β-lactam antibiotics are fundamental in treating most neonatal infections [13]. Among them, cefotaxime and ampicillin/sulbactam are two of the most commonly prescribed antibiotics for neonatal infections [14, 15]. However, antibiotic use is a double-edged sword: while it eliminates harmful pathogens, it also eradicates beneficial bacteria, disrupting the normal development of the gut microbiota. This disruption can lead to the overgrowth of opportunistic pathogens and increase the risk of disease onset [16,17,18]. The disruptive effect of antibiotics on the gut microbiota is more pronounced in neonates, especially within the first week [19]. Furthermore, the type and duration of antibiotic use have varying impacts on the gut microbiota [19, 20]. This study primarily investigates the effects of β-lactam antibiotics, specifically cefotaxime and ampicillin-sulbactam, on the gut microbiota of neonates. We also explored how these effects varied with neonates’ ages and were influenced by different treatment durations, types of antibiotics, delivery modes, and feeding modes.

First, we observed that β-lactam antibiotic administration significantly affected the composition of the gut microbiota. Specifically, the top 5 genera showed significant alterations after β-lactam antibiotic use, with strongly increased proportions of Klebsiella, Enterococcus and Streptococcus and dramatically decreased Escherichia-Shigella, Clostridium sensu stricto 1, Bifidobacterium (Fig. 4). Our findings support an earlier investigation, which found that newborns exposed to antibiotics in their first week of life had higher levels of Klebsiella and Enterococcus than the control group [19].

Klebsiella and Enterococcus, potential pathogens colonizing in neonatal gut, are important organisms for neonatal late onset sepsis [21]. A recent global neonatal sepsis observational cohort study involving 3,195 infants (90.4% neonates aged < 28 days) revealed that of the 17.7% blood culture pathogen positive, Klebsiella pneumoniae was the most common pathogen, accounting for 4.1% [22]. Additionally, Klebsiella spp are frequently enriched in the gut microbiota of preterm neonates, with overgrowth associated with necrotizing enterocolitis (NEC), nosocomial infections and late-onset sepsis [23]. Klebsiella/Enterococcus-dominated fecal microbiota is linked to an increased risk of developing NEC in preterm infants [24]. Klebsiella spp, through Toll-like receptor-4 activation, induce the recruitment of pro-inflammatory T helper 17 cells, resulting in the release of pro-inflammatory cytokines (IL-17, IL-22), leading to erythrocyte death, mucosal injury, and bacterial translocation to the microvasculature beneath the intestinal epithelium [24, 25]. Therefore, the increased abundance of the Klebsiella and Enterococcus induced by β-lactam antibiotic administration may predispose individuals to late infections. On the other hand, Streptococcus, which also significantly increased after β-lactam antibiotic treatment, is a group of Gram-positive bacteria with the potential to cause severe infections associated with significant morbidity and mortality [12]. It has been reported that late-onset neonatal bloodstream infections can be caused by enteric bacteria, including Streptococcus, which commonly resides within the mucosal lining of the intestinal tract and can disseminate to various organs, resulting in severe infections [26, 27]. Therefore, the overgrowth of Streptococcus in the intestine after β-lactam antibiotic treatment may increase the risk for subsequent infection, such as late-onset sepsis. Our results are largely consistent with previous findings, which showed that empirical antibiotic therapy increased harmful bacteria such as Streptococcus and Pseudomonas in preterm infants [12]. However, one inconsistency is that in our study, antibiotic use led to an increase in Streptococcus abundance, whereas in previous research found a decrease [28]. These variations may be due to differences in the sensitivity of Streptococcus to different β-lactam antibiotics or their resistance mechanisms. Specifically, previous studies used amoxicillin/ceftazidime [28] and a combination of penicillin and moxalactam or piperacillin-tazobactam [12], while our study used cefotaxime or ampicillin/sulbactam. Additionally, β-lactam antibiotic administration significantly reduced the abundance of genera such as Escherichia-Shigella, Clostridium sensu stricto 1, and Bifidobacterium. In term infants, an initially aerobic environment primarily hosts aerobes and facultative anaerobes, such as Escherichia, enterococci, Enterobacteriaceae, Staphylococcus, and Streptococcus species [29, 30]. As gut luminal oxygen levels rapidly fall due to consumption by these bacteria and secretory immunoglobulin A, strict anaerobes like Bifidobacterium and Clostridium proliferate [31]. However, antibiotic use significantly decreases the abundance of enteric anaerobic bacteria, including Bifidobacterium, enterobacteria and clostridia [19, 32]. Cefotaxime and ampicillin/sulbactam, broad-spectrum β-lactam antibiotics targeting Gram-positive and -negative bacteria, significantly impact gut microbiota composition. Bifidobacterium species are particularly sensitive to β-lactam antibiotics, and treatment with amoxicillin can greatly influence their composition in infant intestinal microbiota [4, 33]. Moreover, intravenous antibiotic combinations, such as penicillin + gentamicin, co-amoxiclav + gentamicin or amoxicillin + cefotaxime, significantly decreased the abundance of Bifidobacterium, and suggesting that antibiotic treatment may directly eliminate these genera [19]. Therefore, the deceased proportion of Bifidobacterium after cefotaxime or ampicillin/sulbactam treatment in the present study may be due to their direct elimination. However, the specific species affected by these treatments remain unknown and warrant further investigation. We also found that the overgrowth of Klebsiella, Enterococcus and Streptococcus may inhibit the growth of Escherichia-Shigella, as their relative abundances were negatively correlated (Fig. 5A, C, D). Conversely, certain bacteria exhibit synergistic growth effects, such as Klebsiella with Enterobacter, Enterococcus with Streptococcus, and Aeromonas with Streptococcus, which show positive correlations in abundance (Fig. 5B, E, F). Bacteria can inhibit each other’s growth through mechanisms such as spatial and nutritional competition [34]. For example, Klebsiella pneumoniae can produce bacteriocins with antimicrobial effects against closely related species [35]. Furthermore, Klebsiella can use the type VI secretion system to secrete and inject, killing surrounding bacteria and aiding in colonization [36]. On the other hand, the positive correlation between microbiota could result from competition for shared resources and nutrients in similar ecological niches or the possibility of a mutualistic relationship that promotes their growth [37]. Certainly, further research is necessary to clarify the potential mechanisms of interactions between microbial communities.

