Mild Hypothermia Modifies Ammonia-Induced Brain Edema in Rats After Portacaval Anastomosis

Transcription

1 GASTROENTEROLOGY 1999;116: Mild Hypothermia Modifies Ammonia-Induced Brain Edema in Rats After Portacaval Anastomosis JUAN CÓRDOBA, JEFFREY CRESPIN, JEANNE GOTTSTEIN, and ANDRÉS T. BLEI Department of Medicine, Veterans Administration Lakeside Medical Center and Northwestern University, Chicago, Illinois Background & Aims: The pathogenesis of brain edema in fulminant hepatic failure is still unresolved. Mild hypothermia (33 35 C) can ameliorate brain edema after traumatic brain injury. We evaluated mild hypothermia in a model of ammonia-induced brain edema in which accumulation of brain glutamine has been proposed as a key pathogenic factor. Methods: After portacaval anastomosis, anesthetized rats were infused with ammonium acetate at 33, 35, and 37 C or vehicle at 37 C. Water and glutamine levels in the brain, cardiac output, and regional and cerebral hemodynamics were measured when intracranial pressure increased 3 4-fold (ammonia infusion at 37 ) and matched times (other groups). Results: Mild hypothermia reduced ammonia-induced brain swelling and increased intracranial pressure. Brain glutamine level was not decreased by hypothermia. Brain edema was accompanied by a specific increase in cerebral blood flow and oxygen consumption, which were normal in both hypothermic groups. When the ammonia infusion was continued in hypothermic rats, plasma ammonia levels continued to increase and brain swelling eventually developed. Conclusions: Mild hypothermia delays ammonia-induced brain edema. In this model, an increase in cerebral perfusion is required for brain edema to become manifest. Mild hypothermia could be tested for treatment of intracranial hypertension in fulminant hepatic failure. Brain edema is a potentially fatal complication in patients with fulminant hepatic failure, 1 found in some cirrhotic patients with hepatic encephalopathy 2 and a miscellaneous group of patients with acute hyperammonemia. 3 5 Despite progress in intensive care management and liver transplantation, there is no specific treatment for brain edema in these conditions. Once brain swelling occurs, rapid development of intracranial hypertension can lead to brain herniation and death. 6 Experimental studies have shown that a low body temperature ( 30 C) doubles survival time and prevents development of brain edema in a model of acute liver failure. 7 Hypothermia was used as a last resource together with other multiple therapies in a patient with fulminant hepatic failure. 8 However, deep hypothermia has been considered inadequate for therapeutic purposes because of several complications associated with its use. 9 Recently, mild hypothermia (33 35 C) was shown to be neuroprotective in experimental models of traumatic 10 and ischemic brain injury. 11 Although the institution of mild hypothermia is also associated with specific side effects, a randomized controlled trial in brain trauma showed improved survival and less residual deficits when patients were kept at 34 C for 24 hours. 12 Even if clinically ineffective, experiments using hypothermia may help to unravel the pathophysiological mechanisms of brain edema in acute liver failure. Several theories have been proposed to account for brain swelling in fulminant hepatic failure. 13 One hypothesis proposes that brain edema results from the osmotic effects of the accumulation of brain glutamine. 14 We have observed in different experimental preparations of acute liver failure that the development of brain edema follows the increase in brain glutamine level. 15 This sequence can be reproduced with the infusion of ammonia in rats after portacaval anastomosis (PCA), 16,17 a model that has the advantage of a well-standardized preparation performed under strictly controlled physiological conditions. We postulated that mild hypothermia may reduce the formation of brain glutamine, an energy-requiring reaction that occurs in astrocytes. We therefore studied in this model the effects of mild hypothermia on the development of brain edema and changes in brain glutamine. Cerebral hemodynamics were also measured in view of previous studies that have shown an effect of hypothermia on cerebral blood flow and brain oxygen consumption. 18 Abbreviations used in this paper: A VO 2, arteriovenous oxygen difference; CBF, cerebral blood flow; CMRO 2, cerebral metabolic rate of oxygen consumption; CPP, cerebral perfusion pressure; ICP, intracranial pressure; MAP, mean arterial pressure; NH 3, ammonium acetate-infused rats; PCA, portacaval anastomosis; PCO 2, partial pressure of carbon dioxide; PO 2, partial pressure of oxygen; SO 2, fractional oxygen saturation of hemoglobin by the American Gastroenterological Association /99/$10.00

