Europ.J.clin.Invest. 5 , 1-6 (1975)
Canalicular Bile Production in Man* D. Prandi,
S.
Erlinger, J.-C. Glasinovit, and M. Dumont
Clinique Chirurgicale, and Unitk de Recherches de Physiopathologie Hipatique (INSERM), H6pital Beaujon, Clichy, France Received: January I , 1974, and in revised form: April 19, 1974
Abstract. Bile flow, erythritol clearance (measured after administration of l4C-erythrito1) and bile acid excretion were studied in 8 non-icteric patients operated upon for gallstone disease without biliary obstruction in whom a T-tube had been inserted. Bile flow increased from the 2nd (0.27 2 SEM 0.01 ml/min) to the 7th (0.50 f SEM 0.05 ml/min) postoperative days and remained stable thereafter. Because the values obtained between the 7th and the 15th postoperative days (average: 0.46 ? SEM 0.02 ml/min) were not lower and were of the same order of magnitude as those previously reported in the literature for quantitative bile collections, it was assumed that bile collection had been complete. Bile-to-plasma concentration ratio of erythritol was 0.77 f SEM 0.04. The regression line between erythritol clearance and bile flow had a slope of 0.99 (indicating that increments in clearance were followed by equivalent increments of flow) and an intercept for a zero-clearance of 0.11, suggesting a ductular/ductal secretion of 0.11 ml/min. The relationship between erythritol clearance and bile acid excretion suggested that canalicular bile was formed both of a bile acid-dependent flow ( 1 1 p 1 of water being excreted per pmol of bile acid secreted) and a bile acid-independent flow of 0.16 ml/min (or 35 % of the average bile flow). These findings indicate that erythritol clearance provides a valid estimate of canalicular bile flow in man and that the three bile fractions previously postulated in animals (canalicular bile acid-dependent, canalicular bile acid-independent and ductular/ductal) may also be present in man. Key words: Bile, bile formation, bile acids, choleresis, man, biliary clearance, erythritol.
Most of our current knowledge on bile formation comes from animal experiments. It has been proposed that bile was formed of three fractions (1,2,3): a) the hepatocytic (canalicular) bile'acid-dependent fraction, which is considered to be the result of the active secretion of bile acids by the hepatocytes into the canaliculi,with subsequent osmotic flow of water and electrolytes; b) the hapatocytic (canalicular) bile acid-independent fraction, which may be due to active inorganic ion transport; c)the ductal or ductular fraction due to the secretion, presumably by the ductal or ductular epithelium, of a bile-acid free electrolyte solution, mainly in response to secretin administration. Considerable species variations in the process of bile formation have been observed. fie existence in man of basic mechanisms of bile secretion established in animals remains therefore uncertain and extrapolation toman of animal data is hazardous. me present studies, in patients with gallstone disease withcarried Out bile duct obstruction, were undertaken to obtain a better definition of the mechanisms of bile floy in man, and, in particular, to measure bile flow, *Presented in part at the 8th meeting of the European Association for the Study of the Liver, Vittel, France, September 7-9, 1973 (Digestion, 8, 437 (1973).
bile acid excretion and canalicular bile production, estimated by the biliary clearance of 14C-erythritol. The results suggest: 1 ) that 14C-erythritol clearance provides a valid estimate of canalicular bile flow in man, and 2 ) that the free fractions of bile postulated in animals seem to bepresent in man.
Patients and Methods Patients. Eight patients aged 47 to 83 yearswere operated upon for gallstone disease. None was jaundiced or had any other clinical sign of biliary obstruction. Their selection depended upon the surgeon's indication for choledochotomy: a cholecystectomy was performed and, because of dilatation of the c m i ~ ~ obile n duct, a choledocotomy was done, the paduct was verified and a Ttency Of the common tube (external diameter: 3.4 m, internal diameter: 2.3 mm) inserted. Pertinent clinical, biochemical and histological data are sumharized in Table I . During the post-operative period, the patients received a routine ward care: they,were given 5 per cent glucose and/or electrolyte intravenous infusions during 3-4 days, and were put on the ordinary hospital diet as soon as intestinal peristalsis was complete.
~.Prandie t a l . Canalicular Bile Production in Man
L
Table I . Clinica1,histological and biochemical findings Age
Operative findings
Hepatic histology
Sex
Liver function tests ('1 Serum alkaline phosphatase SGPT bilirubin mg Z
Haz..
100
0.8
3.5
50
I00
0.8
5
20
-
acute pancreatitis normal multiple gall-stones
0.8
2
60
I00
gallbladder empyema
0.8
9
95
I00
0.8
4.5
45
87
2.5
4
10
1 00
gall-stones, chronic cholecystitis
-
perforated duodenal ulcer; multiple gall-stones
-
Szy.. 6 4
acute cholecystitis
42 m
Thie.. 50
hepatic granuloma
slight cholestasis
f Am..
-
11
Sai.. 4 9
f
x
2
multiple gall-stones
Pi0... 69
Frankel units
0.3
47 f
m
Bodansky units
Prothrombin concentration (Quick)
normal
f
Ben.. 55
gallbladder empyema
-
m
Che.. 83 f (O)
multiple gall-stones cholangitis
slight cholestasis
obtained at time of study
Bile collections were carried out once a day, in the morning, after an overnight fast, from the 2nd to the 15th postoperative days, during 1 to 2 hours, in sucessive 15-minute samples, from the external branch of the T-tube, which had previously been occluded for twelve hours. Bile flow (taken as the mean flow during each 15minute collection period) and bile acid excretion were obtained during all collections; biliary clearance of erythritol (determined as indicated under "methods") was obtained once in each patient, on days 7 to 15 postoperatively, from two 15-minute samples obtained at the end of the 2-hour collection period, at least 90 minutes after the beginning of the infusion of erythritol (see "methods").
