You are here

Article Text

Article menu

Download PDFPDF

Pregnancy and peri-partum
Stress and mental health symptoms in early pregnancy are associated with the oral microbiome
  1. Ann M Alex1,
  2. Alytia A Levendosky2,
  3. G Anne Bogat2,
  4. Maria Muzik3,
  5. Amy K Nuttall4,
  6. Rebecca C Knickmeyer1,5,
  7. Joseph S Lonstein2
    1. 1 Institute for Quantitative Health Science & Engineering, Michigan State University, East Lansing, Michigan, USA
    2. 2 Department of Psychology, Michigan State University, East Lansing, Michigan, USA
    3. 3 Department of Psychiatry and Department of Obstetrics and Gynecology, University of Michigan, Ann Arbor, Michigan, USA
    4. 4 Department of Human Development and Family Studies, Michigan State University, East Lansing, Michigan, USA
    5. 5 Department of Pediatrics and Human Development, Michigan State University, East Lansing, Michigan, USA
    1. Correspondence to Dr Joseph S Lonstein, Department of Psychology, Michigan State University, East Lansing, Michigan, USA; lonstein{at}msu.edu

    Abstract

    Background Research has revealed associations between microbes of the gastrointestinal tract and stress, anxiety and depression in pregnant or postpartum women. While these studies suggest a gut-brain-behaviour axis, no studies have examined microbes of the oral cavity in relation to maternal mental health.

    Objective To explore a potential oral-brain-behaviour axis related to maternal mental health.

    Methods Microbes were measured in saliva obtained from 224 second-trimester (mean±SD = 17±2 weeks) women oversampled for stress. Oralome data were associated with women’s recent or cumulative pregnancy stress, trait and state anxiety, depression symptoms and post-traumatic stress disorder (PTSD) symptoms. Covariates explored included age, income, alcohol and tobacco use, dental issues and physical health problems.

    Results Pregnant women in the high trait anxiety or depression symptom groups had higher oral alpha diversity, indicating higher richness of species within samples. Groups with high and low PTSD symptoms differed in beta diversity, reflecting differences in community composition. Linear discriminant analysis showed differently abundant microbes in women with high stress versus low life stress, anxiety, depression, and PTSD, with the affected microbes mostly differing by symptom. Notably, members of phylum Proteobacteria were more abundant in women with high recent life stress and Spirochaetes was more abundant in women with high depression symptoms. Members of phylum Firmicutes were more abundant in the high trait anxiety and high depression groups. Genus Dialister (previously found to be lower in the gut of depressed non-pregnant people) was higher in women experiencing either high trait or state anxiety, or experiencing high depression symptoms, while genus Eikenella was elevated with high trait anxiety, depression or PTSD.

    Conclusions The oral microbiome is associated with stress and mental health in pregnant women, in ways different from the gut microbiome or what has been found in non-pregnant people.

    Clinical implications Understanding oral microbiome–mental health relations may reveal future microbial targets to improve maternal psychological well-being.

    • Depression
    • Anxiety disorders
    • Adult psychiatry

    Data availability statement

    Data are available upon reasonable request. Data are available upon reasonable request to JSL or AMA.

    http://creativecommons.org/licenses/by-nc/4.0/

    This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

    https://doi.org/10.1136/bmjment-2024-301100

    Statistics from Altmetric.com

      Request Permissions

      If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

      WHAT IS ALREADY KNOWN ON THIS TOPIC

      • High life stress and poor mental health are known to be associated with changes in the gut microbiome.

      • A clinical trial targeting the gut microbiome of pregnant and postpartum women with probiotics reported improved anxiety and depression symptoms.

      • Many fewer studies have examined relations between life stress and mental health symptoms and the oral microbiome; none have done so in the context of maternal mental health in pregnancy.

      WHAT THIS STUDY ADDS

      • This is the first study to associate life stress and mental health symptoms with the oral microbiome of 224 second-trimester women oversampled for high life stress.

      • The oral microbiome was differentially associated with life stress, anxiety, depression and post-traumatic stress disorder symptoms.

      • The results differed from studies of the gut microbiome in pregnancy and studies of non-pregnant samples.

      HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

      • This study highlights the need for further focus on oral cavity microbes in order to understand maternal mental health, and possibly mental health outside the context of motherhood.

      • The findings point to the oral microbiome as a potential target for ameliorating poor psychological well-being during pregnancy; this could have important consequences for both mother and child.

      Background

      Poor mental health affects >15–20% of pregnant women worldwide1 and significantly reduces maternal well-being, impairs mothers’ capacity for infant caregiving, and imperils child development.2 The biological determinants of poor maternal mental health are numerous and include circulating steroid and peptide hormones, neurosteroids, monoamine neurotransmitters and inflammatory cytokines.3 Recent studies have further proposed relations between maternal mental health and the bacteria existing on and within the human body.4 These microorganisms in the upper and lower gastrointestinal tracts, oral cavity, lungs, genitourinary tract, skin and elsewhere have been thought to be both causes and consequences of our physical health.5 These relations have been extended to mental health. For instance, a meta-analysis confirmed differences between people with and without major depression in their lower gut microbiome and also that altering the gut microbiome with probiotics significantly improved depression symptoms.6

      A handful of studies have investigated links between the lower gut microbiome and maternal mental health. They found that high emotional distress or generalised anxiety during pregnancy is associated with concurrent changes in levels of particular faecal microbes, but not with either alpha diversity (reflecting microbe species richness and/or evenness within samples) or beta diversity (reflecting differences in microbe composition between samples).6–8 In contrast, postpartum depression is associated with changes in particular faecal microbes as well as beta diversity.8 Causal relations are suggested by a double-blind, placebo-controlled study chronically administering the probiotic Lacticaseibacillus rhamnosus HN001 starting in the second trimester of pregnancy, resulting in small but statistically significant reductions in maternal anxiety and depression symptoms when assessed 6 months post partum.9

