Early embryonic arrest is defined as the ultrasound diagnostic of an average diameter of the gestational sac diameter larger than 25 mm, no heartbeat, a crown-rump length exceeding 7 mm, and no fatal cardiac activity by the original ultrasound examination (Hendriks and MacKenzie, 2019). Early embryonic arrest is a component of early spontaneous abortion and a common complication in early pregnancy that can lead to adverse pregnancy outcomes. At present, the main treatment types for early embryonic developmental arrest are expectant management, medicinal management with mifepristone and misoprostol and uterine aspiration (Garcia-Grau et al., 2019; Hendriks and MacKenzie, 2019). In recent years, the risk of embryonic arrest in women of childbearing age has increased, leading to widespread concern. Previous studies have reported that early embryonic arrest is related to chromosomal abnormalities, immunity, infection, endocrine dysfunction, environmental physical and chemical factors, among others. However, in 20–40% of cases, the causes are unknown (Fang et al., 2018). In recent years, studies have found that the occurrence of adverse pregnancy outcomes is significantly related to bacterial pathogens and other pathogenic microorganisms in the reproductive tract (Joseph Davey et al., 2016). In particular, the detection of Gardnerella vaginalis in vaginal secretions of patients with bacterial vaginosis (BV) is directly related to embryo arrest. The concentration of G. vaginalis in patients with embryo arrest is significantly higher than that in individuals without embryo arrest (Seo et al., 2017; Garcia-Grau et al., 2019).

The vaginal microbiota of women is complex, can change rapidly and dramatically and has a significant impact on women’s health. As we all know, the vaginal secretions of women of normal childbearing age are dominated by lactobacilli, which account for more than 95% of the microbial content of the vagina (Antonio and Hillier, 1999). The protective bacteria in the vagina, can inhibit the excessive growth and reproduction of harmful anaerobic and aerobic bacteria, thus maintaining the microecological balance of the vagina. The application of probiotic lactobacilli in the clinical treatment and prevention of recurrent reproductive tract infections and vaginal microflora imbalance has become widespread (Chee et al., 2020). Therefore, the abundance and type of lactobacilli are important criteria for evaluating the imbalance of vaginal flora (Cohen et al., 2020). Vaginal flora imbalance caused by changes in the vaginal flora is closely related to diseases affecting not only the reproductive health of women during pregnancy but also the embryo. This may be because some infectious pathogens promote the growth of BV-related organisms from the lower genital tract, causing the destruction of the barrier and immune defense mechanism in the inflammatory environment (MacPhee et al., 2010).

With the development of next-generation sequencing (NGS), it has become evident that the endometrium not a sterile environment and that the species diversity and richness of the endometrial community are significantly higher than those of the vaginal environment (Verstraelen et al., 2016; Baker et al., 2018). The uterine microbiomes of women in reproductive age with the high abundance of Lactobacillus species, Lactobacillus iners (L. iners), Prevotella spp. and Lactobacillus crispatus (L. crispatus), and the endometrial bacteria are characterized by a polymicrobial ecosystem in the pregnant women (Mitchell et al., 2015; Riganelli et al., 2020). Changes to the abovementioned differences in the microbiome suggest that the intrauterine microbiome plays an important role in human health and diseases, and it helps to clarify the pathogenesis of some common gynecological and obstetrical diseases (Toson et al., 2022).

Previous literature has reported the isolation of vaginal microorganisms from the endometrium in 31.4% of women with acute salpingitis and 14.3% of women with pregnancy loss (den Heijer et al., 2019). Bacterial infection can activate the innate immune system, promote the synthesis of nitric oxide and prostaglandins and affect uterine receptivity and embryo implantation. Early embryo arrest not only affects the emotional and mental health of couples seeking to conceive but also causes serious reproductive tract infections. At present, the abundance of Lactobacillus is used to evaluate vaginal flora imbalance (Cohen et al., 2020). Prospective and retrospective cohort studies to assess the impact of vaginal and intrauterine microbiota on embryonic arrest are still lacking. Here, we aimed to identify microbiota differences among women with early embryo arrest and abortion, to improve the microbiota markers related to early embryo arrest.

