Adverse pregnancy outcomes (APOs) include all pathological pregnancy and childbirth complications. Many public health and medical interventions have been implemented to reduce APOs, yet the incidence of APOs remains high. APOs are a leading cause of maternal, fetal, and neonatal deaths; severe maternal morbidity; and maternal intensive care unit admissions (Rana et al., 2019). APOs are widely recognized as major public health problems due to short- and long-term adverse impacts on mothers and infants.

The gut microbiota is a massive complex community of microbial species living in the digestive tract that not only participates in the digestion of food and the absorption of nutrients but also is involved in physiological regulation of the host through the production of hormonally active substances (Luo et al., 2022). Increasing evidence indicates that microbiome dysbiosis might play an important role in the pathogenesis of APOs (Qi et al., 2021). Zheng et al. (2020) found that dynamics in gut microbiota during the first half of pregnancy differed significantly between gestational diabetes mellitus (GDM) and normoglycemic pregnant women. Another case–control study showed that pregnant women with pre-eclampsia (PE) had a high abundance of Fusobacterium and Veillonella and a relatively low abundance of Faecalibacterium and Akkermansia compared with normotensive pregnant women (Chen C. et al., 2020; Chen X. et al., 2020). Yin et al. (2021) reported that women with preterm birth (PTB) showed a distinct gut microbiome dysbiosis compared with those who delivered at term. Opportunistic pathogens, particularly Porphyromonas, Streptococcus, Fusobacterium, and Veillonella, were enriched, while Coprococcus and Gemmiger were markedly depleted in the preterm group (Yin et al., 2021). However, whether these associations are causal has not been established because of potential biases including residual confounding factors and reverse causality.

Mendelian randomization (MR) is an approach integrating summary data of the genome-wide association study (GWAS) and resembles a randomized controlled trial by leveraging allelic randomization during meiosis and subsequent irreversible exposure to genotype at conception (Lawlor et al., 2008; Ellervik et al., 2019; Ference et al., 2019). Using genetic variants as instrumental variables (IVs), the MR design is typically less likely to be affected by residual confounding factors and reverse causation than conventional observational analysis, thereby strengthening the causal relationship between exposure and outcomes (Chen C. et al., 2020; Chen J. et al., 2020; Yuan and Larsson, 2022).

Exploring causal associations between gut microbiota and APOs not only deepens our understanding of pathogenesis derived from enteric bacteria but also facilitates the development of personalized treatments for pregnancy complications based on dietary or microbiome interventions (Di Simone et al., 2020). Elucidating the causal inference on the gut microbiome and its association with APOs is warranted. Thus, in this study, we conducted a two-sample MR analysis to evaluate the causal association between gut microbiota and six common APOs, namely, postpartum hemorrhage (PPH), abruptio placenta (AP), spontaneous abortion (SAB), premature rupture of membrane (PROM), gestational diabetes mellitus (GDM), and pre-eclampsia (PE) or eclampsia (Figure 1). To our knowledge, this is the first study assessing the causal role of gut microbiome in the development of pregnancy complications.

In the present study, we found causal relationships between genetically predicted abundance of specific bacterial taxa and six common pregnancy complications (PPH, AP, SAB, PROM, GDM, and PE or eclampsia) using large-scale GWAS summary data.

Considerable interest in the potential molecular mechanisms of the gut microbiome in the pathogenesis of PPH was evidenced by an abundance of prior studies. The gut microbiome composition was found to be involved in the pathogenesis of PPH by regulating inflammation, which are closely associated with development of PPH (Farhana et al., 2015; Yang et al., 2021). We successfully identified that Butyricimonas, Marvinbryantia, Oscillibacter, and Peptococcus were negatively associated with the risk of PPH. Butyricimonas, Marvinbryantia, and Oscillibacter can produce butyrate, which participates in regulating gut permeability and induces the inflammatory response (Jang et al., 2020; Chen B. D. et al., 2021; Chen Z. et al., 2021; Liu et al., 2022). Prior studies have reported that the abundance of Peptococcus increases when intestinal inflammation occurs, indicating its pro-inflammatory effects (Sun et al., 2022). The exact mechanism of Peptococcus on the pathogenesis of PPH warrants verification studies. This study also suggested that increased abundance of Deltaproteobacteria was causally related to a higher risk of PPH. Deltaproteobacteria is a common pathogenic bacterium, which initiates inflammation by activating the Notch signaling pathway and induces the expression of pro-inflammatory factors including pro-IL-1β and SOCS3 (Jin et al., 2016; Singh et al., 2021).

