This study was a single-center, randomized, double-blind, and placebo-controlled clinical trial. Among the subjects who visited the health promotion center of Ajou University Hospital from June 2015 to October 2015, those who met the Rome III criteria for IBS were enrolled in this study (Figure 1). IBS patients aged ≥19 years were recruited, and initial screening included medical history taking, physical examination, and laboratory testing. We excluded subjects who met one of the following conditions: subjects with organic GI disease such as inflammatory bowel disease and a history of gluten intolerance (celiac disease); subjects taking antibiotics, synbiotics (including probiotics only), or any other drug that could influence bowel function within 1 month of consumption; the subjects with previous abdominal surgery except for appendectomy; pregnant or breastfeeding females; subjects with evidence of other organic, metabolic, or psychiatric diseases that could impact compliance as well as those with poor medication compliance.

Of the 38 subjects who underwent the screening process, eight applicants were excluded according to the inclusion/exclusion criteria. Therefore, 30 subjects were initially enrolled in the study; of these, 28 subjects completed this study, as two were withdrawn for unknown reasons.

All the subjects completed an informed consent form before participating in the study, and the study was conducted after obtaining approval from the institutional review board of Ajou University Hospital (AJIRB-MED-DRU-14-317). This study was conducted in compliance with the standards of the Declaration of Helsinki and current ethical guidelines.

The subjects were randomly assigned at a 1:1:1 ratio to the high-dose, low-dose, or placebo groups according to block randomization. We did not control any variables during randomization. The patients, investigators, doctors, and study nurses were blinded using randomization codes, which were kept confidential until the end of data analysis. The placebo group took 2 capsules of placebo, the low-dose group received 1 capsule of placebo and 1 capsule of synbiotics, and the highdose group received 2 capsules of synbiotics. The study drugs were administered once a day for 8 weeks. All patients filled in fatigue and bowel symptom questionnaires at the first and last visit. After 4 weeks of the study, all subjects were revisited and their physical condition, vital signs, study compliance, and adverse events were evaluated. Their laboratory test results, study compliance, and adverse events were then re-checked after the completion of 8 weeks of the study. During the entire study period, all patients were prohibited from consuming yogurt or taking pre- or probiotic pills or other GI drugs.

The size of the study sample was not based on a power calculation owing to the absence of reliable calculation data about synbiotics in IBS. Instead, it was considered adequate to characterize a potential effect with sufficient accuracy based on previous experience gained through similar studies [8-10]. In other words, the previous two studies were performed with 25 subjects and 24 subjects with a double-blind randomized design, respectively [8,10]. In another single-blind randomized controlled trial (RCT) of 30 patients that investigated the effects of probiotics on intestinal permeability in IBS [9], the authors estimated that the sample size of 30 patients had a power of 80% to detect a difference of 30% in the proportion of increased small bowel permeability between the placebo and the probiotics group after treatment at the significance level of 5% with a low dropout rate (<5%).

The investigational drug was Ultra-Probiotics-500, supplied by B&A Health Products Inc., Los Angeles, CA, USA. Each capsule contains 10 billion colony-forming units of probiotic bacteria comprising six strains of Lactobacillus (rhamnosus, acidophilus, casei , bulgaricus, plantarum, and salivarius) and two strains of Bifidobacterium (bifidum and longum). Each capsule also contains 175 mg of fructooligosaccharides, 150 mg of Ulmus davidiana (Slippery elm bark powder), 10 mg of Geum urbanum (herb bennet) powder, and 100 mg of inulin powder as prebiotics. The placebo was encapsulated in the same material as the synbiotic capsule, and the content of the capsule was manufactured such that it could not be distinguished by taste or color.

The data were examined using intention-to-treat (ITT) and per-protocol (PP) analyses. For the ITT analysis, all data from the study subjects were included. For missing or incomplete data, the most recently available data were used for a last-observation-carried-forward (LOCF) analysis. For the PP analysis, only the data obtained from the subjects who completed the study on schedule were included, and the analysis was performed only when the data were available at weeks 0 and 8. Of the 30 subjects who enrolled in the study, two subjects were withdrawn (from the high-dose group) and none were omitted due to noncompliance. All side effects were reported and noted in detail, regardless of their relevance to the study drug.

