Probiotic Supplementation during the Perinatal and Infant Period: Effects on gut Dysbiosis and Disease

The perinatal period is crucial to the establishment of lifelong gut microbiota. The abundance and composition of microbiota can be altered by several factors such as preterm delivery, formula feeding, infections, antibiotic treatment, and lifestyle during pregnancy. Gut dysbiosis affects the development of innate and adaptive immune responses and resistance to pathogens, promoting atopic diseases, food sensitization, and infections such as necrotizing enterocolitis (NEC). Recent studies have indicated that the gut microbiota imbalance can be restored after a single or multi-strain probiotic supplementation, especially mixtures of Lactobacillus and Bifidobacterium strains. Following the systematic search methodology, the current review addresses the importance of probiotics as a preventive or therapeutic tool for dysbiosis produced during the perinatal and infant period. We also discuss the safety of the use of probiotics in pregnant women, preterm neonates, or infants for the treatment of atopic diseases and infections.


Introduction
Recent studies have indicated that microbiota colonization of the human body starts during pregnancy, altering the paradigm of the fetus as a sterile organism [1][2][3]. Microbial species such as Staphylococcus and Bifidobacterium have been identified in the meconium of neonates [4], the placenta (Escherichia, Shigella, Propionibacterium, and Enterobacteriaceae) [5], and the amniotic fluid (Streptococcus spp. (several species) and Fusobacterium nucleatum) [6] of healthy pregnancies. The microbiota is established in early pregnancy and varies depending on maternal nutritional habits, infections, and gestational age. Furthermore, the delivery mode as well as breastfeeding or formula feeding strongly influences the abundance and diversity of infant microbiota, which modulates the immune system response. For that, the standard profile of healthy infant microbiota is the fecal microbiota of Materials: Methodology): Section 3.1. "(pregnancy AND microbiome) AND (colonization OR development)"; Section 3.2. "(Microbiota OR dysbiosis) AND (Intrapartum antibiotic prophylaxis OR neonatal antibiotic therapy OR neonatal antimicrobial therapy)"; Section 3.3. "(Dysbiosis) AND (microbiota) AND (allergy OR disease OR food allergy OR atopic dermatitis OR asthma) AND (bifidobacterium)"; Section 3.4. "Probiotics OR bifidobacterium) AND (human milk OR breastfeeding OR breast milk) AND (transfer OR translocation)" Section 3.5. "(Food Hypersensitivity OR Food allergy) AND Probiotics AND (Lactobacillus OR Bifidobacterium)"; Section 3.6. "Probiotics AND (asthma OR wheezing OR rhinitis) AND prevention"; Section 3.7. "(Probiotics OR paraprobiotics) AND (preterm infant OR preterm neonate)"; Section 3.8. "Probiotics AND safety AND (pregnancy OR newborn/infant)".
Inclusion criteria were papers written in English and Spanish (with no geographical restrictions) published from 1 January 2004 to 15 April 2020; the presence of the selected terms in the title or as keywords; the use of a PICOS (patient, intervention, comparators, outcome, and study design) approach; original research performed in humans; selected experimental designs including clinical trials, case-control, longitudinal cohort, and cross-sectional studies. Sample size ≥10. In the section related to preterm infants, eligible studies had to include exclusively neonates ≤32 weeks or birth weight ≤1500 g. The quality of controlled studies referring to randomized, nonrandomized, and pre-post treatment was critically appraised following the Cochrane Collaboration's Risk of Bias Tool [18]. Exclusion criteria were non-pregnant adult population, children older than two years of age, and neonates with significant birth defects. Interventions using only prebiotics or immunotherapy were also excluded (see details in the Supplementary Materials: Methodology).
The selection of original manuscripts started by screening titles and abstracts for inclusion, creating a reference list of relevant papers for the topics explored in this review. Two investigators (A.-F.V. and N.-T.E.) conducted each stage of the studies selection, deleted duplicate inputs and reviewed studies as excluded or requiring further assessment. All data were extracted by one investigator (A.-F.V.) and cross-checked by a second investigator (N.-T.E.). In case of discrepancies in the selected studies, we opted for reconciliation through team discussion. The information obtained from each study was: first author, experimental design, number of participants, and control groups; intervention period (prenatal and/or postnatal); dose/duration and strains of probiotics administered; and main outcomes/findings. The eligibility criteria followed the PICOS approach. Population: pregnant women, newborns and infants; intervention: any doses, strains or species of probiotics administered prenatally and/or postnatally within the first year of life; comparators: placebo or no probiotics; outcome: the primary outcome was allergies or food sensitization. All authors performed a critical appraisal for the studies selected following the inclusion criteria, also analyzing the methodology and key results.
The characteristics and number of participants, as well as the results, including strengths, weaknesses, conclusions, and biases, were evaluated.
However, some bias was expected due to the heterogeneous results observed in the literature selected; the different populations compared; the distinct health conditions; the reduced number of randomized trials in pregnant women; the use of questionnaires to detect allergies; and the small sample size observed in some studies. Finally, the studies indicated in Figure 1 for each section of this review were identified through database searching and other sources. Furthermore, the following studies were evaluated and selected after meeting the inclusion criteria, the application of the exclusion criteria, and an eligibility assessment: Section 3.1 (35); Section 3.2 (15); Section 3.3 (16); Section 3.4 (20); Section 3.5 (12); Section 3.6 (10); Section 3.7 (57); Section 3.8 (21).

Prenatal Development of the Microbiome and Early Colonization
Human microbiome colonization can be understood as a progressive process. In puberty and adulthood, the microbiota shows a higher diversity than in newborns [19]. Avershina et al. analyzed the microbiome of stool samples from a cohort of 86 mothers and their children, concluding that the personal diversity of microbiota increased according to the age of the subjects. However, interindividual diversity decreased with age, being more individually diverse among newborns [20].
