Secondary Logo

Journal Logo

Original Articles: Gastroenterology

Altered Bile Acid Metabolism in Childhood Functional Constipation: Inactivation of Secretory Bile Acids by Sulfation in a Subset of Patients

Hofmann, Alan F*; Loening-Baucke, Vera; Lavine, Joel E; Hagey, Lee R*; Steinbach, Joseph H*; Packard, Christine A§; Griffin, Terrance L§; Chatfield, Dale A§

Author Information
Journal of Pediatric Gastroenterology and Nutrition: November 2008 - Volume 47 - Issue 5 - p 598-606
doi: 10.1097/MPG.0b013e31816920a6
  • Free


Constipation is a common problem of children worldwide. Estimates of the prevalence of functional constipation in the pediatric population have varied from 1% to 30%, with a median of 9% (1). Constipation is a frequent cause of medical visits (2) and has been reported to be a predictor of irritable bowel syndrome in later life (3). Symptoms of constipation vary from mild and short lived to severe and chronic and are sometimes accompanied by fecal impaction, fecal and urinary incontinence, urinary tract infections, and abdominal pain (4–6).

The cause of childhood functional constipation is not known. We hypothesized that some cases of childhood constipation are caused by a deficiency of secretory bile acids in the colon. Several lines of evidence have provided support for the assumption that dihydroxy bile acids act as endogenous laxatives. First, 2 dihydroxy bile acids—chenodeoxycholic acid (3α,7α-dihydroxy, CDCA), a primary bile acid, and deoxycholic acid (3α,12α-dihydroxy, DCA), a secondary bile acid—induce concentration-dependent secretion when perfused into the human (7) or canine (8) colon. Second, in in vitro studies, using polarized monolayers of T84 cells, secretion is induced by these 2 bile acids and by lagodeoxycholic acid, the 12β-hydroxy epimer of DCA (9). Secretion is not induced by ursodeoxycholic acid (3α,7β-dihydroxy, UDCA). This study, together with other in vivo studies (10,11) indicate that there are strict structural requirements for secretion induced by bile acid: of the 18 possible dihydroxy or hydroxy-oxo bile acids, only these 3 bile acids induce secretion. Third, patients with bile acid malabsorption have a secretory diarrhea that responds to bile acid sequestrant administration (12). Fourth, ingestion of CDCA induces a dose-related diarrhea (13). Fifth, both CDCA (14) and cholic acid (15), which is metabolized by bacterial 7α-dehydroxylation to DCA, are effective in the treatment of adult constipation. Finally, a frequent side effect of cholestyramine, a bile acid sequestrant, is constipation (16).

We postulated 2 possible mechanisms by which altered bile acid metabolism would reduce the concentration of CDCA or DCA in the colon. The first possibility was an inborn deficiency of sterol 12 hydroxylase (cyp8B1). Were such to occur, primary bile acids would consist entirely of the conjugates of CDCA. When mature colonic flora develop, the CDCA would be converted by bacterial 7α-dehydroxylation to lithocholic acid (LCA), which is devoid of secretory activity (9). The second possibility was that bacterial epimerization of the 3α-hydroxy group of DCA to a 3β-hydroxy group would be greatly increased. Such epimerization converts DCA to isoDCA, a bile acid that is devoid of secretory activity in T84 cells and presumably in vivo.

The aim of our study was to determine fecal bile acid composition in a group of children with functional constipation and in a group of nonconstipated control children using state-of-the-art analytic techniques to test the hypothesis that altered bile acid metabolism is present and possibly causal in some children with functional constipation.


Children With Functional Constipation

Children with functional constipation, from 1 month to 12 years old, who were seen at the University of Iowa Children's Hospital or the Rady Children's Hospital in San Diego were eligible for the study provided that constipation met the criteria for being functional constipation and constipation was not present at birth but had developed during the first year of life. In functional constipation, there is no evidence for the presence of an inflammatory, anatomic, metabolic, or neoplastic process. Functional constipation was defined in children 2 years old or older by 2 or more of the following characteristics during the previous 8 weeks: frequency of bowel movements of fewer than 3 stools per week, large stools in the rectum or felt on abdominal examination, passing of stools so large that they obstructed the toilet, retentive posturing (withholding behavior), painful defecation, and in children older than 4 years old 1 more episodes of fecal incontinence per week (4–6,17,18). These criteria have been developed at the University of Iowa for the definition of constipation and have been used for the past 20 years in its primary and tertiary care clinics. This definition was also recently agreed upon by an international group of pediatric gastroenterologists and pediatricians gathered at the Second World Congress of Pediatric Gastroenterology, Hepatology and Nutrition in Paris in July 2004. For constipation occurring in children younger than 2 years of age, the criteria were passage of hard or scybalous or pebble-like stools with straining/withholding or painful defecation (19).

