Effects of Fenofibrate on Bile Lipid Composition

Transcription

1 Effects of on Bile Lipid Composition Robert H. Palmer In humans, clofibrate increases the degree of bile cholesterol saturation and predisposes patients to cholesterol gallstone formation. To determine if this activity exends to the related hypolipidemic agent fenofibrate, duodenal bile lipid composition was studied in 15 subjects before they participated in a double-blind study of that drug. Eight subjects were studied again on fenofibrate and six on placebo; five placebo patients were also studied later on open-label fenofibrate. The results were similar in the double-blind and open-label studies, and changes in bile lipid composition were comparable to those seen in studies of clofibrate. caused a significant decrease in the molar percentage of bile acids and increases in the molar percentage of phospholipids and cholesterol. The changes in bile acids and phospholipids had opposing effects on the cholesterol-holding capacity of bile. A statistically significant increase in the cholesterol saturation index was only apparent when all fenofibrate bile analyses were compared with all untreated bile analyses. The results demonstrated that fenofibrate has clear effects on bile lipid composition that may be associated with an increased propensity for gallstone formation, and when fenofibrate is used, patients should be monitored for this possibility. (Arteriosclerosis 5: , November/December 1985) It has been established that clofibrate use is associated with changes in biliary lipid composition 1 " 9 and an increased incidence of gallstones. 5 ' 10 " 12 The present investigations were undertaken to determine if fenofibrate, an effective new hypolipidemic drug, causes similar changes in bile lipid composition and hence should be studied further with respect to possible effects on cholesterol gallstone formation. ethods Subjects The subjects were hypercholesterolemic patients participating in a 6-month double-blind study of fenofibrate compared to placebo. This was followed in some patients by an open-label 6-month treatment period. To be eligible for the double-blind study, patients had to have levels of total cholesterol greater than 250 mg/dl and, on at least one occasion, levels of LDL cholesterol greater than 175 mg/dl as deter- rom the Department of edicine, Columbia University College of Physicians and Surgeons, New York, New York. This work was supported in part by NHLBI Grant HL (SCOR in Arteriosclerosis) and Grant RR (General Clinical Research Center). Address for reprints: Robert H. Palmer,.D., Clinical R & D, Smith Kline and rench Laboratories, 1500 Spring Garden Street, L217, Philadelphia, Pennsylvania Received October 15,1984; revision accepted June 17,1985. mined by the LRC methodology. 13 All patients were men and women between 18 and 65 years of age; the women were incapable of becoming pregnant and not on estrogen or contraceptive steroid therapy. Excluded were patients with a history of important medical illness, including liver or gallbladder disease, myocardial infarction or unstable angina within the previous 3 months, a history of alcohol or drug abuse, secondary hyperlipidemia, or a body mass index (weight/height 2 ) greater than Ib/in 2. Subjects were told of the bile study at four centers, and if willing to participate they underwent oral cholecystography. If a functioning gallblader without stones was found, those patients were referred to Columbia-Presbyterian edical Center for the bile lipid studies. Appropriate informed consent was obtained in accordance with provisions of the Human investigation Committee of the Columbia Presbyterian edical Center. Bile Studies Subjects were hospitalized overnight in the General Clinical Research Center where, after informed consent had been obtained, a small, mercuryweighted, nasoduodenal tube (Hydromer-Dobbhoff Enteric eeding Tube No , Biosearch edical Products, Incorporated, Somerville, New Jersey) was passed and allowed to progress into the duodenum. At bedtime, the patients drank 5 /xci of 14 C- cholic acid (NEC-241, New England Nuclear, Bos- 631