The prenatal colonization in newborns is under study. Some authors indicate that the process may be initiated by microbiota located in the placenta and amniotic fluid. Collado et al. analyzed microbiota from placenta and amniotic fluid samples obtained from 15 full-term neonates born by caesarean section. Their main findings were the low diversity of the microbiota as well as the predominance of Proteobacteria. This microbiome showed common features with the microbiome of the meconium in neonates, suggesting that the colonization process was initiated prenatally [5].
Tapiainen et al. analyzed 212 first-pass meconium samples of near-term and full-term newborns, sequencing regions of the bacterial 16S rRNA gene. The most abundant phyla were Firmicutes, Proteobacteria, and Bacteroidetes, with a relative abundance of 44%, 28%, and 15%, respectively. The diversity of the microbiome seemed to be influenced by the home environment but not by perinatal factors, leading to the conclusion that microbiota were not altered by immediate perinatal factors but by maternal factors before and during pregnancy, suggesting a transfer of the microbiome from the uterus to the fetal gut [21]. Moreover, some authors showed that maternal diabetes [22], diet [23,24], prenatal stress [25], and mother's weight [26,27] had an impact on the early gut microbiome [22][23][24][25][26][27]. However, De Goffau et al. found no evidence of the microbiome in 537 placental biopsies from complicated and uncomplicated pregnancies analyzed using various methods of DNA detection However, Streptococcus agalactiae, a pathogen that causes neonatal sepsis, was found in 5% of placental samples [28].

Gut Colonization
It is widely accepted that the early neonatal gut microbiome comes from maternal strains [29]. Makino et al. identified maternal monophyletic Bifidobacterium strains in the intestines of 11 out of 12 infants born vaginally [30]. The correlation between the early colonization and the delivery mode has been analyzed by many authors. In gut samples of newborns delivered vaginally, the Bifidobacterium genus was predominant (especially species such as Bifidobacterium longum and Bifidobacterium catenulatum) [31], followed by Bacteroides and enterobacteria [32]. The first colonizing bacteria enter the intestine through the oral cavity. In spite of some authors proposing that the birth canal microbiota is the most decisive modulator on infants' oral and gut microbiota, recent studies suggest that the microbiota of the oral cavity in the neonate might have a prenatal origin, preceding the birth canal exposure [33]. Although an imbalance in the birth canal can affect the neonate's oral microbiota. Li et al. studied whether vulvar disinfection with povidone iodine had an effect on the neonatal oral microbiota in 30 infants. Their results showed a lower presence of Lactobacillus and more opportunistic pathogens such as Staphylococcus, Klebsiella, and Escherichia in the disinfected group compared to the non-disinfected and C-section groups [34].
In relation to breastfeeding, some authors revealed that breastfed children had a high presence of Bifidobacterium in their gut and a low abundance of Clostridiales [35]. Conversely, formula-fed infants showed fewer Bifidobacteria and significantly higher proportions of Bacteroides, Clostridium coccoides and Lactobacillus groups [32]. Korpela et al. studied the presence of human milk oligosaccharides (HMO) in mothers with a functional fucosyltransferase 2 (FUT2) allele, and no differences in microbiota composition was observed in the vaginally born infants. However, in C-section-born infants (caesarean section), the functional FUT2 allele partially made up for the lack of microbiome diversity [37]. Similarly, Hill et al. showed that prolonged breastfeeding significantly increased the microbiota diversity of children born by caesarean section after 24 weeks, but had no effect on vaginal delivered infants [39].
The early microbiota are influenced by infant maturity at birth [39,[44][45][46]. Chernikova et al. analyzed 78 stool samples from premature infants and 189 samples from full-term infants and found that the extremely premature infants, delivered before 28 gestational weeks (GW), had lower bacterial intrapersonal diversity compared to babies born at 28-32 or 32-37 GW. They also observed that, in preterm infants, the proportion of Bifidobacterium, Bacteroides, and Streptococcus was decreased compared to full-term infants [47]. Forsgren et al. also found that the prevalence of Bifidobacterium differed in the gut microbiota between the 34-37 GW and >37 GW groups, with a delayed bifidobacteria colonization in late preterm infants [48].
Other factors have been proposed to explain the interindividual variability of the neonatal microbiome. Fallani et al. demonstrated that northern European countries were associated with higher proportions of Bifidobacterium in the feces of newborns, whereas more diverse microbiota with more Bacteroides were obtained in southern European countries [32]. Martin et al.'s analysis revealed that, after birth colonization, the microbiota can also be influenced by the presence of siblings and type of feeding. Interestingly, Bifidobacterium breve or Bifidobacterium longum subspecies infantis were early colonizers regardless of these factors. They found gender as an unexpected confounding factor, with girls being more quickly colonized by Lactobacillus in early life [41].

Respiratory Colonization
The respiratory microbiome has been less studied than the gut microbiome. Shilts et al. analyzed the nasal microbial community in 33 full-term infants, which was dominated by the Firmicutes, Actinobacteria, Proteobacteria, and Bacteroides lineages. They also found that the richness was higher in infants delivered by C-section and fed with formula compared to the vaginally delivered and breastfeeding group [49]. Stokholm et al. found no differences in the microbiome of hypopharyngeal aspirates depending on the delivery method [42]. On the other hand, a different study showed a delay in respiratory microbiota development, with a late colonization of commensal bacteria, such as Corynebacterium and Dolosigranulum, in infants born by C-section [50]. The airway microbiome at birth seems to be similar in preterm and full-term infants. However, infants which developed chronic lung disease showed reduced bacterial diversity at birth [51].

Skin
Recent studies have shown that 84% of healthy neonates had their skin colonized 24 h after birth [52], and the major bacterial growth was produced by coagulase-negative Staphylococcus. Newborns' skin barriers interact with microbiota and express certain antimicrobial peptides, such as cathelicidin antimicrobial peptide LL37, to inhibit Staphylococcus epidermidis growth. Thus, the constant host-microbe interaction contributes to the stability of the skin microbiota [53].
Moreover, Soeorg et al. showed that the skin of breastfed preterm neonates admitted to a neonatal intensive care unit was colonized with distinct Staphylococcus epidermidis strains compared to those found in breast milk. However, neonates gradually acquired strains genetically similar to those found in breast milk, similar to full-term neonates [54].