Control Children

Control children were healthy siblings of constipated children or healthy children. The majority of the children were white, but children of many races and nationalities were included. The study was approved by the Institutional Review Boards of the University of Iowa, the Rady Children's Hospital of San Diego, San Diego State University, and the University of California, San Diego. All of the parents gave written informed consent, and at the University of Iowa, children 7 years old and older also gave written assent.

Analytic Methods

Chemical Standards

Bile acid standards were of high purity and were obtained from a variety of sources, as described (20).

Fecal Samples

Fecal samples were obtained from 207 children, 103 with functional constipation and 104 who had normal bowel habits. The majority of samples (n = 172) were from the University of Iowa, and the remainder (n = 35) were from the Rady Children's Hospital, San Diego.

Stool samples (∼1 g) were placed in polyethylene liquid or glass liquid scintillation vials containing 20 mL reagent grade isopropanol. The vials were sealed tightly, and the cap–vial juncture was wrapped with parafilm. The vial was labeled with a code permitting subsequent patient identification. Vials containing the suspension of fecal material were kept in the refrigerator and, when convenient, sent by express mail to UCSD. On receipt at UCSD, samples were thoroughly homogenized with a high-speed homogenizer (Tissue Tearor, Biospec Products, Bartlesville, OK). Samples were then stored at a refrigerated temperature.

Determination of bile acid classes by electrospray ionization–single ion monitoring–mass spectroscopy (ESI-SIM-MS) and ESI-MS-MS. An aliquot of the isopropanol supernate was taken for analysis. It was assumed that extraction of bile acids from the fecal matrix by isopropanol would be identical in samples from constipated and nonconstipated children. The large number of samples to be analyzed precluded the use of more elaborate extraction procedures (21). The aliquot was diluted with at least 5 volumes of 0.1 mol/L sodium bicarbonate. Bile acids were then adsorbed to a C-8 reversed-phase column (Isolute C8 SPE; International Sorbent Technology, Mid Glamorgan, UK), rinsed with water, and desorbed with methanol.

The ESI-SIM-MS was performed at San Diego State University on a Thermo TSQ Quantum quadrapole mass spectrometer operating in the negative mode, as described previously (20). The sheath gas flow was 70, and the surface-induced dissociation voltage was 30 V. The following mass per charge (m/z) values were measured: for C24 unconjugated bile acids, 373, mono-oxo-cholanoate; 375, monohydroxy cholanoate (presumably LCA or isoLCA); 389, mono-oxo, hydroxy cholanoates; 389, unsaturated or mono-oxo, monohydroxy cholanoates; 391, dihydroxy cholanoates; 405, mono-oxo, dihydroxy cholanoates; and 407, trihydroxy cholanoates. For C24 sulfated nonamidated bile acids, m/z values were as follows: 455, sulfate of monohydroxy-cholanoates; 471, monosulfate of dihydroxy cholanoates; and 487, monosulfate of trihydroxy cholanoates. For C24 amidated nonsulfated bile acids, m/z values were as follows: 498, taurine conjugated dihydroxy cholanoates, and 514, taurine conjugated trihydroxy cholanoates. Also measured were m/z values corresponding to 4 C27 bile alcohol monosulfates: 499, trihydroxy; 515, tetrahydroxy; 531, pentahydroxy; and 547, hexadroxy. Finally, m/z values for 3 taurine conjugated C27 bile acids were also measured: 540, dihydroxy; 556, trihydroxy, and 572, tetrahydroxy cholestanoates. Glycine amidates were not measured because they were considered unlikely to be present.

Spectra were quantified using the detector response from samples injected into a flow of methanol/water, 95:5, vol/vol, and calibration factors were obtained from model mixtures, after background subtraction. All of the samples were run in duplicate with excellent agreement between duplicates.

The ESI-MS-MS was performed at the University of California, San Diego, to confirm the identity of assigned m/z values by use of a Perkin-Elmer Sciex API-III instrument (Perkin-Elmer, Alberta, Canada) modified with a nanoelectrospray source from Protana A/S (Odense, Denmark) as described (20). Chemical identity of peaks was confirmed by the fragmentation pattern of selected ions (Q3 mode) by use of argon collision gas and by comparison with known standards. In particular, conjugates losing sulfate (m/z 97) or taurine (m/z 124) were identified.