beginning of the infusion, and two plasma samples taken at the midpoint of each bile collection period were used to measure bile flow, and bile and plasma 14C activity for each clearance determination. 50 to 100 p1 of bile or plasma were deposited on Whatman glass fiber discs (6), dried, and counted in 1 ml of NE 213 (Nuclear Enterprises, Edinburgh, Scotland) scintillation medium;correction for quenching was made by internal standardization. Erythritol clearance was calculated as bile flow times the bile-to-plasma ratio of I4C activity. Statistical analysis of the results was performed using the Student's t test or theWilcoxon T test (7). Results
Methods. Bile acid concentration in bile was measured by an enzymatic technique using 3a hydroxysteroid dehydrogenase (Worthington Biochemical Corporation, Freehold, N.J., USA). Details of the technique are given elsewhere ( 4 , 5). When biliary clearance of erythritol was measured, I4C-labelled erythritol (The Fadiochemical Centre, Amersham, Buchinghamshire, U.K.) was administered intravenously as a single injection of 1 VCi, followed by a constant infusion of 0.1 vCi/min in glucose (50 g per 1 ) at a rate of 0.1 mllmin; two 15-min bile samples taken at least 90 minutes after the
The evolution of bile flow in the postoperative period is shown in Fig. 1 . Bile flow increasedfrom the 2nd (0.27 2 SEM 0.01 ml/min) to the 7th (0.50 f SEM 0.05 mllmin) postoperative days and tended to remain stable thereafter. The average flow was significantly lower on the 10th than on the other days: this finding was considered to be due to a random fluctuation. The average bile flow from day 7 t o day 15 was 0.46 f SEM 0 . 0 2 ml/min. The relationship between bile flow and bile acid excretion is represented in Fig. 2 . All values ob-
D.Prandi et a l . Canalicular Bile Production in Man
3
Table 2. Bile flow, bile acid concentration, erythritol clearance and canalicular flow in eight patients (at the time of erythritol clearance measurement) Patient Day
Haz..
10
Sai..
11
Bile flow
Bile acid excretion
Erythritol.(B)/(P)
Erythritol clearance
Canalicular flow
ml/min
umo l/min
0.30 0.32
5.1 6.6
0.68 0.76
0.20
0.19
0.24
0.21
0.62
37.4 28 .o
0.85
0.80
0.53 0.38
0.51
0.48
ml/min
ml/min
0.37
Pio..
9
0.44 0.44
9.8 9. I
0.68 0.16
0.30 0.33
0.33 0.33
Say..
7
0.31 0.42
9.2 9.7
0.78 0.78
0.29 0.33
0.26 0.31
ha..
14
0.41
0.48
34.8 34.7
0.ia 0.12
0.31 0.35
0.36 0.31
The..
13
0.44 0.36
22.4 17.8
0.66 0.12
0.29 0.26
0.33 0.25
.
15
0.88
50.3 50.1
0.82 0.95
0.12 0.81
0.71 0.14
0.10
22.8 29.1
0.12 0.19
0.41 0.55
0.46 0.59
0.50 0.04
22.2 3.9
0.11 0.02
0.40
0.40 0.04
Ben.
0.85
Che..
13
Mean SEM
0.51
0.04
Calculated as total bile flow minus ductal/ductular ("distal") flow obtained from Fig. 3
~
1-
. .... .. .. .- . . . .. .. X
Bile flow
'
BILE FLOW
~
-
(ml/min)
i
(ml/min) . Q5-
'ii+
.dF
0
.
=
0.807
0 .
Fig. 1 . Evolution of bile flow in the postoperative period. Each point represents flow during a j5-minute period. 4 to 8 measurements were obtained on a given day, during a 1 to 2 h collection Period. The mean flow (f SEM) was on day 2: 0.27 f 0.01, on day 3: 0.27 f 0.01, on day 5: 0.28 +- 0.01, on day 6: 0.42 2 0.02, on day 7: 0.50 f 0.05, on day 8: 0.32 f 0.03, on day 9: 0.53 f 0.06, on day 10: 0.29 f 0.03, on day 1 1 : 0.48 f 0.05, on day 13: 0.54 f 0.13 and on day 15: 0.60 f 0.05 ml/min
r
rC. .a
'0
mtx.
0 '
0.
'
.
y= a +
-
bx
a 0.27 ml/min b 0.010 ml/pmol
Fig. 2. Relationship between bile flow and bile acid excretion in man. Each point represents one measurement on successive 15-minute periods, obtained during each daily collection between the 7th and the 15th postoperative days in the 8 patients
D.Prandi et a l . Canalicular Bile Production in Man
4
a75
i
Erythritol clearance is linearly related to bile acid excretion rate. The slope of the regression line indicates that erythritol clearance increases 1 1 ~1 when bile acid excretion increases 1 pmol. The extrapolated clearance for a zero bile acid excretion is 0.16 ml/min.
/
Bile flow (mllmin)
Discussion r = 0.98 y.a+bx a * 0.n (mlAin)
/
I' 0
I
/
b = 0.99 (ml/ml) 0.25
I as0 0.75 Erythritol claarana (ml/min~
Fig. 3 . Relationship between bile flow and erythritol clearance. Each point represents one measurement (obtained during a 15-minute period)
1. V a l i d i t y of measurements. The present study was designed to characterize normal human bile formation. Three possible limitations in reaching this purpose will be discussed in this section: a) can the patients be considered normal? b) was bile collection complete? c) is erythritol clearance a valid estimate of canalicular bile flow?
Hepatic bile collection in man is only possible in patients with T tubes; therefore it has to be limited to patients with gallstone disease. However, none of the patients had biliary obstruction; routine liver function tests were normal (or almost normal) at the time of study and the liver biopsy Erythritol . - when performed - revealed no abnormalities or only minor ones. In addition to the disease, two clearance other shortcomings should be considered: the postoperative alterations in biliary function, and the interruption of the enterohepatic circulation. Postoperative alterations in biliary function were reflected in our study by a low bile flow from the 2nd to the 6th postoperative days; a low flow has already been observed in the early postoperative 1 0' 10 20 30 40 50 60 period in man (8, 9) and in the rhesus monkey (10); in the rhesus monkey, the decreased flow was asBile acid excretion ( pmol/min) sociated with a decreased secretion of bile acids, Fig. 4 . Relationship between erythritol clearance phospholipids and cholesterol and returned to norand bile acid excretion. Each point represents one mal within 2 weeks (10). In man, biliary function measurement obtained in the same samples as those has been considered to return to normal within apdepicted on Fig. 3 proximately one week (8, 1 1 , 12); this is in agreement with the stabilization of bile flow 7 days after oDeration in our patients and representative tained between the 7th and the 15th postoperative days have been included. There is a linear relation- data (including erythriiol clearance) have been ship between bile flow and bile acid excretion rate; obtained between the 7th and the 15th postoperative days. Interruption of the enterohepatic circulation the slope of the calculated regression line indicates a flow of 10 ~1 per pmol of bile acid exis followed within 1-3 hours by a decrease of bile flow and of secretion of total solids, phospholicreted. The extrapolated flow for a zero bile acid excretion is 0.27 ml/min. pids and cholesterol ( 1 2 ) ; these changes are reversible within 1-3 hours after refeeding of bile ( 1 2 ) ; bile was collected in our patients while the Bile flow, bile acid concentration, bile-toexternal branch of the T-tube had been occluded for plasma concentration ratios of I4C-erythritol and at least 12 hours, thus minimizing changes related erythritol clearance, recorded when erythritol to interruption of the enterohepatic circulation. clearance was measured in the 8 patients are indicated in Table 2 . The bile-to-plasma concentration The measurement of erythritol clearance (and bile flow) has been made after approximately 2 hours of ratio of erythritol is 0.77 2 SEM 0.02, and siginterruption of the enterohepatic circulation: bile nificantly lower than 1 (p < 0.001). As a result, flow was, on the average, lower at the time of erythritol clearance is significantly lower than bile flow (p c 0.001). The relationship between measurement than at the beginning of bile collection. Our values are therefore slightly underestibile flow and erythritol clearance is represented mated. Moreover, changes due to partial chronic in Fig. 3 . The slope of the regression line (0.99 interruption of the enterohepatic circulation canf SEM 0.01) is not significantly different from 1 . not be excluded. Thus, our patients cannot be conThe extrapolated flow for a zero clearance is 0.11 ml/min. sidered entirely normal; however, major abnormalities due to obstruction of bile ducts, surgery or The relationship between erythritol clearance interruption of the enterohepatic circulation have and bile acid excretion is represented in Fig. 4 . been avoided as much as possible.
