Evaluation of Intestinal Microbial Metabolites in Preterm Infants with Different Initial Feeding Methods by In Vitro Fermentation Modeling System

Abstract

We aim to explore the intestinal microbial metabolites in preterm infants with noninvasive methods and analyze the effects of initial feeding methods. Preterm infants with gestational weeks lower than 34 were recruited for fecal sample collection every 7 days. Fecal pH, ammonia, bile acid, and secretory IgA (sIgA) were tested. A 1:10 fecal slurry was inoculated into different culture media containing different carbohydrates as the only carbon source: lactose (LAT), fructooligosaccharide (FOS), galactooligosaccharide (GOS), and 2′-fucosyllactose (FL2). After 24 h of anaerobic culture through an in vitro fermentation system, air pressure difference, carbohydrate degradation rate, and short-chain fatty acids (SCFAs) content in fermentation pots were measured. Preterm infants were assigned into two groups: group A, preterm infants fed by human milk, including mother’s own milk and donor human milk (DHM); group B, preterm infants fed by preterm formula at first 3 days and fed by human milk (including mother’s own milk and DHM) from day 4 to discharge. Group A included 90 samples and group B included 70 samples. Group A had lower fecal pH (p = 0.023), ammonia (p = 0.001), and bile acids (p = 0.025). Group B also had higher fecal sIgA levels, both in OD (p = 0.046) and concentration (p < 0.0001) methods. Carbohydrates degradation rates in group A were higher than group B, especially in LAT medium (p = 0.017) and GOS medium (p = 0.005). Gas production amount had no significant difference in all four media. Several different SCFAs in four kinds of different culture media in group A were higher than in group B, but valeric acid was lower in group A. The initial feeding methods may affect the preterm infants’ intestinal microecology and microbial metabolites for at least several weeks.

Keywords: human milk; initial feeding methods; intestinal metabolites; preterm infants.

Figure 1

The rate of carbon source degradation by in vitro fermentation modeling system based on different initial feeding methods. Group A, preterm infants fed by human milk, including mother’s own milk and DHM; group B, preterm infants fed by preterm formula at first 3 days and fed by human milk (including mother’s own milk and DHM) from day 4 to discharge. LAT, lactose; FOS, fructooligosaccharide; GOS, galactooligosaccharide; FL2, 2′-fucosyllactose. Data are shown as median with interquartile and error bars represent 5–95 percentile (* p < 0.05, ** p < 0.01).

Figure 2

Heat map analysis of fecal samples based on gas production in different media by in vitro fermentation modeling system. Rows for different samples. Columns for different fermentation media. LAT, lactose; FOS, fructooligosaccharide; GOS, galactooligosaccharide; FL2, 2′-fucosyllactose.

Figure 3

The concentration of fecal SCFAs, total SCFAs (A), acetic acid (B), propionic acid (C), butyric acid (D), isobutyric acid (E), valeric acid (F), and isovaleric acid (G) in different media by in vitro fermentation modeling system based on initial feeding methods. Group A, preterm infants fed by human milk, including mother’s own milk and DHM; group B, preterm infants fed by preterm formula for first 3 days and fed by human milk (including mother’s own milk and DHM) from day 4 to discharge. Slurry, 1:10 fresh fecal slurry; LAT, lactose; FOS, fructooligosaccharide; GOS, galactooligosaccharide; FL2, 2′-fucosyllactose. Data are shown as median with interquartile and error bars represent 5–95 percentile (* p < 0.05, ** p < 0.01).