Gut Supplement, 1991, S86-S91

Acute liver failure

R D Hughes, J Wendon, A E S Gimson

Institute of Liver Studies,King's College Hospitaland School ofMedicineand Dentistry, LondonR D HughesJ WendonA E S GimsonCorrespondence to:Dr Robin Hughes,Institute of Liver Studies,King's College School ofMedicine and Dentistry,Bessemer Road, London SE5.

80-

70

60-

iF 50

c8 40

ol

30

20

10

Acute liver failure is a rare disorder with hepaticencephalopathy occurring within eight weeks ofthe onset ofsymptoms, without previous chronicliver disease, it has a high mortality of up to80% in many series. In 1973 the Liver FailureUnit, the first of its kind, was built at King's asa specific intensive care facility to treat suchcritically ill patients. Since then over 1000 casesof fulminant hepatic failure have been referredand the Unit has acquired an internationalreputation as a centre for developing treatmentstrategies and research. This has been facilitatedby two Medical Research Council ProgrammeGrants and more recently by recognition fornational funding as a specialised SupraregionalService.

In 1973 the survival from fulminant hepaticfailure with grade 4 encephalopathy was about20%, but as is shown in the Figure, since thena steady improvement has been observed withsurvival rates of over 50% in some aetiologicalgroups. This chapter will highlight some of themajor advances that have appeared from themore than 100 published papers that have beenproduced from the liver failure research groupand that have led to the improved survivalfigures. It is, of course, not possible to acknow-ledge the efforts of all the research and clinicalstaff involved in this over the years in this shortreview. For a comprehensive account up to 1986see ref 1.A major milestone in liver failure research at

King's was the International Symposium on

gParacetamol

a Non-A, non-B hepatitis

Drug/halothane hepatitis

III

Improvements in survival, according to aetiology, ofpatients withfulminant hepatic failure(grades III and IVencephalopathy) admitted to the Liver Failure Unit at King's CollegeHospitalfrom 1973 to 1990. Thefigure includes patients who underwent liver transplantation.

Artificial Liver Support for Acute Liver Failureheld in 1974, where many of the problemsassociated with this condition and its manage-ment were identified. Recently, in September1990, the 11th BSG/SK & F International Work-shop was held on Acute Liver Failure:Improved Understanding and Better Therapy.A small group of people all working in this fieldmet to review the progress made to date, some ofwhich is included in this chapter and also thechapter on liver transplantation in this volume.There was no doubt considerable progress hasbeen made.

Aetiology and pathogenesisParacetamol hepatotoxicity was recognised earlyto be an important cause of fulminant hepaticfailure2 and remains the most common cause inthe United Kingdom. Paracetamol is metabol-ised by the cytochrome P450 system with pro-duction of an unstable reactive metabolite N-acetyl-p-benzoquinoneimine which is rapidly in-activated by conjugation with glutathione. Inoverdosage, as glutathione stores are depletedaccumulation of the active metabolite causes celldamage. This cell damage can be prevented bysulphydryl donors such as cysteamine, meth-ionine, and N-acetylcysteine. The latter hasusually been recommended only if it can be givenwithin 16 hours of the overdose, but a retrospec-tive survey showed significant benefit even ifgiven at up to 48 hours. Fewer patients pro-gressed to grade 4 encephalopathy and themortality was lower in those cases receiving lateN-acetylcysteine administration compared withother cases presenting at a similar time after theoverdose and with similar prolongation of theprothrombin time.3 More recently preliminaryanalysis of a randomised clinical trial of late N-acetylcysteine administration has also shown alower mortality in the treated cases. Improve-ments in survival with the later use of N-acetylcysteine, when most ofthe paracetamol hasbeen cleared, probably results from mechanismsunconnected with the direct binding of N-acetylcysteine to the reactive metabolite, as whengiven early.