2 March 1999 MILD HYPOTHERMIA PREVENTS BRAIN EDEMA 687 Materials and Methods Experimental Model Male Sprague Dawley rats (Charles River, Wilmington, MA) weighing g were housed in Plexiglas cages in a temperature- and humidity-controlled room in a 12:12 light/dark cycle and allowed free access to water and Rat Chow (Ralston Purina, St. Louis, MO) until the time of experiment. The experiments were approved by the Animal Care Committee at the Lakeside Veterans Administration Medical Center. One day before the infusion of ammonia, a surgical PCA was constructed. Rats were anesthetized with methoxyflurane (Mallinckrodt Veterinary Inc., Mundelein, IL), and an end-toside PCA was constructed under aseptic conditions using a continuous suture technique. 19 The abdomen was closed in two layers, and the animals were returned to their individual cages. The ammonia infusion experiment was performed on PCA rats infused at different temperatures (33 C, 35 C, and 37 C). A group of PCA rats infused with vehicle at 37 C served as control. At the time of the experiment, animals were anesthetized with pentobarbital (50 mg/kg). Catheters (polyethylene [PE]-50; Intramedic, Parsippany, NJ) were placed in the femoral artery and femoral vein. A PE-10 catheter was placed in the cisterna magna through a small posterior scalp incision to monitor intracranial pressure (ICP). A tracheostomy was performed, and the rat was artificially ventilated (model 683; Harvard, South Natick, MA). Arterial and intracisternal pressures were monitored through a pressure transducer (Spectramed PE23XL; Spectramed, Oxnard, CA) with the zero placed at the level of the midbody and tracings continuously recorded in a Sensormedics R611 recorder (Sensormedics, Anaheim, CA). Core body temperature, continuously monitored through a rectal thermometer, was maintained at the desired temperature with the aid of a temperature-controlled blanket. Ammonium acetate (NH 3 ) (55 µmol kg 1 min 1 ) or vehicle (a mixture of sodium acetate and chloride) was infused intravenously. In PCA rats at 37 C, the ammonia infusion was stopped when ICP reached a 3-fold increase above baseline values. In the other experimental groups, ammonia or vehicle infusion was stopped at equivalent times. At time of death, plasma ammonia was measured enzymatically. Experimental Design Experiment A: brain water and brain glutamine. The study was performed in 4 groups (n 8/group). One group of rats was infused with vehicle at 37 C (vehicle-37 ). The other 3 groups consisted of PCA rats infused with ammonia at 37 C (NH 3-37 ), 35 C (NH 3-35 ), and 33 C (NH 3-33 ). At the end of the infusion ( minutes), rats were decapitated and the brain was quickly removed. One half of the brain was used for measuring water in cortical gray matter in fresh tissue. The other half was immediately frozen using liquid nitrogen and stored at 70 C until measurement of brain glutamine by an enzymatic method. 20 Experiment B: systemic and cerebral hemodynamics. Experiment B was conducted in the same 4 experimental groups as in experiment A (vehicle-37, NH 3-37, NH 3-35, and NH 3-33 ) including 6 rats/group. In addition to the cisterna magna (ICP monitoring), femoral vein (intravenous infusion), and femoral artery catheters (arterial pressure monitoring), additional catheters were placed for injection of radioactive microspheres and estimation of cerebral oxygen consumption. This included a PE-10 catheter inserted in the superior sagittal sinus for sampling cerebral venous blood as well as a PE-50 catheter placed into the left ventricle through the right carotid artery, as described previously. 