Article Text

Download PDFPDF

Cohort profile
Cohort profile: the Swedish Maternal Microbiome project (SweMaMi) – assessing the dynamic associations between the microbiome and maternal and neonatal adverse events
  1. Emma Fransson1,2,
  2. Unnur Gudnadottir1,
  3. Luisa W Hugerth1,3,
  4. Eva Wiberg Itzel4,5,
  5. Marica Hamsten1,
  6. Fredrik Boulund1,3,
  7. Alexandra Pennhag1,
  8. Juan Du6,
  9. Ina Schuppe-Koistinen1,3,
  10. Nele Brusselaers1,7,
  11. Lars Engstrand1,3
  1. 1Centre for Translational Microbiome Research, Department of Microbiology, Tumor and Cell Biology, Karolinska Institute, Stockholm, Sweden
  2. 2Department of Women’s and Children's Health, Uppsala University, Uppsala, Sweden
  3. 3Science for Life Laboratory (SciLifeLab), Stockholm, Sweden
  4. 4Department of Obstetrics and Gynecology, Södersjukhuset AB, Stockholm, Sweden
  5. 5Department of Clinical Science and Education, Karolinska Institutet, Stockholm, Sweden
  6. 6Centre for Translational Microbiome Research, Karolinska Institutet, Stockholm, Sweden
  7. 7Global Health Institute, University of Antwerp, Antwerpen, Belgium
  1. Correspondence to Dr Ina Schuppe-Koistinen; ina.schuppe.koistinen{at}ki.se

Abstract

Purpose The Swedish Maternal Microbiome (SweMaMi) project was initiated to better understand the dynamics of the microbiome in pregnancy, with longitudinal microbiome sampling, shotgun metagenomics, extensive questionnaires and health registry linkage.

Participants Pregnant women were recruited before the 20th gestational week during 2017–2021 in Sweden. In total, 5439 pregnancies (5193 unique women) were included. For 3973 pregnancies (73%), samples were provided at baseline, and for 3141 (58%) at all three timepoints (second and third trimester and postpartum). In total, 31 740 maternal microbiome samples (vaginal, faecal and saliva) and 3109 infant faecal samples were collected. Questionnaires were used to collect information on general, reproductive and mental health, diet and lifestyle, complemented by linkage to the nationwide health registries, also used to follow up the health of the offspring (up to age 10).

Findings to date The cohort is fairly representative for the total Swedish pregnant population (data from 2019), with 41% first-time mothers. Women with university level education, born in Sweden, with normal body mass index, not using tobacco-products and aged 30–34 years were slightly over-represented.

Future plans The sample and data collection were finalised in November 2021. The next steps are the characterisation of the microbial DNA and linkage to the health and demographic information from the questionnaires and registries. The role of the microbiome on maternal and neonatal outcomes and early-childhood diseases will be explored (including preterm birth, miscarriage) and the role and interaction of other risk factors and confounders (including endometriosis, polycystic ovarian syndrome, diet, drug use). This is currently among the largest pregnancy cohorts in the world with longitudinal design and detailed and standardised microbiome sampling enabling follow-up of both mothers and children. The findings are expected to contribute greatly to the field of reproductive health focusing on pregnancy and neonatal outcomes.

  • fetal medicine
  • maternal medicine
  • bacteriology
  • microbiology
  • public health

Data availability statement

Data are available on reasonable request. Register data cannot be shared, due to confidentiality.

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/.

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.

Strengths and limitations of this study

  • The major strength of the Swedish Maternal Microbiome (SweMaMi) project is the large-scale and well-characterised longitudinal cohort of pregnant women with child follow-up planned until age 10.

  • The project includes comprehensive questionnaire data, microbiome sampling from three body sites and registry data, enabling analysis on pregnancy outcomes such as preterm birth with good statistical power.

  • Sample collection was performed in a longitudinal and standardised manner including both pregnant women and their offspring from all over Sweden.

