Serum bile acids and cholestasis in alcoholic hepatitis. Relationship with usual liver tests and histological features

Journal o f Hepatology 1994; 21 : 235-240 Pr#~ted hi Demnark . ,,111rights reserved Munksgaard. Copenhagen

Copyright © Journal of Hepatology 1994

Journal of Hepatology I S S N 0168-8278

Serum bile acids and cholestasis in alcoholic hepatitis. Relationship with usual liver tests and histological features J e a n - C l a u d e Trinchet 12, M a r i e - F r a n q o i s e G e r h a r d t 3, Beverley B a l k a u ~, C6cile M u n z 4 and Ren6e E. P o u p o n t tlNSERM Unit-21, Villejuif :Service d'HOpato-gastroentErologie, Hdpital Jean Verdier. Bondy. ~Service de Biochbnie, Hdpital Saint Joseph, Paris and ~Service d'Anatomopathologie. Hdpital Jean l"erclier. Bondy, France

(Received 10 June 1993)

Cholestasis is a biochemical and/or histological feature observed in some patients with alcoholic liver disease and is mainly related to alcoholic hepatitis. Accumulation of bile acids in the liver could be pathogenic in alcoholic hepatitis. The aim of this study was to assay serum bile acids in patients with alcoholic hepatitis and to assess the relationship between these parameters, the usual liver tests and the histological features of alcoholic hepatitis. Thirty-six patients (median 51 years, 19 females and 17 males) with biopsy-proven alcoholic hepatitis were included in the study. Cirrhosis was present in 27 patients. Serum bile acids were assayed by high performance liquid chromatography. Three histological scores (alcoholic hepatitis, fibrosis, and cholestasis) were established on each liver sample by two independent pathologists. Serum bile acid concentrations were increased in 35 patients (97%). The median concentration of total serum bile acids was 41.6/2mol/l (range 3-293), with an increase in primary bile acids (95.7% of total bile acids), mainly chenodeoxycholic acid (median 27.5 /2tool/l, range 3-184). In contrast, serum bilirubin levels were increased in only 26 patients (72%). Histological cholestasis was present in 14 patients (38%). There was no significant correlation between the alcoholic hepatitis and cholestasis scores (r=0.01, p=0.9). A significant correlation was noted between the alcoholic hepatitis score and serum total bile acid (r=0.34, p=0.04), cholic acid (r=0.38, p=0.03) and chenodeoxycholic acid (1"=0.32, p=0.05) levels. No correlation was noted between serum bile acid levels and cholestasis (r=0.13, p=0.4) or fibrosis scores (r=0.12, p=0.5). Conversely, the bilirubin level was correlated with the cholestasis score (r=0.40, p=0.02), but not with the alcoholic hepatitis (r=0.10, p=0.5) or fibrosis scores (r=0.17, p=0.31). This study suggests that the histological lesions of alcoholic hepatitis could be related, in part, to an increase in primary bile acids. Further studies are required to assess the effect of ursodeoxycholic acid as a treatment for alcoholic hepatitis. © Journal of Hepatology. Key words: Alcoholic liver disease; Chenodeoxycholic acid; Cholic acid

Alcoholic liver disease is characterized by the association, in various degrees, of three histological lesions, namely steatosis, fibrosis and alcoholic hepatitis (AH) (1). Biochemical and/or histological cholestasis may occur in alcoholic liver disease (2-5) and seems mainly related to AH (6,7). In alcoholic liver disease, an increase in serum bile acids has been reported in several studies of patients with alcoholic cirrhosis (8-18), but the presence of AH was usually undetermined because of the lack of a simultaneous liver biopsy. In some studies where histological

injury was assessed, the greatest increase in serum bile acids was observed in patients with alcoholic liver disease in whom necrosis or inflammation was present (8,12,14). In other respects, an increase in serum bile acid levels is an important feature in chronic cholestasis, as well as in biliary primary cirrhosis. It is now accepted that endogenous bile acids play a pathogenic role in this disease (19). By analogy, endogenous bile acids could play a role in the pathogenesis of alcoholic liver disease, especially in AH. In a recent study of patients with AH, histological

Correspondence to: Ren6e Poupon, INSERM U-21, 16, avenue Paul-Vaillant-Couturier, 94807 Villejuif Cedex, France.