21 The catheter in the latter vessel was advanced into the left ventricle using hemodynamic monitoring to signal its correct position. This catheter was used for the injection of radiolabeled microspheres at the end of the ammonia or vehicle infusion ( minutes). Brain Water Brain water was measured in experiment A after ammonia or vehicle infusion using a gravimetric method. 22 Brain hemispheres kept at 4 C and processed within 15 minutes of decapitation were cut into coronal slices, and cortical punch samples of approximately 10 mg in weight were placed in a bromobenzene-kerosene column. The column had been previously calibrated with varying concentrations of K 2 SO 4 to assure a linear relation (r 0.998). Two minutes after placing samples in the column, the equilibration point was read. A minimum of 12 samples per animal were averaged for the final calculation of water content. Cardiac Output and Regional Blood Flows Cardiac output and regional blood flows were calculated in experiment B using the microsphere technique at the time of intracranial hypertension (3-fold increase in ICP) in NH 3-37 and at matched times in the other groups (NH 3-35, NH 3-33, and vehicle-37 ). At that time, 15-µm microspheres labeled with 85 Sr (New England Nuclear, Boston, MA) were injected into the left ventricle. During the experiment, 85,000 microspheres (in 0.7 ml) were ultrasonicated and aspirated into a 1-mL syringe, vigorously vortexed for 1 minute, and injected over 20 seconds into the left ventricle. Five seconds before injection, blood was continuously withdrawn for the next minute from the femoral artery into a syringe at a rate of ml/min. Animals were killed with an overdose of pentobarbital. Splanchnic organs, kidneys, and brain were removed. Organs were desiccated at 160 C over 24 hours, crushed, and placed in counting tubes at equal heights to avoid geometric distortion. Right and left renal blood flow, portal vein inflow (flow that reaches portal vein equals flow of all splanchnic organs), cerebral blood flow (CBF), and cardiac output were calculated from the radioactivity measured in those tissues by a gamma counter. For all tissues, organ blood flow [cpm organ]/[cpm blood]. 23 In addition, cardiac output [cpm injected]/[cpm blood]. Cardiac index cardiac output/body weight.

3 688 CÓRDOBA ET AL. GASTROENTEROLOGY Vol. 116, No. 3 Cerebral Oxygen Consumption For the calculation of cerebral oxygen consumption, few seconds before the injection of radiolabeled microspheres, 0.2 ml of blood was obtained from the femoral artery and superior sagittal sinus of the brain. Hemoglobin was measured in blood from femoral artery using an enzymatic method (Sigma Chemical Co. kit, St. Louis, MO). Partial pressure of oxygen (PO 2 ), partial pressure of carbon dioxide (PCO 2 ), and ph of these samples were measured in an Instrumentation Laboratory Gas Analyzer (model 1640; Lexington, MA). The fractional O 2 saturation of hemoglobin (SO 2 ) in rat blood was calculated at measured values for ph, PCO 2, and body temperature using the Severinghaus formula with the correction of Cartheuser. 24 For these calculations, arterial bicarbonate concentration was estimated using the Henderson Hasselbach equation with pk of carbonic acid and CO 2 solubility coefficient corrected to appropriate body temperature. The systemic arterial and venous oxygen content were calculated according to the formula: O 2 content (vol%) (1.39 Hb SO 2 ) (PO ). Cerebral arteriovenous oxygen difference (A VO 2 ) was calculated as the difference in oxygen content between systemic arterial and superior sagittal sinus blood. Cerebral oxygen consumption (CMRO 2 ) was calculated as cerebral blood flow multiplied by arteriovenous oxygen difference (CMRO 2 CBF A VO 2 ). Statistical Analysis Results are expressed as mean SE. Baseline and final measurements in the same animal were compared using the paired Student t test. One-way analysis of variance (ANOVA) was used for intergroup comparisons. For significant differences (P 0.05), multiple comparisons were performed using the Student Newman Keuls method. When 2 groups of animals were compared, analysis was performed with the unpaired Student t test. The nonparametric Mann Whitney rank sum test and Kruskal Wallis ANOVA on ranks were used for comparisons between groups that did not follow a normal distribution. Results General Aspects In both experiments, baseline parameters did not