  • One limitation is that the SweMaMi project is based on home sampling, therefore blood sampling and samples for metabolomics are lacking.

  • As in many research studies, participants with high socioeconomic status are overrepresented.

Introduction

Microbial communities (including bacteria, viruses and fungi) inhabit our bodies and actively contribute to the homeostasis and health of every individual.1 The microbiome also seems crucial for a healthy pregnancy, both for the mothers and their offspring.2–5

The microbiome during pregnancy

There are specific microbiota in different anatomical niches, which could all affect our health. Regarding pregnancy and reproductive health, the vaginal microbiome is currently most studied.6 In general, a low-diversity, lactobacillus-dominated vaginal microbiome is considered most healthy, while in the gut, a high-diversity gut microbiome (containing a large variety of species) is associated with health benefits.7 The microbiome during pregnancy is dynamic and behaves differently in each body site. In the vagina, the variation of microbes within a single sample (alpha diversity) and the variation of microbial communities between samples (beta diversity) are both reduced during pregnancy.8 In the gut, alpha-diversity also appears to decrease, while the beta diversity, or distance between samples, increases.9 Proinflammatory cytokine levels in faecal samples also seem to increase during pregnancy.9 Yet, large studies have found that individual differences between women overrun any effect of gestational age10 although distinct patterns in the gut microbiome can still be observed in association to pathological conditions.11 12 Gingival inflammation is known to increase and saliva pH to decrease during pregnancy.13 The gingival and saliva microbiome have been shown to differ significantly between pregnant and non-pregnant women, with an increase in evenness and total diversity among pregnant women.14–16 Several demographic, lifestyle and health-related factors might impact on the maternal microbiome, including gynaecological conditions such as polycystic ovary syndrome (PCOS) and endometriosis, which are common causes of subfertility and pregnancy complications.17–21 Obesity, excessive weight gain and gestational diabetes all leave an apparently distinct signature on the microbiome.10 22–24 There seems to be a crosstalk between the host metabolism, microbiome and adipose tissue, which seems particularly important during pregnancy.24–26

The vaginal microbiome

Lactobacillus species (spp) are the most common colonisers of the vaginal tract in women of reproductive age.8 27 28 Some women lack this Lactobacillus dominance in the vagina and have a diverse composition of other bacteria, including anaerobic bacteria, referred to as vaginal dysbiosis, which is asymptomatic in up to 50%.29 30 During a healthy pregnancy, the Lactobacillus dominance in the vaginal microbiome becomes even more pronounced, likely due to the large increase in oestrogens.8 For postpartum samples, when sex hormones drop again and lochia flows, the vaginal microbiome has been described to be more diverse compared with samples taken during the first and third trimester, with lower proportions of Lactobacilli.31 32

Vaginal dysbiosis, and particularly the presence of anaerobic bacteria, has been associated with higher risks of sexually transmitted infections,33 poor fertility treatment outcome,34 human papillomavirus infection and gynaecological cancers.27 35 Vaginal dysbiosis has also been associated with miscarriage, preterm birth, preeclampsia, gestational diabetes and excessive gestational weight gain, which are all common gestational complications.36–40 Some specific bacterial species linked to vaginal dysbiosis such as BVAB1, Sneathia spp, Gardnerella vaginalis and Prevotella spp, have been associated with increased risks of spontaneous preterm birth, yet may be linked to ethno-geographical differences.37 38 41Two recent studies (one with Chinese women and one mixed cohort with mostly Hispanic women) failed to identify associations between specific vaginal bacterial species and preterm birth highlighting the pathogenic complexity.42 43 Yet, as our recent network meta-analysis on preterm birth suggests, there may also be risk differences associated to the dominant Lactobacilli spp, with L. crispatus dominant vaginal microbiome presenting with the lowest risk.44

Prior studies investigating the vaginal microbiome and pregnancy outcomes are often inconclusive due to small sample sizes, cross-sectional designs and without sufficient adjustment for confounding factors such as comorbidities, prescribed drug use and lifestyle factors. Therefore, many research questions remain to be answered.