Viral hepatitis is the most common cause offulminant hepatic failure worldwide with ful-minant hepatitis B virus the most prevalentvariety in many areas. A fulminant course ismore common in the presence of coinfection ofhepatitis D with hepatitis B virus, but may alsobe related to an enhanced antibody response tothe virus. Compared with cases of uncomplica-ted acute viral hepatitis, those with a fulminantcourse have a higher antiHBc titre and morerapid clearance of HBsAg from serum whichmay lead to antigen-antibody complex deposi-

S86

on Novem

ber 26, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.32.Suppl.S

86 on 1 Septem

ber 1991. Dow

nloaded from

Acute liverfailure

tion in hepatic sinusoids and ischaemic necrosisof hepatocytes. Pathogenetic mechanisms offulminant hepatitis A and non-A, non-B areunknown but significant differences occur in theclinical course and complication rates betweenthe three types.4 Thus fulminant viral hepatitis Ahas a shorter duration before the onset of en-cephalopathy, a lower peak prothrombin timeand cerebral oedema is less common comparedwith presumed non-A, non-B where the patientstend to be older, with a slower onset of encephal-opathy and a worse mortality.

Idiosyncratic drug reactions are an importantcause of fulminant hepatic failure and have beenrecorded after a wide range of drugs includingisoniazid, non-steroidal anti-inflammatorydrugs, monoamine oxidase inhibitors, valproateand ketoconazole, but the most commonoccurring in up to 5% of all cases referred tothe Liver Failure Unit follows halothane anaes-thesia. Liver failure may develop up to 21 daysafter halothane exposure often in obese womenwith a history of atopy and previous minorpyrexial reactions to anaesthetics. Hepato-toxicity may be immunologically mediated asantibodies to halothane exposed rabbit hepato-cytes are found in up to 80% of cases.5

Coagulation abnormalitiesStudies on the severe abnormalities in coagula-tion in relation to bleeding stem from the oftenreferred to seminal work of Rake et al onintravascular coagulation in liver failure.6 In-creased susceptibility to bleeding in acute liverfailure results from both consumption and lackof synthesis of clotting factors and their inhibi-tors, as well as a low platelet count. Originallya number of trials to replace depleted clottingfactors using fresh frozen plasma and factor IXconcentrates to reduce bleeding and heparin tostop intravascular coagulation were carried outby Clark, Gazzard and others with little overallsuccess. Further studies were stimulated as aresult of the problems with biocompatibility ofartificial liver support. In these Weston et alinvestigated platelet function and found thatplatelet aggregation to classical stimuli was im-paired and platelet ultrastructure was abnormal.Platelet stickiness was found to be increasedconsistent with activated platelets remaining inthe circulation because oflack ofclearance by thedamaged liver. Factor VIII is involved in plateletadhesion to foreign surfaces and is the onlyclotting factor found at increased concentrationsin fulminant hepatic failure, probably as itoriginates from vascular endothelial cells. Allthree components of factor VIII were found byLangley et al7 to be at concentrations over threetimes normal in liver failure, suggesting involve-ment of factor VIII in the platelet reactions.Difficulties of anticoagulation during extracor-poreal circulation in these patients led Sette et alto study heparin clearance and response in liverfailure and it was found that there was morerapid elimination of heparin and increasedheparin sensitivity which helps explain the basisof this problem. Further confirmation of on-going activation of the coagulation and plateletsystems was obtained when highly sensitive

assays became available; fibrinopeptide A forthrombin action, fibrinopeptide BP3 1-42 forplasmin action, ,3-thromboglobulin and plateletfactor 4 for platelet activation. Plasma levels ofall these were higher than normal, but not to thedegree seen in true disseminated intravascularcoagulation.8