/./
I
D.Prandi e t aZ. Canalicular Bile Production in Man Since the T-tube was not fitted with a balloon, the possibility that bile collection was not complete has been considered. However, the average flow obtained in this study (0.46 2 SEM 0 . 0 2 ml/ min) was not lower and was of the same order of magnitude as those previously reported by Preisig e t aZ. ( 0 . 3 7 2 SEM 0.10 ml/min) (13) and Shersth e t a l . (0.41 ml/min) ( 1 4 ) in patients with balloonequipped catheters in whom quantitative collections were presumably obtained. The average bile flow in our study would correspond to a daily output of bile of approximately 660 ml, a figure again similar to previous estimates of daily bile production in patients in whom quantitative collections had been possible (8, 1 1 , 12). It was therefore assumed that bile collection was complete or nearly complete. Moreover, it should be noted that the loss of a certain amount of bile would not modify the proportions of each constituent of bile flow and, therefore, would not alter the conclusions of the study
.
Measurement of erythritol clearance has been used to estimate canalicular bile flow in guineapigs ( 1 5 , 16), dogs ( 1 7 ) , rabbits (18) and rats ( 4 , 19, 2 0 ) . The validity of this measurement rests on the assumption that erythritol, an inert 5-carbon carbohydrate, enters canalicular bile through simple passive processes and does not move into or out of the biliary channels. As seen in Table 2 , the bile-to-plasma concentration ratio of erythritol was significantly lower than one, on the average 0 . 7 7 . Theoretically the finding of a biliary concentration of the solute lower than its plasma concentration might have been the result of either a limited "permeability" of the canalicular membrane to erythritol relative to water (with incomplete equilibrium of erythritol between plasma and canalicular bile), or, alternatively, of a dilution of canalicular bile by distal (ductular or ductal) fluid secretion. Since it is impossible to obtain canalicular bile, direct evidence for either possibility is lacking. However indirect evidence for the second explanation (dilution by distal fluid secretion) is obtained from the finding that the slope of the relationship between erythritol clearance and bile flow (Fig.3) was close to unity: this means that increments in clearance were followed by equal increments in flow, a finding which would not occur if the "permeability" of the canalicular membrane to erythritol was significantly lower than that of water; the extrapolation of the regression line between clearance and flow to a zero clearance gave a positive intercept of 0.11 mllmin: this would be expected if a distal secretion of 0.11 ml/min had occurred. Calculation of canalicular flow (as total bile flow minus "distal" flow) yielded values very close to the measured erythritol clearance (Table 2 ) . These observations are consistent with the hypothesis that erythritol clearance measures canalicular flow in man. It is interesting to note that in another study in patients (21) the bile-to-plasma concentration ratio and the clearance of mannitol, a somewhat larger solute, were slightly lower than those of erythritol in our study. This is consistent with
5 the view that the "permeability" of the canalicular membrane to mannitol is lower than that of erythritol, a finding already suggested by studies in the guinea-pig (16) and in the rabbit (18). 2. CanaZicuZar b i l e production in man. The study of the relationship between canalicular bile production (as estimated by erythritol clearance) and bile acid excretion sugges.ts the following comments. Firstly, canalicular bile production was proportional to bile acid excretion rate. This is presumably the result of a bile acid-dependent flow, as demonstrated in the dog ( l 7 ) , the rabbit (18) and the rat ( 4 , 19, 2 0 ) . Our study indicates that 1 1 p1 of canalicular bile are formed per pmol of bile acid excreted, a figure of the same order of magnitude as those reported in dogs and rats ( 1 7 , 19, 20) and, also, of the same order of magnitude as that obtained by Scherstin e t aZ. in man ( 1 4 ) from a study of the relationship between bile flow and bile acid excretion. Values obtained in the rabbit (18) and in man by Preisig e t a t . (13) were somewhat higher; the reason for this difference is not clear, but may have been related to methodological differences, regarding in particular, bile acid measurements in bile.