236

J-C. TRINCHET et al.

cholestasis had a negative prognostic value for patient survival (5). To test this hypothesis, we performed a study in patients with AH to assess the relationships between total and individual serum bile acids, usual liver tests and histological liver injury.

Patients and Methods

Thirty-six patients with biopsy-proven AH were included in the study. Alcohol intake was more than 100 g/day in men and 80 g/day in women. Patients with an associated factor able to induce alcohol-unrelated liver disease were excluded. Serum AgHBs was negative in all patients, and none had evidence of histological viral chronic hepatitis. No patients were being treated with drugs capable of inducing AH-like histological lesions, and none had signs of obstructive biliary disease, in particular pancreatic disease. All patients had previously been included in one of three clinical trials (20-22). They were selected to include cases of both "minimal or moderate" and "severe" histological alcoholic hepatitis in order to obtain a wide spectrum of alcoholic hepatitis lesions. Patient characteristics are reported in Table 1. Selection was made without knowledge of the serum bile acid results. Serum and liver samples were simultaneously collected as part of standard patient management. Serum samples were collected after an overnight fast and were frozen at - 8 0 ° C until analysis. Serum levels of total bile acids, cholic acid, chenodeoxycholic acid, deoxycholic acid and lithocholic acid were measured by high performance liquid chromatography (HPLC), using a previously described method (23). Briefly, the method was as follows:

TABLE 1 Characteristics of 36 patients with biopsy-proven alcoholic hepatitis

Age (years) Sex (F) Encephalopathy Ascites Bilirubin (pmol/1) AST (n<40 IU/I) Alkaline phosphatase 01<200 IU/I] Gamma-glutamyltranspeptidase (n<45 IU/I) Albumin (g/l) Prothrombin time (%) Cirrhosis Alcoholic hepatitis score (0-8) Cholestasis score (0-12) Fibrosis score (0--4)

Median or no.

Range or %

51 19 10 16 44 101 248

(28-72) (53%) (28%) (44%) (6--477) (36--425) (110-570)

325 31 63 27 7 0 3.5

(30-2875) (19--49) (12-100) (75%) (I-8) (0-6) (0.54)

a specific three-step extraction with Bond Elut C18 cartridge (Varian, Les Ulis, France) and Lipidex column (Pacqard, Rungis, France), an H P L C separation with a reverse phase column and direct spectrophotometric detection at 199 nm; each conjugated bile acid was identified and quantified in comparison with the H P L C profile of a standard solution; the extraction rate was calculated by reference to an internal standard. The upper limit of normal for fasting total serum bile acids was 6 Mmol/l. Liver biopsy was performed percutaneously with the Menghini technique in 26 cases and by a transjugular route in 10 cases. All specimens were fixed in Bouin's fluid, paraffin embedded and stained with hematoxylin and eosin, trichrome and Sirius red techniques for quantification of fibrosis. All samples were larger than 10 mm. The minimal histologic criteria for diagnosis of AH were those established by an international group (1). Each biopsy was independently assessed by two liver pathologists, who were unaware of the clinical data. Three histological scores were established. A H was graded 0 to 8 by adding three values: hepatocellular necrosis and ballooning degeneration (0, none; 1, sparse; 2, some; 3, many), polynuclear infiltration of the lobule (0, none; 1, sparse; 2, some; 3, many) and alcoholic hyalin (0. none; 1, some; 2, many). Fibrosis was graded 0 to 4 (no fibrosis, 0; portal and periportal fibrosis, 1; portal and periportal fibrosis with perivenular fibrosis, 2; incomplete cirrhosis, 3: cirrhosis, 4). Cholestasis was graded 0 to 12 by adding four values (canalicular bile plugs, 0 to 3; canalicular dilatation, 0 to 3; bile accumulation in hepatocytes, 0 to 3; bile accumulation in macrophages, 0 to 3). The AH, fibrosis and cholestasis scores used in the analysis were the mean of the scores of the two pathologists. Reproducibility of tile scores was assessed by a blind rereading of 30 samples by the two pathologists and was judged as satisfactory (r=0.88 for the AH score, 0.86 for the cholestasis score and 0.83 for the fibrosis score). Numerical data are summarized by the median and the range. Distributions were compared by the Mann-Whitney test. Spearman correlation coefficients were used and the least squares regression line is shown on the scatter diagrams, p<0.05 was chosen as the limit of statistical significance.

