e., lactate and pyruvate, the oxygen-carbohydrate index decreased only in patients with FHF because of net cerebral uptake of lactate exclusively in these patients, which also was shown by the positive LGI and LOI only in patients with FHF (Table 2). Substrate-covered oxygen consumption compared with measured cerebral oxygen consumption are shown in Figure 3. In patients with FHF and healthy subjects, substrate-covered CMRO 2 equaled measured CMRO 2, i.e., v mol/(100 g min) (P NS) and v mol/(100 g min) (P NS), whereas in patients with cirrhosis, substrate-covered CMRO 2 was significantly lower than measured CMRO 2, i.e., v mol/(100 g min) (P.05; Fig. 3). The cerebral oxygen to glucose ratio was significantly greater in patients with cirrhosis than healthy subjects and patients with FHF, i.e., v and , respectively (P.05, ANOVA). Arterial ammonia levels did not correlate with CMRO 2 (P.15), CMR gluc (P.35), or CBF (P.68) in patients with FHF (data not shown). There was no difference in any reported values between survivors and nonsurvivors. Mechanical Hyperventilation During mechanical hyperventilation, PaCO 2 was kpa, and jugular venous saturation decreased to CBF decreased in all patients with FHF during intervention with mechanical hyperventilation. 14 Average cerebral CO 2 reactivity was %/mm Hg in patients with FHF. Intervention with mechanical hyperventilation increased arterial concentrations of lactate and acetoacetate (P.05) and decreased arterial oxygen concentrations (P.05), Table 2. AvDO 2 and avd gluc and Indices of Cerebral Oxidative Metabolism in Healthy Subjects, Patients With Cirrhosis, and During Rest and Mechanical Hyperventilation in Patients With FHF avdo 2 ( mol/ml) avd gluc ( mol/ml) OGI LGI LOI Oxygen-Carbohydrate Control (n 8) Cirrhosis (n 5) * * * FHF (n 15) Rest * * * Hyperventilation NOTE. Data presented as mean SD. Abbreviations: avdo 2, arteriovenous difference in oxygen; avd gluc, arteriovenous difference in substrate; OGI, oxygen-glucose index; LGI, lactate-glucose index; LOI, lactate-oxygen index; FHF, fulminant hepatic failure. *P 0.05 versus control (ANOVA). P 0.05 versus OGI (paired Wilcoxon s). P 0.05 versus Rest (paired Wilcoxon s).

6 Cerebral Metabolism in FHF 1249 Figure 4. Arterial concentrations of (A) oxygen, (B) glucose, (C) lactate, (D) pyruvate, (E) -OHB, and (F) acetoacetate in patients with FHF before (rest) and during hyperventilation. whereas arterial concentrations of glucose, pyruvate, and -OHB remained unchanged (Fig. 4). Cerebral fluxes of glucose (CMR gluc ), oxygen (CMRO 2 ), and -OHB remained unchanged during hyperventilation, whereas cerebral fluxes of lactate, pyruvate, and acetoacetate changed significantly during hyperventilation (Fig. 5). L-P ratio was similar during normoventilation and hyperventilation, i.e., v (P NS). avdo 2 increased to mol/ml (P.001), and avd gluc increased to mol/ml (P.05) during hyperventilation (Table 2). LGI and LOI decreased significantly, but did not become negative, whereas OGI and oxygen-carbohydrate index remained unchanged from normoventilation to hyperventilation (Table 2). Discussion This study is the first exhaustive account of cerebral oxidative substrate consumption in patients with FHF. In summary, we found that CMR gluc and CMRO 2 decreased in parallel because the ratio between CMRO 2 and CMR gluc was normal, and the normal relationship between CBF and CMRO 2, expressed as avdo 2, was retained in patients with FHF investigated before the development of intracranial hypertension. There was striking net cerebral lactate uptake despite ample glucose supply (i.e., high cerebral glucose delivery) in patients with FHF. Finally, we found that short-term intervention with mechanical hyperventilation slightly increased CMRO 2 and decreased lactate uptake in patients with FHF. Thus, our findings reflect normal, but reduced, cerebral glucose and oxygen metabolism