2 i- C CO CO CO CO - C CO 0.0- LL *- OJ CO *- C O. LL LL *- OJ 632 ARTERIOSCLEROSIS VOL. 5, No 6, NOVEBER/DECEBER 1985 Table 1. Serum and Bile Llplds In Patients Receiving or enoflbrate Patient no. Age group group Sex Weight (lbs) Sample date 06/22/82 10/06/82 08/13/82 03/11/83 07/16/82 11/23/82 08/16/83 07/29/82 02/28/83 08/26/83 08/13/82 02/28/83 08/23/83 01/07/83 06/30/83 11/22/83 10/06/82 01/10/84 05/04/82 11/09/82 06/29/82 01/07/83 06/30/82 11/24/82 08/13/82 03/11/83 08/13/82 03/02/83 08/25/82 04/11/83 12/07/82 03/02/83 12/10/82 06/30/82 Treatment* P2 P li. 0. Q. li. - C CO Li. 2 Tl 13 to -» Choi Serum lipids (mg/dl)t TG LDL-C HDL-C Treatment: = baseline placebo; P2 = double-blind placebo; 2 = double-blind fenofibrate; 3 = open label fenofibrate. fbaseline value () is the mean of two determinations. Treatment values (P2, 2, 3) are values obtained at the time of the bile study, or the average of two determinations when the bile study occurred between visits. ^Cholesterol saturation is the mol% of cholesterol required to saturate gallbladder bile at a total lipid concentration of 15 g/dl (see references 18, 19). Cholesterol saturation index is expressed as cholesterol content (mol%) divided by the cholesterol saturation. Choi = cholesterol; TG = triglycerides; BA = bile acids; PL = phospholipids CO CO en co ton, assachusetts) in 2 ml of ethanol and approximately 50 ml of orange juice to enable measurement of bile acid pool size by the method of Duane et al. 14 In the morning, the position of the tube was checked by x-ray, and if the tip was not in the duodenum, the tube was repositioned and gastric emptying was facilitated by the intravenous injection of 5 mg of metoclopromide (A. H. Robbins Company, Richmond, Virginia). As close as possible to 14 hours after the evening meal, gallbladder contraction was stimulated by intravenous injection of 0.02 /xg/kg sincalide (E. R. Squibb & Sons, Incorporated, Princeton, New Jersey). Bile was aspirated, and 1 to 2 ml of the darkest bile was immediately pipetted into 20 ml of chloroform/methanol (2:1). Bile Analyses The chloroform/methanol solution was brought to boiling in a hot-water bath and cooled to room temperature. The precipitated proteins were removed by filtration through Whatman No. 1 filter paper. The

3 ENOIBRATE AND BILE LIPID COPOSITION Palmer 633 Table 1. BA (Continued) Bile lipids (m) PL Choi BA Bile lipids (mol%) PL Choi Choi* sat. (%) Chol sat. index paper was washed with the solvent, and the combined solvent was made up to 20.0 ml. Exactly 5 ml was taken for cholesterol determination by the method of Hanel and Dam. 15 Aliquots of 0.1 and 0.2 ml were taken to dryness for total phosphorus determinations by the method of Bartlett, 18 and additional aliquots were taken for enzymatic analysis of total bile acids as described previously. 17 Separate aliquots were counted for radioactivity using external standards for quench correction Experimental Design All subjects (seven women and eight men) had an initial bile analysis either before beginning the medication or during an initial single-blind placebo period. Patients then entered the double-blind study and received either placebo or 300 mg/day of fenofibrate according to an allocation scheme that reserved certain study numbers at each center for bile-study patients. Patients then had a second bile analysis between 3 and 6 months of the double-blind study, except for Patient 212 who was studied again only after entering the open study. An additional four patients in the placebo group were also studied a third time after 3 to 6 months of fenofibrate administration during the open study. Calculations Cholesterol saturation indices (CSI) were calculated from the tables of Carey 18 or from the equations of Carey and Small, 19 based on a gallbladder bile total lipid of 15 g/dl (a concentration typical of healthy controls but slightly higher than usually seen in patients with gallstones). Comparisons between groups were made by paired or unpaired Student's f tests as applicable.