      Importantly, the gut is not the only source of microbes capable of influencing maternal mental health. The oral cavity is another rich source of bacteria and, while the oral microbiome is most often studied for associations with dental problems, it is now thought to also influence neurobiological function including mental illness and other neuropsychiatric disorders.10–14 The mechanisms through which oral microbes alter brain function are probably numerous and have been suggested to include stress-induced pro-inflammatory oral cytokines impinging on the brain, oral microbes modifying central neurotransmitter synthesis, or stress-altered oral microbes infiltrating the gut to then affect the brain.15 Saliva is a convenient source of oral microbes and contains microorganisms from salivary fluid itself in addition to those from the tongue, tonsils and throat.10 No extant studies of the oral microbiome and mental health have involved pregnant or postpartum samples. Research with other groups indicates that levels of a number of specific bacterial taxa in saliva obtained from adolescents and young adults differ between those with high anxiety, depression symptoms or suicide ideation compared with unaffected individuals.16–18 Furthermore, adults with high psychological distress (including measures of life stress, anxiety and depression symptoms) differed from individuals with low psychological distress in the alpha and beta diversities of their oral microbiome and levels of numerous individual microbes.19 Perhaps surprisingly, there is very little overlap in the oral microbes affected by anxiety symptoms and those affected by depression symptoms,16 20 indicating the value of using multiple measures of psychological distress to independently associate with the oral microbiome.

      Objective

      This study’s objective was to explore relations between measures of life stress, anxiety, depression and post-traumatic stress disorder (PTSD) symptoms and the oral microbiome (alpha diversity, beta diversity, microbe profile) in a large, ethnically diverse, at-risk sample of women in their early second trimester of pregnancy. We studied this time of pregnancy because for many women their earlier pregnancy stress and poor mental health persist throughout pregnancy and post partum21 and predict more severe symptoms post partum.22 We hypothesised that (1) the oral microbiome would differ between women with high and low stress and mental health symptoms and (2) high and low stress, anxiety, depression and PTSD symptoms would differentially affect the levels of particular oral microbes.

      Methods

      Participants

      Participants were a subsample of 224 pregnant women participating in the Michigan Prenatal Stress Study, a large longitudinal study of the influence of the timing of pregnancy stress on women and children (see Levendosky et al 23 and online supplemental figure 1 for details). Women were recruited in early pregnancy using social media, organisations serving pregnant women and flyers posted in the community. Women contacted project offices for screening and the inclusion criteria included: (1) English fluency, (2) 18–34 years old, (3) no endocrine disorders, cancer or cancer treatments affecting stress hormones, (4) in a heterosexual romantic relationship during the current pregnancy, (5) <20 weeks pregnant with a singleton, (6) employment that does not involve working at night or shift work. To increase the range of stress represented in our sample, part way through recruitment eligibility was extended to require recent intimate partner violence (the most common traumatic stressor in pregnancy) and household financial stress (below US Medicaid cut-off point). Participants with current antibiotic use were excluded from the present analyses. Current dental problems (sores, bleeding gums, untreated cavities) were reported by 33% of the participants, 20.5% reported anaemia, 10.7% high blood pressure and 5.8% diabetes. See online supplemental table 1 for demographics.

      Supplemental material

      Supplemental material

      Saliva sampling

      During in-person laboratory visits to our university research sites, women were instructed in how to collect saliva via passive drool. They were given kits containing plastic tubes, straws and written instructions on how to collect saliva at home across 2 consecutive days within 1 week of their in-person interview. Saliva samples were kept in participants’ home refrigerators until retrieved by study staff within 1 week, after which samples were frozen and stored at −80°C until microbe analysis. The present study’s microbe analysis used samples provided immediately on awakening before eating, drinking or performing oral hygiene behaviours.

      Life stress and mental health variables

      Recent life stress

      Recent stress experienced during the same week of saliva sample collection was assessed by asking ‘How stressed were you this past week?’. Participants ranked their stress level on a 1–7 Likert scale, with 7 the highest stress level. The median response was 4 (72 participants), so participants with scores >4 were placed in a high stress group (n=65) and participants with scores ≤4 were in a mild-to-moderate stress group (n=155).

      Pregnancy negative life events

      In addition to recent stress, we obtained information on the perceived impact of stressful events since the beginning of pregnancy (previous ~17 weeks) by using self-reported negative rating scores on the 47-item Life Experiences Survey. Each item was rated on a scale of 0 (no impact) to −3 (extremely negative). A median split resulted in low negative impact (n=112; < −7 score) and high negative impact (n=111; ≥ −7 score) groups.

      State and trait anxiety

      The 20 questions of the state and the 20 questions of the trait subscales of the State-Trait Anxiety Index (STAI) were self-rated on a Likert scale from from 1 (not at all) to 4 (very much) to provide a range of total summed scores of 20–80. The STAI-State asks about anxiety symptoms ‘right now—that is, at this very moment.’ while the STAI-Trait asks about anxiety symptoms related to how responders ‘generally feel’. An established clinical cut-off point of summed score >4024 was used to create the high (n=96) and low state anxiety symptom (n=126) groups and to create the high (n=122) and low (n=99) trait anxiety symptom groups.

      Depression

      The 10-item Edinburgh Postnatal Depression Scale assessed depression symptoms for the previous week. Each question is rated on a 4-point Likert scale ranging from 0 to 3 to provide total summed scores of 0–30. An established clinical cut-off summed score of 1125 was used to create groups with high depression symptoms (n=104) and with low depression symptoms (n=118).