during the first 24 hours of FHF, i.e., well before increases in intracranial pressure. In the present study, midazolam was administered as sedation. In dogs, infusion of midazolam (40 mg/h) resulted in a 75% decrease in CBF and CMRO 2 within 10 to 15 minutes after the infusion was started. Thereafter, additional midazolam had no additional effect on CBF and CMRO In patients with FHF, it is possible that midazolam reduced CBF and cerebral oxidative metabolism to some extent, but we found no direct relationship between midazolam infusion rate and levels of CBF, CMRO 2, and CMR gluc.wefind it more likely that level of unconsciousness accounts for the reduced cerebral oxidative metabolism and CBF in our patients. This is supported to some extent by a previous study of healthy subjects that showed a 25% reduction in CMRO 2 and CBF during deep sleep. 20 The finding of a normal OGI and oxygen-carbohydrate index in patients with FHF indicates that all glucose taken up by the brain is aerobically metabolized. Furthermore, this study shows that although cerebral oxidative metabolism is significantly reduced, it is sufficient for metabolic demands at this stage of FHF because there were no indications of cerebral hypoxia,

7 1250 Strauss et al Figure 5. Cerebral metabolic rates of (A) oxygen, (B) lactate, (C) acetoacetate, (D) glucose, (E) pyruvate, and (F) betahydroxybuturate in patients with FHF before (rest) and after hyperventilation. Values expressed as mean SD. * P <.05 versus normoventilation. i.e., internal jugular bulb saturation was greater than 50% in all cases. The reduced CMRO 2 found in the present study is in accordance with previous studies of patients with FHF 9,10,12 and experimental studies in animal models of HE 7 and hyperammonemia. 21 However, in previous studies, reported values for cerebral glucose metabolism ranged from reduced 9,10 to normal 10 to increased CMR gluc in patients with FHF. 11 This discrepancy may be caused by time differences because the latter study was performed during ongoing intracranial hypertension, i.e., a later stage in the neurological course of FHF. 11 It is not possible from the present study to make firm conclusions about cerebral glucose and oxygen metabolism later during the course of FHF. However, Tofteng et al 13 found that patients with FHF and severe hyperammonemia also has accelerated glycolysis, 22 a finding that supports experimental studies indicating that ammonia compromises mitochondrial function in astrocytes. 23,24 In several reports, we have stressed that cerebral hyperemia is of importance for aggravation of cerebral edema. 25 However, CBF first is low, but then tends to increase as hyperammonemia decreases cerebral arteriolar tone. 14 Mean values for CBF previously reported in patients with FHF ranged from 30 to 42 ml/(100 g min). 9,10,12,26 Only one study reported much higher CBF, but these values were obtained during ongoing intracranial hypertension. 11 Accordingly, high CBF is not an initial phenomenon in patients with FHF, but seems to gradually increase as cerebral ammonia metabolism increases. Although CBF is low, it may still be too high for metabolic demands, i.e., relative cerebral hyperemeia. 27 In the present study, we found no signs of cerebral luxury perfusion within the first 24 hours after the development of stage 3 to 4 HE because avdo 2 in patients with FHF was normal. Although avd gluc in patients with FHF was lower than in healthy subjects, this can be explained by the switch toward lactate oxidation because adjustment of avd gluc to take into regard cerebral lactate uptake, i.e., the ratio between cerebral carbohydrate (glucose lactate) metabolism and CBF, results in normalization of this ratio. These results indicate that cerebral perfusion tightly matches cerebral oxidative metabolism in patients with FHF within the first 24 hours after the appearance of stage 3 to 4 HE.

