Bile acid diarrhoea: Current and potential methods of diagnosis

Abstract

Chronic diarrhoea is common and mostly due to diarrhoea predominant irritable bowel syndrome. Diarrhoea predominant irritable bowel syndrome affects about 11% of the population; however, up to a third of these patients actually have bile acid diarrhoea. There are, therefore, more than one million sufferers of bile acid diarrhoea in the UK. Bile acid diarrhoea is caused by small bowel malabsorption of bile acids and the increased bile acids in the large intestine cause diarrhoea. Once diagnosed, the treatment of bile acid diarrhoea is simple and effective. Bile acid diarrhoea , however, is often not diagnosed because of a lack of easily available and reliable diagnostic methods. In the United Kingdom, the radiolabelled 23-seleno-25-homotaurocholic acid test is the gold-standard method of diagnosis. 23-seleno-25-homotaurocholic acid test, however, is expensive, inconvenient to the patient, involves radiation exposure and has limited availability. As such, a laboratory biomarker is desirable. This review briefly discusses the pathophysiology and management of bile acid diarrhoea and critically evaluates methods for its diagnosis, including serum 7α-hydroxy-4-cholesten-3-one, faecal bile acid measurement, serum fibroblast growth factor 19, urine-2-propanol, and the ¹⁴C-glycocholate breath and stool test.

Introduction

Irritable bowel syndrome (IBS) is estimated to affect 11% of the global population. ¹ IBS is a chronic, relapsing, often lifelong disorder with typical symptoms of abdominal pain, diarrhoea or constipation associated with bloating. IBS is classified into four subtypes depending on the primary stool composition; diarrhoea-predominant IBS (IBS-D), constipation-predominant IBS, mixed IBS and un-subtyped IBS. ² IBS is not life-threatening, but significantly diminishes the quality of life for sufferers. ³

It, however, is estimated that 15 to 50% of patients with IBS-D have bile acid diarrhoea (BAD),^4–6 but the lack of diagnostic tests makes it difficult to determine the exact prevalence. BAD may be associated with other gastrointestinal disorders, such as Crohn’s Disease (CD) or post-cholecystectomy, but may also be idiopathic (primary BAD). BAD is caused by bile acid malabsorption (BAM) in the small intestine. In healthy individuals, approximately 95% of bile acids (BAs) are reabsorbed in the ileum and returned to the liver via the enterohepatic circulation. ⁷ If bile acid (BA) reabsorption is reduced or small bowel BA concentrations overwhelm the resorptive capacity of the ileum, increased concentrations of BAs move into the large intestine and, by stimulating colonic motility and secretions, cause diarrhoea. BAM and BAD are often used interchangeably.

Physiology of bile acids and the enterohepatic circulation

In the classic BA synthesis pathway, the primary BAs, cholic acid and chenodeoxycholic acid are synthesized in the liver from cholesterol. The rate limiting step is catalysed by the enzyme cholesterol-7-alpha-hydroxylase P4507A1 (CYP7A1), which converts cholesterol to 7-alpha-hydroxy cholesterol. ⁸ CYP7A1 expression is upregulated by high cholesterol concentrations and down-regulated by high BA concentrations in the intestines. CYP7A1 is under negative feedback from fibroblast growth factor-19 (FGF19) via the ileal Farnesoid X Receptor (FXR). ⁹ It is important to note that there is an alternative extra-hepatic pathway that may be used if required. This pathway uses oxysterols rather than cholesterol as the starting substrate for BA synthesis. ¹⁰

The primary BAs are conjugated at the carboxyl residue with glycine or taurine in the liver. ¹¹ Conjugation has several benefits, including minimizing passive absorption and promoting resistance to pancreatic enzyme cleavage. The conjugated BAs are secreted from the liver via a bile salt export pump into bile ducts and then transported, concentrated and stored in the gallbladder. ¹⁰ In response to feeding, primary BAs are secreted into the duodenum via the common bile duct to enable digestion of fat and fat soluble vitamins by promoting emulsification. This aids the action of lipase and the formation of micelles in the intestine. ¹² Some reabsorption of BAs occurs passively throughout the length of the small intestine, but 95% of bile acids are reabsorbed by active transport in the terminal ileum, and returned to the liver via the enterohepatic circulation for recycling. ¹³