The gut and saliva microbiome

The gut and oral microbiota have also been shown to have systemic effects during pregnancy,4 45 but the potential interaction with maternal and child health remains unclear. Specifically, the number of studies investigating the maternal gut microbiota in association with pregnancy outcomes are sparse.4 10 46 47 There are claims of correlation between oral disease and the risk of preterm birth, including periodontal disease and the abundance of Porphyromonas gingivalis in subgingival plaque.16 36 Interestingly, an intervention study providing pregnant women with xylitol chewing gum showed a reduction of periodontitis and lower incidence of preterm birth.48 However, results from a recent meta-analysis on pregnancy and the oral microbiome were inconclusive and called for further studies of the topic.16

The maternal and infant microbiome

The maternal and early infant microbiome could also be associated with neonatal outcomes, such as fetal growth, metabolic acidosis, early metabolic or endocrine disturbances and outcome of neonatal intensive care.49–51 There are also studies investigating infant microbiome and later disease, including allergies, asthma, type I diabetes, epilepsy and autism.52–54 Still, large and longitudinal cohorts with high quality data are needed.

Summary of current literature and rational for the current cohort

In summary, research to date points at an association between the vaginal microbiome and poor pregnancy outcomes, but the results show large differences between countries and different ethno-geographical groups. The role of the gut and oral microbiome remains unclear. With increased knowledge, the microbiome could be considered as a possible treatment target for prevention of adverse pregnancy and child outcomes. The Swedish Maternal Microbiome (SweMaMi) project was initiated in 2017 to investigate the pregnancy microbiome and its association with pregnancy outcome and child health. In this cohort description, we give an overview of the SweMaMi study design, participants registered, and how the participants compare to the Swedish population.

Primary aims

The SweMaMi project was designed to deepen knowledge regarding the microbiome in reproductive health and pregnancy outcomes. The overall aim is to assess the association between the maternal vaginal, faecal and oral microbiome during pregnancy and postpartum, and maternal and neonatal adverse events. The main research question is whether certain microbiome compositions during pregnancy are associated with an increased risk for adverse pregnancy outcomes, including preterm birth and pregnancy loss.

Other research questions regard characterisation of the maternal microbiome in pregnancies of women with risk factors for suboptimal pregnancy outcomes, such as PCOS, pre-eclampsia and endometriosis.17 18 In addition, we will investigate the infant microbiome, infant outcome and early-childhood diseases. The study design also allows for further clarification of the ‘healthy’ or ‘normal’ microbiome during pregnancy and postpartum and of the neonate.

Cohort description

Study design and recruitment

The SweMaMi project enrolled participants between November 2017 and February 2021. The last babies of participating women were born at the end of 2021. All pregnant women (before gestational week 20) residing in Sweden with a personal identification number and understanding Swedish or English were eligible to participate. Pregnant women (primiparous and multiparous) were recruited through online advertisement on social media and in pregnancy-related mobile applications.

Information about the study was also shared via posters at Karolinska Institutet campus and in antenatal clinics in Sweden, especially in the greater Stockholm area. Since no healthcare visits were required to participate in the study, women all over Sweden could participate. Women could participate more than once if pregnant several times during the study enrolment period.

Informed consent was obtained from participants through the study webpage before answering the first online questionnaire and could be withdrawn at any time by contacting our team by email or telephone. In association with the home sampling of the infant, both parents (if applicable) signed a consent form. The study complied with the Declaration of Helsinki and the General Data Protection Regulation (GDPR) regulations.

Data collection

Data were collected at three different timepoints, starting with an online questionnaire, followed by home-sampling for microbiome assessment: (1) between pregnancy weeks 10 and 20; (2) pregnancy week 28 and 30 and (3) 5 and 8 weeks after expected delivery date: (including faecal sample from the infant) (see figure 1). Self-collected samples have previously been shown to match the quality of physician-collected samples.55 56

Figure 1

Overview of the SweMaMi study data and sample collection. SweMaMi, Swedish Maternal Microbiome.