differ between the 4 groups (Tables 1 and 2). In experiment A, infusion of ammonia caused a drop in arterial pressure and heart rate that was more pronounced in hypothermic rats (NH 3-35 and NH 3-33). In experiment B, baseline arterial pressure and heart rate were approximately 10% lower than in experiment A. In this experiment, the drop in heart rate and arterial pressure after ammonia infusion, although present in normothermic rats, was only statistically significant in hypothermic animals (NH 3-35 and NH 3-33). Table 1. General Aspects in Experiment A Group (n 8) Weight (g) Baseline MAP (mm Hg) Final MAP (mm Hg) Baseline HR (bpm) Final HR (bpm) Vehicle-37 C a a NH 3-37 C a a NH 3-35 C a a,b NH 3-33 C a,c a,c HR, heart rate. a P 0.02 between baseline and final values (paired Student t test). b P 0.05 between NH 3-35 C vs. vehicle-37 C and NH 3-37 C (ANOVA and multiple comparisons). c P 0.05 between NH 3-33 C vs. vehicle-37 C and NH 3-37 C (ANOVA and multiple comparisons). ICP As previously described in this model, 16 ammonia infusion caused a marked increase in ICP (experiment A, to , P 0.001; experiment B, to , P 0.001) (Figure 1). Comparisons between final values of ICP (ANOVA followed by Student Newman Keuls method) showed that hypothermia prevented the increase in ICP after ammonia infusion. Significantly lower values of ICP were found for the group NH 3-33 C in experiments A and B, and for the group NH 3-35 C in experiment B. Plasma Ammonia PCA rats receiving vehicle exhibit mild hyperammonemia ( µmol/l; n 6) compared with sham animals (38 10 in our laboratory). All PCA rats receiving an ammonia infusion had similar levels: (37 C), (35 C), and µmol/l (33 C). Brain Water and Brain Glutamine Ammonia infusion caused an increase in brain water and brain glutamine levels (P 0.001, ANOVA) (Figure 2). Multiple comparisons showed that the infusion of ammonia under hypothermic conditions Table 2. General Aspects in Experiment B Group Weight (g ) Baseline MAP Final MAP (mm Hg ) (mm Hg ) Baseline HR (bpm) Final HR (bpm) Vehicle-37 C NH 3-37 C NH 3-35 C a,b a,b NH 3-33 C a,c a,c HR, heart rate. a P 0.02 between baseline and final values (paired Student t test). b P 0.05; NH 3-35 C vs. vehicle-37 C, NH 3-35 C, and NH 3-37 C (ANOVA and multiple comparisons). c P 0.05; NH 3-33 C vs. vehicle-37 C, NH 3-33 C, and NH 3-37 C (ANOVA and multiple comparisons).

4 March 1999 MILD HYPOTHERMIA PREVENTS BRAIN EDEMA 689 Figure 1. Baseline and final values of ICP in (A) experiment A and (B) experiment B among vehicle ( ), NH 3-37 C ( ), NH 3-35 C ( ), and NH 3-33 C ( ) groups. Baseline values did not differ among the 4 groups (ANOVA). Final ICPs were significantly different among the 4 groups in experiment A (P 0.001, ANOVA) and in experiment B (P 0.001, ANOVA). In experiment A, multiple comparisons showed significant differences (P 0.05) between NH 3-37 C vs. NH 3-33 C and NH 3-37 C vs. vehicle-37 C. In experiment B, multiple comparisons showed significant differences (P 0.05) between NH 3-37 C vs. NH 3-33 C, NH 3-37 C vs. NH 3-35 C, and NH 3-37 C vs. vehicle. (NH 3-35 and NH 3-33 ) was associated with a lower increase in brain water. However, hypothermia did not modify the increase in brain glutamine after ammonia infusion. Systemic Hemodynamics The infusion of ammonia did not cause changes in cardiac index, renal blood flow, and portal vein inflow (Figure 3). Ammonia infusion was associated with a marked increase in CBF at the time of intracranial hypertension. The increase in CBF resulted in an absolute value of CBF that was 2.7% of cardiac output Figure 2. (A) Water and (B) glutamine levels in the brain at the end of intravenous infusion in experiment A. Brain water (%) was significantly different among the 4 groups (P 0.001, ANOVA). Multiple comparisons showed significant differences (P 0.05) between NH 3-37 C (81.00% 0.09%) vs. vehicle-37 C (79.96% 0.03%), NH 3-37 C vs. NH 3-35 C (80.35% 0.15%), NH 3-37 C vs. NH 3-33 C (80.33% 0.09%), vehicle-37 C vs. NH 3-35 C, and vehicle-37 C vs. NH 3-33 C. Brain glutamine level (µmol/g wet brain tissue) was significantly different among the 4 groups (P 0.001, ANOVA). Multiple comparisons showed significant differences (P 0.05) between vehicle-37 C ( µmol/g) and the 3 other groups: , , and µmol/g. (1% in vehicle-37 ). Rats infused with ammonia under hypothermic conditions (NH 3-35 and NH 3-33 ) did not experience an increase in CBF. The coefficient of variation between left and right renal blood flow (a measure of the reliability of the assessment of regional blood flow) was below 3% for the 4 groups of rats. Cerebral Hemodynamics Table 3 shows cerebral hemodynamics of experiment B. The increase in CBF after ammonia infusion occurred despite a decrease in cerebral perfusion pressure,

