Article Text

Download PDFPDF

Increased deoxycholic acid absorption and gall stones in acromegalic patients treated with octreotide: more evidence for a connection between slow transit constipation and gall stones
  1. A F Hofmann
  1. Correspondence to:
    Professor A F Hofmann
    Department of Medicine, MC 0813, University of California, San Diego, La Jolla CA 92093-0813;

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Acromegalic patients treated with octreotide have prolonged colonic transit, increased bacterial formation, and subsequent absorption of deoxycholic acid that is indicated by an increased proportion of deoxycholic acid in plasma bile acids. Enrichment of deoxycholic acid in the circulating bile acid pool leads to supersaturated bile and cholesterol gall stones.

Thomas and colleagues,1 in this issue of Gut, extend previous work from the laboratory of R Hermon Dowling by showing that acromegalic patients treated with octreotide have an increase in faecal anaerobic bacteria that convert cholic acid, the major primary bile acid in humans, to deoxycholic acid, the major secondary bile acid (see page 630). The paper also confirms previous reports from the Dowling laboratory2,3 that octreotide treatment increases colonic transit time in acromegalic patients. In addition, the paper of Thomas et al shows that octreotide treatment results in enrichment of deoxycholic acid in fasting state plasma bile acids.

The present paper is the 11th and could well be the last in a series of papers spanning more than a decade from the Dowling laboratory that have dealt with the pathogenesis of gall stones in acromegalic patients treated with octreotide, a potent somatostatin agonist. Octreotide is a synthetic peptide whose amide bonds are resistant to hydrolysis by plasma peptidases, thus resulting in its having a much longer duration of action than that of somatostatin.4 In acromegalic patients, octreotide administration suppresses the release of growth hormone and insulin growth factor by the pituitary and results in clinical improvement.5

The initial paper in this series showed that such patients have an increased prevalence of gall stones.6 The gall stones were shown to be rich in cholesterol7 and, as would be expected, bile was found to be supersaturated in cholesterol8 and to dissolve when patients received ursodiol orally,7 an agent that is well known to decrease the cholesterol saturation of bile and induce gradual dissolution of cholesterol gall stones.9

In addition, previous papers have shown that such patients have impaired gall bladder contraction2 and prolonged small intestinal and colonic transit,2,3 as well as an increased proportion of deoxycholic acid in biliary bile acids.8 To define a mechanism for the increased biliary deoxycholic acid, methods for assessing bacterial deconjugating and dehydroxylating activity in intestinal or faecal content were developed,10 as these enzymatic steps are necessary for the conversion of cholic acid to deoxycholic acid. Such methods were used in the present paper to show increased dehydroxylating activity (per gram of faecal protein or wet weight). A previous study11 showed that colonic absorption of newly formed deoxycholic acid was increased in acromegalic patients treated with octreotide. This excellent study used the isotope dilution method for characterisation of bile acid kinetics developed a half century ago by Lindstedt (reviewed by Hofmann and Hoffman12). Moreover, there was a linear correlation between colonic transit (presumably the independent variable) and deoxycholic acid input.11

The increased colonic absorption was explained by two factors. Firstly, the residence time of colonic content was prolonged. Secondly, it was proposed, based on previous studies of others,13 that slow colonic transit leads to more alkaline colonic content (although this was not actually measured) and that if luminal contents of the colon were more alkaline, more deoxycholic acid should be in solution. If more deoxycholic acid is in solution, more should be absorbed, as deoxycholic acid is highly membrane permeable. To prove that the solubility of deoxycholic acid increases as pH is raised, in vitro studies of deoxycholic acid solubility in caecal content were performed. These studies showed that the solubility of deoxycholic increased linearly with increasing pH.14 However, binding of deoxycholic acid to luminal contents appeared to influence the pH solubility relationship, because theory predicts that solubility should increase exponentially rather than linearly with increasing pH—that is, in theory, for each unit increase in pH, solubility should increase 10-fold.15 Finally, studies on gall bladder bile samples showed that the greater the deoxycholic acid proportion in biliary bile acids, the greater the cholesterol saturation and the shorter the crystal detection time.16 It was shown that bile that was enriched in deoxycholyl conjugates had vesicles containing a higher cholesterol/phospholipid ratio, presumably because dihydroxy bile acids have a lower critical micellisation concentration than trihydroxy bile acids. Therefore, for a given bile acid concentration, dihydroxy bile acids solubilise more phospholipid in mixed micelles than trihydroxy bile acids. Such phospholipid solubilisation depletes the phospholipid from phospholipid-cholesterol vesicles in bile, leaving cholesterol rich vesicles that nucleate cholesterol more rapidly.17

