Consecutive patients with chronic non-bloody watery diarrhoea were prospectively included during the period from September 2014 to December 2018 based on the following inclusion and exclusion criteria, which were selected to maximize the pre-test probability of MC.
Inclusion criteria: (1) Women 50 years or older and men 70 years or older; (2) chronic watery diarrhoea with two or more daily liquid stools (Bristol scale = 6 or 7) or frequent episodes (at least three times a week) of watery diarrhoea (Bristol scale = 6 or 7), with a duration of at least one month; (3) normal blood test and biochemistry (including C reactive protein and TSH), negative anti-transglutaminase antibodies, and negative faecal ova and parasites; (4) patients with an indication for a diagnostic colonoscopy by their physician at charge, mainly to rule out MC; and (5) signature of the study informed consent.
Exclusion criteria: (1) patients with alternating diarrhoea-constipation and self-limiting diarrhoea at the time of colonoscopy; (2) patients receiving antibiotic treatment from three months prior to the study until its completion; (3) patients who had travelled to developing or underdeveloped countries from three months before the start of the study until its completion; (4) patients on low-calorie diets, vegan diets, gluten-free diets and other ‘special’ diets; (5) patients consuming probiotics or herbal remedies from three months before the start of the study until its completion; (6) bacterial or parasite intestinal infection (including Blastocystis hominis) in the previous three months; (7) previous history of coeliac disease, inflammatory bowel disease or other types of enteropathy; (8) previous gastrointestinal surgery (excluding appendectomy or inguinal herniorrhaphy); (9) alcoholism or drug addiction; and (10) inability to understand the instructions for participating in this study.
Healthy volunteers aged 18–75 years without digestive symptoms and none of the exclusion criteria above described were included after signing the informed consent and formed the healthy control group (HC).
Faecal sample collection
In all included patients, faecal samples were collected both three days before split-dose PEG colonic cleansing and at 30 days after colonoscopy (always before starting specific treatment for the diarrhoeal illness). Faecal samples were collected by the patients at home in appropriate sterile plastic containers and immediately frozen at − 20 °C. To avoid sample thawing, portable thermal systems were used to transport the faecal samples from home to the laboratory. The samples were then stored at − 80 °C until processing.
Healthy volunteers also performed PEG colonic cleansing and collected faecal samples following the same protocol: 3 days before PEG and 30 days after PEG. The same procedure for sampling, freezing, and storing faecal samples was used.
Diagnostic work-up of chronic watery diarrhoea
A complete colonoscopy was performed under conscious intravenous sedation on all included patients. Multiple biopsy specimen samples were obtained when the macroscopic appearance of the colonic mucosa was normal or mildly abnormal (mild erythema or oedema). Routinely, four samples from the ascending colon, and two each from the transverse, descending, and sigmoid zones, were taken. MC diagnosis was based on both clinical and histological criteria as previously defined [1, 2]. Histological MC diagnosis was reviewed in all cases by experienced pathologists at the participating centres.
When histological examination of colonic samples was normal and diarrhoea persisted, a 75SeHCAT (Se-homotaurocholate) abdominal retention test was performed to assess BAD . BAD was defined as a seventh day retention value < 10%. A value < 5% was considered to be severe BAD. Capsule endoscopy or intestinal MRI was performed on patients with normal ileocolonoscopy and increased levels of faecal calprotectin to completely rule out either Crohn’s disease or other small bowel enteropathies. FD and diarrhoea-predominant IBS were diagnosed when the results of all specific tests performed were normal and the patient fulfilled the Roma III criteria for each functional disease.
Clinical remission definition
Patients were visited at 30 days after colonoscopy. Clinical remission was defined as the absence of watery stools (Bristol scale ≤ 5) in the last week before visit. Patients were followed up by a phone call at month four to rule out diarrhoea relapse.