      PTSD

      PTSD symptoms were measured using the post-traumatic stress checklist for DSM-5, a 20-item checklist measuring PTSD symptom severity. Women rated how much each symptom had disturbed them in the past 30 days on a scale ranging from 0 (‘not at all’) to 4 (‘extremely’). A cut-off score of 32 was used to create high (n=75) and low (n=142) PTSD symptom groups.26

      DNA sequencing

      Frozen saliva samples underwent 16S rRNA gene sequencing. DNA was extracted according to the manufacturer’s protocol. Extracted DNA was quantified using Qubit dsDNA high-sensitivity assay kits (Thermo-Fisher Invitrogen) and normalised to 5 ng/µL. The V3–V4 variable region of the 16S rRNA gene was amplified using 341F and 805R primer pair by following the Illumina 16S metagenomic sequencing library preparation protocol (Illumina), but with increased input of 25 ng template DNA in a 50 µL-volume reaction for initial PCR, then cleaned and indexed using Nextera XT v2 indices. Final indexed libraries were assessed on a 1.5% agarose gel and quantified using Qubit dsDNA high-sensitivity assay kits (Thermo-Fisher Invitrogen). Cleaned libraries were pooled equally and quality control of the final pool was carried out on an Agilent Bioanalyzer 2100 using the high-sensitivity DNA chip assay (Agilent) and quantified by qPCR with KAPA Library Quantification kits for Illumina (Roche KAPA). Sequencing was performed on a MiSeq system using a 2×300 bp v3 kit.

      Covariates

      Twenty-two demographic and health variables assessed in the larger study23 were evaluated for possible selection as covariates in the oral microbe profiling (online supplemental table 2). Cigarette smoking, dental problems, total intimate partner violence perpetration and whether the pregnancy was planned or not explained significant proportions of the variance in the oral microbial profiles (7.2%, 3.1%, 4.1% and 2%, respectively; p<0.05). Two of these variables, cigarette smoking and dental problems, were significant after false discovery rate correction for multiple comparisons and were included as covariates in the permutational multivariate analysis of variance (PERMANOVA) and regression analysis.

      Analytic strategy

      The microbiome analysis was performed in QIIME2_2021.4. The fastq sequences were imported into QIIME2, the raw sequences were demultiplexed and quality filtered using the q2‐demux plugin and visual summary obtained. The primers were trimmed using the cutadapt plugin. Denoising was done using the DADA2 plugin via q2-dada2 to identify all observed amplicon sequence variants. A phred >20 was considered for sequence trimming. The amplicon length was 301 nucleotides. The mean number of reads was 89 613.2875, with a median of 87 660 and range of 104–136 779. All amplicon sequence variants were aligned with mafft (via q2 alignment) and used to construct a phylogeny with fasttree 2 (via q2‐phylogeny).

      Alpha and beta diversity metrics were estimated using q2‐diversity after samples were rarefied (subsampled without replacement). Sequencing depth was kept at 27 000, which retained 6 399 000 (52.02%) features in 223 (99.6%) samples at the specified sampling depth. One sample was omitted owing to low feature count. We generated four measures of alpha diversity: Chao1, Shannon Entropy, Faith’s Phylogenetic Diversity and Observed Species. A principal component analysis was performed on the diversity measures and the resulting alpha diversity principal component 1 (AD PC1) and alpha diversity principal component 2 (AD PC2) used as outcome measures in further analysis. Groups were compared using Kruskal-Wallis tests. The following beta diversity indices were calculated: Bray-Curtis dissimilarity, unweighted UniFrac and weighted UniFrac. PERMANOVA analysis assessed beta diversities between groups while controlling for cigarette smoking and dental problems. In addition, the first principal coordinates for each beta diversity index were used as outcome measures in regression analysis.

      The envfit function in vegan package calculated the effect size and significance of each potential covariate. Ordination was performed using non-metric multidimensional scaling based on Bray-Curtis dissimilarity. Significance was calculated after performing 10 000 permutations and p values adjusted for multiple comparisons using false discovery rate adjustment. The relations between stress, anxiety, depression and PTSD symptoms and microbial diversity were further assessed using multivariate linear regression analysis, controlling for the two significant covariates in the envfit analysis.

      Taxonomic classification was done based on the Human Oral Microbiome Database, version 15.23. Differential abundance of microbial features between different classes or groups derived from stress, anxiety, depression, and PTSD symptoms scores were analysed by linear discriminant analysis (LefSe)27 using the MicrobiomeMarker package in R and multivariable association analysis using MaAsLin2 package in R.28 29 Statistical significance was indicated by a p value <0.05. Cigarette smoking and current dental problems were controlled for in the MaAsLin analysis. We also performed a follow-up analysis to investigate whether the participants clustered in groups based on the six mental health variables we used in our analysis. We used a hierarchical clustering method implemented using the hclust function of cluster R package30 and our clustering analysis resulted in two clusters with weak-to-moderate clustering. To further investigate if the participants' oral microbiota is associated with these symptom-based clusters, we then performed PERMANOVA using adonis function from the vegan R package.31

      Results

      Participants’ life stress, anxiety, depression and PTSD symptom measures were almost all significantly correlated with each other (online supplemental table 3). Analysis of the microbiome composition revealed the three most abundant families in our samples to be Streptococcaceae (26.2%), Prevotellaceae (20.0%) and Veillonellaceae (12.0%). Within-individual microbial diversity was assessed by principal component analysis of the four alpha diversity metrics (Chao1, Shannon Entropy, Faith’s Phylogenetic Diversity and Observed Species) to create AD PC1 and AD PC2. AD PC1 appears to capture species richness and AD PC2 appears to capture evenness and taxonomic richness. AD PC1 and AD PC2 explained 82.4% and 12.7% of the variance in the oralome, respectively (figure 1).

