"';I"d- 3'.?

Eur. surg.Res.18: 283-292:(1986)
C 1986 S. Karger AG, Base! 0014-312X/86/0~5-0283$2. 75/0
A Predictable Pathophysiological Model of Porcine Hepatic Failure

R. Hickman,M.H. Alp

MRC Liver Research Group, Departments ofMedicine and Surgery, University ofCape Town and the Gastrointestinal Clinic, Groote Schuur Hospital, Observatory, South Africa
Key Words. Pig.Experimentalliver failure.Carbon tetrachloride.Ischaemia Phenobarbitone.Amino acids Abstract. A technique is described in which acute hepatic failure may be uniformly induced in pigs using a combination of oral phenobarbitone pretreatment and 2-hour interruption of hepatic arterial blood supply followe;d by intragastric-carbono tetrachloride. This results in deephepatic coma ayer a narrowly predictable time spanand death within 12-52 h. Amino acid leyels in plasma and cerebrospinal fluid oqtained 24 h after the induction of hepatic failure were similar tothose reported in human and experimental encephalopathy.

lntroduction

Acute hepatic failure remains a potentially lethal complication of viTal or drug-induced hepatitis and methods of treatment are still unsatisfactory in general. While the use of galactosamineby Blitzer et al. [1] in the rabbit and by Decker and Keppler [5] in the rat has resulted in adequatesmall animal models of experimental liver failure, there is still a need for a suitable large animal version to allow for repeated sampling or for the application of methods of treatment such as resin perfusion, plasma exchange,etc.
The anhepatic models of Hickman et al. [11] and Herlin et al. [10] and the combined porta-caval shunt/hepatic artery ligation model arDe Groot et al. [6] sutTered from the defect of excluding the damaged liver from the circulation. Hickman et al. have found the thermochemotherapeutic rat model of Miyazaki et al. [17] to be unreliable and not reproducible in the pig (unpublished observations) as has been the administration of galactosamine. Acetaminophen which produces hepatic necrosis in hamsters [3] was proved unreliable in pigs [15]. Yellow phosphorus has a predictably lethal etTectin dogs

l~~i~~ ~

Hickman/ Alp
butremains dangerousto its users even after administration to the animal [2]. Thus these methods did not fulfil completely the criteria suggestedby Terblanche et al. [23] which stress not only the need for a large animal with death from liver failure using methods with minimal hazard to personnel but also reproducibility and potential reversibility. To these five criteria could be added a sixth, videlicet reasonablylong time interval from onset of acute failure to death which would allow for the initiation of methods of treatmentoThus in one way or another t~esevarious methods were unsatisfactory. A suitable model in the pig which fulfils these criteria has been achieved by a combination of pretreatment with phenobarbitone followed by timed reversible hepatic artery occlusionand a graded oral dose of carbon tetrachloride as describedin 1974 by van Leenhoffet al. [25]. Hepatic failure and coma occurred predictably within narrow time limit~ and were invariably followed by death. Plasmaand cerebrospinal fluid (CSF) free amino acids were assayed before and after the induction of hepatic coma.
For groups I and 11 CCI4 was given after a smalI dose of intravenous sodium pentothal to induce anaesthesia.Anaesthesia for hepatic artery Iigation in groups 11, 111and IV was induced with 1-3 mg/kg sodium pentothal and maintained with oxygen and nitrous oxide. AlI pigs were intubated for any procedure to prevent puImonary contamination with
At laparotomy, the hepatic artery was dissected free in the porta hepatis. After all ligamentous and vascular connections passing the diaphragm had been divided, the hepatic artery above the right gastric branch afilie coeIiac artery was occluded with a snare ofpolyethylene tubing. The abdomen was closed with a single layer of DermaIon. Two hour~ Iater the animals were reopened rapidly under minimal nitrous oxide anaesthesia,the snare was removed and the pulsation ofthe hepatic artery was confirmed. Then CCI4 was injected directIy into the stomach. The abdomen was closed in layers. Cathters (8 Fr.) were placed into the femoraI artery and vein for measurement ofbIood pressure, blood sampling and infusion. Intravenous fluids were given to maintain urine output, and the pigs were housed in individual warmed cages.Blood and CSF samples were taken pre-operatively, immediately after the ligation and after 24 and 48 h if the animal survived. Postmortem examination was carried out in all pigs and sampIes ofliver and lung.were

examined histologicaIly.

