P–385 The relationship between systemic oestradiol and vaginal microbiota composition in miscarriage and normal pregnancy

Study question Is there an association between serum oestradiol, vaginal microbial composition and pregnancy outcome in the early first trimester? Summary answer In women with a vaginal microbiome deplete of Lactobacillus species at the time of Pregnancy of Uncertain Viability (IPUV), higher serum oestradiol associates with livebirth. What is known already During pregnancy, oestradiol mediates vaginal mucosal properties and increases glycogen deposition in epithelial cells which is thought to support colonisation of Lactobacillus species. Low levels of Lactobacillus associates with adverse outcomes such as miscarriage and preterm birth. The direct relationship between systemic oestradiol and the vaginal microbiome has never been studied in pregnancy. However studies have shown a positive correlation between serum oestrone, vaginal glycogen and Lactobacillus abundance in menopausal women. Study design, size, duration This was a prospective cohort study where one-hundred women were recruited in early pregnancy at the time of IPUV and donated paired blood and vaginal samples. 40 women had an eventual miscarriage, 58 had a livebirth and two pregnancies were terminated. All 100 women donated one paired serum and vaginal sample at this time point, and 22 women with Lactobacillus depletion at the time of IPUV donated further longitudinal vaginal samples. Participants/materials, setting, methods Participants were recruited from an Early Pregnancy Unit and underwent transvaginal ultrasound assessment of their pregnancy. Serum samples were analysed with an immunoassay on a ROCHE COBAS E411 analyser for Oestradiol (pg/ml) and Progesterone (ng/ml). Bacterial DNA was extracted from paired vaginal swabs and sequenced using Illumina MiSeq sequencing of 16S rRNA gene amplicons. Main results and the role of chance Lactobacillus dominance of the vagina was associated with higher serum levels of E2 and progesterone compared to depletion (E2=398pg/ml vs 302pg/ml(p = 0.02), P4=23.1ng/ml vs 17ng/ml(p = 0.02)). E2 and P4 were positively correlated (r = 0.6, p < 0.05). At species level, L. crispatus dominance associated with significantly higher levels of E2 compared to high-diversity communities (468pg/ml vs 302pg/ml(p = 0.03) but no such relationship was observed for P4. Both E2 and P4 levels were lower in women who eventually miscarried. However there was no significant difference in the vaginal bacterial composition at genera or species level at this early gestational age (P = 0.08) regardless of per vaginal bleeding. However in women with Lactobacillus depleted microbiota, livebirth was associated with significantly higher E2 levels compared to women suffering miscarriage (212pg/ml in miscarriage vs 395pg/ml in livebirth, p = 0.003) (OR = 22.4 P = 0.004). In 22 women who had Lactobacillus depletion at the time of IPUV (7 with an eventual outcome of miscarriage, and 15 with an eventual outcome of livebirth), longitudinal vaginal bacterial DNA sequencing was performed. In 7/15 women with livebirth, and higher E2 levels, the microbial composition changed to become more Lactobacillus dominant during pregnancy, whereas in those with miscarriage, only 1/7 changed to become Lactobacillus dominant. Limitations, reasons for caution In this study, serum oestradiol levels were compared to the local vaginal bacterial environment. The ideal would be to study local vaginal oestradiol, glycogen and the bacterial composition. Wider implications of the findings: In contrast to previous studies in menopause where low oestrogen levels associate with the vaginal microbial composition, this study uses the high oestradiol environment of early pregnancy to study the mechanistic relationship between oestradiol and vaginal Lactobacillus abundance. Trial registration number NA