Identification of Individual Unconjugated and Conjugated Bile Acids by Liquid Chromatography

The liquid chromatography (LC)-MS/MS analysis of selected samples was also performed on the TSQ Quantum at San Diego State University as described previously (20). The samples prepared as described above were evaporated to dryness, then redissolved in methanol containing 5 internal standards [tauroursocholic acid (3α7β,12α-trihydroxy-5β-cholan-24-oyl-taurine), glycolagodeoxycholic acid (3α,12β-dihydroxy-5β-cholan-24-oyl-glycine), 2,2′,4,4′-2H-CA, 2,2′,4.4′-2H-DCA) and 7α-hydroxy-5β-cholan-24-oic acid)]. These compounds acted as retention time standards. The samples were then separated on a Hypersil-Keystone BetaBasic C18 column and eluted with an ammonium acetate buffer/methanol gradient. Detection was by selected reaction monitoring, and quantitation was performed against a 5-point standard curve prepared for each compound; the detector response for the reference compounds was linear in relation to concentration but varied widely between individual bile acids. Samples were analyzed for unconjugated bile acids (ursodeoxycholic acid, CA, CDCA, DCA, and LCA) and for their corresponding taurine N-acyl amidates, glycine N-acylamidates, and sulfate conjugates. The 3β-hydroxy epimers of the unconjugated bile acids were also measured; however, the taurine, glycine, and sulfate conjugates of these 3β-hydroxy epimers were not measured because standards were not available. In addition, isoCDCA (3β,7α-dihydroxy) and isoDCA (3β,12α-dihydroxy) could not be quantified because the 2 compounds had identical retention times on the C18 columns. Consequently, only 23 analytes were measured.

Identification of Individual Unconjugated Bile Acids by Gas Chromatography-Mass Spectrometry

Bile acids were determined by gas chromatography (GC)-MS of the methyl ester per-acetates (22). The triacetate of methyl hyocholate was used as an internal standard. Because GC-MS is not nearly as sensitive as ESI-MS, several samples had insufficient bile acid content for analysis. In addition, samples from 6 patients who were found to have predominantly 3-sulfo-CDCA by LC-MS were removed from the GC-MS database because the ester bond of 3-sulfo-CDCA underwent little hydrolysis during the acetylation procedure and the GC-MS results for these patients were not accurate. Therefore, the sample size for GC-MS analysis was considerably smaller than the sample size for ESI-MS analysis. In the 31 children 2 years old or younger for whom GC-MS data were available, 7-dehydroxylation was incomplete in 7. However, these samples were left in the database, even though they increased the variance.

Bile acids were isolated from the isopropanol supernatant as described above. Methyl esterification was performed by use of ethereal diazomethane, and per-acetates were prepared as described by Roovers et al (22). Chromatography was performed on a capillary column 30 m × 0.25 internal diameter with a stationery phase of methyl polysiloxane (65%)–phenylpolysiloxane (35%) coated at 0.25-μm thickness (J & W Scientific, Folsom, CA). The gas chromatograph was a Hewlett-Packard HP-5890 coupled to a mass selective detector Hewlett Packard HP-5970. Compounds were identified by their retention times and fragmentation pattern based on an MS library of some 40 bile acids. The library was composed of most major and minor bile acids previously identified in fecal bile acids (21).

Data Analysis

Results are expressed as mean ± standard deviation. Data from the 2 groups (constipated and control) were analyzed without and with separate analyses of children younger than 2 years of age, this being the age by which adult colonic flora should have developed. Inasmuch as there was no statistically significant difference between the 2 groups, they were combined. We found no effect of laxative ingestion, and data from children ingesting laxatives (most frequently polyethylene glycol) were included. Samples from children receiving antibiotics were also removed from the database, with 1 exception (see Results). In some cases, more than 1 sample was submitted for the same individual; when that was done, only data from the first sample were included.

Data are shown as box or whisker plots. Differences between the constipated group and the nonconstipated control group were tested for statistical significance using the Student t test and the Mann-Whitney test, which is a test for nonparametric data. Differences with a P value of <0.05 by both tests were considered to be significant.



The number of analyses entered in the final database was considerably smaller than the number of samples collected because of censoring (see above) and for various technical reasons such as inadequate specimen size. The ESI-SIM-MS analyses were based on samples from 73 children with functional constipation (46 boys and 27 girls) and 92 control children (57 boys and 35 girls). In the constipation group, the mean age was 4.9 years; the median age was 4.8 years; and ages ranged from 1 month to 12.9 years. In the control group, the mean age was 3.7 years; the median age was 3.3 years; and the ages ranged from 4 months to 10.8 years.

For GC-MS, data were obtained from 64 constipated children (39 boys and 25 girls). The mean age was 5.5 years; the median age was 5.0 years; and the ages ranged from 5 months to 12.9 years. Samples were analyzed from 64 control children (39 boys and 25 girls). The mean age was 4.3 years; the median age was 3.9 years; and the ages ranged from 1 year to 10.8 years.

