Low birth weight (LBW), defined as birth weight less than 2500 g, is a significant public health concern with a prevalence of 15 % globally and a leading cause of infant mortality, morbidity, and long-term sequelae related to complications of prematurity [1]. Poor maternal nutrition, placental problems, infections during pregnancy [2], and chronic health conditions can increase the risk of LBW or premature birth (infants born before 37 weeks of gestation) [3]. Through epidemiological studies, the majority of current evidence suggests that poor oral health, particularly periodontitis, may be associated with LBW [[4], [5], [6], [7], [8], [9]]. However, it is reported that treatment of periodontitis in pregnant women improves periodontal disease but there is low-quality evidence that periodontal treatment may reduce LBW [10,11]. Further research is needed to identify and address risk factors to improve maternal and infant health outcomes.

The oral microbiome, specifically certain periodontopathogens, has been linked to adverse birth outcomes. Pregnant women with higher levels of bacteria such as Fusobacterium nucleatum and Porphyromonas gingivalis which are red and orange complex pathogens in periodontitis are more likely to have adverse birth outcomes [12,13]. Periodontitis-related species with tissue-invasive tendencies (e.g., species of P.Gingivalis) have been identified in the amniotic fluid of women who deliver prematurely, potentially causing damage to the developing fetus [14]. Being carried in the blood to the uteru, oral microorganisms and products of cytokines or endotoxins, would induce an inflammatory/immune response in the fetal-placental unit, that triggers preterm labor [[15], [16], [17], [18]]. Other oral microbiome species, including Streptococcus mutans, Prevotella intermedia, and Gardnerella vaginalis, have also been linked to adverse pregnancy outcomes [19,20]. However, one study reported that a higher burden of periodontal pathogenic bacteria was only associated with increased periodontal disease activity and not with preterm birth [21].

The majority of studies examining the microbiome and adverse pregnancy outcomes are cross-sectional, longitudinal studies tracking the microbiome's dynamic changes throughout pregnancy are needed. Furthermore, many studies examining the association between the oral microbiome and adverse birth outcomes have primarily focused on periodontopathogens, while largely neglecting the potential contribution of other medical conditions and confounding factors. It is essential to consider additional confounders beyond the microbiome and explore the entire oral microflora during pregnancy to construct a prediction model before clinical application. Therefore, utilizing a nested case-control design within a prospective cohort and adjusting for potential confounding factors, such as collecting sociodemographic and general health information, would provide a higher level of evidence by providing a more precise evaluation of the association between the oral microbiome and adverse pregnancy outcomes.

NGS-based 16S rRNA sequencing techniques allow for the analysis of the entire microbial community within a sample, enabling the identification of species that may not be found using previous methods. This provides an opportunity to identify potential pathogens and gain a more comprehensive understanding of the oral microbiome's role in adverse pregnancy outcomes [22].

Machine learning-based preterm birth prediction models using vaginal microbiome data have achieved high accuracy rates (ranging from 75 % to 85 %), highlighting the potential of microbiome data in predicting adverse pregnancy outcomes [[23], [24], [25]]. While there are similarities between the oral and vaginal microbiomes [26], further research is needed to determine whether certain bacterial species found in the oral cavity can assist to predict adverse pregnancy outcomes.

This study aimed to investigate changes in maternal oral microbiome composition and diversity during pregnancy, identify potential oral microbiome predictors for low-birth-weight delivery, and develop a machine-learning model to predict adverse pregnancy outcomes. The study will contribute to a better understanding of the role of the oral microbiome in pregnancy outcomes and potentially provide a tool for early prediction and prevention of adverse birth outcomes.