Figure 1.

Sample collection and general sample microbiome composition. (A) The general study design is outlined describing the patient cohorts, per hospital feeding protocol, and prospective fecal sampling. Infants born in two hospitals were enrolled shortly after birth and, per hospital protocol, assigned to a feeding protocol based on gestational age at birth and weight at birth. Infants born at >32 weeks corrected gestational age (cGA) and 1500 g were assigned to a feeding protocol which did not include a probiotic (gray, top). Infants born at <32 weeks cGA and/or <1500 g were assigned to a feeding protocol which included daily feedings of B. longum subsp. infantis EVC001 (8 × 10⁹ CFU once per day) starting with the initiation of trophic feeding (teal, bottom) and delivered in a single serving of 0.5 mL medium chain triglyceride oil. Nurses collected fecal samples (brown circles) opportunistically throughout the patient's stay in the neonatal intensive care unit (NICU), aiming for two or more samples per week. Infants received human milk (maternal or donor) through 34 weeks corrected gestational age, as well as human milk-based human milk fortifier. After 34 weeks, infant formula was used in addition to human milk. (B) Shotgun metagenome sequencing performed on samples collected from all 77 infants was used to classify the average community structure of the infant microbiome in all samples. Individual sample compositions were generally similar and typified by either: (C) a high abundance of Bifidobacterium longum species (inclusive of B. longum subsp. infantis; Bifidobacteriaceae, blue), (D) a high abundance of Staphylococcus epidermidis (Staphylococcaceae, gray) and Enterobacteriaceae (red), or (E) an overwhelming abundance of Escherichia coli and unclassified Escherichia (Enterobacteriaceae, red).