Background & Aims
Some studies have found that patients with idiopathic bile acid diarrhea (BAD) present with symptoms of diarrhea-predominant irritable bowel syndrome (D-IBS). However, these studies either were retrospective, did not define D-IBS according to current criteria, or included pa- tients with chronic functional diarrhea. We performed a prospective study of the prevalence of idiopathic BAD in consecutive patients fulfilling the Rome III criteria for D-IBS.
Authors
Methods
Results
Twenty-eight patients were found to have BAD (23.7% of total), with similar percentages at each study site (25.3% and 20%; P </=5%). Eigh
t patients had mild BAD (28.6%), 8 patients had moderate BAD (28.6%), and 12 patients had severe BAD (42.8%). There was no statistical difference in age, sex, depression, patient health questionnaire-12 responses, or SF-36 scores between individuals with vs without BAD. However, patients with BAD had a higher mean body mass index than patients without BAD (31.6 vs 26.4; P [ .003). Physical activity (based on the Short Form-36) was significantly lower in subjects with moderate (43.8) or severe BAD (41.7), compared with patients with mild BAD (87.5) (P [ .046).
Conclusions
Abbreviations used in this paper:
Guidelines for the management of irritable bowel syndrome (IBS) recommend the use of symptom-based criteria to aid clinicians in making a positivediagnosis without the need for performing extensive investigations.1 In 1978, Manning et al2 first described 6 key symptoms commonly seen in individuals who ultimately were diagnosed with IBS. These later were incorporated by a multinational working party to form the Rome I criteria,3 and subsequently have been revised on 2 occasions to produce the Rome II and, most recently, the Rome III criteria.4,5
However, validation of such criteria shows that they perform only modestly in distinguishing IBS from organic diseases.6 Several studies have reported that gastrointestinal symptoms in organic conditions, such as celiac disease,7 exocrine pancreatic insufficiency,8 inflammatory bowel disease,9 and small-bowel bacterial overgrowth,10 significantly overlap and mimic those of diarrhea-predominant IBS (D-IBS). Most recently, interest has focused on the potential role of bile acids in evoking gastrointestinal symptoms compatible with D-IBS.11
Bile acids are synthesized in the liver, stored in the gallbladder, and released into the small bowel, where they aid in the digestion of lipids. Normally, 95% of bile acids are re-absorbed in the terminal ileum (TI) and are recycled back to the liver via the enterohepatic circula- tion. If there is a failure of re-absorption, excess bile acids escape into the colon where they stimulate electrolyte and water secretion, resulting in chronic watery diarrhea.11 The conditions causing bile acid diarrhea (BAD) can be classified according to the etiology. Type 1 represents failure to absorb bile acids from the TI as a result of either resection or localized disease (ie, Crohn’s disease). Type 3 incorporates miscellaneous causes such as previous cholecystectomy, fibrosis after abdominal or pelvic radiotherapy, celiac disease, or microscopic colitis. Type 2, otherwise known as idiopathic BAD, occurs when there is no anatomic abnormality or other risk factor apparent.11 In the latter case, evidence suggests that TI absorption of bile acids is normal or increased,12 but impaired negative feedback of bile acids on fibroblast growth factor-19 leads to dysregulation of the enter- ohepatic circulation and excessive bile acid synthesis.13
A relatively simple and highly sensitive method of testing for BAD is via the 75selenium homocholic acid taurine (75SeHCAT) test, in which retention of radio-labeled bile acids of less than 10% to 15% after 7 days is abnormal.14,15 The 75SeHCAT retention test is available in the United Kingdom, certain European countries, and Canada. However, the test is not available in the United States, where alternate means of detecting BAD can be performed, albeit with difficulty and in selected laboratories only.16 Measurement of total stool bile acid can be used to diagnose BAD, but stool collections for 48 hours or longer are required to account for variations, which can be demanding and unpopular for both patients and laboratory staff. There have been recent advances in the applicability of measuring serum 7a-hydroxy-4- cholesten-3-one (C4), an intermediary in bile acid synthesis. Increased C4 levels represent a surrogate marker for increased bile acid synthesis and increased fecal bile acid loss. The disadvantages of serum C4 are that it requires a fasting sample, can lead to false-positive results in patients with liver disease, or in patients taking medication that alters bile acid synthesis such as statins, or in individuals with an altered circadian rhythm.16 Direct comparisons between the 75SeHCAT retention test and serum C4 show that low 75SeHCAT values correlate more closely with altered