Background: Blenderized tube feeds are associated with reduced symptoms of nausea and vomiting. Current data on the impact of these diets on the oral or gastric microbiome is extremely limited. We hypothesize that patients receiving blenderized tube feeds have less full column gastroesophageal reflux, and therefore anticipate that there will be less gastric-oropharyngeal microbial overlap compared to patients on conventional formula.

Methods: Children age 0.5-18 years requiring enteral tube support were recruited from the Gastroenterology Clinic prospectively. Participants were excluded if exposed to treatment course of antibiotics within 4 weeks, or with history of cystic fibrosis or organ transplant. Diet category was defined as blenderized tube feeds if this constituted > 50% of calories and otherwise categorized as formula. Microbial samples were obtained from the posterior tongue via swab and stomach via gastric aspirate. 16S sequencing was performed, and analyzed using Qiime2, Phyloseq, and custom Python scripts. Alpha and beta diversity, taxonomic relative abundances and distance between the gastric and oral microbiomes were compared between the two groups. Statistical comparisons were performed by Student’s t-tests or Mann-Whitney U with Benjamini-Hochberg corrections and false discovery rate <0.05 when necessary.

Results: One hundred and ten children were recruited, with clinical characteristics presented in the Table. The tongue microbiomes differed significantly by Jaccard distances between diet groups, even after correcting for proton-pump inhibitor use, age, and sex (Figure). The similarity between gastric and tongue microbiome sites differed by diet, such that these sites were much more similar to each other in the formula group, and more dissimilar to each other in the blenderized tube feed group. Moreover, the gastric and tongue microbiome sites were most similar in patients receiving amino acid, extensively hydrolyzed or polymeric formulas. Several oral commensal organisms were present in higher relative abundance in the tongue in the blenderized tube feed group compared to the formula group, including Neisseria, Rothia, Pasteurellaceae and Lacobacillales. Prevotella and Veillonella, were present in higher relative abundance in the tongue in the formula group (Figure).

Conclusions: In the largest study to date, oral microbiomes of children receiving blenderized tube feeds differed significantly from those receiving formula, and were enriched in normal oral commensals compared to patients receiving formula. In keeping with our hypothesis, children on blenderized tube feeds have tongue microbiomes that are more distinct from their gastric microbiomes when compared to children receiving formula feeds, suggesting that reduction of reflux with blenderized feeds may mediate a healthier composition of the oral microbiome.

Christopher Chalmers

Appearances