Embryonic arrest refers to the death of the embryo in early pregnancy, with an incidence of 10–15%, and 80% of miscarriages occur within the first 12 weeks of pregnancy (Peters et al., 2020). The incidence of early embryonic arrest has been increasing in recent years, and the high incidence of embryonic arrest has caused serious harm to many women and families (Fang et al., 2018). Therefore, determining the causes of embryonic arrest is the key to preventing the incidence of embryonic arrest from increasing. Previous studies have reported that in women with repeated embryo arrest, the relative abundance of L. crispatus in the endometrium was significantly lower than that in healthy women, while the abundance of Gardnerella vaginalis in the vagina and endometrium was significantly higher than that in healthy women. Changes in the microecology of the reproductive tract are significantly associated with embryonic development and prognosis (Stout et al., 2017; Peuranpaa et al., 2022). Therefore, the development and prognosis of the reproductive tract microbiome and embryos have attracted increasing attention.

It is known that during pregnancy, due to the dramatic changes in estrogen and progesterone in women of childbearing age, the epithelial cells at the upper end of the vagina grow, and the production of lactic acid in the body increases, which results in changes in the reproductive tract, the pH and natural barriers of the vagina (Chen et al., 2021). The female genital tract accounts for approximately 9% of the human microbiome (Group NHW et al., 2009). Most previous research has focused on the vaginal microbiome, and less research has been conducted on the microbiome in the endometrium. However, dysbiosis of the endometrial microbiota was associated with lower rates of implantation, pregnancy, ongoing pregnancy, and increased spontaneous abortion obstetric complications (Garcia-Grau et al., 2019). The exact relationship between embryonic arrest and the reproductive tract microbiome is still unclear, and there are few correlation studies. Therefore, it is an urgent problem to determine the reason for embryonic arrest to reduce its risk.

Changes in reproductive tract bacteria during pregnancy are characterized by a decrease in the abundance of Lactobacillus and the diversity of the microflora (Di Simone et al., 2020). The most common clinical manifestation of dysbacteriosis during pregnancy is BV of the reproductive tract (Kovachev, 2018; Jayaram et al., 2020). BV is closely related to perinatal maternal and infant health, such as premature rupture of membranes and preterm delivery (Kaambo, 2017; Juliana et al., 2020). In our study, the incidence of BV in the embryo arrest group was significantly higher than that in the normal pregnancy group. In recent years, the clinical incidence of embryo arrest has increased, and some related studies have confirmed that it is closely related to the microflora of the reproductive tract. In the pathogenesis of BV, which is characterized by a decrease in Lactobacillus vaginalis and an increase in Gardnerella, there is no characteristic inflammatory change in the vagina (Yan et al., 2016). Therefore, we further analyzed vaginal swabs for Lactobacillus and Gardnerella. To our surprise, the abundance of Lactobacillus decreased significantly in the embryo-arrest group (shown in Figure 2). However, there was no significant difference in the abundance of Gardnerella. Based on the above research, we speculate that the abundance of Lactobacillus may be more closely related to embryo arrest, which is also consistent with previous studies (Al-Memar et al., 2019).

The dominant genus Lactobacillus in the vagina of women of childbearing age is a protective bacterium of the vaginal microbiome (Kalia et al., 2020). In patients undergoing in vitro fertilization (IVF), the success rate of IVF is higher if only Lactobacillus is present in the vagina at different time points. Lactobacillus may play a certain role in pregnancy, embryo colonization and normal development during pregnancy (Leitich and Kiss, 2007; Hyman et al., 2012). Previous studies have also confirmed that the bacteria in the female vagina are dominated mainly by lactobacillus species, namely, L. crispatus, L. iners, L. jensenii, and L. gasseri (Romero et al., 2014; DiGiulio et al., 2015; Witkin et al., 2019, 2021). The vaginal microbiome of healthy women has been classified into five community state types (CSTs) according to their community structure. Four types of CST are dominated by Lactobacillus, namely L. crispatus (CST I), L. gasseri (CST II), L. Iners (CST III) and L. jenesii (CST V). CST IV is characterized by low levels or absence of Lactobacillus (Ravel et al., 2011; Liu et al., 2021). Previous studies have reported that lactobacilli in the vagina during pregnancy are dominated mainly by L. crispatus (CST I). In addition, the higher the abundance of L. crispatus is, the lower the abundance of other Lactobacillus species will be (Witkin et al., 2021). To explore the role of the community state types (CSTs) of Lactobacillus in embryonic development, we further classified Lactobacillus in the vagina. In our study, both groups were dominated by L. iners (CST III), but the abundance of L. iners in the embryo-arrest group was significantly lower than that in the normal pregnancy group (shown in Figure 3). At present, research on L. iners is still controversial, and its exact function in the vagina is uncertain. Studies have confirmed that L. iners is a beneficial species and is related to vaginal dynamic balance. It can survive and help restore vaginal homeostasis (Ferris et al., 2004; Petrova et al., 2017). In the transitional stage of the vaginal microbiota, it may become the dominant species related to term delivery in pregnant women and associated with a high risk of adverse pregnancy outcomes (Jakobsson and Forsum, 2007; Romero et al., 2014; Stout et al., 2017). Our research shows that the balance of vaginal microbiology may play a certain role in the occurrence of embryo arrest and that domination of the internal environment by different lactobacilli would have different effects vaginal health in women.