Fecal Bile Acid Composition: Constipated Children Versus Control Children

Table 1 compares the ESI-SIM-MS composition of fecal bile acids in the constipated and control groups. C24 bile acids predominated in the 3 classes of cholesterol catabolites. No patient had an obvious defect in side chain oxidation, inasmuch as the proportion of C27 bile acids was low in all of the patients. Bile acids were largely in unconjugated form. The proportion of taurine conjugated bile acids was significantly lower in the constipated group, but the difference was small and could be explained by the more prolonged exposure of bile acids to colonic bacteria in the constipated patients. The percent sulfation was 9.5 ± 12.5% for monohydroxy bile acids in the constipated group and 7.0 ± 6.4% in the control group (NS). Monohydroxy- and dihydroxy- bile acids constituted 80% of fecal bile acids, indicating efficient 7α-dehydroxylation.

Fecal (unesterified) bile acid composition as determined by ESI-SIM-MS, %

The proportion of monosulfated dihydroxy bile acids in the constipated group (5.5%) was significantly greater (P < 0.05) than that in the control group (3.2%). In the constipated group, there were 6 outliers with a greatly increased proportion of monosulfated dihydroxy bile acids. These data are illustrated in Fig. 1.

FIG. 1
FIG. 1:
Monosulfated dihydroxy cholanoates (m/z 471), as percentage of measured bile acids, in children with functional constipation (left) and healthy control children (right) by ESI-SIM-MS. Box denotes 25% to 75% confidence limits. Bars of whiskers indicate 10% and 90%. Median is indicated by horizontal line inside box. Asterisk denotes the mean. Monosulfated dihydroxy cholanoates were significantly greater in children with functional constipation than in nonconstipated control children (P < 0.05). In the 6 children with the highest values, the monosulfated dihydroxy cholanoate was identified as the 3-sulfate of CDCA by LC-MS.

To confirm the high proportion of sulfated dihydroxy bile acids in the constipated group, and to identify which dihydroxy bile acids were sulfated, samples from the constipated patients with a high proportion of sulfated dihydroxy bile acids were examined by LC-MS, a technique permitting definitive assignment of structure. Results from the 6 patients, all younger than 3 years of age, are shown in Table 2. The dominant dihydroxy sulfate was the 3-sulfate of CDCA. Figure 2 shows an ESI-MS of one of these patients and that of a constipated child having a normal fecal bile acid composition.

Fecal bile acid composition as determined by LC-MS in 6 constipated children with a high proportion of sulfated dihydroxy bile acids*
FIG. 2
FIG. 2:
ESI-MS of fecal bile acids from a constipated child with predominantly sulfated dihydroxy bile acids (left) and a constipated child with a normal fecal bile acid pattern (right). The major fecal bile acid in the constipated child has anm/z value of 471, corresponding to a monosulfate of a dihydroxy C24 bile acid that was shown by LC-MS to be the 3-sulfate of CDCA. In the child with a normal fecal bile acid profile, the major peak has an m/z value of 391, corresponding to that of a dihydroxy bile acid. For structures corresponding to other m/z values, see Methods.

Two of these 6 children had growth retardation. Five of the 6 were receiving no medications other than laxatives at the time of the study; 1 child was receiving Septra at the time of collection for a urinary tract infection. A second sample was received from 1 child 6 months later; in both, 3-sulfo-CDCA was the major compound present. On follow-up averaging 1 year later, constipation had resolved in 3 of 5 children, had probably resolved in 1, and was unchanged in 1. One child was lost to follow-up.

Table 3 gives GC-MS data for the 2 groups. The following bile acids composed more than 95% of bile acids in all samples: LCA, isoLCA, DCA, isoDCA, CDCA, isoCDCA, UDCA, isoUDCA, and cholic acid. Hydroxy-oxo bile acids were not present in our samples. The 4 bile acids—isoLCA, LCA, isoDCA, and DCA—composed nearly 100% of all fecal bile acids in 88% of constipated children and in 91% of control children. The remaining bile acid was mostly CDCA, with smaller proportions of UDCA and cholic acid. There was no difference in the fecal bile acid pattern between samples from children with functional constipation and those from nonconstipated control children. No child was identified with a total absence of DCA, thus eliminating our first hypothesis.

Fecal bile acid composition by GC-MS, % (mean ± SD)

Our second hypothesis was that increased epimerization of DCA to isoDCA may occur, thus abolishing the secretory activity of DCA. Data showing the percentage isoLCA and LCA for individual children is illustrated in the left panel of Fig. 3; data for isoDCA and DCA are shown in the right panel. The proportion of iso- bile acids was identical in both groups, refuting our second possible mechanism. Epimerization of LCA was much greater than that of DCA.