Emerging evidence has shown the role of perturbations in mitochondrial biogenesis, oxidative phosphorylation, and oxidative stress pathways in the pathogenesis of AP (Workalemahu et al., 2018). The relationship between gut microbiota dysbiosis and AP remains unclear. This study found that the abundance of Butyricimonas and Holdemanella reduced the risk of AP. The association may be mediated by the short-chain fatty acid butyrate as Butyricimonas and Holdemanella are involved in its production (Zagato et al., 2020; Jing et al., 2021). Butyrate has a key role in maintaining mucosal integrity, reducing inflammation via macrophage function, and regulating oxidative stress (Hu et al., 2020; Jing et al., 2021). We also showed that abundance of Eggerthella and Escherichia Shigella was positively associated with the risk of AP. Eggerthella is a pro-inflammatory genus, which has been associated with the depletion of butyrate (Nikolova et al., 2021). Escherichia Shigella can also mediate inflammation and oxidative stress by regulating the lipopolysaccharide/TLR4/NF-kB signaling pathway (Peng et al., 2020). This study demonstrated that Roseburia and Porphyromonadaceae, butyrate-producing bacteria, were positively related to the risk of AP (Wang et al., 2022). It can suggest that a single microbial taxon may not affect the development of AP, and the overall role of gut microbiota may have a more important role in the onset of AP. We showed that some novel gut bacteria including Acidaminococcaceae, Victivallaceae, Howardella, and Prevotella9 were closely associated with the pathogenesis of AP, but the exact mechanism of these gut microbiome on the development of AP warrants additional investigations.

We discovered that the abundance of Actinomyces, Subdoligranulum, and Veillonella was associated with an increased risk of SAB, while a high abundance of Lactococcus was related to a lower risk of SAB. Immunologic dysfunction and thrombophilia are known key factors in the pathogenesis of SAB (Liu et al., 2021a). A previous study showed that the relative abundance of Actinomyces and Subdoligranulum is closely related to multiple autoimmune diseases, including systemic lupus erythematosus and chronic spontaneous urticaria, suggesting that these bacteria play an important role in the host immune response (Chen B. D. et al., 2021; Liu et al., 2021). The protective effect of Lactococcus on SAB was likely due to lactic acid bacteria reducing autoimmunity and recovering immune homeostasis (Xu et al., 2020). Increased thrombophilia is one of the clinical signs of SAB (de Jong et al., 2014). Veillonella may influence the host trimethylamine N-oxide (TMAO) level by mediating carnitine production and metabolism, which contribute to platelet hyperreactivity and thrombosis (Yang et al., 2019). The exact the mechanism of gut microbiota leading to SAB remains poorly understood and warrants additional studies.

Although exact mechanisms underlying the occurrence of PROM are not fully elucidated, oxidative stress and inflammation have been reported to play an important role (Liu et al., 2021b). Our study revealed that the genetically predicted abundance of Turicibacter was positively correlated to the risk of PROM. Animal studies have reported that the abundance of Turicibacter was negatively associated with propionic acid and butyric acid, which can inhibit the inflammatory response (Wan et al., 2021). We have found that an increased abundance of Marvinbryantia and Ruminococcaceae, which are known to produce short-chain fatty acid, was negatively associated with the risk of PROM (Chen B. D. et al., 2021; Chen Z. et al., 2021).