Recent studies by Langley et al have concen-trated on the role of coagulation inhibitors in theconsumptive coagulopathy of liver failure. Anti-thrombin III, is the major inhibitor of coagula-tion and was found to occur at very low levels inpatients with acute liver failure. Two controlledclinical trials of replacement therapy have beencarried out with a heat treated antithrombin IIIconcentrate produced by Behring. As antithrom-bin III is a heparin cofactor, we evaluatedsupplementation during haemodialysis in liverfailure. Antithrombin III levels were normal-ised, less heparin was administered and losses ofplatelets were significantly reduced. In the othertrial, the effect of antithrombin III supplement-ation was assessed in patients with septic shockassociated with liver failure. The rationale forthis being that intravascular coagulation is mostsevere in this group of patients and could under-lie multiple organ failure. Despite correction ofantithrombin III levels the high mortality in thisgroup was not improved and it was consideredthat earlier therapy might have been moresuccessful. Protein C is a potent inhibitor offactor V and VIII and as with other clottingfactors was found at low levels in fulminanthepatic failure, suggesting that supplementationmight be worthy of study.The relationship between infection and the

coagulation derangement was investigatedfurther from the plasma levels of leucocyteelastase measured by the presence of elastasea1-antitrypsin complexes. Greater release ofelastase was found in liver failure patients whohad bacterial infection and this correlated withreduced plasma fibrinogen and increasedthrombin-antithrombin complexes another sen-sitive marker of thrombin formation. Thus it islikely that infection with the associated cytokinerelease is interrelated with activation of thecoagulation system.

Artificial liver supportEarly attempts at liver support at King'sincluded exchange transfusion and pig liverperfusion, which were impractical to perform ona regular basis. The concept of artificial liversupport based on haemoperfusion throughadsorbents to remove the wide range of toxinsaccumulating in the circulation is still attractive.Experimental work by Wilson, Dunlop andothers led to the clinical use of charcoal haemo-perfusion, as pioneered by TMS Chang inCanada. A column of polymer-coated activatedcharcoal to remove water soluble toxins wasdeveloped in association with Smith andNephew Research. The exciting results in thefirst patients were reported in the Lancet byGazzard et al,9 with 10 of the 22 patients in gradeIV encephalopathy treated surviving. Furtherpatients had severe hypotensive episodes, how-ever, which were shown by Weston et al to be

S87

on Novem

ber 26, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.32.Suppl.S

86 on 1 Septem

ber 1991. Dow

nloaded from

Hughes, Wendon, Gimson

related to the formation of platelet aggregates inthe extracorporeal circuit. Alternative means ofpatient treatment were needed and hence the useof polyacrylonitrile dialysis with a high per-meability membrane. After laboratory studies onagents to inhibit the platelet reactions; charcoalhaemoperfusion with prostacyclin (PGI2) wasstarted in 1980. Subsequently Gimson et al'0reported results in 76 patients, with a survivalrate of up to 65% in patients treated early, whenstill in grade III coma. There had been consider-able debate about the question of performing aclinical trial of charcoal haemoperfusion over theyears, original plans were shelved due to theproblems of blood compatibility. This wasfinally resolved when two clinical trials wereperformed, five hours versus 10 hours haemo-perfusion in patients with grade III and 10 hourshaemoperfusion versus intensive care alone inpatients with grade IV encephalopathy. Nosignificant differences were found in these twogroups in 137 patients entered. Overall survivalwas good and was related to aetiology, being bestin paracetamol overdose and viral hepatitis Aand B and worst in non-A, non-B hepatitis andadverse reactions to drugs." It was consideredthat these results were the result of the improv-ing background of intensive care managementwhich masked any earlier benefits of charcoalhaemoperfusion.

Together with the original work on charcoalthere was interest in an adsorbent to remove lipidsoluble toxins. A resin column was developed inthe Unit based on Amberlite XAD-7 coated withhuman serum albumin to confer good bloodcompatibility, which was confirmed in in vitroexperiments. A successful pilot study wascarried out in patients and larger scale produc-tion of the columns by Immuno AG in Viennaallowed evaluation of serial resin haemoperfu-sion in 19 patients with fulminant hepaticfailure.'2 Blood compatibility was confirmed andremoval of bilirubin, bile acids and substances inthe middle molecular weight range (1000-5000)demonstrated. A larger column was produced toincrease removal, but with some loss of bio-compatibility. Unfortunately, at this point com-mercial development had to be stopped. We arestill interested in artificial liver support forremoval of specific substances using plasmarather than whole blood perfusion to try andavoid blood cell activation.A number of laboratory studies on the toxins

