Phage Sequences and Resistance Genes Predict Necrotizing Enterocolitis Eight Days Before Onset

Researchers reanalyzed existing shotgun metagenomic data from 43 infants who developed necrotizing enterocolitis and 86 matched controls born at 23–33 weeks gestation. Advanced assembly algorithms extracted previously discarded phage sequences and resistance genes; these features were fed into machine-learning models that achieved higher predictive performance than models limited to bacterial species composition. The eight-day lead time creates a potential intervention window before intestinal necrosis begins.

Prior work on necrotizing enterocolitis has focused almost exclusively on bacterial overgrowth, yet this study demonstrates that phage–bacteria interactions and mobile resistance elements are earlier and more specific signals. The finding aligns with observational data showing frequent antibiotic exposure in NICUs, which may enrich both resistance genes and phage-mediated gene transfer. It also suggests why probiotic trials have produced inconsistent results: they rarely account for the resident viral and resistome context.

Next steps require prospective validation across additional NICUs and integration of real-time sequencing into clinical workflows. If confirmed, the biomarkers could trigger targeted interventions such as narrowed antibiotic spectra or phage-based therapies before irreversible tissue damage occurs. A key remaining question is whether these signatures are causal drivers or merely correlated markers of an already perturbed ecosystem.