Descriptive statistics and a frequency analysis were performed to describe the clinical characteristics of the study subjects. We used the Wilcoxon signed rank test to calculate the difference in clinical variable scores between the baseline and after the study within the groups. Then, we used the Kruskal-Wallis test for the comparison of mean clinical characteristics scores between the three groups at baseline and the end of the study (placebo, low-dose, and high-dose groups). We also performed the Kruskal-Wallis test to compare the differences in variables from baseline at 8 weeks among the groups. A P-value <0.05 was considered to indicate statistical significance. For variables with a significant difference between the three groups in the Kruskal-Wallis test at the end of study, we performed the Mann-Whitney test with Bonferroni correction. According to Bonferroni correction of the three groups, P-values <0.017 were considered to indicate statistical significance. Subsequently, we used the Spearman rank correlation test to establish the dose–response trends. All analyses were carried out using IBM SPSS ver. 19.0 (IBM Corp., Armonk, NY, USA).

Table 1 shows the baseline clinical characteristics of the study subjects in the 3 groups by PP analysis (n=28). Overall, the mean age of the study population was 50±8.6 years, the median age was 50.5 years, and 53.6% of the study population was female. The average duration of IBS was 11±9.5 years, and the main subtype was the diarrheal subtype at 71.4% (constipation subtype was 21.4%, mixed subtype was 7.1%). The average abdominal discomfort score was 5.4±1.7, abdominal bloating score was 5.6±2.2, fatigue VAS was 4.5±1.9, and MFI was 89.3±13.1.

There were no significant differences between the three groups with respect to gender, abdominal discomfort, abdominal bloating, frequency of formed stool, epigastric soreness, and nausea; only the age of the subjects showed differences (P=0.086, 0.286, 0.255, 0.384, 0.506, 0.399, and 0.020, respectively). The differences in age (and gender proportion, even it was not significant) at baseline were considered a random error due to the small numbers of subjects in each study group. In addition, there were no significant differences between the groups in terms of baseline biochemistry characteristics, including white blood cell count, hemoglobin, platelet count, AST, ALT, GGT, BUN, and creatinine (P=0.231, 0.452, 0.612, 0.893, 0.115, 0.569, 0.492, and 0.500). All these values were within the normal range. The mean FSS, VAS, and MFI scores were not significantly different between the groups at baseline (P=0.609, 0.256, and 0.438).

Abdominal discomfort, abdominal bloating, fatigue VAS, and MFI in the low-dose and high-dose groups were significantly improved after intervention (P=0.005, 0.007, 0.008, and 0.008 in the low-dose group; P=0.010, 0.011, 0.012, and 0.011 in the high-dose group, respectively), but not in the placebo group (P=0.271, 0.180, 0.196, and 0.681) by PP analysis. Frequency of formed stool showed a significant difference in the placebo, low-dose, and high-dose groups after intervention (P=0.005, 0.007 and 0.010) by PP analysis. In the ITT with LOCF analysis (the high-dose group only), none of these variables showed any significant difference, including abdominal discomfort (P=0.059) and frequency of formed stool (P=0.066). In the Kruskal-Wallis test, abdominal discomfort, abdominal bloating, frequency of formed stool, fatigue VAS, and MFI showed statistical significance (P=0.002, 0.006, 0.007, 0.028, and 0.041) in all three groups at 8 weeks (Table 2). However, after the Mann-Whitney test with Bonferroni correction, abdominal discomfort, abdominal bloating, frequency of formed stool, and fatigue VAS showed significant improvement in the high-dose group only (P=0.002, 0.003, 0.002, and 0.013) compared to the placebo group (Table 3). A statistically significant difference in FSS was not found between the groups at 8 weeks (P=0.115). In the ITT with LOCF analysis, none of the differences were statistically significant. The differences in five variables, abdominal discomfort, bloating, frequency of formed stool, fatigue VAS, and MFI, from baseline at 8 weeks also showed no statistical significance among the groups by PP analysis (P=0.497, 0.253, 0.125, 0.194, and 0.967, respectively). Furthermore, there were no significant differences between the groups at 8 weeks with respect to white blood cell count and levels of hemoglobin, platelet count, AST, ALT, GGT, BUN, and creatinine (P=0.116, 0.382, 0.144, 0.720, 0.359, 0.701, 0.186, and 0.181 by PP analysis).

Correlations between study groups and variables after intervention were determined by Spearman’s rank correlation test (Table 4). Abdominal discomfort, abdominal bloating, formed stool frequency, fatigue VAS, and MFI showed significant correlations with the 3 study groups by PP analysis (r=-0.668, P<0.001; r=-0.604, P=0.001; r=0.605, P=0.001; r=-0.513, P=0.005; and r=-0.408, P=0.031, respectively). In the ITT with LOCF analysis, there were no significant correlations.

In this clinical trial, there were no reports of any significant adverse events during the study or any subjects experiencing either serious adverse events or withdrawal due to adverse events.