accumulating in liver failure were carried out atthe same time as the haemoperfusion work.Fulminant hepatic failure serum was shownto contain substances inhibitory to membranesodium transport both in leucocytes by Alam andPoston'3 and in brain by Seda et al. '4 Membraneultrafiltration (<10 kD) followed by columnchromatography on Sephadex G-25 producedfour fractions which contained the inhibitoryactivity. Resin and charcoal haemoperfusionwere shown to be complementary in removingthese Na+,K+-ATPase inhibitors. Subsequentlysome of the peaks were shown to crossreactwith antibodies to digoxin, an inhibitor ofNa+,K+-ATPase, making this radioimmuno-assay useful for monitoring toxin removal inartificial liver support.'5

Cerebral oedemaThe mechanisms involved in the development ofhepatic encephalopathy and the associatedcerebral oedema in acute liver failure remain illdefined. Cerebral oedema remains the mostcommon cause of death in fulminant hepaticfailure as originally reported from the Unit,'6although there has been considerable progress inits management. The importance of this compli-cation was first shown in post mortem studiesand in an animal model of liver failure, but in1980 Hanid et al reported the first use ofan intra-cranial pressure monitor in patients with fulmin-ant hepatic failure. Dramatic rises in intracranialpressure were observed, often spontaneous butoccasionally precipitated by external events,which importantly preceeded clinical evidenceof cerebral oedema (opisthotonus, pupillaryabnormalities, respiratory arrest).

Rises in intracranial pressure occur in up to80% of cases of fulminant hepatic failure andmanagement rests on maximising cerebral per-fusion pressure and cerebral blood flow, both ofwhich tend to fall as intracranial pressure rises.The patient is nursed flat with minimal stimula-tion, pCO2 maintained at 3-5-4 kPa and coretemperature normalised. Attempts to preventthe development of cerebral oedema withdexamethasone as used in head injury wereunsuccessful. In a large controlled clinical trialCanalese et all7 showed that cases that did notdevelop cerebral oedema had a 60% survivalwhereas in those inwhom it did develop and whodid not receive mannitol survival was 10%.Mannitol therapy was highly effective in reduc-ing a raised intracranial pressure and signific-antly improved survival. This was the firstrandomised trial in fulminant hepatic failure toshow an improved survival with a defined treat-ment modality and represented an importantadvance.

Current studies are attempting to investigatethe relationship between intracranial pressure,mean arterial pressure, cerebral perfusion pres-sure and cerebral blood flow as well as lessinvasive methods to monitor cerebral blood flow.Jugular bulb cannulation with continuous mea-surement of venous oxygen saturation showsencouraging results as a continuous index ofcerebral blood flow.

Parallel laboratory studies were performed toinvestigate the possible mechanisms involved inthe development of cerebral oedema. Changes inblood-brain barrier permeability could allowsolutes into brain which would give rise to anincreased water content. Zaki et all8 showed anincrease in blood-brain barrier transport ofinulin in a rat model of liver failure. Later,ultrastructural studies showed that the tightjunctions in the cerebral capilliaries were intactindicating that the breakdown of the blood-brainbarrier was functional rather than structuralsuggesting vasogenic mechanisms of oedemaformation are less important. Inhibition ofmem-brane transport would be consistent with acytotoxic mechanism ofcerebral oedema. Inhibi-tion of Na+,K+-ATPase activity, a key enzymeof sodium transport, and increased brain watercontent were found in experiments using thegalactosamine rat model of liver failure,'9

S88

on Novem

ber 26, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.32.Suppl.S

86 on 1 Septem

ber 1991. Dow

nloaded from

Acute liverfailure

although dehydration of the animals limited thedegree of oedema.

