I brain edema from that of intracranial hypertension,

Transcription

1 Cerebral Edema and ICP Monitoring 187 Juan Cordoba and Andres T. Blei With the wide acceptance of liver transplantation as a therapeutic alternative in fulminant hepatic failure (FHF), the successful management of patients with this syndrome has acquired a new urgency. Topping the list of medical problems is the development of brain swelling. Two decades after the recognition of its importance,' brain edema and intracranial hypertension still constitute a major cause of death in these patients. In a more recent classification of FHF, brain edema was especially prominent in those subjects with "hyperacute failure,"* in whom a period of 7 days or less elapsed between the development of jaundice and encephalopathy. The goal of this review is to discuss two aspects of this clinical problem. On one hand, elucidation of its pathogenesis should lead to a more rational therapeutic approach; such an information would also be valu- able to understand the relationship between hepatic encephalopathy and brain edema, a source of controversy. Studies of pathogenic mechanisms are difficult to perform in humans and animal models of FHF have proven valuable, as brain swelling can be detected with some regular- On the other hand, an increasing array of techniques is now available in the intensive care setting to monitor patients with FHF. Of these, intracranial pressure monitoring has received the most critical attention. However, concerns with the risks of craniotomy and the need to acquire more dynamic information has led several groups to explore non-invasive methods that evaluate the consequences of intracranial hypertension. Their role, though potentially exciting, is still uncertain. Copyright by the American Association for the Study of Liver Diseases t is important to differentiate the development of I brain edema from that of intracranial hypertension, whereas one represents a net increase in the water content of the brain, the other denotes a rise of pressure within the skull.4 Brain edema, when uncontrolled, evolves into intracranial hypertension; however, other factors can contribute to the rise in pressure. From the three compartments whose volume can potentially increase within the skull-brain tissue, cerebrospinal fluid (CSF) and blood volume (Fig 1) brain tissue is the only one that shows unequivocal signs of expansion. An expansion of CSF is highly improbable in view of the repeated demonstration of normal or diminished ventricular size on CT ~canning.~ The potential contribution of cerebral blood volume (CBV) to the development of intracranial hypertension has been highlighted by recent studies of cerebral perfusion in FHF.6.7 A rise in CBV can by itself increase intracranial pressure (ICP) without causing brain swelling.8 In FHF, the role of CBV is probably important in the final stages, when a decrease in brain compliance is present and small changes in volume can markedly affect ICP (Fig 1). There is no system for an in vivo continuous measurements of CBV; therefore, cerebral blood flow (CBF) has been used as an indirect parameter. Even this is equivocal, because an increased blood volume can exist with a reduced blood flow, as seen with venous obstruction. In any case, clinical studies have shown a high intra and interindividual variability of cerebral blood flow in FHF, a variability that reflects the complex relationship between hepatic encephalopathy, brain metabolism, and intracranial pressure. There is evidence of a decrease in cerebral blood flow with progression of the disea~e.~.'~ An improvement in the neurological status or an optimization of oxygen use is linked to an increase in However, hyperemia is seen in some patients, especially in those who develop more severe intracranial hypertension.6 In this situation, the increase of blood flow is not coupled to a neurological or metabolic improvement. Loss of cerebrovascular autoregulation (the ability to maintain a constant flow in the face of variations of perfusion pressure) may be present, as suggested in an animal model of FHF." A sudden increase of blood volume as a result of arteriolar vasodilatation may explain the appearance of pressure waves, a sign that may signal imminent brain herniation, l2 Brain Edema An increase in the water content of cerebral tissue (brain edema) is the event that leads to intracanial From Lakeside VA Medical Center and Northwester University, Chicago, IL. Supported by the Veterans Administration Research Service and Fi (Spanish Government). Address reprint requests to Andres T. Blei, MD, Northwestern Memorial Hospital, Passavant Pavilion, Suite 726, Supenor St and Fairbanks Ct, Chicago, IL Copyright by the Amencan Association for the Study of Liver Diseases /95/ $3.00/0

