Hyperammonemic coma after hepatectomy in germ-free rats

GASTROENTEROLOGY

77231~234,1979

Hyperammonemic Coma After Hepatectomy in Germ-Free Rats SW.

SCHALM,

and T. VAN DER MEY

Department of Internal Medicine II, Erasmus University Hospital, Rotterdam, Experimental Surgery, University Hospital, Leyden, The Netherlands

Current theories on the pathogenesis of hepatic coma indicate that intestinal bacteria produce cerebral toxins, such as ammonia, mercaptans, and short-chain fatty acids. To test the hypothesis that elimination of anaerobic and aerobic intestinal bacteria retards the onset and alters the biochemical profile of acute hepatic coma, we determined the onset of coma and the ammonia concentration in blood and cecal contents in 7 germ-free and 10 normal dehepatized rats. Ammonia levels were also determined in a further group of 7 germ-free and 12 normal rats 24 hr after hepatic vascular exclusion was accomplished. Onset of coma for germ-free rats (2: 34 hr) was identical to that for normal rats (2: 36 hr). Arterial ammonia was equally elevated in germ-free rats (2: 834 pmoi/Jiter) and in normal rats (x: 854 pmoJ/liter), although the ammonia concentration in the cecal contents was significantly lower in germ-free rats (2: 2762 pmoJ/Jiter) than in normal rats (2: 5572 J.tmoJ/ liter). Jn germ-free animals, portal venous blood contained more ammonia than arterial blood (x A-V difinf erence: -87 pmoJ/Jiter), indicating nonbacterial testinal ammonia release. We conclude that intestinal bacteria toxins are of minor importance in the mechanism of acute hepatic coma of the liverless rat and that presumably bacterial toxins, such as ammonia, can be products of nonbacterial metabolism. Since hyperammonemia Received June 6,1976. Accepted April 9.1979. Address requests for reprints to: Dr. S. W. Schalm, Department of Internal Medicine II, Academisch Ziekenhuis Dijkzigt, Rotterdam, The Netherlands. This study was supported by a grant from The Netherlands organization for the advancement of pure research (ZWO). The authors are grateful to all contributors from the Section for Germ-Free Animals, Pathology and Bacteriology, Radiobiological Institute, TNO, Rijswijk, The Netherlands: to Professor J. L. Terpstra, Dr. H. Stol and the staff of the Laboratory of Experimental

Surgery for their continued support; and to Mr. J. Boot and Dr. S. T. Ypma for their excellent technical assistance in ammonia determinations. 0 1979 by the American Gastroenterological Association 0016-5085/79/0sO231-~~2.~

and Laboratory

of

could be considered an important determinant of coma in our model, prevention of hyperammonemia in functionally anhepatic animals should be the next objective in unraveling the pathogenesis of acute hepatic coma. In severe liver insufficiency, the failing liver is not able to clear ammonia, mercaptans, and short-chain fatty acids adequately so that these substances can accumulate in the brain; coma may eventually develop through synergistic action.’ Intestinal bacteria are generally thought to produce these cerebral toxins.2.3 Ammonia is produced mainly by bacterial degradation of proteins, amino acids, and urea4.5; mercaptans are the metabolic product of methionine”; short-chain fatty acids are derived from the breakdown of medium-chain triglycerides and sugar.’ Bowel cleansing to remove substrates and the administration of nonabsorbable antibiotics have therefore become the cornerstones of the treatment of hepatic encephalopathy.‘,’ In chronic portosystemic encephalopathy, such treatment is usually effective; however its efficacy in acute hepatic coma has never been demonstrated. In experimental hepatectomy studies, preoperative oral neomycin or complete elimination of intestinal toxins by evisceration has been associated with unchanged”*’ or markedly prolonged survival” as well as diminished8 or unchanged hyperammonemia.” Using a germ-free animal as experimental model rather than evisceration, we tested the hypothesis that the onset of coma after hepatectomy should be delayed in germ-free rats, because elevation of plasma intestinal toxins such as ammonia would be prevented by prior removal of intestinal substrates and the complete absence of bacteria.

