A single faecal bile acid stool test demonstrates potential efficacy in replacing SeHCAT testing for bile acid diarrhoea in selected patients

A systematic review looking at clinical utility of tests for bile acid diarrhoea demonstrated an average reported sensitivity and specificity of SeHCAT to be 87.3% and 93.2%8. However, large variation is observed based on the cut off values for this test. A cut off of 8% gives a sensitivity and specificity of 67% and 97% respectively8, whilst a cut off value of 15% gives a sensitivity and specificity of 100% and 91% respectively11, and a cut off of 5% gives a result of 86% and 100% respectively8. Currently there are no clinical validity results for SeHCAT and the diagnostic accuracy has only been determined by treatment response with bile acid sequestrants10. As stated previously, clinical response to treatment is not an accurate measurement of bile acid diarrhoea as medication compliance and adherence is a serious concern for patients who cannot tolerate the medications1,6,10. Whilst the SeHCAT scan is the current gold standard for diagnosing bile acid diarrhoea in the UK, it has several disadvantages including radiation exposure, limited availability and time restraints for the patient who is required to come into hospital twice which may be difficult if they are working and/or are self-employed, have child-care responsibilities or have limited access to transportation. Jackson et al. further demonstrated that although a SeHCAT scan result can indicate the presence of malabsorption, it may not predict the variety or severity of patients’ symptoms as these are dependent on other factors, such as the amount of dietary fat intake12. Additionally, the SeHCAT test has limited accuracy at the indeterminate levels, with ill-defined normal ranges and thus likely variability between sampling. Therefore, there is a need for a test that is more easily accessible but still reliable and dependable in its results. In our study, we attempted to demonstrate efficacy with a convenient single, random stool sample that would be able to correctly identify bile acid diarrhoea and be the least disruptive to a patient’s lifestyle. Whilst our study demonstrated efficacy in patients with severe bile acid diarrhoea and/or in post-operative Crohn’s disease patients, there were several limitations to this test which we discuss below in greater detail.

In this study, faecal bile acid concentrations were significantly increased in post-terminal ileal resected Crohn’s patients with confirmed bile acid diarrhoea, compared to the other three groups. This is to be expected, given that in these patients, the primary location for bile acid reabsorption is removed. The results also demonstrated a significant difference in faecal bile acid results in patients with SeHCAT retention of < 5% compared to SeHCAT retention of > 15%, suggesting utility in diagnosing severe bile acid diarrhoea. There was, however, no significant difference in FBA concentrations in patients with idiopathic bile acid diarrhoea, post-cholecystectomy, or control group, suggesting that faecal bile acid measurement cannot reliably distinguish idiopathic bile acid diarrhoea or post-cholecystectomy patients from the control group cohort. On closer inspection, however, the post-operative Crohn’s cohort all had severe bile acid diarrhoea (< 5%) while 26/33 (78.8%) patients of idiopathic bile acid diarrhoea and 13/20 (65%) patients with post-cholecystectomy had a mixture of mild and moderate bile acid diarrhoea severity. The differences in severity grade between the groups may explain why there was such variability in the faecal bile acid concentrations and larger studies may help to expand on these findings.

The ROC curve analysis demonstrated that faecal bile acid analysis provided some utility in the diagnosis of bile acid diarrhoea, as the AUC was > 0.5, however no cut off provided an acceptable balance of sensitivity and specificity. It was therefore considered whether faecal bile acid measurement may be used to stratify patients requiring SeHCAT. A faecal bile acid concentration of 10.1 μmol/g demonstrated 91% specificity for severe, moderate or mild bile acid diarrhoea, therefore it may be that results above this value could be treated as bile acid diarrhoea, without requirement for SeHCAT referral. This may be of particular use in diagnosis of Type 1 bile acid diarrhoea in patients in whom there is a high diagnostic likelihood, such as those in our post-terminal ileal resected Crohn’s disease group. However, at this cut-off, in our cohort, there would have been 50/113 (44.2%) false negatives and therefore all patients with a result below 10.1 μmol/g would still require a SeHCAT referral; this would have been 90/113 (80%) patients in total, and therefore would not significantly reduce SeHCAT referral burden. The sensitivity and specificity achieved in this study using a faecal bile acid cut off of 1.6 μmol/g (using ≤ 15% SeHCAT retention as diagnostic of bile acid diarrhoea) were 90% and 25% respectively. At this cut off, there would have been 33 false positive results and 7 false negative results, out of 113 patients. As a result, 33 patients would have been unnecessarily treated with medication, which on top of everything else has its own cost implications.

The difficulty in achieving an acceptable balance of sensitivity and specificity may partly be due to the variability observed between individual stool samples, and as stated above, measuring faecal bile acids in a single random stool sample may result in a missed diagnosis. Camilleri et al. demonstrated that the faecal bile acid excretion in a random stool sample was closely correlated with the mean excretion in all samples from individual patients13. However, due to the variation in bile acid excretion per gram faecal weight in each bowel movement, and in the faecal pellet and supernatant of three of their patient samples, they concluded that individual stool samples were not representative of the total faecal bile acid excretion. Thus, more research may be required to determine the optimum stool sample for analysis, balancing both variation in excretion, and patient acceptability. A more recent paper suggested that a random stool sample could be used in conjunction with serum C4 for predicting bile acid diarrhoea14; however, this study used the random stool sample to assess percentage of primary bile acids, using a cut-off of > 10% primary bile acids in the sample, rather than total faecal bile acids. The Immunodiagnostik faecal bile acid kit being assessed in this study measures total faecal bile acids, not individual bile acids, and thus may not have the same utility.