Samples for microbiome analyses

Standardised microbiome sampling, processing and storage was performed to ensure comparability and quality of the samples.

Vaginal samples were collected by inserting a FLOQ swab (Copan, Italy) 2–3 cm into the vagina and rotating for 20–30 s. The tip of the swab was then broken off into a 1.9 mL FluidX tube (Brooks Life Sciences, Massachusetts, USA) which had 0.8 mL DNA/RNA shield (Zymo Research, California, USA).

The faecal samples were collected in DNA/RNA Shield-Faecal collection tube (Zymo Research, California, USA). Participants collected faecal samples from toilet paper with a spoon attached to the lid of the tube and put in the DNA/RNA shield.

Saliva samples were collected using a SalivaGene Collector (Invitek Molecular, Germany) tube according to the manufacturer’s instructions.

All three sample collection kits ensure that the samples would be stable during transportation at room temperature from the participants and until registration and storage at −80°C until DNA extraction for at least 30 days (sent back by post).

Questionnaire data

The questionnaires were designed to gather detailed information on demographic characteristics (eg, age, educational level, body mass index (BMI) and country of birth), health characteristics (eg, detailed questions about menstruation before pregnancy, gynaecological and other comorbidities and characteristics, prescribed and non-prescribed drug use, pregnancy and delivery variables) and lifestyle (eg, diet and tobacco use).

The questionnaire also included standardised questions on mental health, alcohol use and vomiting: The Edinburgh Postnatal Depression Scale,57 the Cohen Perceived Stress Scale,58 the Alcohol Use Disorders Identification Test59 and the 24-hour Pregnancy-Unique Quantification of Emesis Scale.60 In addition, the Bristol Stool Chart, a frequently used measure in gastroenterology practice and research, was used to categorise stools into one of seven types61 and Ferriman-Gallwey scoring was used to assess possible undiagnosed PCOS.62

All study information, questionnaires and sampling instructions were available both in Swedish and English.

Registry data

In addition, a registry linkage will be conducted to obtain more detailed information on the mothers and to be able to follow-up the health of the children. In Sweden, prenatal, delivery and early child healthcare at the Swedish child welfare centres are available to all residents, tax funded (free of charge), and participation is close to 100%.

Data from the following nationwide registries will be collected for the mother: The Medical birth registry (Medicinska födelseregistret), the In-Patient and Out-Patient registry (Patientregistret), the Prescribed Drug registry (Läkemedelsregistret), the Cause of death registry (Dödsorsaksregistret), the Swedish Quality Registry for caries and periodontal disease (SKaPa) and the Pregnancy Quality Register (Graviditetsregistret).63–69 All these registries except the last, are nationwide and recording of information is compulsory for every Swedish resident. Although the pregnancy quality registry is not compulsory, coverage is high (about 90%).69

The Patient, Cancer, Death and Drug registries for the included children will be used to assess the health of the children up to the age of 10.64 65 67 70

Data analyses

Extraction and sequencing of microbial DNA

DNA extraction will be performed with standardised and automated pipelines as previously optimised by our group for vaginal,71 saliva72 and faecal samples.73 The extraction method allows for the detection of all domains of life, namely bacteria, archaea, small eukaryotes including fungi and viruses. Longitudinal profiling of vaginal, faecal and oral microbiome samples will be performed at the Centre for Translational Microbiome Research (CTMR), Karolinska Institutet, Sweden, with a focus on microbial diversity and the correlation between the relative abundances of taxonomic groups with external factors. Taxonomical profiling will be based on metagenomic sequencing to identify important organisms on species/strain level as well as to characterise their functional information such as enzymes involved in metabolic pathways and antibiotic resistance. Additionally, unlike 16S rRNA gene sequencing, metagenomics can provide relative abundances viruses and fungi as well as bacterial species. Sequencing will be performed on MGI sequencers, MGI Tech Co., Ltd (G400 or T7 models) on paired 150 base pair reads, aiming for at least 20 million reads for faecal and saliva samples, or 60 million for vaginal samples due to the high amount of human DNA in vaginal swabs.71