bowel habit,suggesting it is the better method to use for detecting BAD.17 However, in the current climate, because none of these tests routinely are available in countries such as the United States, there is interest in the development of a simple, cheap, and readily available biomarker that will help screen for patients who potentially have BAD and who then subsequently can be referred for confirmatory tests. Of these, measurement of serum fibroblast growth factor-19,18 identification of the primary bile acids cholate and chenodeoxycholate in a single stool sample,19,20 or detection of volatile organic compounds in a single urine sample (presumed to be a consequence of dysbiosis secondary to increased levels of colonic bile acids),21 all have shown initial promising results, but require further work to validate their clinical applicability.22
A systematic review and meta-analysis suggested that idiopathic BAD may be the underlying cause of symptoms in 30% of patients with presumed D-IBS.23 However, many of the included studies either were retrospective in nature, did not use symptom-based criteria to define D-IBS, or included patients with chronic functional diarrhea. Nevertheless, a recent prospective evaluation showed that 23% of patients with D-IBS have a 75SeHCAT retention test of less than 10%, and that colonic bile acid exposure correlates with bowel habit and colonic transit time.17 Furthermore, in patients with BAD, open-label treatment with bile acid sequestrants led to a positive clinical response, thereby supporting the role of bile acids in IBS symptomatology.17 However, this novel study used the Rome II criteria, which can show differences in agreement when subtyping against the current Rome III criteria.24 We now report our experience of testing for idiopathic BAD, using the 75SeHCAT retention test, in a rigorously defined cohort of patients meeting the Rome III criteria for IBS, who otherwise would be labeled as having D-IBS.
Materials and Methods
Participants and Setting
This prospective dual-center study was performed at the Royal Hallamshire Hospital (Sheffield, South York- shire, UK) and St. James’s University Hospital (Leeds, West Yorkshire, UK). Both are located in northern England and provide secondary care services to a local population of 500,000 and 800,000 people, respectively. On average, each of the general luminal gastroenterology clinics that took part in this study sees 200 new out- patients per gastroenterologist per year, with diarrhea and abdominal pain being present in more than a quarter of all referrals and accounting for the major workload.25 After ethical approval from local research committees, the study was conducted between January 2013 and April 2014.
The inclusion criteria were adults who fulfilled the Rome III criteria for D-IBS (Supplementary Table 1).5
These patients had reported similar symptoms when initially referred by their primary care physician several weeks earlier, suggesting that they were relatively stable in terms of their IBS subtype.26 We excluded any subjects with known risk factors for BAD, such as a previous history of cholecystectomy, TI Crohn’s disease, TI resection or right hemicolectomy, celiac disease, abdominal or pelvic radiotherapy, or microscopic colitis. The latter was via normal colonic biopsies in 84.7% of the cohort, and clinical judgement in patients deemed at low risk or who refused a colonoscopy.
Baseline Questionnaires and 75Selenium Homocholic Acid Taurine Retention Test
Information concerning baseline questionnaire data is provided in the Supplementary Materials and Methods section. The 75SeHCAT retention test was perfor- med according to the manufacturer’s instructions (GE Healthcare, Buckinghamshire, UK). A 370-KBq capsule of 75SeHCAT was administered with water on day 1. Patients underwent a scan of their abdomen using a gamma camera to obtain baseline counts at 3 hours. A further scan of the abdomen then was performed on day 7 to determine the percentage of 75SeHCAT retention. We used a retention rate of less than 15% to define the presence of BAD in our primary analysis, but we used a lower threshold of less than 10% in a secondary analysis. In our primary analysis, the degree of BAD was classified further as severe if retention was 5% or less, as mod- erate if retention was 5.1% to 9.9%, and as mild if retention was 10% to 14.9%.
Data Analysis
Statistical analysis was performed using SPSS soft- ware (version 19.0; SPSS, Inc, Chicago, IL). Categoric variables were summarized by descriptive statistics, including total numbers and percentages, and compared between patients with and without BAD using the Pearson chi-squared test. Continuous variables were summarized by means and standard deviations, and compared between patients with and without BAD using the Student t test, and across the 3 categories of BAD severity using a 1-way analysis of variance. Statistical significance was set at a P value of less than .05.
