This study utilized a longitudinal design with a sample of 59 participants drawn from among a larger group of participants taking part in the Emory University African American Vaginal, Oral, and Gut microbiome in Pregnancy cohort study. Institutional review board approval was obtained for this study the primary purpose of which was to characterize the subgingival microbiome at two points during pregnancy: early gestation (8-14 weeks) and late gestation (24-30 weeks).

Participants for this study were recruited from an ongoing investigation of the association between a woman’s microbiome and preterm birth (Corwin et al., 2017). The parent study prospectively enrolled a socioeconomically diverse cohort of pregnant women between 8-14 weeks gestation who were US born Black or African American and followed them through delivery. Participants who were enrolled in the parent study were approached at the initial data collection period within the 8-14 week gestation period to ascertain their interest in and eligibility for this study. Additional eligibility criteria for this study included having: 1) minimum of 20 natural teeth and 2) no dental cleaning in the past three months. These requirements are routinely used in oral health studies (Eick and Pfister, 2002; Griffen et al., 2012; Matthews et al., 2013) to ensure adequate oral sampling sites and an undisturbed subgingival environment. Use of antibiotics was collected to address potential confounding effects on the microbiome, however, none of the participants whose samples were included in our microbiome analyses reported recent use of antibiotics.

The mean age of the participants was 26.04 ± 5.37 years. Other sociodemographic characteristics and birth outcomes may be found in Table 1 .

Only one participant self-reported gingivitis at the second time point. Other oral health symptoms and behaviors are in Table 2 .

There was no difference in measures of alpha or beta diversity for samples from early and late gestation, controlling for participants. The top twenty taxa ( Figure 1 ) represented in the subgingival microbiome of participants across gestation include a spectrum of bacteria representative of various stages of biofilm progression leading to periodontal disease. These include Porphyromonas gingivalis and Tannerella forsythia. Other bacteria associated with periodontal disease that were reflected in the microbiome profile of our participants included several Prevotella spp., and Campylobacter spp.

The relative abundance of six taxa differentiated the subgingival microbiome of our cohort between early and late pregnancy. Five organisms (Porphyromonas sp._HMT_284, Catonella sp._HMT_164, Peptostreptococcaceae_[XI][G-5] [XI][G-5]_saphenum, Johnsonella ignava, Actinomyces massiliensis were more abundant in early pregnancy, while one, Cryptobacterium curtum, was more abundant in late pregnancy ( Table 3 ).

No association between microbiome features and spontaneous abortion or spontaneous preterm birth at either time point were identified. We did, however, find that in late pregnancy both alpha and beta diversity distinguished the subgingival microbiome of women who spontaneously delivered early term from women who delivered full term ( Figures 2 , 3 ). Several taxa also differentiated women who spontaneously delivered early term vs full term. In early pregnancy, eight taxa including Actinomyces israelii, Cardiobacterium hominis, Treponema putidum, and Lautropia mirabilis were more abundant among women who delivered full term compared to early term ( Table 4 ). In late pregnancy, four taxa (Campylobacter gracilis, Prevotella melaninogenica, and two unnamed species belonging to the genera Tannerella and Catonella) were differentially abundant between women who spontaneously delivered early term and women who delivered full term. Of these, P. melaninogenica was the only organism that was more abundant among women who spontaneously delivered early term ( Table 5 ).

This study describes the subgingival microbiome of a group of pregnant Black women and explores associations between the microbiome and preterm birth. Although the parent study, from which this study leveraged its participants, aimed to enroll a socioeconomically diverse cohort of pregnant women, the majority of our convenience sub-sample were socioeconomically vulnerable with less than high school education, and incomes at less than the federal poverty level. Most of the participants were on Medicaid insurance. Evidence suggests that socioeconomic inequalities are strongly associated with oral health disparities, both self-reported and clinically determined (Mejia et al., 2018). This holds true for pregnant populations as well (Azofeifa et al., 2014; Chung et al., 2014; McNeil et al., 2016). None of the women in the study self-reported gingivitis across pregnancy, although one woman reported red or swollen gums, and around a quarter of the participants reported bleeding gums across pregnancy. Both red/swollen gums and bleeding gums are symptoms of gingivitis. Self-reported gingivitis and symptomatology is less than expected within our cohort since we know that gingivitis is present in 50-70% of pregnant women (Anil et al., 2015), and tends to increase in severity across pregnancy (Cohen et al., 1969). Self-reported gingivitis and bleeding gums, while useful for public health screening, lacks validity for individual screening for gingivitis as defined by a dental professional (Kallio et al., 1994; Abbood et al., 2016). It is likely that underreporting of gingivitis and associated symptoms was present in our study.