Analyses of microbial sequencing data

Raw sequencing reads will be trimmed to remove low-quality basepairs, and human reads will be removed by mapping to the reference human genome. All sequencing data, depleted of human DNA, will be deposited in the European Nucleotide Archive. Generated sequencing data will be handled in accordance with Karolinska Institutet’s commitments to secure handling of research data. Raw sequencing data will be primarily stored on the university’s on-premises S3 system and will be downloaded to CTMR’s private high-performance compute cluster Gandalf for bioinformatics processing and downstream bioinformatics and statistical analyses.

Reads will be taxonomically classified using Kraken V.2 or later on an appropriate database for each sample type, for example, the Human Oral Database74 for saliva samples, OptiVag for vaginal samples71 or Genome taxonomy database (GTDB)75 for faecal samples. Since the Human Oral Database does not currently include viral or fungal sequences, these will be annotated on the GTDB as well.

Statistical methods

The associations between the microbiome, other risk factors and poor pregnancy outcomes will be investigated through comparisons of within-sample and between-sample diversity. Within-sample diversity will be estimated using standard ecological measures such as Simpson’s and Chao1 indexes. Distances between samples will be estimated with non-metric, compositionality aware measures such as Bray-Curtis. We will also use evolutionary-aware measures, that is, weighted and unweighted Unifrac.

When we find community-level changes (ie, changes in diversity), we will identify differentially abundant organisms. All tests associating sequencing data to metadata will be conducted with tests that are compositionality-aware, such as beta-regressions. All tests will be corrected for multiple comparison with the Benjamini-Hochberg procedure. Additional confounders (such as maternal age, BMI, parity and prescribed drug use) will be considered. Advanced statistical methods such as generalised linear modelling and mediation analysis will then be applied with identified environmental risk factors.

The planned work includes several substudies where sub-group and sensitivity analyses will be performed to ensure adequate use of confounding factors. The maternal microbiome composition will also be investigated as the outcome, with medical conditions (eg, PCOS, endometriosis) as the exposure. For other conditions, such as gestational diabetes, the associations could be hypothesised to be bidirectional. The maternal microbiome will also be used as the exposure variable for infant microbiome composition, while both the maternal and the infant microbiome will be investigated as the exposure for early childhood outcomes.

Outcomes

Pregnancy health characteristics and outcomes

The extensive questionnaire and registry data will be used to characterise the pregnancy health of the participating women and to associate these factors with microbial data. Information about the main outcome measure, preterm birth, defined as pregnancy shorter than 37 completed weeks of gestation, will be collected both from the postpartum questionnaire and from registry data.

Infant and child outcomes

The SweMaMi project includes investigation of associations between the microbiome and neonatal outcomes, such as small or large for gestational age (babies who weigh significantly less or more than expected for the pregnancy length), intrauterine fetal death, low Apgar score, metabolic acidosis, early metabolic or endocrine disturbances and neonatal intensive care. Using the registry linkage, long-term follow-up of child outcomes will be applied, to investigate pregnant and infant microbial profiles, and potential associations with children’s development and health conditions, defined as diagnostics of allergies, asthma, diabetes type I, epilepsy and autism.