Nonantibiotic medications also showed an impact on the development of the gut microbiota. Nevertheless, there is still a lack of evidence in the neonatal cohort. Le Bastard et al. conducted a systemic review assessing the impact of proton pump inhibitors (PPIs), metformin, nonsteroidal anti-inflammatory drugs (NSAIDs), opioids, statins, and antipsychotics. PPIs and antipsychotic medication decreased, whereas opioids increased microbial diversity indices. PPIs decreased Clostridiaceae and increased Actinomycetaceae, Micrococcacceae, and Streptococcaceae, changes associated with gut dysbiosis and risk for Clostridium difficile infection. Enterococcaceae or Gammaproteobacteria (Enterobacter, E. coli, Klebsiella) counts were not increased [78]. In conclusion, IPA and postnatal antibiotic therapy affect the composition of the neonatal gut microbiota and may increase the risk for NEC and sepsis by Enterobacteriaceae predominance and reduction in protective phyla. Therefore, an antibiotic stewardship may be of utmost importance to reduce unnecessary and harmful antibiotic consequences. Future investigations should focus on: (1) the potential long-term effects of neonatal gut dysbiosis, (2) the effects of nonantibiotic medications in neonatal age, and (3) the definition of biomarkers of induced gut dysbiosis. Sustained effects on the gut microbiota by intensive antibiotic therapy in preterm infants. A brief course of antibiotics suppresses the microbiota diversity temporarily. Short-and long-term treatment with amoxicillin/ceftazidime during the first postnatal week drastically disturbs the normal colonization pattern. ST but not LT allows the recovery of Bifidobacterium levels in the first 6 w. Bifidobacterium dominance allows higher richness and diversity in gut microbiota. Prolonged antibiotic therapy affects the early development of gut microbiota in preterm infants. Antibiotic treatment generates a reduction in bacterial diversity and an enrichment of harmful bacteria such as Streptococcus and Pseudomonas.  IPA had a significant impact on the early gut microbial composition, which could partially be reversed after 30 days of life. No persistent microbiota differences at one year among infants exposed to IPA with elective CS or vaginal delivery.
IPA in C-section and vaginal delivery are associated with neonatal gut microbiota dysbiosis. IPA was associated with reduced microbiota richness and a depletion of Bacteroidetes and increased colonization of Enterococcus and Clostridium.
Breastfeeding modifies some of these effects. IPA modifies gut microbiota by reducing Bifidobacteria, which is further affected in infants receiving formula feeding. Long-term consequences require further investigation. week4: Shannon-Wiener index > 3.6 in all samples (high level of biodiversity). Increased Proteobacteria colonization (54%) in the treatment group compared to 37% in the control group (p = 0.0049). Bacteroidetes were detected in less than half of infants treated with antibiotics, notably low levels if present. Actinobacteria levels significantly lower in the antibiotic-treated group (3% vs. 24%; p = 0.00001).
Postnatal antibiotic therapy induces alterations in the gut microbiota, over eight weeks. The combined use of ampicillin and gentamicin in early life may have significant effects on gut microbiota, but the long-term health implications remain unknown. IPA affected all aspects of gut microbial ecology including species richness, diversity, community structure, and the abundance of colonizing bacterial genera.

Early Aberrant Microbiota and Its Effect on Pediatric Diseases
Allergic diseases occur at any stage of life, although some allergic manifestations, such as allergies to food, are most likely to develop during the first few years of life [79]. After three years of age, the prevalence of IgE, specific to inhalant allergen, becomes predominant [80]. Recent findings demonstrated that variables such as the antibiotic consumption during pregnancy, mode of delivery, feeding and mother's lifestyle during pregnancy strongly influence the neonatal gut microbiome modulating the development and function of the immune system in the host [23,64,81,82]. This microbiome will interact with receptors of the intestinal immune cells, causing the maturation of the intestinal mucosa and its gut-associated lymphoid tissues (GALT) [83]. GALT is functionally connected to the mesenteric lymph nodes and is able to identify pathogens from non-pathogenic microorganisms or antigens and defense against pathogens. Thus, a proper crosstalk between the immune system and microbiota will establish a Th1/Th2 balance, while an early developmental dysbiosis may underlie allergies or intolerances [84,85]. Moreover, the capacity to trigger the Th1 response is specially limited in neonates and infants, due to maternal IgGs in the placental barrier which partially protect the fetus during the last stages of pregnancy [86], and by secretory IgA (SIgA) in breast milk if breastfed. For that, a deficit of human milk intake, which contains several Igs including IgA, SIgA, IgM, secretory IgM, and IgG can produce low levels of SIgA at the intestinal barrier in infants [87]. A low diversity and abundance of bacterial populations promotes an impaired stimulation of SIgA, which targets an extensive number of gut bacteria modulating their growth [87,88] Therefore, reduced diversity of gut microbiota, low levels of mucosal IgA (total) and an aberrant IgA responsiveness to the gut microbiota during infancy are associated with the allergic diseases development [89]. Breast milk contains not only immunological components to protect infants against infections and allergies, but also human milk oligosaccharides (HMOs). These complex sugars stimulate the growth and/or activity of beneficial bacteria such as Bifidobacterium [90]. The genus Bifidobacterium represents one of the dominant bacterial groups in the gut microbiota during early life due to its ability to metabolize different forms of HMO [91]. In the context of allergic diseases, several studies based on murine and in vitro models have highlighted the potential role of Bifidobacterium in reducing inflammation through the production of anti-inflammatory cytokines and suppressing Th2 immune response and IgE production [92][93][94]. Moreover, Bifidobacterium produces short-chain fatty acids (SCFA) that decrease intestinal permeability and maintain the integrity of the intestinal barrier [95] to prevent the triggering of the immune response against antigens in the bloodstream [96]. In patients with atopic dermatitis (AD), the proportion of Clostridia, Clostridium difficile, Escherichia coli, and Staphylococcus aureus in the gut microbiome is higher than in healthy controls, and a reduction in short-chain fatty acid (SCFA)-producing bacteria (Bifidobacterium, Blautia, Coprococcus, Eubacterium, and Propionibacterium) is also observed [97,98]. Interestingly, a new observational study with 94,929 children from both genders showed that gastroenteritis (GE) during infancy could affect the intestinal microbiota in early life and increase rates of asthma, allergic rhinitis, and atopic dermatitis in later life (6 months-5 years). However, the authors did not perform a clinical evaluation of biological samples, thus further studies to find the association between early-infectious-GE, early-noninfectious-GE, and allergic disease are needed [99]. In a recent population-based study of 4.7 million people in British Columbia, the authors identified a 260% decrease in asthma incidence between 2000 and 2014 in young children, which correlated with a large decrease in antibiotic prescriptions. Moreover, antibiotic use in the first year of life was associated with around a doubled risk of asthma diagnosis at five years of age. Additionally, the authors identified a decrease in Faecalibacterium prausnitzii, Roseburia, and Ruminococcus bromii and an increase in Clostridium perfringens, associated with asthma and antibiotic exposure [100]. All these studies showed how intestinal dysbiosis could be a possible origin of future diseases in later stages of life.