Serum bilirubin and serum glutamate oxaloacetate transaminase (SGOT) were estimated by standard techniques. Prothrombin index was estimated by a manual Quick one-stage method using a Dade standard kit containing rabbit brain thrombopIastin. Plasma ammonia leveIs were estimated using a Hyland blood ammonia test kit. The samples were collected directly into heparin-containing tubes on ice, centrifuged within 30 min, stored at 20.C and tested within 48 h of collection. Samples were measured in duplicate and compared with known standards supplied with the kit. CSF was obtained by lumbar puncture from pigs in group 11, 111and IV. Sgme animals needed 2-3 mI sodium pentothaI as a light anaesthetic for this procedure but many were so deeply unconscious that this was not necessary.This waj taken as additional evidence of encephaIopathy. Amino acid concentrations in the serum and CSF both before and after induction ofhepatic faiIure were estimated using a Waters HPLC reverse-phase C-I8 coIumn (3.9 mI X 30 cm) with a

I Methods

Large White pigs, 6-8 weeks old weighing 1520 kg were used throughout. Four groups of animals were used which had been starved for 24 h (table 1). Group I (10 pigs)was given 0.5 ml/kg oral carbon tetrachloride (CCI4) alone. Group 11 (16 pigs) was given 0.5 ml/kg CCI4 followed by 2 h of hepatic artery ligation 1-3 days later. Group 111.( O pigs)was treated for days with 8 mg/kg phenobarbitone prior ~o hepatic artery ligation for 2 h followed immediately by 0.5 ml/kg CCI4. Group IV (8 pigs) was treated in a similar fashion as group 111 prior administration of but phenobarbitone was omitted. In all groups CCI4 was instilled into the stomac;-hby gastric tube or direct injection.

Model of HepaticFailure

Tablel. Survival fol1owing induction ofliver failure Group Method of induction Survival

11 (n = 16)

CC!4 0.5 rn!/kg hepatic arteria! ischaernia for 2 h !-3 days !ater phenobarbitone, 8 rng/kg for 3 days hepatic arterial occlusion for 2 h CCl4 0.5 rnl/kg hepatic arterial occlusion fOI 2 h CCI4 0.5 ml/kg -

(n = 10) IV (n = 8)

short-termsurviva1(15-52 h)
postoperative death pneumonia short-termsurvival >10 days
linear acetonitrile gradient (10-55%) over 32 min in 15 mM phosph~te buffer (pH 7.2) [12]. The equipment consisted of two Waters Model 6000 A pumps, a Waters -Model 680 automatic gradient controller and a Wisp 710 B injector. Precolumn derivation and fluorescente were obtained by using ethanethiol and o-phthaldialdehyde [14], the eluted amino acids being identified and quantitated using a Cecil continuousflow UV monitor set at a wavelength of 330 nm and 0.02 absorbance by their retention times and absorbance peaks on the column as read by a Waters Model M730 Data Module. Homocysteic acid (10 mM) was used as an internal standard. Concentrations of internal standard were as follows: 5 ~ for norIrlal CSF, 7.5 ~ for norIrlal plasma and 10~ for hepatic failure plasma and CSF. Two hundred microliters of plasma or CSF was always derived but the amount of derived plasma applied to the column also varied: 15 ~ for hepatic failure; 25 ~ for norIrlal plasma and hepatic failure CSF and 50 ~ for norIrlal CSF. The method is limited by the similar retention times of glycine and threonine which are therefore

Results

Survival (table 1) Group lo Of the 10 pigs given 0.5 ml/kg CCI4via a gastric tube under light pentothal anaesthesia,2 died 8 h later without regaining consciousness and another died from se'veTe pneumonic consolidation after 36 h. The remainderlived for 10daysbefore sacrifice with no evidence of coma. AII showed histological features of mild hepatic damage at post-~ortem (fig. la). Group 11. Of the 16 pigs givn 0.5 ml/kg CCI4 1-3 days before hepatic artery occlusion, 4 died before operation and 2 died during the operation from excessive bleeding. Of the remaining lO, 3 were still alive after 1 week and were sacrificed, while 7 survived from 12 h to 6 days (average3.1 days). The

.~~,~.c,

.;r,1r

"k

Table 11. Biochemical changes in porcine induced. liver failure

Normal

(n = 45)

Group II (n = 16)

Group 111 (n = 10)

Group IV

(n = 8)
Arnrnonia Prothrombinindex

75 (53-100)