Secondly, extrapolation of the relationship between erythritol clearance and bile acid excretion gave a positive intercept of 0.16 ml/min. This is consistent with the existence of a bile acid-independent flow of canalicular origin of 0.16 ml/min or 35 X of total bile flow. An estimate of 0.13 ml/min, thus very similar to ours, has been obtained by mannitol clearance measurement (21). A canalicular bile acid-independent flow has already been postulated in the dog ( 1 7 ) , the rabbit ( l a ) , the rat (19, 20) and suggested (although not supported by erythritol clearance measurements) in the hamster (22) and the rhesus monkey ( 2 3 ) . Quantitatively, this fraction of canalicular bile, although ranging from approximately 5 pl/min/kg in the dog (17) to 60 pl/min/kg in the rabbit (18) , represents, in all species studied, 35 to 6 0 X of "basal" bile flow. It would therefore appear that, in man as in other species studied so far, canalicular bile is formed o f : a) a bile acid-dependent fraction, amounting to, on the average, 0.19 ml/min, and resulting from the excretion.of I 1 p1 of water per umol of bile acid secreted; b) a bile acid-independent fraction amounting to 0.16 ml/min. 3. Ductular or ductal f l u i d secretion. When the regression line between bile flow and erythritol clearance was extrapolated to the origin, a flow of 0.11 ml/min was obtained; another way of expressing this finding is to compare the extrapolated values for a zero bile acid excretion of bile flow ( 0 . 2 7 ml/min) and erythritol clearance (0.16 ml/min). These observations are consistent with secretion of 0.11 ml/min of fluid at sites distal to the d c u l i , i.e. in the ductules or
6
ducts. A ductular or ductal secretion has been strongly suggested in dogs (17, 24), but has not been found in rabbits or rats (18, 19, 20). In man, a "total" bile acid-independent bile flow (including the canalicular bile acid-independent and the ductular/ductal fractions) of 0.I5 to 0.20 ml/&n has been found in similar patients (13, 14), part of which has been postulated to be of ductular/ ductal origin and stimulated by biliary obstruction (13). In our patients, the moderate dilatation of the bile ducts may have contributed to the ductular/ ductal secretion. In the dog, the ductular/ductal fraction is stimulated by secretin administration (17); a similar observation has been made in man (21).
D.Prandi e t al. Canalicular Bile Production in Man
following choledochotomy. Acta chir. scand. 133, 483 (1967) 10. Redinger, R.N., Herman, A.R., Small, D.M.: The effects of surgery on hepatic physiology in the -primate. (abstract). Gastroenterology 60, 795 (1971) 1 1 . Mollowitz, G.: Beobachtungen der Gallensekretion des Menschen. Langenbecks Arch. klin. Chir. 291, 359 (1959) 12. Thureborn, E.: Human hepatic bile. Composition changes due to altered enterohepatic circulation. Acta chir. scand. (supp1.303), 1 (1962) 13. Preisig, R.,Bircher, H., Stirnemann, H., Tauber, J.: Postoperative choleresis following bile duct obstruction in man. Rev. franc. Etud. clin. biol. 14, 151 (1969) 1 4 . Scherstgn, T., Nilsson, S., Cahlin, E., FilipIn conclusion, erythritol clearance seems to son, M . , Brodin-Persson, G.: Relationship beprovide a valid estimate of canalicular bile flow tween the biliary secretion of bile acids and in man; canalicular bile production seems to be the the secretion of water, lecithin, and cholesresult of both a bile acid dependent-flow and a bile acid-independent flow; canalicular bile is terol in man. Europ. J. clin. Invest. 1 , 242 (1971) diluted by ductular or ductal fluid secretion. These observations support the view that basic de15. Forker, E.L.: Two sites of bile formation as determined by mannitol and erythritol clearance terminants of bile formation previously established in animals are also present in man. However, in the guinea pig. J. clin. Invest. 46, 1189 ( 1967) further studies are required to better define the mechanisms of elaboration and control of the vari16. Forker, E.L.: Bile formation in guinea pigs: ous fractions of bile in man. analysis with inert solutes of graded molecular radius. Amer. J. Physiol. 215, 56 (1968) 17. Wheeler, H.O., Ross, E.D., Bradley, S.E.: CanaAcknowledgements. This study was supported in licular bile production in dogs. h e r . J. part by a grant from INSERM (ATP no 9). Physiol. 214. 866 (1968) 18. Erlinger, S., Dhumeaux, D., Berthelot, P., Dumont, M.: Effect of inhibitors of sodium transport on bile formation in the rabbit. Amer. J. References Physiol. 219, 416 (1970) 1. Wheeler, H.O.: Secretion of bile acids by the 19. Boyer, J.L., Klatskin, G.: Canalicular bile flow liver and their role in the formation of hepand bile secretory pressure. Evidence for a nonatic bile. Arch. intern. Med. 130, 533 (1972) bile salt dependent fraction in the isolated 2. Erlinger, S.: Physiology of bile flow. In: Properfused rat liver. Gastroenterology 59, 853 gress in Liver Diseases, H. Popper and F. ( 1970) Schaffner (Eds.), Vol. 4, p. 63. New York and 20. Boyer, J.L.: Canalicular bile formation in the London: Grune and Stratton, 1972 isolated perfused rat liver. Amer. J. Physiol. 3. Erlinger, S., Dhumeaux, D.: Mechanisms and con221, 1156 (1971) trol of secretion of bile water and electro21. Boyer, J.L., Bloomer, J.R., Canalicular bile selytes. Gastroenterology 66, 281 (1974) cretion in man. Studies utilizing the biliaryclear4. Berthelot, P., Erlinger, S., Dhumeaux, D., ance o f [14C] mannitol. J.clin.Invest. 54, 773 Prgaux, A.M.: Mechanisms of phenobarbital in(1974) duced hypercholeresis in the rat. Amer. J. 22. King, J.E., Schoenfield, L.J.: Cholestasis inPhysiol. 219, 809 (1970) duced by sodium taurolithocholate in isolated 5. Erlinger, S.. Dumont, M.: Influence of theohamster liver. J. clin. Invest. 50, 2305 (1971) phylline on bile formation in the dog. Biomedi23. Redinger, R.N., Small, D.M.: Primate biliary c;ne 19, 27 (1973) physiology. VIII. The effect of phenobarbital 6. Davies, J.W., Cocking, E.C.: Liquid scintilupon bile salt synthesis and pool size, biliary lation counting of I4C and 3 H samples using lipid secretion, and bile composition. J. clin. glass-fibre or filter paper discs. Biochim. Invest. 52, 161 (1973) biophys. Acta Amst. 115, 511 (1966) 24. Wheeler, H.O., Mancusi-Ungaro, P.L.: Role of 7. Snedecor, G.W., Cochran, W.G.: Statistical bile ducts during secretin choleresis in dogs. methods, Ames (Iowa): The Iowa University Press, Amer. J. Physiol. 210, 1153 (1966) 1967 8. Rundle, F.F., Cass, M.H., Robson, B., Middleton, M.: Bile drainage after choledocostomy in man, Dr. Serge Erlinger with some observations on biliary fistula. SurUnit& de Recherches de gery 37, 903 (1955) Physiopathologie Hspathique 9. Guldhammer, E.H., Kjeldsen, K.: Bile flow, pH, Hbpital Beaujon electrolytes, and bilirubin in hepatic bile F-92110 Clichy/France
Canalicular Bile Production in Man* D. Prandi,
S.