Virtually all the women reported brushing their teeth in the last two days, however, use of other oral hygiene measures such as flossing, and mouthwash were not as ubiquitous. This can be understood in the light of previous research which indicates racial, ethnic, and economic disparities related to oral hygiene practices during pregnancy (Boggess et al., 2010). Even controlling for income, education, and insurance status, racial disparities in oral health experiences persist; an important contextual consideration for our cohort (Hwang et al., 2011).

Microbial community diversity can be described in two ways. Alpha diversity describes how many ASVs are present in a community and how evenly they are distributed; beta diversity describes the diversity between two different environments. In studies conducted among an Asian population, pregnancy itself appears to increase the alpha diversity of the subgingival microbiome (Balan et al., 2021), although both species richness and diversity remain stable throughout pregnancy (Balan et al., 2018). Our findings support the consistency of alpha diversity throughout pregnancy among African American women since we found no difference in either species richness or diversity between early and late pregnancy. More research among diverse racial and ethnic populations of pregnant women is needed to further assess the generalizability of this finding.

To contextualize our description of the most prevalent organisms identified in the subgingival space of our cohort of pregnant women, it is important to understand that periodontal disease has a polymicrobial etiology. The current scientific understanding of periodontal disease is that of progressive dysbiosis within the biofilm, a polymicrobial ecosystem which attaches to the surface of the tooth in the subgingival pocket (Guthmiller and Novak, 2002). The dysbiotic shift results in a predominantly gram negative environment (Mohanty et al., 2019) caused by microbial succession that occurs with rapid accumulation of plaque on the teeth (Socransky and Haffajee, 2005). According to the seminal work of Socransky and Haffajee (2005), the postulated scheme of microbial succession occurs according to complexes, or co-occurring clusters of organisms. Initially, members of the yellow, green, and purple complexes dominate along with Actinomyces species (Socransky and Haffajee, 2005). The establishment of these three complexes contributes to the alteration of the subgingival environment such that the next two, more pathogenic, complexes become dominant ( Figure 4 ) (Socransky and Haffajee, 2005). These are the “orange complex” organisms (Prevotella intermedia, Prevotella nigrescens, Peptostreptococcus micros, Fusobacterium nucleatum, Eubacterium nodatum, Streptococcus constellatus, and several Campylobacter species). The orange complex of organisms contain bridge species and generally precedes the growth of the red complex which consists of Treponema denticola, Porphyromonas gingivalis, and Tannerella forsythia (Guthmiller and Novak, 2002). The red complex organisms are known as periopathogens and are present at sites presenting with symptoms of periodontal disease.

Notably, several of the organisms represented among the top 20 most abundant organisms in the subgingival microbiome of our cohort include members of both the red and orange complexes: Porphyromonas gingivalis, Prevotella intermedia, Prevotella nigrescens, Campylobacter gracilis, and Campylobacter showae, suggesting a shift toward a pathogenic subgingival ecosystem. This finding is supported by a previous study by Balan et al. (2018) which also identified a dominance of periopathogens during pregnancy from the genera Prevotella, Veillonella, and Streptococcus, and at the species level, P. gingivalis, P. nigrescens, and P. oris.