      Figure 1
      Figure 1

      Alpha diversity correlation coefficients for principal component 1 of the oralome of pregnant women.

      For AD PC1, differences between the groups of women with high or low trait anxiety or with high or low depression symptoms were statistically significant. Pairwise comparisons between high and low symptom groups revealed significantly higher species richness in pregnant women with high trait anxiety or high depression symptoms (table 1). In contrast, AD PC2 was not associated with any of the stress or mental health variables (table 2). Unlike the between-group linear regression analysis just described, the multivariate regression analysis revealed that the PC1 and PC2 alpha diversity measures were not associated with any of the predictor variables (online supplemental tables 4 and 5).

      View this table:
      Table 1

      Kruskal-Wallis tests for change in PC1 between low-symptom and high-symptom groups

      View this table:
      Table 2

      Kruskal-Wallis tests for change in PC2 between low-symptom and high-symptom groups

      The weighted unifrac and Bray-Curtis measures of beta diversity showed significant differences in microbe community structure between the groups with high and low PTSD symptoms (WU - PERMANOVA, F=2.12, p=0.037; BC - F=2.12, p=0.006). Differences between PTSD symptom groups were also found to be marginally significant in the multivariate association analysis with the PC1 of Bray-Curtis measure of beta diversity. However, the oral microbial community structure did not differ between participants with high or low life stress, high and low state or trait anxiety, or high and low depression symptoms (online supplemental table 6). None of the stress, anxiety, or depression symptom measures were associated with beta diversity PCs in the multivariate association analysis (online supplemental table 7). Our follow-up analysis that clustered participants based on their overall stress and mental health status (i.e., collapsed across the six predictor variables) revealed that they group into two clusters of moderate strength. PERMANOVA analysis based on the Bray-Curtis measure of the clusters to assess for relationships with the oral microbiome, however, revealed non-significant results (F=0.75, p=0.68).

      Linear discriminant analysis and multivariable association analysis revealed several distinct, differently abundant bacterial genus between women with high or low recent life stress, negative life events since the beginning of pregnancy, state or trait anxiety, depression symptoms, or PTSD symptoms (figure 2) with limited overlap among them. As examples, phylum Proteobacteria was more abundant in women experiencing high recent life stress (but intrestingly not higher impact of negative life events since the beginning of pregnancy), some members of phyla Bacteroidetes and Firmicutes were more abundant in women with high trait anxiety, and other members of Firmicutes and phylum Spirochaetes were more abundant in women with high depression symptoms. However, genus Dialister (family Bacillota) was in higher abundance in women with either high anxiety (state or trait) or depression symptoms. Genus Eikenella was abundant in the women with high trait anxiety, depression symptoms or PTSD symptoms. The several significant associations between stress or mental health and particular microbes revealed by both the univariate and multivariate analysis instil greater confidence in the differences observed. See online supplemental tables 8 and 9 for differentially abundant bacteria for each of the six stress and mental health predictor variables identified from the LefSe and MaAsLin2 analyses.

      Figure 2
      Figure 2

      Cladograms showing differentially abundant microbial genera in groups of pregnant women with high or low recent stress, stress since the beginning of pregnancy, state anxiety, trait anxiety, depression symptoms or post-traumatic stress disorder (PTSD) symptoms.

      Discussion

      Numerous studies have linked features of the gut microbiome to maternal stress and mental health,6–8 whereas research on linkages between the oral microbiome and psychosocial variables is just emerging.10–20 To date, no published work has reported on the relation of stress or mental health with the oral microbiome in pregnant women. Our novel research associated the oral microbiome assayed from saliva with life stress, anxiety, depression and PTSD symptoms in a large sample of second-trimester pregnant women. We hypothesised that the oral microbiome would differ between groups of women with high or low stress, anxiety, depression or PTSD symptoms and that membership in each high-symptom group would be associated with mostly unique changes in microbial taxa. Our results generally supported our hypotheses. Women in the high trait anxiety or depression symptom groups showed higher alpha diversity; women in the high PTSD group showed altered beta diversity; and membership in the high recent life stress, anxiety, depression or PTSD symptom groups was associated with mostly different changes in the abundance of particular oral microbes.

      Maternal oral microbial population diversity

      Our finding of greater oral microbiome alpha diversity accompanying membership of the high trait anxiety or depression symptom group is consistent with a study investigating oral microbiome–mental health relations in men and non-pregnant women,19 but two other studies did not find such a relationship.16 17 These inconsistencies may be due to the older age of the participants in our study and in that by Kohn et al.19 This possibility is supported by a cross-sectional healthy community sample study finding that some alpha diversity measures became more homogeneous, and phylogenetic diversity and species richness increased, with age.32 In addition to age, reproductive history is highly relevant, with early pregnancy a unique time for the oral microbiome.33 Our results then caution that no particular group’s oral microbiome is likely to be universal. Any advantage of our greater alpha diversity in higher trait anxiety or depression symptoms is unknown, given that the gut microbiome literature suggests that greater microbe diversity reflects positive mental and physical health, although Kohn et al 19 highlight that higher oral alpha diversity is in fact associated with poorer rather than better oral health. Our lack of a relationship between state anxiety and alpha diversity also suggests greater relevance of trait-level or chronic mental health symptoms with this diversity feature of the oral microbiome.