Erlinger, J.-C. Glasinovit, and M. Dumont
Clinique Chirurgicale, and Unitk de Recherches de Physiopathologie Hipatique (INSERM), H6pital Beaujon, Clichy, France Received: January I , 1974, and in revised form: April 19, 1974
Abstract. Bile flow, erythritol clearance (measured after administration of l4C-erythrito1) and bile acid excretion were studied in 8 non-icteric patients operated upon for gallstone disease without biliary obstruction in whom a T-tube had been inserted. Bile flow increased from the 2nd (0.27 2 SEM 0.01 ml/min) to the 7th (0.50 f SEM 0.05 ml/min) postoperative days and remained stable thereafter. Because the values obtained between the 7th and the 15th postoperative days (average: 0.46 ? SEM 0.02 ml/min) were not lower and were of the same order of magnitude as those previously reported in the literature for quantitative bile collections, it was assumed that bile collection had been complete. Bile-to-plasma concentration ratio of erythritol was 0.77 f SEM 0.04. The regression line between erythritol clearance and bile flow had a slope of 0.99 (indicating that increments in clearance were followed by equivalent increments of flow) and an intercept for a zero-clearance of 0.11, suggesting a ductular/ductal secretion of 0.11 ml/min. The relationship between erythritol clearance and bile acid excretion suggested that canalicular bile was formed both of a bile acid-dependent flow ( 1 1 p 1 of water being excreted per pmol of bile acid secreted) and a bile acid-independent flow of 0.16 ml/min (or 35 % of the average bile flow). These findings indicate that erythritol clearance provides a valid estimate of canalicular bile flow in man and that the three bile fractions previously postulated in animals (canalicular bile acid-dependent, canalicular bile acid-independent and ductular/ductal) may also be present in man. Key words: Bile, bile formation, bile acids, choleresis, man, biliary clearance, erythritol.
Most of our current knowledge on bile formation comes from animal experiments. It has been proposed that bile was formed of three fractions (1,2,3): a) the hepatocytic (canalicular) bile'acid-dependent fraction, which is considered to be the result of the active secretion of bile acids by the hepatocytes into the canaliculi,with subsequent osmotic flow of water and electrolytes; b) the hapatocytic (canalicular) bile acid-independent fraction, which may be due to active inorganic ion transport; c)the ductal or ductular fraction due to the secretion, presumably by the ductal or ductular epithelium, of a bile-acid free electrolyte solution, mainly in response to secretin administration. Considerable species variations in the process of bile formation have been observed. fie existence in man of basic mechanisms of bile secretion established in animals remains therefore uncertain and extrapolation toman of animal data is hazardous. me present studies, in patients with gallstone disease withcarried Out bile duct obstruction, were undertaken to obtain a better definition of the mechanisms of bile floy in man, and, in particular, to measure bile flow, *Presented in part at the 8th meeting of the European Association for the Study of the Liver, Vittel, France, September 7-9, 1973 (Digestion, 8, 437 (1973).
bile acid excretion and canalicular bile production, estimated by the biliary clearance of 14C-erythritol. The results suggest: 1 ) that 14C-erythritol clearance provides a valid estimate of canalicular bile flow in man, and 2 ) that the free fractions of bile postulated in animals seem to bepresent in man.
Patients and Methods Patients. Eight patients aged 47 to 83 yearswere operated upon for gallstone disease. None was jaundiced or had any other clinical sign of biliary obstruction. Their selection depended upon the surgeon's indication for choledochotomy: a cholecystectomy was performed and, because of dilatation of the c m i ~ ~ obile n duct, a choledocotomy was done, the paduct was verified and a Ttency Of the common tube (external diameter: 3.4 m, internal diameter: 2.3 mm) inserted. Pertinent clinical, biochemical and histological data are sumharized in Table I . During the post-operative period, the patients received a routine ward care: they,were given 5 per cent glucose and/or electrolyte intravenous infusions during 3-4 days, and were put on the ordinary hospital diet as soon as intestinal peristalsis was complete.
~.Prandie t a l . Canalicular Bile Production in Man
L
Table I . Clinica1,histological and biochemical findings Age
Operative findings
Hepatic histology
Sex
Liver function tests ('1 Serum alkaline phosphatase SGPT bilirubin mg Z
Haz..
100
0.8
3.5
50
I00
0.8
5
20
-
acute pancreatitis normal multiple gall-stones
0.8
2
60
I00
gallbladder empyema
0.8
9
95
I00
0.8
4.5
45
87
2.5
4
10
1 00
gall-stones, chronic cholecystitis
-
perforated duodenal ulcer; multiple gall-stones
-
Szy.. 6 4
acute cholecystitis
42 m
Thie.. 50
hepatic granuloma
slight cholestasis
f Am..
-
11
Sai.. 4 9
f
x
2
multiple gall-stones
Pi0... 69
Frankel units
0.3
47 f
m
Bodansky units
Prothrombin concentration (Quick)
normal
f
Ben.. 55
gallbladder empyema
-
m
Che.. 83 f (O)
multiple gall-stones cholangitis
slight cholestasis
obtained at time of study
Bile collections were carried out once a day, in the morning, after an overnight fast, from the 2nd to the 15th postoperative days, during 1 to 2 hours, in sucessive 15-minute samples, from the external branch of the T-tube, which had previously been occluded for twelve hours. Bile flow (taken as the mean flow during each 15minute collection period) and bile acid excretion were obtained during all collections; biliary clearance of erythritol (determined as indicated under "methods") was obtained once in each patient, on days 7 to 15 postoperatively, from two 15-minute samples obtained at the end of the 2-hour collection period, at least 90 minutes after the beginning of the infusion of erythritol (see "methods").