2 188 Cordoba and Blei ICP I B Brain Volume Figure 1. (A) The diagram separates three compartments in the brain by their relative volume: brain tissue (70%), cerebrospinal fluid (CSF, 25%), and blood volume (BV, 5%). (B) In an early stage of increase in brain volume, large changes in volume result in small changes in intracranial pressure (ICP); at a later stage, brain compliance is reduced, and small changes in volume cause large changes in pressure. hypertension in FHF.4 Conceptually, brain edema can develop by an increase in vascular permeability to water or osmogenic metabolites (vasogenic) or by an accumulation of osmolytes generated in the brain (cellular or cytotoxic). Although an increase in bloodbrain barrier permeability for certain substances has been observed in animal models of FHF, it appears late in the course of FHF. l3 The normal appearance of the capillary endothelium on human brain biopsy specimens14 and the observation in animal models that swelling of cortical astrocytes is a consistent finding15 support the notion that brain edema in this condition is fundamentally of a cytotoxic nature. Additional indirect evidence is the absence of focal involvement on CT scanning (as seen in cases of vasogenic edema), normal protein levels in the CSF and the finding that in animal models, the cortical gray mater is the main site of swelling.16 Several ideas have been developed to explain the pathogenesis of brain swelling. Recent data in hepatectomized rats support the concept that brain edema is a consequence of liver failure rather than the effect of toxins released from the necrotic 1i~er.l~ Thus, interest has focused on identifjmg a circulating toxin(s) as well as in the localization of altered molecular mechanism(s). A postulated inhibition of brain Na+-K+ adenenosine triphosphatase (ATPase) by a circulating factor1* does not explain the selective involvement of astrocytes and has not been supported by data in the experimental animal.19 Ammonia may be involved in the pathogenesis of brain edema. Intracranial hypertension occurs in other hyperammonemic conditions, such as children with urea cycle deficiencies,20 postchemotherapy2 and in Reye s syndrome.** Brain edema is not an effect of ammonia per se but arises after the generation of osmogenic metabolites during ammonia detoxification. The glutamine hypothesis postulates that brain edema appears as the consequence of the accumulation of this amino acid. Glutamine is a product of glutamate amidation in astrocytes, the sole cell of the brain that contains glutamine synthetase, which catalyzes this reaction.23 Several studies have shown a protective effect of methioninesulfoximine, an inhibitor of glutamine synthethase, in the development of ammonia-induced brain swelling.24,25 Unfortunately, inhibition of glutamine synthesis with this inhibitor may not be a practical therapeutical approach because of drug toxicity and direct neurotoxicity from the high levels of ammonia reached in this circumstance. A criticism of the glutamine hypothesis has been the lack of clinical evidence of intracranial hypertension in acute episodes of encephalopathy in chronic liver disease, where brain glutamine also rises. In spite of preliminary reports to the contrary,16 most hepatologists do not recognize brain edema in cirrhosis. A recent study using nuclear magnetic resonance (NMR) spectroscopy sheds light on this matter. A patient in whom a transjugular intrahepatic portosystemic shunt (TIPS) was placed, a procedure that can

3 Cerebral Edema and ICP Monitoring 189 be viewed as an acute ammonia load, had an increase in brain glutamine with a concomitant decrease in the content of other organic osmolytes, such as myoinosit01.l~ Thus, the development of brain edema in FHF, and especially in those with the hyperacute form of the syndrome,2 may arise from the accumulation of an osmole(s) as well as the absence of a compensatory mechanism that allows brain osmolarity to be restored. Such mechanisms operate in other conditions that affect brain osmolarity, such as acute and chronic hyponatremia or hypernatremia.18 Effects of lntracranial Hypertension A major consequence of intracranial hypertension is the effect on cerebral perfusion. The maintenance of cerebral blood flow is critical to assure an adequate