Yuthana Kong

The effect of a nonabsorbed oral antibiotic (rifaximin) on the symptoms of the irritable bowel syndrome: a randomized trial.

Context Few trials have evaluated the effects of antibiotics on symptoms of the irritable bowel syndrome (IBS). Contribution In this double-blind trial, 87 patients with IBS were randomly assigned to either rifaximin (400 mg 3 times daily) or placebo for 10 days. Over a 10-week follow-up period, the rifaximin recipients reported global improvements in overall symptoms and less bloating more frequently than the placebo recipients. No major differences in abdominal pain, diarrhea, or constipation were observed between the groups. Cautions The duration of therapy and follow-up was short. Implications Rifaximin may improve some symptoms in some patients with IBS. The Editors The irritable bowel syndrome (IBS) is one of the most common chronic medical conditions (16), yet its cause is unknown. Among other contributors, alterations in gut flora have been identified as potentially important. Results of recent studies indicate that up to 84% of patients with IBS have an abnormal lactulose breath test result, suggesting small-intestinal bacterial overgrowth (7, 8). On the basis of this concept, the antibiotic neomycin can statistically significantly improve the symptoms of IBS (7, 8). In addition, the effect of neomycin correlates with the elimination of bacterial overgrowth, as indicated by the normalization of the lactulose breath test result (7, 8). Although neomycin seems to improve symptoms, it effectively eliminates bacterial overgrowth in only about 25% of patients with IBS (8). Furthermore, side effects limit the use of neomycin. Low efficacy also applies to other antibiotics (for example, doxycycline and amoxicillinclavulanate) that have been previously investigated for treating bacterial overgrowth (9). An ideal antibiotic for IBS is, arguably, one with negligible systemic absorption, minimal side effects, and high efficacy for bacterial overgrowth. Rifaximin is a gut-selective antibiotic with negligible systemic absorption (0.4%) and broad-spectrum activity in vitro against gram-positive and gram-negative aerobes and anaerobes (10). On the basis of this broad spectrum, eradication rates with rifaximin in bacterial overgrowth are as high as 70% (11). Furthermore, rifaximin has a similar tolerability profile to that of placebo and has known activity against Clostridium difficile (12). These properties make it a good candidate for treating a condition that is as common as IBS. Our study aimed to determine whether the nonabsorbed antibiotic rifaximin is more effective than placebo in reducing symptoms in adults with IBS. Methods Setting and Participants Our study was conducted at the Cedars-Sinai Medical Center, Los Angeles, California, and the University of Chicago, Chicago, Illinois. We recruited patients with IBS through advertising in local media (radio and news publications). We did not recruit patients from the IBS clinics of the Cedars-Sinai Gastrointestinal Motility Program to avoid enrollment of tertiary care patients. The institutional review board of both centers approved the study, and all patients provided written informed consent. Patients between 18 and 65 years of age who met Rome I criteria (13) were eligible. Exclusion criteria were the presence of underlying conditions that are known to predispose to bacterial overgrowth, including diabetes; narcotic use; previous bowel resection; inflammatory bowel disease; cirrhosis; known bowel adhesions; or any known chronic gastroenterological disease, such as celiac disease. We excluded patients who were taking tegaserod and antidepressants unless these treatments were discontinued before study entry. We also excluded participants who reported taking an oral antibiotic within the previous 3 months. After participant inclusion and exclusion, we recruited 84 participants from the Cedars-Sinai Medical Center and 3 participants from the University of Chicago. We followed participants in special research clinics at both centers. Randomization and Interventions Eligible patients completed a 7-day stool diary that was based on the Bristol stool form scale (14). Patients returned to the clinic after a 12-hour fast and completed a symptom questionnaire about the preceding 7 days of symptoms. We then randomly assigned patients to double-blind treatment with 400 mg of rifaximin 3 times daily for 10 days or a matching placebo. We chose this dosage on the basis of a previous study that demonstrated the efficacy of rifaximin in bacterial overgrowth (11). The randomization of rifaximin versus placebo was conducted outside of Cedars-Sinai Medical Center in a 1:1 ratio into blocks of 4 patients. The allocation sequence was determined and coded at Salix Pharmaceuticals, Morrisville, North Carolina. Since this was an investigator-initiated study, the rifaximin and placebo were distributed to the Cedars-Sinai Medical Center, and nonstratified medicine and placebo were sent to the University of Chicago