beginning of the infusion, and two plasma samples taken at the midpoint of each bile collection period were used to measure bile flow, and bile and plasma 14C activity for each clearance determination. 50 to 100 p1 of bile or plasma were deposited on Whatman glass fiber discs (6), dried, and counted in 1 ml of NE 213 (Nuclear Enterprises, Edinburgh, Scotland) scintillation medium;correction for quenching was made by internal standardization. Erythritol clearance was calculated as bile flow times the bile-to-plasma ratio of I4C activity. Statistical analysis of the results was performed using the Student's t test or theWilcoxon T test (7). Results
Methods. Bile acid concentration in bile was measured by an enzymatic technique using 3a hydroxysteroid dehydrogenase (Worthington Biochemical Corporation, Freehold, N.J., USA). Details of the technique are given elsewhere ( 4 , 5). When biliary clearance of erythritol was measured, I4C-labelled erythritol (The Fadiochemical Centre, Amersham, Buchinghamshire, U.K.) was administered intravenously as a single injection of 1 VCi, followed by a constant infusion of 0.1 vCi/min in glucose (50 g per 1 ) at a rate of 0.1 mllmin; two 15-min bile samples taken at least 90 minutes after the
The evolution of bile flow in the postoperative period is shown in Fig. 1 . Bile flow increasedfrom the 2nd (0.27 2 SEM 0.01 ml/min) to the 7th (0.50 f SEM 0.05 mllmin) postoperative days and tended to remain stable thereafter. The average flow was significantly lower on the 10th than on the other days: this finding was considered to be due to a random fluctuation. The average bile flow from day 7 t o day 15 was 0.46 f SEM 0 . 0 2 ml/min. The relationship between bile flow and bile acid excretion is represented in Fig. 2 . All values ob-
D.Prandi et a l . Canalicular Bile Production in Man
3
Table 2. Bile flow, bile acid concentration, erythritol clearance and canalicular flow in eight patients (at the time of erythritol clearance measurement) Patient Day
Haz..
10
Sai..
11
Bile flow
Bile acid excretion
Erythritol.(B)/(P)
Erythritol clearance
Canalicular flow
ml/min
umo l/min
0.30 0.32
5.1 6.6
0.68 0.76
0.20
0.19
0.24
0.21
0.62
37.4 28 .o
0.85
0.80
0.53 0.38
0.51
0.48
ml/min
ml/min
0.37
Pio..
9
0.44 0.44
9.8 9. I
0.68 0.16
0.30 0.33
0.33 0.33
Say..
7
0.31 0.42
9.2 9.7
0.78 0.78
0.29 0.33
0.26 0.31
ha..
14
0.41
0.48
34.8 34.7
0.ia 0.12
0.31 0.35
0.36 0.31
The..
13
0.44 0.36
22.4 17.8
0.66 0.12
0.29 0.26
0.33 0.25
.
15
0.88
50.3 50.1
0.82 0.95
0.12 0.81
0.71 0.14
0.10
22.8 29.1
0.12 0.19
0.41 0.55
0.46 0.59
0.50 0.04
22.2 3.9
0.11 0.02
0.40
0.40 0.04
Ben.
0.85
Che..
13
Mean SEM
0.51
0.04
Calculated as total bile flow minus ductal/ductular ("distal") flow obtained from Fig. 3
~
1-
. .... .. .. .- . . . .. .. X
Bile flow
'
BILE FLOW
~
-
(ml/min)
i
(ml/min) . Q5-
'ii+
.dF
0
.
=
0.807
0 .
Fig. 1 . Evolution of bile flow in the postoperative period. Each point represents flow during a j5-minute period. 4 to 8 measurements were obtained on a given day, during a 1 to 2 h collection Period. The mean flow (f SEM) was on day 2: 0.27 f 0.01, on day 3: 0.27 f 0.01, on day 5: 0.28 +- 0.01, on day 6: 0.42 2 0.02, on day 7: 0.50 f 0.05, on day 8: 0.32 f 0.03, on day 9: 0.53 f 0.06, on day 10: 0.29 f 0.03, on day 1 1 : 0.48 f 0.05, on day 13: 0.54 f 0.13 and on day 15: 0.60 f 0.05 ml/min
r
rC. .a
'0
mtx.
0 '
0.
'
.
y= a +
-
bx
a 0.27 ml/min b 0.010 ml/pmol
Fig. 2. Relationship between bile flow and bile acid excretion in man. Each point represents one measurement on successive 15-minute periods, obtained during each daily collection between the 7th and the 15th postoperative days in the 8 patients
D.Prandi et a l . Canalicular Bile Production in Man
4
a75
i
Erythritol clearance is linearly related to bile acid excretion rate. The slope of the regression line indicates that erythritol clearance increases 1 1 ~1 when bile acid excretion increases 1 pmol. The extrapolated clearance for a zero bile acid excretion is 0.16 ml/min.
/
Bile flow (mllmin)
Discussion r = 0.98 y.a+bx a * 0.n (mlAin)
/
I' 0
I
/
b = 0.99 (ml/ml) 0.25
I as0 0.75 Erythritol claarana (ml/min~
Fig. 3 . Relationship between bile flow and erythritol clearance. Each point represents one measurement (obtained during a 15-minute period)
1. V a l i d i t y of measurements. The present study was designed to characterize normal human bile formation. Three possible limitations in reaching this purpose will be discussed in this section: a) can the patients be considered normal? b) was bile collection complete? c) is erythritol clearance a valid estimate of canalicular bile flow?
Hepatic bile collection in man is only possible in patients with T tubes; therefore it has to be limited to patients with gallstone disease. However, none of the patients had biliary obstruction; routine liver function tests were normal (or almost normal) at the time of study and the liver biopsy Erythritol . - when performed - revealed no abnormalities or only minor ones. In addition to the disease, two clearance other shortcomings should be considered: the postoperative alterations in biliary function, and the interruption of the enterohepatic circulation. Postoperative alterations in biliary function were reflected in our study by a low bile flow from the 2nd to the 6th postoperative days; a low flow has already been observed in the early postoperative 1 0' 10 20 30 40 50 60 period in man (8, 9) and in the rhesus monkey (10); in the rhesus monkey, the decreased flow was asBile acid excretion ( pmol/min) sociated with a decreased secretion of bile acids, Fig. 4 . Relationship between erythritol clearance phospholipids and cholesterol and returned to norand bile acid excretion. Each point represents one mal within 2 weeks (10). In man, biliary function measurement obtained in the same samples as those has been considered to return to normal within apdepicted on Fig. 3 proximately one week (8, 1 1 , 12); this is in agreement with the stabilization of bile flow 7 days after oDeration in our patients and representative tained between the 7th and the 15th postoperative days have been included. There is a linear relation- data (including erythriiol clearance) have been ship between bile flow and bile acid excretion rate; obtained between the 7th and the 15th postoperative days. Interruption of the enterohepatic circulation the slope of the calculated regression line indicates a flow of 10 ~1 per pmol of bile acid exis followed within 1-3 hours by a decrease of bile flow and of secretion of total solids, phospholicreted. The extrapolated flow for a zero bile acid excretion is 0.27 ml/min. pids and cholesterol ( 1 2 ) ; these changes are reversible within 1-3 hours after refeeding of bile ( 1 2 ) ; bile was collected in our patients while the Bile flow, bile acid concentration, bile-toexternal branch of the T-tube had been occluded for plasma concentration ratios of I4C-erythritol and at least 12 hours, thus minimizing changes related erythritol clearance, recorded when erythritol to interruption of the enterohepatic circulation. clearance was measured in the 8 patients are indicated in Table 2 . The bile-to-plasma concentration The measurement of erythritol clearance (and bile flow) has been made after approximately 2 hours of ratio of erythritol is 0.77 2 SEM 0.02, and siginterruption of the enterohepatic circulation: bile nificantly lower than 1 (p < 0.001). As a result, flow was, on the average, lower at the time of erythritol clearance is significantly lower than bile flow (p c 0.001). The relationship between measurement than at the beginning of bile collection. Our values are therefore slightly underestibile flow and erythritol clearance is represented mated. Moreover, changes due to partial chronic in Fig. 3 . The slope of the regression line (0.99 interruption of the enterohepatic circulation canf SEM 0.01) is not significantly different from 1 . not be excluded. Thus, our patients cannot be conThe extrapolated flow for a zero clearance is 0.11 ml/min. sidered entirely normal; however, major abnormalities due to obstruction of bile ducts, surgery or The relationship between erythritol clearance interruption of the enterohepatic circulation have and bile acid excretion is represented in Fig. 4 . been avoided as much as possible.
