Study of the gut microbiome profile in full-term infants with necrotizing enterocolits

Abstract

BACKGROUND

Necrotizing enterocolitis (NEC) is a severe gastrointestinal disease that primarily affects preterm. Recently, cases with similar clinical manifestations have been observed in full-term infants without Hirschsprung’s disease. Dysbiosis suggested to have role in development of NEC.

AIM

To compare the gut microbiome composition between full term infants with NEC and healthy controls and to evaluate the impact of feeding sources on microbial diversity.

METHODS

Ten full term infants with NEC and six matched healthy breastfed control infants were enrolled in the study. Triplicate stool samples were collected from the enrolled infants. Genomic DNA was isolated and subjected to polymerase chain reaction analysis.

RESULTS

Bacteroidetes and Bacteroides were more abundant in control infants than in NEC cases, although the differences were statistically insignificant (P = 0.118 and P = 0.147, respectively), with large and moderate effect sizes. Bifidobacteria levels were significantly greater when a relaxed threshold (P ≤ 0.1) was used in both the control and NEC breastfed groups than in the NEC formula-fed group (P = 0.098). Additionally, alpha diversity was significantly reduced at the 0.1 level in NEC patients, especially among formula-fed infants (P = 0.094).

CONCLUSION

Full-term Egyptian infants with NEC exhibit reduced microbial diversity and alterations in bacterial abundance, supporting a potential link between dysbiosis and NEC. Feeding practices, particularly breastfeeding, appear to influence the gut microbiome profile regardless of NEC status. Although Bacteroidetes and Bacteroides did not reach statistical significance, their effect sizes suggest a need for further investigation into their roles in NEC pathogenesis in full-term infants.

Key Words: Necrotizing enterocolitis; Full-term infants; Gut microbiome; Bifidobacteria; Dysbiosis

Core Tip: The study aims to map the gut microbiome profile of full-term infants with non-Hirschsprung’s [necrotizing enterocolitis (NEC)] in Egypt and to relate the gut microbiome profile of non-Hirschsprung’s NEC in full-term infants to the source of the feeding factor. The study aims to compare the gut microbiome profile between NEC infants and healthy controls. The study was conducted in the Department of Pediatrics, Alexandria University.

INTRODUCTION

Necrotizing enterocolitis (NEC) is a life-threatening gastrointestinal condition that is most observed in preterm neonates, but cases have been also reported in full-term infants without Hirschsprung’s disease. It involves inflammatory processes within the intestine, resulting in bacterial translocation, tissue injury, and, in severe cases, necrosis of the intestinal wall can occur[1]. Globally, 10% of NECs occur in full-term infants[2]; however, more statistics should be obtained to determine the actual percentage in Egypt. Although the exact cause of NEC is unknown, it is thought to be a multifactorial condition with prematurity (< 32 weeks gestation), and very low birth weight (< 1500 g) are among the most common contributing factors. Furthermore, microbial dysbiosis of the infant gut has long been associated with NEC[3].

The gut microbiome is the entire genetic material derived from all microorganisms living in the human gastrointestinal system, it differs from “gut microbiota”, which refers to actual organisms[4,5]. This microbial ecosystem begins at birth and evolves with age, it is influenced by genetics, environment, nutrition, antibiotic usage, and even stress[6]. The gut microbiome has a major impact on our overall health. The microbiome plays an important role in immune system development, growth and gut maturation. It also helps in metabolism and digestion of nutrients, as well as the vitamins and amino acids synthesis[7].

The composition of the infant gut microbiota is heavily influenced by early-life exposures, including the mode of delivery and type of feeding. Natural birth and breastfeeding encourage healthy bacterial colonization, such as Bifidobacterium, Lactobacillus, and Prevotella, while changes in these parameters can cause microbial imbalances[8]. Such imbalances, known as dysbiosis, have been linked to several health conditions, including inflammatory and metabolic disorders. Consequently, the gut microbiome has been increasingly recognized as a crucial factor in NEC pathogenesis, with studies suggesting that microbial dysbiosis may contribute to intestinal inflammation and disease progression. Although dysbiosis of the gut microbiota is linked to several human diseases, until now, there has been insufficient information to link it with NEC, especially in full-term infants with non-Hirschsprung’s disease. Hirschsprung’s disease is a congenital disorder characterized by the absence of ganglion cells in the mesenteric and submucosal plexuses of the intestine, leading to functional bowel obstruction[9].

