Biochemical epidemiology of gallbladder cancer

Biochemical Epidemiology of Gallbladder Cancer BRIAN L. STROM,1 ROGER D. SOLOWAY,2 JAIME L. RIOS-DALENZ,3 HECTOR A. RODRIGUEZ-MARTINEZ,4 SUZANNE L. WEST,1 JUDITH L. KINMAN,1 ROGER S. CROWTHER,2 DONALD TAYLOR,2 MARCIA POLANSKY,5 1 AND JESSE A. BERLIN

To evaluate the a priori hypotheses that an increased level of glyco and tauro lithocholic acid, perhaps because of a decreased capacity for hepatic sulfation, contributed to the biochemical epidemiology of gallbladder cancer, a case-control study was undertaken at four hospitals in La Paz, Bolivia, and at one hospital in Mexico City, Mexico. Eighty-four cases with newly diagnosed histologically confirmed gallbladder cancer were compared with 264 controls with cholelithiasis or choledocholithiasis in the absence of cancer and with 126 controls with normal biliary tracts. All study subjects were undergoing abdominal surgery. Interview data were collected for all study subjects, as well as blood, bile, and gallstone specimens when feasible. Sera were analyzed for carcinoembryonic antigen, cholesterol concentration, and total bile acids. Bile specimens were analyzed for carcinoembryonic antigen; and for concentration of bile salts; cholesterol; phospholipids; and the glycine and taurine conjugates of cholic, ursodeoxycholic, chenodeoxycholic, deoxycholic, and lithocholates; sulfoglycolithocholate; and sulfotaurolithocholate. Gallstone specimens were analyzed for the percentage of cholesterol content, the percentage of calcium bilirubinate content, and the percentage of calcium carbonate content. Serum bile acids were increased in cases versus the two control groups (median 11.7 nmol/mL vs. 9.3 nmol/mL for stone controls and 8.2 nmol/L for nonstone controls, P ° .02 for each pairwise comparison). Biliary bile acids were markedly decreased in the cases (median 3.98 mmol/mL vs. 33.09 mmol/mL and 154.0 mmol/L, reAbbreviations: HDL, high-density lipoprotein; HPLC, high-pressure liquid chromatography; TUDC, taurousodeoxycholate; GUDC, glycoursodeoxycholate; GC, glycocholate; TCDC, taurochenodeoxycholate; GCDC, glycochenodeoxycholate; CDC, chenodeoxycholate; DC, deoxycholate; UDC, ursodeoxycholate; UDCA, ursodeoxycholic acid. From1 the Center for Clinical Epidemiology and Biostatistics, Department of Biostatistics and Epidemiology, and Division of General Internal Medicine of the Department of Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA; 2the Division of Gastroenterology, University of Texas Medical Branch at Galveston, Galveston, TX; 3Facultad de Medicina, Universidad Mayor de San Andres, La Paz, Bolivia; 4Hospital General de Mexico, Unidad de Patologia, Mexico; and 5Hahnemann University, Philadelphia, PA. Received December 1, 1994; accepted January 19, 1996. Supported by the National Institutes of Health, Bethesda, MD, Grant R01CA-35934. Address reprint requests to: Brian L. Strom, M.D., M.P.H., 824 Blockley Hall, Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania School of Medicine, Philadelphia, PA 19104-6021. Copyright q 1996 by the American Association for the Study of Liver Diseases. 0270-9139/96/2306-0016$3.00/0

spectively, P ° .0001 for each comparison), even after excluding those with a serum bilirubin higher than 2.0 mg/dL. Bile cholesterol was lower for the cases as well (median 1.70 mmol/mL vs. 4.90 mmol/mL and 16.81 mmol/ mL, respectively, P ° .02), as was the concentration of bile phospholipids (median 2.97 mmol/mL vs. 6.26 mmol/ mL and 52.69 mmol/mL, P Å .1 and .0004, respectively). Contrary to our a priori hypothesis, there was no difference between the cases and either control group in their bile concentrations of lithocholate, the proportion of bile acids which were sulfated, or the concentration of nonsulfated lithocholate. However, the cases had a higher concentration of ursodeoxycholate (UDC) (P õ .004 for both control groups), especially glycoursodeoxycholate (P õ .001 for both control groups). A previously published suggestion that gallstone size differed between cases and controls was not confirmed. In conclusion, cases with gallbladder cancer differed from controls with stones and from controls with normal biliary tracts in their serum and bile biochemistries. These findings may be a reflection of the disease process, or may provide useful clues to its pathogenesis. (HEPATOLOGY 1996;23:1402-1411.)

The study of gallbladder cancer provides a unique opportunity to combine biochemistry and epidemiology to investigate the possible cause of cancers that are uncommon in the United States’ black and white populations, but of considerable international importance among Hispanic, and especially Amerind, populations.1 Different parts of the biliary tract are bathed in the same material: bile. Yet, there are important differences in the epidemiology of cancers of the liver, the intrahepatic bile ducts, the extrahepatic bile ducts, the ampulla of Vater, and the gallbladder.2 In addition, there are marked differences in the incidence of these diseases across countries.1,3 These findings have led to many hypotheses about cause. The ability to collect bile and gallstones for biochemical analysis; and the development of technology which allows the measurement of potential carcinogens such as bile acid conjugates (the major solids in bile), as well as other serological cancer markers, made it feasible to combine epidemiology and biochemistry in an international collaborative study. Preliminary work showed intriguing differences in bile composition among gallbladder cancer patients, gallstone patients, and nonstone controls.4 In particular, the bile of patients with gallbladder can-

1402

5p0e$$0003

05-21-96 21:45:19

hepas

WBS: Hepatology

HEPATOLOGY Vol. 23, No. 6, 1996

STROM ET AL.