Figures in parentheses indicate range. * p < 0.05; ** p < 0.0 1; *** P < 0.00 I
reduced response to painful stimuli before regaining consciousness. Histologically all showed evidence of varying degrees of hepatic necrosis at postmortem. Group III~ Ten pigs were given 8 mg;kg phenobarbitone orally for 3 days followed by 2 h of hepatic artery ligation and intragastric CCl4 (0.5 ml/kg). All 10 animals recovered from the procedure"and were alert but not standing 6 h postoperatively. Death occurred in all pigs within 15-52 h (average32 h). This was preceded by 6-36 h of deep hepatic coma, the pigs being unconscious and only slightly responsiveto painful stimuli. Death was frequently preceded by twitching and convulsions. No evidence ofintracerebral or CSF infection could be found at :
postmortem. Pulmonary consolidation was not seenalthough on occasionthe lungs were dotted with 2-4-mm haemorrhagic aTeaS. The.livers in all pigs were pale and swollen. Histology showedseverecentrilobular necrosis typical of CCl4 poisoning (fig. lb). No other intraabdominallesion was found. Group IV. Eight pigs underwent 2 h of hepatic artery ligation followed by intragastric CCl4(0.5 ml/kg) without prior phenobarbitone treatment. Three died within 36 h, 2 from respiratory failure and 1 from wound dehiscence.One died from an overdosageof sodium pentothal usedto induce anaesthesia for lumbar puncture at 24 h, 1 from no obvious cause at 48 h, and there were 3 longterm survivors. As for group 1, the hepatic histological changeswere variable but generally mildo Biochemical Resulls(table II) There were no significant biochemical changesin group l. Levels of SGOT in group II pigs 24 h postoperativelywere moderately elevated compared with those obtained pre-

Fig. 1. a Biopsy from a pig sacrificed 4 days after administration of oral CCl4 alone. Note th minimal pericentral destruction. HE. X 40. b A biopsy from a pig from group 1II (oral phenobarbitone, hepatic arterial ischaemia and intragastric CCI4) showing severe centrilobular necrosis destroying two thirds of the hepatic lobule. HE. X 40.

." \'

'""~-"""~~

Hickman/Alp

Table 111.Plasma amino acid levels (J1mol/l) in induced hepatic damage (mean :!: SEM)

Amino acid

(n = 16)

Group 11 (o = 8)

Group 1II (n = 10)
Group IV (o = 7) 38:1:6 111:f18 88:1:25 234:1:22 502:1:60 95:1:12 1,242:1:132 643:1:75 69:1:12 188:1:16 403:1:33 88:1: 15 226:1:25 114:1:13 253:1:22 316:1:55 330:1:55 560:1:97
Aspartic acid Glutamicacid. AsparagIne Serine Glutamine Histidine Glycine and threonine.-. Alanme TyrosineArginine Valine' Methionine' Isoleucine Tryptophan Leucine PhenylalanineOmithine' Lysine '."
50:1:14 208:1:46 77:1:16 108:1:17 287:1:27 59:1:10 873:1:81 590:1:73 55:1:9 192:1:15 446:1:54 56:1:14 161:1:21 58:1:8 262:1:42 120:1:26 233:1:40 352:1:59
34:1::1:20 155:1:25 293:1:77 72:1:19 184:1:43* 117:1:115 280:1:54* 373:1:77 1,115:1:371* 65:1:15 210:1:42" 713:1:72 2,213:1:484* 403:1:66 1,672:1:585 37:1:8 213:1:47" 112:1:23377:1:96 245:1:32 448:1:81 35:1:3 214:1:44" 103:1:23 154:1:36 84:1:13 132:1:18** 188:1:10 330:1:98 60:1:8 624:1:72*** 191:1:45 604:1:118* 321:1:17 1,264:1:333*
.p < 0.05;.p < 0.01;.p < 0.001
operatively but fell rapidly to be within normallimits within 48-72 h. Blood ammonia was not significantly elevated at any time nor was serum bilirubin. In group 111there were highly significant changesin all samplesstudied. In 2 pigs surviving for longer than 48 h there was a further rise in bilirubin levels at the time with a 50 % fall in ammonia and SGOT levels. Prothrombin indices fell to a mean of 45 % (normal >80%). Group IV pigs showed a moderate yet statistically significant rise in SGOT levels 24 h after operation which persisted for 48-72 h. No significant rise occurred in ammonia or bilirubin levels although the latter rase further in the subsequent h. Prothrombin in48 dex was not recorded.
Amino Acid Changes(tablesIII, IV) There were no significant differences in plasma or CSF amino acid levels in group I (CCI4 treated) pigs as compared with preoperative values obtained from 16 animals. There were significant elevations of gIutamine, arginine and methionine in CSF of group 11pigs, while in group m animals, the amino acids were even more abnormal, with only gIutamic acid, aspa!agine, serine and omithine failing to- show statistically significantrises. Althougt "branched-chain aroand matic amino acids as well asmethionine were significantly elevated~ greatestrise was in t'\te the latter two groups, markedly lowering the branched-chain:aromatic amino acid ratio. This was common both to the plasmaand the