One must express admiration for the tenacity with which this multifaceted problem has been pursued as well as for the great variety of experimental approaches that have been used. Certainly new insights into deoxycholic acid metabolism have been developed. In addition, a hypothesis for the pathogenesis of the gall bladder stones occurring in acromegalic patients receiving octreotide has been advanced, and a number of lines of evidence supporting this reasoning have been advanced. The purpose of this brief commentary is to highlight the achievements to date and point out a few remaining experimental approaches that seem to be needed to clinch the argument.

The first question is whether octreotide administration per se induces the formation of supersaturated bile in the acromegalic patient. Supersaturated bile is not uncommon in the healthy Caucasian adult18 and appears to be frequent in acromegalic patients, at least in one study.19 Growth hormone administration to healthy subjects20 or to children deficient in growth hormone21 does not alter biliary lipid composition. In a small prospective study, Erlinger et al found that acromegalic patients had bile that was supersaturated in cholesterol and that octreotide administration did not increase the degree of cholesterol supersaturation but was associated with the appearance of cholesterol crystals.19 These results differ from those of the Dowling group who found that octreotide administration caused biliary cholesterol saturation to increase from unsaturated to supersaturated.8 Thus it seems that most acromegalic patients who are treated with octreotide begin with, or because of octreotide therapy develop, gall bladder bile that is supersaturated in cholesterol and will rapidly form cholesterol crystals.

In addition to its effect on biliary lipid composition, octreotide administration causes a profound disturbance of biliary tract motility. Octreotide abolishes both cholecystokinin release from the small intestine22 as well as the contractile response of the gall bladder to infused cholecystokinin.23 Therefore, at first glance, the effect of octreotide should be to cause a “non-functioning” gall bladder. However, an additional effect of octreotide should be to inhibit the usual prandial relaxation of the sphincter of Oddi, an action now known to be mediated by local release of nitric oxide.24 The coordinated contraction of the gall bladder and relaxation of the sphincter of Oddi leads to controlled emptying of gall bladder contents into the duodenum during digestion, which is evidenced by the postprandial rise in the plasma bile acid level.25 Remarkably, the flow and ebb of the bile acid pool stimulated by meal ingestion is perturbed relatively little by cholecystectomy, as in cholecystectomised patients the pool is stored in the small intestine rather than in the gall bladder, and quickly transported to the distal ileum and absorbed when a meal is ingested.25

Bile acid secretion can be assessed indirectly by changes in the plasma bile acid level as fractional hepatic extraction of bile acids remains constant irrespective of the load.26 At present, we have no information on the dynamics of bile acid secretion in the acromegalic patient treated with octreotide. However, it is reasonable to propose that bile acid secretion is markedly impaired in acromegalic patients because of the paralysis of the gall bladder and sphincter of Oddi. An intraluminal bile acid deficiency is likely to contribute to the steatorrhoea that has been observed in patients with somatostatinomas27 as well as in patients with other maladies treated with octreotide.28 Patients with the AIRE syndrome may also have impaired cholecystokinin release and have been shown to have an intraluminal bile acid deficiency.29 Indeed, malabsorption of fatty acids as such has been shown to result from octreotide administration,30 an effect that is best explained by a decreased intraluminal bile acid concentration.