There are several transport molecules involved in ileal reabsorption. ¹⁴ Within the brush-border membranes of the terminal ileum are apical sodium-linked bile salt transporters (ASBT), which move BAs into the enterocyte in a sodium-dependent manner. The BAs are then intracellularly transported by ileal bile acid binding protein and are extruded into portal circulation through the organic solute transporter heterodimer (OSTα and β).^15,16 Within the enterocyte, the BAs activate FXR which promotes gene expression and synthesis of FGF19. FGF19 is then transported via the portal circulation to the liver and activates fibroblast growth factor-receptor 4 (FGF-R4) which suppresses expression of CYP7A1, ¹⁷ thereby inhibiting BA synthesis. If BA concentrations in the intestine and/or terminal ileal cells are low, then FXR is not activated and FGF19 is not synthesized. This reduction in FGF19 leads to absence of feedback inhibition, increasing CYP7A1 expression and BA synthesis.

BAs not reabsorbed in the ileum enter the colon and are excreted in the faeces. In the colon, the primary BAs are converted by bacteria into the secondary BAs, deoxycholic acid and lithocholic acid. About 75% of BAs reaching the colon are passively reabsorbed. In the colon, BAs stimulate secretion of fluids and colonic motility.^5–7

The BA pool is recycled up to 12 times per day, owing to the high energy cost of new bile acid production. This results in approximately 200–600mg of BAs being lost into the faeces each day. ¹³ Every day around 800 mg of cholesterol is synthesized and approximately 400 mg of this is used for BA synthesis. ¹⁸

Classification of BAD

BAD is classified into three types, depending on the cause of the BA malabsorption. Type I BAD is secondary to ileal resection or inflammation, as may occur in CD. Type 2 BAD is also known as primary or idiopathic BAD, and is hypothesized to be caused by increased BA synthesis, potentially as a result of defective feedback inhibition or ASBT mutations. ⁴ Type 3 BAD is secondary to various other gastrointestinal disorders such as cholecystectomy, coeliac disease, small intestinal bacterial overgrowth, post radiation enteritis and chronic pancreatitis.

Pathophysiology of BAD

In Type I BAD, the loss of the primary location for active and passive reabsorption of bile acids owing to ileal resection or disease results in increased BA delivery to the large intestine. ¹⁵ Type 2 (primary) BAD is not clearly understood. Mutations in SLC10A2, ¹⁹ which codes for the ASBT, leading to failure of BA reabsorption have been reported but are very rare. Rather than malabsorption, it has been suggested that the primary mechanism for Type 2 BAD is increased BA production due to defective negative feedback on BA synthesis. ¹⁵ The excessive delivery of BAs to the ileum exceeds its capacity for reabsorption leading to increased BAs in the colon causing diarrhoea. Although less likely, Type 2 BAD may also result from rapid small bowel transport bypassing the active BA reabsorption. ⁷ As discussed, FGF19 is a regulator of BA synthesis in the liver. A recent study reported that patients with Type 2 BAD had significantly lower FGF19 concentrations compared to controls, and this was inversely correlated with serum 7α-hydroxy-4-cholesten-3-one (C4) concentrations, ²⁰ a marker of BA synthesis. This suggests that defective or reduced FGF19 may lead to increased BA synthesis and subsequently BAD.

Since the aetiology of Type 3 BAD is varied it is likely that the underlying pathophysiological mechanisms will also be diverse. In small bowel bacterial overgrowth, increased or altered bacterial deconjugation of BAs reduces the efficiency of BA reabsorption. Other gastrointestinal diseases, through increasing intestinal motility, may reduce the time for ileal reabsorption of BAs ¹⁵ and reduce the time for bacterial conversion of primary to secondary BAs thereby reducing the efficiency of BA reabsorption. Further research is, therefore, required to fully establish the pathophysiological and molecular mechanisms underlying Types 2 and 3 BAD.