Results
A total of 127 consecutive individuals meeting the Rome III criteria for D-IBS were approached, and 118 (92.9%) participated in the study: 83 individuals from Sheffield and 35 individuals from Leeds. Combined data analysis from both sites showed the mean age of D-IBS patients to be 41.7 years, with 72.9% (n 1⁄4 86) women, 91.5% (n 1⁄4 108) Caucasian, with a mean body mass index (BMI) of 27.6. The mean scores for assessments of well-being were as follows:
Hospital Anxiety and Depression Scale–anxiety subscale (HADS-A) was 8.4, HADS–depression subscale (HADS-D) was 5.03, Patient Health Questionnaire (PHQ)-12 was 6.94, Short Form-36 (SF-36) physical component score was 56.1, and SF-36 mental component score was 56.7. Subgroup analyses of D-IBS patients, between both recruitment sites, showed no statistical differences in basic demographics, HADS-A, HADS-D, PHQ-12 somatization, or SF-36 quality- of-life scores.
The total prevalence of BAD, using a 75SeHCAT score of less than 15%, was 23.7% (n 1⁄4 28). The prevalence was statistically similar across both recruitment sites, at 25.3% (n 1⁄4 21 of 83) in Sheffield and 20.0% (n 1⁄4 7 of 35) in Leeds (P 1⁄4 .54). Had we used a lower 75SeHCAT retention test cut-off value of less than 10% to diagnose BAD, then the total prevalence would have been 16.9% (n 1⁄4 20), with 18.1% (n 1⁄4 15) at Sheffield and 14.3% (n1⁄45)atLeeds(P1⁄4.62).
The characteristics of patients diagnosed with BAD, compared with those without, are reported in Table 1. Individuals with BAD were significantly more likely to have a higher mean BMI than patients without BAD (31.6 vs 26.4; P 1⁄4 .003). However, other demographic variables or assessments of well-being such as HADS, PHQ-12, or SF-36 did not differ between patients with or without BAD.
The breakdown of the 28 patients diagnosed with BAD showed that 8 (28.6%) had mild BAD, 8 (28.6%) had moderate BAD, and 12 (42.8%) had severe BAD. The characteristics of individuals with BAD according to each of the 3 severities are reported in Table 2. Interestingly, subjects with severe/moderate BAD reported significant role limitations owing to their physical health compared with patients with mild BAD. However, other variables assessing physical well-being did not differ according to the severity of BAD, and there were no difference in the overall SF-36 physical component score. Finally, there were no differences in baseline demographics, PHQ-12, HADS-A, HADS-D, or SF-36 mental component score detected between the various BAD severities.
Discussion
This study has shown that almost 1 in 4 individuals presenting for secondary care who fulfill Rome III criteria for D-IBS have evidence of idiopathic BAD. Furthermore, in nearly 70% of cases the severity of BAD was within the moderate to severe category, and was associated with significant physical role limitations, compared with patients with only mild BAD. We also found that individuals with BAD had a significantly higher mean BMI compared with patients without BAD. These findings have important implications for the investigation and management of this group of patients.
Strengths of our study include the methodology used. Several previous studies have claimed an association between symptoms that are compatible with D-IBS and idiopathic BAD, yet have been limited by their retrospective design, choice of patients recruited, or use of non–gold standard criteria to define IBS. In contrast, we evaluated the prevalence of idiopathic BAD in a prospective cohort of consecutive patients fulfilling the Rome III criteria for D-IBS. Furthermore, we also ensured that we did not overestimate the prevalence of BAD by excluding any individuals who had risk factors associated with type 1 or type 3 BAD. The high participation rate also was a strength, and may reflect the fact that 75SeHCAT testing was accessible only via this study, and patients were keen to explore this as a potential cause for their symptoms. Finally, we showed similar results at 2 independent centers, in secondary care populations. We therefore believe that, having used state-of-the-art definitions and testing modalities, our findings are contemporary and provide compelling evidence that can be generalized to other centers seeing a similar population of patients. In fact, a recent study from the United States identified that 38% of patients with D-IBS were noted to have higher serum C4 and fecal bile acid levels compared with healthy volunteers or patients with constipation-predominant IBS,27 suggesting that, if 75SeHCAT testing were available in the United States, the yield in patients with D-IBS likely would be similar to ours.