The composition of neonatal gut microbiota, influenced by antibiotic treatment, is determined by various factors, including the timing, duration, and specific type of antibiotics administered [10]. Our study assessed the effects of β-lactam antibiotic treatment on neonatal intestinal microbiota across specific age ranges: ≤ 7 days, 8–14 days, 15–21 days, and 22–28 days. Remarkably, we observed a significant increase in the abundance of Klebsiella and Enterococcus following β-lactam antibiotic treatment at all four stages compared to the healthy group. Our findings were consistent with a previous study where 147 infants born at ≥ 36 weeks of gestational age received intravenous antibiotic treatment (penicillin + gentamicin, co-amoxiclav + gentamicin or amoxicillin + cefotaxime) in the first week of life, resulting in increased abundance of Klebsiella and Enterococcus spp and decreased abundance of Bifidobacterium spp [19]. During the first days of life, the gut microbiota primarily consists of aerobic/facultative anaerobic bacteria belonging to the phyla Proteobacteria (e.g., Enterococcus spp) and Firmicutes (e.g., Staphylococcus, and Streptococcus) [31, 38]. The abundance of these facultative bacterial taxa decreases rapidly due to oxygen consumption and intestinal secretory immunoglobulin A, along with the expansion of anaerobic bacteria such as Bifidobacterium and Clostridium during the first months of life [31]. However, antibiotic use, one the most disruptive factors for neonatal gut microbiota development, strongly interferes with normal gut microbiota development at every neonatal period [39]. Our research provides more detailed insights into the influence of antibiotic usage on neonates of varying age groups, indicating that antibiotic use in neonates at any age can substantially impact their gut microbiota.

Additionally, the duration of antibiotic treatment significantly affects the structure of enteric microbiota. Rooney et al. suggested that 1 week of discontinuation of antibiotic treatment, each additional day of antibiotics was associated with lower richness of obligate anaerobes [20]. Zwittink et al. observed a significant reduction in Bifidobacterium levels in preterm infants (35 ± 1 week’s gestation) following short (≤ 3 days) or long (≥ 5 days) antibiotic treatment, which persisted until the third week after birth (P = 0.028). For long antibiotic treatments, this reduction continued until the sixth postnatal week (P = 0.009) [40]. These studies, along with our findings, suggest that longer durations of antibiotic use have a greater impact on the gut microbiota. Simultaneously, prolonged antibiotic treatment led to the emergence and overgrowth of antibiotic-resistant microbes [16, 41]. Therefore, it is crucial to minimize the duration of antibiotic use as much as possible while treating neonatal infections.

