Occurrence of Neonatal Necrotizing Enterocolitis in Premature Neonates and Gut Microbiota: A Case-Control Prospective Multicenter Study

Abstract

Background: Necrotizing enterocolitis (NEC) is still one of the leading causes of neonatal death. The present study reports the data from a French case-control prospective multicenter study.

Methods: A total of 146 preterm neonates (PNs) with or without NEC were included. Bacterial 16S rRNA gene sequencing was performed on stool samples (n = 103). Specific culture media were used to isolate Escherichia coli, Clostridium butyricum, and Clostridium neonatale, and strains were phenotypically characterized.

Results: The gut microbiota of PNs was dominated by Firmicutes and Proteobacteria, and five enterotypes were identified. The microbiota composition was similar between NEC cases and PN controls. However, differences were observed in the relative abundance of Lactobacillus genus, which was significantly lower in the NEC group, whereas that of the Clostridium cluster III was significantly higher (p < 0.05). Within enterotypes, several phylotypes were significantly more abundant in NEC cases (p < 0.05). Regarding perinatal factors, a statistical association was found between the gut microbiota and cesarean delivery and antifungal therapy. In NEC cases and PN controls, the carriage rates and virulence genes of uropathogenic E. coli were equivalent based on culture. No correlation was found between E. coli, C. butyricum, and C. neonatale carriages, beta-lactam resistance, and antibiotic treatment.

Conclusions: At disease onset, our data support a microbiota dysbiosis between NEC and control infants at the genus level. In addition, it provides valuable information on bacterial antimicrobial susceptibility.

Keywords: Clostridium butyricum; Clostridium neonatale; NEC; UPEC; antimicrobial susceptibility; gut microbiota; multicenter study; necrotizing enterocolitis; uropathogenic Escherichia coli.

Figure 1

ClosNEC cohort flow chart.

Figure 2

Specificities of the preterm gut microbiota. (A). Unsupervised cluster analysis based on gut microbiota composition described based on 97% sequence similarity and demonstrated by principal coordinate analysis. Microbiota can be stratified into five clusters (color coded). (B). Relative abundance of the major driving genera in each microbiota cluster. (C). Bacterial load expressed as CFU/g of stool samples based on qPCR analysis of the “All Bacteria” domain. (D). Bacterial diversity highlighted by the Simpson diversity index. (E). Cluster distribution as a function of age (in weeks of amenorrhea; WA) and body weight (in grams; g). * p-value < 0.05.

Figure 3

Comparison of the gut microbiota between the NEC cases and the control preterm neonates. (A). Richness (number of observed OTUs) and alpha-diversity (Simpson index) of the gut microbiota in NEC cases and control preterm neonates. No significant difference was found. (B). Bacterial genera are significantly different between the NEC cases and control preterm neonates. (C). Specific phylotypes (OTUs) significantly associated with the microbiota of NEC cases or controls based on LEfSe analysis (Linear discriminant analysis Effect Size). NEC: infants with necrotizing enterocolitis; CTRL: infants without NEC; C: Cluster; LDA: Linear discriminant analysis; Bacterial taxonomy is color coded as indicated; ** p-value < 0.001, * p-value < 0.05.

Figure 4

Impact of antifungal treatment on the preterm gut microbiota. (A). Antifungal treatments were significantly more frequent in preterm neonates with the lowest gestational age (in weeks) and birth weight (in grams). (B). Proportions of bacterial genera were significantly different between antifungal-treated and untreated PNs. **** p-value < 0.0001, ** p-value < 0.001, * p-value < 0.05; ns: non-significant; No: untreated group; Yes: Treated group.