cer had a higher concentration of lithocholate, a known co-carcinogen, than the bile of patients with gallstones or normal biliary tracts, and a smaller proportion of this lithocholate was sulfated, which normally would detoxify the lithocholate and allow its excretion. The purpose of the current study was to confirm these earlier findings, used as a priori hypotheses, and investigate other biochemical differences between cases and controls. PATIENTS AND METHODS Study and Design A case-control study was conducted, comparing cases with gallbladder cancer to two groups of controls. The hospitals included as the study sites were the Hospital Generale in Mexico City, Mexico and the Hospital Obrero, Hospital de Clinicas, Hospital Metodista, and Instituto Gastroenterologia in La Paz, Bolivia. Cases were all individuals who had newly diagnosed, histologically confirmed, primary biliary tract cancer undergoing abdominal surgery. Although we initially intended to include cancers of the extrahepatic bile duct and ampulla of Vater, no such cases were identified during this time period. For each case, up to eight controls were chosen. Up to four individuals were sought with cholelithiasis or choledocholithiasis in the absence of biliary tract malignancy (hereafter referred to as stone controls). Up to four were sought who did not have biliary tract malignancies, cholelithiasis, or choledocholithiasis. Controls were all individuals who were also undergoing abdominal surgery within the same hospital and were matched to cases for sex and age ({5 years). Study subjects were interviewed regarding demographic characteristics, prior medical history, family history, and exposure to agents presumed to be risk factors for biliary cancer. Interviews were obtained for 474 subjects; 84 cases (84.5% female, average age 57.0 { 11.1 SD), 264 stone controls (88.6% female, average age 55.2 { 10.3 SD), and 126 nonstone controls (81.0% female, average age 54.5 { 11.2 SD). The results of these interviews are reported elsewhere.5 In addition to the interviews, blood, bile, and, where appropriate and possible, gallstone specimens were obtained during surgery. Some study subjects refused to provide specimens or desired to keep their own stones. For some cases, stones were not available because they were embedded in tumor tissue that was not removed. The specimens were frozen at 0207C and hand carried back to the United States on ice every 6 months, following visits to the study sites by two of the American investigators (B.L.S. and S.L.W.). Sera were subjected to analysis for the following: carcinoembryonic antigen, cholesterol concentration (total and esterified), and total serum bile acids. Bile specimens were subject to analyses for bile salt concentration, carcinoembryonic antigen, cholesterol concentration, phospholipid concentration, concentration of the glycine and taurine conjugates of cholic, ursodeoxycholic, chenodeoxycholic, deoxycholic, and lithocholates; sulfoglycolithocholate and sulfotaurolithocholates. Gallstone specimens were analyzed for the percentage of cholesterol content, the percentage of calcium bilirubinate content, and the percentage of calcium carbonate content. Informed consent was obtained from each study subject. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki, and was approved by the Insti-

5p0e$$0003

05-21-96 21:45:19

hepas

1403

tutional Review Board at each participating institution (Philadelphia, PA; Mexico City, Mexico; and La Paz, Bolivia). Laboratory Measurements Serum Chemistries. A chemzyme evaluation was performed on serum samples by a commercial laboratory (SmithKline Laboratories, Norristown, PA). Results are reported by case-control status. The following were measured: alkaline phosphatase, serum glutamic oxaloacetic transaminase, serum glutamic pyruvic transaminase, bilirubin, albumin, globulin, triglycerides, cholesterol, high-density lipoprotein (HDL)cholesterol, low-density lipoprotein cholesterol, and cholesterol HDL. The latter two measures are calculated results based on the cholesterol and HDL concentrations. Serum lipid data were available for 49 cases, 86 stone controls, and 48 nonstone controls. Fewer results were obtained for the HDL cholesterol, low density lipoprotein cholesterol, and cholesterol/HDL ratios. Several reasons were specified by SmithKline: insufficient serum, interference from non-HDL lipoproteins, and inability to calculate low-density lipoprotein cholesterol when triglyceride levels were in excess of 400 mg/ dL. Results for HDL cholesterol were based on 43 cases, 57 stone controls and 35 nonstone controls; for low-density lipoprotein cholesterol: 37 cases, 55 stone controls, and 34 nonstone controls; for cholesterol/HDL ratio: 43 cases, 57 stone controls, and 35 nonstone controls.

Serum Biochemistry Total Serum Bile Acids. Total serum bile acids were measured with a single beam fluorescence spectrophotometer using the method described by Siskos et al.,6 modified for single beam analysis. The standards used widely bracket the normal serum levels of 5 to 13 mL. Analyses of Drift Over Time and Dehydration. The possible differential effect of the lag time between the on-site collection of serum in Bolivia and Mexico and the testing in the laboratory was examined by: (1) calculating Spearman rank correlations between serum values and the time elapsed between drawing the sample and their analysis; and (2) observing the effects on the odds ratios for the serum levels of adding elapsed time to the conditional logistic regression. We also tested for the statistical significance of the interaction term between elapsed time and serum value. The Spearman correlation coefficients between the chemzyme evaluation and lapsed time were clearly nonsignificant (P ¢ .1) for 8 of the 11 analyses performed, with only the following exceptions: serum glutamic pyruvic transaminase (00.23, P Å .002), serum glutamic oxaloacetic transaminase (00.13, P Å .08), and cholesterol/HDL ratio (0.19, P Å .03). The results of the analyses of odds ratios for serum variables in relation to elapsed time parallels these figures. Only one of the 22 serum analyses showed a borderline significant (P Å .04) interaction between serum level and elapsed time: HDL cholesterol in the case-stone control comparison. The odds ratio obtained in the conditional logistic regression was 1.0 (0.97-1.0) before adjustment for elapsed time and 1.0 (0.998-1.01) after adjustment, so there is no confounding by elapsed time either. Given the number of nonsignificant analyses, and the very modest magnitude of the results in the remainder, these data are evidence against significant drift and/or dehydration, especially drift that differs among the study groups. Bile Salts, Cholesterol, and Phospholipids. Bile specimens were analyzed for bile salts, cholesterol, and phospholipids