Recently, and for the first time, a human study using culture-independent techniques to investigate the relationship between the mother's gut microbiota during pregnancy and allergic disease in the offspring showed that maternal carriage of Prevotella copri is associated, in a dose-response manner, with a decreased risk of food allergy during infancy [101]. This finding is relevant since, until now, food allergy (FA) has been related to a reduced bacterial diversity and an increased Enterobacteriaceae to Bacteroidaceae ratio. Vuillermin et al. (2020) presented a new during-pregnancy predictor of food allergy in offspring, Prevotella copri, probably associated with its important role in stimulating fetal immune development via the Toll-like receptor 4-dependent pathway and SCFA production [101].

Transfer of Probiotic Bacteria from Mother to Child
Despite human milk being classically considered sterile, irrefutable evidence has demonstrated that human milk contains a diverse bacterial community. Moreover, retrograde transfer of external bacteria into the mammary gland has also a strong role in milk inoculation during lactation [102]. Human milk from healthy women contains approximately 10 3 -10 5 CFU/mL (where CFU is colony-forming unit) and constitutes one of the main sources of bacteria to the breastfed infant gut. Although culture-dependent techniques have identified some key genera in breast milk, such as Staphylococcus, Streptococcus, Lactobacillus, and Bifidobacterium spp., culture-independent techniques, based on the amplification of the gene coding for bacterial 16S ribosomal RNA (rRNA), have allowed a more comprehensive assessment of the bacterial diversity in human milk. Thus, several studies have described a "core" microbiome of breast milk, consisting of Streptococcus, Staphylococcus, and Propionibacterium, although these genera vary depending on the population studied, the hypervariable region selected, and the milk extraction method used [103,104]. Despite this, potentially beneficial genera such as Lactobacillus and Bifidobacterium, widely used as probiotics in children for a wide variety of conditions, appear in most published studies, regardless of the use of next-generation sequencing (NGS) or culture-dependent techniques [104][105][106][107][108][109]. A high prevalence of these genera is found in the colostrum and milk following vaginal full-term deliveries [110]. Bacterial translocation from the digestive tract has been proposed as a source of bacteria for the mammary gland during late pregnancy and lactation. This route, called the enteromammary pathway, involves dendritic (DCs) and CD18+ cells, which would be able to take up nonpathogenic bacteria from the GI lumen through the tight junctions and, subsequently, carry them to other locations, including the lactating mammary gland, through the lymphatic system. Thus, this pathway implies a close communication between the gut microbiota and the immune system in all its stages [111,112]. Thus, the intake of probiotics during lactation could be a source of these beneficial bacteria in the infant, aiding the maturation of the intestinal epithelium and the neonatal immune system. Several studies have demonstrated the translocation of probiotic bacteria from the gastrointestinal tract to breast milk via the enteromammary pathway. Jimenez et al. (2008) showed that two probiotic strains isolated from human milk, L. salivarius CECT5713 and L. gasseri CECT5714, were present in the human milk of six out of 10 women after 30 days of oral intake, although L. salivarius CECT5713 appeared in a higher proportion [113]. Two years later, Arroyo et al. (2010) showed that not only L. salivarius CECT5713 but also L. fermentum CECT5716 appeared in milk samples after 21 days of oral intake [114].
Similar results were obtained with the L. salivarius PS2 strain, which was detected in 17 of 29 maternal milk samples from women who took the probiotic from week~30 of pregnancy until birth [115]. This strain also prevented infectious mastitis in this population, selected for having suffered recurrent episodes in previous pregnancies. Lactobacillus reuteri, a probiotic widely used to decrease colic in breastfed babies, has also been isolated in the colostrum of women who have taken it during the last four weeks of pregnancy, showing a significant presence compared to the placebo (12% vs. 2%). However, no difference between groups was observed in the prevalence of probiotics in breastmilk one month after delivery [116].
Although all these studies showed an increased presence of the administered bacteria in the breast milk of mothers, several studies did not demonstrate this enteromammary route in the studied strains. Simpson et al. (2018) found that maternal supplementation of Lactobacillus rhamnosus GG, Lactobacillus acidophilus La-5, and Bifidobacterium animalis ssp. lactis Bb-12 did not significantly affect the general breastfeeding-associated microbiota at 10 d or three months postpartum after four months of probiotic intake [117]. Based on these results, the authors concluded that strains isolated from breast milk would have a "natural affinity" and a greater ability to be transferred to the milk microbiota. Interestingly, Mastromarino et al. (2015) showed that women with vaginal delivery obtained higher amounts of lactobacilli and bifidobacteria in colostrum and mature milk compared to probiotic supplemented women who had a caesarean section. Thus, the authors established that the type of birth also influences the structure of the milk microbiota [118].