5 690 CÓRDOBA ET AL. GASTROENTEROLOGY Vol. 116, No. 3 Figure 3. CBF, portal vein inflow, renal blood flow, and cardiac index (in ml min g 1 of tissue or 100 g body wt) in experiment B. At the end of intravenous infusion, cardiac index, renal blood flow, and portal vein inflow did not differ among the 4 groups of rats. CBF was significantly higher in group NH 3-37 C than in the other 3 groups (*P 0.001, ANOVA; P 0.05, multiple comparisons). Values of CBF (ml min g 1 body wt) are vehicle-37 C ( ) vs. NH 3-37 C ( ) vs. NH 3-35 C ( ) vs. NH 3-33 C ( ). explained by the combined effect of a decrease in mean arterial pressure (MAP) (Table 1) and an increase in ICP (Figure 1) (cerebral perfusion pressure MAP ICP). The increase in CBF after ammonia infusion was associated with a decrease in cerebral vascular resistance. Rats infused with ammonia under hypothermic conditions did not experience changes in either CBF or cerebral vascular resistance. Table 3. Cerebral Hemodynamics in Experiment B Group CBF (ml min g 1 ) CPP (mm Hg ) CVR (wood units/ 100 g ) Vehicle-37 C a NH 3-37 C b b NH 3-35 C NH 3-33 C CPP, cerebral perfusion pressure (CPP MAP ICP); CVR, cerebral vascular resistance (CVR CPP/CBF). a P 0.05 vehicle-37 C vs. NH 3-37 C, NH 3-35 C, and NH 3-33 C (ANOVA and multiple comparisons). b P 0.05 NH 3-37 C vs. vehicle-37 C, NH 3-35 C, and NH 3-33 C (ANOVA and multiple comparisons). Cerebral Oxygen Consumption There were no significant differences in the arteriovenous oxygen difference between the 4 groups of rats (vehicle-37 C, ; NH 3-37 C, ; NH 3-35 C, ; and NH 3-33 C ; P 0.25, ANOVA) (Figure 4). Accordingly, the increase in CBF after ammonia infusion was accompanied by an increase in CMRO 2. In rats infused with ammonia at 35 C and 33 C, the values of CMRO 2 did not differ from those of rats infused with vehicle at 37 C. Prolonged Mild Hypothermia We maintained the ammonia infusion in a group of hypothermic rats at 33 C (n 4). At minutes, ICP increased 3-fold above baseline. Brain water increased (81.1% 0.1%). Two vehicle-infused rats had no increase in ICP and brain water at equivalent time was 80.1% 0.1%. However, plasma ammonia was also markedly increased in the hypothermic group, reflecting the effects of a more prolonged infusion: µmol/l.

6 March 1999 MILD HYPOTHERMIA PREVENTS BRAIN EDEMA 691 Figure 4. Cerebral oxygen consumption at the end of intravenous infusion in experiment B. Cerebral oxygen consumption was significantly higher in the group NH 3-37 C than in the other 3 groups (*P 0.04, ANOVA; P 0.05, multiple comparisons). Discussion This study shows that mild hypothermia delays the development of brain edema and the increase in intracranial hypertension induced by infusion of ammonia in rats with PCA. We also observed that ammoniainduced brain edema was associated with an increase in CBF and cerebral oxygen consumption that was ameliorated by mild hypothermia. The neuroprotective effects of profound hypothermia after brain trauma or brain ischemia have been known for a long time. 25 Similarly, profound hypothermia has been observed to be neuroprotective in rodents after ammonia administration, 26 hepatectomy, 27 and hepatic devascularization. 7 Recently, many experimental studies have shown that even small variations in brain temperature determine the extent of histopathologic injury after ischemic and traumatic brain injury. 28 The present study extends the prior observations of a beneficial effect of hypothermia in experimental models of acute liver failure. The results show that mild reductions in body temperature, at levels that may be used in clinical practice, prevent the development of brain edema after ammonia infusion. Brain edema has been observed in different situations associated with acute hyperammonemia 1 5 and can be reproduced in different species after the infusion of ammonia. 