WBS: Hepatology

1404 STROM ET AL.

HEPATOLOGY June 1996

according to the techniques of Talalay,7 Abell et al.8 and Fiske and Subbarow,9 respectively. The validity of the techniques is well-established.7-9 Each of these methods has a coefficient of variation in this laboratory of 5% between runs of each assay and 3% between duplicate specimens within a run. Repetitive measurement of the samples for cholesterol and total bile acids allowed determination of long-term drift. In the National Cooperative Gallstone Study,10 samples studied repeatedly over 12 to 18 months gave reproducible results, showing no significant long-term drift for bile acids and cholesterol. However, phospholipids were not stable enough to test. A number of attempts at measurement of long-term drift were unsuccessful in that study because of instability of the base sample because of breakdown of the phospholipids. Given the absence of prior data on long-term drift, in the current study a pool of bile was stored in aliquits at 0707C and samples were tested blindly by the technician at monthly intervals for the purpose of measuring changes/degradation over time. Bile Acid Conjugates. The individual bile acid conjugates were measured using high performance liquid chromatography, as described by Rossi et al.1 with the following modifications: 100 mL of native bile was mixed with 8 mL 0.1 NaOH and 1.0 mmol of cholate (internal standard), and put through an octadecyl (C18 ) extraction column (Baker bond spe, J. T. Baker, Inc., Phillipsburg, NJ). The column was washed with 20 mL of water and the bile salts were eluted with 8 mL high-pressure liquid chromatograph (HPLC) grade methanol. The eluate was collected, dried under N2 , and reconstituted to 0.5 mL with methanol. A 5-mL sample was then automatically injected into a Waters HPLC system (Waters Chromatograph, Milford, MA) consisting of a solvent delivery system, a variable wavelength ultraviolet detector set at 200 nm, a WISP automatic sample injector (Waters Chromatograph), and a computing integrator to calculate area under the sample curve. The column used was a Radial-Pak Nova-Pak C18 (Waters Chromatograph), 4-m particle size, 10 cm 1 8-mm internal diameter, housed in a radial compression module. The mobile phase contained 68% methanol and 32% 0.01 mol potassium phosphate buffer, pH adjusted to 5.5 with phosphorate. The mobile phase was freshly prepared before each run and filtered through a 0.45-mm membrane filter. The eluant flow rate was 1.0 mL/min at a pressure of 1,400 psi. A standard solution containing 12 commercially purchased conjugated bile acids (P-L Biochemicals, Inc., Milwaukee, WI) plus the cholate internal standard was prepared in methanol. The purity of each bile acid, checked by thin-layer chromatography, is greater than 99%. The bile acids are listed in their order of appearance in the HPLC chromatogram: (1) tauroursodeoxycholate (TUDC); (2) glycoursodeoxycholate (GUDC); (3) sulfotaurolithocholate; (4) taurocholate; (5) sulfoglycolithocholate; (6) glycocholate (GC); (7) taurochenodeoxycholate (TCDC); (8) taurodeoxycholate; (9) glycochenodeoxycholate (GCDC); (10) glycodeoxycholate, (11) taurolithocholate; and (12) glycolithocholate. The retention times drifted slowly with time, but this variation was less than 10% and is considered acceptable for this method.12 The identity of the samples is not lost, because, in a given sample, the retention times of all of the bile acids vary in the same manner. The method used is the most sensitive method for detection and separation of bile acid conjugates developed to date, and detects small concentrations of lithocholate conjugates that are not detected by previous methods.

5p0e$$0003

05-21-96 21:45:19

hepas

A pooled sample was divided into small aliquots, frozen at 0207C, and used monthly throughout the study to test for long-term drift. There is no problem in storing bile acids in this fashion. Values for retention time and for area under the detection curve (peak) as calculated by the automatic integrator check for deterioration of the column (shortening of the retention time) and for sample degradation (decrease in peak height). Analyses of Drift Over Time and Dehydration. The same procedure as described above was used to evaluate whether there was dehydration over time in the bile specimens, examining Spearman (nonparametric) correlations between bile values and the time elapsed between drawing the sample and their analysis. The Spearman correlation coefficients between bile values and lapsed time were: bile phospholipids, 00.38, P Å .0001; bile cholesterol, 0.34, P Å .0001; and biliary bile salts, 0.16, P Å .02. For the HPLC measurements, of the 12 bile salt measurements, 8 were clearly nonsignificant (P ¢ .1). The exceptions were GC: 00.33, P Å .0001, GCDC: 00.31, P Å .0001; glycodeoxecholate: 00.27, P Å .0003; and GUDC: 00.27, P Å .0004. Again, given two thirds of the analyses were nonsignificant, and the correlations of the remainder were inconsistent in direction, these data are evidence against meaningful drift and/or dehydration. A multivariate ANOVA was also performed with the 12 HPLC values as the outcome variables to obtain a global P value for the effect of elapsed time.13 The P value was .2511. Therefore, overall, the passage of time did not appear to have had any effect on the bile specimens used for HPLC testing. In a further effort to control for the possible effect of deliquescence on the bile specimens, we used for our calculations the proportion of each bile conjugate relative to the sum total of HPLC values. Using proportions rather than the actual values assumes that, even if deliquescence occurred, the proportion of each conjugate relative to the total of HPLC values would remain constant. In addition, the bile results were tested in a conditional logistic model that obtained odds ratios unadjusted, and then adjusted for, lapsed time. The results were similar, ruling out confounding by lapsed time. The interactions between lapsed time and bile measurements were not significant either. Gallstone Measurements Measurements of stones consisted of the number of stones, cholesterol content as percent of stone, stone volume, and stone diameter. The formula for stone volume is based on stone diameter.14 Size of stone was based on measurements made on an ‘‘average’’ stone; average size was based on visual inspection of submitted stones, which tend to be quite homogeneous for a given set. Note that stones were not available for all patients with stones, some stone sets were incomplete, and for some stones only fragments were available. Thus, information on number of stones was available for 76 cases and 214 stone controls. Cholesterol concentration was obtained for 28 cases and 130 stone controls. Stone volume and stone diameter were based on analyses of the specimens available from 25 cases and 92 stone controls. For stone biochemical analyses, a portion of the gallstone was pulverized, the powder dessicated to constant weight, and a 0.5-mg aliquot mixed with potassium bromide for 20 minutes and pressed into a pellet. The pellet was allowed to dessicate for 1 week and then spectra were obtained using

WBS: Hepatology

HEPATOLOGY Vol. 23, No. 6, 1996

STROM ET AL.

quantitative infrared spectroscopy.15 The peak height at 2,920 cm01 was measured for cholesterol and compared with a set of standards which form a straight line. The percentage of cholesterol per stone weight was calculated. Calcium bilirubinate and calcium carbonate were measured using infrared spectroscopy, measuring peaks at 1,600 cm01 and 870 cm01, respectively. The percentage of each stone constituent per stone weight was calculated. The within assay coefficient of variation of this assay, between duplicate specimens, is 8 { 3% (SD).