All these results demonstrate that there are many factors influencing the transfer of probiotics from mother to child through breast milk. Although the existence of an enteromammary route has been demonstrated, the strains isolated from breast milk reach the mammary glands more easily than other strains not present in breast milk. Significant differences in immunoregulatory factors such as cytokines and hormones are seen before childbirth in women with vaginal delivery compared to C-section [119]. Furthermore, a significant efflux of intestinal immune cells to the mammary glands during late pregnancy and lactation has been shown [120]. Thus, the type of delivery, as well as the time and duration of the treatment, will determine the greater or lesser translocation of the probiotic from the intestines to the breast milk. Moreover, the variability of sample collection and the DNA extraction and identification techniques of the strains can generate bias in the results. More studies are needed to determine the effects of specific probiotic strains on the breast milk composition.

Probiotics for the Prevention of Food Sensitization in Infants: Administration to Mothers Versus Infants
Food allergies (FA) have become a common problem that affects approximately 6% of infants under two and 9% of children aged 3 to 5 [79]. Eggs, milk and peanuts are the most common food allergens, and skin problems, such as eczema, are closely associated with FA [121,122]. The use of probiotics to prevent food reactions has gained popularity in clinical practice, considering that the gastrointestinal microbiota may modulate the mucosal physiology, the barrier function, and systemic immunologic and inflammatory responses. For that, the evaluation of probiotic supplementation in the prenatal and/or postnatal stage during the first months of life, based on randomized, double-blind, and placebo-controlled trials, is necessary to provide the latest evidence about food hypersensitivity in young children. A total of 10 studies published between 2004 and 2020 were included in the analysis: five of them referred to the supplementation of probiotics during the prenatal or pre-and postnatal stages (Table 2) and five only during the postnatal stage (Table 3). In reference to probiotic intake during the postnatal stage, four of the five studies failed to obtain significant results regarding the incidence of food allergy or allergen sensitization in children with a cow's milk allergy (CMA) or with a high risk of allergies. Intervention periods showed differences between studies, from treatments starting at birth [123] to probiotic administrations at ages up to one year, not showing administrations beyond a year and a half of age [124][125][126]. Strains, dose, mode of administration (mixed with water, food, formula, or infant cereals), and treatment period also differed among studies. Only one study performed on newborns with a high risk of allergy showed a significant decrease in sensitization to cow's milk (CM), due to using for one year a nonhydrolyzed formula fermented with Bifidobacterium breve C50 and Streptococcus thermophilus 065. A significant decrease in positive IgE tests against other foods (hen's eggs, codfish, wheat flour, soy flour, and peanuts) was also observed [127]. Interestingly, this study was the longest of the five postnatal studies. In addition, the formula contained nonhydrolyzed milk proteins. Probiotic strains were heat-inactivated after milk fermentation, that is, no live bacteria were used during intervention (named paraprobiotics). The authors attributed its effect to cell wall components such as peptidoglycans, which are thermoresistant and able to activate Toll-like receptor (TLR) 2. TLR2 activates the production of mediators such as IL-6, inducing IgA differentiation from naive B cells [128]. Regarding the effect of probiotic supplementation in both the prenatal and postnatal stages, probiotics were administered during the last weeks of pregnancy (weeks 32-36) and probiotic intake in newborns with a high risk of allergy was extended to six months in three of the four studies [129][130][131]. The exception was the study by Abrahamsson et al. (2007), which was prolonged up to one year after delivery [132]. Remarkably, all studies showed a lower sensitization to common food allergens in the probiotic group compared to the placebo, although only two showed statistical significance [129,131]. Both mothers and infants took the same probiotic mixture in all cases, and in only one study the mother also took probiotics after delivery [130]. Kim et al. supplemented a probiotic mix in Korean infants at high risk of food allergy, concluding that there were no changes in the frequency of positive food antigen-specific IgE sensitization and food allergies. However, sensitization against any one of the common food allergens (egg whites, cow's milk, wheat, peanuts, soybeans, and buckwheat) appeared to be lower in the probiotic (38.7%) than in the placebo group (51.7%). However, this study was limited by the high drop-out rate: there were only 31 individuals in the probiotic group and 29 individuals in the placebo group that were compared for sensitization and prevalence of food allergy [130]. Interestingly, Kuitunen et al. (2009) observed a significant decrease in atopic sensitization (positive food skin prick test (SPT) response and/or food-specific IgE >0.7 kU/L) in caesarean-delivered children compared to vaginally delivered children [131] after prenatal and postnatal probiotic treatment. Similar to Morisset et al., the authors inferred that the transient protection offered by probiotics against IgE-associated allergic diseases is based on the stimulation of Toll-like receptors. Additionally, in the subgroup of caesarean-delivered children, the authors noticed a delayed increase in bifidobacteria recovery in placebo-treated children.
Only one of the treatments in the prenatal and postnatal period was performed with a single strain. Lactobacillus reuteri ATCC 55730 showed the ability to significantly decrease the levels of circulating IgE to egg white at two years of age after a prenatal and postnatal treatment in infants at a high risk of allergy [132]. However, Lactobacillus reuteri ATCC 55730 had no effect on other food allergens, such as cow's milk, cod, wheat, peanuts, and soybeans. Interestingly, the authors observed that the effect of the treatment was more pronounced in infants whose mothers (and not fathers) have allergic disease. This highlights the significance of the supplementation to mothers in late pregnancy. Finally, no significant differences were observed in positive food SPT in infants at high risk of allergy after a prenatal treatment with Lactobacillus rhamnosus GG (Gorbach-Goldin) during the last month of pregnancy [133].