According to some articles, the pathophysiology of NEC can be explained in different ways; in particular, it is argued that the phylum Proteobacteria is the dominant microorganism in premature NEC infants and that it probably leads to deterioration in the defensive mechanisms of the infant gut[10]. As a result, complications such as peritonitis, sepsis, intraventricular hemorrhage, intestinal failure, bowel perforation, neurodevelopmental disorders, poor growth, stricture and dysmotility in the gastrointestinal tract, cholestasis and liver disease can occur in the short or long term[11]. Furthermore, studies had detected that abundance of Bacillota (firmicutes) and Bacteroidota are decreased in premature infants with NEC, which further explains the characteristics of the disease[10]. However, most studies and review articles did not include the cause of the disease in mature infants. To bridge this knowledge gap, this study aims to map the gut microbiome profile of full-term infants with non-Hirschsprung’s (necrotizing enterocolitis) in Egypt and to relate the gut microbiome profile of non-Hirschsprung’s NEC in full-term infants to the source of the feeding factor.

MATERIALS AND METHODS

This prospective case control study was carried out at Alexanderia University Children’s Hospital from February 2025 to June 2025. During the study period 114 children were presented to the emergency department with symptoms suggestive of enterocolitis (lethargy, poor feeding, vomiting, diarrhea, and abdominal pain). The current study included only full-term infants who were presented with enterocolitis. The following group of infants were excluded: (1) Hirschsprung’s NEC; (2) Children older than two years; (3) Premature children; and (4) Infants received antibiotics or intestinal antiseptics in the previous 14 days. After verifying this exclusion criteria, finally ten full term infants with NEC and six matched healthy breastfed control infants were enrolled in the study. The following data was collected from all included infants: (1) Basic demographic information; (2) Information about the infant’s feeding (whether breast or formula feeding, start of weaning and type of weaning food) and mode of delivery; and (3) Presenting symptoms in NEC group, such as their chief complaint and when the first symptom appeared. Triplicate stool samples were collected from each infant enrolled in the study. The samples were kept in sterile containers at -20 °C upon defecation at the hospital in the case of diseased infants or at home in the case of the control group and delivered frozen to the Alexandria University Gut Microbiome Laboratory and stored at -80 °C for further processing[12].

The real-time polymerase chain reaction protocol will be performed as previously described[13]. Specific polymerase chain reaction primers were used to target selected phyla, genera, or species constituting the gut microbiota (e.g., Bacteroides, Prevotella, Ruminococcus, Firmicutes, Bacteroidetes, Lactobacilli, Bifidobacteria, Akkermansia muciniphila, Faecalibacterium prausnitzii, Clostridium difficile, Bacteroides fragilis, Escherichia coli, Lactobacillus plantarum, Lactobacillus reuteri, and Bifidobacterium longum)[14-22] (Table 1). In addition, a broad-range primer targeting the conserved 16S rRNA sequence of total bacteria was used, the amplification of which served as the denominator against which the amplification of other bacteria was estimated[14-22] (Table 1).

Table 1 Primers used in the present study.

Target	Primer name	Primer sequence (5’-3’)	Ref.
Lactobacillus plantarum	L. plantarum; FL. plantarum R	ATTCATAGTCTAGTTGGAGGT; CCTGAACTGAGAGAATTTGA	[14]
Lactobacillus reuteri	L. reuteri F; L. reuteri R	GAAGATCAGTCGCAYTGGCCCAA; TCCATTGTGGCCGATCAG	[14]
Bifidobacterium longum	B. longum F; B. longum R	CAG TTG ATC GCA TGG TCT T; TAC CCG TCG AAG CCA C	[14]
Total bacteria	UnivF; UnivR	TCCTACGGGAGGCAGCAGT; GGACTACCAGGGTATCTATCCTGTT	[15]
Akkermansia muciniphila	AM1-F; AM2-R	CAG CAC GTG AAG GTG GGG AC; CCT TGC GGT TGG CTT CAG AT	[16]
Bacteroides	B3F; B3R	CGATGGATAGGGGTTCTGAGAGGA; GCTGGCACGGAGTTAGCCGA	[17]
Bacteroidetes	Bact934F; Bact1060R	GGARCATGTGGTTTATTCGATGAT; AGCTGACGACAACCATGCAG	[17]
Firmicutes	Firm934F; Firm1060R	GGAGYATGTGGTTTAATTCGAAGCA; AGCTGACGACAACCATGCAC	[17]
Bacteroides fragilis	Bact F; Bact R	GAA AGC ATT AAG TAT TCC ACC TG; CGG TGA TTG GTC ACT GAC A	[18]
Bifidobacterium	Bif-F; Bif-R	TCGCGTC(C/T)GGTGTGAAAG; CCACATCCAGC(A/G)TCCAC	[19]
Faecalibacterium prausnitzii	FPR-2F; Fprau645R	GGAGGAAGAAGGTCTTCGG; AATTCCGCCTACCTCTGCACT	[19]
Prevotella	PrevF; PrevR	CACCAAGGCGACGATCA; GGATAACGCCYGGACCT	[17]
Lactobacilli	Lacto-F; Lacto-R	AGCAGTAGGGAATCTTCCA; CACCGCTACACATGGAG	[19]
Ruminococcus	Rflbr730F; Clep866mR	GGCGGCYTRCTGGGCTTT; CCAGGTGGATWACTTATTGTGTTAA	[20]
Clostridioides difficile	C. diff F; C. diff R	TTGAGCGATTTACTT CGGTAAAGA; TGTACTGGCTCACCTTTGATATTCA	[21]
Escherichia coli	E. coli F; E. coli R	CATGCCGCGTGTATGAAGAA; CGGGTAACGTCAATGAGCAAA	[22]