1405

The median level of serum bile acid among the cases (11.7 nmol/mL) was higher than that of the stone controls (9.3 nmol/mL; P Å .02) or the nonstone controls (8.2 nmol/mL; P Å .01). Note that, using values for all subjects, the correlation (Spearman) between serum bile acid and serum bilirubin was 0.41, P Å .0003. Bile Results

Bile results are presented only for subjects without clinical evidence of obstruction; a serum bilirubin level Statistical Analysis greater than 2.0 mg/dL is used as evidence of obstrucThe results are presented as means ({SD), medians, and tion (Table 1). The proportions of excluded subjects P values. Because of the very skewed distribution of some of were 44.9% of the cases, 14.9% of the stone controls, the variables, the P values were calculated using the Wil- and none of the nonstone controls (P õ .001). Analyses coxon rank sum test.16 Because biochemical data were un- are based on subjects for whom both the serum biliruavailable for many members of matched sets of patients, leav- bin test and the specific bile test are available. ing many sets with either no case or no controls, the matching Bile Biochemistry. There were extremely large difhad to be abandoned in the analysis. The statistical software ferences in the concentrations of bile salts in the cases package SAS (SAS Institute Inc., Cary, NC, 1988) was used (median level 3.98 mmol/mL) versus the stone controls for the analysis. (median 33.09 mmol/mL, P Å .0001) and the nonstone Inasmuch as bile results are problematic in subjects with controls (154.0 mmol/mL, P Å .0001). Smaller, although obstructed bile ducts, bile analyses are presented just for subjects showing no evidence of obstruction. A serum biliru- still large, differences were apparent for bile cholesbin level greater than 2.0 mg/dL is defined as evidence of bile terol and phospholipids, as well. The median level of bile cholesterol for the cases (1.70 mmol/mL) was much obstruction. smaller than that of the stone controls (4.90 mmol/mL, P Å .02) or, especially, the nonstone controls (16.81 RESULTS mmol/mL, P Å. 001). The median level of bile phosphoSerum Chemistries lipids was 2.97 mmol/mL in the cases, 6.26 mmol/mL As expected, the cases were consistently different among the stone controls (P Å .1), and 52.69 mmol/mL from both control groups in their liver function tests, among the nonstone controls (P Å .0004). with the median level of alkaline phosphatase among Bile Acid Conjugates. Among the case-stone control the cases about three times that of the stone controls comparisons, we found significantly lower values and four times higher for the cases than the nonstone among the cases than among the stone controls on the controls (P Å .0001); the median level of serum glu- following HPLC measures (Table 1): TUDC, TCDC, tamic pyruvic transaminase more than twice as high taurodeoxycholate, GCDC, glycolithocholate, chenodeamong the cases than among the stone controls and oxycholate (CDC), and deoxycholate (DC) as proporalmost four times as high as among the nonstone con- tions of total HPLC. Total (UDC) (TUDC / GUDC) and trols (P ° .001); the median level of serum glutamic especially GUDC proportions were significantly higher pyruric transaminase in the cases twice as high as the among the cases than among the stone controls, and stone controls and three times as high as the nonstone significantly higher relative to the combination of DC, controls (P õ .04). The median level of albumin among CDC, and lithocholate. Sulfotaurolithocholate proporthe cases was 20% lower than that among either group tion had a significant P value, even though the median of controls (P Å .0001). In contrast, the median level showed the same level for both cases and stone controls. of globulin for the cases was only 5% higher than that The remaining variables showed either borderline of the stone controls (P Å .15) and 8% higher than that (where indicated) or nonsignificant results: proportions of the nonstone contols. The median level of bilirubin of taurocholate, sulfoglycolithocholate, GC, taurodeoxywas slightly higher for the cases than for the stone cholate, total sulfates, total lithocholate, nonsulfated controls (1.3 vs. 0.8, P Å .09), but about two and half lithocholate, total cholates, and sulfate concentration. times as high as the median of 0.5 for the nonstone Similar results were found for the case versus noncontrols (P Å .0001). There were 147 subjects with se- stone control comparisons (Table 1). Significantly lower rum bilirubin of 2.0 mg/dL or lower, and 36 subjects values were found for the cases than for the nonstone with serum bilirubin higher than 2.0 mg/dL. controls on the following measures: TUDC, STLC, tauroThere were no systematic differences between the deoxycholate, glycodeoxycholate, total sulfates, and DC case group and the stone control group in cholesterol, as proportions of total HPLC. Significantly higher valtotal or fractionated, or triglycerides, although the ues of GUDC and total UDC proportions were found cases had a higher concentration of triglycerides (me- for cases compared with nonstone controls, as well as dian 145 vs. 108, P õ .004) and a lower concentration the ratio of total UDC to the combination of DC, CDC, of HDL cholesterol (25.0 vs. 33.5, P õ .03) than the and lithocholate. The remaining variables showed einonstone controls. ther borderline or nonsignificant results: proportions

5p0e$$0003

05-21-96 21:45:19

hepas

WBS: Hepatology

1406 STROM ET AL.

HEPATOLOGY June 1996 TABLE 1. Bile Constituents: Subjects With Serum Bilirubin °2.0 mg/mL N

Biliary Bile Acid (mmol/mL) Cases Stone controls Nonstone controls Bile Cholesterol (mmol/mL) Cases Stone controls Nonstone controls Phospholipids (mmol/mL) Cases Stone controls Nonstone controls Total lithocholate (%)* Cases Stone controls Nonstone controls Glycolithocholate (%)* Cases Stone controls Nonstone controls Taurolithocholate (%)* Cases Stone controls Nonstone controls Sulfotaurolithocholate (%)* Cases Stone controls Nonstone controls Sulfoglycholithocholate (%)* Cases Stone controls Nonstone controls Total ursodeoxycholate (%)* Cases Stone controls Nonstone controls Tauroursodeoxycholate (%)* Cases Stone controls Nonstone controls Glycoursodeoxycholate (%)* Cases Stone controls Nonstone controls Chenodeoxycholate (%)* Cases Stone controls Nonstone controls Taurochenodeoxycholate (%)* Cases Stone controls Nonstone controls Glycochenodeoxycholate (%)* Cases Stone controls Nonstone controls Deoxycholate (%)* Cases Stone controls

Mean (SD)

05-21-96 21:45:19

Mann-Whitney P Value

14 51 13

4.97 48.49 166.14

(4.12) (46.99) (70.94)

3.98 33.09 154.00

.0001 .0001

14 52 14

3.50 12.77 24.41

(4.56) (23.43) (39.26)

1.70 4.90 16.81

.0234 .0014

14 45 11

4.36 12.56 56.54

(4.87) (19.14) (25.28)

2.97 6.26 52.69

.1030 .0004

13 27 7

0.1143 (0.2313) 0.0534 (0.0925) 0.0766 (0.1191)

0.0000 0.0134 0.0230

.3131 .1640

13 27 7

0.0000 (0.0000) 0.0055 (0.0115) 0.0029 (0.0053)

0.0000 0.0000 0.0000

.0092 .0573

13 27 7

0.0718 (0.2138) 0.0008 (0.0017) 0.0024 (0.0049)