The 48-h total faecal bile acid collection is the diagnostic method for bile acid diarrhoea in the United States where the 75SeHCAT scan is not licensed for use15. The diagnostic accuracy of this test has demonstrated an average sensitivity and specificity of 66.6% and 79.3% respectively. However, there are multiple problems with this method. The processes used to extract bile acids from faeces are complex and generally restricted to research laboratories. Furthermore, patients need to provide a 48-h stool sample and are required to undertake a 4-day strict 100 g fat-intake diet, making it inconvenient and unpleasant for the patient16. This requirement is due to the hypothesis that there is variable excretion in bile acids throughout the day in different stool samples, thus a random stool sample may result in a missed diagnosis17. The Immunodiagnostik assay uses a random stool sample, improving convenience for both patient and laboratory staff. However, use of a random sample may not provide a representative view of total faecal bile acid excretion, which may have contributed to the lack of significant results observed in the post-cholecystectomy and idiopathic bile acid diarrhoea groups, and also to the poor correlation observed with the SeHCAT scan results. In comparison, Sagar et al.18 found the median total faecal bile acids in bile acid diarrhoea to be higher than our results (9.17 (IQR 7.79–14.12) vs 4.8 (IQR 2.5–10.0) respectively). Whilst they used a random stool sample to analyse their results, their protocol differed from ours by using high performance liquid chromatography coupled to tandem mass spectrometry (HPLC–MS/MS). Their results also showed a greater concentration of secondary faecal bile acids compared to primary faecal bile acids (7.13 vs 1.5, respectively). The Immunodiagnostik assay is marketed as measuring total faecal bile acids, and consequently we were unable to separate primary and secondary faecal bile acids. Thus, our results are consistent with previous publications that a random stool sample may not provide sufficient information on faecal bile acid excretion19,20.

Measurement of faecal bile acids before and after treatment was performed to assess whether faecal bile acid measurement could assess therapeutic concordance and response. No significant difference was seen pre- and post-treatment, regardless of clinical improvement of symptoms. Camilleri et al.19 investigated the effect of colesevelam on faecal bile acids and found that post-treatment, there was an increase in deoxycholic acid with a reduction in cholic acid. There were not any differences, however, in the proportion of chenodeoxycholic acid, lithocolic acid or ursodeoxycholic acid. Their methodology also used a separate extraction technique using methanol which allowed them to extract bile acids that were both free and colesevelam-bound. As our assay did not have this critical methanol extraction step and only measures total faecal bile acids, it is likely that sequestered bile acids that are excreted are also measured in the assay, and therefore this method would not be appropriate for monitoring treatment efficacy.

Due to the global limited availability of the SeHCAT scan and the inconvenience of collecting a 48-h faecal bile acid sample, other methods to diagnose bile acid diarrhoea are currently being investigated, including the use of C4 and FGF-19 biomarkers. Vijayvargiya et al. recently demonstrated high negative predictive values and specificities for C4 (79%/83%) and FGF-19 (78%/78%) but found that combining the biomarkers, the sensitivity increased to 50% and the specificity was 65%7. Interestingly, their results improved when they excluded the cholecystectomy patients from their study. In our random stool samples, the post-cholecystectomy group did not show much variation in median faecal bile acid concentrations compared to the idiopathic bile acid diarrhoea cohort but it would be interesting to correlate these findings with C4 results.

The main strengths of our study include comparing our results against a control group and directly with the SeHCAT scan, the current gold standard diagnostic test for bile acid diarrhoea. As mentioned above, the main limitations of this study are the inability to differentiate between total and individual bile acids and from colesevelam-bound and free faecal bile acids with our assay. The use of a random stool sample may explain some of the negative results obtained, however this in itself is not considered a limitation, as the purpose of the study is to investigate whether the Immunodiagnostik kit could be used with a random stool sample for diagnosis of bile acid diarrhoea. Our study also did not control for other possible variables that could have affected our results, including dietary intake and definitively excluding other malabsorption conditions. Studies have shown that a high-fibre and low-fat diet results in lower faecal bile acid concentrations21,22. However, as patients undergoing SeHCAT scans are not advised to undergo any dietary restrictions before or during the testing time, we wanted to maintain the same guidance for consistency. As this was the first study to utilise this kit to diagnose bile acid diarrhoea, we appreciate that future larger studies exploring dietary modifications to changes in faecal bile acid concentrations would be beneficial. Faecal bile acid concentrations can also be elevated in the presence of small intestinal bacterial overgrowth (SIBO) and pancreatic insufficiency23. Whilst in an ideal world, all patients should have all of these conditions excluded, this is not practical in the real world and moreover, falls outside of the national BSG guidance on management of chronic diarrhoea1. Further investigations for chronic diarrhoea were left to the discretion of the patients’ standard care physician and was outside of the trial remit.

In this small study, we were able to demonstrate the potential efficacy of using faecal bile acids in diagnosing severe bile acid diarrhoea and/or in patients with post-operative Crohn’s disease, without the need for SeHCAT testing. This can prove to be beneficial in patients who do not have access to SeHCAT testing but are also unable to tolerate the bile acid sequestrant when given an empirical trial of treatment. Considering that patients with severe bile acid diarrhoea have a higher response rate to treatment, confirming their diagnosis via a single random faecal bile acid test could motivate them to persevere with their medication, with the goal to improve their underlying symptoms and quality of life. Whilst the correlation between the SeHCAT retention and faecal bile acid excretion is not strong enough to replace the utilisation of SeHCAT testing in our other patient groups, our calculations deduced that our study was not adequately powered. Thus, larger studies are now required to confirm the potential efficacy of using a single random faecal bile acid test to accurately diagnose bile acid diarrhoea in these patients in the absence of SeHCAT testing, establish a normal range, and determine whether a cut off can be set at which an acceptable sensitivity and specificity may be achieved.

Source link

Back to top button