FIG. 3
FIG. 3:
Left, Percent of isoLCA (3β-hydroxy) and LCA (3α-hydroxy) in children with functional constipation and control children. Right, Percent of isoDCA (3β,12α-dihydroxy) and DCA (3α,12α-dihydroxy) in children with functional constipation and control children.

Age and 7-Dehydroxylation

There were 31 children 2 years old or younger whose bile acids were examined by GC-MS. In 24 of those children, fecal bile acids were fully (>90%) dehydroxylated. Six children 4 months to 19 months old had limited (10–90%) 7-dehydroxylation. One child 2.0 years old had no 7-deoxy bile acids These data, which are illustrated in Fig. 4, indicate that by 1 year an anaerobic flora that dehydroxylates bile acids is present in most but not all children.

FIG. 4
FIG. 4:
Relation between age and percent bile acid 7-dehydroxylation in fecal bile acid samples from children ≤2 years old with functional constipation (&.cirf;) and control children (○).


Fecal Bile Acid Composition

This study indicates that the majority of children with functional constipation have a fecal bile acid profile that does not differ from control subjects. Nonetheless, the dominant fecal bile acid of a small fraction of constipated children is the 3-sulfate of CDCA. To our knowledge, such an abnormality in bile acid metabolism has not been reported previously in studies of fecal bile acid composition from adults in whom the sulfate fraction of fecal bile acids was estimated after class separation by ion exchange chromatography directly (23–25). Sulfation of CDCA and DCA is known to abolish their secretory activity in the perfused rat colon (26), raising the possibility that the absence of DCA in the fecal bile acids of these children whose bile acids were largely in sulfated form contributed to their functional constipation.

The route by which the 3-sulfate of CDCA is formed is not known. We think it is unlikely that it was secreted in bile, based on LC-MS analyses of biliary bile acids for some 60 children younger than 1 year old (unpublished data) in which sulfated bile acids were only trace constituents. The literature also indicates that sulfated dihydroxy bile acids are minor constituents of biliary bile acids in adults (27,28). In human cecal content, the proportion of DCA considerably exceeds that of CDCA because of rapid bacterial 7α-dehydroxylation (20). By contrast, in these fecal samples CDCA sulfate greatly exceeded DCA sulfate.

Our working hypothesis is that (unconjugated) CDCA is generated by bacterial deconjugation in the distal small intestine. CDCA is readily absorbed by passive mechanisms (29,30) and could undergo sulfation in the ileal enterocyte, inasmuch as sulfotransferase has been shown to be present in the ileal epithelium (31,32) The resulting 3-sulfate could then be pumped back into the intestinal lumen by 1 or more apical transporters such as multidrug resistance–associate protein 2 (MRP2), which are known to be highly expressed in the human small intestine (33–35). In addition, the sulfo-CDCA may also be transported across the basolateral membrane of the enterocyte, return to the liver, and be secreted in bile (36). In any event, during transit through the colon, most of the 3-sulfo-CDCA escapes bacterial hydrolysis and is excreted as such in fecal bile acids. Bile acid biosynthesis is well known to be controlled in a negative feedback manner by the return of bile acids to the hepatocyte. Whether formation of 3-sulfo-CDCA, a bile acid that should not undergo enterohepatic cycling, leads to increased bile acid biosynthesis is not known.

A problem with this hypothesis is that it should also apply to orally administered CDCA or UDCA, and to our knowledge, sulfates of these administered bile acids have not yet been detected in plasma or feces in people receiving these bile acids as therapeutic agents. Moreover, given that bacterial deconjugation occurs normally in the distal ileum (37), this possible pathway should be present to some extent in everyone. It may not be obvious, however, because bacterial hydrolysis of bile acid sulfates is well known to occur during colonic transit in human beings (38) and in rodents (39). The spontaneous resolution of constipation in 3 of the 5 children with high sulfation is not inconsistent with the proposed role of bacteria in the formation of the sulfo-CDCA.

The proportion of sulfated bile acids did not differ greatly from values reported in the literature when sulfated bile acids were measured after being separated as a class by ion exchange chromatography (23–25). The proportion of DCA was lower than that reported in most studies of fecal bile acids (21), but the lower proportion of DCA and isoDCA is likely to be the result of the presence in the fecal sample of ester derivatives of DCA (40) that were not measured because of the absence of an alkaline hydrolysis step. Our data nonetheless provide information on the fecal (nonesterified) bile acid profile of healthy children, and they indicate the consistent presence of iso (3β-hydroxy) bile acids in both constipated and nonconstipated children. It is likely that isoLCA and isoDCA are absorbed in part from the colon; if so, they will be reepimerized to LCA and DCA, respectively, during transport through the hepatocyte, on the basis of studies in rats (41,42) and humans (43).