To evaluate the environmental sustainability of our laboratory methods, we conducted a carbon footprint analysis via an online tool[23,24]. Relevant data regarding energy use, materials consumed, and laboratory activity were input to estimate the associated greenhouse gas emissions, expressed in metric tons of CO₂ equivalent. This assessment aimed to quantify the carbon impact of the methods employed in our study.

RESULTS

The current study included ten infants with NEC (mean age of 12 months) and six controls (mean age of 12.6 months). Females made up to 60% of the NEC group, while both sexes were equally distributed in control group. 90% of the NEC group delivered via cesarean section, compared to 50% in the control group. Their feeding pattern varied; breast feeding was the predominant practice in all control groups, but only 60% of NEC infants were nursed. Weaning was uncommon in both groups (one infant each). Notably, all NEC infants had diarrhea, and vomiting was present in most of the majority (90%). Furthermore, 60% of cases experienced fever (Table 2).

Table 2 Demographic and clinical data of the participants.

	NEC	Control
	10	6
Age (months)
Minimum-maximum	3-19	3-21
Mean	12.1	12.6
Sex
Male	4	3
Female	6	3
Mode of delivery
Normal	1	3
Cesarean section	9	3
Feeding
Breast-fed	6	6
Formula-fed	4	0
Weaning	1	1
Food introduction	9	5
Age of introduction (months)
Minimum-maximum	4-12	6-8
Mean	7.11	6.8
Eating duration (months)
Minimum-maximum	1-11	3-14
Mean	6	7.8
Type of food
Rice	6	3
Boiled potato	5	3
Boiled zucchini	3	3
Boiled carrot	1	1
Full cream milk	1	1
Yogurt	2	2
Cerelac (wheat)	2	2
Biscuit	2	2
Homemade food	3	3
Cooked jute mallow	1	2
Boiled egg	1	2
Clinical presentation
Vomiting	9	N/A
Fever	6	N/A
Diarrhea	10	N/A
Anorexia	2	N/A
Lethargy	1	N/A
Dehydration	1	N/A

NEC: Necrotizing enterocolitis; N/A: Not applicable.

Gut microbiome analysis

Specific bacterial DNA is measured relative to total bacterial DNA in stool samples, rather than in absolute numbers. To ensure clarity with extremely small microbial numbers, the median relative abundance values via exponential notation were employed in reporting the results. A value of 6.83 × 10^-4 was represented as 6.83E-04 in data presentation. This standard provides clarity when working with extremely small microbiological numbers.

Phylum level analysis: Although there is no statistically significant difference between the two groups in terms of microbiome analysis, Firmicutes were abundant in NEC group (5.23E-01 vs 4.01E-01 in the control group), with a minor effect size [risk ratio (RBC) = -0.233]. More specifically, the Bacteroidetes tended to increase in the controls (3.48E-01 vs 1.10E-01 in NEC), with a large effect size (RBC = +0.5). Furthermore, while the Firmicutes/Bacteroidetes ratio was higher in NEC group, it was statistically insignificant (median 5.42 vs 1.25 in controls, P = 0.118); yet, it had a considerable size effect (RBC = -0.5) (Table 3, Figure 1).

Figure 1 Gut microbiome profile for all groups. NEC: Necrotizing enterocolitis.

Table 3 Comparison of the gut microbiome between necrotizing enterocolitis patients and controls.