      In contrast to alpha diversity, beta diversity was unrelated to anxiety or depression symptoms in our participants, but was instead related to women’s total PTSD symptom scores. Beta diversity differences in the gut microbiome have been found between groups of women with no trauma experience and those with both trauma and probable PTSD,34 but not found in other studies involving both men and women.35 36

      Maternal oral microbial population abundance

      Our results revealed greater abundance of phylum Proteobacteria in women experiencing high recent life stress compared with women with low/moderate life stress. The fact that this was not found between groups with high and low impact of negative life events since the beginning of pregnancy suggests greater importance of recent over more distal stress. Recent life stress has elsewhere been associated with the oral microbiome, including increased alpha diversity in terms of richness (i.e., Chao-1) in healthy medical students,37 and a trend for relations between cortisol levels or self-reported stress with particular oral microbes in patients with periodontal disease.38 Causal evidence for possible involvement of recent stress-released cortisol in our study could be inferred from an in vitro study showing changes in oral microbe gene expression after acute incubation of the samples with cortisol.39 While the cortisol response of the hypothalamic-pituitary-adrenal axis to stress is suppressed in pregnancy, this effect is most evident in the third trimester, so we were probably able to reveal positive associations between oral Proteobacteria and stress-induced rises in cortisol from our second-trimester saliva sampling.

      Proteobacteria becomes the dominant oral cavity phylum in healthy pregnant women, while in non-pregnant women it is Firmicutes.33 Proteobacteria are thought to cause neuroinflammation by affecting blood-brain-barrier permeability to other bacteria, and oral Proteobacteria levels are positively related to systemic inflammatory markers in adolescents.16 Inflammation markers are also higher in pregnant women with recent life stress.40 Salivary Proteobacteria levels are lower rather than higher in young adults who received institutionalised care before being adopted in early infancy,41 again suggesting that recent stress is more likely than distal stress to drive an increase in salivary Proteobacteria. Proteobacteria was not associated with state anxiety in our sample, which was assessed in the same week as saliva sampling, indicating that not all measures of recent psychological distress are similarly associated with the oral microbiome.38 Furthermore, high levels of recent life stress, anxiety, depression symptoms or PTSD symptoms were each associated with mostly different changes in the levels of particular microbes, with the greatest number of differences between the high and low depression symptom and PTSD symptom groups (online supplemental table 8). This was found despite the significant correlations among almost all our stress and mental health variables. Using a latent variable of overall distress created from individual general life stress, trait anxiety and depression scales, Kohn et al 19 found that overall distress is related to oral microbe diversity and taxonomy. Our findings suggest that the manifest distress scales might have driven different components of their overall microbiome results. In fact, our own follow-up analysis revealed no microbe differences between high-symptom and low-symptom groups when all our mental health measures were combined into a single variable.

      Our increased Bacteroidetes and Firmicutes abundance in women with high trait anxiety aligns with some gut microbiome data from non-pregnant people with high generalised anxiety—for example, Wang et al. 42 The fact that these differences were not found between our groups of women with high and low state anxiety again indicates greater relevance for trait-like, rather than immediate, anxiety symptoms on particular oral microbe levels, which was also suggested in a study of dental procedure anxiety.38 Our elevated Spirochaetes levels in the high depression symptom group has been observed in an oral microbiome study of adolescents, where it was additionally positively associated with high anxiety.16 We found genus Dialister from phylum Firmicutes to be one of relatively few microbes affected by more than one stress or mental health variable, being higher in pregnant women experiencing high state or trait anxiety or high depression symptoms. Oral Dialister is higher in young adults with suicidal ideation,18 which is often a symptom of depression and comorbid with anxiety. However, gut Dialister levels are lower in non-pregnant people with high anxiety or clinical depression.43 44 This contradictory finding suggests that, while the oral and lower gut microbiomes are known to be highly related, they are not identical45 and unique relations with stress and mental health are not unexpected.

      The genus Eikenella was also elevated in more than one of our high-symptom groups (trait anxiety, depression symptoms, PTSD symptoms). Oral-derived Eikenella corrodens increases after in vitro treatment with the adrenal medulla stress hormones, epinephrine or norepinephrine,46 but in a mixed-sex group of patients being treated for moderate to severe chronic periodontitis salivary levels, neither E. corrodens nor E. nodatum differed between groups with high or low trait anxiety, or between groups with low or high perceived life stress.38 Eikenella was also unrelated to general anxiety or depression scores in a large national study involving both men and women.20 It is difficult to ascribe any meaning to these apparent differences in Eikenella results between studies involving mixed-sex and/or clinical dentistry samples and our study of a community sample of pregnant women, especially considering that we controlled for dental problems in our analyses of bacteria abundance.

      Strengths and limitations

      The strengths of this study include the large sample size relative to most previous studies of distress and the oral microbiome, our sample’s racial diversity that allows us to better generalise results to the larger population of pregnant women, recruiting a higher-risk sample to ensure we captured a wide range of life stress and mental health symptoms, using numerous indicators of recent and more distal stress and mental health, and evaluating many potential covariates.

      Limitations also exist. We had insufficient data to examine covariates related to participants’ diet or body weight; we did not inquire about longer-term oral health history; the maternal anxiety and depression measures were self-reported, although our groups were created using cut-off scores that have high validity with diagnostic interviews24 25; we targeted only one timepoint in pregnancy; and we did not analyse other sources of microbes, such as the lower gut, which might have a bidirectional relationship with the oral microbiome.47 We also do not know the direction of our results (oral microbiome to mental health, vice versa or both), although experimental studies in mice report that changing the oral microbiome by infusing donor saliva alters the recipient’s anxiety- and depression-related behaviours.48 49 Finally, we did not test possible mechanisms underlying our results which could include stress-induced pro-inflammatory oral cytokines impinging on the brain, oral microbes modifying central neurotransmitter synthesis, or stress-altered oral microbes infiltrating the gut to then affect the brain.20

      Conclusions

      Our study shows that numerous aspects of the oral microbiome in pregnancy are associated with women’s life stress and mental health. Importantly, these associations differ from studies of the gut microbiome and studies of non-pregnant people. Successful targeting of the gut microbiome with probiotic treatment to improve maternal mental health9 could be extended in future studies to target oral cavity microbes through dietary changes, making recommendations for improving oral health, and probiotic treatments that might benefit mothers struggling with high life stress and poor mental health.