Limitations of the Study

A defect in our study is that fecal bile acid excretion was not measured. Bile acid excretion in the steady state is equal to bile acid biosynthesis, and it is entirely possible that bile acid biosynthesis is decreased in some constipated children. To measure bile acid biosynthesis by measuring fecal bile acid output would require prolonged hospitalization for complete fecal collections or the use of a steady-state fecal recovery marker (44). Measurement of fecal bile acid proportions provides no information on the aqueous concentration of bile acids in the proximal colon, where CDCA and DCA may modulate fluid content by their concentration-dependent secretory activity (9).

Methodological Considerations

Each of the analytic methods used in this study has limitations. First, the extraction procedure, homogenization in isopropanol, may not have given complete recovery of all bile acids. Second, for ESI-MS-MS, reference compounds were not available for all m/z values, and each peak could contain multiple compounds, each with differing detector responses. In addition, any peak measured by ESI-MS could contain non–bile acid compounds (i.e., unabsorbed dietary constituents). Therefore, the ESI-MS technique is sensitive, but not specific; and the results must be considered semiquantitative. They do, however, permit comparisons of the 2 groups.

In the GC-MS method used, deconjugation and solvolysis procedures were omitted, as was a weak alkaline hydrolysis procedure, because of the large number of samples to be processed. Deconjugation was considered unnecessary because multiple studies have indicated that fecal bile acids are predominantly in unconjugated form (21), and this was confirmed by our ESI-SIM-MS analyses. A solvolysis procedure was judged to be unnecessary because the acetylation procedure was shown to hydrolyze the 3-sulfate of LCA. However, in the 6 samples containing high proportions of the 3-sulfate of CDCA by LC-MS, CDCA was not present in the GC-MS results, indicating that this sulfate is not cleaved during the acetylation procedure. As noted above, these samples were excluded from the GC-MS results. Omission of a weak alkaline hydrolysis procedure meant that bile acids in esterified form may not have been measured, with the exception of the 3-sulfate of LCA. Esters that have been reported include ethyl esters (45) and long-chain fatty acid esters (46) of LCA, and nucleic acid esters of DCA (47) or undefined esters of DCA and isoDCA (40). We judged that such esters would have no secretory activity, and they are estimated to constitute only about 25% of fecal bile acids (23,40). Omission of a weak alkaline hydrolysis procedure and solvent extraction also meant that plant sterols were present in the methyl ester per-acetates. These had different retention times and were readily distinguished from bile acids by their fragmentation patterns. Spectra were not recorded from peaks occurring before isoLCA (3βOH), so Δ3-cholenic acid, known to be formed by bacterial dehydroxylation of LCA (48,49) was not measured. However, this compound should have no secretory activity (9). On balance, the analytic methods should have been sufficiently robust to detect any difference between the study and control groups.

It seems that LC-MS is the optimal analytic method because it permits unequivocal identification of individual bile acids. The technique requires a large set of bile acid standards and high resolution by LC. Separation of fecal bile acids into individual classes (21,23–25) before LC-MS is desirable but is labor intensive.

In conclusion, we have identified a small subset of children with functional constipation with a new abnormality in bile acid metabolism that may play a role in childhood functional constipation. In these children, fecal bile acids are composed mostly of the 3-sulfate of CDCA, a conjugated bile acid devoid of secretory activity. We found 6 such children in a cohort of 73 patients, indicating that the defect is uncommon in children with functional constipation. Formation of 3-sulfo-CDCA can be explained by abnormal bacterial deconjugation in the small intestine, formation of the sulfate in the enterocyte, followed by extrusion back into the intestinal lumen. The children were identified by the finding of increased monosulfated, dihydroxy bile acids using ESI-SIM-MS, but we cannot define the sensitivity and specificity of this method because LC-MS, the gold standard, was not performed on all of the specimens. Detection of this abnormality in bile acid metabolism requires mass spectrometry, but this technique is more and more frequently used in clinical chemistry. If the formation of this sulfated bile acid results from bacterial overgrowth in the small intestine, then such children may benefit from treatment with antibiotics.


The authors thank Prof Takashi Iida, Department of Chemistry, Nihon University, Japan, for samples of the 3- and 7 sulfates of CDCA.