The main limitation of our study was that, despite finding a high prevalence of idiopathic BAD in patients fulfilling the Rome III criteria for D-IBS, we did not assess the clinical response to bile acid sequestrants. Nevertheless, other studies have shown a beneficial response to open treatment, particularly in patients with moderate to severe BAD.17,23 A systematic review noted that a response to cholestyramine occurred in 96% of patients with a 75SeHCAT retention rate of less than 5%, in 80% with a 75SeHCAT retention rate of less than 10%, and in 70% with a 75SeHCAT retention rate of less than 15%.23 In our study, of patients who had BAD, 28.6% were mild, 28.6% were moderate, and 42.8% were severe. This suggests that making a positive diagnosis will be of clinical benefit, and serves as a mandate for randomized placebo-controlled trials of bile acid sequestrants in D-IBS. Another limitation was that we did not assess symptom severity and its relationship with BAD, although a recent study suggested that perceived symptom severity is not linked to 75SeHCAT retention values.17 Nevertheless, because a low 75SeHCAT retention test value is characterized by having more frequent stools and accelerated colonic transit time, future studies in patients presenting with D-IBS–type symptoms may wish to explore whether quantifying stool frequency or consistency can be used as a predictive tool to diagnose idiopathic BAD.
The study findings have both major clinical and research implications because they call into question whether symptom-based criteria can be used as a tool to positively diagnose D-IBS without the need to exclude possible underlying organic pathology. Such uncertainties already are apparent in clinical practice in which health care professionals vary greatly in their beliefs as to whether IBS is a diagnosis of exclusion.28 A survey conducted in the United States among primary care physicians, community gastroenterologists, and nurse practitioners (collectively referred to as non-IBS experts) compared with an internationally recognized panel of IBS key opinion leaders, showed that nonexperts were far more likely to endorse IBS as a diagnosis of exclusion (72% vs 8%, respectively). As a result, non- experts showed an increased utilization of resources, particularly in patients presenting with D-IBS.28 This highlights that, despite decades of expert opinion, clinicians remain concerned about the possibility of a missed organic pathology in patients meeting criteria for IBS. This is reinforced by the fact that celiac disease may account for up to 4% of IBS cases,7 which subsequently led to a paradigm shift in international guidelines.1 Similarly, studies elsewhere have shown D-IBS to be associated with other organic pathologies, such as exocrine pancreatic insufficiency, inflammatory bowel disease, and small-bowel bacterial overgrowth.8–10
Our findings only serve to heighten this concern further by providing clear evidence that idiopathic BAD can masquerade as D-IBS. It is interesting to note that the magnitude of this problem appears far greater than that of celiac disease in suspected IBS. This may be explained by changing referral patterns, with primary care physicians in the United Kingdom adhering to recent National Institute of Health and Care Excellence guidelines that advocate excluding celiac disease, via serologic testing, in patients with suspected IBS. Patients who test positive are then likely to be referred directly for endoscopy and duodenal biopsies to confirm the diagnosis, whereas patients with negative serology, and therefore suspected IBS, will be referred to secondary care for further evaluation and treatment of their symptoms.29,30 Clearly, patients with positive celiac serology were excluded from this study, but during a similar period of time among 240 consecutive patients referred with chronic diarrhea to the same general luminal gastroenterology clinics, and who were ineligible for the study, because they did not meet the Rome III criteria for IBS, 25 (10.4%) were found to have BAD, compared with only 10 (4.2%) with celiac disease.31,32 This suggests that secondary care physicians now are posed with different diagnostic challenges when evaluating patients with suspected D-IBS or chronic diarrhea.
In conclusion, almost 1 in 4 cases presenting with D-IBS will have idiopathic BAD. In the majority, this will be moderate to severe. This suggests that future guide- lines should advocate diagnostic testing to exclude BAD before a diagnosis of D-IBS is made.
Supplementary Material
Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at www.cghjournal.org, and at http://dx.doi.org/10.1016/j.cgh.2015.03.002
References