in groups of 4 as enrollment progressed. The medicine and placebo were prepackaged to conceal content at all times. Research personnel who were involved in product distribution were also blinded to package content. Assessments and Follow-up After completing the 10-day course of study medication, patients immediately began another stool diary for 7 days then returned to complete a follow-up questionnaire and to return their pill container for a pill count to determine adherence. Patients then entered the follow-up phase, during which they completed a weekly self-administered symptom questionnaire at home that documented their symptoms for an additional 9 weeks (for a total of 10 weeks of post-treatment follow-up). During this time, we asked participants to fax their responses to the research office. When a fax was not received on the appropriate day, research assistants called patients to ensure adherence. During this phase of study, no physician interaction occurred. During the last week of follow-up, patients completed a daily stool diary. At the end of the follow-up period, patients returned to the clinic for a final visit, which included another symptom questionnaire. For the initial symptom questionnaire, patients were asked to indicate the severity of each of 9 symptoms (abdominal pain, diarrhea, constipation, bloating, urgency, incomplete evacuation, mucus, sense of incomplete evacuation, and gas) on a visual analogue scale (VAS) ranging from 0 mm to 100 mm, with 100 mm being extreme. We used all 9 symptoms to verify IBS criteria in patients, but we assigned only diarrhea, constipation, abdominal pain, and bloating a priori as treatment end points. We asked patients to rate the severity of their symptoms on the VAS again 7 days after the completion of rifaximin treatment or placebo. Furthermore, we asked patients to provide a percentage of global improvement in their overall IBS symptoms from 0% to 100%. We chose global improvement since the Rome Consensus Group considers it to be the preferred end point measure in IBS treatment studies (15). Patients then rated the severity of their symptoms on the VAS and rated global improvement again each week for 8 weeks of follow-up and at the final visit to provide a total of 10 weeks of follow-up data. Table 1 depicts the number of patients with outcomes at various time points during the study. Table 1. Study Recruitment and Enrollment Summary At the first follow-up visit, physicians evaluated adverse events by asking patients, in an open-ended manner, whether they had experienced adverse events while receiving therapy and to elaborate on any that occurred. Although breath testing and breath methane level determinations were performed, we do not report them in our paper. Statistical Analysis We determined the number of patients for the study on the basis of the neomycin effect in a recent double-blind study for IBS on global improvement (8). To detect a difference of 35% (SD, 50%) with a power of 90%, we needed to assign 44 participants per group. Assuming a dropout rate of 10%, we calculated that approximately 96 patients would need to be recruited. The primary end point was global improvement in IBS symptoms during follow-up. As seen in Figure 1 , data were not available for all 10 weeks of follow-up. Figure 1. Study flow chart. We assessed the primary end point (percentage of global improvement) across the 10 weeks of follow-up by using an approach analogous to a repeated measures analysis of variance. Specifically, we used a mixed model with visit week (at 10 levels), treatment group (rifaximin or placebo), and group-by-week interaction as the fixed factors and patient as the random factor. The interaction and group factors were the main factors of interest in the analyses. We estimated mixed models by using the restricted maximum likelihood method. Because the global improvement percentage varied widely across week for most individuals, we considered week to be a categorical variable in the mixed model. Within-patient correlation across time was addressed by an autoregressive (first-order) model for the covariance structure. Missing data were mostly intermittent, and we assumed them to be missing at random. The normality assumption was rarely satisfied in either group at any week. However, at least 34 observations were recorded per group per week and the sample sizes were well-balanced, so we used the mixed-model analysis. We analyzed models with a single covariate (baseline diarrhea, constipation, abdominal pain, or bloating severity score). The covariate models did not improve the fit nor did they change the substantive results. Hence, we presented the simpler (no covariate) model results. We used a similar mixed-model approach to assess the secondary end points of abdominal pain, bloating, diarrhea, and constipation. Within-patient correlation was modeled by an autoregressive covariance structure. The normality assumption was rarely satisfied for the diarrhea outcome, with a similar floor effect for the primary outcome.