Methods Male germ-free rats (Wistar, Radiobiologic Institute Rijswijk, The Netherlands), 20 wk old and weighing

232

SCHALM

AND

VAN

about 250-300 g were used. Control rats, which were of the same strain and approximately the same age and weight, were raised in a specific pathogen-free environment; the intestinal flora showed a normal anaerobic bacteria concentration but a limited number of aerobic gram-negative strains. The techniques for raising, handling, feeding, and maintaining germ-free rats were described by van der Waay.” The experiments were performed in a laminar flow box under sterile conditions. Once a week end-to-side portacaval shunts were performed under ether anesthesia in 2 germ-free and z normal rats. Postoperatively, the animals were kept in metabolic cages to prevent coprophagy; 80 ml of 5% dextrose were given orally per 24 hr. Twenty-four hours after surgery, 500 mg of magnesium sulfate were given orally to cleanse the bowel. Forty-eight hours after surgery, the abdomen was reopened under ether anesthesia: Either total hepatectomy was performed by ligating and removing the various lobes, or the hepatic artery was ligated to create hepatic vascular exclusion. Initially total hepatectomy and hepatic artery ligation were alternated weekly; however since no difference was observed after the first series of experiments, only hepatic artery ligation was performed in subsequent experiments. The jugular vein was cannulated with polyethylene tubing which was connected to an infusion system, permitting the animal complete freedom of movement in his cage. 1.3 Milliliters of 0.9% NaCl solution containing 5 or 15% dextrose, 10’ U penicillin G and 20 mg gentamycin per 100 ml were infused hourly. In the first series, the animals were tested three times a day for signs of encephalopathy (ability to walk, to lift head, and to react to sound or pain); when coma ensued, or 48 hr after dehepatization, the animals were sacrificed by bleeding from the aorta. Autopsy was performed, blood samples collected, and cecal contents taken for culture. In the second series of experiments, all animals were sacrificed by bleeding 24 hr after hepatic vascular exclusion; blood was taken from the aorta and, when feasible, from the portal vein. Cecal contents were also obtained for analysis. Blood for biochemical analysis was collected in icechilled plastic tubes containing disodium EDTA (approximately 1 mg/ml of blood). Within 1 hr, plasma was separated by centrifugation at 4’C and stored at -2O’C for 2 to a maximum of 18 hr. Deproteinization was performed by ultrafiltration (Amicon Centriflo membranes CF 25) at 4°C. The ultrafiltrate was frozen at -20°C until analysis, which took place within 3 days. The same methodology was used for the preparation of the cecal contents. Ammonia was determined by the enzymatic method (Monotest, Boehringer, Mannheim),13 adapted for small samples (200 ~1 for duplicate measurement). The method was precise (coefficient of variation: 7%) and accurate (recovery of added ammonia: 98 + 2%). Addition of 5 mM of glutamine and/or 1 mM of tyrosine did not influence the test results. Normal values for nonfasted rats were below 86 pmol/ liter. Twenty abnormal rat plasma samples were simultaneously assayed by us and in another laboratory according to two different methods14; results correlated extremely well (r = 0.99).

Vol. 77, No. 2

GASTROENTEROLOGY

DER MEY

Blood samples were cultured in 60 ml of trypticase soy broth and were observed for at least 1 wk. Cecal contents were transferred in dry sterile tubes to the microbiologic laboratory; within 2 hr, samples were placed in brainheart infusion medium for the aerobic cultures and in Brewer medium for the anaerobic cultures and incubated at 21°C and 37’C, respectively. After l-2 wk, the culture material was transferred to sheep blood agar plates. The plates containing the anaerobic culture were left in a high CO,-low 0, environment (Gaspak system); plates were read after 72 hr. This methodology has been followed routinely for the past several years for checking the germ-free status of the rats in the Radiobiologic Institute.” Glucose and clotting factors were measured by the hexokinase methodI and the normotest method,” respectively. Statistical methods included Student’s t-test and Wilcoxon rank sum test for unpaired samples.”

Results Adequacy of animal model: all 7 germ-free rats in the initial series and 7 out of 10 normal rats developed signs of encephalopathy and went into coma. Three normal rats were still in good condition when they were sacrificed 48 hr after dehepatization. Two of these had undergone complete hepatectomy as verified at autopsy; total liver necrosis was found in the third rat as a result of the hepatic vascular exclusion. Blood glucose values at sacrifice ranged from 5 to 20 mmol/liter at a constant glucose infusion of 250 mg/kg/hr; mmol/liter) was occasionally

glucose in control ments, a constant was subsequently nate hypoglycemia. depressed (~20%)

since hypoglycemia (<2 found at this dosage of

rats of the second series of experiglucose infusion of 750 mg/kg/hr introduced to completely elimiClotting factors were severely

in all animals tested 24 hr after surgery or later. Blood cultures were negative except in 1 germ-free and 1 normal rat; cultures of cecal

contents were negative for aerobic and anaerobic microorganisms for all germ-free animals, and positive for all normal rats tested (Table 1). Onset of coma: coma developed 24-45 hr (mean +SD: 33.7 + 6.6 hr) after induction of the anhepatic state in 7 germ-free rats: this period was identical (mean f SD: 35.8 -C 9.8) for 10 anhepatic rats with Table

1.