/./
I
D.Prandi e t aZ. Canalicular Bile Production in Man Since the T-tube was not fitted with a balloon, the possibility that bile collection was not complete has been considered. However, the average flow obtained in this study (0.46 2 SEM 0 . 0 2 ml/ min) was not lower and was of the same order of magnitude as those previously reported by Preisig e t aZ. ( 0 . 3 7 2 SEM 0.10 ml/min) (13) and Shersth e t a l . (0.41 ml/min) ( 1 4 ) in patients with balloonequipped catheters in whom quantitative collections were presumably obtained. The average bile flow in our study would correspond to a daily output of bile of approximately 660 ml, a figure again similar to previous estimates of daily bile production in patients in whom quantitative collections had been possible (8, 1 1 , 12). It was therefore assumed that bile collection was complete or nearly complete. Moreover, it should be noted that the loss of a certain amount of bile would not modify the proportions of each constituent of bile flow and, therefore, would not alter the conclusions of the study
.
Measurement of erythritol clearance has been used to estimate canalicular bile flow in guineapigs ( 1 5 , 16), dogs ( 1 7 ) , rabbits (18) and rats ( 4 , 19, 2 0 ) . The validity of this measurement rests on the assumption that erythritol, an inert 5-carbon carbohydrate, enters canalicular bile through simple passive processes and does not move into or out of the biliary channels. As seen in Table 2 , the bile-to-plasma concentration ratio of erythritol was significantly lower than one, on the average 0 . 7 7 . Theoretically the finding of a biliary concentration of the solute lower than its plasma concentration might have been the result of either a limited "permeability" of the canalicular membrane to erythritol relative to water (with incomplete equilibrium of erythritol between plasma and canalicular bile), or, alternatively, of a dilution of canalicular bile by distal (ductular or ductal) fluid secretion. Since it is impossible to obtain canalicular bile, direct evidence for either possibility is lacking. However indirect evidence for the second explanation (dilution by distal fluid secretion) is obtained from the finding that the slope of the relationship between erythritol clearance and bile flow (Fig.3) was close to unity: this means that increments in clearance were followed by equal increments in flow, a finding which would not occur if the "permeability" of the canalicular membrane to erythritol was significantly lower than that of water; the extrapolation of the regression line between clearance and flow to a zero clearance gave a positive intercept of 0.11 mllmin: this would be expected if a distal secretion of 0.11 ml/min had occurred. Calculation of canalicular flow (as total bile flow minus "distal" flow) yielded values very close to the measured erythritol clearance (Table 2 ) . These observations are consistent with the hypothesis that erythritol clearance measures canalicular flow in man. It is interesting to note that in another study in patients (21) the bile-to-plasma concentration ratio and the clearance of mannitol, a somewhat larger solute, were slightly lower than those of erythritol in our study. This is consistent with
5 the view that the "permeability" of the canalicular membrane to mannitol is lower than that of erythritol, a finding already suggested by studies in the guinea-pig (16) and in the rabbit (18). 2. CanaZicuZar b i l e production in man. The study of the relationship between canalicular bile production (as estimated by erythritol clearance) and bile acid excretion sugges.ts the following comments. Firstly, canalicular bile production was proportional to bile acid excretion rate. This is presumably the result of a bile acid-dependent flow, as demonstrated in the dog ( l 7 ) , the rabbit (18) and the rat ( 4 , 19, 2 0 ) . Our study indicates that 1 1 p1 of canalicular bile are formed per pmol of bile acid excreted, a figure of the same order of magnitude as those reported in dogs and rats ( 1 7 , 19, 20) and, also, of the same order of magnitude as that obtained by Scherstin e t aZ. in man ( 1 4 ) from a study of the relationship between bile flow and bile acid excretion. Values obtained in the rabbit (18) and in man by Preisig e t a t . (13) were somewhat higher; the reason for this difference is not clear, but may have been related to methodological differences, regarding in particular, bile acid measurements in bile.