Methane Production During Lactulose Breath Test Is Associated with Gastrointestinal Disease Presentation

It has recently been determined that there is an increased prevalence of bacterial overgrowth in IBS. Since there are two gases (hydrogen and methane) measured on lactulose breath testing, we evaluated whether the different gas patterns on lactulose breath testing coincide with diarrhea and constipation symptoms in IBS and IBD. Consecutive patients referred to the gastrointestinal motility program at Cedars-Sinai Medical Center for lactulose breath testing were given a questionnaire to evaluate their gastrointestinal symptoms. Symptoms were graded on a scale of 0–5. Upon completion of the breath test, the results were divided into normal, hydrogen only, hydrogen and methane, and methane only positive breath tests. A comparison of all subjects and IBS subjects was undertaken to evaluate diarrhea and constipation with regards to the presence or absence of methane. This was further contrasted to Crohns and ulcerative colitis (UC) patients in the database. After exclusion criteria, 551 subjects from the database were available for comparison. Of the 551 subjects (P < 0.05, one-way ANOVA) and in a subgroup of 296 IBS subjects (P < 0.05, one-way ANOVA), there was a significant association between the severity of reported constipation and the presence of methane. The opposite was true for diarrhea (P < 0.001). If a breath test was methane positive, this was 100% associated with constipation predominant IBS. Furthermore, IBS had a greater prevalence of methane production than Crohns or UC. In fact, methane was almost nonexistent in the predominantly diarrheal conditions of Crohns and UC. In conclusion, a methane positive breath test is associated with constipation as a symptom.

Neomycin Improves Constipation-Predominant Irritable Bowel Syndrome in a Fashion That Is Dependent on the Presence of Methane Gas: Subanalysis of a Double-Blind Randomized Controlled Study

Recent studies have shown that normalization of the lactulose breath test (LBT) with neomycin leads to a significant reduction in irritable bowel syndrome (IBS) symptoms. This subanalysis was done on the constipation-predominant IBS subgroup of patients (C-IBS) to test the ability of neomycin to improve constipation and its correlation with the elimination of methane on breath test. IBS subjects underwent LBT in a blinded fashion. They were then randomly allocated to neomycin or placebo groups. For the purpose of this analysis, only the C-IBS subjects were identified. They were then evaluated for global improvement, abdominal pain, and constipation severity. The ability of neomycin to eliminate methane and its associated improvement in constipation was also determined. One hundred eleven subjects meeting Rome I criteria for IBS were included in the study. Thirty-nine of these had C-IBS. Of these, 20 received placebo and 19 received neomycin. With neomycin, a global improvement of 36.7±7.9% was seen, compared to 5.0±3.2% for placebo (P < .001) in the intention-to-treat analysis. Constipation was improved by 32.6±9.9% with neomycin compared to 18.7±7.2% for placebo (P=.26). Of the original 111 subjects, 12 demonstrated methane on breath test. All 12 of these patients were constipation predominant. In the methane producers receiving neomycin or placebo, improvement in constipation was significantly greater in those receiving neomycin (44.0±12.3%) compared to placebo (5.0±5.1%) (P < .05). Treatment with neomycin improves constipation in C-IBS. This improvement depends on the presence and elimination of methane on breath test.

Lower frequency of MMC is found in IBS subjects with abnormal lactulose breath test, suggesting bacterial overgrowth.