      Supplemental material

      Data availability statement

      Data are available upon reasonable request. Data are available upon reasonable request to JSL or AMA.

      Ethics statements

      Patient consent for publication

      Ethics approval

      This study involves human participants and the study procedures were approved by the Institutional Review Boards of Michigan State University (#16-1090M) and University of Michigan (#HUM00121071). Participants gave informed consent to participate in the study before taking part.

      Acknowledgments

      This research was funded by NIH grant R01HD085990 to AAL, GAB and JSL, as well as a grant from the Michigan State University College of Social Science Faculty Initiative Fund to JSL, AKN, RCK, AAL and GAB. The authors thank Katherine Gurdziel at the Wayne State University Genomic Sciences Core for conducting the 16S rRNA gene sequencing and Alexandra Ballinger, Jewelian Fairchild and other members of the Michigan Prenatal Stress Study team for their assistance with various other aspects of this study.

      References

      1. ↵
        2. Howard LM ,
        3. Khalifeh H
        . Perinatal mental health: a review of progress and challenges. World Psychiatry 2020;19:313–27. doi:10.1002/wps.20769
        OpenUrlCrossRefPubMed
      2. ↵
        2. Slomian J ,
        3. Honvo G ,
        4. Emonts P , et al
        . Consequences of maternal postpartum depression: A systematic review of maternal and infant outcomes. Womens Health (Lond) 2019;15:1745506519844044. doi:10.1177/1745506519844044
      3. ↵
        2. Brummelte S ,
        3. Galea LAM
        . Postpartum depression: Etiology, treatment and consequences for maternal care. Horm Behav 2016;77:153–66. doi:10.1016/j.yhbeh.2015.08.008
        OpenUrlCrossRef
      4. ↵
        2. Rackers HS ,
        3. Thomas S ,
        4. Williamson K , et al
        . Emerging literature in the Microbiota-Brain Axis and Perinatal Mood and Anxiety Disorders. Psychoneuroendocrinology 2018;95:86–96. doi:10.1016/j.psyneuen.2018.05.020
        OpenUrlCrossRefPubMed
      5. ↵
        2. Jovel J ,
        3. Dieleman LA ,
        4. Kao D , et al
        . The human gut microbiome in health and disease. Metagenomics (Cairo) 2018;197–213. doi:10.1016/B978-0-08-102268-9.00010-0
      6. ↵
        2. Hechler C ,
        3. Borewicz K ,
        4. Beijers R , et al
        . Association between Psychosocial Stress and Fecal Microbiota in Pregnant Women. Sci Rep 2019;9:4463. doi:10.1038/s41598-019-40434-8
      7. ↵
        2. Peñalver Bernabé B ,
        3. Maki PM ,
        4. Cunningham JL , et al
        . Interactions between perceived stress and microbial-host immune components: two demographically and geographically distinct pregnancy cohorts. Transl Psychiatry 2023;13:3. doi:10.1038/s41398-022-02276-3
      8. ↵
        2. Zhou Y ,
        3. Chen C ,
        4. Yu H , et al
        . Fecal Microbiota Changes in Patients With Postpartum Depressive Disorder. Front Cell Infect Microbiol 2020;10:567268. doi:10.3389/fcimb.2020.567268
      9. ↵
        2. Slykerman RF ,
        3. Hood F ,
        4. Wickens K , et al
        . Effect of Lactobacillus rhamnosus HN001 in Pregnancy on Postpartum Symptoms of Depression and Anxiety: A Randomised Double-blind Placebo-controlled Trial. EBioMedicine 2017;24:159–65. doi:10.1016/j.ebiom.2017.09.013
        OpenUrlCrossRefPubMed
      10. ↵
        2. Maitre Y ,
        3. Micheneau P ,
        4. Delpierre A , et al
        . Did the Brain and Oral Microbiota Talk to Each Other? A Review of the Literature. J Clin Med 2020;9:3876. doi:10.3390/jcm9123876
      11. ↵
        2. Qiao Y ,
        3. Wu M ,
        4. Feng Y , et al
        . Alterations of oral microbiota distinguish children with autism spectrum disorders from healthy controls. Sci Rep 2018;8:1597. doi:10.1038/s41598-018-19982-y
      12. ↵
        2. Hicks SD ,
        3. Uhlig R ,
        4. Afshari P , et al
        . Oral microbiome activity in children with autism spectrum disorder. Autism Res 2018;11:1286–99.
        OpenUrlCrossRefPubMed
      13. ↵
        2. Yolken R ,
        3. Prandovszky E ,
        4. Severance EG , et al
        . The oropharyngeal microbiome is altered in individuals with schizophrenia and mania. Schizophr Res 2021;234:51–7. doi:10.1016/j.schres.2020.03.010
        OpenUrlCrossRefPubMed
      14. ↵
        2. Castro-Nallar E ,
        3. Bendall ML ,
        4. Pérez-Losada M , et al
        . Composition, taxonomy and functional diversity of the oropharynx microbiome in individuals with schizophrenia and controls. PeerJ 2015;3:e1140. doi:10.7717/peerj.1140
      15. ↵
        2. Paudel D ,
        3. Uehara O ,
        4. Giri S , et al
        . Effect of psychological stress on the oral-gut microbiota and the potential oral-gut-brain axis. Jpn Dent Sci Rev 2022;58:365–75. doi:10.1016/j.jdsr.2022.11.003
        OpenUrlCrossRefPubMed
      16. ↵
        2. Simpson CA ,
        3. Adler C ,
        4. du Plessis MR , et al
