The predictive value of a multivariable model based on vaginal Lactobacillus relative abundance and microecological features at 24 weeks gestation and before delivery in spontaneous preterm birth: A Prospective cohort study

Abstract

Background Vaginal microecological imbalance, particularly reduced Lactobacillus dominance, is a significant risk factor for spontaneous preterm birth (sPTB).However, the predictive value of microbial composition, pathogen colonization, and vaginal inflammation biomarkers at different pregnancy stages remains unclear. This study aims to evaluate the predictive potential of vaginal microecology at 24 weeks of gestation and before delivery for spontaneous preterm birth. Methods This was a prospective multicenter cohort study. Vaginal swabs were collected at 24 weeks of gestation and before delivery to assess Lactobacillus relative abundance(LRA), pathogen colonization, and biochemical markers (pH, hydrogen peroxide, leukocyte esterase, and sialidase). Multivariable logistic regression models, ROC curves, and Kaplan-Meier survival analysis were used to identify predictors of spontaneous preterm birth. Results At 24 weeks of gestation, the preterm group showed significantly lower Lactobacillus relative abundance, higher AV and Nugent scores, and higher vaginal pH. Logistic regression revealed that an increase in Lactobacillus relative abundance(LRA) was negatively associated with sPTB (OR = 0.97, 95% CI = 0.96–0.98, P < 0.001), while higher parity (OR = 2.28, 95% CI = 1.56–3.32, P < 0.001) and a history of late miscarriage/preterm birth (OR = 6.96, 95% CI = 2.77–17.47, P < 0.001) increased the risk. Lactobacillus relative abundance(LRA) below 57.5% was linked to increased preterm birth risk in univariate analysis, but this was no longer significant after multivariable adjustment. Parity, late miscarriage/preterm birth history, vaginal pH > 4.5, and GBS positivity remained independent risk factors (AUC = 0.775). Kaplan-Meier survival analysis showed that women with LRA > 72.5% at 24 weeks and LRA > 57.5% before delivery had significantly higher term delivery probabilities (log-rank P < 0.001). Conclusion At 24 weeks, reduced vaginal LRA is an independent predictor of spontaneous preterm birth. Early vaginal microbiome assessment, combined with clinical features, can help identify high-risk pregnant women and improve outcomes.