Bacteria	NEC	Control	P value	Rank-biserial correlation
Firmicutes			0.492	-0.233
Minimum-maximum	1.89E-01-8.98E-01	1.97E-02-6.01E-01
Median	5.23E-01	4.01E-01
IQR	2.98E-01-6.92E-01	3.12E-01-5.49E-01
Bacteroidetes			0.118	+0.5
Minimum-maximum	1.52E-02-6.84E-01	4.47E-02-7.72E-01
Median	1.10E-01	3.48E-01
IQR	4.19E-02-1.82E-01	2.04E-01-5.45E-01
F/B			0.118	-0.5
Minimum-maximum	0.31-35.79	0.05-6.78
Median	5.42	1.25
IQR	2.31-10.84	0.52-3.25
Prevotella			0.713	+0.133
Minimum-maximum	1.47E-03-1.00E-01	3.06E-04-6.22E-01
Median	6.95E-03	8.33E-02
IQR	4.19E-02-1.82E-01	2.04E-01-5.45E-01
Bacteroides			0.147	+0.467
Minimum-maximum	5.66E-03-6.02E-01	1.64E-02-5.92E-01
Median	4.24E-02	2.17E-01
IQR	3.98E-03-2.86E-02	2.66E-03-2.31E-01
P/B			0.713	+0.133
Minimum-maximum	0.003-0.80	0.001-3.49
Median	0.48	0.51
IQR	0.17-0.64	0.07-1.08
Ruminococcus			0.956	+0.033
Minimum-maximum	0.00E+00-3.41E-02	0.00E+00-6.48E-03
Median	1.11E-03	2.80E-03
IQR	6.38E-06-5.42E-03	6.55E-04-4.43E-03
Lactobacilli			0.875	+0.067
Minimum-maximum	1.34E-02-4.45E-01	1.22E-04-4.87E-01
Median	1.55E-01	2.37E-01
IQR	6.43E-02-2.51E-01	3.79E-02-4.30E-01
Bifidobacteria			0.875	-0.067
Minimum-maximum	8.89E-04-5.58E-01	4.51E-03-1.95E-01
Median	1.24E-01	1.11E-01
IQR	3.92E-03-1.71E-01	9.32E-02-1.24E-01
Akkermancia muciniphila			0.669	-0.117
Minimum-maximum	0.00E+00-8.95E-02	0.00E+00-2.80E-03
Median	0.00E+00	0.00E+00
IQR	0.00E+00-1.17E-05	0.00E+00-0.00E+00
Faecalibacterium prausnitzii			0.492	+0.233
Minimum-maximum	5.37E-04-3.81E-01	0.00E+00-2.54E-01
Median	2.43E-02	8.53E-02
IQR	3.25E-03-3.47E-02	1.50E-02-1.84E-01
Clostridium difficile			-	-
Minimum-maximum	0.00E+00-0.00E+00	0.00E+00-0.00E+00
Median	0.00E+00	0.00E+00
IQR	0.00E+00-0.00E+00	0.00E+00-0.00E+00
Escherichia coli			0.428	-0.267
Minimum-maximum	1.59E-01-8.63E-01	2.15E-02-4.13E-01
Median	3.63E-01	2.84E-01
IQR	2.39E-01-5.90E-01	2.12E-01-3.77E-01
Bacteroides fragilis			0.697	+0.133
Minimum-maximum	0.00E+00-3.00E-01	0.00E+00-3.21E-01
Median	2.43E-03	1.54E-02
IQR	0.00E+00-4.54E-02	2.55E-03-2.63E-02
Lactobacillus reuteri			0.850	+0.05
Minimum-maximum	0.00E+00-1.73E-05	0.00E+00-1.64E-05
Median	0.00E+00	0.00E+00
IQR	0.00E+00-0.00E+00	0.00E+00-0.00E+00
Lactobacillus plantarum			0.947	+0.033
Minimum-maximum	0.00E+00-2.17E-02	0.00E+00-4.81E-03
Median	0.00E+00	0.00E+00
IQR	0.00E+00-2.82E-05	0.00E+00-3.86E-05
Bifidobacterium longum			0.745	-0.117
Minimum-maximum	0.00E+00-2.06E-01	0.00E+00-4.27E-02
Median	1.43E-02	1.64E-02
IQR	3.74E-04-8.09E-02	3.90E-03-2.58E-02
Diversity index			0.092a	+0.533
Minimum-maximum	1.00-1.94	1.60-1.95
mean ± SD	1.529 ± 0.286	1.787 ± 0.147
Median	1.49	1.8
IQR	1.43-1.76	1.68-1.90
Dissimilarity index			-	-
Minimum-maximum	32%-47%	-
mean ± SD	40.100% ± 4.701%	-
Median	40.5%	-

^aP ≤ 0.10 was considered statistically significant.

