Spontaneous preterm birth (SP^PTB) represents the first cause of perinatal mortality and a huge public health cost in many countries, including Italy.

SP^PTB is a complex multifactorial obstetric and neonatological complication, and little is known about its heterogeneous pathogenesis (1). Although a proven infection can only be documented in about 40% of preterm births with premature membrane rupture or spontaneous labor, the colonization of the genital tract by some pathogens and chorioamnionitis have been traditionally considered as the most prevalent causes of SP^PTB.

Accumulating evidence suggests that the uterine environment is not sterile but populated by a set of different microbial communities present in the placenta, in the fetal membranes and in the amniotic fluid that could contribute to the formation of the neonatal microbiota before birth (2–4).

The formation of the microbiota begins early at the time of organogenesis and can promote protection or predisposition toward certain insults and pathologies (5–7).

Nevertheless, it is known that fetal and maternal colonization process can be influenced by several factors such as the mode of delivery [vaginal delivery (VD) vs. cesarean section (CS)] (8, 9), maternal bonding, admission to intensive care unit, administration of antibiotic therapy, H₂ antagonists, dietary modifications with formulated milk, use of fortifying breast milk and fasting periods (10, 11). These are frequent circumstances in premature infants, who often undergo treatments that can alter the eubiotic colonization process.

The evolution of infants’ gut microbiota was a very investigated topic in recent years, while few researchers have addressed the development of the microbiota of the respiratory tract, particularly lower airways one. Pulmonary colonization starts in utero and can be subsequently influenced by the exposition to the birth canal and the skin of the mother, as well as by microorganisms, introduced through ventilation or micro-aspiration (2, 8, 12–14). The lung microbiota shows changes in the first week of life (15, 16) and an evolution of the colonization pattern is found over the first week of life with evidence of the prevalence of organisms of the species Staphylococcus sp. (Firmicutes) or Ureaplasma spp. (Tenericutes) within 7 days (17).

Furthermore, lung microbiota can be influenced by lung diseases, particularly by the development of bronchopulmonary dysplasia (BPD) (16).

The aim of the study is to investigate the role of maternal and neonatal microbiota in SP^PTB cases. We investigated maternal vaginal microbiota, and neonatal lung and intestinal microbiota. We compared preterm infants born by spontaneous birth for suspected chorioamnionitis and/or premature prolonged rupture of the membranes (p-PROM) and/or preterm labor (SP^PTB group) with preterm infants born by cesarean section performed for maternal or fetal medical indications, with intact membranes and in the absence of labor (Medically Indicated Preterm Birth–MI^PTB).

The identification of a specific maternal and preterm newborns microbiota signature associated to SP^PTB and to adverse neonatal outcomes, could open to the investigation of new markers (a) in pregnancy, to select cases at higher risk of SP^PTB and (b) in preterm newborns, to early identify cases at higher risk of adverse pulmonary outcomes.

From August 2019 to August 2020, a total of 30 mothers and 35 newborns were enrolled at birth after receiving informed parental consent. Six mother-newborns couples were excluded because newborns were not intubated in the first 24 h of life. The study group consisted of 29 newborns, including 3 pairs of twins, and 26 mothers.

Twenty-nine BALF, 29 early meconium and 26 vaginal swabs samples were analyzed. Six BALF samples were excluded due to the low bacterial biomass. A total of 20 mothers (11 = SP^PTB; 9 = MI^PTB) and 23 neonates (12 = SP^PTB; 11 = MI^PTB), were included in the final statistical analysis.

The demographics and clinical characteristics of infants included in the analysis are shown in Table 1. No significant differences were observed between the two groups, except for earlier initiation of antibiotic therapy in infants of the SP^PTB group [1 (1–2) vs. 2.5 (2–7) days of life, p = 0.01] and for the occurrence of premature prolonged rupture of the membranes (pPROM) which only registered in the SP^PTB group (p < 0.0001). In this group of neonates, only in two cases labor was activated without rupture of the membranes. A trend in terms of lower gestational age (27.0 ± 1.9 vs. 28.3 ± 1.1 weeks, p = 0.07) and of higher incidence of late-onset sepsis [6 (50%) vs. 1 (9%), p = 0.07] was found in the SP^PTB group compared to the MI^PTB group. All 23 infants survived to 36 weeks’ postmenstrual age (only 1 infant in MI^PTB died after 36 weeks postmenstrual age). Based on their pulmonary outcomes, the 23 infants were further classified as those who did not develop BPD (No BPD group, n = 15) or those who later developed BPD (BPD group, n = 8). The demographics and clinical characteristics of patients in these two categories are shown in Table 2, with significant differences in some characteristics and outcomes. In particular, infants in the BPD group had a lower GA and birth weight, had a greater number of episodes of blood stream infection and pneumonia during hospitalization and received more cycles of antibiotic therapy. In addition, infants in the BPD group needed oxygen therapy and non-invasive ventilation for a longer time than infants in the No BPD group, had a hemodynamically significant patent ductus arteriosus (PDA) more frequently, and their hospital stay was significantly longer.