At present, the study of the vaginal microbiome is still the focus of research, and an increasing number of studies have focused on changes in the vaginal microbiome during pregnancy, to infer its impact on pregnancy outcomes. Previous studies have confirmed that the abundance of Lactobacillus increases and that the diversity of bacteria decreases during pregnancy. In our study, the diversity of bacteria in the embryo arrest group decreased significantly, and Firmicutes, Bacteroidetes, Proteobacteria and Actinobacteria were dominant in the two groups at the phylum level; however, at the genus level, the abundance of Lactobacillus in the embryo arrest group decreased significantly, and the corresponding Bacteroides and Helicobacter increased significantly. To identify the bacteria related to embryo arrest in the vagina of the embryo arrest group and the normal pregnancy group, we used random forest analysis to identify the different bacteria and further predict embryo arrest by ROC curve analysis. We found that Ensifer and Bosea had high values in embryo arrest. In this study, we detected two special strains, Ensifer and Bosea. Ensifer was isolated in 1982 and located in the soil. Previously, scholars have reported that Ensifer microbes may be temporary in the skin, and have been found to be present in human skin for hours to days (Suwarsa et al., 2021). Bosea was detected in cervical microflora in a study of the relationship between human papillomavirus infection and age, and it was significantly associated with HPV infection (Hu et al., 2022). Interestingly, these two strains were detected in the vagina of pregnant women with embryo arrest and normal pregnant women, their abundances were low, and there was a significant difference between the two groups; however, their specific effects need to be further studied. The development of next-generation sequencing techniques has improved our understanding of the vaginal microbiota, which represents a unique microbiota (Franasiak et al., 2016; Verstraelen et al., 2016).

For a long time, research on the microbiota of the female reproductive tract during pregnancy has focused mainly on the vagina. Due to the limited collection of intrauterine samples during pregnancy, there are few studies on intrauterine flora during pregnancy (Moreno and Simon, 2018; Riganelli et al., 2020). Previous studies have shown that bacteria can enter and reproduce in the uterus from the vagina and cervix (Ramos Bde et al., 2015). The changes in the endometrial microbiome, characterized by a low abundance of lactobacilli, seem to be related to the maintenance and implantation of embryos. However, the current reports of the microbiota in the uterus vary widely. Due to the low abundance of the microbiome in the uterus, fewer changes in the vagina and uterus, and the possibility of sample contamination is inevitable (Baker et al., 2018). The new study is still exploring the interaction between the microbial profiles in the uterine cavity and the host during pregnancy. At present, the study of the uterine microbiome still needs to overcome some technical challenges to avoid misleading conclusions. The abortion caused by the stagnation of early embryonic development usually occurs in the early stage of pregnancy, so the changes in microflora in the uterine cavity of pregnant women have an important impact on the changes in embryos. We further attempted to use high-throughput sequencing to discover the unique microbiota in the uterine cavity of patients with embryonic arrest and women during normal pregnancy. The bacterial diversity of in the uterine cavity was not significantly different between the embryo-arrest group and normal pregnancy group. The core assemblage microbiota at the phylum level were Proteobacteria, Bacteroidetes, Firmicutes and Actinobacteria similar to the vaginal microbiota. However, the abundances were significantly different from that of the vaginal microbiome. The microbiota in the uterus has the highest proportion of Proteobacteria, and the highest proportion of Firmicutes in the vagina (Toson et al., 2022). At the genus level, we observed that Bacteroides, Pseudarthrobacter, Lactobacillus, Ralstonia and Pseudomonas accounted for the main proportion between the embryo-arrest group and the normal pregnancy group. This is consistent with previous studies reporting that non-Lactobacillus was the dominant genus (Mei et al., 2019; Moreno et al., 2020). However, it has also confirmed that the microbiome in the uterus is not dominated by lactobacilli (Leoni et al., 2019). The bacteria in the endometrium mentioned above not only include Lactobacillus, but are also significantly related to pregnancy.