8 Cerebral Metabolism in FHF 1251 Interestingly, we found net cerebral lactate uptake exclusively in patients with FHF despite an ample glucose supply. Only the ionized fraction of lactate is transported across the blood-brain barrier by the monocarboxylate carrier transport system, whereas the unionized fraction is assumed to cross the blood-brain barrier by passive diffusion. In the physiological ph range, as in the present study, only a very small fraction of lactate is unionized (pk a for lactate is 3.83), indicating that diffusion of free lactate acid is unlikely to provide an explanation for cerebral lactate uptake. Cerebral lactate uptake despite ample glucose supply has been reported in other conditions with hyperlactemia, such as exercise 28 and after cardiac arrest. 29 Likewise, a previous study of patients with FHF and ongoing intracranial hypertension also reported cerebral lactate uptake. 11 This phenomenon could be the passive consequence of increased arterial lactate content because it seems to be a general trait that brain metabolism shifts toward oxidation of available substrates, e.g., during hyperketonemia, the brain shifts toward oxidation of -OHB. 30 Alternatively, it could be the result of neuronal activation because lactate seems to serve as the primary energy substrate for neurons. 31 From the present study, we cannot make a conclusion about the fate of lactate taken up by the brain. Carbon dioxide is a powerful regulator of CBF. Hypocapnia induces cerebral vasoconstriction, resulting in decreases in CBF and cerebral blood volume. Because of its pronounced and immediate effect on CBF, intervention with mechanical hyperventilation often has been used as rescue therapy in conditions with increased intracranial pressure. One of the major concerns has been that hypocapnia may reduce CBF to critically low levels. Wendon et al 10 found that mechanical hyperventilation increased cerebral efflux of lactate slightly in patients with FHF. However, the measured cerebral lactate efflux reported by Wendon et al 10 during both normoventilation and hyperventilation is of the same magnitude as that observed in healthy subjects during normoventilation and after voluntary hyperventilation. 32 During mechanical hyperventilation, we found that cerebral lactate uptake decreased despite a slight increase in arterial lactate concentration. This could be caused by inhibition of cerebral lactate uptake mediated by the increase in ph during hyperventilation. 33 However, increased outward flux compared with inward flux of lactate at the monocarboxylate carrier transport system cannot be excluded. In the present study, LOI decreased, but remained positive, indicating aerobic cerebral metabolism during short-term hyperventilation. Thus, there was no indication of compromised cerebral oxidative metabolism during short-term moderate mechanical hyperventilation in the present study, suggesting that an acute reduction of approximately 1.3 kpa can be regarded as safe. Although patients with cirrhosis of the liver primarily functioned as a control group, some results in this group need to be discussed further separately. To date, no study has measured global cerebral oxidative metabolism in patients with cirrhosis of the liver without HE. In the present study, CBF and cerebral oxidative metabolism in patients with cirrhosis were not significantly different from those of healthy subjects; however, measured cerebral oxygen consumption could not be accounted for by substrate-covered oxygen consumption (Fig. 3). This indicates that other substrates must be oxidized in addition to glucose in the brain of patients with cirrhosis. Because there was no net cerebral consumption of any measured substrate except glucose, this extra oxygen could be used for oxidation of fatty acid or, alternatively, oxidation of carbon skeletons arising from degradation of cerebral proteins. 14 Acknowledgment The authors thank Gerda Thomsen, Karin Stahr, Mie Poulsen, Nine Scherling, and Liselotte Hansen for skillful technical assistance. References 1. Blei AT, Larsen FS. Pathophysiology of cerebral edema in fulminant hepatic failure. J Hepatol 1999;31: Knudsen GM, Hasselbalch SG, Hertz MM, Paulson OB. High dose insulin does not increase glucose transfer across the bloodbrain barrier in humans: A re-evaluation. Eur J Clin Invest 1999; 29: Hasselbalch SG, Knudsen GM, Jakobsen J, Hageman LP, Holm S, Paulson OB. Brain metabolism during short-term starvation in humans. J Cereb Blood Flow Metab 1994;14: Clemmesen JO, Larsen FS, Kondrup J, Hansen BA, Ott P. Cerebral herniation in patients with acute liver failure is correlated with arterial ammonia concentration. Hepatology 1999; 29: Hindfelt B, Siesjo BK. Cerebral effects of acute ammonia intoxication. II. The effect upon energy metabolism. Scand J Clin Lab Invest 1971;28: Gjedde A, Lockwood A, Duffy TE, Plum F. Effect of ammonia on cerebral metabolism of rats with portocaval shunts. Trans Am Neurol Assoc 1976;101: Jessy J, DeJoseph MR, Hawkins RA. Hyperammonaemia depresses glucose consumption throughout the brain. Biochem J 1991;277: Cruz NF, Duffy TE. Local cerebral glucose metabolism in rats with chronic portacaval shunts. J Cereb Blood Flow Metab 1983;3: Larsen FS, Ejlersen E, Clemmesen JO, Kirkegaard P, Hansen