Diagnosis of BAD

Diarrhoea is defined as the abnormal passage of liquid/loose stools, either greater than three times per day, or generating more than 200 g of stool per day. ²¹ Chronic diarrhoea is defined by the National Institute for Health and Care Excellence (NICE) as diarrhoea lasting more than four weeks. ²² Diagnosis of BAD is, however, often missed owing to difficulties in accessing a suitable diagnostic test, and subsequently patients are diagnosed with IBS. The four most explored options for diagnosis of BAD are radiolabelled 23-seleno-25-homotaurocholic acid (SeHCAT) test, measurement of faecal bile acids (FBA), serum-C4 measurement and a trial of bile acid sequestrants. Other proposed options include serum FGF19 measurement, ¹⁴C-glycocholate breath and stool test and urine-2-proponol. These are discussed below.

SeHCAT scan

The current gold standard for the diagnosis of BAD is the SeHCAT test, which assesses BA retention in the body. ²¹ After an overnight fast, the patient takes a capsule containing radiolabelled SeHCAT, a synthetic BA. This is followed by two full-body gamma camera scans, the first 1 to 3 h after taking the capsule, and the second seven days later. The gamma count from the second scan is then divided by the count from the first scan, and expressed as percentage retention of BAs. In health, the majority of BAs are reabsorbed and recirculated but retention of BAs is reduced in BAD. Retention values of greater than 15% are generally classed as normal; ²¹ however, different centres may have different cut-off points, including an indeterminate range between 12 and 19% retention.

It has been suggested that the test may not detect some patients with idiopathic BAD, as the causative mechanisms of idiopathic BAD are diverse and not well understood. Some patients may reabsorb relatively normal amounts of BAs in the ileum, but overproduction of BAs results in saturation of the retention capacity of the ileum. ²³ These patients may demonstrate a relatively normal SeHCAT retention, but increased production means that excess BA still move into the large intestine.

Whilst the SeHCAT test is considered the gold standard test in the UK, it is time consuming, expensive and results in exposure to radiation. Furthermore, it is not widely available, so patients often have to travel to have the test performed. The SeHCAT test is not currently licenced for use in the USA, which complicates comparison of diagnostic methods.

Faecal bile acid measurement

Several laboratories, particularly in the USA, have developed methods for measuring faecal bile acid (FBA) excretion. FBA are expected to be higher in patients with BAD, as decreased reabsorption in the ileum results in increased faecal excretion. There are several published enzymatic and liquid chromatography tandem mass spectrometry (LC-MS/MS) methods for measurement of FBA.^24–26

Enzymatic methods generally utilize an NAD-dependent steroid dehydrogenase to oxidize deconjugated BAs and produce NADH, which is measured. ²⁴ An extraction process is required to remove the BAs from the stool. Owing to the variety of conjugation processes that BAs undergo, this method may underestimate total FBA, particularly if hydrolysis time is not sufficient, as different conjugates undergo hydrolysis at different rates. ²⁴ Furthermore, stool samples may contain significant amounts of 3β-hydroxy and 3-oxo bile acids, which would not be oxidized by the commonly used 3α-hydroxysteroid dehydrogenase (3α-HSD) and thus would not be measured. ²⁷ More recently, Immunodiagnostik (IDK) have developed a kit marketed to measure FBA on a random stool sample, in contrast to other methods which generally require at least a 48-h stool collection.

The processes used to extract BAs from faeces for LC-MS/MS analysis are complex due to the requirement for a very clean sample. Amplatz et al. described a method that takes several hours, and uses a small amount of freeze-dried stool which is incubated in sodium hydroxide (NaOH), followed by addition of distilled water, homogenization and addition of acetonitrile. Samples are vortexed, centrifuged and the supernatant dried down and re-suspended, prior to a solid phase extraction (SPE) stage with washing and elution. ²⁵ Wong et al. proposed incubating a homogenized stool sample with NaOH and sodium chloride, followed by centrifugation. The subsequent supernatant is collected and the extraction process twice repeated, and the pellet is then re-suspended in methanol and centrifuged. The extract is purified using SPE with several washing steps. ²⁶ These lengthy and complex processes make them impractical for use in a routine laboratory.