If bile acid secretion is impaired in acromegalic patients treated with octreotide, this could provide an explanation for the formation of supersaturated bile. Older clinical studies showed that during overnight fasting, bile acid secretion decreased to a greater extent than cholesterol secretion, increasing the saturation of bile.31 The one problem with this line of reasoning is that when bile is stored for prolonged periods in the gall bladder, cholesterol is absorbed to a greater extent than phospholipid or cholesterol, decreasing the saturation of bile.32,33 The gall bladder bile samples obtained by the Dowling group were obtained by percutaneous puncture, and we have no idea how long the bile had been present in the gall bladder. It would have been possible to determine such by using a biliary recovery marker such as indocyanine green,34 but this was not done.

Does octreotide have a direct effect on hepatocyte secretion of cholesterol? Biliary cholesterol secretion is still poorly understood although the two ATP stimulated efflux pumps that mediate biliary cholesterol secretion into canalicular bile, ABC5 and ABC8, have recently been identified.35 The effect of somatostatin on bile flow has received considerable attention36 but the effect of this hormone or its agonists on biliary cholesterol secretion has had relatively little attention. None the less, Marteau et al noted that when somatostatin was administered to a patient with a biliary fistula, bile became supersaturated in cholesterol.37

Another question is why gall bladder atony and its colonic equivalent—prolonged colonic transit—develop when the acromegalic patient is treated with octreotide. As noted above, failure of gall bladder contraction during meals is easily explained by the dual action of octreotide. Constipation is presumed to result from a direct inhibitory effect of somatostatin on colonic propulsive motility.38

The extensive studies of the Dowling group have shown elegantly that prolonged intestinal transit leads to increased deoxycholic acid absorption from the colon. The paper by Thomas et al in this issue of Gut1 shows that faecal samples contain increased dehydroxylating activity in faecal samples. In a previous study, she and her coworkers showed that bacterial activity of faecal samples is similar to that of caecal samples.39 None the less, as bacterial density is usually expressed in logarithmic units, the increased bacterial density reported by Thomas et al, although statistically significant, is really quite modest.

The Dowling group has argued strongly for a key role of deoxycholic acid in the pathogenesis of supersaturated bile and cholesterol gall stones in these patients based on their observation that increased biliary deoxycholic acid was associated with increased biliary cholesterol saturation and rapid cholesterol crystal formation in vitro. Older studies of Marcus and Heaton show clearly that slowing colonic transit increases biliary deoxycholic acid and increases biliary cholesterol saturation.40 The Carulli group showed that enriching bile in deoxycholic acid by feeding cholic acid, its precursor, induced the formation of supersaturated bile.41 Berr et al identified a group of gall stone patients with greatly increased deoxycholic acid formation. When these patients were given ampicillin to decrease deoxycholic acid formation, bile became less saturated in cholesterol.42 Additional mechanisms by which increased deoxycholic acid absorption can promote the formation of cholesterol gall stones are summarised in a scholarly review by Van Erpecum and Van Berge-Henegouwen.43 A novel hypothesis by which increased deoxycholic acid input could cause increased biliary cholesterol saturation was first advanced by Einarsson et al44 based on studies in healthy subjects. These workers found that deoxycholic acid feeding causes inhibition of cholesterol 7α-hydroxylase activity (presumably by activating FXR, the nuclear receptor that modulates bile acid synthesis). At the same time, deoxycholic acid did not inhibit HMG CoA reductase activity (presumably by not activating the PPARα nuclear receptor45). This disassociation of cholesterol synthesis from cholesterol catabolism into bile acids provides a possible mechanism for the induction of supersaturated bile secretion by deoxycholic acid.