P value was calculated via the Mann-Whitney test. Rank-biserial correlation: ± 0.00-0.09 (irrelevant), ± 0.10-0.29 (small effect), ± 0.30-0.49 (moderate effect), and ≥ ± 0.50 (large effect)[23]. Negative sign in rank-biserial correlation: Higher rank in the necrotizing enterocolitis group. Positive sign in rank-biserial correlation: Higher rank in control group. NEC: Necrotizing enterocolitis; IQR: Inter quartile range; F/B: Firmicutes/Bacteroidetes; P/B: Prevotella/Bifidobacteria.

Genus level analysis: At the genus level analysis, no statistically, significant difference was identified between the two groups. Bacteroides were more abundant in the control group (2.17E-01) than in the NEC patients (4.24E-02), with a moderate effect size (RBC = +0.467). In contrast, the abundance of Prevotella, Ruminococcus, Lactobacilli, and Bifidobacteria in the control group was insignificant due to small effect sizes, as was the Prevotella/Bifidobacteria ratio. (Table 3, Figure 1).

Species level analysis: Among the beneficial bacteria, Akkermansia muciniphila, Faecalibacterium prausnitzii, and Bacteroides fragilis were not significantly different between the groups (P > 0.1), with small effect sizes. Similarly, the probiotic bacteria Lactobacillus reuteri, Lactobacillus plantarum, and Bifidobacterium longum exhibited no significant changes (P > 0.1), with small effect sizes. Clostridioides difficile was undetectable in both groups, precluding statistical analysis. Additionally, Escherichia coli was more abundant in NEC patients than in controls (3.63E-01 vs 2.84E-01, P = 0.428), with a small effect size (RBC = -0.267), although this difference did not reach statistical significance (Table 3, Figure 1).

Alpha diversity: The Shannon diversity index revealed notable differences in microbial diversity between NEC patients and controls. The mean diversity was lower in NEC patients (1.529 ± 0.286) than in controls (1.787 ± 0.147), and this difference was statistically significant (P = 0.092) using relaxed threshold (P ≤ 0.1) with a large effect size (RBC = +0.533) (Table 3).

Dissimilarity index: Bray-Curtis dissimilarity index analysis revealed important differences in microbial community composition between NEC patients and controls. The mean dissimilarity between the NEC and control groups was 40.1% ± 4.7%, with values ranging from 32% to 47% (Table 3).

Microbiome analysis between breast feeding and formula feeding in NEC and control group

Phylum level analysis: Firmicutes was not significantly different between the groups (P > 0.1), with only a weak effect size (ε² = 0.038), although it was more abundant in the NEC formula-fed group than in the NEC breastfed and control groups. More notably, Bacteroidetes presented a moderate trend toward greater abundance in the control group than in the NEC breastfed and NEC formula-fed groups, with this difference approaching but not reaching statistical significance (P = 0.148) and showing a moderate effect size (ε² = 0.14) (Table 4, Figure 1). The Firmicutes/Bacteroidetes ratio was markedly greater in NEC patients, especially NEC formula-fed patients, than in controls and NEC breastfed patients, with these differences approaching but not reaching significance (P = 0.148) and demonstrating a moderate effect size (ε² = 0.14).

Table 4 Comparison of the gut microbiome between breast- and formula-fed necrotizing enterocolitis patients and controls.