0.0000 0.0000 0.0000

.4195 .3486

13 27 7

0.0000 (0.0000) 0.0022 (0.0045) 0.0034 (0.0033)

0.0000 0.0000 0.0042

.0058 .0002

13 27 7

0.0425 (0.0989) 0.0449 (0.0922) 0.0678 (0.1184)

0.0000 0.0053 0.0111

.3178 .1932

13 27 7

0.3236 (0.3763) 0.0649 (0.1242) 0.0180 (0.0092)

0.0870 0.0239 0.0216

.0015 .0034

13 27 7

0.0008 (0.0030) 0.0189 (0.0527) 0.0031 (0.0025)

0.0000 0.0025 0.0029

.0010 .0029

13 27 7

0.3228 (0.3769) 0.0460 (0.0764) 0.0149 (0.0077)

0.0870 0.0217 0.0177

.0006 .0007

13 27 7

0.1619 (0.2042) 0.4009 (0.1376) 0.2909 (0.2397)

0.0000 0.4317 0.3122

.0018 .0891

13 27 7

0.0485 (0.0693) 0.0931 (0.0501) 0.1111 (0.1113)

0.0000 0.0876 0.0627

.0202 .0626

13 27 7

0.1134 (0.1574) 0.3078 (0.1448) 0.1797 (0.2302)

0.0000 0.3266 0.0000

.0018 .5915

13 27

0.0230 (0.0616) 0.1905 (0.1279)

0.0000 0.1893

.0001

(Continued on following page)

5p0e$$0003

Median

hepas

WBS: Hepatology

HEPATOLOGY Vol. 23, No. 6, 1996

STROM ET AL.

1407

TABLE 1 (cont’d). Bile Constituents: Subjects With Serum Bilirubin °2.0 mg/mL N

Nonstone controls Taurodeoxycholate (%)* Cases Stone controls Nonstone controls Glycodeoxycholate (%)* Cases Stone controls Nonstone controls Cholate (%)* Cases Stone controls Nonstone controls Glycocholate (%)* Cases Stone controls Nonstone controls Taurocholate (%)* Cases Stone controls Nonstone controls Total sulfate (%)* Cases Stone controls Nonstone controls Sulfate Concentration (%)* Cases Stone controls Nonstone controls Total Nonsulfated lithocholate (%)* Cases Stone controls Nonstone controls Ursodeoxycholate/(Deoxycholate / Chenodeoxycholate / Lithocholate) Cases Stone controls Nonstone controls Cholesterol Saturation (Bile Cholesterol/[Biliary Bile Acid / Phospholipids]) Cases Stone controls Nonstone controls

Mean (SD)

Median

Mann-Whitney P Value

7

0.2861 (0.2525)

0.1662

.0007

13 27 7

0.0082 (0.0243) 0.0421 (0.0970) 0.1832 (0.2282)

0.0000 0.0207 0.0183

.0156 .0452

13 27 7

0.0148 (0.0532) 0.1485 (0.1062) 0.1029 (0.0978)

0.0000 0.1425 0.0878

.0001 .0022

13 27 7

0.3771 (0.3139) 0.2903 (0.1409) 0.3285 (0.1386)

0.4667 0.3195 0.3444

.4700 .7504

13 27 7

0.2275 (0.2489) 0.1804 (0.1093) 0.1803 (0.1358)

0.1917 0.2087 0.1982

.7254 .7433

13 27 7

0.1497 (0.2003) 0.1099 (0.0771) 0.1482 (0.1662)

0.0799 0.0896 0.1010

.8623 .7504

13 27 7

0.0425 (0.0989) 0.0471 (0.0921) 0.0712 (0.1200)

0.0000 0.0081 0.0156

.2228 .0462

5 20 7

0.7000 (0.4472) 0.6566 (0.4005) 0.7914 (0.3671)

1.0000 0.8553 1.0000

.6217 .8563

13 27 7

0.0718 (0.2138) 0.0063 (0.0115) 0.0053 (0.0067)

0.0000 0.0000 0.0000

.1440 .3486

9 27 7

0.5473 (0.8292) 0.2019 (0.6188) 0.0272 (0.0133)

0.0870 0.0340 0.0289

.0085 .0010

14 45 11

0.4103 (0.4222) 0.3121 (0.4868) 0.1386 (0.2188)

0.2431 0.1148 0.0762

.2579 .0950

* Å Proportion of HPLC sum.

of taurocholate, sulfoglycolithocholate, GC, TCDC (borderline), GCDC, taurodeoxycholate, glycolithocholate (borderline), total lithocholate, nonsulfated lithocholate, CDC (borderline), total cholates, as well as sulfate concentration. In data presented elsewhere5 we document that patients of Indian ancestry are at markedly increased risk of gallbladder cancer and that, within the Bolivian subjects, patients of Aymaran Indian ancestry are at increased risk, whereas patients of Quechua Indian ancestry are at decreased risk. Examining the HPLC measurements, comparing Indians to non-Indians across both countries, of the 27 analyses, only one was

5p0e$$0003

05-21-96 21:45:19

hepas

close to statistical significance: Indians had a lower concentration of bile salts (median 15.6 vs. 48.2 mmol/ mL, P Å .02). This suggests that race is not an important confounding variable in these analyses. Comparing Aymaran to non-Aymaran patients of Bolivia, again ignoring case-control status, Aymaran patients had a lower concentration of bile salts (median 13.1 vs. 49.4 mmol/mL, P Å .02). They also tended to have lower concentrations of nonsulfated lithocholate (0 vs. 0.005 mmol/mL, P Å .07), glycolithocholate (0 vs. 0.003 mmol/ mL, P Å .08), and higher concentrations of GUDC (0.045 vs. 0.02 mmol/mL, P Å .03) GC (0.235 vs. 0.158 mmol/mL, P Å .07), and UDC (0.056 vs. 0.024 mmol/

WBS: Hepatology

1408 STROM ET AL.

HEPATOLOGY June 1996

mL, P Å .08). When the Aymaran versus non-Aymaran comparison was restricted to controls, only the GC result remained significant, suggesting that the apparent racial differences are attributable to case-control differences, and not vice versa. In analyses using the square roots of these bile variables to correct for the skewed distributions, differences between cases and controls remained highly significant for bile salts, GUDC, and UDC, even after controlling for race using two-way ANOVA. Stones