Results

The results of the serum bile acid assessment are shown for each patient in Fig. 1. For the whole group, serum concentrations of total and individual bile acids are reported in Table 2. Serum total bile acids were above normal in 35 of the 36 patients, and were often markedly elevated. Serum total bile acids were not significantly dif-

SERUM BILE ACIDS AND ALCOHOLIC HEPATITIS

237

300 - -

250

ml.c

~" 200

F-l c~c

,.... 0

E

m

~c 150 m

I

I

E

2 0

lOO

50

Patients

Fig. 1. Serum bile acids in 36 patients with alcoholic hepatitis (C cholic acid, CDC chenodeoxycholic acid, DC deoxycholic acid, LC lithocholic acid).

ferent between patients with cirrhosis (31.3 pmol/l, range

was 0.48 (range 0-1.5), and was not significantly different

3.3-293)

between patients with cirrhosis (0.41, range 0-1.3) and pa-

and

patients

without

cirrhosis

(53.3 pmol/l,

range 6.8-244). S e r u m cholic, c h e n o d e o x y c h o l i c , deoxycholic, and

tients w i t h o u t cirrhosis (0.65, range 0.15-1.5).

lithocholic acids were, respectively, 29.8%

S e r u m bilirubin levels were increased ( > 1 7 / . t m o l / l ) in

(0-59.2), 62.2% (28.5-90.4), 2.2% (0-32.5), and 0% (0-9.1)

26 patients. S e r u m bilirubin and total bile acids were sim-

o f total s e r u m bile acids. S e r u m c h e n o d e o x y c h o l i c acid was p r e d o m i n a n t in 32 patients (89%), and d e o x y c h o l i c acid was absent in 16 patients (44%) (Fig. 1). S e r u m prim a r y bile acids, n a m e l y cholic acid and c h e n o d e o x y c h o l i c acid, were 95.7% (range 65.5-100) o f serum total bile acids, with no difference between patients with (95.4%,

TABLE 3 Correlations between biochemical and histological parameters in 36 patients with biopsy-proven alcoholic hepatitis Correlation

r

p

rhosis. T h e ratio o f cholic acid to c h e n o d e o x y c h o l i c acid

Bilirubin- total bile acids Bilirubin - alkaline phosphatase Total bile acids - alkaline phosphatase

0.59 0.0005 0.30 NS (0.07) 0.29 NS (0.09)

TABLE 2

Bilirubin - cholestasis score Bilirubin - alcoholic hepatitis score Bilirubin - fibrosis score

0.40 0.02 0.10 NS (0.5) 0.17 NS (0.3)

Total bile acids - cholestasis score Total bile acids - alcoholic hepatitis score Total bile acids - fibrosis score

0.13 NS (0.4) 0.34 0.04 0.12 NS (0.5)

range 65.5-100) and w i t h o u t (96"/0, range 85.3-100) cir-

Serum bile acids in 36 patients with biopsy-proven alcoholic hepatitis

Total bile acids (/Jmol/I) Cholic acid (umol/I) Chenodeoxycholic acid (pmol/l) Deoxycholic acid (,umol/l) Lithocholic acid (umol/l)

Median

Range

41.6 8.9 27.5 1.15 0

(3-293) (0-129) (3-184) (0-6) (0-7)

Alkaline phosphatase - cholestasis score 0.06 NS (0.7) Alkaline phosphatase- alcoholic hepatitis score 0.37 0.03 Alkaline phosphatase - fibrosis score -0.23 NS (0.2) Alcoholic hepatitis score - cholestasis score NS: not significant.

