Advertisement

104: The association between bacterial vaginosis and spontaneous preterm birth is modified by history of PPROM

      Objective

      To estimate if the risk of spontaneous preterm birth (SPTB) associated with bacterial vaginosis (BV) is modified by three specific components of pregnancy history.

      Study Design

      We performed a 4-year prospective cohort study of urban pregnant subjects who were universally screened for BV at <13 weeks with vaginal Gram stain and the Nugent criteria interpreted by a trained technician. The primary outcome was spontaneous preterm delivery (<37 weeks). Providers and investigators were blinded to the BV screen results. Data on multiple obstetric, medical and demographic variables were collected using a questionnaire, mid-pregnancy interview and postpartum chart review. Bivariable and multivariable statistical methods were used to assess the exposure-outcome association and pertinent interactions identified a priori: preterm premature rupture of membranes (PPROM), symptomatology, and antibiotic therapy.

      Results

      Among 1826 enrolled patients, 1452 progressed beyond 22 weeks and had complete data. The prevalence of BV was 37% (n=537): 13% symptomatic and 24% asymptomatic. The prevalences of spontaneous and any preterm birth were 7.3% (n=106) & 11.4% (n=153), respectively. BV was not significantly associated with spontaneous preterm birth in the overall population in the unadjusted or multivariable analysis (Adjusted OR 0.8, 95%CI 0.5-1.2). However, the association between BV and SPTB was modified by pregnancy history, such that women with prior PPROM and current BV had a markedly increased odds of preterm birth compared to women with BV without prior PPROM (Interaction adjusted OR 8.3, 95%CI 1.5-47.6). The BV-SPTB association was not modified by symptomatology or prenatal antibiotic treatment.

      Conclusion

      In this general obstetric population, BV conveyed increased risk for spontaneous preterm birth only if preceded by a history of PPROM in a prior pregnancy. This interaction could be the result of genetic variation or environmental factors that predispose the high-risk patient to a hyper-inflammatory response to BV.