Information on number of stones was missing for many of the cases. Evidence of the existence of stones was noted for 22 of these ‘‘unknown’’ cases, either in the reports of surgery or pathology, on ultrasound, or otherwise noted on the reporting forms. ‘‘Multiple’’ stones were noted in several instances. In others, the cancer was so extensive that it obscured the gallbladder; there were cases for whom the surgery was palliative only, and the gallbladder was not removed. The total number of cases for whom information was available on the specific number of stones was 53. Given these limitations, the number of stones documented among the cases ranged from 0 (n Å 5) to 295, with a mean of 29.9 (SD, 65.5) and a median number of 3. Four of the five cases with no stones showed some indication of stones on ultrasound, but no stones were found on surgery. In the operative report of the fifth case, there was a clear statement that no stones were observed. In comparison, the number of stones among the stone controls (n Å 144) ranged from 1 to 404, with a mean of 18.6 (SD, 48.2) and a median of 3 (P Å .77 vs. cases). The median cholesterol level for the cases’ stones was similar to that of the controls (81.9% vs. 84.0%, P Å .2). The median volume of the stones of the cases was the same as that of the controls (0.4 mL; P Å .7). Similarly, the median diameter of the average stone of the cases was the same as that of the controls (8.5 mm; P Å .8). DISCUSSION

The results of this study did not confirm the a priori hypotheses that there was an increased proportion of lithocholic acid and that these species were inadequately sulfated, increasing their toxicity and carcinogenic potential. In particular, when considering patients without clinical evidence of obstruction, there was no difference between the cases and the controls in their bile concentrations of lithocholate, in the proportion of bile acids that were sulfated, or, in particular, in the concentration of nonsulfated lithocholate. In addition, there was no difference in the cholesterol concentration of the stones between the cases and the stone controls. Thus, while the power to examine these findings was not enormous, given that specimens were not obtained from all subjects, and given the variability inherent in the measurements, this study, clearly more

5p0e$$0003

05-21-96 21:45:19

hepas

powerful and more systematic than the previously available data,4 did not confirm the a priori hypotheses about bile lithocholate, nonsulfated bile salts, and stone cholesterol concentration being important to the cause of gallbladder cancer. Note that the bile acid HPLC methodology used in this study was clearly an improvement in sensitivity and reliability over the previously used nonautomated system,3 which also used a slightly different, less stable, and less reproducible solvent system. The formerly used system contained an initial isocratic phase followed by a gradient, while in the present system the separation was completely isocratic and thus more reproducible. However, a number of other differences between cases and controls was observed. This study reveals the expected differences between the cases and the controls in their liver function tests and other biochemical markers of hepatic obstruction and/or infiltration. Interestingly, it does not reveal a difference in either the serum cholesterol, the stone cholesterol content, or the bile cholesterol saturation. However, serum triglycerides are elevated in cases, and serum HDL is decreased (vs. nonstone controls only), as is bile cholesterol. Biliary bile salts and phospholipids are also decreased in cases, while serum bile acids are increased. The decreased biliary bile salts, phospholipids, and cholesterol in cases may be because of decreased bile secretion and cholestasis. The increased serum bile acids may be because of increased cholehepatic shunting, which should also make ursodeoxycholate the predominant bile acid. An alternative view is that ursodeoxycholic acid (UDCA) has the highest intrinsic hepatic clearance among all of the naturally occurring bile acids, contributing to its increased proportion in mildly cholestatic bile. The wide decrease in biliary bile acids in the cancer cases, once clinical obstruction is removed by eliminating those patients with a bilirubin greater than 2.0 mg/dL, remains incompletely explained. These may be patients with subclinical cholestasis of biliary lipids based on the low biliary bile acids, but whose bilirubin is not affected. The elevation of alkaline phosphatase suggests that there is at least partial obstruction to the biliary system, because the level is much higher than in either stone controls or normals. Additional support for the theory that some cholestasis was present is that the median level of serum bile acid in cases was higher than that of stone controls or nonstone controls. It is of interest that the concentration of biliary cholesterol for stone controls was significantly lower than that of nonstone controls. Clinically, these patients did not show biliary obstruction and, yet, bile composition is in contrast to the findings in the National Cooperative Gallstone Study in which patients with gallstones had a somewhat higher biliary cholesterol saturation than the nonstone controls that were available.10 The same phenomenon is present for phospholipids and, because bile salt, phospholipids and cholesterol are measured by separate methods, this seems to be a real finding.

WBS: Hepatology

HEPATOLOGY Vol. 23, No. 6, 1996

STROM ET AL.

Note, however, that bile in the cases was considerably diluted. Total bile acid concentration was about 4 to 5 mmol/mL; mean UDCA concentration was thus approximately 2mmol/mL in the cases. If we consider the stone (or the nonstone controls), UDCA concentration is 0.015 (% of total bile acids) 1 166 mmol Ç2 mmol, thus virtually the same absolute concentration as found in the cases. In the nonstone controls, similar UDCA concentration is observed. In contrast to previous data,17,18 we observed no differences between cases and controls in volume of stones or diameter of stones. This may have been an artifact because of incomplete collection of stones in our study, because of differences in definition (average size in our study vs. maximum size in the prior studies), or reliance of the previous studies on medical record data. Note that Moerman, using radiological methods, also was unable to find a relationship between stone diameter and cancer risk.19 Assuming it is a real observation, the fact that the mean diameter and volume of the stones was small and similar to that of controls suggests that the stones were not present for longer periods of time in cases and that we were not looking at gallstone disease that was significantly more advanced in patients with cancer than in patients with stones. Of the individual bile salts, the only one significantly increased in the cases was GUDC. In fact, the cases had an increased concentration in UDC in general. Virtually all the other individual bile salts were in lower concentration in the cases versus the controls, as were the total biliary bile salts. However, this increased concentration of UDC in particular GUDC, was quite striking. It would suggest that either there is selective alteration of intestinal bacteria in these patients, such that the bacteria that remain metabolize bile salts differently, that there is preferential UDC clearance by the cholestatic liver, or there is a predisposition for the liver to manufacture more of this bile acid. The mechanism for the latter could be genetic in origin (Aymaran Indians), or could be because of an increased available substrate in the diet, or it could operate through some other environmental mechanism. Perhaps this observation results from cholehepatic recycling as a defense mechanism against cholestasis.20,21 In view of the literature that suggests that unconjugated UDC has a beneficial effect on liver disease, especially with cholestasis,22-27 it seems plausible that it may be serving a protective function in this pathological condition. Although GUDC does not cause hypercholeresis,28 it may be a marker for increased UDC secretion. Unfortunately, we did not measure UDC directly. The effects of UDC on the gallbladder mucosa, however, are less clear. In this study, the remarkable finding on examining all 12 bile acid species is that GUDC shows a striking relative increase in patients with cancer of the gallbladder, magnifying a trend seen in patients with cholelithiasis alone. In fact, this increase was so pronounced that TUDC is virtually eliminated and GCDC