Bacteria	NEC breastfed	NEC formula-fed	Control	P value	Epsilon squared (ε2)
Firmicutes				0.288	0.038
Minimum-maximum	1.89E-01-6.97E-01	3.17E-01-8.98E-01	1.97E-02-6.01E-01
Median	3.96E-01	6.23E-01	4.01E-01
IQR	2.32E-01-6.32E-01	4.87E-01-7.50E-01	3.12E-01-5.49E-01
Bacteroidetes				0.148	0.14
Minimum-maximum	4.14E-02-6.84E-01	1.52E-02-1.86E-01	4.47E-02-7.72E-01
Median	1.23E-01	9.12E-02	3.48E-01
IQR	5.18E-02-4.04E-01	3.34E-02-1.54E-01	2.04E-01-5.45E-01
F/B				0.148	0.14
Minimum-maximum	0.31-11.52	1.70-35.79	0.05-6.78
Median	4.34	12.04	1.25
IQR	1.48-7.74	5.14-22.29	0.52-3.25
Prevotella				0.749	0.00
Minimum-maximum	1.47E-03-3.41E-02	3.70E-03-1.00E-01	3.06E-04-6.22E-01
Median	6.95E-03	2.14E-02	8.33E-02
IQR	4.05E-03-1.08E-02	4.53E-03-5.35E-02	2.66E-03-2.31E-01
Bacteroides				0.299	0.032
Minimum-maximum	5.66E-03-6.02E-01	1.31E-02-1.51E-01	1.64E-02-5.92E-01
Median	4.00E-02	5.12E-02	2.17E-01
IQR	1.35E-02-3.41E-01	2.98E-02-8.81E-02	1.86E-01-2.55E-01
P/B				0.894	0.00
Minimum-maximum	0.003-0.80	0.14-0.66	0.001-3.49
Median	0.49	0.42	0.51
IQR	0.11-0.66	0.25-0.59	0.07-1.08
Ruminococcus				0.754	0.00
Minimum-maximum	0.00E+00-3.05E-02	2.55E-05-3.41E-02	0.00E+00-6.48E-03
Median	4.41E-04	1.45E-03	2.80E-03
IQR	0.00E+00-5.25E-03	1.01E-03-9.70E-03	6.55E-04-4.43E-03
Lactobacilli				0.607	0.00
Minimum-maximum	2.62E-02-4.45E-01	1.34E-02-2.76E-01	1.22E-04-4.87E-01
Median	1.67E-01	9.06E-02	2.37E-01
IQR	1.54E-01-2.77E-01	3.19E-02-1.76E-01	3.79E-02-4.30E-01
Bifidobacteria				0.098a	0.203
Minimum-maximum	1.11E-02-5.58E-01	8.89E-04-1.82E-01	4.51E-03-1.95E-01
Median	1.37E-01	1.34E-03	1.11E-01
IQR	1.18E-01-3.59E-01	1.09E-03-4.66E-02	9.32E-02-1.24E-01
Akkermancia muciniphila				0.631	0.00
Minimum-maximum	0.00E+00-8.95E-02	0.00E+00-7.31E-05	0.00E+00-2.80E-03
Median	0.00E+00	7.80E-06	0.00E+00
IQR	0.00E+00-0.00E+00	0.00E+00-3.00E-05	0.00E+00-0.00E+00
Faecalibacterium prausnitzii				0.344	0.01
Minimum-maximum	1.14E-03-3.81E-01	5.37E-04-2.69E-02	0.00E+00-2.54E-01
Median	3.26E-02	1.32E-02	8.53E-02
IQR	9.10E-03-1.84E-01	3.79E-03-2.29E-02	1.50E-02-1.84E-01
Clostridium difficile				-	-
Minimum-maximum	0.00E+00-0.00E+00	0.00E+00-0.00E+00	0.00E+00-0.00E+00
Median	0.00E+00	0.00E+00	0.00E+00
IQR	0.00E+00-0.00E+00	0.00E+00-0.00E+00	0.00E+00-0.00E+00
Escherichia coli				0.656	0.00
Minimum-maximum	1.59E-01-8.63E-01	1.82E-01-8.49E-01	2.15E-02-4.13E-01
Median	3.87E-01	3.58E-01	2.84E-01
IQR	2.44E-01-5.90E-01	2.23E-01-5.71E-01	2.12E-01-3.77E-01
Bacteroides fragilis				0.898	0.00
Minimum-maximum	0.00E+00-3.00E-01	0.00E+00-5.40E-02	0.00E+00-3.21E-01
Median	9.70E-03	2.43E-03	1.54E-02
IQR	0.00E+00-1.23E-01	5.19E-05-1.71E-02	2.55E-03-2.63E-02
Lactobacillus reuteri				> 0.99	0.00
Minimum-maximum	0.00E+00-0.00E+00	0.00E+00-1.73E-05	0.00E+00-1.64E-05
Median	0.00E+00	0.00E+00	0.00E+00
IQR	0.00E+00-0.00E+00	0.00E+00-4.33E-06	0.00E+00-0.00E+00
Lactobacillus plantarum				0.831	0.00
Minimum-maximum	0.00E+00-2.17E-02	0.00E+00-3.76E-05	0.00E+00-4.81E-03
Median	0.00E+00	0.00E+00	0.00E+00
IQR	0.00E+00-6.62E-04	0.00E+00-9.40E-06	0.00E+00-3.86E-05
Bifidobacterium longum				0.413	0.00
Minimum-maximum	0.00E+00-2.06E-01	1.47E-05-8.79E-02	0.00E+00-4.27E-02
Median	3.73E-02	5.01E-04	1.64E-02
IQR	1.41E-02-1.16E-01	1.89E-04-2.25E-02	3.90E-03-2.58E-02
Diversity index				0.094a	0.209
Minimum-maximum	1.47-1.94	1.00-1.78	1.60-1.95
mean ± SD	1.648 ± 0.194	1.350 ± 0.334	1.787 ± 0.147
Median	1.61	1.31	1.8
IQR	1.49-1.76	1.15-1.51	1.68-1.90
Dissimilarity index				-	-
Minimum-maximum	32.00%-45.00%	37.00%-47.00%	-
mean ± SD	40.000% ± 5.215%	40.250% ± 4.573%	-
Median	42.5%	38.5%	-

^aP ≤ 0.10 was considered statistically significant.