Secondly, extrapolation of the relationship between erythritol clearance and bile acid excretion gave a positive intercept of 0.16 ml/min. This is consistent with the existence of a bile acid-independent flow of canalicular origin of 0.16 ml/min or 35 X of total bile flow. An estimate of 0.13 ml/min, thus very similar to ours, has been obtained by mannitol clearance measurement (21). A canalicular bile acid-independent flow has already been postulated in the dog ( 1 7 ) , the rabbit ( l a ) , the rat (19, 20) and suggested (although not supported by erythritol clearance measurements) in the hamster (22) and the rhesus monkey ( 2 3 ) . Quantitatively, this fraction of canalicular bile, although ranging from approximately 5 pl/min/kg in the dog (17) to 60 pl/min/kg in the rabbit (18) , represents, in all species studied, 35 to 6 0 X of "basal" bile flow. It would therefore appear that, in man as in other species studied so far, canalicular bile is formed o f : a) a bile acid-dependent fraction, amounting to, on the average, 0.19 ml/min, and resulting from the excretion.of I 1 p1 of water per umol of bile acid secreted; b) a bile acid-independent fraction amounting to 0.16 ml/min. 3. Ductular or ductal f l u i d secretion. When the regression line between bile flow and erythritol clearance was extrapolated to the origin, a flow of 0.11 ml/min was obtained; another way of expressing this finding is to compare the extrapolated values for a zero bile acid excretion of bile flow ( 0 . 2 7 ml/min) and erythritol clearance (0.16 ml/min). These observations are consistent with secretion of 0.11 ml/min of fluid at sites distal to the d c u l i , i.e. in the ductules or
6
ducts. A ductular or ductal secretion has been strongly suggested in dogs (17, 24), but has not been found in rabbits or rats (18, 19, 20). In man, a "total" bile acid-independent bile flow (including the canalicular bile acid-independent and the ductular/ductal fractions) of 0.I5 to 0.20 ml/&n has been found in similar patients (13, 14), part of which has been postulated to be of ductular/ ductal origin and stimulated by biliary obstruction (13). In our patients, the moderate dilatation of the bile ducts may have contributed to the ductular/ ductal secretion. In the dog, the ductular/ductal fraction is stimulated by secretin administration (17); a similar observation has been made in man (21).
D.Prandi e t al. Canalicular Bile Production in Man
following choledochotomy. Acta chir. scand. 133, 483 (1967) 10. Redinger, R.N., Herman, A.R., Small, D.M.: The effects of surgery on hepatic physiology in the -primate. (abstract). Gastroenterology 60, 795 (1971) 1 1 . Mollowitz, G.: Beobachtungen der Gallensekretion des Menschen. Langenbecks Arch. klin. Chir. 291, 359 (1959) 12. Thureborn, E.: Human hepatic bile. Composition changes due to altered enterohepatic circulation. Acta chir. scand. (supp1.303), 1 (1962) 13. Preisig, R.,Bircher, H., Stirnemann, H., Tauber, J.: Postoperative choleresis following bile duct obstruction in man. Rev. franc. Etud. clin. biol. 14, 151 (1969) 1 4 . Scherstgn, T., Nilsson, S., Cahlin, E., FilipIn conclusion, erythritol clearance seems to son, M . , Brodin-Persson, G.: Relationship beprovide a valid estimate of canalicular bile flow tween the biliary secretion of bile acids and in man; canalicular bile production seems to be the the secretion of water, lecithin, and cholesresult of both a bile acid dependent-flow and a bile acid-independent flow; canalicular bile is terol in man. Europ. J. clin. Invest. 1 , 242 (1971) diluted by ductular or ductal fluid secretion. These observations support the view that basic de15. Forker, E.L.: Two sites of bile formation as determined by mannitol and erythritol clearance terminants of bile formation previously established in animals are also present in man. However, in the guinea pig. J. clin. Invest. 46, 1189 ( 1967) further studies are required to better define the mechanisms of elaboration and control of the vari16. Forker, E.L.: Bile formation in guinea pigs: ous fractions of bile in man. analysis with inert solutes of graded molecular radius. Amer. J. Physiol. 215, 56 (1968) 17. Wheeler, H.O., Ross, E.D., Bradley, S.E.: CanaAcknowledgements. This study was supported in licular bile production in dogs. h e r . J. part by a grant from INSERM (ATP no 9). Physiol. 214. 866 (1968) 18. Erlinger, S., Dhumeaux, D., Berthelot, P., Dumont, M.: Effect of inhibitors of sodium transport on bile formation in the rabbit. Amer. J. References Physiol. 219, 416 (1970) 1. Wheeler, H.O.: Secretion of bile acids by the 19. Boyer, J.L., Klatskin, G.: Canalicular bile flow liver and their role in the formation of hepand bile secretory pressure. Evidence for a nonatic bile. Arch. intern. Med. 130, 533 (1972) bile salt dependent fraction in the isolated 2. Erlinger, S.: Physiology of bile flow. In: Properfused rat liver. Gastroenterology 59, 853 gress in Liver Diseases, H. Popper and F. ( 1970) Schaffner (Eds.), Vol. 4, p. 63. New York and 20. Boyer, J.L.: Canalicular bile formation in the London: Grune and Stratton, 1972 isolated perfused rat liver. Amer. J. Physiol. 3. Erlinger, S., Dhumeaux, D.: Mechanisms and con221, 1156 (1971) trol of secretion of bile water and electro21. Boyer, J.L., Bloomer, J.R., Canalicular bile selytes. Gastroenterology 66, 281 (1974) cretion in man. Studies utilizing the biliaryclear4. Berthelot, P., Erlinger, S., Dhumeaux, D., ance o f [14C] mannitol. J.clin.Invest. 54, 773 Prgaux, A.M.: Mechanisms of phenobarbital in(1974) duced hypercholeresis in the rat. Amer. J. 22. King, J.E., Schoenfield, L.J.: Cholestasis inPhysiol. 219, 809 (1970) duced by sodium taurolithocholate in isolated 5. Erlinger, S.. Dumont, M.: Influence of theohamster liver. J. clin. Invest. 50, 2305 (1971) phylline on bile formation in the dog. Biomedi23. Redinger, R.N., Small, D.M.: Primate biliary c;ne 19, 27 (1973) physiology. VIII. The effect of phenobarbital 6. Davies, J.W., Cocking, E.C.: Liquid scintilupon bile salt synthesis and pool size, biliary lation counting of I4C and 3 H samples using lipid secretion, and bile composition. J. clin. glass-fibre or filter paper discs. Biochim. Invest. 52, 161 (1973) biophys. Acta Amst. 115, 511 (1966) 24. Wheeler, H.O., Mancusi-Ungaro, P.L.: Role of 7. Snedecor, G.W., Cochran, W.G.: Statistical bile ducts during secretin choleresis in dogs. methods, Ames (Iowa): The Iowa University Press, Amer. J. Physiol. 210, 1153 (1966) 1967 8. Rundle, F.F., Cass, M.H., Robson, B., Middleton, M.: Bile drainage after choledocostomy in man, Dr. Serge Erlinger with some observations on biliary fistula. SurUnit& de Recherches de gery 37, 903 (1955) Physiopathologie Hspathique 9. Guldhammer, E.H., Kjeldsen, K.: Bile flow, pH, Hbpital Beaujon electrolytes, and bilirubin in hepatic bile F-92110 Clichy/France
