Characterization of Microbiota in Children with Chronic Functional Constipation

Abstract

Objectives: Disruption of the intestinal microbiota is considered an etiological factor in pediatric functional constipation. Scientifically based selection of potential beneficial probiotic strains in functional constipation therapy is not feasible due to insufficient knowledge of microbiota composition in affected subjects. The aim of this study was to describe microbial composition and diversity in children with functional constipation, compared to healthy controls.

Study design: Fecal samples from 76 children diagnosed with functional constipation according to the Rome III criteria (median age 8.0 years; range 4.2-17.8) were analyzed by IS-pro, a PCR-based microbiota profiling method. Outcome was compared with intestinal microbiota profiles of 61 healthy children (median 8.6 years; range 4.1-17.9). Microbiota dissimilarity was depicted by principal coordinate analysis (PCoA), diversity was calculated by Shannon diversity index. To determine the most discriminative species, cross validated logistic ridge regression was performed.

Results: Applying total microbiota profiles (all phyla together) or per phylum analysis, no disease-specific separation was observed by PCoA and by calculation of diversity indices. By ridge regression, however, functional constipation and controls could be discriminated with 82% accuracy. Most discriminative species were Bacteroides fragilis, Bacteroides ovatus, Bifidobacterium longum, Parabacteroides species (increased in functional constipation) and Alistipes finegoldii (decreased in functional constipation).

Conclusions: None of the commonly used unsupervised statistical methods allowed for microbiota-based discrimination of children with functional constipation and controls. By ridge regression, however, both groups could be discriminated with 82% accuracy. Optimization of microbiota-based interventions in constipated children warrants further characterization of microbial signatures linked to clinical subgroups of functional constipation.

Conflict of interest statement

P.H.M Savelkoul and A.E. Budding have proprietary rights on the IS-pro platform technology (patent ‘Microbial population analysis’, WO2008/125365) and are co-founders of a spin-off company developing this technique into a clinical diagnostic product. This patent is only related to the microbiota profiling technique IS-pro. This does not alter our adherence to all PLOS ONE policies on sharing data and materials.

Figures

Fig 1. Clustered heat map with IS profiles of children with functional constipation and controls.

Clustered heat map displaying IS profiles of 76 children with functional constipation and 61 healthy controls. Individual subjects are shown on the X axis; children with constipation in red, healthy controls in green. On the Y axis, IS fragment lengths are expressed (in number of nucleotides), corresponding with bacterial strain type (OTU). Blue peaks represent Firmicutes, Actinobacteria, Fusobacteria, Verrucomicrobia (FAFV), red peaks represent Bacteroidetes and yellow peaks represent Proteobacteria. Intensity of colors reflect relative dominance of each indicated bacterial strain, grey signals represent less prevalent IS fragment lengths. No disease-specific clustering was observed, indicating that the groups could not be distinguished based on IS profiles using this unsupervised method. The most abundant OTUs in both study groups were observed within the phylum Bacteriodetes, corresponding to the species Bacteroides vulgatis (480 nt), Alistipes putredinis (220 nt) and Alistipes finegoldii (401 nt).

Fig 2. Principle coordinate analysis of microbial profiles of children with and without functional constipation.

Principle coordinate analysis scatterplot displaying overall bacterial community composition, showing no separate clustering of microbial profiles of children with functional constipation (red dots) and controls (green dots) for all phyla together (A) and per phylum (B: Bacteroidetes; C: Firmicutes, Actinobacteria, Fusobacteria, Verrucomicrobia (FAFV); D: Proteobacteria).

Fig 3. Diversity indices of children with functional constipation and healthy controls.

Shannon diversity index of children with functional constipation (dark colors) and healthy controls (light colors), showing similar indices, both when taken all phyla together, as well as on phylum level. Green: all phyla taken together. Red: Bacteroidetes, blue: represent Firmicutes, Actinobacteria, Fusobacteria, Verrucomicrobia (FAFV), yellow: Proteobacteria.

Fig 4. Receiver operating characteristic (ROC) curves for the discrimination of children with and without functional constipation.

Receiver operating characteristic (ROC) curves summarizing the predictive power of the cross-validated logistic ridge regression model for clinical status per phylum and for all phyla combined.