P value was calculated via the Kruskal-Wallis test. Epsilon squared (ε²): < 0 (no effect), 0.00-0.01 (negligible), 0.01-0.04 (weak effect), 0.04-0.16 (moderate effect), 0.16-0.36 (relatively strong effect), 0.36-0.64 (strong effect), 0.64-1.00 (very strong effect) based on Rea and Parker[24] in 2014. NEC: Necrotizing enterocolitis; IQR: Inter quartile range; F/B: Firmicutes/Bacteroidetes; P/B: Prevotella/Bifidobacteria.

Genus level analysis: There was no significant difference in Bacteroides abundance between the groups (P = 0.299), with a negligible effect size (ε² = 0.032). The abundance of Bifidobacteria significantly differed (P = 0.098) at the 0.1 level, with a relatively strong effect size (ε² = 0.203), with a greater abundance in both the control (1.11E-01) and NEC breastfed groups (1.37E-01) than in the NEC formula-fed group (1.34E-03). In contrast, the abundances of Prevotella, Ruminococcus, and Lactobacillus were not significantly different and had negligible effect sizes (Table 4, Figure 1). The Prevotella/Bifidobacteria ratio was not significantly different, with a negligible effect size.

Species level analysis: Akkermansia muciniphila, Faecalibacterium prausnitzii, Escherichia coli, Bacteroides fragilis, Lactobacillus reuteri, Lactobacillus plantarum, and Bifidobacterium longum were not significantly different between the groups (all P > 0.1), with negligible effect sizes (all ε² < 0.01). Clostridioides difficile was undetectable in all groups, precluding statistical analysis (Table 4, Figure 1).

Alpha diversity: The Shannon diversity index revealed statistically significant differences at the 0.1 level between groups (P = 0.094), with a relatively strong effect size (ε² = 0.209). Diversity was markedly lower in NEC patients (breastfed mean: 1.648; formula-fed mean: 1.350) than in controls (mean: 1.787). The formula-fed NEC group presented the most pronounced diversity reduction, with effect size larger than the threshold for strong clinical relevance (ε² > 0.16) (Table 4).

Dissimilarity index: Bray-Curtis dissimilarity index analysis revealed distinct microbial community differences between the NEC groups and the control groups. The mean dissimilarity was 40.0% ± 5.2% between NEC breastfed infants and controls and 40.3% ± 4.6% between NEC formula-fed infants and controls, with values ranging from 32% to 47% across all comparisons (Table 4).

Green assessment

The carbon footprint of our laboratory methods was estimated at 0.21 metric tons of CO₂ equivalent. This value reflects the total greenhouse gas emissions associated with the laboratory activities involved in this study. Compared with common emission benchmarks, such as emissions from short-distance travel or household electricity use, this value represents a modest environmental impact. These findings suggest that the methods employed are relatively sustainable from a carbon emissions perspective.

DISCUSSION

The pathogenesis of NEC is closely linked to the composition of the intestinal microbiota. Gut microbiota development in infants is shaped by early-life factors such as delivery mode and feeding practices[8]. Dysbiosis that represent abnormal colonization patterns, characterized by a predominance of pathogenic bacteria had been implicated in the development of NEC. That disrupts the delicate balance required for intestinal homeostasis, leading to inflammation and tissue injury[25].

The present study aimed to map the gut microbiome profile of full-term infants with non-Hirschsprung’s NEC in Egypt and to relate the gut microbiome profile of non-Hirschsprung’s NEC in full-term infants to the source of the feeding factor. In the present study, the majority (90%) of NEC infants were delivered by the cesarean section, compared to 50% in control group. All control infants were breastfed while 60% of NEC infants were breastfed. All our patients had late-onset NEC. Matched results were also reported by Li et al[26], where 39.9% of the cases studied had vaginal delivery and 70.8% (179/253) of cases were formula-fed. However, a study by Abbo et al[27] higher prevalence (70%, n = 19/27) of vaginal delivery among full-term patients with NEC With respect to their feeding, 20 patients had enteral feeding; 70% of them were breastfed. In our study, weaning was uncommon in both groups (one infant each). Difference in the mode of delivery could affect bacterial colonization so cesarean section could be a risk factor for NEC, but further studies are needed to study risk factors for NEC in full term infants.