Confounders

Possible confounders that may affect the outcomes will be checked using the above-mentioned registries and questionnaire data. Some data will be available from multiple sources, increasing the validity of the information. Those confounders include pregnancy related characteristics (eg, conception (natural or induced), duration of pregnancy, singleton or multiple pregnancy, parity and C-section or vaginal delivery) and maternal characteristics (eg, age, prescribed drug use, chronic comorbidities, mental health, lifestyle and socioeconomic-factors). Environmental factors such as unemployment, low income or poor mental health as well as reproductive health problems could impact on pregnancy outcome.76 Low level of education and psychological stress have also been associated with a less optimal microbiome in pregnant women.77 78

Approximately 1/6 of the microbiome variation on population level can be attributed to the use of prescribed medications.79 Therefore, we will also assess the role of maternal and early life intake of potential microbiome-modulating drugs including antibiotics, proton-pump inhibitors, metformin, anti-inflammatory drugs and others.80

Power analysis

A power calculation for preterm birth was performed prior to the start of the SweMaMi project based on the incidence of spontaneous preterm birth being 3% in Sweden. The initial power analysis indicated that a sample of 1300 participants, a possible difference of 2% in the proportion of preterm birth between women with different bacterial colonisation types could be detected (power 0.8, alpha 0.05). However, due to the planned use of subgroup analyses, and potentially large microbiome variation as well as dropouts over time we initially aimed for 2500 women. Moreover, to enable detection of small variations in the proportion of spontaneous preterm deliveries in different diversity/abundance groups, the inclusion was increased to reach a sample of at least 3300 women with full data.

Representativeness of SweMaMi cohort

To assess the representativeness of our study cohort, we compared it to the nationwide pregnancy data as obtained from The National Board of Health and Welfare for the same study period.81 The most recent statistics for Sweden are from 2019 and the following characteristics will be compared: maternal age, parity, country of birth, highest level of education, BMI, tobacco use (smoking and snuff) and Swedish Region.

Basic functions in Excel and RStudio were used for all comparisons and overviews. Descriptive overview of the cohort was created with Excel, and a choropleth map of Sweden using R studio (V.1.3.1093), adjusted from.81 82 χ² test was conducted to compare the SweMaMi cohort with the overall population in Sweden using R studio.

Patient and public involvement

The research team for this study on pregnancy has included pregnant women and mothers. Study participants has been encouraged to contact the team regarding any questions or comments on the study procedures. During the study design time period, the team also had close contact with different stake holders, including clinical staff and patient organisations.

Findings to date

Cohort description

In total, 5193 unique women (5439 pregnancies) registered for the study by answering the first questionnaire. For 3973 pregnancies (73%), samples were provided at baseline and for 3141 (58%) at all three timepoints (second and third trimester and postpartum). In total, 38 591 maternal microbiome samples (vaginal, faecal and saliva) and 3109 infant faecal samples were collected (total 41 700 samples) (figure 2).

Figure 2

Flowchart showing the number of participants which completed each stage of the study with three questionnaires and three sample kits, with response rate as % of the women that answered questionnaire 1.

The SweMaMi cohort characteristics are presented in table 1. On average, women were 31.7 (±4.3) years old at recruitment and the majority were born in Sweden (90%). Participants lived all over Sweden, with a majority residing in Stockholm (42%) (figure 3). Most of the women had a university level education (78%).

Table 1

Overview of the Swedish Maternal Microbiome (SweMaMi) cohort

Figure 3

Choropleth map showing the number of SweMaMi participants per 10 000 inhabitants in each county of Sweden. Darker colour indicates higher number of participants per county. SweMaMi, Swedish Maternal Microbiome.

Most women (96%) reported having only one sexual partner in the previous year and were in a relationship (98%). For the majority of women (72%), this was not the first pregnancy. Most children were naturally conceived (90%), while 10% of women reported assisted reproductive technology (ART) use. Of those, in vitro fertilisation (IVF) was the most common (6%). Most women had no reported prior miscarriages (69%), but 4% had suffered recurrent pregnancy loss, defined as three or more consecutive miscarriages.83

Regarding gynaecological characteristics, 83% reported regular menstruation prior to conception, 4% had diagnosed endometriosis and 8% diagnosed polycystic ovary syndrome (PCOS).

