Preterm birth, the leading cause of death in children under five, may be caused by inflammation triggered by ascending vaginal infection. About two million cervical cerclages are performed annually to prevent preterm birth. The procedure is thought to provide structural support and maintain the endocervical mucus plug as a barrier to ascending infection. Two types of suture material are used for cerclage: monofilament or multifilament braided. Braided sutures are most frequently used, though no evidence exists to favor them over monofilament sutures. In this study we assessed birth outcomes in a retrospective cohort of 678 women receiving cervical cerclage in 5 UK university hospitals and showed that braided cerclage was associated with increased intrauterine death (15% v 5%, P = 0.0001) and preterm birth (28% v 17%, P = 0.0006) compared to monofilament suture. To understand the potential underlying mechanism, we performed a prospective, longitudinal study of the vaginal microbiome in women at risk of preterm birth because of short cervical length (≤25 mm) who received braided (n=25) or monofilament (n=24) cerclage under otherwise comparable circumstances. Braided suture induced a persistent shift towards vaginal microbiome dysbiosis characterized by reduced Lactobacillus spp. and enrichment of pathobionts. Vaginal dysbiosis was associated with inflammatory cytokine and interstitial collagenase excretion into cervicovaginal fluid and premature cervical remodeling. Monofilament suture had comparatively minimal impact upon the vaginal microbiome and its interactions with the host. These data provide in vivo evidence that a dynamic shift of the human vaginal microbiome toward dysbiosis correlates with preterm birth.