Effects of One-Week Empirical Antibiotic Therapy on the Early Development of Gut Microbiota and Metabolites in Preterm Infants

Abstract

The early postnatal period is the most dynamic and vulnerable stage in the assembly of intestinal microbiota. Antibiotics are commonly prescribed to newborn preterm babies and are frequently used for a prolonged duration in China. We hypothesized that the prolonged antibiotic therapy would affect the early development of intestinal microbiota and their metabolites. To test this hypothesis, we analyzed the stool microbiota and metabolites in 36 preterm babies with or without antibiotic treatment. These babies were divided into three groups, including two groups treated with the combination of penicillin and moxalactam or piperacillin-tazobactam for 7 days, and the other group was free of antibiotics. Compared to the antibiotic-free group, both antibiotic-treated groups had distinct gut microbial communities and metabolites, including a reduction of bacterial diversity and an enrichment of harmful bacteria such as Streptococcus and Pseudomonas. In addition, there was a significant difference in the composition of gut microbiota and their metabolites between the two antibiotic-treated groups, where the piperacillin-tazobactam treatment group showed an overgrowth of Enterococcus. These findings suggest that prolonged antibiotic therapy affects the early development of gut microbiota in preterm infants, which should be considered when prescribing antibiotics for this population.

Figure 1

Venn diagrams of the distribution of Operational Taxonomic Units (OTUs) among three study groups on day 3 (d3) and day 7 (d7). The number in the intersection represent the number of OTUs shared among different groups while the number out of the intersection represents the number of unique OTUs in each group. AF, antibiotic free group (red color); PT, piperacillin-tazobactam group (green color); PM, combination of penicillin and moxalactam group (blue color).

Figure 2

Comparison of gut microbial diversity among PT and PM treatment groups and AF control group on d3 and d7. A bigger Shannon index indicates a higher microbial diversity. The inside bar represents the median, the outer horizontal line of the box represents the 25th and 75th percentiles. Error bars represent the standard error. *Indicates significant differences in the Shannon index between PT and AF groups (P = 0.028) and between PM and AF groups (P = 0.008). For the abbreviation of group names, see Fig. 1 legend.

Figure 3

Principal coordinate analysis (PCoA) of microbial communities among PT and PM treatment groups and AF control group. (A) Distribution of all samples on d3; (B) Distribution of all samples on d7. Samples that are positioned close to each other are similar to each other in microbial compositions whereas samples that are positioned distantly from each other are distinct to each other in microbial compositions. AF, antibiotic free group; PT, piperacillin-tazobactam group; PM, combination of penicillin and moxalactam group.

Figure 4

Comparison of the relative abundance of bacterial communities among PT and PM treatment groups and AF group on d3 and d7. (A) Relative abundance at the phylum level. (B) Relative abundance at the genus level. Different phyla and genera are color-coded. AF, antibiotic free group; PT, piperacillin-tazobactam group; PM, combination of penicillin and moxalactam group.

Figure 5

Partial least squares-discriminant analysis (PLS-DA) score plots of metabolite profiles on d3 and d7. (A and B) Negative ion mode. (C and D) Positive ion mode. The ellipse area represents the 95% confidence interval by Hotelling T2 test. The red circle represents each sample in the antibiotic free group; the green triangle represents each sample in the piperacillin-tazobactam group; the blue square represents each sample in the combination of penicillin and moxalactam group.

Figure 6

Heatmap of significantly different metabolic pathways of the three study groups on d3 (A) and d7 (B). One-way ANOVA test was used for the comparison of metabolic pathways among the three groups (P < 0.05). The cluster analysis was based on the processes of metabolism. The activity score (AS) calculated by Pathway Activity Profiling (PAPi) ranged from −1 to 1. For the abbreviation of group names, see Fig. 1 legend.