4 634 ARTERIOSCLEROSIS VOL 5, No 6, NOVEBER/DECEBER 1985 Results Data on the subjects and the results of the bile and plasma lipid analyses are shown in Table 1. Seven patients were assigned to the placebo group (six of whom were studied for the second time during placebo administration), and eight were assigned to the fenofibrate group. ive of the placebo patients were also studied again during open-label fenofibrate administration. Statistical analyses of paired bile data from the double-blind study are presented in Table 2, and analyses of paired data from the open-label study are presented in Table 3. Unpaired comparisons of all data are given in Table 4. The placebo group and the fenofibrate group were comparable at baseline. There were three women and four men in the placebo group, with a mean age of 50.1 ± 13.5 years () and a mean weight of 148 ± 34.6 pounds. There were four women and four men in the fenofibrate group, with a mean age of 56.3 ± 7.3 years and a mean weight of 174 ± 44.7 pounds. The baseline cholesterol and triglycerides during the single-blind placebo period were 297 ± 72 mg/dl and 160 ± 62 mg/dl in the placebo group, and 289 ± 47 mg/dl and 239 ± 125 mg/dl in the fenofibrate group. Thus, the fenofibrate group was slightly older, heavier, and had higher triglyceride levels than the placebo group, but the differences were not statistically significant. Bile Acid Pool Size The bile acid pool averaged 9.4 ± 8.4 mmol in placebo determinations (n = 21) and 8.4 ± 3.4 mmol in fenofibrate determinations (n = 13). However, measurements of bile acid pool size were quite varied, and it is doubtful that the method gave reliable information when used in this manner. Certainly, no firm conclusions can be drawn, but there was no suggestion that fenofibrate influenced pool size. Table 2. Bile Lipids: Double-Blind Study Baseline Bile adds (mol%) (n = 6) (n = 8) Phospholipids (mol%) Cholesterol (mol%) Cholesterol saturation (mol%) Cholesterol saturation index 'Unpaired t = 1.81; p = funpaired t = -1.94; p = Bile Lipids When the individual bile lipids were examined, it was clear that fenofibrate produced substantial changes in the relative amounts of the three constituents. In the double-blind study, fenofibrate decreased the relative proportion of bile acids from 70.5% to 59.1 % (p = 0.01), increased the proportion of phospholipids from 19.3% to 27.9% (p = 0.001), and increased cholesterol from 10.2% to 13.0% (not significant). In the placebo group, neither the change in bile acids from 69.5% to 66.2%, in phospholipids from 20.7% to 23.5%, nor the change in cholesterol from 9.8% to 10.3% was significant. When the treatment values for the placebo group and the fenofibrate group were compared by unpaired f tests, the differences in bile acids and phospholipids were not significant (p = 0.09 and 0.08 respectively). Similar results were obtained in five placebo patients when the baseline bile compositions were compared with the compositions on open-label fenofibrate, except that the differences were slightly larger and statistically more significant; during open-label fenofibrate, the increase in cholesterol was significant (p = 0.01), although not as striking as the increase in phospholipids (p = 0.008) or the decrease in bile acids (p = 0.003). inally, similar differences were observed when all 21 untreated samples (obtained prior to the study or while on placebo) were compared with all 13 fenofibrate samples. In this comparison, the changes were highly significant (Table 4). Bile Cholesterol Saturation and Cholesterol Saturation Index The increased proportion of phospholipids that resulted from fenofibrate administration increased the cholesterol-holding capacity of bile in spite of the decreased proportion of bile acids. This was a con- Treatment * 6.85* 4.08t 4.37f Difference Paired t P