There are advantages and disadvantages for both enzymatic methods and LC-MS/MS methods. The LC-MS/MS method used by the Mayo Clinic in the US requires a 48-h stool sample collection following a four-day fat-controlled diet, making it inconvenient and unpleasant for the patient. The enzymatic assay developed by IDK requires a random stool sample, which is more convenient and less unpleasant for the patient and the laboratory. Mitchell et al. ²⁸ suggested that there is variable excretion in BAs throughout the day in different stool samples, therefore measurement of FBA in a random stool sample may result in missed diagnosis. This will need further clarification. The extraction methods prior to LC-MS/MS analysis are cumbersome and time consuming for the laboratory. Enzymatic methods are generally quicker and easier, but have the potential to underestimate total BAs within the stool. ²⁴ One benefit of LC-MS/MS analysis is that it allows quantitation of individual BAs within the stool, as opposed to the total BA measurement obtained in enzymatic methods. ²⁵

FBA measurement as a direct biomarker of BA excretion offers an attractive proposition for the diagnosis of BAD. The diagnostic accuracy of measurement of FBA, however, in comparison to SeHCAT testing has not yet been fully assessed due to unavailability of the former in the United Kingdom (UK) and the latter in the United States (US).

Bile acid sequestrant trial

A therapeutic trial of bile acid sequestrants in suspected BAD leading to an improvement in patient symptoms would support a diagnosis of BAD. BA sequestrants, however, are often poorly tolerated, making them a less attractive diagnostic option. Furthermore, patient non-compliance may result in misdiagnosis. A trial of BA sequestrants is, therefore, not recommended by the British Society of Gastroenterology for the diagnosis of BAD. ²⁹

Serum 7α-hydroxy-4-cholesten-3-one measurement

Cholesterol is converted by CYP7A1 to 7α-hydroxycholesterol and is the rate-limiting step in the classic BA synthesis pathway. The enzyme 3β-hydroxy-D5-C27-steroid dehydroxylase converts 7 α-hydroxycholesterol to 7α-hydroxy-4-cholesten-3-one (C4), which is the common precursor for the primary BAs, cholic acid and chenodeoxycholic acid. Serum C4 is, therefore, utilized as a biomarker of BA synthesis. Serum C4 would be expected to be higher in patients with BAD, as BA synthesis increases to compensate for the increased faecal BA loss. Several studies have looked at the utility of serum C4 as a biomarker of BAD;^7,30–32 however, its adoption as a routine test has been limited. Reported methods use LC-MS/MS or high performance liquid chromatography (HPLC); with extraction protocols varying between studies. Donato et al. and Kent et al. used acetonitrile precipitation in conjunction with a deuterated internal standard. The supernatant is then injected directly onto the LC-MS/MS system for quantification.^33,34 This method is rapid and efficient and does not require specialist equipment for solvent evaporation unlike other published methods, making it an attractive option for a routine laboratory. This method may, however, be unsuitable with less sensitive mass spectrometers. Camilleri et al. ³¹ used acetonitrile and ammonium sulphate precipitation followed by incubation and separation, before evaporation under nitrogen and reconstitution. Gothe et al. ³⁰ utilized a HPLC method; however, this method required several washing steps on octadecylsilane-bonded silica columns, using hexane-chloroform for elution making it unsuitable for routine use.

The major benefit of C4 measurement is that it requires a single blood sample, rather than a long stool collection or two visits to hospital for the SeHCAT test, making it more convenient for patients. BA synthesis and therefore C4 levels, have diurnal variation and increase post-prandially, thus a fasting morning sample is preferred. ³⁴ Gothe’s study demonstrated that children with CD and persistent diarrhoea had significantly higher serum C4 concentrations compared to those with formed stools. Furthermore, C4 concentrations were increased in patients with ileal resection compared to those with intact ileum. ³⁰ Serum C4 concentration had 90% sensitivity and 79% specificity, respectively, in the diagnosis of BAD when SeHCAT testing was used as the gold standard. ³¹ C4 has also been shown to be inversely related to SeHCAT retention, ³⁵ and several studies have shown its correlation with FBA loss.^5,7 Owing to its high negative predictive value, C4 has been proposed as a rule-out test for BAD, reducing the number of patients requiring referral for SeHCAT testing. ³⁶ However, research is still limited owing to the small number of centres that measure the analyte.