1. Van den Berg MM, Benninga MA, Di Lorenzo C. Epidemiology of childhood constipation: a systematic review. Am J Gastroenterol 2006; 101:2401–2409.
2. Chitkara DK, Talley NJ, Weaver AL, et al. Incidence of presentation of common functional gastrointestinal disorders in children from birth to 54 years: a cohort study. Clin Gastroenterol Hepatol 2007; 5:186–191.
3. Khan S, Campo J, Bridge JA, et al. Long-term outcome of functional childhood constipation. Dig Dis Sci 2007; 52:64–69.
4. Loening-Baucke V. Functional fecal retention with encopresis in childhood. J Pediatr Gastroenterol Nutr 2004; 38:79–84.
5. Loening-Baucke V. Urinary incontinence and urinary tract infection and their resolution with treatment of chronic constipation of childhood. Pediatrics 1997; 100:228–232.
6. Loening-Baucke V. Factors determining outcome in children with chronic constipation and faecal soiling. Gut 1989; 30:999–1006.
7. Mekhjian HS, Phillips SF, Hofmann AF. Colonic secretion of water and electrolytes induced by bile acids: perfusion studies in man. J Clin Invest 1971; 50:1569–1577.
8. Mekhjian HS, Phillips SF. Perfusion of the canine colon with unconjugated bile acids: effect on water and electrolyte transport, morphology, and bile acid absorption. Gastroenterology 1970; 50:120–129.
9. Keely S, Scharl M, Bertelsen LS, et al. Bile acid induced secretion in polarized monolayers of T84 colonic epithelial cells: structure-activity relationships. Am J Physiol 2007; 292:G292–G297.
10. Chadwick VS, Gaginella TS, Carlson GL, et al. Effect of molecular structure on bile acid-induced alterations in absorptive function, permeability, and morphology in the perfused rabbit colon. J Lab Clin Med 1979; 94:661–674.
11. Gordon SJ, Kinsey MD, Magen JS, et al. Structure of bile acids associated with secretion in the rat cecum. Gastroenterology 1979; 77:38–44.
12. Hofmann AF, Poley JR. Role of bile acid malabsorption in pathogenesis of diarrhea and steatorrhea in patients with ileal resection: I. Response to cholestyramine or replacement of dietary long chain triglyceride by medium chain triglyceride. Gastroenterology 1972; 62:918–934.
13. Schoenfield LJ, Lachin JM, Steering Committee, and National Cooperative Gallstone Study Group. Chenodiol (chenodeoxycholic acid) for dissolution of gallstones: the National Cooperative Gallstone Study. Ann Intern Med 1981; 95:257–282.
14. Bazzoli F, Malavolti M, Petronelli A, et al. Treatment of constipation with chenodeoxycholic acid. J Int Med Res 1983; 11:120–123.
15. Hepner GW, Hofmann AF. Cholic acid therapy for constipation: a controlled study. Mayo Clin Proc 1973; 48:356–358.
16. Knodel LC, Talbert RL. Adverse effects of hypolipidemic drugs. Med Toxicol 1987; 2:10–32.
17. Voskuijl WP, Heijmans J, Heijmans A, et al. Use of Rome II criteria in childhood defecation disorders: applicability in clinical and research practice. J Pediatr 2004; 145:213–217.
18. Benninga MA, Candy DCA, Catto-Smith AG, et al. The Paris consensus on childhood constipation terminology (PACCT) group. J Pediatr Gastroenterol Nutr 2005; 40:273–275.
19. Hyams J, Colletti R, Faure C, et al. Functional gastrointestinal disorders: Working Group Report of the First World Congress of Pediatric Gastroenterology. Hepatology and Nutrition. J Pediatr Gastroenterol Nutr 2002; 35(Suppl 2):S110–S117.
20. Hamilton JP, Xie G, Raufman J-P, et al. Bile acids of human cecal contents: concentration and spectrum. Am J Physiol 2007; 293:G256–G263.
21. Setchell KDR, Street JM, Sjövall J. Fecal bile acids. In: Setchell KDR, Kritchevsky D, Nair PP, editors. The Bile Acids: Chemistry, Physiology, Metabolism Vol 4. New York: Plenum Press; 1988. pp. 441–570.
22. Roovers R, Evrard E, Vanderhaeghe H. An improved method for measuring human blood bile acids. Clin Chim Acta 1986; 19:449–457.
23. Korpela JT, Fotsis T, Adlercreutz H. Multicomponent analysis of bile acids in faeces by anion exchange and capillary column gas-liquid chromatography: application in oxytetracycline treated subjects. J Steroid Biochem 1986; 25:277–284.
24. Tanida N, Hikasa Y, Shimoyama T, et al. Comparison of faecal bile acid profiles between patients with adenomatous polyps of the large bowel and healthy subjects in Japan. Gut 1984; 25:824–832.