0.01 NS (0.9)

238

J-C. TRINCHET et al.

ultaneously normal in only one patient. N o n e of the patients showed increased serum bilirubin without increased serum bile acids. Serum alkaline phosphatase was increased in 23 patients (64%), but was more than twice the upper limit of normal in only three. Histological scores are reported in Table 1. Histological cholestasis (score > 0) was present in 14 patients (38%). Correlations between biochemical and histological parameters are reported in Table 3. The A H and cholestasis scores were not significantly correlated. Total serum bile acids were significantly correlated with serum bilirubin (Fig. 2) and the A H score (Fig. 3), but not with the cholestasis score. A significant correlation between the A H score and total primary serum bile acids (r=0.33, p=0.04) was noted. Correlations between the A H score and serum cholic and chenodeoxycholic acid levels were respectively 0.38 (p=0.03) and 0.32 (o=0.05). Individual serum bile acid levels were not correlated with either the cholestasis or fibrosis scores. Serum bilirubin was significantly correlated with the cholestasis

300 -

o o

o

250

.~,

r = 0.59

P = 0.0005

o 2oo 3 150 _.m IOO

E 50 O 0

•

0

i

•

50

i

I00

•

i

•

150

i

•

200

i

•

250

i

•

300

i

,

i

350

•

400

i

450

•

i

500

Serum bilirubin (pmol/L) Fig. 2. Correlation (Spearman test) between serum bile acids and serum bi]irubin in 36 patients with alcoholic hepatitis.

score, but not with the A H score (Fig. 4). The albumin level was not significantly correlated with any histological score.

Discussion

In this study of patients with biopsy-proven A H , total serum bile acids were increased, often markedly, in 97% o f the patients. Such an increase is well known in alcoholic cirrhosis (8-18), but few studies have assessed the correlations between bile acid concentrations and histological injury (8,12,14). In the present report, the increase in total serum bile acids was similar in patients with or without cirrhosis. Moreover, no significant correlation was noted between the fibrosis score and total serum bile acids. Conversely, a significant correlation was noted between the A H score and both total and primary bile acids. In the study by Milstein et al. (8), serum bile acids were increased to the same extent in inactive cirrhosis and in patients with A H without cirrhosis. Moreover, in the Aim et al. study (14), serum bile acids were more correlated with inflammation or necrosis than with fibrosis or cirrhosis. Conversely, serum bile acids were moderately elevated in the study of Von Bergmann et al. (11) in which none of the patients had histological AH. These results suggest that A H is an important cause of increased serum bile acids, independent of fibrosis and cirrhosis. The mechanisms of increases in serum bile acids are not completely known. In alcoholic cirrhosis, the pool o f bile acids is usually reduced (9). Increased serum bile acid levels are attributed to a decrease in hepatic captation of bile acids, and/or intra- or extrahepatic porto-caval shunts (13). In our study, primary serum bile acids were predominant, representing more than 95% of total bile acids.

500 "

300

O

450 " 250

400 -

r = 0.34 P = 0.04

.d

"-

"~ 200

r = 0.10 NS

350

.~ 300

-~ 15o

0

0

o

0 0

2 .D

0

.~ IOO E

o

t/}

O

50

O

0

o

o

I 0 i

0 i

I

2

o

8

o

o

i

i

i

i

i

.1"

4

5

6

7

8

3

Alcoholic

E

08

68

hepatitis score

Fig. 3. Correlation (Spearman test) between serum bile acids and alcoholic hepatitis score in 36 patients with alcoholic hepatitis.

250 200

O

150 o

lOO 0

0

50

0

o o t

i

1

2

o

o

g i

~1

3

4

5

i

~

i

i

6

7

o

@

§

~ 8

Alcoholic hepatitis score

Fig. 4. Correlation (Spearman test) between serum bilirubin and alcoholic hepatitis score in 36 patients with alcoholic hepatitis.