5 ENOIBRATE AND BILE LIPID COPOSITION Palmer 635 Table 3. Bile Llplds: Open-Label Study Difference Bile acids (mol%) Phospholipids (mol%) Cholesterol (mol%) Cholesterol saturation (mol%) Cholesterol saturation index Paired t P Results shown are for five patients studied first at baseline and again during open label fenofibrate (see Table 1). sistent finding in both studies, and also was seen when the results of all analyses were pooled. Nevertheless, the increased cholesterol content in the open-label study was still sufficient to increase the cholesterol saturation index (CSI) significantly, although the increase in the double-blind study was less and not statistically significant. When all of the results were pooled, the difference in the CSI was statistically significant at the p = 0.02 level. Changes In Plasma Llplds and Llpoprotelns At the time of the second bile study (during the double-blind period), the plasma cholesterol in the fenofibrate group had fallen from 289 mg/dl to 226 mg/dl (p = 0.02), whereas the plasma cholesterol in the placebo group was unchanged (285 mg/dl vs 287 mg/dl). Similarly, plasma triglycerides decreased in the fenofibrate group from 239 mg/dl to 121 mg/dl (p = 0.04), whereas they were unchanged in the placebo group (174 mg/dl vs 180 mg/dl). When changes in plasma lipoproteins were examined, LDL cholesterol increased in the placebo group (n = 6) from 191 mg/dl to 202 mg/dl (5.6%) and decreased in the fenofibrate group (n = 8) from 177 mg/dl to 133 mg/dl (-25%); the difference was highly significant (p < 0.001). Conversely, HDL cholesterol in the placebo group was unchanged at 46 mg/dl, while it rose in the treated group from 46 mg/dl to 53 mg/dl (15%). The mean plasma cholesterol value for the total group of the 21 untreated samples was 291 mg/dl, and the mean value for the total group of 13 fenofibrate samples was 236 mg/dl. The corresponding triglyceride values were 186 mg/dl and 110 mg/dl. There was no apparent relationship in the treated patients between the extent of plasma lipid reduction and the change in biliary lipids. Discussion The major finding of this study was a highly significant change in the relative proportions of bile lipids during fenofibrate therapy. The proportion of bile acids decreased, whereas the proportions of phospholipids and cholesterol increased. Because the proportions of the two solubilizers of cholesterol (bile acids and lecithin) changed in opposite directions, the effects on bile saturation with cholesterol were inconclusive. The experimental design of this study resulted in moderate variability in the determination of bile lipid composition. The analyses were based on only one sample of bile, whereas some studies (particularly those in bile fistula patients where sample collection is simple) have utilized multiple bile samples. In addition, it is not possible to be sure that all samples were strictly comparable, as it is impossible to quantitate the extent to which bile was diluted with intestinal and/or gastric contents (particularly when metaclopromide was used to facilitate tube passage) and to know whether the collected bile was representative gallbladder bile, diluted gallbladder bile, or hepatic bile. Thus, even though three samples had bile acid concentrations of less than 5 m, all samples were used in the analysis. Attempts were made to standardize the amount of time between the last meal (and hence the last stimulated gallbladder contraction) and the bile sample, as it has been shown that bile saturation changes with the length of time the enterohepatic circulation has been interrupted by fasting; 20 ' 21 nevertheless, some variation in this interval was inevitable. The variables just discussed did not prevent the demonstration of highly significant changes in bile iipid composition, and even omitting the dilute bile Table 4. Bile Llplds: All Studies Combined Bile acids (mol%) Phospholipids (mol%) Cholesterol (mol%) Cholesterol saturation (mol%) Cholesterol saturation index (n = 21) (n = 13) Unpaired t P <0.001 < <

6 636 ARTERIOSCLEROSIS VOL 5, No 6, NOVEBER/DECEBER 1985 samples from the analysis did not alter the findings appreciably. In general, these changes are qualitatively similar to those seen in previous studies examining the effect of clofibrate on bile lipids, 3-5 ' " 9 which have shown a decrease in the molar percentage of bile acids and an increase in the molar percentage of phospholipid and/or cholesterol. In air clofibrate studies, the changes in bile lipid composition resulted in an increased CSI, whereas the qualitatively similar changes that we observed with fenofibrate were quantitatively slightly different, and the resulting increase in cholesterol-holding capacity prevented the demonstration of a significant increase in the CSI except when the results of all the studies were pooled. There have been limited studies on the effect of other clofibrate derivatives on bile composition. ost, 7 ' but not all, 24 of the studies demonstrated similar changes. Two of the studies included fenofibrate. Schlierf et al. 7 found that the increase in the CSI produced by fenofibrate