Moreover, the choice of antibiotics administered to neonates can distinctly affect the composition of their intestinal microbiota. We found that ampicillin/sulbactam significantly increased the richness of Enterobacter, Citrobacter, Lachnospiraceae_Unclassified, and Staphylococcales_Unclassified compared to cefotaxime, indicating that the impact of each antibiotic on neonatal microbiota is not uniform. This suggests that ampicillin/sulbactam may be more harmful to neonatal microbiota due to the enrichment of opportunistic pathogen [42]. A previous study revealed that broad-spectrum antibiotics for suspected early-onset neonatal sepsis, such as amoxicillin + cefotaxime, had the largest effects on microbial community composition and antimicrobial resistance gene profiles, whereas penicillin + gentamicin exhibited the least effects [19]. Antibiotic treatment, especially with broad-spectrum antibiotics, disrupts the gut microbiota and colonization resistance [43]. The choice of antimicrobials for neonatal infection depends on the most frequent causative microorganisms and is often empirical until culture results and antibiograms are available [44]. Ampicillin and gentamicin are the WHO’s first-line regimen for empiric antibiotic combinations, with cefotaxime as the second-line regimen [22]. Adding a β-lactamase inhibitor like sulbactam to β-lactam antibiotic like ampicillin can broaden the spectrum of activity to cover gram-negative extended-spectrum β-lactamase producers [45]. However, we found that ampicillin/sulbactam led to an increase in the richness of Enterobacter, Citrobacter. We speculated that its effect may be associated with the increased antibiotic resistance microbiota, but the exact mechanisms need further investigation.

Delivery mode significantly affects the colonization and development of neonatal intestinal microbiota [9, 46, 47]. Compared to vaginally delivered infants, those born by cesarean section showed decreased relative abundance of Bacteroides and Parabacteroides and enrichment of Clostridium_sensu_stricto_1, Enterococcus, Klebsiella, Clostridioides, and Veillonella [9]. We further found that, compared to vaginal delivery, β-lactam antibiotic treatment in cesarean-delivered infants further increased the abundance of Klebsiella, Enterobacteriaceae_Unclassified, Lactobacillales_Unclassified and Pectobacterium (Fig. 7C), indicating that β-lactam antibiotics seem to exacerbate intestinal flora disturbance caused by cesarean section. In addition, feeding mode also significantly affects neonatal intestinal composition [48, 49]. However, it did not alter the overall adverse effects of antibiotic use on intestinal flora, as the abundance of the main genera influenced by β-lactam antibiotic treatment, such as Klebsiella, Enterococcus and Streptococcus, showed no significant differences among artificial feeding, breast feeding, and mixed feeding groups. Only small proportion of generadiffered (Fig. 7D). These results suggest that regardless of previous feeding patterns, β-lactam antibiotic treatment significantly impacts the compositions of neonatal microbiota. Van Daele et al. indicated that antibiotic exposure in the first week perturbated fecal microbiota of term infants, with the perturbation still notable at one month in formula-fed infants, but only until two weeks in breast-fed infants [50]. The results suggests that breast feeding can help restore dysbiosis. Mechanically, breastmilk is abundant in bioactive components, including human milk oligosaccharides, immune cells, lactoferrin, cytokines, antibodies, and antimicrobial proteins and peptides, which aids restoration by stimulating the growth of bifidobacteria and reducing (potential) pathogens [29, 51]. A recent study showed that breast feeding and antibiotics have opposing effects on the infant microbiome, and that breast feeding enrichment of Bifidobacterium longum subsp. infants is associated with reduced antibiotic-associated asthma risk [52]. Taken together, while breast feeding may not prevent the adverse effects of antibiotic on the gut microbiota, it can help restore gut microbiota after antibiotic treatment.

Notable, the study has several limitations. It is limited to examining the effects of ampicillin/sulbactam and cefotaxime, preventing conclusions about which antibiotic treatment causes the least ecological damage or has the shortest duration of impact on the gut microbiome. Future research should explore a broader range of antibiotics to determine the therapies with minimal ecological impact.