supply of oxygen. The driving force in maintaining a stable blood flow is the cerebral perfusion pressure (CPP), the arithmetical difference between mean arterial pressure (MAP) and ICP (intracranial pressure). When cerebral perfusion pressure is less than 50 mm Hg, cerebral blood flow falls precipitously and flow thresholds-initially for electrical derangement and later for structural cell damage-are reached.2y Although these thresholds are in general valid for FHF, they have been challenged by the observation of patients that recovered from this situation without neurological damage.30 Indeed it has been shown that in FHF low rates of brain oxygen consumption (25% of normal) are not infreq~ent,~ probably reflecting a decrease of brain energetic needs because of the encephalopathy. Ischemia may be present in others, as increased lactate production has been observed in some patients with intracranial hypertension. An0 ther consequence of intracranial hypertension is the compression of nervous structures. The increase in pressure causes displacement of brain tissue, resulting in herniation and direct compression of the temporal lobe or the cerebellum. Pupillary abnormalities may reflect compression of the 111 nerve. Brain stem compression can result in sudden respiratory arrest and circulatory collapse. lntracranial Pressure Monitoring Diagnosis Brain edema does not result in clinical manifestations unless intracranial hypertension is present, as the displacement of brain tissue is the factor that results in neurological symptoms. The diagnosis of intracra- nial hypertension based on clinical signs (decerebrate rigidity, myoclonus, seizures, pupillary abnormalities) is unreliable, because clinical signs can be absent with pressures as high as 60-mm Hg.31 When manifest, symptoms can appear precipitously, and consequences can be irreversible (respiratory arrest, hypotension unresponsive to therapy). Furthermore, because of the risk of sudden respiratory arrest, these patients are intubated and paralyzed when they are in coma, making a diagnosis based on pupillary signs difficult to monitor; arterial hypertension, a sign of the Cushing reflex, is inconsistently present. Patients who exhibit an elevation of arterial pressure have a better prognosis,32 because they are able to maintain a better cerebral perfusion pressure. There are no methods to monitor water accumulation in the brain. CT scans are seldom useful to detect brain edema in patients with FHF, even at a time when intracranial hypertension is present.* The only accurate method to measure intracranial pressure is the placement of an intracranial pressure transducer. When ICP monitoring was compared with frequent bedside exploration, the latter did not diagnose more than 25% of episodes of intracranial hypertension, leading to a fewer number of therapeutic intervention^.^^ Although ICP monitoring is the most sensitive method for assessment of intracranial events, concern has risen about its true value. Intracranial hypertension appears when the prognosis is poor, and the procedure itself can be harmful. In one study in which subdural bolts where placed in patients with FHF in coma without contraindications for liver transplant, ICP monitoring resulted in the identification of episodes of intracranial hypertension responsive to therapy in 17% of patients, whereas in 21% it was associated with complication^.^^ However, in patients suffering from severe head trauma, ICP monitoring modifies the In patients with FHF, its effect on survival has not been studied in a prospective randomized fashion. In one study using epidural transducers, ICP monitoring did not result in an increase on global survival but with its use the median time to death was delayed to 50 hours.32 Early identification and treatment of mild rises of ICP delays the onset of severe episodes, allowing time for a new liver to become available for transplantation. Thus, ICP monitoring should only be considered in patients that are candidates for liver transplantation. In this setting, a prognosis for neurological recovery can be established and perioperative management will be optimized3* (Table 1).