5p0e$$0003

05-21-96 21:45:19

hepas

1409

and TCDC are reduced reciprocally, compared with stone controls. In stone controls, UDC is 6.5% and CDC is 40.1%, whereas in cases UDC is 32.4% and CDC is 16.2%. Thus, the reciprocal nature of this relationship is indicated by the sum of UDC and CDC conjugates, which remains stable at 48% in cases and 46% in stone controls. The lack of increase in DC or LC and the lack of increase in their glycine conjugates indicate that there is no generalized increase in the enterohepatic circulation of bile acids leading to an increase in secondary bile acid or glycine conjugates. The question remains concerning why the proportion of GUDC is so markedly increased in patients with gallbladder cancer. Because of hyperconjugation after feeding unconjugated bile acid, it is the GUDC form that is in the highest proportions in patients who are treated with UDC for gallstone dissolution.29,30 UDC is known to undergo a cholehepatic short-circuit cycling from biliary ducts to the liver.30,31 Whether the glycine conjugate can undergo this cycling is uncertain.31 GUDC is less polar and more lipid soluble than TUDC and may, therefore, undergo greater short circuiting and enterohepatic cycling, accounting for its greater concentration in these patients. It is more likely that this bile acid would be the most protective of the bile acids because of its anti-inflammatory22-27 and hypercholeretic effects.32-36 The increased amount of glycoursodeoxycholate present could reflect this increased cholehepatic circulation and be a small proportion of the fluxing bile acids that are conjugated with glycine. Once conjugated, because it is less likely to undergo cholehepatic cycling, it would be excreted in bile. Because glycine dihydroxy bile acids are more likely to be reabsorbed in the duodenum and upper jejunum by passive reabsorption, in effect this is recycling to some degree and explains why glycine is retained and is increased relative to the taurine conjugates. The theory of Gurantz and Hofmann36 suggests that unconjugated UDC may be reabsorbed in a protonated form in the canaliculi and ductules and be recycled. It is this undetectable recycling that leads to the hypercholeresis. Anwer37 suggested that HCO30 , left in the lumen after H/ donation to the bile acid, exerts osmotic activity itself, not requiring recycling. Howard and Murphy28 indicate that this recycling is necessary to have a molecule to provide proton acceptors. Farges et al.31 show directly using retrograde injections of bile acids that UDCA is more readily absorbed than cholate and the taurine conjugate of UDCA. A number of questions are raised by these data: (1) Is UDC being formed in the liver because of the tumor and is some factor inverting the reduction of the normal intermediate 3a, 7 keto bile acid in the liver to UDC, rather than CDC? (2) Is there selective reabsorption of the bile acid into the enterohepatic circulation through the tumor or the biliary tree? Does mild intrahepatic obstruction, before the serum bilirubin rises above 2.0, lead to this type of change? (3) Does cholelithiasis, leading to an injured gallbladder, cause this type of abnor-

WBS: Hepatology

1410 STROM ET AL.

HEPATOLOGY June 1996

mality? The increase in GUDC coupled with a decrease in DC and LC suggests that a combination of cholehepatic shunting amplifying the concentration of UDC, perhaps of hepatic origin, is occurring simultaneously with a decrease in enterohepatic circulation of bile salts. In addition, the constant proportion of bile salt represented by UDC and CDC suggests that UDC feeds back on CDC to cause a decrease in CDC production. This study suffers from a number of limitations. While it is the largest and most complete consecutive series of patients with gallbladder cancer that we know of, specimens were not available on all patients, and these missing data could have lead to a selection bias. Further, multiple statistical tests were performed. We have not corrected for multiple comparisons, because these were a priori defined contrasts. However, most of these P values were sufficiently small that they would not have been affected by such a correction. Nevertheless, the primary hypotheses of this study were not confirmed by these data and the primary positive findings were not a priori hypotheses, so these positive findings remain to be confirmed elsewhere before being considered definitive. Most centrally, we cannot differentiate whether this increased concentration of UDC is a cause of the cancer, an incidental result of the cancer, or part of the body’s attempt to defend against the cancer. In addition to the treatment of cholelithiasis, great interest has been evidenced in the use of UDC in chronic liver disease.20,21 Will its regular use cause chronic liver damage or lead to long-term effects on the biliary tree, eventuating in cancer? The data generated in this study do not allow us to differentiate cause from effect. Studies involving hepatocyte cell lines from patients with Indian background may help in making this differentiation, as will studies of the long-term outcome of patients given UDC therapeutically. Acknowledgment: We are grateful to the surgeons and gastroenterologists at the participating hospitals for enrolling their patients in the study, and to Dianne Marie Greer for her excellent manuscript assistance. REFERENCES 1. Strom BL, Nelson WL, Henson DE, Albores-Saavedra J, Soloway RD. Carcinoma of the gallbladder. In: Cohen S, Soloway RD, eds. Gallstones. New York: Churchill Livingstone, 1985:275-298. 2. Strom BL, Hibberd PL, Soper KA, Stolley PD, Nelson WL. International variations in epidemiology of cancers of the extrahepatic biliary tract. Cancer Res 1985;45:5165-5168. 3. Rios-Dalenz J, Takabayashi A, Henson D, Strom BL, Soloway RD. The epidemiology of cancer of the extra-hepatic biliary tract in Bolivia. Int J Epidemiol 1983;12:156-160. 4. Rios-Dalenz J, Takabayashi A, Henson D, Strom BL, Soloway RD. Cancer of the gallbladder in Bolivia: suggestions concerning etiology. Am J Gastroenterol 1985;80:371-375. 5. Strom BL, Soloway RD, Rios-Dalenz JL, Rodriguez-Martinez HA, West SL, Kinman JL, Polansky M, et al. International collaborative case-control study of gallbladder cancer. Cancer 1995; 76:1747-1756. 6. Siskos PA, Cahill PT, Javitt NB. Serum bile acid analysis: a rapid direct enzymatic method using dual-beam spectrophotofluorimetry. J Lipid Res 1977;18:666-671.