25. Breuer N, Dommes P, Tandon R, et al. Isolierung und Quantifizierung nicht-sulfatierter und sulfatierter Gallensaeure im Stuhl. J Clin Chem Biochem 1984; 22:623–631.
26. Breuer NF, Rampton DS, Tammar A, et al. Effect of colonic perfusion with sulfated and non-sulfated bile acids on mucosal structure and function in the rat. Gastroenterology 1983; 84:969–977.
27. Stiehl A, Raedsch R, Rudolph G, et al. Biliary and urinary excretion of sulfated, glucuronidated and tetrahydroxylated bile acids in cirrhotic patients. Hepatology 1985; 5:492–495.
28. Takikawa H, Beppu T, Seyama Y. Profiles of bile acids and their glucuronide and sulphate conjugates in the serum, urine, and bile from patients undergoing bile drainage. Gut 1985; 26:38–42.
29. Schiff ER, Small NC, Dietschy JM. Characterization of the kinetics of the passive and active transport mechanisms for bile acid absorption in the small intestine and colon of the rat. J Clin Invest 1972; 51:1351–1362.
30. van Berge Henegouwen GP, Hofmann AF. Pharmacology of chenodeoxycholic acid: II. Absorption and metabolism. Gastroenterology 1977; 73:300–309.
31. Dew MJ, Hawker PC, Nutter S, et al. Human intestinal sulfation of lithocholate: a new site for bile acid metabolism. Life Sci 1980; 27:317–323.
32. Chen G, Zhang D, Jing N, et al. Human gastrointestinal sulfotransferases: identification and distribution. Toxicol Appl Pharmacol 2003; 187:186–197.
33. Wang Q, Bhardwaj RK, Herrera-Ruiz D, et al. Expression of multiple drug resistance conferring proteins in normal Chinese and Caucasian small and large intestinal tissue samples. Mol Pharm 2004; 1:447–454.
34. Hilgendorf C, Ahlin G, Seithel A, et al. Expression of thirty-six drug transporter genes in human intestine, liver, kidney, and organotypic cell lines. Drug Metabol Dispos 2007; 35:1333–1340.
35. Berggren S, Gall C, Wollnitz N, et al. Gene and protein expression of P-glycoprotein, MRP1, MRP2 and CYP3A4 in the small and large human intestine. Mol Pharm 2007; 4:252–257.
36. Gartner U, Goeser T, Stiehl A, et al. Transport of chenodeoxycholic acid and its 3-alpha and 7-alpha-sulfates by isolated perfused rat liver. Hepatology 1990; 12:738–742.
37. Northfield TC, McColl I. Postprandial concentrations of free and conjugated bile acids down the length of the normal human small intestine. Gut 1973; 14:513–518.
38. Cowen AE, Korman MG, Hofmann AF, et al. Metabolism of lithocholate in healthy man: II. Enterohepatic circulation. Gastroenterology 1975; 69:67–76.
39. Robben J, Caenepeel P, Van Eldere J, et al. Effects of intestinal microbial bile salt sulfatase activity on bile salt kinetics in gnotobiotic rats. Gastroenterology 1988; 94:494–502.
40. Norman A. Faecal excretion products of cholic acid in man. Br J Nutr 1964; 18:173–186.
41. Shefer S, Salen G, Hauser S, et al. Metabolism of iso-bile acids in the rat. J Biol Chem 1982; 257:1401–1406.
42. Cronholm T, Makino I, Sjövall J. Steroid metabolism in rats given 1,2-2H-ethanol: oxidoreduction of isomeric 3-hydroxycholanoic acids and reduction of 3-oxo-4-cholenoic acid. Eur J Biochem 1972; 26:251–258.
43. Marschall HU, Roeb E, Yildiz Y, et al. Study of human isoursodeoxycholic acid metabolism. J Hepatol 1997; 26:863–870.
44. Wilkinson R. Polyethylene glycol 4000 as a continuously administered nonabsorbable faecal marker for metabolic balance studies in human subjects. Gut 1971; 12:654–660.
45. Kelsey MI, Sexton SH. The biosynthesis of ethyl esters of lithocholic acid and isolithocholic acid by rat intestinal microflora. J Steroid Biochem 1976; 7:641–647.
46. Norman A, Palmer RH. Metabolites of lithocholic acid-24-C-14 in human bile and feces. J Lab Clin Med 1964; 63:986–1001.
47. Benson GM, Haskins NJ, Eckers C, et al. Polydeoxycholate in human and hamster feces: a major product of cholate metabolism. J Lipid Res 1993; 34:2121–2134.
48. Pacini N, Albini E, Ferrari A, et al. Transformation of sulfated bile acids by human intestinal microflora. Artzneimittelforschung 1987; 37:983–987.
49. Kelsey MI, Molina JE, Huang SK, et al. The identification of microbial metabolites of sulfolithocholic acid. J Lipid Res 1980; 21:751–759.

Bile acid conjugation; Bile acid sulfation

© 2008 Lippincott Williams & Wilkins, Inc.