Probiotics for Prevention of Asthma/Wheezing and Rhinitis: Administration to Mothers Versus Infants
Allergic diseases such as asthma, atopic dermatitis (AD), and allergic rhinoconjunctivitis (ARC) are among the main health problems in children and are particularly abundant in western countries. The prevalence of allergic diseases varies based on the population studied, from 9.5% in asthma up to 10-20% in allergic dermatitis in American children [134,135]. The risk factors are multiple, including parental history of allergies, early childhood allergen exposure, lack of breastfeeding, or an immune predisposition to Th2 [136]. Although many follow-ups of RCTs (randomized clinical trials) revealed a lower risk for eczema after probiotic treatment [137], recent studies regarding its effect in preventing asthma, rhinitis, or wheezing need to be reviewed. The latest evidence from randomized, double-blind, and placebo-controlled trials evaluated the preventive properties of probiotic supplementation in the prenatal and/or postnatal stages in asthma/wheezing and rhinitis. Since Elazab et al. (2013) published a meta-analysis about the effect of probiotics in atopy and asthma in early life [138], our analysis contains all the prenatal/postnatal or postnatal probiotic prevention studies in asthma, wheezing, and rhinitis indexed in Pubmed in the last 10 years. A total of 12 randomized, double-blind, placebo-controlled trials were included in the analysis: eight of them referred to the supplementation of probiotics during the preand postnatal stage (Table 4) and four during the postnatal stage (Table 5). Several studies administered probiotics during the first 6 months of life and no study of the postnatal period showed administrations beyond two years of age. Regarding probiotic treatment during the postnatal stage, all of them failed to obtain significant results in the prevention of asthma, rhinitis, or conjunctivitis. Three of the four studies agreed on the duration of treatment (six months), except for the study of West et al. (2013), which extended the treatment to nine months. The age at intervention was also different, as well as the strains, the doses, and the mode of administration (mixed with water, infant cereals, cow's milk-based formula, partially hydrolyzed whey formula, or breast milk), although half of the studies agreed to start treatment after birth and in newborns with a high risk of allergies [139,140]. A significant difference in the incidence of eczema was observed in the study of Schmidt (2019) compared to placebo according to a previous meta-analysis that showed a significantly lower risk ratio (RR) for eczema compared to controls, especially those supplemented with a probiotic mixture [137]. However, it is important to note that asthma, rhinitis, and conjunctivitis develop later in childhood [141,142]. In Schmidt's study, the children had a maximum age of 20 months at the end of the treatment and no follow-up was performed, so that the protective effect of the probiotic could have been masked. In the remaining postnatal studies, the authors observed a lower cumulative incidence of asthma and allergic rhinitis in the probiotic groups compared to the placebo group (asthma: 9.7% vs. 17.4% [139]; rhinitis 12.9% vs. 19% [140] at five years of age), but differences were not significant. There was no significant improvement in lung function after a follow-up of 8-9 years in children treated with Lactobacillus paracasei F19 for nine months at four months of age [124]. However, this study lost 60-70% of the original population after follow-up, probably biasing the results. In the same line, the intervention with probiotics during the prenatal and postnatal stages did not prevent rhinitis or asthma/wheezing during childhood. Only Wickens et al. observed a significant reduction in the prevalence of wheezing (64.2% vs. 76.8%), eczema (42.1% vs. 59.4%), and atopic sensitization in (49.5% vs. 62.3%) in the group treated with Lactobacillus rhamnosus HN001 compared to the placebo after 11 years of follow-up. No statistical significance for rhinitis prevention was obtained after probiotic treatment (65.6% vs. 73.5%). This study is the longest related to these outcomes for now [143]. Interestingly, this same study showed a significantly lower incidence for eczema at four years (32.7% vs. 49.3%), but not for asthma or wheezing [144]. However, despite the high rate of participation at four years of age (about 90% in all three groups), no significant results were obtained for the Bifidobacterium lactis HN019 group. One of the eight pre/postnatal studies showed, using electronic follow-up data, a higher prevalence of asthma in the children of the probiotic group versus placebo at five years (31% vs. 17%). These results contradict some meta-analyses that did not find an increased risk of asthma or wheezing [138]. As the authors mentioned, children or breastfeeding mothers could have taken commercial probiotics during the follow-up period [145]. Neither a significantly lower incidence of allergic rhinoconjunctivitis nor asthma/wheezing was seen at two or six years of age after the supplementation of LGG, B. lactis Bb-12, and L. acidophilus La-5 in mothers for four months. Although AD incidence was significantly lower in both studies [146,147], insufficient statistical power and a high proportion of missing data produced nonsignificant results in asthma and ARC.   Significant reduction in the cumulative incidence of AD at two years of age.
No reduction in the incidence of asthma or ARC The nonsignificant results in asthma and ARC may be a result of insufficient statistical power.
No statistically significant differences between the groups.
Loss to follow-up of 60-70% of the original study population.

The Use of Probiotics and Paraprobiotics in Preterm Neonates
Preterm infants experience a delay in bacterial colonization, usually composed by Bifidobacterium and Lactobacillus, causing the settlement of pathogenic bacteria. The colonization of pathogenic microbes, which leads to dysbiosis in preterm infants, is linked to the delayed introduction of human milk, early antibiotic intervention, a high rate of caesarean delivery, and total parenteral nutrition [152]. This last factor is associated with a lower abundance of Bacteroides and Bifidobacterium and a significant loss of biodiversity [153].
Disturbances in the gut microbiota may impair the barrier and immune system, leading to delayed maturation of the humoral immune systems and subsequently to inflammatory reactions. The imbalance between proinflammatory response and insufficient anti-inflammatory protection increases the risk of late-onset sepsis (LOS) and necrotizing enterocolitis (NEC), especially in very-low-birth-weight (VLBW) babies [154]. NEC is a harmful pathology, with a high rate of morbidity and mortality, and occurs prevalently in neonates born weighing less than 1500 g. The pathogenesis may be multifactorial, involving the immune system as a response to an ischemic or infectious insult, with the intestinal microbes playing an important role in the pathogenesis of NEC [155].
Preterm neonates are characterized by immature immune pathways, so they cannot control the extension of pathogenic bacteria. Moreover, the immaturity of the gastrointestinal function, in particular intestinal motility, circulatory regulation, the intestinal permeability barrier, and mechanisms of humoral immune defense, enhance the susceptibility to severe diseases. Supplementation with probiotics may regulate the intestinal microbiota and settle the gut, with beneficial bacteria preventing the development of NEC. Probiotics may act on intestinal permeability, enhance mucosal IgA responses, and increase anti-inflammatory cytokines [156].