We have recently described an association between irritable bowel syndrome (IBS) and abnormal lactulose breath test, suggesting small intestinal bacterial overgrowth (SIBO). However, the mechanism by which SIBO develops in IBS is unknown. In this case–control study we evaluate the role of small intestinal motility in subjects with IBS and SIBO. Small intestinal motility was studied in consecutive IBS subjects with SIBO on lactulose breath test. After fluoroscopic placement of an eight-channel water-perfused manometry catheter, 4-hr fasting recordings were obtained. Based on this, the number and duration of phase III was compared to 30 control subjects. To test whether there was a relationship between the motility abnormalities seen and the SIBO status of the patient at the time of the motility, subjects with a breath test within 5 days of the antroduodenal manometry were also compared. Sixty-eight subjects with IBS and SIBO were compared to controls. The number of phase III events was 0.7 ± 0.8 in IBS subjects and 2.2 ± 1.0 in controls (P < 0.000001). The duration of phase III was 305 ± 123 sec in IBS subjects and 428 ± 173 in controls (P < 0.001). Subjects whose SIBO was still present at the time of manometry had less frequent phase III events than subjects with eradicated overgrowth (P < 0.05). In conclusion, phase III is reduced in subjects with IBS and SIBO. Eradication of bacterial overgrowth seems to result in some normalization of motility.

The Degree of Breath Methane Production in IBS Correlates With the Severity of Constipation

BACKGROUND:Recent work has demonstrated that among irritable bowel syndrome (IBS) subjects, methane on lactulose breath test (LBT) is nearly universally associated with constipation predominance. This work has been based on subjective constipation outcomes. In this study, methane is compared to constipation in another population of IBS subjects with constipation being determined both subjectively and objectively.METHODS:A nested study was conducted in subjects enrolled in a double-blind randomized placebo-controlled study. After consent, subjects were asked to complete a stool diary for 7 days. This included logging of all bowel movements that week as well as documenting the stool consistency for each during the same period using the Bristol Stool Score. After 7 days, subjects were asked to rate their symptoms on a visual analogue scale (VAS) score (0–100 mm) for diarrhea and constipation. They then had an LBT to evaluate both methane and hydrogen profiles over 180 min. Subjects with methane were compared to those without methane for Bristol Stool Score, stool frequency, as well as VAS scores for diarrhea and constipation. The degree of constipation was then compared to the quantity of methane production on LBT based on area under the curve.RESULTS:Among 87 subjects, 20 (23.8%) produced methane. IBS subjects with methane had a mean constipation severity of 66.1 ± 36.7 compared to 36.2 ± 30.8 for nonmethane producers (P < 0.001). The opposite was noted for diarrhea (P < 0.01). On LBT, the quantity of methane seen on breath test was directly proportional to the degree of constipation reported (r = 0.60, P < 0.01). In addition, greater methane production correlated with a lower stool frequency (r = −0.70, P < 0.001) and Bristol Stool Score (r = −0.58, P < 0.01).CONCLUSION:Methane on LBT is associated with constipation both subjectively and objectively. The degree of methane production on breath test appears related to the degree of constipation.

Breath testing to evaluate lactose intolerance in irritable bowel syndrome correlates with lactulose testing and may not reflect true lactose malabsorption.