Previous studies on GDM or post-GDM women with impaired glucose tolerance discovered a broad range of gut microbiota dysbiosis, which was associated with Collinsella (Crusell et al., 2018), Olsenella (Crusell et al., 2018), Prevotella9 (Hasain et al., 2020), and Ruminococcus2 (Crusell et al., 2018). On the other hand, beneficial butyrate-producing bacteria including Methanobrevibacter (Kuang et al., 2017) and Oscillibacter (Crusell et al., 2018) were depleted. A prior study also indicated that individuals with diminished insulin sensitivity had a lower abundance of Tenericutes, suggesting it might be associated with the occurrence and development of GDM (Yuan X. et al., 2021). Larsen et al. found that the relative abundance of Betaproteobacteria was higher in the type 2 diabetes group and positively correlated with the plasma glucose level (Larsen et al., 2010). These results were consistent with the present study. The mechanisms through which these bacteria exert beneficial or detrimental effects on the pathogenesis of GDM deserve to be studied.

Our study also showed that the abundance of Eubacterium ruminantium, Eubacterium ventriosum, Ruminococcaceae UCG002, Ruminococcus1, and Enterobacterales was negatively linked to the risk of PE or eclampsia. Prior studies indicated that chronic peripheral and placental inflammation were critical to the onset of PE (Black and Horowitz, 2018). Eubacterium, Ruminococcaceae, and Enterobacterales are involved in the production of short-chain fatty acids, such that their depletion is causally linked to inflammation (Tran et al., 2019; Zhuang et al., 2020). Wang et al. (2019) showed that fecal microbiota showed a significant reduction of Ruminococcus in the PE group, which was important in maintaining gut homeostasis. This study also demonstrated that some bacteria such as Prevotellaceae, Enterobacteriaceae, Enterorhabdus, and Euryarchaeota were involved in the pathogenesis of PE. Future studies are warranted to elucidate the functional significance of these bacteria in the onset of PE.

The advantages of this study as follows: First, the MR design diminishes the residual confounding and reverse causality and thereby improves the causal inference in relationships of gut microbiota and APOs. Second, we comprehensively studied up to six common APOs. Third, the identified causal associations in our MR analysis may provide candidate microbial taxa for subsequent functional studies and help in the development of new approaches for the prevention and treatment of pregnancy complications by targeting specific gut bacteria.

The limitations of our study must also be acknowledged. First, SNPs obtained based on the genome-wide statistical significance threshold (5 × 10⁻⁸) were too limited for further study; therefore, we included the SNPs that met the locus-wide significance level (1 × 10⁻⁵) in this study. Second, while the GWAS summary data in this study were obtained mostly from European populations, only a small amount of the gut microbiome data was taken from other races, which may partially bias our findings. Third, we cannot exclude the possible diet–gene or gene–environment interactions on outcomes, which may influence the observed results.

In conclusion, this MR study comprehensively explores causal effects of gut microbiota on APOs for the first time. Our findings provide new insights into the prevention, course of disease, and treatment of APOs by targeting specific bacteria taxa. Further studies are warranted to identify the exact mechanism of enteric microbiota–pregnancy complication associations.

Summary data used for this study can be accessed through the following links: microbiota, https://mibiogen.gcc.rug.nl/; post-partum hemorrhage, https://r5.finngen.fi/pheno/O15_POSTPART_HEAMORRH; abruptio placenta, https://r5.finngen.fi/pheno/O15_PLAC_PREMAT_SEPAR; spontaneous abortion, https://r5.finngen.fi/pheno/O15_ABORT_SPONTAN; premature rupture of membrane, https://r5.finngen.fi/pheno/O15_MEMBR_PREMAT_RUPT; gestational diabetes mellitus, https://r5.finngen.fi/pheno/GEST_DIABETES; pre-eclampsia or eclampsia, https://r5.finngen.fi/pheno/O15_PRE_OR_ECLAMPSIA.

This study was approved by Ethics Committee of Shengjing Hospital of China Medical University.

CLi and CLiu designed the research. CLi collected and analyzed the data and drafted the manuscript. CLiu and NL supervised the study. CLi, CLiu, and NL were involved in writing the manuscript. All authors contributed to the article and approved the submitted version.