The present paper aims to identify specific vaginal and neonatal lung microbiota profiles associated to the occurrence of SP^PTB and MI^PTB. As preterm births have been historically considered a sign of infection, we compared SP^PTB cases to MI^PTB cases, to understand if an adverse uterine environment could influence preterm birth and neonatal lung microbiota.

Chorioamnionitis, defined as inflammation of fetal membranes, is one of the mechanisms leading to preterm labor. A clinical diagnosis can be made prenatally by maternal findings (fever, uterine contractions, leucocytosis, high serum levels of C-reactive protein, abnormal vaginal secretions), while a pathological diagnosis is obtained postnatally by histological examination of the placenta and the amniochorionic membranes (27). A proven microorganism infection can be detected only in 50% of cases of chorioamnionitis, probably due to the technical limitations of cultural tests. Microbiota profiling by NGS technique might represent a more sensitive diagnostic tool to identify dysbiosis and increased risk of SP^PTB (28, 29).

Recent studies challenged the assumption of a sterile uterine environment, demonstrating the presence of bacteria in the placenta, fetal membranes, and amniotic fluid (2–4). However, these findings are controversial as there is the risk of contamination of low microbial biomass samples.

Vaginal dysbiosis is recognized as a factor that increases the risk of SP^PTB. Several species of Lactobacilli promote the maintenance of an acid pH and the production of antimicrobial substances.

During pregnancy there is a progressive increase in abundance of species such as Lactobacillus vaginalis, L. crispatus, L. gasseri, and L. jensenii, belonging to the Phylum of the Firmicutes. Conversely, bacteria such as Prevotella, Sneathia, Gardnerella Vaginalis, Ruminococcaceae, Parvimonas, and Mobiluncus, often associated with bacterial vaginosis and widely present in non-pregnant women of childbearing age, progressively decrease (30). Hence, the microbiota of women in pregnancy becomes progressively more stable, less prone to changes and protective against genitourinary tract infections (31).

Several studies have shown that greater α-diversity of vaginal microbiota is associated with a progressive reduction in the relative abundance of the genus Lactobacillus, which is known for its role in giving stability to the vaginal microbiota and protecting it from pathogenic insults.

Consistently with previous literature, in this study we documented a greater alpha diversity of vaginal microbiota of the SP^PTB mothers when compared with mothers of the MI^PTB group. Moreover, in the SP^PTB group, vaginal swabs revealed a decrease in Firmicutes and an increase in abundance of Bacteroidetes. At Genus analysis, a significant decreased abundance of Lactobacillus was found in SP^PTB.

The timeline of lung microbial colonization is highly debated. Mourani and colleagues reported the presence of detectable bacterial DNA in only 2 of 10 tracheal aspirates of intubated preterm infants in the first 72 h of life, whereas all samples from the same newborns were positive at day 7 (17). On the other hand, Lohmann et al. found the presence of bacterial DNA in all tracheal aspirates collected in 25 neonates immediately after orotracheal intubation on the first day of life (16). Lal et al. observed that BALF samples of extremely low birth weight infants exposed to chorioamnionitis showed a decreased genus Lactobacillus at birth, which has been associated with development of BPD.

All cases of chorioamnionitis were determined by positive placental histopathology and prenatal antibiotics administration in mothers (32).

We investigated any possible cause of spontaneous birth and possible adverse uterine environment (2 mothers in SP^PTB had chorioamnionitis and 10 had a p-PROM) looking for a microbial signature suggestive of spontaneous birth. However, despite differences observed in maternal vaginal swabs, we did not find in neonatal BALF samples at birth a different pattern of alpha and beta diversity and/or differences either at Phylum and Genus level, comparing SP^PTB and MI^PTB group.

A problem underlying the analysis of neonatal lung microbiota is the difficulty in retrieving material from the lungs and distal airways in neonatal population, due to the low biomass with bacterial loads close to the detection limit of the assays and to the high risk of contamination from the upper airways (33). In these terms, lung microbiota at first day of life could be significantly impaired by contamination during sampling and by sequencing technique. Moreover, samples derived from sites with a low microbial biomass can give results which are difficult to distinguish from DNA traces present in reagents used for extraction, amplification, and sequence library preparation for molecular microbiology studies (34, 35). Hence, we cannot exclude that our findings are the results of high sensitivity of sequencing process and sampling.

Indeed, the presence of bacterial DNA in BALF samples, while lending support to the hypothesis that the process of microbiota establishment might begin antenatally, does not exclude the possibility of post-birth microbiome development, impaired by environmental factors, or sample contaminations.