There is a certain relationship between the imbalance in endometrial microbiota and the determination of the cause of infertility, and changes in the microbiome of may be the reason for disease in women (van Oostrum et al., 2013; Moreno and Simon, 2018). The cervix forms a complete barrier between the uterus and vagina. The mucins in the cervix change their conformation, allowing bacteria to pass through under certain conditions (Brunelli et al., 2007). It has been confirmed that there is a significant correlation between the abundance of lactobacilli and pregnancy outcomes in recurrent pregnancy loss. The abundance of Lactobacillus was significantly correlated with implantation, pregnancy and live birth rate in women. The survival rate of embryos is higher when the microbiome of the uterine cavity comprises Lactobacillus-dominant microbiota at the time of implantation, and pregnancy (Moreno et al., 2016; Shi et al., 2022). However, infection of the pathogenic microbiome in the reproductive tract, especially inflammatory reactions, destroys the structure of the chorion and amniotic membrane, which is toxic to the endometrium and the normal development of embryos, making this an important cause of embryo arrest (Giakoumelou et al., 2016). In our study, we explored the relationship between the endometrial microbiome and pregnancy outcomes and observed significant changes in the intrauterine cavity microbiome in the embryo-arrest group and the normal pregnancy group. We further classified the Lactobacillus, and found that the Lactobacillus in the uterus is also dominated by L. iners, L. crispatus, L. gasseri and L. jensenii. To our surprise, L. jensenii in utero was significantly different between the embryo-arrest group and normal pregnancy groups. In the vagina, the abundance of the above microbiota is significantly related to preterm delivery. In pregnant women with missed abortion, this abundance is significantly decreased (Sun et al., 2022). The exact role of Lactobacillus species in the uterine cavity has not been reported thus far and its exact effect on embryo arrest remains to be explored. More research is still needed on the relationship between the microflora in the uterine cavity and embryo arrest in order to characterize the different microbiomes in utero and evaluate their predictive value for embryo arrest. Through random forest analysis, we analyzed the ROC curves of the bacteria with significant differences in the uterus. We found that the Eubacterium xylanophilum group, Candidatus Symbiobacter, Odoribacter, Blautia, Nocardioies and Ileibacterium have a certain predictive value for embryo arrest. Thus, the specific mechanism of the above bacteria in the implantation and outcome of embryos in the uterus as well as the interactions of bacteria in the uterine cavity need to be further studied.

Thus, it can be seen that the balance of vaginal microecology and uterine microflora during pregnancy is very important for the development of embryos. There are many reports on the balance of microecology in the vagina during pregnancy, but it is also affected by a variety of factors, including the interaction of bacteria to maintain the relative balance of the vagina. Embryo arrest is a complex process, not only in the vaginal environment, but also in the uterine cavity. Our study explored the effect of the environment in the reproductive tract on embryo arrest from two aspects of the difference in vaginal and uterine flora, to provide some research ideas for the prevention of embryo termination.

With the support of high-throughput sequencing technology, we performed a detailed study on the uterine and vaginal microbiomes of patients with embryo arrest and women with normal pregnancies. There were significant differences in the uterine and vaginal microbiomes between these two groups. The above study of the microbiome provides insights into the pathogenesis of embryo arrest and, according to our evidence. Additionally, our findings indicate that some bacteria have a certain predictive value for embryo arrest. However, due to the limitation of sample size and specimen collection, the predictive value of the microbiome in the uterus and vagina for embryo termination is limited, and further prospective studies are needed to determine the predictive value of the microbiome.

The data presented in the study are deposited in the NCBI repository, accession number PRJNA933192.

The studies involving human participants were reviewed and approved by Chengdu Women’s and Children’s Central Hospital (grant no. 2021(88)). The patients/participants provided their written informed consent to participate in this study.

LW and ML contributed to the study design. LH, JC, HLiu, and YL completed all tests. LW, ML and WL analyzed the data and drafted this paper. HLi and ZL contributed to reagent management. All authors contributed to the article and approved the submitted version.

The clinical case screening and data collection were supported by the Chengdu Science and Technology Bureau: technology innovation research and development project [2021-YF05-00648-SN] and Chengdu Medical Research Project [2022203].

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.