was slight and statistically not significant (lipid composition not given). In recent studies published while the present results were being analyzed, von Bergmann and Leiss 23 found that fenofibrate decreased bile acids from 72.7% to 5%, increased bile cholesterol from 8.2% to 14.3%, and increased bile phospholipids from 19.1 % to 27.1 %. The lipid changes were similar to those seen in the present study, but did result in a statistically significant increase in the CSI from 1.31 to The change in CSI found by von Bergmann and Leiss was almost identical to the change observed in the present study for the pooled analyses (1.35 vs 1.72). The mechanism by which these compounds alter bile lipid composition has not been completely established, in part because the mechanisms governing the secretion of bile constituents are not entirely understood. The relative decrease in bile acids and increase in both cholesterol and phospholipids would be consistent with either a decrease in bile acid secretion (presumably secondary to decreased synthesis and pool size, or possibly decreased recycling), an increase in the secretion of cholesterol and phospholipids, or both. The present studies provide no direct evidence on this question. Previous work 22 has shown that bile acid secretion rates are unaffected by ciprofibrate. Clofibrate does decrease bile acid synthesis, 2 either directly by affecting enzymatic (7ahydroxylase) conversion of cholesterol to bile acids, or indirectly by reducing the microsomal (perhaps regulatory) pool of cholesterol that serves as substrate. An indirect effect is suggested because: 1) clofibrate inhibits whole body cholesterol synthesis; 2 2) fenofibrate inhibits 3-hydroxy-3-methylglutaryl CoA (GG-CoA) reductase in monocytes; 25 and 3) hepatic LDL receptors are upregulated by bezafibrate and fenofibrate, and whole body receptor-mediated uptake of LDL is increased. 26 ' 27 These studies argue persuasively that the major effect of these drugs is to decrease the regulatory microsomal pool of cholesterol, probably through a primary effect on cholesterol synthesis, with secondary effects of increased LDL uptake and reduced bile acid synthesis. However, the secondary increase in LDL uptake would provide more cholesterol and tend to normalize bile acid synthesis; the extent to which that happens may indicate the extent to which LDL cholesterol ester can be hydrolyzed and compete with newly synthesized cholesterol as a substrate for bile acid synthesis. The second possible explanation for the change in bile lipid composition would be enhanced secretion of cholesterol and phospholipids. Studies with ciprofibrate in eight patients showed a statistically significant increase in bile cholesterol secretion for the five patients in whom bile cholesterol saturation increased. 22 In the studies of Grundy et al., 2 clofibrate appeared to mobilize cholesterol from tissues. This hypothesis is supported by the early studies of Grundy and ok, 4 in which the bile changes induced by clofibrate appeared to be transient, and by the studies of Angelin et al. 8 in which the lithogenic effect of clofibrate that was apparent after 6 weeks of treatment had largely disappeared in studies done after 6 months of treatment. Since the half-time for transfer of cholesterol between the slowly turning over pool and the plasma is 35 days, 28 6 months of treatment might be expected to result in mobilization of most of the excess tissue stores and lead to a new equilibrium. If these inferences are correct, it might explain why the present results were quantitatively somewhat less significant than those of other studies, as most of the analyses in this study were done after 3 to 6 months of therapy (150 ± 41 days for the eight fenofibrate patients during the double blind study: 179 ± 34 days for the five patients on open-label fenofibrate). Thus, it seems reasonable to suggest that the period of greatest vulnerability with respect to developing cholelithiasis resulting from the administration of clofibrate derivatives may be limited to a few months. On the other hand, the decreased proportion of bile acids shown in the present study could still be demonstrated after 3 to 6 months. It is possible that agents such as clofibrate and fenofibrate or ciprofibrate may share a common mechanism of action with respect to reducing serum lipids and mobilizing cholesterol from tissues, but may differ in having different effects on the coupling of bile acid secretion to cholesterol and phospholipid secretion; these effects may result in differing effects on the bile cholesterol saturation index. Such differences in coupling might lead to changes in the ratio of cholesterol/phospholipids in bile, and increases have been seen in studies where the hypolipidemic agent was associated with increases in bile cholesterol saturation indices. 8 - ^ a Such changes were not marked in the present study, where the ratio changed from 0.47 to 0.49, and the changes did not correlate well with changes in the saturation indices in individuals. The present study indicates that fenofibrate use should be monitored to detect any possible tendency