Table IV. CSF amino acid levels ~ol/l)
in induced hepatic damage (mean :t SEM) Group 11 (o = 5) 20:tS 20:t6 76:t4S 94:t10 1,162:t 122* IS:t4 126:t21 IIS:t31 29:t4 27:tS* 29:t6 24:tS* 17:t3 22:t6 ~

Amino Acid

Group III
(n = 5) 36:t lO 227:t 107 117:tSS 192:t37* I,SSO:t212"* 94:t1166S:t166"* S61:t172* 101 :t29* 164:t67* 12S:t3S* S6:t 13** 67:t20* -~ 67:t2S*
Group IV (n = S) 10:t3 -20:t7 SI:t16 SO:t9 767:t66 216:t70 364:t32 S:t 1 2:tS 11:tS 4:t2 12:t3 6:t 1
Aspartic acid Glutamic acid Asparagine Serine Glutamine Histidine Glycineandthreonine Alanine Tyrosine Arginine Vaiine Methionine Isoleucine Tryptophan
6:t3 7:t3 112:t12 116:t9 60S:t29 11:t4 12S:tIS 109:tIO 14:t7 S:t2 -14:t3 S:t2 12:t3' __IS:t3
Leucine.~PhenylalanlIle Omithine Lysine
11:t3 10:t 2 186:t46 187:t 29
30:t12 16:t 7 236:tS3 209:t 34
128:t41* ~~127:t 20*** 2S8:t37 514:t 60-*

S:tl13:t 3

* p < 0.05; ** p < 0.01; *** P < 0.001
In group IV pigs, apart from a rise in CSF glutamine, Done of the amino acids showed any significant change. Changesin plasma amino acids in groups 11 and IV were not significant but in group 111, aromatic amino acids and methionine all were markedly elevated (p < 0.001) but the branched-chain amino acid levels remained within normallimits.

Discussion

e.g. portacaval shunt with irreversible arterial ischaemiaas described by De Groot et al. [6] or the partial hepatectomytechnique of Minato et al. [16]. The presentmodel is theoretically reversible becausethe damaged liver remains in the circulation and both insults are of a sublethalnature (the dose of CC14 as describedby van Leenh<;>ff al. [25] and the et arterial ischaemia chosen fiom the experience of Kahn et al. [13]). In further preliminary studies, the administration of PGF2a resuted in survival in about 40% ofaimals, thus confirming the possibility of re-
versibility. Amino acid profiles for plasma have Many ofthe previous large animal models been published for rabbits [4], rats [7], dogs of hepatic failure have not been predictable or were irreversible because the damaged [20], pigs [18] and humans [19]. Our results liver had beenexcluded from the circulation, show little variations from the spectropho-