Host defences and sepsisIt has long been recognised that a significantproportion of the mortality from fulminanthepatic failure can be attributed to the highincidence of sepsis that is so frequently seen.Investigation and documentation of this in-creased susceptibility to infection has been atopic ofmuch research at the Liver Unit over theyears. Studies by Wyke et al showed defectiveopsonisation of E coli and yeasts by serumof patients with fulminant hepatic failure, theaetiology for this defect was ascribed to comple-ment deficiency with components of both theclassical and alternative pathways being reducedto below 40% of that found in control serum.Neutrophil adherence was also noted to besignificantly impaired by Altin et al, and thisimpaired adherence along with complementdeficiency and depressed opsonisationresult in an overall significant impairment in hostdefences against infection. Both the plasma levelof fibronectin, an important component of thereticuloendothelial system and in vitro plasmaopsonisation activity were found by Imawari et alto be significantly depressed in patients withfulminant hepatic failure and the level of fibro-nectin correlated with the degree of impairmentof opsonisation activity. Significant impairmentof Kupffer cell function was elegantly demon-strated by Canalese et al in patients withfulminant hepatic failure using clearance ofmicroaggregated albumin.20 Impaired Kupffercell function was associated with encephalopathyand a significant improvement in function at 48hours was seen in those patients who survived.By contrast hepatocyte dysfunction was equallyimpaired in patients with acute hepatic necrosiswith and without encephalopathy, highlightingthe possible importance of the relationshipbetween Kupffer cells and hepatocytes inthe aetiology of encephalopathy. Fibronectinreplacement in patients with fulminant hepaticfailure was studied by Hughes et al, this resultedin an increase in in vitro plasma opsonisationactivity, however, clearance of microaggregatedalbumin remained unchanged as did oxygenconsumption, suggesting that other factors areresponsible. The incidence of infection inpatients with fulminant hepatic failure hasrecently been reported by Rolando et al,2 show-ing a 80% infection rate in patients with fulmin-ant hepatic failure, of great interest was thefinding that infecting organisms were gram posi-tive in 50% of cases and were nosocomial innature. Further work by Rolando et al has alsoindicated a surprisingly high incidence of fungalinfection in fulminant hepatic failure, occurringin 30% of patients.

HaemodynamicsThe haemodynamic and cardiovascular disturb-ances offulminant hepatic failure have long beenan interest of the Liver Unit. In 1977 Weston etal described a frequent incidence of cardiacdysrythmias and electrocardiographic abnorm-

alities in patients with fulminant hepatic failure,malignant arrythmias being more frequent inpatients with hypoxia, acidosis and hyper-kalaemia. Macroscopic cardiac abnormalitieswere found in 64% of patients at necropsy, thesecomprising scattered peticheal haemorrhages,pericardial effusions and fatty pale ventricles.The pathophysiology of the hypotension seen infulminant hepatic failure was first investigatedby Trewby et al. In this paper it was shown thatthe hypotension of fulminant hepatic failure wasrelated to peripheral vasodilation rather thanprimary myocardial failure, and that volumeloading frequently resulted in an increase incardiac output and a resultant increase in meanarterial pressure. Trewby et al also noted thepresence of low pressure pulmonary oedema infulminant hepatic failure, this finding being fre-quently associated with the presence of severecerebral oedema raising the possibility of neuro-genic pulmonary oedema in patients with fulmi-nant hepatic failure. Work by Bihari and Gimsonresulted in detailed study and further investiga-tion of the haemodynamic pathophysiology offulminant hepatic failure, with the identificationof pathological supply dependency for oxygen insuch patients. Significant differences betweensurvivors and non-survivors was noted, withlower systemic vascular resistance indices,oxygen extraction ratio and higher blood lactatesin non-surviving patients.22 Only in survivorswas the in vivo P50 related to the oxygen extrac-tion ratio. In addition it was noted that mixedvenous lactate correlated with systemic vascularresistance index, mean arterial pressure andoxygen extraction ratio.23 These data suggest thatin patients with fulminant hepatic failure there isa disturbance of the microcirculation resulting ina tissue oxygen debt that is greater in patientsthat fail to survive. The use of prostacyclin, amicrocirculatory vasodilator in such patientsresults in an increase in oxygen uptake, with thepossibility of lessening or reversing the coverttissue oxygen debt. More recently the effects ofvasopressor agents have been studied. Althoughthese agents improve mean arterial pressure theyhave a detrimental effect on peripheral oxygenextraction ratio and oxygen consumption result-ing in tissue hypoxia and this effect on peripheraloxygen uptake can be reversed by the use ofprostacyclin in addition to vasopressor agents.The aetiological events that result in a tissueoxygen debt in patients with fulminant hepaticfailure are poorly understood but relate to thedevelopment of interstitial oedema, white celland platelet plugs and functional arteriovenousshunts. Cytokines are almost certainly importantmediators resulting in these microcirculatorychanges. Recent work by Sheron et al has shownraised levels of cytokines in fulminant hepaticfailure with interleukin-6 (a marker of endo-thelial activation) correlating with hypotension,cerebral oedema, systemic vascular resistanceindex, oxygen extraction ratio, oxygen consump-tion and pH.

