(P460) Metagenomic Community State Types and Recurrent Bacterial Vaginosis

Background: Bacterial vaginosis (BV) recurrence following antibiotic treatment ranges from 50-70% within 6 months. We hypothesized that the pre-treatment vaginal microbiome in patients with BV is in part associated with recurrence. Using metagenomic sequencing to characterize the species composition and gene-based functional information of the vaginal microbiome, we determined the association between antecedent vaginal metagenomic community state types (mgCSTs) and recurrent BV (rBV) within six months.



Methods: This pilot study utilized data from a nested case control study of incident STI within the 1999-2002 Longitudinal Study for Vaginal Flora. Participants were followed for up to 12 months with quarterly examinations. BV was diagnosed by ≥ 3 Amsel’s criteria (consistent discharge, pH >4.5, whiff+, clue cells). In this secondary analysis, rBV cases (N=39) were defined by an index visit with Amsel-BV that cleared 3 months later but then recurred 6 months after the index visit. Controls (N=70) had Amsel-BV that did not recur within 6 months. All patients included were gonorrhea, chlamydia and trichomoniasis negative at the time of sampling. Metagenomes were sequenced on the Illumina HiSeq 4000 and mgCSTs assigned using the mgCST-classifier. The association between mgCSTs and rBV was modeled with logistic regression.


Results: Eleven mgCSTs were observed (three dominated by L. iners, three by “Ca. Lachnocurva vaginae”, and five by Gardnerella). Other Lactobacillus-dominated mgCSTs were not observed. After adjustment for age, race, BV treatment regimen, patients with specific Gardnerella vaginalis-dominated mgCSTs (mgCST 20 and 24) had 8- to 10-fold greater odds of rBV at six months (OR:10.8, 95% CI:1.9-91 and OR:8.8, 95% CI:1.7-67, respectively) compared to L. iners-dominated microbiomes. Other mgCSTs dominated by G. vaginalis (mgCST 22 and 23) were not statistically significant.



Conclusion: This is the first study to use metagenomic data to assess the vaginal microbiome as a risk factor for recurrent Amsel-BV. Functionally distinct microbiomes were present in Amsel-BV, even when predominated by the same species. L. iners was the only Lactobacillus species dominant in these Amsel-BV cases. Functionally specific Gardnerella-dominated microbiomes were strongly associated with BV recurrence at six months. Exploration of the vaginal metagenome may yield valuable information on the pathophysiology of recurrent BV.