        . Oral microbiome composition, but not diversity, is associated with adolescent anxiety and depression symptoms. Physiol Behav 2020;226:113126. doi:10.1016/j.physbeh.2020.113126
        OpenUrlCrossRefPubMed
      17. ↵
        2. Wingfield B ,
        3. Lapsley C ,
        4. McDowell A , et al
        . Variations in the oral microbiome are associated with depression in young adults. Sci Rep 2021;11:15009. doi:10.1038/s41598-021-94498-6
      18. ↵
        2. Ahrens AP ,
        3. Sanchez-Padilla DE ,
        4. Drew JC , et al
        . Saliva microbiome, dietary, and genetic markers are associated with suicidal ideation in university students. Sci Rep 2022;12:14306. doi:10.1038/s41598-022-18020-2
      19. ↵
        2. Kohn JN ,
        3. Kosciolek T ,
        4. Marotz C , et al
        . Differing salivary microbiome diversity, community and diurnal rhythmicity in association with affective state and peripheral inflammation in adults. Brain Behav Immun 2020;87:591–602. doi:10.1016/j.bbi.2020.02.004
        OpenUrlCrossRefPubMed
      20. ↵
        2. Li C ,
        3. Chen Y ,
        4. Wen Y , et al
        . A genetic association study reveals the relationship between the oral microbiome and anxiety and depression symptoms. Front Psychiatry 2022;13:960756. doi:10.3389/fpsyt.2022.960756
      21. ↵
        2. Baron E ,
        3. Bass J ,
        4. Murray SM , et al
        . A systematic review of growth curve mixture modelling literature investigating trajectories of perinatal depressive symptoms and associated risk factors. J Affect Disord 2017;223:194–208. doi:10.1016/j.jad.2017.07.046
        OpenUrlCrossRefPubMed
      22. ↵
        2. Fisher SD ,
        3. Wisner KL ,
        4. Clark CT , et al
        . Factors associated with onset timing, symptoms, and severity of depression identified in the postpartum period. J Affect Disord 2016;203:111–20. doi:10.1016/j.jad.2016.05.063
        OpenUrlCrossRefPubMed
      23. ↵
        2. Levendosky AA ,
        3. Bogat GA ,
        4. Lonstein J , et al
        . Longitudinal prospective study examining the effects of the timing of prenatal stress on infant and child regulatory functioning: the Michigan Prenatal Stress Study protocol. BMJ Open 2021;11:e054964. doi:10.1136/bmjopen-2021-054964
      24. ↵
        2. Grant KA ,
        3. McMahon C ,
        4. Austin MP
        . Maternal anxiety during the transition to parenthood: a prospective study. J Affect Disord 2008;108:101–11. doi:10.1016/j.jad.2007.10.002
        OpenUrlCrossRefPubMedWeb of Science
      25. ↵
        2. Levis B ,
        3. Negeri Z ,
        4. Sun Y , et al
        . Accuracy of the Edinburgh Postnatal Depression Scale (EPDS) for screening to detect major depression among pregnant and postpartum women: systematic review and meta-analysis of individual participant data. BMJ 2020;371:m4022. doi:10.1136/bmj.m4022
      26. ↵
        1. National Center for PTSD
        . Using the PTSD checklist for DSM-5 (PCL-5). 2024. Available: http://www.ptsd.va.gov/professional/assessment/adult-sr/ptsd-checklist.asp
      27. ↵
        2. Segata N ,
        3. Abubucker S ,
        4. Goll J , et al
        . Microbial community function and biomarker discovery in the human microbiome. Genome Biol 2011;12:1–25. doi:10.1186/1465-6906-12-S1-P47
        OpenUrlCrossRefPubMed
      28. ↵
        2. Cao Y ,
        3. Dong Q ,
        4. Wang D , et al
        . microbiomeMarker: an R/Bioconductor package for microbiome marker identification and visualization. Bioinformatics 2022;38:4027–9. doi:10.1093/bioinformatics/btac438
        OpenUrlCrossRefPubMed
      29. ↵
        2. Mallick H ,
        3. Rahnavard A ,
        4. McIver LJ , et al
        . Multivariable association discovery in population-scale meta-omics studies. PLoS Comput Biol 2021;17:e1009442. doi:10.1371/journal.pcbi.1009442
      30. ↵
        2. Maechler M ,
        3. Rousseeuw P ,
        4. Struyf A , et al
        . cluster: Cluster Analysis Basics and Extensions [R package version 2.1.6]. 2023.
      31. ↵
        2. Oksanen J ,
        3. Simpson G ,
        4. Blanchet F , et al
        . 2024;vegan.
      32. ↵
        2. Willis JR ,
        3. Saus E ,
        4. Iraola-Guzmán S , et al
        . Citizen-science reveals changes in the oral microbiome in Spain through age and lifestyle factors. NPJ Biofilms Microbiomes 2022;8:38. doi:10.1038/s41522-022-00279-y
      33. ↵
        2. Saadaoui M ,
        3. Singh P ,
        4. Al Khodor S
        . Oral microbiome and pregnancy: A bidirectional relationship. J Reprod Immunol 2021;145:103293. doi:10.1016/j.jri.2021.103293
        OpenUrlCrossRefPubMed
      34. ↵
        2. Ke S ,
        3. Wang X-W ,
        4. Ratanatharathorn A , et al
        . Association of probable post-traumatic stress disorder with dietary pattern and gut microbiome in a cohort of women. Nat Ment Health 2023;1:900–13. doi:10.1038/s44220-023-00145-6
        OpenUrlCrossRef
      35. ↵
        2. Hemmings SMJ ,
        3. Malan-Müller S ,