Faecal microbiome analysis
Genomic DNA extraction
DNA was extracted following the International Human Microbiome Standards (IHMS; http://www.microbiome-standards.org) . A frozen aliquot (250 mg) of each sample was suspended in 250 mL of guanidine thiocyanate, 40 mL of 10% N-lauroyl sarcosine and 500 mL of 5% N-lauroyl sarcosine. DNA was extracted by the mechanical disruption of the microbial cells with beads and nucleic acids were recovered from clear lysates by alcohol precipitation. An equivalent of 1 mg of each sample was used for DNA quantification using a NanoDrop ND-1000 Spectrophotometer (Nucliber). DNA integrity was examined by micro-capillary electrophoresis using an Agilent 2100 Bioanalyzer.
High-throughput DNA sequencing
For profiling microbiome composition, the hyper-variable region (V4) of the bacterial and archaeal 16S rRNA gene was amplified by PCR. For amplification, the universal primers V4F_517_17: 5′GCCAGCAGCCGCGGTAA-3′ (Forward primer) and V4R_805_19: 5′-GACTACCAGGGTATCTAAT-3′ (Reverse primer) were used. The use of these primers guarantees the amplification of practically 100% of the domains of bacteria and archaea. The sequencing process, following standard Illumina platform protocols (Illumina website), was performed as previously described .
Sequence data analysis
The raw sequences were loaded into the QIIME 1.9.1 pipeline . To study diversity information, Operational Taxonomic Units (OTUs) tables were performed. To estimate the microbial richness and evenness of the sample, in terms of what are known as alpha-diversity estimates, we calculated the Chao1 and Shannon diversity indexes [17, 18]. To calculate between-samples diversity or beta-diversity, weighted and unweighted UniFrac metrics were applied to build phylogenetic distance matrices, which were then used to construct hierarchical cluster trees using the Unweighted Pair Group Method with Arithmetic mean and Principal Coordinate Analysis (PcoA) representations.
Microbial dysbiosis index
The microbial dysbiosis index (MD-index) was calculated as previously described . The MD-index is defined as the log of [total abundance in organisms increased in either MC, FD or BAD] over [total abundance of organisms decreased in either MC, FD or BAD]. Increased or decreased organisms were defined as those with a p < 0.05 in comparison with the HC group.
The study protocol was submitted and approved by the local Ethical and Research Committees of both the Hospital Universitari Mutua Terrassa (18 June 2014, Acta 06/14, Terrassa, Barcelona, Spain) and the University Hospital Vall d’Hebron (Barcelona, Spain). The study was conducted in accordance with the Declaration of Helsinki. All patients and healthy volunteers received information concerning their participation in the study and gave written informed consent.
This was an exploratory observational study and no calculation of the sample size was intended. There were no previous data about the remission rate achieved after PEG colonic cleansing, and in this sense present study should be considered as a pilot study. We select the inclusion criteria to maximize the pre-test probability of MC and we stopped patient recruitment when 20 MC patients were included.
Statistical analyses were carried out in QIIME and in R . To work with normalised data, we analysed an equal number of sequences from all groups. The Shapiro–Wilk test was used to check the normality of data distribution. Parametric normally distributed data were compared by the Student’s t test for paired or unpaired data. Otherwise, the Wilcoxon signed rank test was used for paired data and the Mann–Whitney U test for unpaired data. The Kruskal–Wallis one-way test of variance was used to compare the median alpha-diversity and MD-Index between groups, as well as the number of sequences of the groups at various taxonomic levels. The Friedman test was used for one-way repeated measures analysis of variance. We used a permutational multivariate analysis of variance (PERMANOVA), a non-parametric multivariate analysis of variance, to test for differences in microbial community composition adjusted for age and sex. When possible, the analysis provided false discovery rate (FDR)-corrected p values. FDR, q < 0.05 was considered significant. Spearman correlation was used to evaluate significant associations between alpha-diversity, MD-index and the daily stool number. MedCalc statistical software, version 18.2.1 (MedCalc Software bvba, Ostend, Belgium) was used to construct the figures.