16,18,29 One possible explanation for brain edema is an increase in intracellular osmolarity resulting from the metabolism of ammonia in the brain to form glutamine. 14 This reaction is catalyzed by glutamine synthetase, an enzyme that in the brain is localized exclusively in astrocytes. Administration of methionine sulfoximine, an inhibitor of glutamine synthetase, prevents brain swelling in rats infused with ammonia 16,30 and inhibits cellular swelling in cultures of astrocytes incubated with ammonia. 31 Because temperature affects enzyme activity, we hypothesized that mild hypothermia could decrease the rate of glutamine synthesis. Although mild hypothermia reduced brain edema, it did not affect brain glutamine levels. We did not directly measure the activity of glutamine synthetase or of the degrading activity of neuronal glutaminase; glutamine levels are the net result of the activity of both. Studies in isolated cells may be required to provide a definitive answer on temperature and enzymatic activity. However, the marked and similar increase of brain glutamine seen with mild hypothermia argues for an alternative mechanism to explain the protective effects of a lower temperature. Several abnormalities of CBF and cerebral oxygen consumption have been observed in patients with fulminant hepatic failure. A reduction of CBF may be an initial event, as seen in human (reviewed by Larsen 32 ) and experimental 33 studies. However, a wide interindividual variation exists, 32 reflecting a complex and dynamic phenomenon. The role of these abnormalities in the genesis of brain edema and intracranial hypertension is controversial. One group interprets that brain edema develops because of insufficient oxygen supply to the brain, 34 and another group has proposed that intracranial hypertension is the consequence of gradual cerebral hyperemia. 35 Our results suggest that ammonia-induced brain edema may be related to an increase in CBF. This effect was limited to the brain, as shown by the lack of changes in systemic hemodynamics. Our results point at the development of cerebral vasodilatation in this model but not of insufficient oxygen supply to the brain. CBF was increased because of a reduction in cerebrovascular resistance and was coupled to an increase in CMRO 2,as shown by lack of differences in A VO 2. This interpretation is supported by the observation that rats treated with hypothermia did not develop changes in CBF or CMRO 2. Furthermore, mild hypothermia prevents brain injury in models of brain ischemia via a mechanism that does not reduce oxygen demands and preserves energy supplies. 11 Does mild hypothermia prevent or delay the development of brain edema? At first glance, mild hypothermia seems to prevent the development of brain edema when all animals were studied at matched times. We continued the infusion in hypothermic rats beyond such a time point. More than 1 hour later, the brain became swollen; however, plasma ammonia levels were almost doubled. The ideal experiment would have maintained a steady-

7 692 CÓRDOBA ET AL. GASTROENTEROLOGY Vol. 116, No. 3 state ammonia level in the hypothermic rats, but this is extremely difficult to model with the need to maintain constant blood and brain parameters. In fact, this additional observation further supports our contention that an increase in CBF is necessary for edema to become manifest in our model. Cerebral vasodilatation and the development of ammonia-induced brain edema does not invalidate the glutamine hypothesis. As astrocytes directly interact with endothelial cells, 36 it is possible that astrocyte swelling as a result of increased glutamine levels may determine secondary alterations of the cerebral circulation. This may explain the restoration of cerebrovascular CO 2 responsiveness after inhibition of glutamine synthesis in rats infused with ammonia. 37 In this model, the initial event may be astrocyte swelling induced by glutamine accumulation and secondarily cerebral vasodilatation that contributes to the progression of brain edema, especially to the development of pressure waves. 38 An important mechanism by which mild hypothermia reduces brain injury in experimental models of ischemia is by attenuating excitotoxicity. 