O-276 Implantation failure is not associated with seven different vaginal bacterial community-state types; however, a distinct signature, including Ureaplasma parvum, is strongly linked to treatment failure

Abstract

Abstract Study question Is microbial diversity or vaginal bacterial composition associated with embryonic implantation in women undergoing frozen-thawed embryo transfer? Summary answer While global diversity measures of microbial composition do not correlate with positive implantation, specific taxa may serve as predictors of treatment outcomes. What is known already Clinical studies have indicated a reduced likelihood of embryo implantation in women diagnosed with microbial dysbiosis. The application of Next-Generation Sequencing for microbiota analysis has significantly advanced our understanding of vaginal microbiome composition. Recent studies have proposed distinct vaginal microbiome patterns as potential predictors of IVF treatment success or failure. However, the evidence regarding the biological basis of test algorithms (such as the independent association of distinct community state types with outcomes) and the reliability of diagnostic test performance (including replicability and external validity) in predicting implantation success or failure in women undergoing embryo transfer remains less robust. Study design, size, duration In a prospective, single-centre observational cohort study (05/2017-03/2019, NCT03507673, local ethical approval reference no.18-005), conducted at a university-affiliated centre for reproductive medicine in Germany, vaginal swab samples were collected from infertile, asymptomatic female patients (n = 260), aged 18-45, undergoing frozen-thawed embryo transfer. Patients, each contributing a single treatment cycle, received sequential oral estradiol and oral estradiol/dydrogesterone in an HRT regimen for FET. Vaginal swabs were taken from baseline visit before transfer day and stored at –80 °C. Participants/materials, setting, methods 16S rRNA gene amplicon sequencing was performed targeting V3/V4 hypervariable region using Illumina MiSeq platform. Raw data assembly, filtering and taxonomic assignment were performed using mothur against EzBioCloud database. Statistical analyses, and visualization were conducted using R. Community structure via hierarchical clustering, alpha and beta diversity, and indicator species via LefSe analysis were assessed. Main results and the role of chance Based on a hierarchical clustering using the 25 most abundant bacterial taxa, seven separate community state types (CSTs) being either dominated by species of the genus Lactobacillus (CST-A, CST-B, CST-C, CST-E, CST-G) or Gardnerella (CST-F) or presenting a variable community composition (CST-D) were identified. A community state type dominated by Lactobacillus fornicalis was found in a small fraction of women and enhanced resolution among the Gardnerella genus impacts CST definitions in this sample. Embryonic implantation had no significant association to either alpha diversity measures or CST distribution; likewise, the previously suggested ratio of relative abundances of Lactobacillus iners vs. Lactobacillus crispatus was not associated with implantation. Lactobacillus fornicalis is positively linked to implantation, while Lactobacillus iners and Ureaplasma parvum were linked to implantation failure: above 50% relative abundance of Lactobacillus fornicalis, was associated with 75% positive implantation. Conversely, just 0.5% relative abundance of Ureaplasma parvum was associated with a decrease in implantation rate to approximately 20%. LefSe and binary regression analysis identified a relatively rare (6% in our sample) microbiome signature, mostly from CST-C (Lactobacillus iners dominated) with a very low likelihood of implantation, with Ureaplasma parvum as the common denominator. Limitations, reasons for caution Studies on the vaginal microbiome in IVF face issues such as study design variability, small sample-sizes, limited control for temporal dynamics, and patient heterogeneity. Attempts at establishing causal relationships through interventional studies in IVF have been limited and inconclusive. Understanding the microbiome’s functional aspects is crucial for reproductive outcome elucidation. Wider implications of the findings Current study findings don’t confirm previously reported associations between common vaginal bacterial CSTs and outcome. Therefore, screening for vaginal microbiome composition in asymptomatic women with subfertility is not recommended outside research. Signatures/taxae with an independent and strong negative association with implantation are likely to be rare yet deserve further attention. Trial registration number NCT03507673