Aim
To compare quantities of predominant and pathogenic bacteria in mucosal and faecal samples.
Conclusion
Non-invasive faecal samples do not reflect bacterial counts on the mucosa at the individual level, except for bifidobacteria often analysed in probiotic intervention studies.
Methods
Twenty patients undergoing diagnostic colonoscopy with endoscopically and histologically normal mucosa were recruited to the study, 14 subjects of which also supplied faecal (F) samples between 15 d to 105 d post colonoscopy. Mucosal biopsies were taken from each subject from the midportion of the ascending colon (right side samples, RM) and the sigmoid (left side samples, LM). Predominant intestinal and mucosal bacteria including clostridial 16S rRNA gene clusters IV and XIVab, Bacteroidetes, Enterobacteriaceae, Bifidobacterium spp., Akkermansia muciniphila (A. muciniphila), Veillonella spp., Collinsella spp., Faecalibacterium prausnitzii (F. prausnitzii) and putative pathogens such as Escherichia coli (E. coli), Clostridium difficile (C. difficile), Helicobacter pylori (H. pylori) and Staphylococcus aureus (S. aureus) were analysed by quantitative polymerase chain reaction (qPCR). Host DNA was quantified from the mucosal samples with human glyceraldehyde 3-phosphate dehydrogenase gene targeting qPCR. Paired t tests and the Pearson correlation were applied for statistical analysis.
Results
The most prominent bacterial groups were clostridial groups IV and XIVa+b and Bacteroidetes and bacterial species F. prausnitzii in both sample types. H. pylori and S. aureus were not detected and C. difficile was detected in only one mucosal sample and three faecal samples. E. coli was detected in less than half of the mucosal samples at both sites, but was present in all faecal samples. All detected bacteria, except Enterobacteriaceae, were present at higher levels in the faeces than in the mucosa, but the different locations in the colon presented comparable quantities (RM, LM and F followed by P(1) for RM vs F, P(2) for LM vs F and P(3) for RM vs LM: 4.17 ± 0.60 log(10)/g, 4.16 ± 0.56 log(10)/g, 5.88 ± 1.92 log(10)/g, P(1) = 0.011, P(2) = 0.0069, P(3) = 0.9778 for A. muciniphila; 6.25 ± 1.3 log(10)/g, 6.09 ± 0.81 log(10)/g, 8.84 ± 1.38 log(10)/g, P(1) < 0.0001, P(2) = 0.0002, P(3) = 0.6893 for Bacteroidetes; 5.27 ± 1.68 log(10)/g, 5.38 ± 2.06 log(10)/g, 8.20 ± 1.14 log(10)/g, P(1) < 0.0001, P(2) ≤ 0.0001, P(3) = 0.7535 for Bifidobacterium spp.; 6.44 ± 1.15 log(10)/g, 6.07 ±1.45 log(10)/g, 9.74 ±1.13 log(10)/g, P(1) < 0.0001, P(2) ≤ 0.0001, P(3) = 0.637 for Clostridium cluster IV; 6.65 ± 1.23 log(10)/g, 6.57 ± 1.52 log(10)/g, 9.13 ± 0.96 log(10)/g, P(1) < 0.0001, P(2) ≤ 0.0001, P(3) = 0.9317 for Clostridium cluster XIVa; 4.57 ± 1.44 log(10)/g, 4.63 ± 1.34 log(10)/g, 7.05 ± 2.48 log(10)/g, P(1) = 0.012, P(2) = 0.0357, P(3) = 0.7973 for Collinsella spp.; 7.66 ± 1.50 log(10)/g, 7.60 ± 1.05 log(10)/g, 10.02 ± 2.02 log(10)/g, P(1) ≤ 0.0001, P(2) = 0.0013, P(3) = 0.9919 for F. prausnitzsii; 6.17 ± 1.3 log(10)/g, 5.85 ± 0.93 log(10)/g, 7.25 ± 1.01 log(10)/g, P(1) = 0.0243, P(2) = 0.0319, P(3) = 0.6982 for Veillonella spp.; 4.68 ± 1.21 log(10)/g, 4.71 ± 0.83 log(10)/g, 5.70 ± 2.00 log(10)/g, P(1) = 0.1927, P(2) = 0.0605, P(3) = 0.6476 for Enterobacteriaceae). The Bifidobacterium spp. counts correlated significantly between mucosal sites and mucosal and faecal samples (Pearson correlation coefficients 0.62, P = 0.040 and 0.81, P = 0.005 between the right mucosal sample and faeces and the left mucosal sample and faeces, respectively).