The development of microbiomes in full-term neonates has been well described in a study including 900 neonates from 6 centers in Europe and the United States. They outlined three phases of microbiome development. Neonatal gut microbiome development progresses through distinct stages, beginning with the dominance of facultative anaerobes such as Escherichia coli and streptococci in the first two weeks, which helps establish the anaerobic environment needed for subsequent colonization by obligate anaerobes such as Bifidobacterium and Bacteroides species. Anaerobic G+ cocci (peptococci, peptostreptococci) and Bacteroides species become more diverse in the last stage[28].

With respect to our results of the gut microbiome analysis, Bacteroidetes appeared more abundant in the control group (3.48E-01 vs 1.10E-01 in NEC), showing a clear trend that, while not statistically significant (P = 0.118), was marked by a large effect size (RBC = +0.5), suggesting potential biological relevance. Similarly, the Firmicutes-to-Bacteroidetes ratio was elevated in NEC patients (median 5.42 vs 1.25 in controls), but this was insignificant (P = 0.118), despite a large effect size (RBC = -0.5), indicating a meaningful shift in microbial balance that warrants further investigation. Moreover, Bacteroides were the most predominant but did not reach statistical significance (P = 0.147), with a moderate effect size (RBC = +0.467) favoring higher median abundance in controls (2.17E-01) than in NEC patients (4.24E-02), suggesting a potential but inconclusive role in NEC-associated dysbiosis.

In contrast, the NEC group presented a greater median abundance of Firmicutes than the control group did (5.23E-01 vs 4.01E-01); however, this difference did not reach statistical significance (P = 0.492) and was associated with a small effect size (RBC = -0.233). Additionally, the other detected species, Prevotella, Ruminococcus, Lactobacilli and Bifidobacteria, were not significantly different (P > 0.1), regardless of their small effect sizes. A meta-analysis conducted by Pammi et al[10] characterized the microbial composition of preterm infants following the onset of NEC revealed a notable increase in the relative abundance of Proteobacteria and a concurrent decrease in both Firmicutes and Bacteroidetes. While in control infants there was a reduction in Proteobacteria and an increase in Firmicutes and Bacteroidetes. Furthermore, the study indicated that the onset of NEC coincided with this microbial shift, typically occurring at approximately 30 weeks of corrected gestational age[10].

A comparison of the gut microbiomes of the breast and formula-fed NEC patients and control revealed that Bifidobacteria demonstrated statistically significant differences (P = 0.098) using a relaxed threshold (P ≤ 0.1) with a relatively strong effect size (ε² = 0.203), which was greater in both the control group (1.11E-01) and the NEC breastfed group (1.37E-01) than in the NEC formula-fed group (1.34E-03), suggesting that breastfeeding supports the colonization of Bifidobacteria in the infant gut. Masi et al[29] reported no correlation between the mother’s own milk microbiota and NEC development compared with the mother’s own milk of healthy preterm infants via 16S rRNA gene sequencing.

The current study has some limitations; small sample size (10 NEC cases and 6 controls) limits the results’ statistical power and applicability, making it more difficult to draw firm conclusions. The effect of confounding factors related to delivery mode and differences in disease stages not studied in the current study.

Moreover, Sample collection was laborious, particularly when dealing with very sick infants, which occasionally delayed processing and made logistics more difficult. The use of more comprehensive techniques, such as metagenomic analysis or next-generation sequencing, which could provide a deeper understanding of microbial function, was further restricted by financial constraints. Furthermore, while targeted real-time polymerase chain reaction makes it possible to quantify specific bacterial taxa, its taxonomic range is too small to detect less prevalent or ecologically important microbial groups. Finally, short duration of study precluded the evaluation of seasonal effects that might have altered the composition of the microbiome. Despite its limitations, this study provides important preliminary clues regarding the association between intestinal dysbiosis and NEC in full-term infants.

CONCLUSION

This research represents one of the initial comprehensive analyses of the gut microbiome composition of full-term Egyptian infants with non-Hirschsprung’s NEC. NEC patients, especially those who received milk formula, presented a significant decrease in alpha diversity at the 0.1 level, providing preliminary clues regarding the association between intestinal dysbiosis and NEC in full-term infants. Although several specific bacterial genera and species showed no significant differences, the overall microbiome profile, especially between breastfed and formula-fed NEC patients, indicates that feeding practices could affect microbial development regardless of NEC. The NEC patients presented lower levels of Bacteroidetes and Bacteroides and an increased Firmicutes/Bacteroidetes ratio, suggesting a microbial imbalance.