None the less, the two largest studies of biliary bile acid composition and biliary cholesterol saturation46,47 have found little relationship between the proportion of deoxycholic acid and the extent of supersaturation with cholesterol. Moreover, when deoxycholic acid was given to healthy volunteers at rates exceeding the usual input by a factor of 7–10, little change in biliary cholesterol saturation occurred.48

We are left with the intriguing possibility, also proposed by others, that deoxycholic acid may have a biphasic dose-response curve: low doses (or input) in the range of the normal daily input from the colon increase biliary cholesterol saturation whereas higher (pharmacological) doses do not alter biliary cholesterol saturation. Can this be tested? What is needed is a simple dose-response study of oral deoxycholic acid, as has been performed for chenodeoxycholic acid and ursodeoxycholic acid.49 This should be done first in acromegalic patients receiving octreotide, to establish the veracity of the hypothesis developed by the Dowling group, and second, in cholesterol gall stone patients.

A final question is whether the formation of deoxycholic acid matters for any other reason. Bile acids are bacteriostatic and their high aqueous concentration in the small intestine contributes to its relative sterility.50 In the colon, bacterial deconjugation and dehydroxylation converts conjugates of cholic acid to deoxycholic acid (and chenodeoxycholyl conjugates to lithocholic acid); these secondary bile acids are poorly soluble at caecal pH, permitting bacteria to flourish. For herbivores, bacteria mediate the conversion of unabsorbed carbohydrate to short chain fatty acids, which are an important and even essential caloric source for hind gut fermenters such as the horse. Thus, in herbivores, formation of deoxycholic acid which results in a profound decrease in the aqueous concentration of luminal bile acids appears to be beneficial to the whole organism. Humans, however, obtain few calories from their colon,51 and probably most adults wish that they absorbed none.

Experimental oncologists however take a dim view of deoxycholic acid. A recent study by Pai et al has claimed that low concentrations of deoxycholic acid increase tyrosine phosphorylation of beta catenin and enhance colon cancer cell proliferation and invasiveness.52 Deoxycholic acid has also been reported to induce COX-2 activity53 and to activate epidermal growth factor receptor.54 A cautionary note seems in order as these cancer promoter activities of deoxycholic acid have been shown in colon cell lines. Such an experimental situation may not mimic the in vivo situation because bile acids may not accumulate in the colonic epithelium if the blood supply is ample. Bovids presumably have a continuing flux of deoxycholic acid across their colonic epithelium, and to the best of my knowledge colon cancer is rare in bovids.

The extensive body of work from the Dowling group has greatly enriched our knowledge of the factors influencing the absorption of deoxycholic acid from the colon, a silent process that is occurring in all of us. The acromegalic patient treated with somatostatin has multiple “lithogenic” defects in his or her biliary tract—bile that is supersaturated with cholesterol, a short nucleation time, and impaired gall bladder emptying that provides time for cholesterol crystals to form and sink to the bottom of the gall bladder. In addition, the sphincter of Oddi is unlikely to relax and as a consequence bile acid secretion is likely to be grossly deficient, an additional factor promoting the secretion of bile that is supersaturated in cholesterol. As discussed above, a number of lines of evidence suggest that increased absorption of deoxycholic acid from the colon promotes the secretion of bile that is supersaturated in cholesterol. None the less, deoxycholic acid input needs to be varied experimentally in order clinch the view that increased deoxycholic acid is a major contributor to gall stone formation in both acromegalic patients receiving octreotide as well as in patients who have idiopathic cholesterol gall stone disease.

Deoxycholic acid and ammonia appear to be the two toxins that we absorb continuously from our colon and that cause little impairment of health so long as the liver is detoxifying normally and the gall bladder is contracting nicely. Based on older work of Heaton55 and the recent studies of the Dowling group, we now have a hypothesis supported by a great deal of experimental and clinical evidence that can explain the association between constipation caused by slow colonic transit and cholesterol gall stones.56 Development of this principle is a major achievement in gastroenterology in my judgment.


The author’s work is supported by a grant-in-aid from the Academic Senate of the University of California.

Acromegalic patients treated with octreotide have prolonged colonic transit, increased bacterial formation, and subsequent absorption of deoxycholic acid that is indicated by an increased proportion of deoxycholic acid in plasma bile acids. Enrichment of deoxycholic acid in the circulating bile acid pool leads to supersaturated bile and cholesterol gall stones.



  • Conflict of interest: None declared.

Linked Articles