Is the Vaginal Microbiome and Metabolome Associated With Spontaneous Preterm Birth (sPTB) in Multiple Pregnancies?

Overview

The aim of this study is to try and find links between the microscopic organisms (such as bacteria, yeasts and viruses) in the vagina, and twin pregnancies that deliver too early (preterm birth). Being born earlier than expected (preterm birth) happens in over half of twin pregnancies with 1 in 10 sets of twins delivering before 32 weeks gestation. Sometimes, when birth happens very early, babies can be at risk of serious harm including damage to the brain, lungs and bowel - all of which can result in life changing disabilities. How severe these problems are is related to how early they are born. Unfortunately, tests used to find women at risk of preterm birth have only been proven to work when the woman is carrying one baby, not twins, and at present no treatment has been shown to be effective in stopping a twin pregnancy from delivering early. Preventing twins from being born too early is therefore a target for research by the NHS and patient groups including the James Lind Alliance. It is normal for every woman to have microscopic organisms (such as bacteria, yeasts and viruses) in the vagina. New interest has been shown at looking closely at these organisms during pregnancy. These organisms can change and may be related to the number of weeks a woman will go into labour, however to date all research on this has been conducted in pregnancies with only one baby. We want to explore these organisms in twin pregnancies; taking swabs from the vagina at 16- and 28-weeks of your pregnancy, along with at the time of birth. Information will be gathered on the organisms present in the vagina (both of women that deliver too early and those that deliver on time), hoping this information will help us understand why preterm birth happens and help predict the chances of preterm labour in twin pregnancies. By identifying specific organisms linked with preterm birth, we also hope to be able to guide new targets for treatments to prevent preterm birth in twins in future. Due to the small number of twin pregnancies, measurements of how 'stiff' the neck of the womb (cervix) are along with blood samples will be taken. Research has shown that there may be links with how stiff the neck of the womb is and premature birth as well as markers within the blood that may help us predict preterm birth that are yet to be discovered. This will provide the foundations for a future research study.