Comparison to the Swedish population

Compared with the national pregnancy population in 2019 (see table 2), the SweMaMi participants were older (30.3% vs 41.9% younger than 30 years) and smoked less during early pregnancy (1.3% vs 3.9%). SweMaMi participants were more often normal weight (65.9% vs 51.8% with BMI between 18.5 and 24.9), university-level educated (77.5 vs 54.7%) and born in Sweden (90.5% vs 68.1%).

Table 2

Characteristics of the Swedish Maternal Microbiome (SweMaMi) cohort, compared with all women giving birth in Sweden Swedish population in 2019

Strengths and limitations

The major strength of the SweMaMi project is the large-scale and well characterised cohort of pregnant women, including comprehensive questionnaire data, microbiome sampling from three body sites and registry data, enabling analysis on pregnancy outcomes such as preterm birth with good statistical power. In addition, sample collection was performed in a longitudinal and standardised manner including both pregnant women and their offspring from all over Sweden, with the possibility of a 10-year follow-up for the children. The project uses cutting-edge sequencing technology for metagenomic analyses of microbial DNA performs analysis on viral and fungal components and generates detailed genetic functional analysis of the microbiome, including metabolic pathways. The linkage to National registries is unique within the microbiome field, where Swedish registries are generally highly complete and up to date. The questionnaire and the participant information were translated and provided in both Swedish and English to enable a broader recruitment.

The data collection also has limitations. The SweMaMi project is based on home sampling, and we therefore lack samples that were not feasible to collect at home or send by mail, such as blood samples or samples for bacterial culture. There would have been an advantage with more frequent sampling including a sample prior to pregnancy, but this was not feasible for this setup. We also lack samples for metabolomics to measure biochemical products produced on the microbiota in the mouth, gut and vagina, also because home sampling precludes samples that needs to stay frozen at all times. However, our protocol supports functional analyses of the microbiome, which will answer some questions regarding mechanistic pathways. In hindsight, the study would have benefited from even more detailed questionnaire data on diet and physical exercise. However, the number of questions had to be restricted to ensure that participants would agree to contribute and continue participation. In addition, as in many research studies, participants with high socioeconomic status are overrepresented.

Data availability statement

Data are available on reasonable request. Register data cannot be shared, due to confidentiality.

Ethics statements

Patient consent for publication

Ethics approval

The study protocol has been approved by the Regional Board of Ethics, Stockholm, Sweden (2017/1118-31).

Acknowledgments

During the data collection, a number of people have contributed to the work and need to be acknowledged: Kim Wong, Linn Flygare Wallén, Laura Poole, Caroline Bjurnemark, Maike Seifert and André Lauber. Figures 1 and 2 are made with BioRender.

References

Footnotes

  • EF and UG are joint first authors.

  • Twitter @lafransson, @NeleBrusselaers

  • Contributors The study was initiated by LE, IS-K and NB in close collaboration with the team at CTMR (EF, JD, LH, MH and AP) and with LE and IS-K as the guarantors. EWI took part in clinical considerations during the study. FB was responsible for building the database and systems for sample registration. UG has been responsible for the data and sample collection since 2019.

  • Funding Ferring Pharmaceuticals has funded this work with an unrestricted research grant. The study was also financed by grants from the Swedish state under the agreement between the Swedish government and the county councils, the ALF-agreement (Region Stockholm).

  • Map disclaimer The inclusion of any map (including the depiction of any boundaries therein), or of any geographic or locational reference, does not imply the expression of any opinion whatsoever on the part of BMJ concerning the legal status of any country, territory, jurisdiction or area or of its authorities. Any such expression remains solely that of the relevant source and is not endorsed by BMJ. Maps are provided without any warranty of any kind, either express or implied.

  • Competing interests The Centre for Translational Microbiome Research (CTMR) and associated authors are partly funded by Ferring Pharmaceuticals.

  • Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details.

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

  • Author note IS-K and LE are joint last authors.