Hepatic regenerationLiver regeneration is ultimately responsible forsurvival in liver failure, but the processes in-

S89

on Novem

ber 26, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.32.Suppl.S

86 on 1 Septem

ber 1991. Dow

nloaded from

S90 Hughes, Wendon, Gimson

volved are only recently beginning to becomeelucidated at the molecular level. Serum frompatients with fulminant hepatic failure was shownby Hughes et al to be cytotoxic to rabbit hepato-cytes and subsequently in 1985 Gove et al'4showed that fulminant hepatic failure serum wasinhibitory to DNA synthesis in isolated regener-ating rat hepatocytes. This was confirmed in invivo experiments in partially hepatectomised ratsby Yamada et al." Injection of fulminant hepaticfailure serum at the time of maximal DNAsynthesis inhibited the incorporation of 3H-thymidine into hepatic DNA. These effects weregreater with sera from patients with fulminanthepatic failure because ofnon-A, non-B hepatitisand adverse reactions to drugs compared withpatients with fulminant hepatic failure caused byparacetamol overdose and hepatitisA and B. Theinhibitory activity was associated with a less than10kD fraction ofserum and this could be furtherresolved into two fractions by gel chromato-graphy corresponding to approximate molecularweights of 2 kD and 5 kD.The isolation and purification of a highly

potent human hepatocyte growth factor (hHGF)was recently reported by workers from Japan.'6High concentrations of this factor were found insera of patients with fulminant hepatic failure,particularly in those who died. This suggests thatthe poor liver regeneration in certain groups ofpatients is not caused by lack of this stimulatoryfactor. In which case identification of the inhibi-tory factors found at King's is of even greaterimportance.

Recommendations for treatmentThe previous sections of this chapter have givena selected overview of the main areas of thera-peutic study and have led to certain generalisa-tions about overall treatment strategies for ful-minant hepatic failure. All patients with grade3 or 4 hepatic encephalopathy should receivedextrose infusion and H2-receptor blockersas originally described by the late BrianMacdougall. Intermittant positive pressureventilation to maintain PaCO2 at 4-4 5 kPashould be started and the patient nursed with100 head up position. Mannitol, 20% solution,should be given for episodes of cerebral oedemadetected either by clinical criteria or intracranialpressure monitoring and supplemented byhaemofiltration if renal failure is present.Machine driven dialysis should be avoided andcontinuous arteriovenous haemofiltration is thepreferred renal replacement therapy. Closemicrobiological surveillance will reveal septicepisodes and dictate the appropriate antibioticpolicy. Finally, in those cases in whom theprognostic criteria indicate a survival of less than10%, orthotopic liver transplantation remainsthe best chance of prolonged survival.