        4. van den Heuvel LL , et al
        . The Microbiome in Posttraumatic Stress Disorder and Trauma-Exposed Controls: An Exploratory Study. Psychosom Med 2017;79:936–46. doi:10.1097/PSY.0000000000000512
        OpenUrlCrossRefPubMed
      36. ↵
        2. Zeamer AL ,
        3. Salive M-C ,
        4. An X , et al
        . Association between microbiome and the development of adverse posttraumatic neuropsychiatric sequelae after traumatic stress exposure. Transl Psychiatry 2023;13:354. doi:10.1038/s41398-023-02643-8
      37. ↵
        2. Stoy S ,
        3. McMillan A ,
        4. Ericsson AC , et al
        . The effect of physical and psychological stress on the oral microbiome. Front Psychol 2023;14:1166168. doi:10.3389/fpsyg.2023.1166168
      38. ↵
        2. Dubar M ,
        3. Clerc-Urmès I ,
        4. Baumann C
        . Relations of Psychosocial Factors and Cortisol with Periodontal and Bacterial Parameters: A Prospective Clinical Study in 30 Patients with Periodontitis Before and After Non-Surgical Treatment. Int J Environ Res Public Health 2020;17:7651. doi:10.3390/ijerph17207651
      39. ↵
        2. Duran-Pinedo AE ,
        3. Solbiati J ,
        4. Frias-Lopez J
        . The effect of the stress hormone cortisol on the metatranscriptome of the oral microbiome. NPJ Biofilms Microbiomes 2018;4:25. doi:10.1038/s41522-018-0068-z
      40. ↵
        2. Keane JM ,
        3. Khashan AS ,
        4. McCarthy FP , et al
        . Identifying a biological signature of prenatal maternal stress. JCI Insight 2021;6:e143007. doi:10.1172/jci.insight.143007
      41. ↵
        2. Charalambous EG ,
        3. Mériaux SB ,
        4. Guebels P , et al
        . Early-Life Adversity Leaves Its Imprint on the Oral Microbiome for More Than 20 Years and Is Associated with Long-Term Immune Changes. Int J Mol Sci 2021;22:12682. doi:10.3390/ijms222312682
      42. ↵
        2. Wang Z ,
        3. Liu S ,
        4. Xu X , et al
        . Gut Microbiota Associated With Effectiveness And Responsiveness to Mindfulness-Based Cognitive Therapy in Improving Trait Anxiety. Front Cell Infect Microbiol 2022;12:719829. doi:10.3389/fcimb.2022.719829
      43. ↵
        2. Chen H-M ,
        3. Chung Y-CE ,
        4. Chen H-C , et al
        . Exploration of the relationship between gut microbiota and fecal microRNAs in patients with major depressive disorder. Sci Rep 2022;12:20977. doi:10.1038/s41598-022-24773-7
      44. ↵
        2. Caso JR ,
        3. MacDowell KS ,
        4. González-Pinto A , et al
        . Gut microbiota, innate immune pathways, and inflammatory control mechanisms in patients with major depressive disorder. Transl Psychiatry 2021;11:645. doi:10.1038/s41398-021-01755-3
      45. ↵
        2. Maki KA ,
        3. Kazmi N ,
        4. Barb JJ , et al
        . The Oral and Gut Bacterial Microbiomes: Similarities, Differences, and Connections. Biol Res Nurs 2021;23:7–20. doi:10.1177/1099800420941606
        OpenUrlCrossRefPubMed
      46. ↵
        2. Roberts A ,
        3. Matthews JB ,
        4. Socransky SS , et al
        . Stress and the periodontal diseases: effects of catecholamines on the growth of periodontal bacteria in vitro. Oral Microbiol Immunol 2002;17:296–303. doi:10.1034/j.1399-302x.2002.170506.x
        OpenUrlCrossRefPubMedWeb of Science
      47. ↵
        2. Tan X ,
        3. Wang Y ,
        4. Gong T
        . The interplay between oral microbiota, gut microbiota and systematic diseases. J Oral Microbiol 2023;15:2213112. doi:10.1080/20002297.2023.2213112
      48. ↵
        2. Qiao Y ,
        3. Gong W ,
        4. Li B , et al
        . Oral Microbiota Changes Contribute to Autism Spectrum Disorder in Mice. J Dent Res 2022;101:821–31. doi:10.1177/00220345211070470
        OpenUrlCrossRefPubMed
      49. ↵
        2. Lou F ,
        3. Luo S ,
        4. Kang N
        . Oral microbiota dysbiosis alters chronic restraint stress-induced depression-like behaviors by modulating host metabolism. Pharmacol Res 2024;204:107214. doi:10.1016/j.phrs.2024.107214
        OpenUrlCrossRefPubMed

      Supplementary materials

      Footnotes

      • Contributors AMA, AAL, GAB, MMM, AKN, RCK, and JSL were involved in the conception and design of the current study. JSL, AAL, GAB and MMM were involved in data acquisition. AMA and AKN performed data analyses and all authors were involved in interpreting the results. JSL and AMA wrote the first draft of the manuscript. All authors (AMA, AAL, GAB, MMM, AKN, RCK, JSL) then contributed to subsequent drafts of the manuscript and approved the final version. JSL and AMA are the guarantors of this paper.

      • Funding This research was funded by NIH grant R01HD085990 to AAL, GAB and JSL, as well as a grant from the Michigan State University College of Social Science Faculty Initiative Fund to JSL, AKN, RCK, AAL and GAB.

      • Competing interests None declared.

      • Provenance and peer review Not commissioned; externally peer reviewed.

      • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.