4 ~ 190 Cordoba und Blei - Table 1. Advantages of ICP Monitoring Early diagnosis (before clinical signs) Accurate diagnosis (identification of subclinical episodes) Objective measurement of the effects of therapy (mandatory with barbiturates) Information about the effects of nursing maneuvers (avoidance of certain postures) Possible estimation of brain compliance (with appropriate computerized analyses of pressure waves) Optimization of anesthetic and surgical management before and during liver transplantation Identification of individuals with poor neurological prognosis The choice of the intracranial pressure transducer has important consequences. In these patients, a severe coagulopathy increases the risk of intracranial hemorrhage, that can be fatal. From the observations of a multicentric survey36 and the experience of single institution^,^^,^^ epidural transducers (5% hemorrhage) are preferable to ones where the dura is pierced, such as subdural bolts or intraparenchymal monitors (up to 20% hemorrhage, Fig 2). Epidural monitors provide less-accurate readings, but in addition to fewer risks of severe complications, their placement is easier and can even be performed by medical staff of intensive care units. The correct interpretation of ICP readings requires familiarity with monitor techniques and handling. Medical staff should be able to identify false readings and their of 25 INFECTION HEMORRHAGE Cases source. Frequent checks of the equipment and correct calibration of monitors, under neurosurgical supervision, is necessary. Recordings should be continuously registered on paper. This permits a better assessment of the tracing as well as evaluation of current or newer therapies. In a recent experience with a liver assist device, the reduction of intracranial pressure appeared as one of the beneficial effects of this therapy.39 Because of the complexity of the syndrome and its low incidence in the United States, patients with FHF should be transferred to centers that are capable of performing emergency liver transplantation. ICP monitors should not be inserted before transfer, in view of the potential harm of the procedure and the need for specialized management. Placement of intraparenchymal monitors at the time of intracranial hypertension increases the risk of hemo~hage.~~ Treatment It would be highly desirable to have a treatment that prevents brain water accumulation before the development of intracranial hypertension. Unfortunately, no specific therapy to prevent brain swelling is currently available. Treatment of intracranial hypertension is currently focused on (1) Proper treatment of multiorgan complications associated with FHF40; (2) avoidance of factors that increase ICP (Table 2); (3) administration of drugs that decrease ICP (Table 3); and (4) liver transplantation. With ICP monitoring, an objective goal for therapy can be established. Extrapolating from the experience with severe head trauma and Reye s syndrome, it is common to start treatment in FHF with an ICP over 25 mm Hg (normal 0-10 mm Hg). The factors of Table 2 can cause a transient increase of ICP. For this reason antihypertensive therapy is not started for rises that last less than 5 minutes. However, pressure E S P E S P Figure 2. Summary of complications of intracranial pressure monitoring in 267 patients.36 A low number of infections with epidural (E), subdural (S) and parenchymal (P) monitors reflects the short duration of monitoring. On the other hand, piercing the duramater was associated with a high number of hemorrhagic complications. Table 2. Factors that Increase lntracranial Pressure Valsalva maneuver Head turning and moving Neck vein compression Respiratory suctioning Psychomotor agitation Seizures Vasodilator agents Fever

5 Cevebrul Edema and ICP Monitonng 191 Table 3. Management of lntracranial Hypertension in Fulminant Hepatic Failure General measures Head elevation to 20" Avoid factors of Table 2 and fluid overload Tracheal intubation at onset of stage 3 Keep PC02 below 30 mmhg Placement of an ICP transducer at onset of stage 4 If intracranial hypertension* Mannitol bolus (0.5-1 gm kg IV bolus over 30 min) Pentobarbital infusion according to the effects on CPPt *For ICP > 25 mm Hg for more than 5 minutes. tkeep CPP > 50 mm Hg, using vasopressors if necessary. The usual doses are boluses of 100 to 150 mg IV every 15 minutes during 1 hour followed by 1 to 3 mg/kg/h. waves (also termed Lundberg type A and B waves41) require prompt attention. The mainstay of treatment is osmotherapy with manit~l.