Given that periodontal disease symptoms intensify over pregnancy (Cohen et al., 1969), we might expect to see increases in the abundance of these pathogenic organisms over pregnancy, however, the significant changes in relative abundance of organisms that we identified were not among established periodontal disease associated organisms. Instead, slight but significant differences were identified with six taxa, five of which decreased from early to late pregnancy, and one which increased in abundance across pregnancy. Of the five taxa that decreased in abundance, three were uncultivated phylotypes. Little is known about Porphyromonas sp._HMT_284, however, sister organisms within the genus are P. gingivalis and P. endodontalis, both of which have been associated with periodontal disease (Lombardo Bedran et al., 2012). The genus Catonella is a relatively newly identified genus (Moore and Moore, 1994) with one named species, C. morbi, that has been described as a putative periopathogens (Wu et al., 2017). Bacteria from the family Peptostreptococcaceae are a morphologically diverse group of gram positive organisms (Slobodkin, 2014). Due to the multiple genera and species within this family it is difficult to speculate on the meaning of our finding that this particular taxa decreased in abundance over pregnancy. Johnsonella ignava may be an opportunistic pathogen. There is modest evidence that it may be associated with chronic obstructive pulmonary disease as well as oral squamous-cell carcinoma (Wu et al., 2017). The evidence on whether it is associated with periodontal disease is mixed (Moore and Moore, 1994; Wu et al., 2017). Actinomyces spp. are commonly found in human oral mucosa and many novel species have been described in recent years (Renvoise et al., 2009). Although A. massiliensis is known to reside in the subgingival biofilm, its role within an Actinomyces complex associated with periodontal disease is not known (Vielkind et al., 2015). Cryptobacterium curtum was the only taxa identified to increase from early to late pregnancy. Cryptobacterium curtum is an oral opportunistic pathogen that is involved in dental and oral infections (Mavrommatis et al., 2009).

Our findings suggest that despite the overall stability in alpha and beta diversity over the course of pregnancy, relative abundance shifts within individual taxa do occur over time with organisms that have some associations to periodontal disease and other opportunistic infections. Further investigation of these organisms and their clinical impact on oral disease and pregnancy outcomes are warranted.

Likely due to limited power from the small number of preterm births in our cohort, we did not identify any associations between subgingival microbiome features and spontaneous preterm birth. We did, however, identify associations among microbiome features and spontaneous early term birth (37 0/7 weeks through 38 6/7 weeks). During early pregnancy, eight taxa were more abundant among women who delivered full term compared to those who delivered early term. Two were unnamed species belonging to the genera Leptotrichia and Capnocytophaga. Leptotrichia spp. are facultative anaerobic oral commensals that are commonly found in the oral cavity (Smid et al., 2015) and can act as opportunistic pathogens or stimulate the growth of other pathogenic organisms (Eribe and Olsen, 2017). Various pathologic conditions are associated with Leptotrichia, (Eribe and Olsen, 2017) including poor pregnancy outcomes. Leptotrichia amnionii, L. sanguinegens, and L. buccalis have been associated with outcomes such as miscarriage, chorioamnionitis, preterm labor, pregnancy loss, neonatal infection and postpartum infection (Smid et al., 2015). Several Capnocytophaga spp. belong to the green complex, early colonizers in the microbial succession leading to periodontal disease.

Cardiobacterium hominis was also more abundant among women who delivered full term. C. hominis is not commonly associated with oral disease, although one study did find it to be associated with aggressive periodontitis and suggested it as a potentially new periopathogens (Lourenço et al., 2014). C. hominis is better known as an oral commensal belonging to the HACEK group of five organisms which can cause infective endocarditis (Moreillon, 2010).

Oral treponemes are widely considered to play an important role in periodontal disease etiology and pathogenesis (You et al., 2013). Treponema putidum is a novel treponeme isolated from lesions of periodontal disease and acute necrotizing ulcerative gingivitis (Wyss et al., 2004). The genus Prevotella is the largest genus in the phylum Bacteroidetes (Dewhirst et al., 2010) and contains species such as P. intermedia and P. nigrescens, members of the orange complex. Lesser known are organisms like P. loeschii which is, nonetheless, associated with periodontal disease. P. loeschii is known to produce propionic acid, a major metabolic player in the inflammatory process of gingival tissue (Al-Lahham et al., 2010). Actinomyces spp. are thought to play a significant role as early colonizers in the process of plaque formation and figure prominently in the microbial succession dogma of periodontal disease. The genus contains a total of 42 species, 20 of which are relevant for human health (Vielkind et al., 2015). A. israelii is part of the normal human oral flora and is frequently found in dental caries and at the gingival margins of individuals with poor oral hygiene (Desouches et al., 2019). It is the most common cause of actinomycosis, and has also been associated with periodontal disease (Vielkind et al., 2015). Lautropia mirabilis is a Gram-negative, motile, coccus bacteria that has been isolated from the human oral cavity (Gerner-Smidt et al., 1994). While L. mirabilis may be associated with some disease states (Ben Dekhil et al., 1997; Rossmann et al., 1998), it is currently primarily associated with periodontal health (Colombo et al., 2009; Balan et al., 2018; Papapanou et al., 2020). Finally, Bacteroidaceae_[G-1] [G-1]_bacterium_HMT_272 is a novel, uncultivated, unnamed, and uncharacterized taxon. Little is known about this organism, however, it has been identified as the most prevalent phylotype in apical periodontitis lesions (Amaral et al., 2020).