OBJECTIVES:An increased prevalence of lactose intolerance is seen in irritable bowel syndrome (IBS). Recently, we demonstrated a high prevalence of abnormal lactulose breath test results in IBS suggesting bacterial overgrowth. Because symptoms of lactose intolerance result from bacterial fermentation, the purpose of this study was to determine whether an abnormal lactose breath test is reflective of malabsorption or early presentation to bacteria.METHODS:Subjects with diarrhea-predominant IBS were enrolled. On day 1, subjects underwent a lactulose breath test after an overnight fast. Within 1 wk, subjects returned after fasting for a lactose breath test with simultaneous blood glucose measurements every 15 min to complete a lactose tolerance test (LTT). Symptoms were evaluated 3 h after lactose administration.RESULTS:Twenty subjects completed the study. One subject inadvertently received dextrose through the intravenous and was excluded. Of the remaining 19 subjects, three (16%) had an abnormal LTT suggesting malabsorption. In all, 10 subjects (53%) had an abnormal lactose breath test, 14 (74%) had an abnormal lactulose breath test, and 11 (58%) had symptoms after lactose administration. The agreement with symptoms was moderate (κ= 0.47) and fair (κ= 0.24) when compared to the lactose breath test and LTT, respectively. There was a fair correlation between lactose breath test and LTT (κ= 0.29). However, lactose breath test hydrogen levels >166 ppm were universally predictive of abnormal LTT. Finally, a significant correlation was seen between the hydrogen production on lactose and lactulose breath test (r = 0.56, p = 0.01).CONCLUSIONS:Lactose breath testing in IBS subjects does not seem to reflect malabsorption; it may be an indicator of abnormal lactulose breath test, suggesting bacterial overgrowth.

A 14-day elemental diet is highly effective in normalizing the lactulose breath test.

Treatment of small intestinal bacterial overgrowth is frustrated by the low efficacy of antibiotics. Elemental diets have been shown to reduce enteric flora. In this study, we evaluate the ability of an elemental diet to normalize the lactulose breath test (LBT) in IBS subjects with abnormal breath test findings. Consecutive subjects with IBS and abnormal LBT suggesting the presence of bacterial overgrowth underwent a 2-week exclusive elemental diet. The diet consisted of Vivonex Plus (Novartis Nutrition Corp., Minneapolis, MN) in a quantity based on individual caloric requirement. On day 15 (prior to solid food), subjects returned for a follow-up breath test and those with an abnormal LBT were continued on the diet for an additional 7 days. The ability of an elemental diet to normalize the LBT was determined for days 15 and 21. A chart review was then conducted to evaluate any clinical benefit 1 month later. Of the 93 subjects available for analysis, 74 (80%) had a normal LBT on day 15 of the elemental diet. When those who continued to day 21 were included, five additional patients normalized the breath test (85%). On chart review, subjects who successfully normalized their breath test had a 66.4 ± 36.1% improvement in bowel symptoms, compared to 11.9 ± 22.0% in those who failed to normalize (P<0.001). An elemental diet is highly effective in normalizing an abnormal LBT in IBS subjects, with a concomitant improvement in clinical symptoms.

IBS Subjects with Methane on Lactulose Breath Test Have Lower Postprandial Serotonin Levels Than Subjects with Hydrogen

We have previously shown that methane on lactulose breath test (LBT) is highly associated with constipation in IBS and that methane gas itself slows small bowel transit in dogs. Previous studies suggest that serotonin may have a role in the control of transit in IBS. In this study, we aim to evaluate the role of serotonin in methane producing IBS subjects. Rome I-positive IBS subjects were recruited into the study after exclusion criteria were met. A fasting LBT was performed after subjects filled out a questionnaire rating the degree of constipation and diarrhea. Within 7 days of this test, subjects returned fasting for determination of serotonin before and after a 75-g oral glucose meal. The serotonin response was compared between hydrogen and methane producing IBS subjects. After 2 subjects were excluded for inadequate blood samples, 18 subjects completed the study. Four of 18 subjects produced methane. The postprandial serotonin level in methane producing IBS subjects was lower than in hydrogen producers (P<0.05). Methane producers had a reduction in serotonin after glucose. Methane producing IBS subjects have reduced postprandial serotonin. Whether methane is a surrogate marker of constipation or contributing to the reduced serotonin remains to be determined.

The presence of constipation and methane on lactulose breath test in ibs subjects is associated with lower serotonin levels compared to hydrogen alone

The presence of constipation and methane on lactulose breath test in ibs subjects is associated with lower serotonin levels compared to hydrogen alone