~~ Corticosteroids are not useful and mechanical ventilation provides only a marginal benefit, that disappears with time due to renal metabolic compensation. More recent studies raise the possibility that hyperventilation may actually worsen cerebral ischemia in FHF.' Mannitol should be given as a bolus, 0.5 to 1 mg/kg over 30 minutes, but not at fixed intervals. During treatment, plasma osmolarity should be checked and not allowed to rise above 320 mosm/ kg. If renal failure develops, dialysis is needed to eliminate mannitol and allow subsequent administration of the drug. In case of failure of mannitol, barbiturate infusion is the next step.43 The decrease in brain metabolism brought on by the drug is thought to underlie its beneficial effects. Pentobarbital and its precursor, thiopental, can be administrated at established doses, but should be adjusted, based on the effects on ICP and arterial pressure. Plasma levels should be monitored, and electrophysiological recordings are recommendable, whereas clinical assessment is not possible. With barbiturates a decrease in arterial pressure is a constant feature and vasoconstrictive agents are frequently needed to maintain cerebral perfusion pressure. In cases of refractory intracranial hypertension, maintenance of cerebral perfusion pressure over 50 mm Hg is more important that the normalization of ICP. Supplementary causes of intracranial hyperten- sion (intracranial hemorrhage) or monitor malfunctioning should be excluded. Additional modalities of therapy (liver assist device, hypothermia, hepatectomy) are still investigational and should be considered under research protocols. The proposed beneficial effects of N-acetylcystein in this setting await ~onfirmation.~~ Liver transplantation is the definitive treatment for intracranial hyperten~ion.~~ ICP monitoring helps to optimize penoperative management. During anesthesia, inhalational agents should be avoided, because of their vasodilatory properties. Opiates, benzodiaepines, and neuromuscular blockers are the agents of choice. Normalization of ICP after liver transplant can be delayed for more than 24 hours.3 One of the most difficult decisions is to preclude a liver transplant in FHF because of concerns with irreversible brain damage. Electrophysiological assessment is often difficult to interpret because of the concomitant use of neuroactive agents. The observation that neurological recovery is often absent when CPP has fallen below 40 mm Hg for more than 2 hours has led several groups to use this criteria as a contraindication for tran~plantation.~~ However, recent observations have challenged this statement,30 making a more flexible approach recommendable, where CPP is one of the factors in the decisionmaking process. Future Directions Two are the areas undergoing development: (1) noninvasive procedures to supplement or even potentially replace ICP monitoring; and (2) a more accurate assessment of brain damage secondary to intracranial hypertension. In this setting, transcranial doppler ultrasonography (TCDU) appears to be a promising tool. TCDU provides continuous monitoring of blood velocity in the middle cerebral artery by a transducer attached in a fixed position with a band. During a rise in ICP, it is possible to appreciate a decrease or reversal of the diastolic flow velocity.47 The difference between systolic and diastolic velocity (pulsatility index) has correlated well with CPP in patients with severe head trauma for CPP between 20 and 70 mm Hg,48 and it has been suggested that it could substitute ICP monitoring. However, TCDU results are more reliable in assessing the evolution of a patient than in

6 192 Cordoba and Blei providing a true assessment of cerebral perfusion pressure. Furthermore, with a low perfusion pressure or high cerebral blood flow, this procedure can lead to an erroneous interpretation of the actual cerebral perfusion pressure. Measurements of flow are also subject to error; because flow is the product of velocity times vessel area, small changes in vessel diameter can markedly affect the calculation of flow. Jugular oximetry continuously estimates the cerebral metabolic rate of oxygen consumption by calculating the cerebral arterio-venous difference of hemoglobin sat~ration.~~ A catheter is placed in a cephalic direction in the jugular venous bulb, and measurements are obtained from this site and a peripheral artery. This provides information about the degree of coupling between arterial flow and energetic needs, though the results are more difficult to interpret in the absence of flow measurements. In general, a narrow difference suggests the presence of redundant flow (hyperemia), whereas a low venous saturation suggests the existence of brain ischemia. In two recent studies, measurements of total cerebral blood flow could be compared to the arteriovenous oxygen difference. In one study, the metabolic rate for oxygen (CMR02, the product of blood flow and arteriovenous oxygen difference) was reduced; a state of cerebral ischemia was inferred by the production of lactate in the brain. In another study,6 the arteriovenous oxygen difference was found to be low, a difference that was interpreted as indicative of cerebral hyperemia. In any case, a low CMROz did not preclude a complete neurological recovery. Thus, the exact role of jugular oximetry remains to be defined. If arteriovenous shunts are open in FHF, it is even possible that jugular oximetry underestimates CMR02. Electrophysiological recordings have been always used to assess the viability of damaged brain, but the results have been imprecise. Recently, an interesting study has suggested that sensory-evoked potentials may allow physicians to discriminate among patients who will have a spontaneous recovery as well as those in whom an early liver transplantation is indicated.jo The prolongation of the N70 peak (long latency) as well as the reduction of the amplitude of N2O peaks (short latency) were thought to provide a better discriminant function than clinical criteria in deciding the need for transplantation. This approach needs confirmation, because the original clinical criteria for emergency transplantation set by the Kings College51 as well as the Clichy groups52 are still subject of controversy. j3 In the future, it is possible that with multiple monitoring of patients with FHF, some of the unsolved pathophysiological questions regarding brain swelling and intracranial hypertension will be answered. In the meantime, the importance of intracranial pressure monitoring in the management of severe FHF cannot be dismissed. References 1. Ware AJ, D Agostino AN, Combes B. Cerebral edema: A major complication of massive hepatic necrosis. Gastroenterology 1971 ;61: O Grady J, Schalm SW, Williams R. Acute liver failure: redefining the syndromes. Lancet 1993;342: Blei AT, Traber PG. Brain edema in experimental fulminant hepatic failure. In Butterworth R, Pomier Layrargues G (eds). Hepatic encephalopathy. Clifton, NJ: Humana Press, 1989; Blei AT. Cerebral edema and intracranial hypertension in acute liver failure: Distinct aspects of the same problem. Hepatology 1991 ;13: Mufioz SJ, Robinson M, Northrup B, Bell R, Moritz M, Jarrel B, et al. Elevated intracranial pressure and computed tomography of the brain in fulminant hepatocellular failure. Hepatology 1991 ;13: Agarwall 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: Wendon JA, Harrison PM, Keays R, Williams R. Cerebral blood flow and metabolism in fulminant hepatic failure. Hepatology 1994;19: Langfitt W. Increased intracranial pressure and the cerebral circulation. In: Youmans JR. Neurological Surgery. Philadelphia: Saunders 1982; Shah V, Webster S, Gottstein J, Blei A. Reduction of cerebral perfusion precedes rise of intracranial pressure in rats with ischemicfulminant liver failure. Hepatol- Ogy Almdal T, Shroeder T, Ranek L. Cerebral blood flow and liver function in patients with encephalopathy due to acute and chronic liver diseases. Scand J Gastroenterol 1989;24: Larsen FS, Knudsen GT, Paulson OB, Vilstrup H. Cerebral blood flow autoregulation is absent in rats with thioacetamide-induced hepatic failure. J Hepatol 1994; Hanid MA, Davies M, Mellon PJ, Silk DBA, Strunin L, McCase JJ, Williams R. Clinical monitoring of intracranial pressure in fulminant hepatic failure. Gut 1980;21: Zaki AEO, Ede RJ, Davis M, Williams R. Experimental studies of blood-brain barrier in acute hepatic failure. Hepatology 1984;4: Kato MD, Hughes RD, Keays RT, Williams R. Electron microscopic study of brain capillaries in cerebral edema from fulminant hepatic failure. Hepatology 1992;15: Norenberg MD. The role of astrocytes in hepatic encephalopathy. Neurochem Pathol 1987;6:13-33.

7 Cerebral Edema and ICP Monitoring Traber PJ, Ganger DR, Blei AT. Brain edema in rabbits with galactosamine-induced fulminant hepatitis. Gastroenterology 1986;91: Olafsson S, Gottstein J, Blei AT. Brain edema and intracranial hypertension in rats after total hepatectomy. Gastroenterology 1995:108 (in press). 18. Seda HMW, Hughes RD, Gove CD, Williams R. Inhibition of rat brain Na+-K+ ATPase activity by serum from patients with fulminant hepatic failure. Hepatology 1984; 4: Pappas SC, Ferenci P, Jones EA. Evidence against the hypothesis that cerebral edema in fulminant hepatic failure is due to decreased neural Na+/K+ ATPase activity. Hepatology 1983;3: Brusilow SW. Inborn errors of urea synthesis. In Genetic and Metabolic Disease in Pediatrics. Lloyd J and Scriver C (eds). London: Butterworths; Watson AJ, Chambers T, Karp JE, Risch VR, Walker WG, Brusilow SW. Transient idiopathic hyperammonemia in adults. Lancet 1985;328: Kindt GW, Waldman J, Kohl S, Baublis J, Tucker RP. lntracranial pressure in Reye syndrome. JAMA 1975; Martinez-Hernandez A, Bell KP, Norenberg MD. Glutamine synthethase: glial localization in brain. Science 1977;195: Takahashi H, Koehler RC, Brusilow SW, Traysman RJ. Inhibition of brain glutamine accumulation prevents cerebral edema in hyperammonemic rats. Am J Physiol 1991 ;261:H Blei AT, Olafsson S, Therrien G, Butterworth RF. Ammonia-induced brain edema and intracranial hypertension in rats after portacaval anastomosis. Hepatology 1994; Donovan JP, Quigley EMM, Sorell MF, Zetterman RE, Castaldo P, Markin RS, Shaw BW. Cerebral edema as a complication of chronic liver disease. [Abstract]. Hepatology 1990;12: Haussinger D, Laubenberger J, Von Dahl S, Ernst T, Bayer S, Langer M, Gerok W, Hennig J. Proton magnetic resonance spectroscopy studies on human brain myoinositol on hyposmolarity and hepatic encephalopathy. Gastroenterology 1994;107: Lee JH, Arcinue E, Ross BD. Organic osmolytes in the brain of an infant with hypernatremia. N England J Med 1994;331: Branston NM, Symon L, Crockard HA. Relationship between the cortical evoked potential and local cortical blood flow following acute middle cerebral artery occlusion in the baboon. Exp Neurol 1974;45: Davies MH, Mutimer D, Lowes J, Elias E, Neuberger J. Recovery despite impaired cerebral perfusion in fulminant hepatic failure. Lancet 1994;343: Stone JL. Nonsurgical management of increased intracranial pressure. Semin Neurol 1989;9: Keays RT, Alexander GJM, Williams R. The safety and value of extradural intracranial pressure monitors in fulminant hepatic failure. J Hepatol 1993;18: Lidofsky SD, Bass NM, Prager MC, Washington DE, Read AE, Wright TL, Ascher NL et al. lntracranial pressure monitoring and liver transplantation for fulminant hepatic failure. Hepatology 1992;16: McGilliudicy JE. Cerebral protection: pathophysiology and treatment of increased intracranial pressure. Chest 1985;87: Keays R, Potter D, O Grady J et al. lntracranial and cerebral perfusion changes before, during and immediately after orthotopic liver transplantation for fulminant hepatic failure. Q J Med 1991 ;79: Blei A, Olafsson S, Webster S, Levy R. Complications of intracranial pressure monitoring in fulminant hepatic failure. Lancet 1993;341: Ellis AJ, Wendon J, Williams R. Efficacy and safety of intra-cranial pressure monitoring in fulminant hepatic failure [Abstract]. J Hepatol 1994;24(Suppl):Pl C4/ Aldersley MA, Richardson P, Juniper M, O Grady JG. A prospective study of complications of intracranial pressure monitoring in acute liver failure. [Abstract]. J Hepatol 1994;24(Suppl 1):Pl C4/ Rozga J, Podesta L, LePage E, Morsiani E, Moscioni AD, Hoffman A, Villamil F, et al. A bioartificial liver to treat severe acute liver failure. Ann Surg 1994;219: Mutioz SJ. Difficult management problems in fulminant hepatic failure. Semin Liver Dis 1993;13: Lundberg N. Continuous recording and control of ventricular fluid pressure in neurosurgical practice. Acta Psychiat Scand 1960;36(suppl 149):l Canalese J, Gimson AES, Davies C, Mellon PJ, Davis M, Williams R. Controlled trial of dexamethasone and mannitol for the cerebral edema of fulminant hepatic failure. Gut 1982;23: Forbes A, Alexander GJ, O Grady JG, Keays R, Gullan R, Dawling S, Williams R. Thiopental infusion in the treatment of intracranial hypertension complicating fulminant hepatic failure. Hepatology 1989;10: Harrison PM, Wendon JA, Gimson AES, Alexander GJM, Williams T. Improvement by acetylcysteine of hemodynamics and oxygen transport in fulminant hepatic failure. New Engl J Med 1991 ;324: Bismuth H, Samuel D, Gugenheim J, Castaing D, Beranu J, Rueff B, Benhamou JP. Emergency liver transplantation for fulminant hepatitis. Ann Intern Med 1987;107: lnagaki M, Shaw 8, Shafer D, Phillen T, Langnas A, Stratta R, Donovan J, Sorrel1 M. Advantages of intracranial pressure monitoring in patients with fulminant hepatic failure [Abstrat]. Gastroenterology 1992;102: A Ringelstein EB. Transcranial doppler monitoring. In Aaslid R (ed). Transcranial Doppler Sonography. Wien: Spinger Verlag, 1986:i Chan KH, Miller JD, Dearden NM, Andrews PJD, Midgley S. The effect of changes in cerebral perfusion pressure upon middle cerebral artery blood flow velocity and jugular bulb venous oxygen saturation after severe brain injury. J Neurosurg 1992;77: Sutton LN, McLaughlin AC, Dante S, Kotapka M, Sinwell T, Mills E. Cerebral venous oxygen content as a measure of brain energy metabolism with increased intracranial pressure and hyperventilation. J Neurosurg 1990;73: Madl C, Grimm G, Ferenci P, Kramer L, Yeganehfar W,

