Effect of periconceptual vaginal microbiota disruption on fecundability among Kenyan women planning pregnancies

Abstract

INTRODUCTION: Vaginal microbiota disruption is a common condition that is associated with sexually transmitted infection and HIV acquisition, spontaneous abortion, preterm birth, and poor in-vitro fertilization outcomes. Whether vaginal microbiota disruption is associated with non-medically assisted reproduction has not been thoroughly explored. This dissertation addresses the hypothesis that disruption of the optimal, Lactobacillus dominated vaginal microbiota during the periconception period is associated with reduced fecundability among HIV-negative Kenyan women. METHODS: These analyses leveraged data from a prospective case-cohort study assessing the impact of periconceptual vaginal microbiota on women’s risk of preterm birth (Chapter 2). At monthly preconception visits, participants reported the first day of last menstrual periods and recent sexual behavior, underwent a urine pregnancy test, and provided vaginal swab specimens. Discrete time proportional probabilities models with robust standard errors were used to estimate fecundability ratios (FR) and 95% confidence intervals (CI) among menstrual cycles with and without vaginal microbiota disruption assessed using non-molecular methods of detection. Three non-molecular methods of detection were assessed including bacterial vaginosis (BV) by Nugent score (≥7) (Chapter 3), elevated sialidase activity (≥0.25 g) detected by a point of care test for BV (Chapter 3), and cultivation of Lactobacillus on Rogosa agar (Chapter 4). RESULTS: Among 273 eligible women, 35.5% (n=97) had BV, 33.3% (n=91) had elevated sialidase, and 30.8% (n=84) had no cultivable Lactobacillus at enrollment. Participants contributed 768 menstrual cycles and 128 became pregnant. The six-menstrual cycle cumulative pregnancy rate was 69.3% (95%CI 61.5-82.7). Neither BV (adjusted FR [aFR] 0.95, 95%CI 0.66-1.35), elevated sialidase (aFR 1.12, 95%CI 0.79-1.59), nor cultivable Lactobacillus (aFR 1.14, 95%CI 0.77-1.68) at the visit prior to each pregnancy test were significantly associated with fecundability. However, when considering two consecutive measurements per discrete menstrual cycle, the per-cycle probability of pregnancy among menstrual cycles with persistent BV was associated with a 34% reduction in fecundability that trended toward significance compared to cycles that were BV negative at both visits (aFR 0.66, 95%CI 0.42-1.03, p=0.07). In addition, the per-cycle probabilities of pregnancy in menstrual cycles with cultivable Lactobacillus at one or both consecutive visits were a non-significant 1.1-1.6 fold higher compared to cycles negative for Lactobacillus at both visits. Lastly, menstrual cycles with Lactobacillus morphotypes detected on Gram stain at both consecutive measurements had a 52% increased per-cycle probability of pregnancy that trended toward significance (aFR 1.52, 95%CI 0.97-2.41). CONCLUSIONS: There were no significant associations between BV by Nugent score, elevated sialidase, or cultivable Lactobacillus at the visit prior to pregnancy testing and fecundability. However, the secondary analyses assessing the stability of the vaginal microbiota across two time points for each menstrual cycle provide some evidence that persistent vaginal microbiota disruption may contribute to reduced fecundability, and that Lactobacillus may be associated with modestly increased fecundability. If persistent vaginal microbiota disruption is associated with reduced fecundity, this could have important implications for a large number of women who wish to conceive due to the high global prevalence of BV and infertility. Importantly, these results are preliminary as the target sample size was not reached by the analysis date. Future studies should consider utilizing both culture-dependent and molecular methods of detection to explore the association between the presence, viability, and functionality of vaginal bacteria and fecundability.