9 1252 Strauss et al BA. Preservation of cerebral oxidative metabolism in fulminant hepatic failure: An autoregulation study. Liver Transpl Surg 1996;2: Wendon JA, Harrison PM, Keays R, Williams R. Cerebral blood flow and metabolism in fulminant liver failure. Hepatology 1994;19: Jalan R, Damink SW, Deutz NE, Lee A, Hayes PC. Moderate hypothermia for uncontrolled intracranial hypertension in acute liver failure. Lancet 1999;354: Aggarwal S, Kramer D, Yonas H, Obrist W, Kang Y, Martin M, Policare R. Cerebral hemodynamic and metabolic changes in fulminant hepatic failure: A retrospective study. Hepatology 1994;19: Tofteng F, Jorgensen L, Hansen BA, Ott P, Kondrup J, Larsen FS. Cerebral microdialysis in patients with fulminant hepatic failure. Hepatology 2002;36: Strauss GI, Knudsen GM, Kondrup J, Moller K, Larsen FS. Cerebral metabolism of ammonia and amino acids in patients with fulminant hepatic failure. Gastroenterology 2001;121: Madsen PL, Holm S, Herning M, Lassen NA. Average blood flow and oxygen uptake in the human brain during resting wakefulness: A critical appraisal of the Kety-Schmidt technique. J Cereb Blood Flow Metab 1993;13: Strauss GI, Moller K, Holm S, Sperling BK, Knudsen GM, Larsen FS. Transcranial Doppler sonography and internal jugular bulb saturation during hyperventilation in patients with fulminant hepatic failure. Liver Transpl 2001;7: Simonsen L, Bulow J, Madsen J. Adipose tissue metabolism in humans determined by vein catheterization and microdialysis techniques. Am J Physiol 1994;266:E357-E Williamson DH, Mellanby J, Krebs HA. Enzymatic determination of -hydroxybutyric acid and acetoacetic acid in blood. Biochem J 1961;82: Fleischer JE, Milde JH, Moyer TP, Michenfelder JD. Cerebral effects of high-dose midazolam and subsequent reversal with Ro in dogs. Anesthesiology 1988;68: Madsen PL, Schmidt JF, Wildschiodtz G, Friberg L, Holm S, Vorstrup S, Lassen NA. Cerebral O 2 metabolism and cerebral blood flow in humans during deep and rapid-eye-movement sleep. J Appl Physiol 1991;70: Gjedde A, Lockwood AH, Duffy TE, Plum F. Cerebral blood flow and metabolism in chronically hyperammonemic rats: Effect of an acute ammonia challenge. Ann Neurol 1978;3: Ratnakumari L, Murthy CR. Response of rat cerebral glycolytic enzymes to hyperammonemic states. Neurosci Lett 1993;161: Rama RK, Jayakumar AR, Norenberg MD. Induction of the mitochondrial permeability transition in cultured astrocytes by glutamine. Neurochem Int 2003;43: Rao KV, Norenberg MD. Cerebral energy metabolism in hepatic encephalopathy and hyperammonemia. Metab Brain Dis 2001; 16: Larsen FS. Cerebral circulation in liver failure: Ohm s law in force. Semin Liver Dis 1996;16: Durham S, Yonas H, Aggarwal S, Darby J, Kramer D. Regional cerebral blood flow and CO 2 reactivity in fulminant hepatic failure. J Cereb Blood Flow Metab 1995;15: Lassen NA. The luxury-perfusion syndrome and its possible relation to acute metabolic acidosis localized within the brain. Lancet 1966;2: Ide K, Schmalbruch IK, Quistorff B, Horn A, Secher NH. Lactate, glucose and O 2 uptake in human brain during recovery from maximal exercise. J Physiol 2000;522: Rivers EP, Paradis NA, Martin GB, Goetting ME, Rosenberg JA, Smithline HA, et al. Cerebral lactate uptake during cardiopulmonary resuscitation in humans. J Cereb Blood Flow Metab 1991;11: Hasselbalch SG, Madsen PL, Hageman LP, Olsen KS, Justeseu N, Holm S, Paulson OB. Changes in cerebral blood flow and carbohydrate metabolism during acute hyperketonemia. Am J Physiol 1996;270:E746-E Qu H, Haberg A, Haraldseth O, Unsgard G, Sonnewald U. (13)C MR spectroscopy study of lactate as substrate for rat brain. Dev Neurosci 2000;22: Moller K, Strauss GI, Thomsen G, Larseu FS, Holm S, Sperling BK, et al. Cerebral blood flow, oxidative metabolism, and cerebrovascular carbon dioxide reactivity in patients with acute bacterial meningitis. Acta Anaesthesiol Scand 2002;46: Knudsen GM, Paulson OB, Hertz MM. Kinetic analysis of the human blood-brain barrier transport of lactate and its influence by hypercapnia. J Cereb Blood Flow Metab 1991;11: