Advances in critical care management of hepatic failureand insufficiency

MeiLan King Han, MD; Robert Hyzy, MD

Patients with liver disease repre-sent an important populationwithin in the intensive careunit (ICU) because these pa-

tients experience a particularly high mor-bidity and mortality among the criticallyill. This article addresses advances intherapy and in specific management is-sues related to hepatic dysfunction theintensivist is commonly called on to treatin the ICU, including ascites, hepatorenalsyndrome, hepatic encephalopathy, andfulminant hepatic failure.

Ascites

Cirrhotic patients admitted to themedical ICU have significant mortality,ranging from 40% to 90%. Ascites is themost common complication of cirrhosis

(1) and is suspected based on a history ofabdominal distension, early satiety, short-ness of breath, and by physical examina-tion, which may show “shifting dullness.”However, the sensitivity and specificity ofthe physical examination ranges from 50%to 94% and from 29% to 82%, respectively,when compared with abdominal ultra-sound (2). If ascites is suspected, abdominalultrasound can help not only in confirmingthe diagnosis but also in localizing a site forparacentesis. Ultrasound-assisted abdomi-nal paracentesis yields a higher success ratethan when directed by physical examina-tion alone (95% vs. 61%) (3).

Patients with ascites in the intensivecare setting should undergo a diagnosticparacentesis to rule out infection and toobtain the serum-to-ascites albumin gra-dient. If the ascites is new in onset, im-mediate ultrasound should be obtained torule out acute thrombosis affecting thepatency of the portal and hepatic veins, asseen in acute Budd-Chiari syndrome oracute portal vein thrombosis.

Ascitic fluid analysis is required to de-termine the pathogenesis of ascites. Theserum-to-ascites albumin gradient, cal-culated by subtracting the ascitic fluid

albumin level from the serum albuminlevel, has been shown to be effective indifferentiating portal hypertensive fromnonportal hypertensive ascites (4). A se-rum-to-ascites albumin gradient of �1.1g/dL is seen when portal hypertension ispresent, as with patients who have cirrho-sis, Budd-Chiari syndrome, cardiac disease,portal vein thrombosis, myxedema, or livermetastasis. A serum-to-ascites albumingradient of �1.1 g/dL suggests nonportalhypertensive pathogeneses, including ma-lignancy, pancreatic disease, bile leak, in-fection, or nephrosis.

Ascitic fluid analysis should also in-clude a cell count with differential andculture. In cirrhotic patients, spontane-ous bacterial peritonitis is diagnosedwhen �250 neutrophils/mm3 are foundin the fluid sample (sensitivity, 85%;specificity, 93%; diagnostic accuracy,95%) (5). Studies suggested that reagentstrips may provide a more rapid diagnosisof spontaneous bacterial peritonitis (6),although additional confirmatory studiesare required before widespread accep-tance of this technology (7). Spontaneousbacterial peritonitis has a 1-yr mortalityof 40%, despite treatment with antibiot-

From the Division of Pulmonary and Critical CareMedicine, University of Michigan Health System, AnnArbor, MI.

The authors have not disclosed any potential con-flicts of interest.

Copyright © 2006 by the Society of Critical CareMedicine and Lippincott Williams & Wilkins

DOI: 10.1097/01.CCM.0000231882.85350.71

Background: Chronic liver disease is becoming an increasinglyfrequent diagnosis for patients in the intensive care setting withsuch diagnoses as symptomatic ascites, spontaneous bacterialperitonitis, hepatorenal syndrome, or fulminant hepatic failure.

Objective: To review frequent diagnoses for patients with chronicliver disease admitted to the intensive care unit and discuss currentconcepts in management and investigational modalities.

Results: Patients with new-onset ascites in the intensive caresetting should undergo immediate ultrasound to rule out acutethrombosis. A transjugular intrahepatic portosystemic shunt isindicated when control of the refractory ascites or hepatic hydro-thorax is required. In patients with hepatorenal syndrome, hemo-dialysis can be used as a bridge to liver transplantation. Other-wise, hepatorenal syndrome carries a high mortality. Whenhepatic encephalopathy is present, a precipitating cause shouldbe sought and treated, if identified. Although bioartificial supportsystems are under active investigation, standard treatment for

hepatic encephalopathy is lactulose and alteration of gut flora.Patients with fulminant hepatic failure should be stabilized andtransferred to the intensive care unit of a liver transplant centerand supported with appropriate airway management, close neu-rologic evaluation, glucose monitoring, and correction of coagu-lopathy when there is overt bleeding or an invasive procedure isplanned. Intracranial pressure monitoring is recommended tomaintain an adequate cerebral perfusion pressure of >60 mm Hg.

Conclusion: Review of the literature demonstrates that certaincritically ill patients with chronic liver disease may benefit frominvasive modalities such as transjugular intrahepatic portosys-temic shunting, hemodialysis, and in some cases, liver transplan-tation, which may be offered only at tertiary care centers. (CritCare Med 2006; 34[Suppl.]:S225–S231)

KEY WORDS: liver; cirrhosis; intensive care unit; dialysis; fulmi-nant; hepatic failure; ascites; hepatorenal; spontaneous bacterialperitonitis; transplant

S225Crit Care Med 2006 Vol. 34, No. 9 (Suppl.)

ics (8). A polymicrobial culture raises thesuspicion for intestinal perforation or ab-scess formation. Additional useful testsperformed on ascitic fluid include: glu-cose, which is often elevated in the set-ting of malignancy or gut perforation;amylase, which may be elevated in pan-creatic ascites; and lactate dehydroge-nase, which may be low in cirrhosis butelevated in spontaneous bacterial perito-nitis (9–11). Cytology can be obtainedsubsequently via large-volume paracente-sis, if needed. If there is no obvious causeof ascites, a diagnostic laparoscopic ex-amination may determine whether ma-lignant or infectious peritoneal implanta-tion is present.

The treatment of ascites is directed atthe underlying pathogenesis. Commoncirrhotic ascites can often be managedwith diuretics and sodium restriction.The most successful diuretic regimen is acombination of spironolactone and furo-semide (12). The goal of sodium restric-tion should be to limit intake to 2000mg/day (13). Recent evidence suggestsoctreotide, administered in combinationwith midodrine, may improve both renaland systemic hemodynamics in patientswith ascites (14). Whereas rapid diuresiscan precipitate hepatorenal syndrome andshould be avoided, large-volume paracente-sis (�5 L) has been shown to be safe andeffective, regardless of the cause of as-cites. When performing large-volumeparacentesis in patients with cirrhosis, aninfusion of 6–8 g of albumin per literremoved prevents the development ofparacentesis-induced circulatory dys-function often associated with large fluidshifts (13). Cirrhotics treated with diure-sis should have their electrolytes and kid-ney function closely monitored.

If there are �250 neutrophils/mm3 inan ascitic fluid sample, empirical antibiot-ics should be administered expeditiouslyfor a diagnosis of spontaneous bacterialperitonitis (15). A third-generation cepha-losporin such as cefotaxime is most fre-quently employed in this setting (16, 17).Some authors recommend similar ther-apy in patients with clinical suspicion ofspontaneous bacterial peritonitis, even ifthe ascitic fluid count is �250 neutro-phils/mm3 (15). If spontaneous bacterialperitonitis is diagnosed, albumin infusionand antibiotics have been shown to pre-vent hepatorenal syndrome. In the set-ting of spontaneous bacterial peritonitis,intravenous albumin at 1.5 g/kg of bodyweight at the time of diagnosis, followedby 1 g/kg on day 3, was effective in pre-

venting hepatorenal syndrome in one un-blinded, randomized study (18). Patientswith more advanced liver disease or im-paired renal function may benefit themost (16).

The insertion of a transjugular intra-hepatic portosystemic shunt can be em-ployed for treatment of refractory ascites.A low-resistance channel is created be-tween the hepatic vein and the intrahe-patic portion of the portal vein via angio-graphic placement of a metal stent,allowing portal blood to bypass the liverand return to systemic circulation. Tran-sjugular intrahepatic portosystemicshunt insertion lowers the rate of ascitesrecurrence and the risk of developing he-patorenal syndrome when compared withparacentesis plus albumin administrationin patients with refractory ascites (19).Although the transjugular intrahepaticportosystemic shunt is fairly successfulin the treatment of ascites, a recentmeta-analysis concluded that it can alsobe associated with the development ofincreased encephalopathy and offers nosurvival benefit (20).

The Model for End-stage Liver Disease(MELD) score is a prospectively devel-oped and validated scoring system forchronic liver disease that utilizes a pa-tient’s serum bilirubin, creatinine, andinternational normalized ratio for pro-thrombin time to predict survival. Sev-eral on-line calculators are readily avail-able: http://www.unos.org/resources/MeldPeldCalculator.asp?index�98. AMELD score of �15 is effective at predict-ing a significant risk of hepatic decompen-sation, encephalopathy, and subsequentpoor survival of patients undergoing tran-sjugular intrahepatic portosystemic shuntplacement (21). In general, a MELD scoreof �15 would correspond to a candidatewith a bilirubin at �3 mg/dL, an interna-tional normalized ratio of �2, creatinineof �2 mg/dL, and less than grade II en-cephalopathy. However, when the man-agement of ascites is critical, such as inthe case of ventral hernia rupture or he-patic hydrothorax causing persistent re-spiratory failure, a transjugular intrahe-patic portosystemic shunt can be placedto help control ascites, regardless of theMELD score (22).

Hepatic hydrothorax is usually rightsided, but may be bilateral, and is seenwhen ascitic fluid tracks up into the tho-rax through defects in the diaphragm,potentially causing respiratory embar-rassment. Although usually evident in ad-dition to ascites, it can develop in its

absence. Radionuclide scintigraphy canbe used to confirm the passage of asciticfluid across the diaphragm (23). Treat-ment of hepatic hydrothorax includesusual ascitic care, including salt restric-tion and diuretics. Therapeutic thoracen-tesis with albumin replacement may behelpful. However, tube thoracostomyshould be avoided. Chest tube drainage isoften persistent, making tube removaldifficult and increasing the risk of infec-tion. Transjugular intrahepatic portosys-temic shunt has been shown to be aneffective alternative in the managementof refractory hydrothorax (24).

Hepatorenal Syndrome

Hepatorenal syndrome is the develop-ment of renal failure in a patient withadvanced liver disease. Hepatorenal syn-drome carries a high mortality; therefore,early diagnosis is crucial. The new liverallocation scheme for transplantation pri-oritizes patients with hepatorenal syn-drome. Hemodialysis can be used as abridge to liver transplantation, which offersthe best option for long-term survival.

Hepatorenal syndrome is character-ized by impaired renal function, abnor-malities in the arterial circulation, andactivity of the endogenous vasoactive sys-tem (25). It is the consequence of a re-duction in renal perfusion induced by se-vere hepatic dysfunction. Hepatorenalsyndrome is classified into two types: typeI is more serious and is defined as adoubling of initial serum creatinine to�2.5 mg/dL or a 50% reduction of theinitial 24-hr creatinine clearance to alevel of �20 mL/min in �2 wks. Type IIhepatorenal syndrome does not have arapidly progressive course, displaying aninsidious increase in serum creatinine ora reduction in creatinine clearance overseveral months. The prevalence of hepa-torenal syndrome in patients with end-stage cirrhosis ranges between 7% and15% (26). Predictive factors include so-dium and H2O retention (indicated by aurinary sodium of �5 mEq/L and dilu-tional hyponatremia), low mean arterialblood pressure, poor nutrition, reducedglomerular filtration rate, high plasmarenin activity, and esophageal varices.

Diagnostic criteria for hepatorenal syn-drome established by the International As-cites Club include the following: creati-nine of �1.5 mg/dL or 24-hr creatinineclearance of �40 mL/min; absence ofshock, infection, or fluid losses; no im-provement in renal function after di-

S226 Crit Care Med 2006 Vol. 34, No. 9 (Suppl.)

uretic withdrawal and expansion ofplasma volume with 1.5 L of plasma ex-pander; proteinuria of �500 mg/day; andno evidence of parenchymal or obstruc-tive renal disease (27). Before making thediagnosis, reversible prerenal azotemia,such as secondary to bacterial infectionor drugs such as nonsteroidal anti-inflammatory drugs or aminoglycosides,should be ruled out.

Besides MELD, the severity of chronicliver failure can also be assessed by theChild–Turcotte-Pugh classification sys-tem. The Child–Turcotte-Pugh system isbased on serum bilirubin, serum albu-min, prothrombin time, in addition tomore subjectively assessed variables suchas ascites, and encephalopathy (Table 1).Although the Child–Turcotte-Pugh sys-tem does not correlate with the develop-ment of hepatorenal syndrome, both theChild–Turcotte-Pugh and the MELD scoreshave been found to be predictive of survivalin patients with hepatorenal syndrome (28,29). Nevertheless, survival of patients withhepatorenal syndrome is very poor, with a60% mortality at 2 wks for type I patients.

Patients with hepatorenal syndromeshould be managed by monitoring ofurine output, patient weight, blood pres-sure, evaluation and replacement of elec-trolytes, and the institution of emergentprocedures such as dialysis. Intravenousadministration of clonidine has beenshown to lower renal vascular resistanceand increase the glomerular filtrationrate by as much as 25%, an effect unfor-tunately not able to be sustained withoral therapy (30). Liver transplantationoffers the best treatment, as it resolvescirculatory and renal dysfunction andprovides a 5-yr posttransplantation sur-vival rate of 70% (31). About 5% of pa-tients progress to end-stage renal diseaseafter transplantation and require hemodi-alysis. Unfortunately, few patients with he-patorenal syndrome undergo liver trans-plantation because of the small donor pooland long waiting lists.

The combination of midodrine and oc-treotide may be effective in the treatmentof hepatorenal syndrome by lessening hy-perdynamic circulation without compro-mising glomerular filtration rate, as com-pared with octreotide alone, which doesnot improve systemic hemodynamics andworsens glomerular filtration rate (14).Midodrine is a systemic vasoconstrictor,whereas octreotide inhibits endogenousvasodilator release. In a study of 13 pa-tients with hepatorenal syndrome, threeof the five patients who received mido-

drine (2– 4 �g·kg�1·min�1) and oct-reotide (100 to 200 �g subcutaneouslythree times daily) survived to discharge,whereas seven of the eight patients whoreceived dopamine died (32). Significantimprovements in renal function wereseen in the midodrine/octreotide group,whereas there was a nonsignificant trendtoward renal function deterioration inthe dopamine group. Although this ob-servation awaits confirmation in a largerstudy, given the overall poor prognosis ofthis syndrome, consideration can be givento employing this approach in selected pa-tients with hepatorenal syndrome.

Patients with refractory ascites treatedwith a transjugular intrahepatic portosys-temic shunt may have a lower prevalenceof developing hepatorenal syndrome andbe less likely to progress from type II totype I hepatorenal syndrome (19). Manypatients with hepatorenal syndrome,however, are too ill to undergo tran-sjugular intrahepatic portosystemicshunt placement. Patients with a MELDscore of �18 should probably not un-dergo transjugular intrahepatic portosys-temic shunt placement because these pa-tients have a median predicted survival of�3 months after the procedure (33).

The molecular absorbent recirculatingsystem (MARS), a form of extracorporealalbumin dialysis, has also been proposedas a modality for the treatment of hepa-torenal syndrome. This is an artificialliver support system in which blood isdialyzed against an albumin-enriched di-alysate to facilitate removal of albumin-bound toxins and of bilirubin, aromaticamino acids, and H2O-soluble substances(34). MARS is one of several bioartificialliver support systems that have been orare being developed. Currently, MARS iscommercially available only in Europe. Inthe United States, additional trials arebeing conducted to provide safety data for

obtaining Food and Drug Administrationdevice approval.

A prospective, randomized, controlledtrial of MARS was performed to deter-mine the effect of a MARS on 30-daysurvival in 13 patients with type I hepa-torenal syndrome compared with stan-dard medical treatment (35). The patientstreated with five daily MARS sessions,each lasting 6–8 hrs, displayed signifi-cant decreases in bilirubin and creatininecompared with the nontreatment group.All conventionally treated patients weredead by day 7, but there were two survi-vors at 30 days among eight MARS pa-tients. Larger trials are needed to assessthe efficacy and safety of MARS in hepa-torenal syndrome.

Hepatic Encephalopathy

Hepatic encephalopathy occurs in pa-tients with portal hypertension and cir-rhosis. When severe, hepatic encephalop-athy should be managed in the ICU.Hepatic encephalopathy involves a widerange of neuropsychiatric changes in pa-tients with significant liver dysfunction,ranging from subtle cognitive abnormali-ties to coma (36). Different grades for he-patic encephalopathy are listed in Table 2.There are three clinical patterns for hepaticencephalopathy (37). Type A is related toacute liver failure. Type B occurs in thesetting of normal liver histology and thepresence of a hepatic vascular bypass,such as portocaval shunting. Type C he-patic encephalopathy is due to cirrhosis,entails the majority of cases, and is thetype most commonly seen in the ICUsetting. Type C hepatic encephalopathy isdivided into acute encephalopathy, whichis usually spontaneous and a precipitantis identified, and chronic encephalopathy,which involves a recurrent and fluctuat-ing course.

Table 1. Child-Turcotte-Pugh classification of liver disease severity

Points Assigneda

Parameter 1 2 3Ascites Absent Slight ModerateBilirubin, mg/dL �2 2–3 �3Albumin, g/dL �3.5 2.8–3.5 �2.8Prothrombin time

Seconds over control 1–3 4–6 �6INR �1.7 1.8–2.3 �2.3

Encephalopathy None Grades 1–2 Grades 3–4

INR, international normalized ratio.aA total score of: 5–6 is grade A, 1- and 2-yr patient survival of 100% and 85%, respectively; 7–9

is grade B, 1- and 2-yr patient survival of 80% and 60%, respectively; and 10–15 is grade C, 1- and 2-yrpatient survival of 45% and 35%, respectively.

S227Crit Care Med 2006 Vol. 34, No. 9 (Suppl.)

Diagnosis is usually established basedon a combination of laboratory abnor-malities suggesting severe hepatic dys-function and neurologic deficits. Al-though elevated blood ammonia levelscan be present, they are not required formaking a diagnosis. Early neurologic ab-normalities include disturbance in sleeppatterns such as insomnia or hypersom-nia. Neurologic abnormalities seen inmore advanced presentations include as-terixis and hyperactive deep tendon re-flexes. Focal neurologic signs may be de-tected in some patients during episodesof hepatic encephalopathy, with hemiple-gia being the most common focal deficitseen (38).

The first step in evaluation and manage-ment of these patients is to identify andtreat precipitating factors such as gastroin-testinal bleeding, infection, alkalosis, hypo-kalemia, sedatives/tranquilizers, ingestionof dietary proteins, azotemia, and progres-sive hepatic dysfunction (39). The mainstayof treatment for hepatic encephalopathy islactulose and alteration of gut flora. Lac-tulose, a nonabsorbable disaccharide,should be initiated and titrated to aboutfour bowel movements a day. Lactulose ismetabolized by gut flora, lowering co-lonic pH and thereby favoring ammoniaelimination. Enteric flora modificationwith antibiotics, such as metronidazoleor neomycin, is a second-line treatment,and can be used in combination withlactulose.

Management also includes supportivemeasures such as restoring electrolytebalance, fluid maintenance, aspirationprecautions, and rapid sequence intuba-tion for airway protection in grades 3–4hepatic encephalopathy. A low-proteindiet is required only in patients not im-proving with the above-described mea-sures. The utility of blood ammonia levelsin tracking changes in the depth of he-patic encephalopathy remains controver-sial (40).

Flumazenil has been proposed as apossible therapeutic agent for hepatic en-cephalopathy based on the theory that“endogenous benzodiazepines” may bepresent in patients with hepatic enceph-alopathy (41). In a trial of 560 patientswith hepatic encephalopathy and changesin mental status, intravenous flumazenilimproved mental status in 15% of patients,compared with 3% of placebo-treated con-trol subjects (42). Meta-analyses suggestedthat flumazenil was associated with a sig-nificant improvement in encephalopathycompared with placebo; however, thebenefit was short term and may havebeen confined to patients who otherwisehad a favorable prognosis (43, 44). Alonger-acting intravenous or oral formu-lation is not available. In patients inwhom there is suspected benzodiazepineuse, flumazenil clearly has utility; inother patients, its use is less clear.

Extracorporeal albumin dialysis alsohas been studied for the treatment ofhepatic encephalopathy. Case series haveevaluated this modality in about 60 pa-tients with cirrhosis and hepatic enceph-alopathy (45). Neurologic improvementhas been observed in the majority of pa-tients. More recently, a randomized, con-trolled trial of extracorporeal albumin di-alysis vs. usual supportive care wasperformed in 23 patients with acute-on-chronic liver failure. Bilirubin decreasedand both renal dysfunction and hepaticencephalopathy improved in the treat-ment group. There was also improved30-day survival in the treatment group(46). This suggests extracorporeal albu-min dialysis may serve as an effectivebridge to liver transplantation.

Finally, L-ornithine-L-aspartate ad-ministration has been shown to improveammonia detoxification in several ran-domized trials in patients with hepaticencephalopathy (47). Significant im-provements in neuropsychological test-ing, mental state grade, and portosys-

temic encephalopathy index have beendescribed. L-Ornithine-L-aspartate hasnot been compared with lactulose aloneor in combination and is not presentlyavailable in the United States.

Fulminant Hepatic Failure

Fulminant hepatic failure is a clinicalsyndrome characterized by the rapid on-set of hepatic encephalopathy in conjunc-tion with a marked decline in hepaticsynthetic function. Once a patient is di-agnosed with fulminant hepatic failure,the patient should be stabilized andtransferred to a liver transplant center, asliver transplantation offers the best long-term survival in patients likely to die ofthis condition (48). There the patientshould be cared for in the ICU setting andsupportive measures initiated, includingclose neurologic evaluation and glucosemonitoring.

Temporally related definitions havebeen proposed to classify patients withfulminant hepatic failure as: hyperacute,�7 days; acute, 7–28 days; and subacute,28 days to 6 months (49). The NationalInstitutes of Health Acute Liver FailureStudy Group reported the cause of fulmi-nant hepatic failure in 308 patients asfollows: acetaminophen hepatotoxicity(39%), idiosyncratic drug reaction (13%),hepatitis B (6%), hepatitis A (6%), andindeterminate cause (17%) (50). Overallsurvival is poor without liver transplan-tation, with a reported mortality of 90–97% (51). The advent of liver transplan-tation and aggressive medical care in theICU has improved the mortality rate (52).

Hepatic encephalopathy and severecoagulopathy are important features offulminant hepatic failure (53). Severe co-agulopathy often precedes the evolutionof hepatic encephalopathy to coma. Pa-tients can rapidly progress from mild he-patic encephalopathy to deep coma (Table2) (54). As soon as the diagnosis is made,it is important to establish the cause. Ifthere is no clear cause based on history,urine and serum toxicology screensshould be ordered in addition to hepatitisserologies. Other tests that should beconsidered include ceruloplasmin, anti-nuclear antibodies, smooth-muscle anti-bodies, serum protein electrophoresis,and antibodies to cytomegalovirus andEpstein–Barr virus. Certain pathogenesesdemand immediate specific treatment,including N-acetylcysteine for acetamin-ophen ingestion; penicillin for Amanitamushroom poisoning; delivery of the in-

Table 2. Grading of hepatic encephalopathy

Grade Mental Status Asterixis Electroencephalogram

I Euphoria/depression Yes/no Usually normalMild confusionSlurred speechDisordered sleep

II Lethargy Yes AbnormalModerate confusion

III Marked confusion Yes AbnormalIncoherent, sleeping but arousable

IV Coma No Abnormal

S228 Crit Care Med 2006 Vol. 34, No. 9 (Suppl.)

fant in acute fatty liver of pregnancy; zincand trientine therapy for Wilson’s dis-ease; transjugular intrahepatic portosys-temic shunt, surgical decompression orthrombolysis in patients with acuteBudd-Chiari; and acyclovir in patientswith acute liver failure related to herpes-virus infection.

Management involves supportive mea-sures including nutrition (amino acids, lip-ids, glucose, and essential elements), elec-trolyte balance, frequent glucosemonitoring (more often than every 6hrs), aspiration precautions, and fluidmaintenance. Hypokalemia, hyponatre-mia, and hypophosphatemia are com-mon. Hypoglycemia, seen in up to 45% ofpatients with fulminant hepatic failure,requires aggressive glucose administra-tion, often with 10% dextrose via centralvenous access (55). Infection in patientswith fulminant hepatic failure is a majorsource of mortality, as 44–80% of pa-tients with fulminant hepatic failure de-velop bacterial infections. Empirical,broad-spectrum antibiotics should be ini-tiated on clinical suspicion of infection(56). Fungal infections are also not un-common in these patients, with rates ashigh as 32% having been reported (57).Acute renal failure frequently develops infulminant hepatic failure. Renal failure isparticularly high in the setting of acet-aminophen ingestion, as it can directlydamage the kidneys. Once renal failure isestablished, it often is irreversible andcarries a grave prognosis. Renal replace-ment therapy is generally well toleratedand may provide a bridge to transplant (58).

The development of severe coagulopa-thy is due to the decreased synthesis ofclotting factors II, V, VII, and IX and ismanifested by a prolonged prothrombintime. However, current recommenda-tions are to correct coagulopathy withfresh frozen plasma intravenously onlywhen overt bleeding occurs or when aninvasive procedure is planned. Recombi-nant factor VIIa has been shown to be safeand effective in reversing the coagulopa-thy in patients with fulminant hepaticfailure (59). The protocol is to infuse 80�g/kg after infusion of 4 units of freshfrozen plasma. This can normalize pro-thrombin time for up to 6 hrs.

Neurologic evaluation, a critical guideto therapy, should be performed at leastevery 6 hrs. Medications with sedativeproperties should be avoided if possible.Patients with chronic liver failure shouldbe continued on lactulose. The use oflactulose for acute liver failure differs

widely in ICUs throughout the country,but a review of 23 liver transplant centersacross the United States suggested that lac-tulose provided only a short-term survivaladvantage in this patient population (60).

Cerebral edema is a common compli-cation of fulminant hepatic failure, oc-curring in up to 80% of patients withgrade IV coma, but requires a high levelof clinical suspicion. An emergent headcomputed tomographic scan should beperformed if there is a change in mentalstatus or signs of increased intracranialpressure. The diagnosis may be difficultto establish as head computed tomo-graphic scan is insensitive, being usefulonly to rule out hemorrhage. Clinicalsigns of cerebral edema, such as decere-brate posturing, systemic hypertension,and pupillary abnormalities, are unreli-able and should not be used for clinicaldecision making. Cerebral edema oftenleads to intracranial hypertension andsubsequent herniation of the cerebral un-cus, cerebral ischemic injury, and death(61). Intracranial hypertension can alsocause a reduction in the cerebral perfu-sion pressure (mean arterial pressure mi-nus intracranial pressure), which mayproduce cerebral ischemia. A cerebralperfusion pressure of �60 mm Hg is cru-cial to maintain intact neurologic func-tion (62). Direct intracranial pressuremonitoring is recommended in patientssuspected of cerebral edema or intracra-nial hypertension, with a target intracra-nial pressure of �20 mm Hg (63).

Intracranial pressure monitoring is rec-ommended to maintain an adequate cere-bral perfusion pressure of �60 mm Hg.The placement of extradural intracranialpressure monitors is considered safer thansubdural catheters. Recombinant factorVIIa may be superior to fresh frozen plasmain temporarily reversing coagulopathy inthose patients requiring intracranial pres-sure monitor placement (64). In general,sedation should be avoided so that mental

status may be assessed. However, an agi-tated patient with grade III coma mayrequire the use of short-acting benzodi-azepines, the preferred agents (65). Al-though hyperventilation may also reducecerebral edema, it is effective only for afew hours. Mannitol is first-line therapyfor treating cerebral edema and intracra-nial hypertension, administered at 0.3–0.4 g/kg body weight. In patients withrenal failure, mannitol may accumulatein astrocytes and cause increased re-bound swelling (66). Thiopental may beused in this setting (250 mg over 15mins). In one case series, seven patientswith fulminant hepatic failure were ad-ministered propofol for deteriorating ce-rebral edema (67). Decreased intracranialpressure was seen in five patients, butonly three survived, with one undergoingliver transplantation. Additional studiesare needed before a recommendation touse propofol in this setting can be made.Another therapeutic adjunct, moderatehypothermia to 32–33°C, may be usefulin decreasing intracranial pressure as abridge to liver transplantation (68) orwhile transplantation is being per-formed (69).

Liver transplantation offers the bestlong-term survival, with an overall post-transplantation 1-yr survival of about60% (70). Unfortunately, prediction ofthe need for transplantation remainsproblematic. The King’s College Hospitalcriteria are the most widely used prog-nostic indicator for survival in fulminanthepatic failure (Table 3) (71). These cri-teria include an arterial pH of �7.30 afteradequate fluid resuscitation or the com-bination of a prothrombin time of �100secs, a creatinine level of �3.3 mg/dL,and grade III or IV encephalopathy. Thesecriteria exhibit a sensitivity, specificity,positive predictive value, and negativepredictive value of 55%, 94%, 87%, and78%, respectively. A meta-analysis inves-tigated several prognostic criteria for de-

Table 3. King’s College criteria for liver transplantation in acute liver failure

Acetaminophen Nonacetaminophen

Arterial lactate �3.5, 4 hrs after resuscitation INR of �6.5 (PT of �100 secs) or any threeof the following:

Or INR of �3.5 (PT of �50 secs)pH of �7.3 or arterial lactate of �3.0, 12 hrs

after resuscitationAge of �10 or �40 yrs

Or Serum bilirubin of �17.5 mg/dLINR of �6.5 (PT of �100 secs) Duration of jaundice of �7 daysSerum creatinine of �3.4 mg/dL Pathogenesis: drug reaction

INR, international normalized ratio; PT, prothrombin time.

S229Crit Care Med 2006 Vol. 34, No. 9 (Suppl.)

termining the need for liver transplant inacetaminophen-induced fulminant he-patic failure, including King’s Collegecriteria, pH, prothrombin time, factor Vlevels, and creatinine, but found thatnone of these was sufficiently sensitive topredict the need for liver transplantation(72). Arterial blood lactate of �3.5mmol/L at 4 hrs after presentation to thehospital has been shown to have a sensi-tivity of 67%, a specificity of 95%, a pos-itive likelihood ratio of 13%, and a nega-tive likelihood ratio of 35% for survival inacetaminophen-induced fulminant he-patic failure (73).

Because only a relatively small portionof liver is actually required to supporthepatic function, another potential ther-apeutic alternative that is being investi-gated is auxiliary liver transplantation. Inthis technique, a partial liver graft isplaced either adjacent to the patient’s na-tive liver or in the hepatic bed after aportion of the native liver has been re-moved. Theoretically, this graft may sup-port the patient while the native liverregenerates so that ultimately the patientwould not need chronic immunosuppres-sion. Although case reports and smallcase series have reported success usingthis technique, the procedure is techni-cally very difficult and has not been ade-quately evaluated in controlled trials (74).

Short-term extracorporeal hepatic sup-port for patients with fulminant hepaticfailure may ultimately serve to improveoverall survival and provide support as abridge to liver transplantation, but it re-mains experimental. Two types of systemsare being investigated: cell- and non–cell-based systems. Extracorporeal albumindialysis, such as MARS, is an example ofnon–cell-based systems. Whereas non–cell-based systems aim to adsorb toxinsfrom the patient’s blood, cell-based sys-tems also are designed to provide hepaticsynthetic support. A meta-analysis of ar-tificial and bioartificial support systemsfor fulminant hepatic failure examined atotal of eight randomized controlled tri-als, involving 139 patients, and found noimprovement in mortality compared withstandard supportive care (relative risk,0.95; 95% confidence interval, 0.71–1.29)(75). In addition, the interventions werenot found to be useful as a bridge to livertransplantation (relative risk, 0.60; 95%confidence interval, 0.29–1.23). The sup-port systems seemed to have an increasedrisk of bleeding associated with their use.However, a meta-regression suggestedthat patients with acute-on-chronic liver

failure experienced a 33% reduction inmortality as opposed to those with simplyacute liver failure. Another future optionincludes hepatocyte transplantation. Inone series, three of six patients with ful-minant hepatic failure survived between14 and 52 days after transplantation of1010 human hepatocytes (76). To date, norandomized, controlled studies have ex-amined this therapeutic option.

REFERENCES

1. Gines P, Quintero E, Arroyo V, et al: Com-pensated cirrhosis: Natural history and prog-nostic factors. Hepatology 1987; 7:122–128

2. Cattau EL, Benjamin SB, Knuff TE, et al: Theaccuracy of the physical examination in thediagnosis of suspected ascites. JAMA 1982;247:1164–1166

3. Nzeer SR, Dewbre H, Miller AH: Ultrasound-assisted paracentesis performed by emer-gency physicians vs the traditional tech-nique: A prospective, randomized study.Am J Emerg Med 2005; 23:363–367

4. Pare P, Talbot J, Hoefs JC: Serum-ascitesalbumin concentration gradient: A physio-logic approach to the differential diagnosis ofascites. Gastroenterology 1983; 85:240–244

5. Garcia-Tsao G: Spontaneous bacterial perito-nitis. Gastroenterol Clin North Am 1992; 21:257–275

6. Castellote J, Lopez C, Gornals J, et al: Rapiddiagnosis of spontaneous bacterial peritonitisby use of reagent strips. Hepatology 2003;37:893–896

7. Runyon BA: Strips and tubes: Improving thediagnosis of spontaneous bacterial peritoni-tis. Hepatology 2003; 36:745–747

8. Thuluvath PJ, Morss S, Thompson R: Spon-taneous bacterial peritonitis-in-hospitalmortality, predictors of survival, and healthcare costs from 1988 to 1998. Am J Gastro-enterol 2001; 96:1232–1236

9. Runyon BA, Hoefs JC, Morgan TR: Asciticfluid analysis in malignancy-related ascites.Hepatology 1988; 8:1104–1109

10. Runyon BA: Amylase levels in ascitic fluid.J Clin Gastroenterol 1987; 9:172–174

11. Runyon BA, Hoefs JC: Ascitic fluid analysis inthe differentiation of spontaneous bacterialperitonitis from gastrointestinal tract perfo-ration into ascetic fluid. Hepatology 1984;4:447–450

12. Fogel, MR, Sawhney VK, Neal EA, et al: Di-uresis in the ascitic patient: A randomizedcontrolled trial of three regimens. J ClinGastroenterol 1981; 3(Suppl 1):73–80

13. Runyon BA: Management of adult patientswith ascites caused by cirrhosis. Hepatology1998; 27:264–272

14. Kalambokis G, Economou M, Fotopoulos A,et al: The effects of chronic treatment withoctreotide versus octreotide plus milrinoneon systemic hemodynamics and renal hemo-dynamics and function in nonazotemic cir-

rhotic patients with ascites. Am J Gastroen-terol 2005; 100:879–885

15. Runyon BA: The evolution of ascitic fluidanalysis in the diagnosis of spontaneous bac-terial peritonitis. Am J Gastroenterol 2003;98:1675–1677

16. Mowat C, Stanley AJ: Spontaneous bacterialperitonitis: Diagnosis, treatment and preven-tion [review]. Aliment Pharmacol Ther 2001;15:1851–1859

17. Soares-Weiser K, Paul M, Brezis M, et al:Antibiotic treatment for spontaneous bacte-rial peritonitis. BMJ 2002; 324:100–103

18. Sort P, Navasa M, Arroyo V, et al: Effect ofintravenous albumin on renal impairmentand mortality in patients with cirrhosis andspontaneous bacterial peritonitis. N EnglJ Med 1999; 341:403–409

19. Gines P, Uriz J, Calahorra B, et al: Transjugu-lar intrahepatic portosystemic shunting ver-sus paracentesis plus albumin for refractoryascites in cirrhosis. Gastroenterology 2002;123:1839–1847

20. Deltenre P, Mathurin P, Dharancy S, et al:Transjugular intrahepatic portosystemic shuntin refractory ascites: A meta-analysis. Liver Int2005; 25:349–356

21. Malichok M, Kamath PS, Gordon FD, et al: Amodel to predict poor survival in patientsundergoing TIPS. Hepatology 2000; 31:864–871

22. Peron JM, Barange K, Otal P, et al: Tran-sjugular intrahepatic portosystemic shuntsin the treatment of refractory ascites: Resultsin 48 consecutive patients. J Vasc Interv Ra-diol 2000; 11:1211–1216

23. Bhattacharya A, Mittal BR, Biswas T, et al:Radioisotope scintigraphy in the diagnosis ofhepatic hydrothorax. J Gastroenterol Hepa-tol 2001; 16:317–321

24. Spencer EB, Cohen DT, Darcy MD: Safetyand efficacy of transjugular intrahepatic por-tosystemic shunt creation for the treatmentof hepatic hydrothorax. J Vasc Interv Radiol2002; 13:385–390

25. Arroyo V, Gines P, Gerbes AL, et al: Defini-tion and diagnostic criteria of refractory as-cites and hepatorenal syndrome. Hepatology1996; 23:164–176

26. Schrier RW, Arroyo V, Bernardi M, et al:Peripheral arterial vasodilation hypothesis: Aproposal for the initiation of renal sodiumand water retention in cirrhosis. Hepatology1988; 8:1151–1170

27. Cardenas A, Uriz J, Gines P, et al: Hepatorenalsyndrome. Liver Transpl 2000; 6:S63–S71

28. Gines A, Escorsell A, Gines P, et al: Inci-dence, predictive factors, and prognosis ofthe hepatorenal syndrome in cirrhosis withascites. Gastroenterology 1993; 105:229–236

29. Witzke O, Baumann M, Patschan D, et al:Which patients benefit from hemodialysistherapy in hepatorenal syndrome? J Gastro-enterol Hepatol 2004; 19:1369–1373

30. Esler M, Dudly F, Jennings G, et al: Increasedsympathetic nervous activity and the effects ofits inhibition with clonidine in alcoholic cir-rhosis. Ann Intern Med 1992; 116:446–455

S230 Crit Care Med 2006 Vol. 34, No. 9 (Suppl.)

31. Nair S, Verma S, Thuluvath PJ: Pretransplantrenal function predicts survival in patientsundergoing orthotopic liver transplantation.Hepatology 2002; 35:1179–1185

32. Angeli P, Volpin R, Gerunda G, et al: Reversalof type 1 hepatorenal syndrome with the ad-ministration of midodrine and octreotide.Hepatology 1999; 29:1690–1697

33. Malinchoc M, Kamath PS, Gordon FD, et al:Hepatology: A model to predict poor survivalin patients undergoing transjugular intrahe-patic portosystemic shunts. Hepatology2000; 31:864–871

34. Solanki P, Stange J, Hassaneim TI, et al: Themolecular adsorbents recycling system as aliver support system based on albumin dial-ysis: A summary of preclinical investigations,prospective randomized controlled clinicaltrial and clinical experience from 19 centers.Artif Organs 2001; 26:103–110

35. Mitzner SR, Stange J, Klammt S, et al: Im-provement of hepatorenal syndrome with ex-tracorporeal albumin dialysis MARS: Resultsof a prospective, randomized, controlled clin-ical trial. Liver Transpl 2000; 6:277–286

36. Butterworth RF: Complications of cirrhosisIII: Hepatic encephalopathy. J Hepatol 2000;32:171–180

37. Ferenci P, Lockwood A, Mullen K, et al: He-patic encephalopathy: Definition, nomencla-ture, diagnosis, and quantification. Final re-port of the working party at the 11th WorldCongresses of Gastroenterology, Vienna,1998. Hepatology 2002; 35:716–721

38. Cadranel JF, Lebiez E, Di Martino V, et al:Focal neurological signs in hepatic encepha-lopathy in cirrhotic patients: An underesti-mated entity? Am J Gastroenterol 2001; 96:515–518

39. Blei AT, Cordoba J: Hepatic encephalopathy.Am J Gastroenterol 2001; 96:1968–1976

40. Ong JP, Aggarwal A, Krieger D, et al: Corre-lation between ammonia levels and the se-verity of hepatic encephalopathy. Am J Med2003; 114:188–193

41. Basile AS, Hughes RD, Harrison PM, et al:Elevated brain concentrations of 1,4-benzo-diazepines in fulminant hepatic failure.N Engl J Med 1991; 325:473–478

42. Barbaro G, Di Lorenzo G, Soldini M, et al:Flumazenil for hepatic encephalopathyGrade III and IVa in patients with cirrhosis:An Italian multicenter double-blind, placebo-controlled, cross-over study. Hepatology1998; 28:374–378

43. Als-Nielsen B, Kjaergard LL, Gluud C: Ben-zodiazepine receptor antagonists for acuteand chronic hepatic encephalopathy. Co-chrane Database Syst Rev 2001; 4:CD002798

44. Goulenok C, Bernard B, Cadranel JF, et al:Flumazenil vs placebo in hepatic encepha-lopathy in patients with cirrhosis: A meta-analysis. Aliment Pharmacol Ther 2002; 16:361–372

45. Stange J, Mitzner SR, Klammt S, et al: Liversupport by extracorporeal blood purification:A clinical observation. Liver Transpl 2000;6:603–613

46. Heemann U, Treichel U, Loock J, et al: Albu-min dialysis in cirrhosis with superimposedacute liver injury: A prospective, controlledstudy. Hepatology 2002; 36:949–958

47. Kircheis G, Wettstein M, Dahl S, et al: Clin-ical efficacy of L-ornithine-L-aspartate in themanagement of hepatic encephalopathy.Metab Brain Dis 2002; 17:453–462

48. Bismuth H, Samuel D, Castaing D, et al:Liver transplantation in Europe for patientswith acute liver failure. Semin Liver Dis1996; 16:415–425

49. Ostapowicz G, Lee WM: Acute hepatic failure: AWestern perspective. J Gastroenterol Hepatol2000; 15:480–488

50. Ostapowicz G, Fontana RJ, Schiodt FV, et al:Results of a prospective study of acute liverfailure at 17 tertiary care centers in theUnited States. Ann Intern Med 2002; 137:947–954

51. Mas A, Rodes J: Fulminant hepatic failure.Lancet 1997; 349:1081–1085

52. Hoofnagle JH, Carithers RL Jr, Shapiro C, etal: Fulminant hepatic failure: Summary of aworkshop. Hepatology 1995; 21:240–252

53. Blei AT, Larsen FS: Pathophysiology of cere-bral edema in fulminant hepatic failure.J Hepatol 1999; 31:771–776

54. Wijdicks EF, Plevak DJ, Rakela J, et al: Clin-ical and radiologic features of cerebral edemain fulminant hepatic failure. Mayo Clin Proc1995; 70:119–124

55. Shakil AO, Mazariegos GV, Kramer DJ: Livertransplantation: Current management, ful-minant hepatic failure. Surg Clin North Am1999; 79:77–108

56. Fontana RJ, Quallich L: Acute liver failure.Curr Opin Gastroenterol 2001; 17:291–298

57. Rolando N, Harvey F, Brahm J, et al: Fungalinfection: A common, unrecognized compli-cation of acute liver failure. J Hepatol 1991;12:1

58. Agarwal R, Farber MO: Is continuous veno-venous hemofiltration for acetaminophen-induced acute liver and renal failure worth-while? Clin Nephrol 2002; 57:167–170

59. Shami VM, Caldwell SH, Hespenheide EE, etal: Recombinant activated factor VII for co-agulopathy in fulminant hepatic failure com-pared with conventional therapy. LiverTranspl 2003; 9:138–143

60. Lee WM: Acute liver failure in the UnitedStates. Semin Liver Dis 2003; 23:271–282

61. Munoz SJ, Moritz MJ, Martin P, et al: Rela-tionship between cerebral perfusion pressureand systemic hemodynamics in fulminanthepatic failure. Transplant Proc 1993; 25:1776–1778

62. Lidofsky SD, Bass NM, Prager MC, et al:Intracranial pressure monitoring and liver

transplantation for fulminant hepatic failure.Hepatology 1992; 16:1–7

63. Daas M, Plevak DJ, Wijdicks EF, et al: Acuteliver failure: results of a 5-year clinical pro-tocol. Liver Transpl Surg 1995; 1:210–219

64. Kalicinski P, Kaminski A, Drewniak T, et al:Quick correction of hemostasis in two pa-tients with fulminant liver failure undergo-ing liver transplantation by recombinant ac-tivated factor VII. Transplant Proc 1999; 31:378–379

65. Shakil AO, Kramer D, Mazariegos GV, et al:Acute liver failure: Clinical features, outcomeanalysis, and applicability of prognostic cri-teria. Liver Transpl 2000; 6:163–169

66. Canalese J, Gimson AE, Davis C, et al: Con-trolled trial of dexamethasone and mannitolfor the cerebral oedema of fulminant hepaticfailure. Gut 1982; 23:625–629

67. Wijdicks EF, Nyberg SL: Propofol to controlintracranial pressure in fulminant hepaticfailure. Transplant Proc 2002; 34:1220–1222

68. Jalan R, Damink SW, Deutz NE, et al: Mod-erate hypothermia for uncontrolled intracra-nial hypertension in acute liver failure. Lan-cet 1999; 354:1164–1168

69. Jalan R, Olde Damink SW, Deutz NE, et al:Moderate hypothermia prevents cerebral hy-peremia and increase in intracranial pressurein patients undergoing liver transplantationfor acute liver failure. Transplantation 2003;75:2034–2039

70. McCashland TM, Shaw BW Jr, Tape E: TheAmerican experience with transplantationfor acute liver failure. Semin Liver Dis 1996;16:427–433

71. O’Grady JG, Alexander GJ, Hayllar KM, et al:Early indicators of prognosis in fulminanthepatic failure. Gastroenterology 1989; 97:439–445

72. Bailey B, Amre DK, Gaudrealt P: Fulminanthepatic failure secondary to acetaminophenpoisoning: A systematic review and meta-analysis of prognostic criteria determiningthe need for liver transplantation. Crit CareMed 2003; 31:299–305

73. Bernal W, Donaldson N, Wyncoll D, et al:Blood lactate as an early predictor of out-come in paracetamol-induced acute liver fail-ure: A cohort study. Lancet 2002; 359:558–563

74. Lee DS, Woo JG, Lee HH, et al: Auxiliarypartial orthotopic liver transplantation in thetreatment of acute liver failure: A case report.Transplant Proc 2004; 36:2228–2229

75. Kjaergard LL, Liu J, Als-Nielsen B, et al:Artificial and bioartificial support systems foracute and acute-on-chronic liver failure.JAMA 2003; 289:217–222

76. Bilir BM, Guinette D, Karrer F, et al: Hepa-tocyte transplantation in acute liver failure.Liver Transpl 2000; 6:32–40

S231Crit Care Med 2006 Vol. 34, No. 9 (Suppl.)