The futureNumerous avenues of new treatment andresearch are opening up in fulminant hepaticfailure. Not all can be mentioned in this reviewbut studies into the mechanisms of hepato-cellular necrosis and attempts to prevent this

process will be ofprime importance. Preliminarystudies from Toronto on the use of PGE1 infu-sion as a hepatoprotective agent have beenencouraging start to this approach, but awaitformal trials. When events at the sinusoidal leveland cellular level which lead to cell death arebetter understood, more rational treatmentstrategies could then be developed.The concept of circulating toxins that are not

cleared by the failing liver and are responsible forthe end organ damage of fulminant hepaticfailure has a long history. Nevertheless, recentanecdotal reports of significant improvementsin intracranial pressure and haemodynamics inpatients with fulminant hepatic failure in whomthe liver is removed some hours before ortho-topic liver transplantation have raised the possi-bility that the necrotic liver itself is releasingtoxic substances with systemic effects onorgan function and microcirculatory blood flow.Identification of these substances and attemptsto remove them by extracorporeal devices or byneutralisation with monoclonal antibodies willbe of key importance, as well as earlier removalof the necrotic liver in patients selected for livertransplantation.

Because of the high frequency of clinicallysignificant bacterial and fungal sepsis duringfulminant hepatic failure the possibility of pro-phylactic treatment has been raised. Systemicparenteral and enteral decontamination regimeshave been shown to be of use in critically illpatients in intensive care units. A trial is cur-rently in progress to assess the benefit and sideeffects of prophylactic parenteral and enteraldecontamination.The future investigation of the haemo-

dynamics of fulminant hepatic failure requireselucidation and investigation of agents involvedin the control of the microcirculation, such asthe endothelins, endothelium-derived relaxantfactor, thromboxanes and prostacyclin. Im-proved understanding of the role of these agentsand their interaction with tumour necrosis factorand the interleukins will allow greater manipula-tion of the microcirculation. Compounds topromote nutritive microcirculatory flow could beadministered and detrimental agents removedwith appropriate monoclonal antibodies. Such astrategy would limit tissue hypoxia and hence themultiple organ damage that is so frequently seenin fulminant hepatic failure and results in suchan increase in mortality.

In recent years there has been much study ofthe mechanisms of control and initiation of livercell regeneration. Identification of hepaticstimulatory substances and inhibitors of regen-eration and investigation of the ability of thesurviving hepatocytes to respond to these stimuliwill all yield important results.

Hepatocyte transplantation and heterotopicliver transplantation to tide the patient over aperiod of severe hepatocellular dysfunction, withthe patient retaining their own liver, is alsoanother promising area for future investigation.

1 Williams R, ed. Fulminant hepatic failure. Semnin Liver Dis1986; 6: No. 2.

2 Toghill PJ, Williams R, Stephens JD, Carroll JD. Acutehepatic necrosis following an overdose of paracetamol.Gaseroernerology 1969; 56: 773-6.

on Novem

ber 26, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.32.Suppl.S

86 on 1 Septem

ber 1991. Dow

nloaded from

Acute liverfailure S91

3 Harrison PM, Keays R, Bray GP, Alexander GJM, WilliamsR. Improved outcome of paracetamol-induced fulminanthepatic failure by late administration of acetylcysteine.Lancet 1990; 335: 1572-3.

4 Gimson AES, White YS, Eddleston ALWF, Williams R.Clinical and prognostic differences in fulminant hepatitistype A, B and non-A, non-B. Gut 1983; 24: 1194-8.

5 Vergani D, Mieli-Vergani G, Alberti A, et al. Demonstrationofa circulating antibody reacting with the surface membraneof halothane-altered hepatocytes in patients with fulminanthepatitis following halothane anaesthesia. N Engl J Med1980; 303: 66-71.

6 Rake MO, Flute PT, Pannell G, Williams R. Intravascularcoagulation in acute hepatic necrosis. Lancet 1970; i: 533-7.

7 Langley PG, Hughes RD, Williams R. Increased factor VIIIcomplex in fulininant hepatic failure. Thrombosis Haemo-stasis 1985; 54: 693-6.

8 Hughes RD, Lane DA, Ireland H, Langley PG, Gimson AES,Williams R. Fibrinogen derivatives and platelet activationproducts in acute and chronic liver disease. Clin Sci 1985;68:701-7.