Several quantitative nonrandomized studies have demonstrated that a prophylactic probiotic in preterm babies <1500 g is associated with lower mortality and morbidity, including a lower risk of NEC or LOS [157][158][159][160]. However, other studies with a similar design did not find this association, probably due to the multifactorial causes of these pathologies [161][162][163] (Table 6).
Regarding randomized clinical trials, numerous studies have demonstrated the relationship between probiotic administration and low mortality and low risk of NEC [164][165][166]. Other studies found a beneficial effect only on LOS [167]; others did not find any beneficial effects of probiotics in reducing the risk of NEC [168][169][170] (Table 6). The strengths of such studies were the longitudinal design, the inclusion of a large number of preterm neonates, the strict inclusion criteria, the stratification for weight and gestational age, antibiotic therapy, caesarean section, maternal pathology, and all conditions that might impair the establishment of a beneficial and functionally active neonatal intestinal microbiota.
The limitations of these studies were that the authors did not report the comorbidity of other diseases such as the persistence of arterial ductus (also a multifactorial pathology) or the ratio of small for gestational age neonates who are at high risk of NEC [171].
Furthermore, there is a high heterogeneity among studies, which could be produced by differences in eligibility criteria, the use of a variety of probiotic species, and the different protocols of dosage and timing. Therefore, the included studies had different strategies concerning enteral feeds (e.g., breast milk vs. formula) and the use of antibiotics. Previous studies reported that more effective colonization by supplemented agents could start at birth, in an uncontaminated gut environment, leading to improved short-and long-term benefits [172].
The most recent studies support the use of Lactobacillus and Bifidobacterium combination probiotics as most beneficial for preventing NEC in very preterm neonates [160].
Bifidobacterium species digest components of human breast milk, such as human oligosaccharides (HMOs), enhancing their establishment in the infant gut and maximizing nutrient utilization. Bifidobacterium may create resistance to potentially dangerous pathogens and also stimulate the improvement of the mucosal and systemic immune systems, which play an essential role in enhancing the development of the preterm gut, preventing NEC and LOS [173].
It is important to clarify that the different results of some studies are linked to critical differences between probiotic strains, whose characteristics such as the ability to modulate immunity and infections will be different. Accordingly, randomized clinical trials such as the large UK multicenter Probiotics in Preterm Infants Study (PiPS), which used the probiotic Bifidobacterium breve BBG-001 strain, did not improve the prognosis of NEC or LOS [174].
An updated meta-analysis including placebo-controlled studies to explore the effect of Lactobacillus on the incidence of NEC in preterm infants showed a significant reduction in the incidence of NEC (RR 0.34, 95% CI 0.25-0.46; p < 0.00001) and death (RR 0.48, 95% CI 0.36-0.64; p < 0.00001). No significant difference in the incidence of sepsis was found between the Lactobacillus and placebo groups (RR 0.90, 95% CI 0.72-1.12; p = 0.34) [175]. However, this study had several limitations, such as different doses, strains, duration of supplementation, and variations in the gestational age or birth weight of preterm infants. Another recent meta-analysis based on 23 RCTs (n = 4783) of probiotics in preterm neonates in low-and medium-income countries indicated that probiotics are significantly effective at reducing the risk of all-cause mortality, LOS, and NEC in preterm VLBW [176]. However, nearly 40% of those studies carried a high risk of bias.
Interestingly, a network meta-analysis (NMA) to identify the best prevention strategy for NEC in preterm infants concluded that a probiotic mixture and Bifidobacterium more significantly reduced the incidence of NEC than Lactobacilli, Bacillus, or Saccharomyces [177]. The last published guidelines of The European Society for Paediatric Gastroenterology Hepatology and Nutrition (ESPGHAN) recommend that, if all safety conditions are met, the use of Lactobacillus rhamnosus GG ATCC 53103 at a dose ranging from 1 × 10 9 CFU to 6 × 10 9 CFU might reduce NEC stage 2 or 3, although with low certainty of evidence. Most recently, the panel conditionally recommended using a combination of Bifidobacterium infantis Bb-02, Bifidobacterium lactis Bb-12, and Streptococcus thermophilus TH-4 at a dose of 3.0 to 3.5 × 10 8 CFU (of each strain) as it might reduce NEC stage 2 or 3 (low certainty of evidence) [178].
Nevertheless, it has been reported that probiotics have the potential to cause probiotic-related sepsis [179], but in the analyzed studies the authors did not find any adverse effect. Accordingly, some authors recommend screening the safety of probiotic supplements for antibiotic resistance in commercially manufactured probiotic supplements [180]. Defined as nonviable microbial cells (intact or broken) or crude cell extracts, paraprobiotics have been proposed as a potential alternative, although more clinical trials on this topic are necessary [181]. The incidence of NEC was analyzed in <30 weeks or <1500 g babies, from the national database and from the VON databases separately in all five level III NICUs and additionally in three groups according to the probiotic practice.
LGG had no influence on the clinical course of NEC.   VLBW neonates (BW ≤ 1500 g) treated from the first feed until discharge. The median duration of probiotic supplementation and follow-up was 44 days. The study was conducted in preterm infants (≤ 32 GWs, ≤ 1500 g birth weight). They were randomized either to receive feeding supplementation with S. boulardii 50 mg/kg every 12 h or a placebo, starting with the first feed and continuing until discharge.

Safety of Probiotics in Pregnancy and Neonatal Period
In recent years, probiotics have been routinely used in pregnant women and newborns. Probiotic preparations can be administered alone or in combination with antibiotics, especially for gastrointestinal and genitourinary health. Interestingly, the Food and Drug Administration (FDA) regulated probiotics such as nutritional components; their regulation varies between regions and is focused on the legitimacy of any counter-claims, rather than the efficacy, safety, and quality [183]. Thus, these regulatory deficits may have serious consequences for vulnerable groups, and a careful safety evaluation is required before their use. The European Commission's Scientific Committee on Food (SCF) recommended the exclusion of Enterococcus strains as probiotics [184]. Due to the increased use of probiotics, several studies have focused on the effectiveness and the safety of probiotics in recent years. Of the initial 75 studies selected in this review, after eligibility assessment we included 21 studies that met the inclusion criteria; all clinical trials published in the last 10 years that evaluated the safety of probiotics in pregnancy and the neonatal period.