7 ENOIBRATE AND BILE LIPID COPOSITION Palmer 637 to increase gallstone formation. The changes in bile saturation were barely significant and not consistent, but the changes in lipid composition were definite and similar to those seen with clofibrate. However, it should be noted that changes in lipid composition can not be used to predict with any certainty whether an increased rate of gallstone formation will occur or not, and it is important to emphasize that increases in bile cholesterol content or saturation do not necessarily lead to gallstone formation. It is well established that the bile of most people is supersaturated with cholesterol at some time or other, usually in the morning, 21 and it may well be that the relative proportions of substances that enhance or inhibit nucleation or crystal formation are more important than the cholesterol content of bile in determining whether gallstones will occur in any individual. 29 " 33 It would be of interest to ascertain whether the administration of clofibrate affects the nucleation time of bile and, if so, whether this property is shared by fenofibrate and other clofibrate derivatives. But despite these reservations, the changes we found make it prudent to recognize the possibility that fenofibrate may increase gallstone formation. Even if it should be proven that fenofibrate does predispose to gallstone formation, this fact alone should not constitute a contraindication to the use of this drug in appropriate circumstances. As with any drug, the risks have to be weighed against the benefits. In this case, the possible increased risk would have to be quantified and weighed against the benefits of effective serum lipid reduction. In this small study, fenofibrate reduced LDL cholesterol by 25%, while it increased HDL cholesterol by 15%, resulting in approximately a 30% decrease in the LDL/HDL ratio. (These results are slightly greater than, though not dissimilar to, those obtained in the larger, multicenter trial to be reported separately.) It is possible that enhanced bile secretion of cholesterol is a concomitant of any hypolipidemic agent that mobilizes cholesterol from tissues, and that some increased risk of gallstone formation is inherent in all effective agents. It is also possible that any such risk could be minimized or abolished by concomitant bile acid administration, 7 and that any new stones (presumably being pure cholesterol stones), would be highly susceptible to medical dissolution with bile acids. Clearly, more work needs to be done on the relation of hypolipidemic agents to bile cholesterol secretion. or the present, the changes observed in this study should be construed as a warning, not as a contraindication. Whether they have established a definitive or unusual risk remains to be determined. Acknowledgments The help of arian Adams and the staff of the General Clinical Research Center in doing the clinical studies and the excellent assistance of Dean Gittleman and Lakshmi Gollapudi in the laboratory are gratefully acknowledged. References 1. Thistle JL, Schoenfleld LJ. Induced alterations in composition of bile of persons having cholelithiasis. Gastroenterology 1971;61:488^»96 2. Grundy S, Atirens EH Jr, Salon G, SchreJbman PH, Nestel PJ. echanisms of action of clofibrate on cholesterol metabolism in patients with hyperlipidemia. J LJpid Res 1972; 13: Pertsemlldls D, Panvellwalla D, Ahrens EJ Jr. Effects of clofibrate and of an estrogen-progestin combination on fast- Ing biliary lipids and cholic acid kinetics In man. Gastroenterology 1974;66: Grundy S, ok HYI. Colestipol, clofibrate, and phytosterols in combined therapy of hyperlipidemia. J Lab Clin ed 1977:89: Bateson C, aclean D, Rosa PE, Bouchler IAD. Clofibrate therapy and gallstone induction. Digestive Dis 1978; 7: Angelln B, Elnarsson K, LeIJd B. Biliary