--.'" ;j

tometric assays in any species apart from lower glutamine levels. Normal plasma leucine and valiDe levels in the pig are twice those in human and rabbit plasma but other essentialamino acid concentrationsare similar. However, little is known about the amino acid levels in pig plasma and CSF following induction of hepatic failure. In a. limited study of plasma amino acids in pigs subjected to portacaval shunt and hepatic artery ligation, Wustrow et al. [26] found depressiOI1 the levels of branched-chain of amino acids and elevation of aromatic amino acid levels. OUTresults confirm the rise in most plasma amino acids, particularly the aromatic amino acids, noted in human hepatic failure [8] and hepatectomised dogs [20]. However, branched-chain amino acids were not significantly elevated in OUT study, resulting in a fall in the branchedchain:aromatic amino acid ratio, which is suggestedto be important in the development of hepatic encephalopathy [8]. This lack of change may relate specifically to ibis form of induction of hepatic failure, and possibly contributory factors such as the insulin:glucagon ratio need to be investigated before further interpretation. The normal porcine CSFamino acids differ from human levels [24] in that normal porcine CSF contains much greateramounts of asparagine, glycine and alanine, and reduced arginine. However, the essential amino acid concentrations are similar. In hepatic coma, the CSF profile changesmarkedly with the prominent abnormality being in glutamine levels. CaTewas taken to exclude intracerebral infection in view of the reported changes in CSF amino acids in its presence[22]. In comparison to amino acid levels found in human hepatic encephalopathy [21], porcine CSF contained greater
amounts of glycine and branched-chain
amino acids, but not sufficiently to alter lheir ratio compared to the aromatic amino acids. It is recognisedthat CSF obtained by lumbar puncture may not be representative of that from cisternal puncture but the hazards of infection by the latter method are greater. In OUT experiments, too, pre- and postoperative samples were obtained by the same route. The increasein plasma tryptophan in pigs of group 111 and in CSF arginine may represent only speciesdifference or be specific features of this method of induction of encephalopathy. Further comment could be made if values were available from other models of encephalopathy in pigs. The lack of correlation between amino acid charigesin plasma and CSF needscommentoIn pigs of groups 11and 111, apparently greaterchangesoccurred in the levels in CSF, suggestinga defect in blood brain transport, despite the assertions of Herlin et al. [9] to the contrary. The pig model of phenobarbitone treatment in conjunction with timed hepatic artery ligation and oral CCl4 regime has been shown to fulfil all the criteria expressedby Terblanche et al. [23] except the possibility of reversibility. In preliminary studies with this model, administration of prostaglandin F2 resulted in a survival rate of about 40% for > 10 days, while administration of amino acids, cholestyramine, cysteamine or glutathione proved unsuccessful [Alp, submitted for publicati?n].- Thus, the model as mentioned earlier does appear to be reversible. We are also ~ble 10 demonstrate that the model fulfils the:'Sixth criteria of a reasonable time interval between induction of failure and death, and thus it should prove useful in further studies of therapeutic modalities.

Acknowledgements

Financial assistance was received from the Medical Research Council of South Africa and the Staff Research Fund of the University of Cape Town. Ms. Rose Innes and H. McLeod performed the biochemical and amino acid analyses and Mrs. Grunwald prepared the histological specimens. Mrs. Johns typed the manuscript. Pre- and postoperative care was given by the staff of the J.S. Marais Surgical Laborntory, especially D. Tango.
9 Herlin, P.M.; James, J.H.; Machbauer, C.A.; Fischer, J.E.: The blood brain barrier is intact eighteen hours after total hepatectomy. Hepatology 1: 515 (1981). 10 Herlin, P.M.; James, J.H.; Nachbauer, C.A.; Fischer, J.A.: Effect of total hepatectomy on administration of branched chain amino acids on regional norepinephrine, dopamine and amino acids in rat brain. Ann. Surg. 198: 172 (1983). II Hickman, R.; Dent, D.; Terblanche, J.: The anhepatic model in the pig. S. Afr. med. J. 48: 263
(1974). 12 Hill, D.W.; Walters, F.H.; Wilson, T.D.; Stuart, J.D.: High perfonnance liquid chromatographic detennination of amino acids in the pico-mole References range. Analyt. Chem. 51: 1338 (1979). I Blitzer, B.L.; Waggoner, J.G.; Jones,E.A.; Gral- 13 Kahn, D.; Hickman, R.; Dent, D.M.; Terblanche, J.: For how long can the liver tolerate ischaemia? nick, H.R.; Towne,D.; Butler,J.; Weise,V.; KoEur. surg. Res. 18: 277-282 (1986). pin, l.J.; Walters,l.; Teychenne, P.E.; Goodman, D.U.M.; Berk,P.D.:A modeloffulminant hepatic 14 Lindroth, P.; Mupper, K.: High performance liquid chromatographic determination of sub picofailure with the rabbit. Gastroenterology 664 74: mole amounts of amino acids by fluorescence (1978). derivatization with o-phthaldialdehyde. Analyt. 2 Bumell, J.M.; Dennis, M.B.; Clayson, M.S.; Chem. 51: 1667 (1979). Smuckler, E.A.; Clift, R.A.: Evaluationin dogsof crosscirculation in the treatmentof acutehepatic 15 Miller, D.J.; Hickman, R.; Fratter, R.; Terblanche, J.; Saunders, S.: An animal model of fulminant necrosis inducedbr yellowphosphorus. Gastroenhepatic failure: a feasibility study. Gastroenterolterology71: 827(1976). ogy 71: 109 (1976). 3 Chiu, S.; Bhakthan,H.M.G.: Experimentalacetoaminophen inducedhepaticnecrosis: biochemi- 16 Minato, M.; Houssin, D.; Demma, l.; Morin, J.; Gigou, M.; Szekely, A.M.; Bismuth, H.: Transcal and electronmicroscope study of cysteamine plantation of hepatocytes for treatment of surgiprotection.Lab. Invest. 39: 193(1978). cally induced acute hepatic faiJure in the rato Eur. 4 Daniel, P.M.; Pratt, O.E.; Spargo, Amino acid E.: surg. Res. 16: 162 (1984). pattem in the blood of alloxan diabetic rabbits response glucagon. Physiol.32: 512(1980).17 Miyazaki, M.; Makowka, L.; Falk, J.A.; McDonell, to J. M.; Venturi, D.: Protection ofthermochemotheraS Decker,K.; Keppler,D.: Galactosainine induced peutically induced lethal acute hepatic necrosis in injury. Prog. Liver Dis., vol. 4, p. 183(Grune & the pig by 16,J6-dimethyl-prostaglandin El. J. Stratton, New York 1972). surg. Res. 34: 415 (1983). 6 Groot, G.H. de; Schalm,S.W.; Schicht,l.; Bativier, P.; Jonge,M. de; Lens, J.; Terpstra,J.L.: 18 Montgqmery, G.W.; Flux, D.S.; Greenway, R.M.: Tryptphan deficiency in pigs: changes in food Comparison oflarge pare membranehaemodialyintake and plasma levels of glucose, amino acids, sisand crossdialysisin acutehepaticinsufficiency insulin and growth hormone. Honnone metabol.