Only one of the studies reviewed reported adverse effects due to the use of probiotics. Topcuoglu et al. included 210 neonates born before 32 weeks of gestation for a randomized controlled trial of a probiotic preparation containing Lactobacillus casei, Lactobacillus rhamnosus, Lactobacillus plantarum, Bifidobacterium lactis, fructooligosaccharide, galactooligosaccharide, colostrums, and lactoferrin (no strain names are published). A vancomycin-resistant Enterococcus (VRE) outbreak occurred while the probiotic trial was being conducted. The only difference found in the VRE newborn group was the use of probiotics (p < 0.001). The authors concluded that concomitant probiotic and vancomycin treatment increases the risk of developing VRE, probably by the acquisition of resistance genes of bacteria mediated by probiotics use [185]. Two case reports about Bifidobacterium-related sepsis have been described in LBW and VLBW neonates, one of them after a surgery on a rare abdominal wall defect (omphalocele) and the other one in an extremely low-birthweight infant (600g) [179,186]. Neither case was life-threatening.
Among the other reviewed studies, no adverse effects were reported from the use of probiotics.

Discussion
The colonization of the human microbiome is a progressive process. In recent years, controversy has arisen over whether the colonization starts intrauterine, with the microbiota present in the placenta and the amniotic fluid, or during the delivery process. The available studies agree that the delivery mode modifies the earliest microbiome, with lower proportions of Bifidobacterium and Lactobacillus and a delayed establishment of gut microbiota in newborns delivered by C-section [207]. Infant maturity at birth also influences the gut early microbiome, with Bifidobacterium being one of the genera most affected in late preterm infants [47,48]. The airway microbiome also depends on the delivery mode more than the infant maturity, although further research to clarify this topic is needed. The studies presented in Section 3.2 about IPA and postnatal antibiotic treatment showed a high prevalence of Proteobacteria and Enterococcus and low levels of Bifidobacterium, as well as a delayed colonization of this genus. This early imbalance in the gut microbiota of newborns may be related to future diseases, such as atopic diseases [208]. The combination of prescribed antibiotics and infants born via C-section leads to a larger expression of antibiotic resistance genes in the gut microbiome [209], so the evaluation of the possible overuse of antibiotics in neonatal centers should be performed in future trials.
It is known that the administration of probiotics in early life can stimulate Th1 cytokines to reverse the Th2 imbalance. However, the heterogeneity of studies on the prevention of atopic diseases generates controversy about their effectiveness in prenatal and/or postnatal administration. The few existing studies, together with their heterogeneity and the fact that the properties of one strain cannot be extrapolated to others, make it difficult to establish recommendations for probiotic use for the prevention of food allergies during infancy. Further studies are necessary to verify the specific benefits of individual strains, as several studies have been performed using a mixture of strains. Moreover, differences in the genetic background, mode of delivery, and intestinal microbiota composition among populations could bias the results, limiting the comparability of studies. Despite this, the results suggest that prenatal and postnatal probiotic treatment produce a protective effect on food sensitization compared to the use of probiotics only after delivery. These studies highlight the importance of maternal immunocompetence in utero and breastfeeding microbiota transfer to infants. Furthermore, the study of Morisset et al. highlights the use of heat-inactivated probiotics, also called paraprobiotics, for immune health in infants [127]. Paraprobiotics are inactivated microbial cells (nonviable), are immunologically active, and have been reported to provide health benefits to hosts [210]. Their application in foods could offer interesting advantages such as a longer shelf life, simple storage and transport, easier standardization, and parents' greater confidence in their use. The World Allergy Organization (WAO) guidelines determined that probiotics could have a benefit in eczema prevention when used in pregnant women. However, studies in breastfeeding women and infants at high risk of developing allergies [211] did not show a significant increase in the prevention of other allergies such as rhinitis, wheezing, or asthma. These results are consistent with a previous meta-analysis [137,138]. However, we considered it necessary in this review to analyze the latest clinical studies not included in previous studies. Only the study carried out by Wickens et al. showed prevention against wheezing, atopic sensitization, and eczema in one of the two strains studied, proving, once again, that the immune-modulating effects of bacteria are strain-specific [143]. Interestingly, previous studies showed that L. rhamnosus HN001 mostly benefited infants with a genetic predisposition to poor skin, low intestinal barrier function, and an imbalanced Th1/Th2 response [212].
Some points relative to the preventive effect of probiotics on these diseases must be clarified. These allergies usually do not arise until childhood, so longer follow-up studies are necessary. The use of other probiotics or antibiotics during the follow-up period and the high dropout rate, probably due to the preventive effect of treatment, make it difficult to complete the study successfully. The benefits of electronic health records, showed by Davies et al. in terms of a high retention rate after long follow-ups, could be a promising tool to decrease those biases [145]. Moreover, Lactobacilli are transient colonizers.
Most of the analyzed studies spanned around six months, so it would be desirable to incorporate a longer duration of treatment in further studies. In addition, the delivery mode must be considered as a stratification factor, due to its influence on breastfeeding, infant gut colonization, and allergic risk [213].
Studies have revealed that probiotics are a safe alternative in pregnancy and in full-term newborns. Our analysis detected adverse effects in only three studies, two of them were case studies of bacteremia in VLBW and one after surgery in an LBW infant [179,186]. For that reason, the use of inactivated probiotics in highly vulnerable populations is a promising option for future studies. Probiotics and paraprobiotics are, therefore, a safe therapy with wide applications throughout perinatal life. However, it is imperative to perform rigorous studies and enact regulations to guarantee the safety of the weakest populations. The lack of strict regulations on probiotic manufacturing invites doctors and consumers to demand products whose efficacy and safety have been clearly demonstrated in clinical studies.