Results of Bacteriologic Investigations Anhepotic Rats at Sacrifice Blood culture

Species Germ-free (14) Normal w ” Number

negative

Culture

positive

of

of cecal contents

negative

positive

12

1

13

0

14

1

0

14

of experiments

are indicated

in parentheses.

HYPERAMMONEMIC

August 1979

COMA AFTER HEPATECTOMY

Aorta-Portal

Ammonia

RATS

233

Vein

umol/L

Normal

Germ-free

IN GERM-FREE

+300 48..

Oy

1

.

200

-----_-__ __ 1 1

&I _--

100

__-;__

x

i

o---

0 a . a

0 .

100 200

12

X

;___

X X

1 -300 1

Figure

1 I

,

1. 0nset of coma in rats after hepatectomy (0) or hepatic vascular exlusion (0). No difference is noted between germ-free rats and rats with normal intestinal bacteria, ,with either procedure.

normal intestinal flora. No difference was observed between animals with total hepatectomy and animals with hepatic vascular exclusion (Figure 1). The ammonia concentration in blood and cecal contents: Arterial ammonia concentrations were markedly elevated at the onset of coma, and also 24 hr after induction of the anhepatic state. No significant difference was observed in the arterial ammonia level between germ-free rats (mean + SD: 834 and normal rats (mean f SD: 854 f k 281 pmol/liter) 230 pmol/liter) (Figure 2). In contrast, ammonia concentrations in fecal water were significantly (P < 0.001)lower in germ-free rats (mean f SD: 1762 + 391pmol/liter) than in normal rats (mean + SD: 5572 f 2102 pmol/liter). Portal venous blood, however, contained more ammonia than arterial blood (mean Arterial

Ammonia

Germ

blood

-free

UlllOl/L 1200

800

‘-s-e----

600

Figure

2 Ammonia concentrations in arterial blood in dehepatized rats. No significant difference is noted between germ-free rats and animals with intestinal bacteria regardless of the procedure and time of bleeding. (Hepatectomy, at coma: 0; hepatic vascular exclusion, at coma: e and hepatic vascular exclusion, after 24 hr: x).

Figure

3. Difference

between

the ammonia

concentrations

aorta and the portal vein 24 hr after exclusion. The negative aorta-portal measured in germ-free rats, suggest testinal ammonia release.

in the

hepatic vascular vein differences, nonbacterial in-

zk SD A-V difference: -87 f 133 pmol/liter) in the germ-free animals of the second series (Figure 3).

7

Discussion Germ-free rats went into coma after hepatectomy in the same manner as normal rats. These observations confirm previous suggestions that toxins produced by intestinal bacteria play only a minor role in the pathogenesis of acute hepatic coma in normoglycemic animals. It is of interest that the mean duration of survival after dehepatization was considerably longer than usually reported.B.‘“20 The prolonged survival may be related to incomplete removal or devascularization of the liver around the caval vein, but is more likely attributable to the careful cleansing of the bowel and the absence of sepsis in our models.“‘~21 Ammonia rose markedly after hepatectomy or hepatic vascular exclusion in germ-free rats. Apparently ammonia, which served as a marker of gutderived bacterial toxins, can be the product of nonbacterial metabolism, as we found substantial intestinal ammonia release in germ-free rats. Previous work in germ-free Eck fistula dogs established that hyperammonemia and encephalopathy can occur.” The endogenous source of the increased ammonia was not conclusively established, but intestinal hydrolysis of urea could be ruled out.23.24Since blood ingestion in germ-free Eck fistula dogs caused marked elevation of systemic blood ammoniaZ3 and since a high-protein diet in germ-free guinea pigs was associated with a threefold rise in the portal blood ammonia level,” intestinal luminal contents were indicated as the source of nonbacterial am-

234

SCHALM

AND VAN DER MEY

monia. In our germ-free animals, the bowels were cleansed, and ammonia was still released into the systemic circulation by the portal vein. Deamination of glutamine in the intestinal wa1P2’ is probably the most important mechanism of ammonia production under these circumstances. The elevation of the blood ammonia concentration was such that hyperammonemia could be a major determinant of survival in our hepatic failure model. In our experiments, the ammonia concentrations, like those found by Schlienger and Imler,” were more than twice those reported by Degos et al ‘; the latter concluded that ammonia was of little importance in the mechanism of encephalopathy in the liverless rat, since differences in elevation of arterial ammonia concentrations between anhepatic and eviscerated rats were not reflected in duration of survival. Since for unknown reasons survival was much shorter in Degos et al. experiments” than in our study, additional information is needed concerning the role of ammonia in acute hepatic coma. The development of a functionally anhepatic animal model with normal blood glucose values, cleansed bowels, no sepsis, and normal ammonia concentrations will ultimately reveal the importance of ammonia in the pathogenesis of hepatic coma.

