25. Molecular Bacterial Vaginosis and Prospective Risk of Cervicovaginal Chlamydia Trachomatis Infection in Adolescents


      Adolescents make up only a quarter of the population but account for over half of all incident sexually transmitted infections. Bacterial vaginosis has been associated with cervicovaginal Chlamydia trachomatis (CT), although the mechanisms of their association remain poorly characterized. This study investigated the prospective association between the cervicovaginal microbiome (CVM) and the risk of acquiring CT infection in adolescents.


      We conducted a nested case-control study of CT within an HPV longitudinal cohort study of sexually active adolescents from a large adolescent health center in New York City. Controls were matched to incident (newly diagnosed) CT cases using risk set sampling based on age, date of study enrollment, and prior history of CT infection. For this study, cervicovaginal swab samples collected from three time points were analyzed: approximately six months before CT infection (V1), at the time of CT diagnosis (V2), and six months after infection treatment (V3). Clinical CT was diagnosed using a GEN-PROBE APTIMA assay. The CVM was evaluated using 16SV4 rRNA bacterial gene amplicon next-generation sequencing. Sequence reads were clustered into amplicon sequence variants using DADA2 and taxonomy was assigned using a custom cervicovaginal microbiome specific database employing a Naive Bayesian classifier. We assessed the presence of bacterial vaginosis (BV) using a previously validated molecular score from the 16SV4 reads (molBV) that provides a Nugent-like score on a scale from 0 - 10. At V1, 232/502 (46.2%) cervicovaginal samples were molBV-BV negative (score 0-3), 79/502 (15.7%) were molBV-BV intermediate (4-6), and 191/502 (38.0%) were molBV-BV positive (7-10). The relative risk (RR) of incident CT based on the molBV-BV score was assessed by comparing participants who were BV positive to those intermediate and negative using multivariable conditional regression adjusting for patient BMI, number of sexual partners within the last 3 months, and condom use.


      16SV4 rRNA gene amplicon sequencing yielded an average of 33,339±13,280 bacterial reads with no significant differences based on case status (p=0.08). Subclinical detection of BV by the molBV score at V1 significantly predicted risk of CT diagnosis at V2, 6 months later (RR=1.78, 95%CI=1.05-3.03). Adjusted cross-sectional analysis of V2 samples revealed an even stronger association between molBV-BV positive status and clinical detection of CT (RR=2.90, 95%CI=1.74-4.85), whereas following CT treatment (V3) there were no differences between cases and controls in their BV scores (p=0.5).


      While it is known that having a clinical diagnosis of BV is associated with cervicovaginal CT, to our knowledge this is the first study to prospectively demonstrate an association between subclinical detection of BV by next-generation molecular sequencing of the CVM and the risk of acquiring a cervicovaginal CT infection.

      Sources of Support

      This work was supported in part by the National Institute of Allergy and Infectious Diseases (R01AI072204) and the Einstein Cancer Research Center (P30CA013330), and the Einstein-Rockefeller-CUNY Center for AIDS Research funded by the NIAID (P30AI124414).