RATIONALE FOR CURRENT STUDY The lack of evidence for benefit of any modality for the prevention of sPTB in multiple pregnancies means that novel approaches are urgently needed to address this major health inequality. Despite the growing evidence regarding both vaginal microbiome and vaginal mucosal metabolome and its association with sPTB, no study has analysed these in multiple pregnancies; neither determining any difference to singleton pregnancies or exploring any unique associations with sPTB. Discovering new associations will aide our ability to reliably predict women at highest risk of sPTB and identify those most likely to benefit from screening and intervention. The overall aim of this study is to assess the vaginal microbiome in multiple pregnancies to ascertain if it differs from singleton pregnancies, comparing results to our large and unique cohort of low and high-risk singletons (Harris-Wellbeing Preterm Birth Centre). In addition, the metabolic profile of the vaginal mucosa will be assessed in multiple pregnancies, working with internationally recognised collaborators to utilise novel DESI MS analysis. This study will be the first to assess both the vaginal microbiome and the vaginal mucosal metabolic profile in multiple pregnancies and to evaluate associations between the vaginal microbiome, metabolome and sPTB in a well phenotyped multiple pregnancy population. Furthermore, this study will utilise the unique cohort of multiple pregnancies by collecting blood samples and cervical stiffness measurements for future exploratory analysis, ultimately aiming to develop a successful risk stratification model for multiple pregnancies. RESEARCH QUESTION/AIM(S) To assess the vaginal microbiome in multiple pregnancies compared to low and high-risk singleton pregnancies and to evaluate associations between the vaginal microbiome, vaginal metabolome and sPTB. OBJECTIVES The SPRUCE Study has the following overall objectives: * To compare the vaginal microbiome of multiple pregnancies to singleton pregnancies * To assess the association of the vaginal microbiome to sPTB in multiple pregnancies * To assess the association of the vaginal metabolome to sPTB in multiple pregnancies STUDY DESIGN AND METHODS OF DATA COLLECTION A single site prospective cohort study of 120 women with a multiple pregnancy. High vaginal swabs, blood sampling and CL measurement will be performed at 16- and 28-weeks' gestation. Cervical stiffness will also be measured at 16 weeks gestation (during swab collection). Vaginal swabs will then be collected at the onset of labour/induction/prior to caesarean section (C/S) in women with intact membranes (See Figure 1). Vaginal samples from all women (n=120) will be characterised by 16S rRNA gene sequencing and quantification, metabolomic profiling, Candida quantitative PCR, and FFN, and samples from PTB cases (approximately 23) and matched controls will additionally undergo metagenomic profiling. Blood samples will be stored for future exploratory biomarker and proteomic analysis relating to preterm birth prediction and cervical stiffness measurements will be recorded for future research regarding its potential associations with preterm birth. STUDY DESIGN Participants will be recruited directly from the multiple pregnancy clinic (MPC) at LWH. All women pregnant with twins will receive a patient information leaflet when booking for antenatal care and will be invited to participate at their first visit to the MPC at 16 weeks' gestation by a member of the research team. At this point they will receive further verbal information on the study and an opportunity to ask questions. If the woman agrees to participate, they will sign the study specific informed consent form and be registered onto a bespoke electronic data capture system that will generate a unique participant identification number. Women will be allowed as long as they wish to decide whether to participate in the study. If a woman requires further time to decide on participation in the study a further appointment will be offered. Sample collection will only be arranged once the participant has consented. Following informed consent, women will be asked to complete a health questionnaire providing information on maternal medical and surgical history, basic demographics and socioeconomic background. They will then be asked to provide high vaginal swab samples and cervical stiffness measurements (via speculum examination), blood samples and CL measurements (via transvaginal ultrasound). These assessments will be carried out at 16 weeks' gestation. Vaginal swabs, blood samples and CL measurement will be repeated at 28 weeks' gestation and further vaginal swabs will be collected at the onset of labour/induction/prior to caesarean section (CS) in women with intact membranes. Prior to each collection of blood and speculum examination patients will fill in a brief questionnaire stating if any current medications, recent antibiotic and/or antifungal treatment or sexual intercourse. If the patient has had vaginal sexual intercourse in the last 48 hours sample collection will be postponed to a later appointment. If there is evidence of a shortened CL (\<25mm) at 16 weeks the patient will be treated as per local Liverpool Women's Hospital (LWH) policy. The participant would not be excluded from the research study and planned assessments would still take place. For high vaginal swabs the patient will undergo a speculum examination and swabs will be taken under direct vision from the posterior fornix. If swabs are being taken at the time of delivery, they will be performed prior to digital vaginal examination. Cervical stiffness measurement will be taken simultaneously alongside vaginal swabs. Once recruited, participants will remain in the study until delivery (routinely not later than 38 weeks in any twin pregnancy) and discharge from hospital. Routine clinical data will be collected from all participants' notes and electronic hospital records for maternal and neonatal outcomes. All study visits will coincide with routine care and all participating women will receive routine antenatal care as per NICE Twin and Triplet pregnancy guidelines throughout their pregnancy. SAMPLE SIZE The Multiple Pregnancy Clinic at LWH performs 100 CL scans annually. A recruitment rate of 80% would be expected due to the acceptability of a CL scan. Over an 18-month period a conservative 120 participants would therefore be expected. Based on the estimated rate of preterm deliveries \<34 weeks of 19% (LWH data) then an estimated level of precision is given by a standard error or 3.6% meaning that the early delivery rate should be estimated with a 95% confidence interval defined by +-7% degree of precision. Further this should lead to approximately 23 preterm deliveries within the observational cohort which will allow for multivariable analyses including 2 covariates to establish the strength of association between clinical/demographic factors of interest against early delivery rate. OUTCOME MEASURES The primary outcome will be sPTB \<34 weeks' gestation. Vaginal microbiota predictors based on 16S rRNA gene sequencing will include total vaginal bacterial load; vaginal microbiota composition type; and vaginal relative abundance and estimated concentrations of L. crispatus, L. iners, all lactobacilli combined, all BV-anaerobes combined, and all pathobionts combined. Additional predictors that are measured in all women will include vaginal Candida yeasts concentration, FFN concentration and metabolomic data using 20-30 metabolomic predictors that are of interest based on the results of previous studies in singleton pregnancies by David MacIntyre et al, Imperial College London. In the nested case-control study, additional predictors will be distilled from metagenomics data using established bioinformatic pipelines. Secondary maternal and neonatal outcomes will be recorded for descriptive analyses: * Maternal outcomes: PPROM, maternal Infection (requiring IV antibiotic treatment for suspected chorioamnionitis), gestation at delivery (\<28 weeks/\<32 weeks), mode of delivery, PTB treatment. * Neonatal outcomes: pregnancy complications (TTTS/SFGR/TAPS), livebirth, fetal/neonatal death, birthweight, neonatal morbidity (IVH/NEC/ROP), admission to NICU, ventilator dependence and age at NICU discharge. STATISTICAL METHODOLOGY The full 16S microbiota data (all bacterial taxa in all samples) will first be visualised in heat maps for all women and for women with and without sPTB. The data will then be summarised for use in biostatistical aetiological models by calculating the following for each individual sample: Estimated concentrations of L. crispatus, L. iners, all lactobacilli combined, all BV-anaerobes combined, all pathobionts combined, all bacteria combined (from the 16S gene qPCR) and all Candida species combined (also by qPCR); alpha diversity (Inverse Simpson index); and microbiota composition types by hierarchical clustering. These microbiota variables will be tested one at a time using logistic regression with birth outcome as the dependent variable. The results of these analyses will also be compared to the results of similar analyses in our pre-existing high and low risk preterm singleton cohorts (1). Unsupervised and supervised multivariate analyses will be performed on the metagenome and metabolome datasets. Subsequent analysis using principal component analysis (PCA) will be used for visualisation of general clustering trends and a Random Forest (RF) classifier will be used for group discriminatory analyses. Targeted analyses of specific DESI-MS metabolites features derived from short and long chain fatty acids and biogenic amines that have previously been found to correlate with vaginal microbiome composition and immune/inflammatory activation status will also be performed(2).