On the other hand, the lack of bacterial DNA in the analyzed meconium samples (collected before the start of enteral feeding, with no probiotic intake and in the absence of abdominal pathological processes) supports the idea that microbiota composition is acquired after birth, and it is not related with the cause of preterm birth or transferred from swallowing bacteria previously translocated from uterine niche in the amniotic liquid.

The absence of a specific microbial signature in the BALF samples of the two groups (MI^PTB vs. SP^PTB) may, at least in part, suggest that the cause of preterm birth was not associated with infant lung microbiota composition.

While the two groups differed for some clinical characteristic as GA and birth weight, the analysis conducted on BALF samples of infants who developed or not BPD, did not show statistically significant differences.

Furthermore, the newborns of the BPD group had a higher incidence of hemodynamically significant PDA, blood stream infections and pneumonia during hospitalization and, consequently, are subjected to a greater number of antibiotic therapy cycles (Table 2). However, the two groups did not differ in the use of intrapartum antibiotic prophylaxis, nor in the time elapsed from birth to the start of antibiotic therapy. This may partly justify the lack of a statistically different expression of the microbiota in BALF samples taken in the first 24 h of life, and certainly more information could have been obtained by having a greater number of BALF samples available on the third, and seventh day and in the more advanced periods of life.

Although not statistically significant, a slight reduction in Shannon diversity index is evident in the BALF samples of BPD infants (Figure 5). This trend confirms the results obtained by Lohmann et al. (16). The authors identified a reduction in the number of species of respiratory microflora and Shannon diversity index in the BALF samples obtained in the first 24 h of life from 10 infants diagnosed with BPD compared to 15 other infants who had not developed the disease. A decrease in microecological diversity of the respiratory tract in BPD was observed also by Lal et al. (32) who analyzed 18 newborns with confirmed BPD, of which 5 infants with an extremely low birth weight.

We did not find a characteristic microbiological signature linked to BPD development both at genus and phylum level. However, the results of the studies by Lohmann et al. and Lal et al. showed opposite trends in Proteobacteria (a downward trend in the results of Lohmann et al. compared to an increase in Lal et al.) and Firmicutes abundances (with an upward trend in Lohmann et al. compared to a decrease in Lal et al.).

Collaterally, from cultural analysis, positive Ureaplasma parvum day-one BALF culture resulted significantly more prevalent in BPD group. Moreover, all neonates with positive day-one BALF were born form mother with positive vaginal swabs for Ureaplasma. This result is in line white previous studies, which associated Ureaplasma colonization in the respiratory tract with the subsequent development of BPD (16, 17, 36, 37).

However, there is still not a total agreement on the relationship between respiratory Ureaplasma colonization and BPD development, as well as there is not agreement on the benefit or risk of antibiotic therapy for respiratory Ureaplasma colonization (38).

We are aware that our study may have several limitations, as the small study population, given the difficulty of obtaining vaginal swabs timely before birth and the difficulty of sampling from the distal airways in the neonatal population. Other limitations of the study are represented by the lack of analysis of the placental microbiota and the lack of a histological diagnosis of chorioamnionitis. The implementation of these analysis, in association with further studies on amniotic fluid microbiota are needed to investigate the presence of a link between maternal and neonatal microbiota.

An increase in α-diversity values and a consequent fall in Lactobacillus in vaginal environment could be associated to a higher risk of SP^PTB, probably due to a less stable and protective environment, suggesting that variation in maternal microbiota can predispose to SP^PTB.

We could identify neither a specific neonatal lung or meconium microbiota profiles in preterm infants born by SP^PTB nor a microbiota at birth suggestive of subsequent BPD development. However, accordingly to previous studies, the presence of Ureaplasma bacteria in newborns’ BALF could be considered a risk factor for the development of BPD.

Although a strict match has not been revealed between microbiota of SP^PTB mother-infant couples, a relationship cannot be excluded.

To figure out the reciprocal influence of the maternal-neonatal microbiota and its potential role in the pathogenesis of SP^PTB and BPD further research is needed.

The data presented in this study are deposited in the sequence read archive (SRA) and are available as BioProject number PRJNA849991.

The studies involving human participants were reviewed and approved by Comitato Etico Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, Italy. Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin. Written informed consent was obtained from the individual(s), and minor(s)’ legal guardian/next of kin, for the publication of any potentially identifiable images or data included in this article.

CTi, AP, and FD conducted the statistical analysis and interpretations of results. CTi designed the study. AP, CTi, and FD wrote the manuscript with the support of CTe, SD’I, ND, ALa, GSc, and GV. FD, FM, GSa, DB, BP, and MS carried out the microbiological analysis. CTi, AP, CTe, SD’I, ND, MT, ALi, and NM collaborated in data collection. All authors read and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.