in pigs.Eur. J. clin. Invest. 13: 65-71 (1983). Res. 12:30(1980). 7 Eriksson, Carlsson, Liljequist, S.; Hagman, J.; A.; M.; Jagenburg, Decrease plasma R.: in aminoacids 19 Morgan, M,Y.; Marshall, A.W.; Milsom, J.P.; Sherlock. S.: Plasma amino acids in liver disease. in rat after acute administration of ethanol. J. Gut 23: 362 (1982). Pharm.Pharmac.32: 512(1980). 20 Nagasue,N.; Kanashima, R.; lnokuchi. K.: Altera8 Fischer,J.E.; Rosen,H.M.; Ebeid,A.M.; James, tion in plasma amino acid concentrations followJ.H.; Keane,J.M.; Soeters, P.O.:The elTect ofnoring subtotal hepatectomy in dogs. Annls ChiToGymalization of plasmaamino acids on hepatic ennaec. Fenn. 70: 50 (1981). cephalopathy manoSurgery 7 (1976). in 80:
21 Record, C.; Buxton, B.; Chase, R.; Curzon, G.; Murray-Lyon, I.M.; Williams, R.: Plasma and brnin amino acids in fulminant hepatic failure and their relationship to hepatic encephalopathy. Eur. J. clin. lnvest. 6: 387 (1976). 22 San Joaquin, U.H.; Khai, N.; Seale, T.W.; Rennert, O.M.: lncreased cerebrospinal fluid free amino acid concentrations in children with bacterinl meningitis. Stand. J. infecto Dis. 14: 23 (1982). 23 Tcrblanche, J.; Hickman, R.; Miller, D.: Animal experience with support systems. Are there appropriate animal models of fulminant hepatic necrosis? In Williams, Murray-Lyon, Artificial liver support, p. 163 (Pitman, London 1975). 24 Thalman, R.; Comegys, T.H.; Thalman, l.: Amino acid profiles in inner ear fluid and cerebrospinal fluid. Laryngoscope 92: 32 (1982). 25 Lcenhoff, J.W. van; Hickman, R.; Saunders, S.J.; Tl'rblanche, J.: Massive liver cell necrosis induced in the pig with carbon tetrachloride. S. Afr. med. J. 4.\':1201 (1974).
26 Wustrow, "f.; Hoom-Hickman, R. van; Hoom, W.A. van; Vinik, A.I.; Fischer, M.; Terblanche, J.: Acute hepaticischaemia in the pig -the changesin plasma hormones, amino acids and brain chemistry. Hepato-Gastroenterology 28: 143 (1981).
Received: October 9, 1985.Accepted: November 27, 1985
Dr. R. Hickman, Departmentof Surgery, University of CapeTown, Observatory 7925(South Africa)

" ,.