39 In situations of energy failure, elevated extracellular levels of glutamate can cause neuronal injury and brain edema. 40 However, the role of excitotoxicity in the generation of ammoniainduced brain edema is controversial. Administration of memantine, an antagonist of the binding of glutamate to N-methyl-D-aspartate receptors, has been shown to reduce intracranial hypertension after ammonia infusion in rats. 41 However, these results are difficult to reconcile with the location of edema in astrocytes and the lack of neuronal injury in this model. We did not examine whether brain energy failure, a requisite for excitotoxicity to develop, is present in this model. However, the increase in brain oxygen consumption makes this possibility unlikely. 42 Additional studies, using brain microdialysis, will be necessary to unravel the participation of excitotoxicity in the development of brain edema. Mild hypothermia may be used as a tool to unravel the role of excitotoxicity. This study points at the beneficial effects of mild hypothermia in a preparation that models brain edema in clinical situations of acute hyperammonemia. Our results provide experimental support to test mild hypothermia for the treatment of brain edema complicating fulminant hepatic failure 1 or cirrhosis. 2 A preliminary report in 5 patients awaiting emergency liver transplantation is highly encouraging. 43 Our experimental results support such testing in a situation that frequently results in irreversible brain damage. References 1. Mas A, Rodes J. Fulminant hepatic failure. Lancet 1997;349: Donovan JP, Shafer DF, Shaw BW Jr, Sorrell MF. Cerebral oedema and increased intracranial pressure in chronic liver disease. Lancet 1998;351: Kindt GW, Waldman J, Kohl S, Baublis J, Tucker RP. Intracranial pressure in Reye syndrome. JAMA 1975;231: Mitchell RB, Wagner JE, Karp JE, Watson AJ, Brusilow SW, Przepiorka B, Storb R, Santos GW, Burke PJ, Saral R. Syndrome of idiopathic hyperammonemia after high-dose chemotherapy: review of nine cases. Am J Med 1988;85: Lichtenstein GR, Kaiser LR, Tuchman M, Palevsky HI, Kotloff RM, O Brien CB, Furth EE, Raps EC, Berry GT. Fatal hyperammonemia following orthotopic lung transplantation. Gastroenterology 1997; 112: Munoz SJ. Difficult management problems in fulminant hepatic failure. Semin Liver Dis 1993;13: Traber P, DalCanto M, Ganger D, Blei AT. Effect of body temperature on brain edema and encephalopathy in the rat after hepatic devascularization. Gastroenterology 1989;96: Rozga J, Podesta L, LePage E, Hoffman A, Morsiani E, Sher L, et al. Control of cerebral oedema by total hepatectomy and extracorporeal liver support in fulminant hepatic failure. Lancet 1993;342: Clifton GL. Systemic hypothermia in treatment of severe brain injury. J Neurosurg Anesthesiol 1995;7: Dietrich WD, Alonso O, Busto R, Globus MYT, Ginsberg MD. Post-traumatic brain hypothermia reduces histopathological damage following concussive brain injury in the rat. Acta Neuropathol 1994;87: Busto R, Dietrich WD, Globus MYT, Valdes I, Scheinberg P, Ginsberg MD. Small differences in intra-ischemic brain temperature critically determine the extent of ischemic neuronal injury. J Cereb Blood Flow Metab 1987;7: Marion DW, Penrod LE, Kelsey SF, Obrist WD, Kochanek PM, Palmer AM, et al. Treatment of traumatic brain injury with moderate hypothermia. N Engl J Med 1997;336: Cordoba J, Blei AT. Brain edema and hepatic encephalopathy. Semin Liver Dis 1996;16: Brusilow SW. Hepatic encephalopathy (letter). N Engl J Med 1986;314: Olafsson S, Gottstein J, Blei AT. Brain edema and intracranial hypertension in rats after total hepatectomy. Gastroenterology 1995;108: Blei AT, Olafsson S, Therrien G, Butterworth RF. Ammoniainduced brain edema and intracranial hypertension in rats after portacaval anastomosis. Hepatology 1994;19: Vogels BAPM, Van Steynen B, Maas MAW, Jorning GG, Chamuleau RAFM. The effects of ammonia and portal-systemic shunting on brain metabolism, neurotransmission and intracranial hypertension in hyperammonemia-induced encephalopathy. J Hepatol 1997;26: Lanier WL. Cerebral metabolic rate and hypothermia: their relationship with ischemic neurologic injury. J Neurosurg Anesthesiol 1995;7: Lee SH, Fisher B. Portacaval shunt in the rat. Surgery 1961;50: Bergmeyer HU. Methods of enzymatic analysis. New York: Academic, Fung HL, Chong S, Blei AT. Cardiac output is an apparent determinant of nitroglycerin pharmacokinetics in rats. J Pharmacol Exp Ther 1986;239: Traber PG, Ganger DR, Blei AT. Brain edema in rabbits with galactosamine-induced fulminant hepatitis. Regional differences