SERUM BILE ACIDS AND ALCOHOLIC HEPATITIS Chenodeoxycholic acid was the major serum bile acid in 89% of patients, representing up to 90% of total bile acids. The ratio cholic to chenodeoxycholic acid decreased in patients with cirrhosis (0.41) compared to those without cirrhosis (0.65), but not significantly. Previous studies have reported similar results in patients with alcoholic cirrhosis (8,13). These findings could be explained by a lack of hepatic synthesis of primary bile acids, especially cholic acid (9,24). Nevertheless, primary bile acid synthesis does not seem to be significantly modified in early cirrhosis (18,25). In our study, secondary serum bile acids, namely deoxycholic and lithocholic acids, were virtually absent in serum, comprising less than 5% of total bile acids. This percentage is close to 40% in normal subjects (26). Previous studies in alcoholic cirrhosis have shown a lack of deoxycholic acid in serum and bile (9,25,27). This finding could be explained by a lack of conversion of cholic acid to deoxycholic acid due to a marked decrease in 7a-dehydroxylase activity of fecal bacteria (25,27). In our study, the decrease was not significantly different between patients with or without cirrhosis. This result suggests that a lack of bacterial conversion may occur early in the course of alcoholic liver disease, antedating the development of established cirrhosis. The defect in bacterial 7adehydroxylase is probably related to inhibition by environmental factors (27) because the fecal microflora is not modified in alcoholic cirrhosis (28,29). Nevertheless, the precise mechanism is unknown. The prevalence of cholestasis in alcoholic liver disease varies according to the definition and selection of patients (2--4). In a recent study in 306 patients with alcoholic liver disease, increases in serum bilirubin and glycocholic acid were reported in 24% and 73% of patients, respectively, and significant histological cholestasis was noted in 22% (5). In this study, the selection of patients did not permit an assessment of the prevalence of cholestasis. Nevertheless, several abnormalities suggesting cholestasis were noted, such as increased serum bilirubin, alkaline phosphatase and total bile acids. Histological cholestasis was reported in 38°/,, of patients. Although a significant correlation between serum bile acids and bilirubin was noted, increased serum bile acids without increased bilirubin were reported in about 25°/,, of patients. These findings have already been observed in other studies, suggesting a disassociation between mechanisms responsible for increased serum levels of these substances (5,8,15,17). Serum bilirubin was significantly correlated with histological cholestasis, while serum alkaline phosphatase and total bile acids were not. In another study, no correlation between serum total bile acids and tissue cholestasis was noted (14).

239 In the present study, there was a correlation between total serum bile acid levels and histological lesions of AH, which is explained by correlations with both cholic and chenodeoxycholic acid levels. Previous data are conflicting, with a significant correlation between histological inflammation or necrosis and total serum bile acids in one study (14), but not in another (8). The lack of significant correlation in the present study between histological AH lesions and cholestasis suggests relative independence between the mechanisms responsible for these lesions. Nevertheless, a significant correlation between cholestasis and necrosis was noted in a previous study, in which histological cholestasis had a negative predictive value on survival (5). The pathogenesis of AH is complex and not completely known. Direct toxicity of ethanol, mainly due to its main metabolite, acetalde.hyde, is predominant (30). However, genetic and immunologic factors, nutritional deficiencies, B and C virus, and xenobiotics, probably play a role (30,31). The possible role of the latter two factors can be excluded in this study due to the patient selection criteria. These results suggest that endogenous bile acids could play a pathogenic role in AH, as in chronic cholestasis (19). Experimentally, endogenous bile acids are responsible for cholestasis in animals and have many deleterious effects on cellular functions (19). In particulal:, dihydroxy bile acids, chenodeoxycholic and deoxycholic acids, are hepatotoxic in humans (19). The clinical severity of AH is related both to a high early mortality in severe disease, and to its major role in the genesis of cirrhosis (31). Among the numerous treatments suggested, only corticosteroid therapy has been shown to be effective in reducing early mortality in severe AH (31,32). Recently, the clinical efficacy and tolerance of ursodeoxycholic acid in the treatment of primary biliary cirrhosis were demonstrated (33), and similar effects are likely in other chronic cholestasis (19). Treatment by ursodeoxycholic acid could be beneficial in AH as in other settings in reducing concentrations of endogenous cytotoxic bile acid such as chenodeoxycholic acid (34,35). Moreover, as previously suggested in chronic cholestasis (36), this treatment could favourably influence the immunologic factors which are thought to play a role in AH, such as aberrant HLA class I expression (37). In a recent randomized cross-over trial in 11 patients with alcoholic cirrhosis, a significant decrease in serum bilirubin, gamma-glutamyltranspeptidase and alanine aminotransferase was reported after 4 weeks of treatment with ursodeoxycholic acid (38). In conclusion, serum bile acid concentrations were almost always increased in AH. Total and primary serum bile acids were correlated with histological AH lesions.