Serum fibroblast growth factor 19 measurement

FGF19 inhibits BA synthesis. In BAD, loss of FBA and reduced ileal BA concentrations decrease FGF19 which leads to increased hepatic BA synthesis. Several commercial enzyme-linked immunosorbent assay (ELISA) kits are available for measurement of FGF19 in serum and plasma. The kits generally use several steps, including multiple washing and incubation steps, and thus the assay may be time consuming in a routine laboratory environment.

Vijayvargiya et al. ³⁶ reported FGF19 to have a negative predictive value of 78% and specificity of 78% for BAD when using 48-h FBA measurement as the gold standard. Pattni et al. ³⁵ demonstrated a negative predictive value and positive predictive value of 82% and 61%, respectively, when utilizing an FGF19 cut-off of ≤145 pg/mL, for predicting a SeHCAT of <10%. A positive correlation between FGF19 values and SeHCAT retention was also identified. Thus, like serum C4, FGF19 may have a role in excluding BAD and identifying patients for SeHCAT.

An advantage of ELISA kit availability is that FGF19 assays can be standardized across sites, which may be more difficult with in-house assays such as LC-MS/MS methods for FBA and serum C4, although ELISA lot to lot variability can present problems. Unlike mass spectrometry-based assays, ELISA kits are relatively easy to implement without large corresponding start-up costs, although cost per test may be higher.

¹⁴C-glycocholate breath and stool test

The ¹⁴C-glycocholate breath and stool test measures increased BA deconjugation in the gastrointestinal tract, either as a result of small bowel bacterial overgrowth or due to increased BA presence in the colon. ³⁷ The test involves ¹⁴C-glycocholate ingestion with a standard meal. Exhaled air is collected every hour for 6 h and stool is collected for 24 h. In health, orally administered ¹⁴C-glycocholate is largely reabsorbed in the ileum. In bacterial overgrowth or increased colonic BA due to BAD, bacterial action releases ¹⁴C-glycine from ¹⁴C-glycocholate. ¹⁴C-glycine enters the portal circulation and is metabolized by the liver to ¹⁴CO_2, which is measured in exhaled breath, ²⁴ and ¹⁴C output can be measured in the stool. ³⁸ As such, ¹⁴C-glycocholate has been proposed as an indirect marker BAD. It also has the potential to identify the cause of malabsorption, as exhaled ¹⁴CO₂ peaks will be observed sooner after ingestion with small bowel bacterial overgrowth, with a normal faecal ¹⁴C output. Conversely, if the ¹⁴C-glycocholate is not reabsorbed in the terminal ileum and enters the large intestine, the ¹⁴C-glycocholate will be deconjugated by colonic bacteria, increasing exhaled ¹⁴CO₂ but also increasing faecal ¹⁴C output. However, altered bowel transit time in certain GI conditions means ¹⁴C-glycocholate cannot completely differentiate small bowel bacterial overgrowth from other causes of BAD. ²⁴ Additionally, the complexity and laborious nature of the investigation has made this test of historical interest.

Urine-2-proponol

Urine-2-proponol is produced during bacterial cleavage of BAs in the gastrointestinal tract 7 and is an indirect marker of BA metabolism. Studies have utilized an ‘electronic nose’ system which detects volatile organic compounds, or field asymmetric ion mobility spectrometry (FAIMS) which uses differences in ion mobility to separate chemical compounds, ³⁹ to investigate urine gas profiles. Covington et al. ³⁹ demonstrated that patients with SeHCAT confirmed BAD had a different urine gas profile to both healthy individuals and patients with ulcerative colitis, and that gas-chromatography mass spectrometry analysis of urine identified prominent peaks of 2-proponol and acetamide in the BAD cohort only. This therefore provides a potentially non-invasive testing method, however requires more investigation to determine its utility in diagnosis.