9 Gazzard BG, Weston MJ, Murray-Lyon IM, et al. Charcoalhaemoperfusion in the treatment of fulminant hepaticfailure. Lancet 1974; i: 1301-7.

10 Gimson AES, Braude S, Mellon PJ, Canalese J, Williams R.Earlier charcoal haemoperfusion in fulninant hepaticfailure. Lancet 1982; ii: 681-3.

11 O'Grady JG, Gimson AES, O'Brien CJ, Pucknell A, HughesRD, Williams R. Controlled trials of charcoal haemoper-fusion and prognostic factors in fulminant hepatic failure.Gastroenterology 1988; 94: 1186-92.

12 Bihari D, Hughes RD, Gimson AES, et al. Effects of serialresin haemoperfusion in fuhninant hepatic failure. IntJ ArtifOrgans 1983; 6: 299-302.

13 Alam AN, Poston L, Wilkinson SP, Golindano CG, WilliamsR. A study in vitro of the sodium pump in fulminant hepaticfailure. Clin Sci MolMed 1978; 55: 355-63.

14 Seda HWM, Hughes RD, Gove CD, Williams R. Inhibition ofrat brain Na+,K+-ATPase activity by serum from patientswith fulninant hepatic failure. Hepatology 1984; 4: 74-9.

15 Yang S-S, Hughes RD, Williams R. Digoxin-like immuno-reactive substances in severe acute liver disease due to viral

hepatitis and paracetamol overdose. Hepatology 1988; 8:93-7.

16 Gazzard BG, Portmann B, Murray-Lyon IM, Williams R.Causes ofdeath in fulninant hepatic failure and relationshipto quantitative histological assessment of parenchymaldamage. QJMed 1975; 44: 615-26.

17 Canalese J, Gimson AES, Davis C, Mellon PJ, Davis M,Williams R. Controlled trial of dexamethasone and mannitolfor the cerebral oedema of fulminant hepatic failure. Gut1982; 23: 625-9.

18 Zaki AEO, Ede RJ, Davis M, Williams R. Experimentalstudies of the blood brain barrier in acute hepatic failure.Hepatology 1984; 4: 359-64.

19 Ede RJ, Gove CD, Hughes RD, Marshall W, Williams R.Reduced brain Na+,K+-ATPase activity in rats withgalactosamine-induced hepatic failure: relationship to en-cephalopathy and cerebral oedema. Clin Sci 1987; 72:365-71.

20 Canalese J, Gove CD, Gimson AES, Wilkinson SP, WardleEN, Williams R. Reticuloendothelial system and hepatocytefunction in fulminant hepatic failure. Gut 1982; 23: 265-9.

21 Rolando N, Harvey F, Brahm J, et al. Prospective study ofbacterial infection in acute liver failure: an analysis of fiftypatients. Hepatology 1990; 11: 49-53.

22 Bihari D, Gimson AES, Waterson M, Williams R. Tissuehypoxia during fulminant hepatic failure. Crit Care Med1985; 13: 1034-9.

23 Bihari D, Gimson AES, Lindridge J, Williams R. Lacticacidosis in fulminant hepatic failure. Some aspects ofpathogenesis and prognosis. J Hepatol 1985; 1: 405-16.

24 Gove CD, Hughes RD, Williams R. Rapid inhibition of DNAsynthesis in hepatocytes from regenerating rat livers byserum from patients with fulninant hepatic failure. BrJ ExpPathol 1982; 63: 547-53.

25 Yamada H, Hughes RD, Gove CD, Williams R. Effects offulninant hepatic failure serum on hepatic DNA synthesisin normal and partially hepatectomised rats. Eurj Gastro-enterol Hepatol 1990; 2: 483-8.

26 Gohda H, Tsubouchi H, Nakayama H, et al. Purification andpartial characterization of hepatocyte growth factor fromplasma of a patient with fulminant hepatic failure. J ClinInvest 1988; 81: 414-9.

on Novem

ber 26, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.32.Suppl.S

86 on 1 Septem

ber 1991. Dow

nloaded from