240

J-C. TRINCHET et al.

F u r t h e r studies are r e q u i r e d to assess the effect o f ursod e o x y c h o l i c acid as a t r e a t m e n t for A H .

20.

References

21.

I. Review by an International Group. Alcoholic liver disease: morphological manifestations. Lancet 1981 ; i: 707-11. 2. Harinasuta V, Chomet B, Ishak K, Zimmernlan HJ. Steatonecrosis - Mallory body type. Medicine 1967; 46: 141-62. 3. Glover SC, McPhie JL. Brunt PW. Cholestasis in acute alcoholic liver disease. Lancet 1977; ii: 1305-7. 4. Morgan MY, Sherlock S, Scheuer PJ. Acute cholestasis, hepatic failure, and fatty liver in the alcoholic. Scand J Gastroenterol 1978; 13: 299-303. 5. Nissenbaum M, Chedid A, Mendenhall C, Gartside P. Prognostic significance of cholestatic alcoholic hepatitis. VA Cooperative Study Group. Dig Dis Sci 1990; 35: 891-6. 6. McGill DB. Steatosis. cholestasis, and alkaline phosphatase in alcoholic liver disease. Am J Dig Dis 1978: 23: 1057-60. 7. Perrillo RP, Griffin R, De Schryver-Kecskemeti K, Lander JL, Zuckerman GR. Alcoholic liver disease presenting with marked elevation of serum alkaline phosphatase. A combined clinical and pathological study. Dig Dis 1978; 23: 1061-6. 8. Milstein H J, Bloomer JR, Klatskin G. Serum bile acids in alcoholic liver disease. Comparison with histological features of the disease. Dig Dis 1976: 21: 281-5. 9. Vlahcevic ZR, Prugh MF, Gregory DH, Swell L. Disturbances of bile acids metabolism in parenchymal liver cell disease. Clin Gastroenterol 1977; 6: 25--43. 10. Struthers JE, Metha SJ, Kaye MD, Naylor JL. Relative concentrations of individual nonsulfated bile acids in the serum and bile of patients with cirrhosis. Am J Dig Dis 1977; 22: 861-5. I 1. Von Bergmann K, Mok HY, Hardison WGM, et al. Cholesterol and bile acid metabolism in moderately advanced stable cirrhosis of the liver. Gastroenterology 1979; 77:1183-92. 12. Tobiasson P, Boeryd B. Serum cholic and chenodeoxycholic acid conjugates and standard liver function tests in various morphological stages of alcoholic liver disease. Scand J Gastroenterol 1980; 15: 657-63. 13. Poupon R, Poupon RE, Lebrec D, Le Quernec L, Darnis F. Mechanisms for reduced hepatic clearance and elevated plasma levels of bile acids in cirrhosis. Gastroenterology 1981: 80: 1438--44. 14. Aim R, Carlson J, Eriksson S. Fasting serum bile acids in liver disease. A comparison with histological features. Scand J Gastroenterol 1982; 17: 213-8. 15. Mannes GA. Increased serum bile acids in cirrhosis with normal transaminases. Digestion 1982: 25: 217-21. 16. Islam S, Poupon RE, Barbare JC, Chr6tien Y, Darnis F, Poupon R. Fasting serum bile acid level in cirrhosis. A semi-quantitative index of hepatic function. J Hepatol 1985; 1: 609-17. 17. Einarsson K, Angelin B, Bj6rkhem I, Glaumann H. The diagnostic value of fasting individual serum bile acids in anicteric alcoholic liver disease: relation to liver morphology. Hepatology 1985; 5: 108-11. 18. J6nson G, Hedenborg G, Wis6n, Norman A. Serum concentrations and excretion of bile acids in cirrhosis. Scand J Clin Lab Invest 1992; 52: 599-605. 19. Poupon RE, Poupon R, Ursodeoxycholic acid for the treatment of cholestatic diseases. In: Boyer JL, Ockner RK, eds. Progress

22.

