LAUSR

1 Evaluation of patientversus providercollected vaginal swabs for microbiome analysis during pregnancy S. A. Blankenship, K. M. Wylie, M. G. Tuuli, G. A. Macones, M. J. Stout Department of Obstetrics and Gynecology, Washington University in St. Louis School of Medicine, St. Louis, MO, Department of Pediatrics, Washington University School in St. Louis School of Medicine, St. Louis, MO, The McDonnell Genome Institute, Washington University in St. Louis School of Medicine, St. Louis, MO OBJECTIVES: Prior microbiome analyses in non-pregnant women demonstrate similar vaginal bacterial communities with patientand provider-based sampling. Given changes in the vaginal microbiome and maternal physiology during pregnancy, including the gravid abdomen and leukorrhea of pregnancy, we aimed to evaluate if patientand provider-collected vaginal swabs reflect similar microbial community characteristics in pregnant women. METHODS: Pregnant patients performed a self-collected vaginal swab, and subsequently underwent a provider-collected swab via speculum exam during a routine prenatal care visit. DNA pyrosequencing of V1V3 and V3V5 variable regions of the 16S rRNA gene of patientand provider-collected swabs were performed. Community characteristics of patientand provider-collected swabs were compared, including relative abundance of taxa, alpha diversity and beta diversity. RESULTS: Ninety-four vaginal swabs from 47 women were analyzed, with mean 5572 16S rRNA gene sequences obtained per sample. Median gestational age of sampling was 20.1 weeks (interquartile range 12.4-28.0 weeks) and ranged from 5 to 33 weeks. On nonmetric multi-dimensional scaling plots, patient-collected swabs did not cluster separately from provider-collected swabs; paired patientand physician-collected swabs clustered closely. Among paired samples of 16S rRNA sequences from V1V3, 39 of 47 paired samples (83.0%) showed strong correlation between providerand patientcollected swabs (Pearson correlation coefficient >0.9); the 8 V1V3 sequences with weaker Pearson correlation (<0.9) had correlation coefficients ranging from 0.57 to 0.89 Additionally, 34 of 46 paired samples (73.9%) of 16S rRNA sequences from V3V5 also demonstrated strong Pearson correlation (>0.9); the 12 samples in the V3V5 region with weaker correlation had correlation coefficients ranging from 0.49 to 0.89. No specific taxa were preferentially detected by self-sampling or physician-sampling, with relative abundance of taxa highly conserved between paired samples. No significant difference in Shannon diversity for V1V3 (p1⁄40.22) and V3V5 (p1⁄40.11) sequences among patientand provider-collected samples was demonstrated. CONCLUSIONS: We demonstrate that bacterial communities defined from patientand provider-collected vaginal swabs in pregnant women are similar, validating utilization of patient-collected swabs for vaginal bacterial microbiome sampling during pregnancy.