5p0e$$0003

05-21-96 21:45:19

hepas

7. Talalay P. Enzymatic analysis of steroid hormones. Methods Biochem Anal 1960;8:119-143. 8. Abell LL, Levy BB, Brodie BB, Kendall FE. A simplified method in the estimation of total cholesterol in serum and demonstration of its specification. J Biol Chem 1952;195:357-366. 9. Fiske CH, Subbarow Y. The colorimetric determination of phosphorus. J Biol Chem 1925;66:375-400. 10. Schoenfield LJ, Lachin JM, and the Steering Committee for the National Cooperative Gallstone Study. Chenodiol (chenodeoxycholic acid) for the dissolution of gallstones: the National Cooperative Gallstone Study. Ann Intern Med 1981;95:257-282. 11. Rossi SS, Converse JL, Hofmann AF. High pressure liquid chromatographic analysis of conjugated bile acid in human bile: simultaneous resolution of sulfated and unsulfated lithocholyl amidates and the common conjugated bile acids. J Lipid Res 1987;28:589-595. 12. Hofmann AF, Grundy SM, Lachin JM, Lan SP, Baum RA, Hanson RF, Hersh T, et al. Pretreatment biliary lipid composition in white patients with radiolucent gallstones in the National Cooperative Gallstone Study. Gastroenterology 1982;83:738752. 13. Johnson RA, Wichern DW. Applied multivariate statistical analysis. Ed 2. Englewood Cliffs: Prentice Hall, 1988. 14. Lasser EC, Amberg JR, Baily NA, Varady P, Lachin J, Okun R, Schoenfield L. Validation of a computer-assisted method for estimating the number and volume of gallstones visualized by cholecystography. Invest Radiol 1981;16:342-347. 15. Trotman BW, Morris TA III, Sanchez HM, Soloway RD, Ostrow JD. Pigment vs. cholesterol, cholelithiasis, identification and quantification by infrared spectroscopy. Gastroenterology 1977; 72:495-498. 16. Conover WJ. Practical nonparametric statistics. Ed 2. New York: Wiley, 1980. 17. Diehl AK. Gallstone size and the risk of gallbladder cancer. JAMA 1983;250:2323-2326. 18. Lowenfels AB, Walker AM, Althaus DP, Townsend G, Domello¨f L. Gallstone growth, size, and risk of gallbladder cancer: an interracial study. Int J Epidemiol 1989;18:50-54. 19. Moerman CJ, de Mesquita HBB, Runia S. Dietary sugar intake in the aetiology of biliary tract cancer. Int J Epidemiol 1993;22: 207-214. 20. Abernathy CO, Zimmerman HJ, Ishak KG, Utili R, Gillespie J. Drug-induced cholestasis in the perfused rat liver and its reversal by tauroursodeoxycholate: an ultrastructural study. PSEBM 1992;199:54-58. 21. Ota S, Tsukahara H, Terano A, Hata Y, Hiraishi H, Mutoh H, Sugimoto T. Protective effect of tauroursodeoxycholate against chenodeoxycholate-induced damage to cultured rabbit gastric cells. Dig Dis Sci 1991;36:409-416. 22. Heuman DM, Pandak WM, Hylemon PB, Vlahcevic ZR. Conjugates of ursodeoxycholate protect against cytotoxicity of more hydropholeic bile salts: in vitro studies in rat hepatocytes and human erythrocytes. HEPATOLOGY 1991;14:920-926. 23. O’Brien CB, Senior JR, Arora-Mirchandani R, Batta AK, Salen A. Ursodeoxycholic acid for the treatment of primary sclerosing cholangitis: A 30-month pilot study. HEPATOLOGY 1991;14:838847. 24. Galle PR, Theilmann L, Raedsch R, Otto G, Stiehl A. Ursodeoxycholate reduces hepatotoxicity of bile salts in primary human hepatocytes. HEPATOLOGY 1990;12:486-491. 25. 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;11:12-15. 26. Poupon RE, Balkau B, Eschwe`ge E, Poupon R, and the UDCAPBC Study Group. A multicenter, controlled trial of ursodiol for the treatment of primary biliary cirrhosis. N Engl J Med 1991: 1548-1554. 27. Yoshikawa M, Tsujii T, Matsumura K, Yamao J, Matsumura Y, Kubo R, Fukui H, et al. Immunomodulatory effects of ursodeoxycholic acid on immune responses. HEPATOLOGY 1992;16:358-364.

WBS: Hepatology

HEPATOLOGY Vol. 23, No. 6, 1996

STROM ET AL.

28. Howard PJ, Murphy GM. Bile physiology. Gastroenterology 1992;8:731-743. 29. Garbutt JT, Lack L, Tyor MP. The enterohepatic circulation of bile salts in gastrointestinal disorders. Am J Med 1971;51:627636. 30. Hoffman NE, Hofmann AF. Metabolism of steroid and amino acid moieties of conjugated bile acids in man. Gastroenterology 1977;72:141-148. 31. Farges O, Corbic M, Dumont M, Maurice M, Erlinger S. Permeability of the rat biliary tree to ursodeoxycholic acid. Am J Physiol 1989;256:G653-G660. 32. Erlinger S. Secretion of Bile. In: Schiff L, Schiff ER, eds. Diseases of the Liver. Ed. 7. Philadelphia: Lippincott, 1993. 33. Dumont M, Uchman S, Erlinger S. Hypercholeresis induced by ursodeoxycholic acid and 7-ketolithocholic acid in the rat: possible role of bicarbonate transport. Gastroenterology 1980;79:82-89.

5p0e$$0003

05-21-96 21:45:19

hepas

1411

34. Yoon YB, Hagey LR, Hofmann AF, Gurantz D, Michelotti EL, Steinbach JH. Effect of side-chain shortening on the physiologic properties of bile acids: hepatic transport and effect on biliary secretion of 23-norursodeoxycholate in rodents. Gastroenterology 1986;90:837-852. 35. Lake JR, Van Dyke RW, Scharschmidt BF. Effects of Na replacement and amiloride on ursodeoxycholic acid stimulated choleresis and biliary bicarbonate secretion. Am J Physiol 1987;252: G163-G169. 36. Gurantz D, Hofmann AF. Influence of bile acid structure on bile flow and biliary lipid secretion in the hamster. Am J Physiol 1984;247:G738-G748. 37. Anwer MS. Mechanism of bile acid-induced HCO0 3 -rich hypercholeresis. An analysis based on quantitative acid-base chemistry. J Hepatol 1992;14:118-126.

WBS: Hepatology