23.

24.

25.

26. 27.

28.

29.

30.

31. 32.

33.

34.

35.

36.

37.

38.

in Liver Diseases, Vol. X. Philadelphia: Saunders WB, 1992: 219-38. Trinchet JC, Coste T, L6vy VG, et al. Traitement de l'h6patite alcoolique par la silymarine. Une 6tude comparative en double insu chez 116 malades. Gastroenterol Clin Biol 1989; 13: 120~.. Trinchet JC, Beaugrand M, Callard P, et al. Treatment of alcoholic hepatitis with colchicine. Results of a randomised double blind trial. Gastroenterol Clin Biol 1989: 13: 551-5. Trinchet JC, Balkau B, Poupon RE, et al. Treatment of severe alcoholic hepatitis by infusion of insulin and glucagon. A multicentre sequential trial. Hepatology 1992; 15: 76-81. Labb6 D, Gerhardt MF. Myara A, Vercambre C, Trivin F. High-performance liquid chromatographic determination of tauro- and glyco-conjugated bile acids in human serum. J Chromatogr 1989: 490: 275-284. Patteson TE, Vlahcevic ZR. Bile acid metabolism in cirrhosis. VI. Site of blockage in the bile acid pathways to primary bile acids. Gastroenterology 1980: 79: 620-8. Yoshida T, McCormick WC, Swell L, Vlahcevic ZR. Bile acid metabolism in cirrhosis. IV. Characterization of the abnormality in deoxycholic acid metabolism. Gastroenterology 1975; 68: 335-41. Paumgartner G. Serum bile acids. Physiological deternlinants and results in liver disease. J Hepatol 1986: 2: 291-8. Knodell RG. Kinsey MD, Boedeker EC, Collin DP. Deoxycholate metabolism in alcoholic cirrhosis. Gastroenterology 1976: 71: 196--201. Floch MH, Katz J, Conn HO. Qualitative and quantitative relationships of the fecal flora in cirrhotic patients with portal systemic encephalopathy and following portacaval anastomosis. Gastroenterology 1972; 62: 275-9. Lai D, Gorbaeh SL, Levitan R. Intestinal microflora in patients with alcoholic cirrhosis: urea splitting and neomycin resistance. Gastroenterology 1970; 59: 70-5. Lieber CS. Biochemical and molecular basis of alcohol-induced injury to liver and other tissues. N Engl J Med 1988; 319: 1639-50. Maddrey WC. Alcoholic hepatitis: pathogenesis and approaches to treatment. Scand J Gastroenterol 1990: 25: 118-30. Ramond M J, Poynard T, Rueff B. et al. A randomized trial of prednisolone in patients with severe alcoholic hepatitis. N Engl J Med 1992; 326: 507-12. Poupon RE, Balkau B, Eschw6ge E, et al. A multicenter, controlled trial of ursodiol for the treatment of primary biliary cirrhosis. N Engl J Med 1991; 324: 1548-54. Poupon RE, Chr6tien Y, Poupon R, Paumgartner G. Serum bile acids in primary biliary cirrhosis: effect of ursodeoxycholic acid therapy. Hepatology 1993; 17: 599-604. Ohiwa T, Katagiri K, Hoshino M, Hayakawa T, Nakai T. Tauroursodeoxycholate and tauro-,O-muricholate exert cytoprotcction by reducing intrahepatocyte taurochenodeoxycholate content. Hepatology 1993; 17: 470-6. Calmus Y, Gane P, Rouger P, Poupon R. Hepatic expression of class I and class II major histocompatibility complex molecules in primary biliary cirrhosis: effect of ursodeoxycholic acid. Hepatology 1990; 11: 12-5. Zetterman RK. Autoimmune manifestations of alcoholic liver disease. In: Krawitt EL, Wiesner RH, eds. Autoimmune liver diseases. New York: Raven Press, 1991, 247-60. Plevris JN, Hayes PC, Bouchier lAD. Ursodeoxycholic acid in the treatment of alcoholic liver disease. Eur J Gastroenterol Hepatol 1991; 3: 653-6.