lipid compositions during treatment with different hypolipidaemic drugs. Eur J Clin Invest 1979 #: Schllerf G, Chwat, euerborn E, et al. Biliary and plasma lipids and lipid-lowering chemotherapy. Atherosclerosis 1980; 36: Angelln B, Elnarsson K, Leljd B. Clofibrate treatment and bile choleterol saturation: short-term and long-term effects and influence of combination with chenodeoxycholic acid. Eur J Clin Invest 1981 ;11: Kesanleml YA, Grundy S. Clofibrate, caloric restriction, supersaturatkxi of bile, and cholesterol crystals. Scand J Gastroenterol 1983;18: The Coronary Drug Project Research Group. Gallbladder disease as a side effect of drugs Influencing lipid metabolism. N Engl J ed 1977;296: Cooper J, Geizerova H, Oliver. Clofibrate and gallstones. Lancet 1975;1: Palmer RH. Prevalence of gallstones in hyperlipidemia and incidence during treatment with clofibrate and/or cholestyramine. Trans Assoc Am Phys 1978^1: LJpid Research Cllnlca Program. anual of laboratory operations: LJpid and lipoprotein analysis (NIH) publication Bethesda, aryland: National Heart Lung and Blood Institute, Duane WC, Adler RD, Bennlon LJ, Ginsberg RL. Determination of bile ackj pool size in man: a simplified method with advantages of increased precision, shortened analysis time, and decreased isotope exposure. J LJpid Res 1975;16: Hanoi HK, Dam H. Determination of small amounts of total cholesterol by the Tschugaeff reaction with a note on the determination of lanosterol. Acta Chem Scand 1955;9: Bartlett GR. Phosphorus assay in column chromatography. J Biol Chem 1959;234: Palmer RH. The enzymic assay of bile acids and related 3ahydroxysteroids: its application to serum and other biological fluids. In: Colowtek SP, Kaplan N, eds. ethods in enzymology. V: Steroids and terpenoids. New York: Academic Press, 1969: Carey C. Critical tables for calculating the cholesterol saturation of native bile. J Lipid Res 1978;19: Carey C, Small D. The physical chemistry of cholesterol solubility in bile. Relationship to gallstone formation and dissolution In man. J Clin Invest 1978;61: Bloch H, Thornton JR, Heaton KW. Effects of fasting on the composition of gallbladder bile. Gut 1980;21: etzger AL, Adler R, Heymsfleld S, Grundy S. Diurnal variation in biliary lipid composition. N Engl J ed 1973; 288: Angelln B, Elnarsson K, Leljd B. Effect of ciprofibrate treatment on biliary lipids in patients with hyperiipoproteinaemia. Eur J Clin Invest 1984;14: von Bergmann K, Lelss O. Effect of short-term treatment with bezafibrate and fenofibrate on biliary lipid metabolism in

8 638 ARTERIOSCLEROSIS VOL 5, No 6, NOVEBER/DECEBER 1985 patients with hyperlipoproteinaemia. Eur J Clin Invest 1984; 14: Takeuchi N, Kukita H, Kajiyama G et al. Effect of clinofibrate, a new hypolipidemic agent, on biliary and serum lipids in patients with hyperlipidemia. Atherosclerosis 1982;42: Schneider AG, Ditschuneit HH, Strange EE, Ditschuneit H. Regulation of 3-hydroxy-3-methylglutaryl Coenzyme A reductase in freshly isolated human mononuclear cells by fenofibrate. In: Carlson LA, Olsson AG, eds. Treatment of hyperlipoproteinemia. New York: Raven Press, 1984: Stewart J, Packard CJ, Lorimer AR, Boag DE, Shepherd J. Effects of bezafibrate on receptor mediated and receptor independent low density lipoprotein catabolism in Type II hyperlipoproteinemic subjects. Atherosclerosis 1982;44: Shepherd J, Caslake J, Lorimer AR, Vallance BD, Packard CJ. reduces low density lipoprotein catabolism in hyper- triglyceridemic subjects. Arteriosclerosis 1985; 5: Smith R, Dell RB, Noble RP, Goodman DS. Parameters of the three-pool model of the turnover of plasma cholesterol in normal and hyperlipidemic humans. J Clin Invest 1976; 57: Holzbach RT, Kibe A, Thiel E, Howell JH, arsh, Hermann RE. Biliary proteins. Unique inhibitors of cholesterol crystal nucleation in human gallbladder bile. J Clin Invest 1984;73: Holan KR, Holzbach RT, Hermann RE, Cooperman A, Claffey WJ. Nucleation time: a key factor in the pathogenesis of cholesterol gallstone disease. Gastroenterology 1979; 77: Small D. Cholesterol nucleation and growth in gallstone formation. N Engl J ed 1980;302: Sedaghat A, Grundy S. Cholesterol crystals and the formation of cholesterol gallstones. N Engl J ed 1980;302: Burnstein, llson RG, Petrunka CN, Taylor RD, Strasberg S. Evidence for a potent nucleating factor in the gallbladder bile of patients with cholesterol gallstones. Gastroenterology 1983;85: Index Terms: fenofibrate bile hypolipidemia gallstones