References 1. Zieve L: Pathogenesis of hepatic coma. Ann Rev Med 26:143157,1975 2. Schenker S, Breen KJ, Hoyumpa AM: Hepatic encephalopathy: current status. Gastroenterology 66:121-151,1974 3. Plum F, Hindfelt B: The neurological complications of liver diseaseIn: Handbook of Clinical Neurology 27. Edited by PG Vinken, AW Bruyn. New York, Elsevier, 1976.349-377 4. Wolpert E, Phillips SF, Summerskill WHJ: Ammonia production in the human colon. N Engl J Med 283:159-X4,1970 5. Phillips GB, Schwartz R, Gabuzda GJ, et al: The syndrome of impending hepatic coma in patients with cirrhosis of the liver given certain nitrogenous substances. N Engl J Med 247:239246.1952 6. Chen S, Zieve L, Mahadevan V: Mercaptans and dimethylsulfide in the breath of patients with cirrhosis of the liver. J Lab Clin Med 75:628-635,197O 7. Dawson, AM, McLaren J, Sherlock S: Neomycin in the treatment of hepatic coma. Lancet 2:1263-1269,1957

GASTROENTEROLOGY

Vol. 77, No. 2

8. Degos F, Degos JD, Bourdian D, et al: Experimental acute hepatic encephalopathy: comparison of the electro-encephalographic changes in the liverless and in the eviscerated rat. Clin Sci Mol Med 47:599-608,1974 W, Schwartz AE, Randall HT: Alterations in blood 9. Lawrence ammonia in dogs following total hepatectomy and abdominal evisceration. Surg Gyn Obstet 107:69-78,1958 as a factor causing death in 10. Ingle DJ, Nezamis JE: Infection the eviscerate rat. Proc Sot Exp Biol71:438-439,1949 JL, Imler M: Variations de I’ammoniemie et de la 11 Schlienger glutaminemie chez le rat hepatectomise et eviscere. Bordeaux Medical 10:1833-1940,1977 AH: Prevention of airborne con12 van der Waay D, Andreas tamination and cross-contamination in germ-free mice by laminar flow. J Hyg 69:83-89,197l F: Direkte Plasma-Ammoniakbestim13. Da Fonseca-Wollheim mung ohne Entweissung. Z Klin Chem Klin Biochem 11:426431,1973 BG, Leynse B: Evaluation of the Du Pont 14. Ypma ST, Blijenberg ACA ammonia procedure. Clin Chem 24:489-492,1978 of glucose-6-phosphatase 15. Bondar RJL, Mead DC: Evaluation dehydrogenase from leuconostic mesenteroides in the hexokinase method for determining glucose in serum. Clin Chem 20:586-590,1974 value of coagulation stud16. Veltkamp JJ, Kreuning J: Diagnostic ies in chronic liver disease. Stand J Gastroenterol 8: (Suppl. 18). 93-95,1973 TDV: Statistics at square one. Second edition. Lon17. Swinscow don, Britisch Medical Association, 1977 hepatectomy in the rat. J Appl 18. Bollman JL: A simple two-stage Physiol24:722-723,1968 S, Benhamou JP: L’insuffisance hepato19. Sicot Ch, Erlinger cellulaire experimentale. Press Med 79:1201-1208, 1971 FP: Total hepatectomy in the rat. Am J Physiol 20. Meehan 179:282-284,1954 as a manifesta21. Gans H, Mori K, Lindsey E, et al: Septicemia tion of acute liver failure. Surg Gyn Obstet 132:783-790,197l in germ22. Nance FC, Kline DC: Eck’s Fistula encephalopathy free dogs. Ann Surg 174:856-862,197l HJ, Kline DC: Role of urea in the hyper23. Nance FC, Kaufman ammonemia of germ-free Eck fistula dogs. Gastroenterology 66:108-112.1974 24. Levensom SM, Crowley LV, Horowitz RE, et al: Metabolism of carbon-labeled urea in the germ-free rat. J Biol Chem 234:2061-2062,1959 KS, Newton WL: Portal and peripheral blood am25. Warren monia concentrations in germ-free and conventional guinea pigs. Am J Physiol 197:717-720,1959 26. Felig Ph: Amino acid metabolism in man. Ann Rev Biochem 44:933-956,1975 27. Weber FL: The contribution of the small intestine to gut ammonia production in the fasting dog. Gastroenterology 74:1166, 1978