8 March 1999 MILD HYPOTHERMIA PREVENTS BRAIN EDEMA 693 and effects on intracranial pressure. Gastroenterology 1986;91: Heymann MA, Payne BD, Hoffman JE, Rudolph AM. Blood flow measurements with radionuclide labeled particles. Progr Cardiovasc Dis 1977;20: Cartheuser CF. Standard and ph-affected hemoglobin-o 2 binding curves of Sprague Dawley rats under normal and shifted P 50 conditions. Comp Biochem Physiol 1993;106A: Fay T. Observations on generalized refrigeration in cases of severe cerebral trauma. Assoc Res Nerv Ment Dis Proc 1943;24: Schenker S, Warren KS. Effect of temperature variation on toxicity and metabolism of ammonia in mice. J Lab Clin Med 1962;60: Peignoux M, Bernuau J, Benhamou JP. Total hepatectomy and hepatic vascular exclusion in the rat: a comparison, with special reference to the influence of body temperature. Clin Sci 1982;62: Dietrich WD. The importance of brain temperature in cerebral injury. J Neurotrauma 1992;9(suppl 2):S475 S Chodobski A, Szmydynger-Chodobska J, Urbanska A, Szczepnaska- Sadowska E. Intracranial pressure, cerebral blood flow, and cerebrospinal fluid formation during hyperamonemia in the cat. J Neurosurg 1986;65: Takahashi K, Koehler RC, Brusilow SW, Traystman RJ. Inhibition of brain glutamine accumulation prevents cerebral edema in hyperammonemic rats. Am J Physiol 1991;261:H825 H Norenberg MD, Bender AS. Astrocyte swelling in liver failure: role of glutamine and benzodiazepines. Acta Neurochir Suppl 1994;60: Larsen FS. Cerebral circulation in liver failure: Ohm law in force. Semin Liver Dis 1996;16: Shah V, Webster J, Levy R, Gottstein J, Blei AT. A reduction in cerebral blood flow precedes the development of intracranial hypertension in acute liver failure. Hepatology 1993;17: Wendon JA, Harrison PM, Keays R, Williams R. Cerebral blood flow and metabolism in fulminant liver failure. Hepatology 1994; 19: Larsen FS, Hansen BA, Pott F, Ejlersen E, Secher NH, Paulson OB. Dissociated cerebral vasoparalysis in acute liver failure. A hypothesis of gradual cerebral hyperaemia. J Hepatol 1996;25: Janzer RC, Raff MC. Astrocytes induce blood-brain barrier properties in endothelial cells. Nature 1987;325: Takahashi H, Koehler RC, Hirata T, Brusilow SW, Traytsman RJ. Restoration of cerebrovascular CO 2 responsivity by glutamine synthesis inhibition in hyperammonemic rats. Circ Res 1992;71: Cordoba J, Blei AT. Cerebral edema and intracranial pressure monitoring. Liver Transplant Surg 1995;1: Busto R, Globus MYT, Dietrich WD, Martinez E, Valdes I, Ginsberg MD. Effect of mild hypothermia on ischemia-induced release of neurotransmitters and free fatty acids in rat brain. Stroke 1989; 20: Lipton SA, Rosenberg PA. Excitatory amino acids as a final common pathway for neurologic disorders. N Engl J Med 1994; 330: Vogels BAPM, Maas MAW, Daalhuisein J, Quack G, Chamuleau RAFM. Memantine, an non-competitive NMDA-receptor antagonist improves hyperammonemia-induced encephalopathy and acute hepatic encephalopathy in rats. Hepatology 1997;25: Madsen PL, Linde R, Hasselbach SG, Paulson OB, Lassen NA. Activation-induced resetting of cerebral oxygen and glucose uptake in the rat. J Cerebr Blood Flow Metab 1988;18: Jalan R, Olde Damink SWM, Lee A, Hayes PC. Moderate hypothermia for uncontrolled intracranial hypertension in acute liver failure: a bridge to liver transplantation (abstr). Hepatology 1998;28:492A. Received June 9, Accepted November 23, Address requests for reprints to: Andres T. Blei, M.D., Department of Medicine, Lakeside Veterans Affairs Medical Center Room 111 E, 333 East Huron Street, Chicago, Illinois andyblei@merle.acns.nwu.edu; fax: (312) Supported by Merit Review from the Veterans Administration Research Service and the Stephen B. Tips Memorial Fund. Dr. Córdoba was supported by a grant from the Spanish Government (FIS/BAE). Presented at the meeting of the American Association for the Study of Liver Diseases and published in abstract form (Hepatology 1997;26:182A).