In late pregnancy, four taxa were identified as being differentially abundant among women who spontaneously delivered early term. Women who delivered early term had lower abundances of Catonella sp._HMT_164, Tannerella sp._HMT_286, and Campylobacter gracilis. Both Catonella sp._HMT_164 and Tannerella sp._HMT_286 are uncultivated phylotypes. While these strains are not yet well characterized, we do know that other species within these genera are periopathogens. Tannerella forsythia, for example, belongs to the red complex and is known as a keystone periopathogen. Catonella morbi, the only named species within the genus, has been described as a putative pathogen (Wu et al., 2017). Campylobacter gracilis was one of the top 20 most abundant organisms across pregnancy in our cohort. C. gracilis belongs to the orange complex, known to closely precede colonization by the three dominant periopathogens of the red complex.

Prevotella melaninogenica was the only taxon identified to be more abundant among women who spontaneously delivered early term. P. melaninogenica is an oral commensal, gram-negative anaerobe that is closely related to periodontal disease (Ibrahim et al., 2017). P. melaninogenica has a pro-inflammatory effect that stimulates the release of inflammatory cytokines, suggesting that it may play an important role in promoting chronic inflammation (Larsen, 2017).

Several of the taxa identified to be differentially abundant between women who delivered early term versus full term in both early and late pregnancy were novel, uncultivated species suggesting the importance of looking beyond known pathogens to additional microbial players that may affect oral disease and birth outcomes. Many of the taxa identified to be differentially abundant between early term and full term women have strong known or suspected associations with periodontal disease. That most were elevated in women who delivered full term challenges the hypothetical association between periodontal disease (and periodontal disease-causing organisms) and adverse birth outcomes. Further investigation of these organisms, their role in periodontal disease and early term birth is required.

What our findings do reveal is that there are taxonomic shifts that are associated with early term birth. Two taxa are potential targets for future research. Lautropia mirabilis is a Gram-negative, motile, coccus bacteria that is known as an oral commensal (Gerner-Smidt et al., 1994). Studies have identified L. mirabilis as being associated with periodontal health (Colombo et al., 2009; Papapanou et al., 2020). In our study we identified that a depletion of this organism was associated with spontaneous early term birth suggesting that it may be a potential target of future investigation as an oral commensal associated with both oral and pregnancy health. Prevotella melaninogenica was the only taxa identified to be enriched in women who delivered early term. While not traditionally understood as a periopathogen, it has been identified to be highly abundant in the saliva of pregnant women (Balan et al., 2018), and enrichment of Prevotella spp. is associated with mucosal inflammation leading to systemic dissemination of inflammatory mediators, bacteria, and bacterial products, which in turn, may affect systemic outcomes (Larsen, 2017). Further investigation of this organism and its inflammatory effects on oral health and pregnancy are needed.

Spontaneous early term births are births that occur between 37 and 39 weeks. While infants born at this gestational age do not face the same risks as those born preterm (< 37 weeks), they do remain at increased risk for respiratory distress syndrome, transient tachypnea of the newborn, feeding difficulty, pneumonia, and hypothermia (Parikh et al., 2014). Although reasons for spontaneous early term birth are not fully understood, biological determinants may be at play including placental ischemia, diabetes, and poly- or oligohydramnios (Brown et al., 2015). To our knowledge, no studies have investigated the contribution of the oral microbiome to early term birth or tested the potential hypotheses that an oral microbiome shifted toward periodontal disease might contribute to early term labor. This study lays the groundwork and provides preliminary evidence for future targeted investigations testing this hypothetical association.

Several limitations must be acknowledged. First, this was an exploratory study and was not fully powered to identify microbiome associations with preterm or early term birth. Second, although the Human Oral Microbiome Database offers a well-curated and up-to-date resource for 16S rRNA sequences, marker gene sequencing technology limits the resolution with which to identify taxa. Finally, the lack of periodontal disease clinical diagnosis limits our ability to interpret the clinical impact of the oral microbiome.