Food Insecurity, Gut Microbiome Composition, and Sutterella Enrichment in Ethiopian Children

In a cross-sectional study of Ethiopian schoolchildren, household food insecurity—measured with the Household Food Insecurity Access Scale (HFIAS)—was associated with differences in gut microbiome composition based on fecal samples.

Across analyses, they describe a recurring genus-level signal involving Sutterella among children classified as food-insecure. The paper also presents a proof-of-concept microbiome-based classifier that separated participants by food security status using taxonomic features. Overall, the report links food insecurity metrics with shifts in gut microbial composition in this cohort, with repeated identification of specific taxa in exploratory analyses.

The study included 57 school-aged children recruited in Hawela Tula (Hawassa, Sidama Regional State, Ethiopia) during the 2023 academic year. Eligibility criteria excluded recent anti-helminthic treatment or iron supplementation and specified a minimum local residence period. Fecal samples underwent 16S rRNA amplicon sequencing, and household food insecurity status was assessed using the nine-item HFIAS instrument with a 30-day recall window and score-based classification into food-secure versus food-insecure categories. The authors describe analyses spanning community-level diversity, group separation, and follow-up models aligned to individual HFIAS questions intended as proxies for dietary deprivation. The central comparison was microbiome profiles by overall food security status and by specific reported food-insecurity experiences.

At the community level, the authors report no statistically significant differences in alpha diversity by food security status using Chao1 and Shannon indices (Wilcoxon p > 0.05). They report beta diversity differences between food-secure and food-insecure groups using Bray–Curtis dissimilarity (PERMANOVA p < 0.05), with principal coordinate analysis presented as a visualization of these compositional differences. In analyses of individual HFIAS components, limited dietary variety, consumption of disliked foods, and reduced meal size were each associated with compositional shifts (PERMANOVA; q < 0.05 for each). These results were framed as community composition differences associated with food insecurity measures rather than changes in within-sample richness or evenness.

For taxa-level signals, the authors describe applying differential abundance workflows (including LEfSe and Maaslin2) to identify features associated with food insecurity definitions. They report that Sutterella was more abundant among food-insecure participants, describing a composite-model multiple-testing result of q = 0.11 alongside component-specific models in which Sutterella met q < 0.05. They also note additional taxa in group-enrichment outputs, including Prevotella 9 and the family Christensenellaceae as discriminating features, without presenting these as the primary repeated finding. Across the paper’s reported analyses, Sutterella was the most consistently highlighted genus-level pattern linked to food insecurity classification and selected HFIAS items.

The authors also report a machine-learning experiment evaluating whether microbiome profiles could classify food security status in this dataset. Using an XGBoost classifier with genus- and family-level features, centered log-ratio transformation, and 10-fold cross-validation, they report AUC = 0.81 and balanced accuracy of approximately 84%. Feature-importance outputs ranked taxa including Sutterella and Veillonella among the top contributors, alongside higher-level features such as families represented in the filtered set. They describe this as proof-of-concept and note constraints on broader inference, citing the modest sample size, cross-sectional design, dietary frequency variables not incorporated into the microbiome analyses, and absence of external validation.

In this cohort, the model demonstrates within-sample discrimination by reported food security status, while the paper stops short of demonstrating external generalizability or linkage to clinical endpoints.

Key Takeaways:

• The authors report that food insecurity status was associated with gut microbiome compositional differences (beta diversity), while alpha diversity metrics did not differ by status.
• Across reported differential abundance analyses, Sutterella was repeatedly identified as more abundant among food-insecure children in this cohort.
• A microbiome-feature classifier distinguished food security status in these participants, but the authors describe cross-sectional design, small sample size, dietary frequency variables not incorporated into the microbiome analyses, and no external validation as constraints on broader inference.