Objective

Small for gestational age (SGA) preterm infants (PT) are at greatest risk for growth failure. Our objective was to assess the impact of an exclusive human milk diet (HUM) on growth velocities and neonatal morbidities from birth to discharge in a SGA population.

Study design

Multicenter, retrospective cohort study, subgroup analysis of SGA PT comparing a cow's milk diet (CMD) with HUM diet.

Results

At birth 420 PT were classified as SGA (197 CMD group, 223 HUM group). Demographics and anthropometric measurements were similar. HUM group PT showed improvement in length Z score at discharge (p = 0.024) and reduction in necrotizing enterocolitis (NEC) (p = 0.004).

Conclusion

SGA PT fed a HUM diet had significantly decreased incidence of NEC, surgical NEC, and late-onset sepsis. Due to concerns about growth in a HUM diet, it is reassuring SGA infants fed the HUM diet had similar growth to CMD diet with trends toward improvement.

Necrotizing enterocolitis (NEC) is the leading cause of death caused by gastrointestinal disease in preterm infants. Major risk factors include prematurity, formula feeding, and gut microbial colonization. Microbes have been linked to NEC, yet there is no evidence of causal species, and select probiotics have been shown to reduce NEC incidence in infants. In this study, we evaluated the effect of the probiotic Bifidobacterium longum subsp. infantis (BL. infantis), alone and in combination with a human milk oligosaccharide (HMO)-sialylactose (3'SL)-on the microbiome, and the incidence of NEC in preterm piglets fed an infant formula diet. We studied 50 preterm piglets randomized between 5 treatments: (1) Preterm infant formula, (2) Donor human milk (DHM), (3) Infant formula + 3'SL, (4) Infant formula + BL. infantis, and (5) Infant formula and BL. infantis + 3'SL. NEC incidence and severity were assessed through the evaluation of tissue from all the segments of the GI tract. The gut microbiota composition was assessed both daily and terminally through 16S and whole-genome sequencing (WGS) of rectal stool samples and intestinal contents. Dietary BL. infantis and 3'SL supplementation had no effect, yet DHM significantly reduced the incidence of NEC. The abundance of BL. infantis in the gut contents negatively correlated with disease severity. Clostridium sensu stricto 1 and Clostridium perfringens were significantly more abundant in NEC and positively correlated with disease severity. Our results suggest that pre- and probiotics are not sufficient for protection from NEC in an exclusively formula-based diet. The results highlight the differences in microbial species positively associated with both diet and NEC incidence.