Summary of diagnostic methods

There is currently insufficient evidence to advise on the most appropriate method for diagnosing BAD, taking into account accuracy, cost and feasibility. SeHCAT testing is recommended by NICE, and no other methods are widely available in the UK. However, the British Society for Gastroenterology have recently recommended that a diagnosis of BAD should be made either through SeHCAT testing or serum C4 measurement, and if neither of these are available, then 48-h FBA measurement may be used. ²⁹ The guidelines, notably, do not recommend a diagnostic therapeutic trial of bile acid sequestrants, FGF19 measurement or urine-2-propanol measurement, nor do they mention random FBA measurement.

Treatment of BAD

The primary treatment of BAD is bile acid sequestrants, ⁴⁰ anion exchange resins which bind BAs in the intestine with high affinity, forming an insoluble complex that is excreted in the faeces, preventing re-uptake into the enterohepatic circulation. ⁴¹ There are three drugs in use; Cholestyramine, Colestipol and Colesevelam. The first-line treatment is often Cholestyramine; however, variable results using this treatment have been reported. A systematic review identified 23 studies involving 801 patients that used Cholestyramine alone as first-line treatment. Of these, 70% reported a good clinical response, 16% found the therapy ineffective, 11% could not tolerate the therapy and the remaining 3% were either non-compliant, lost to follow-up or the reason for treatment failure was unclear. ⁴² The same review found that 9% of 90 patients were unable to tolerate Colesevelam, a newer drug which demonstrates greater affinity for BAs than Cholestyramine and Colestipol. It is important to note that these studies are often subjective, with patients having different tolerance levels and different thresholds for defining symptom improvement. There is evidence that there is a dose–response relationship to Cholestyramine and patients with more severe BAD appear more likely to respond to treatment, ²³ although the endpoints in this study were not standardized. The few studies evaluating the effectiveness of Colestipol and Colesevelam have indicated therapeutic benefit, ⁴² but these are not currently licensed for the treatment of BAD in the UK.

Side effects of bile acid sequestrants include flatulence, constipation, bloating, nausea and an unpleasant taste. ⁴⁰ These may reduce patient compliance, which may contribute to the reported level of ineffectiveness, making it difficult to fully assess the efficacy of treatment.

Studies have also reported an improvement in patient symptoms when commencing a low-fat diet. ⁴³ A low-fat diet reduces BA production and their release into the bowel following a meal, thus reducing the intestinal BA concentrations. Patients have reported significant improvements in bowel urgency and frequency, along with bloating and abdominal pain, ⁴³ after commencing a low fat diet. Low fat diets, however, may be unpalatable. Dietary intervention should therefore be commenced where possible alongside BA sequestrants to maximize therapeutic benefit to patients.

Other therapies including anti-diarrhoeals may be useful in those intolerant to bile acid sequestrants. An emerging therapy is obeticholic acid, a potent FXR agonist, which stimulates ileal production of FGF19 and thereby inhibits hepatic BA synthesis and secretion, and has shown significant promise in the management of BAD. ⁴⁴

Conclusion

There is a clear need for further research and development of easily available diagnostic methods for BAD, and strategies to optimize patient, clinical and laboratory convenience. Effective and easily accessible methods of diagnosis should reduce missed diagnosis, increase initiation of appropriate treatment and improve patient quality of life.

Determining correlation of FBA concentration between a 48-h stool sample, and a random sample, will be important in deciding whether kit-based FBA assays may be used to diagnose BAD, as a 48-h stool collection is likely to be unacceptable to patients. Urine-2-propanol offers a promising non-invasive option, but requires a significant amount of further research. Although recommended by the BSG, utility of serum C4 is most likely as a rule out test and identifying patients for referral for SeHCAT testing for diagnosis of BAD. It, however, is not widely available in the UK and further research is needed to identify an appropriate cut-off point for the UK population.

In conclusion, there are several promising methods for the diagnosis of BAD; all require further evaluation in the field before consideration as a replacement for SeHCAT. This should be a priority, owing to the significant under-diagnosis of patients, the availability of relatively simple and reasonably effective treatment for BAD and the potential for improvement in quality of life in these patients.