8 194 Byers W. ShawJr. Oder W, Steininger R, et al. Serial recording of sensory evoked potentials: a noninvasive prognostic indicator in fulminant liver failure. Hepatology 1994;20: O'Grady, Alexander GJ, Hayllar KM, Williams R. Earlyindicators of prognosis in fulminant hepatic failure. Gastroenterology 1989;97: Bernau J, Samuel D, Durand F, Bourliere FSM, Adam R, Gugenheim J, Castaing D, et al. Criteria for emergency liver transplantation in patients with acute viral hepatitis and factor V below 50% of normal: a prospective study [Abstract]. Hepatology 1991 ;14:49A. 53. Pauwells A, Mostefa-Kara N, Florent C, Levy VG. Emergency liver transplantation for acute liver failure: evaluation of London and Clichy criteria. J Hepatol 1993;17: AUXILIARY LIVER TRANSPLANTATION FOR ACUTE LIVER FAILURE Byers W. ShawJv. s the name suggests, an auxiliary liver graft is A one that acts as a supplement to the native liver. In supplementing native liver function, one can elect to place the auxiliary graft in a heterotopic position or, by removing part of the native liver, in an orthotopic position. The advantage of the orthotopic position is that it allows one to attach the venous drainage of the graft to a low-pressure venous outflow tract and, at the same time, to provide the graft with portal venous inflow. The disadvantage is that one has to remove part of the native liver. The advantage of the heterotopic position is that one can stay entirely away from the native liver in finding a location for the allograft. The disadvantage lies in the difficulty of providing space, a low-pressure venous outflow conduit and satisfactory portal inflow. The experience of some groups during the past several years has been important in all but eliminating most of these largely technical considerations. Likewise, for most well-trained liver transplant surgeons, the native hepatectomy is rarely the daunting experience it has been portrayed to be by those advocating its omission. We are still faced with other concerns about auxiliary grafting, concerns that are ultimately more important and perhaps more interesting than either the details of vascular anastomoses or the problem of adequate hollow spaces within the abdomen. These concerns are best addressed in light of the specific From the Department of Surgery, University of Nebraska Medical Center, Omaha, NE. Address reprint requests to Byers W. ShawJr., MD, University of Nebraska Medical Center, Department of Surgery, 600 S. 42nd St, Omaha, NE Copright by the American Association for the Study of Liver Diseases /95/ $3.00/0 reasons that one might opt for an auxiliary as opposed to total liver transplantation. Although a bit like advocating the existence of noise that no one is there to hear, proponents of auxiliary transplantation for acute liver failure claim that in most cases, had the patient survived, the native liver would have recovered.' If true, then removal of the entire native liver during transplantation for acute liver failure eliminates any such recovery and unnecessarily commits the patient to the lifelong risks of immunologic rejection and immunosuppression. Orthotopic Auxiliary Transplantation The reported experience using auxiliary orthotopic liver transplantation for temporary support of patients with acute liver failure consists of one published case report and approximately 15 as yet unpublished cases that were the subjects of two different oral presentation^.^.^ Full functional and histological recovery of the native liver was the rule in these selected cases, with a few important exceptions. Based on this limited experience and on the discussion that follows, the following conclusions seem warranted: (1) any patient with acute liver failure caused by a process that either has or likely will resolve should be considered a candidate for auxiliary transplantation; (2) patients with a subacute course may be more likely to demonstrate early fibrosis of the native liver, a potential contraindication to auxiliary grafting; (3) in the event that the native liver does not recover, the orthotopic position for the auxiliary graft offers the distinct advantage of being ideal for total transplantation so that the residual native liver can be removed without the need to either replace or reposition the allograft; (4) the