Dynamic changes of cervical microbiome during pregnancy for preterm birth risk prediction: A prospective cohort study.

This prospective cohort study investigated the dynamic changes in the cervical microbiome during pregnancy and developed a predictive model for preterm birth risk. Ninety-three singleton pregnant women were enrolled, including 41 with preterm birth and 52 with term delivery. Cervical secretions were collected at four gestational stages and analyzed using 16S rRNA sequencing, alongside ELISA testing for inflammatory markers. The preterm group exhibited significantly lower microbial diversity and a progressively increasing ratio of Lactobacillus iners to Lactobacillus crispatus throughout pregnancy. Early pregnancy IL-6 levels were also significantly elevated in this group. Logistic regression identified the L. iners/L. crispatus ratio, IL-6, history of preterm birth, and short cervical length as independent risk factors. The integrated prediction model demonstrated high accuracy (AUC 0.847), with even stronger performance in predicting births before 34 weeks (AUC 0.892). These findings suggest that microbiome patterns and inflammatory markers can effectively predict preterm birth risk, supporting early clinical intervention.