Meta-Analysis Reveals the Vaginal Microbiome is a Better Predictor of Earlier Than Later Preterm Birth

Abstract

High-throughput sequencing measurements of the vaginal microbiome have yielded intriguing potential relationships between the vaginal microbiome and preterm birth (PTB; live birth prior to 37 weeks of gestation). However, results across studies have been inconsistent. Here we perform an integrated analysis of previously published datasets from 12 cohorts of pregnant women whose vaginal microbiomes were measured by 16S rRNA gene sequencing. Of 1926 women included in our analysis, 568 went on to deliver prematurely. Substantial variation between these datasets existed in their definition of preterm birth, characteristics of the study populations, and sequencing methodology. Nevertheless, a small group of taxa comprised a vast majority of the measured microbiome in all cohorts. We trained machine learning (ML) models to predict PTB from the composition of the vaginal microbiome, finding low to modest predictive accuracy (0.28-0.79). Predictive accuracy was typically lower when ML models trained in one dataset predicted PTB in another dataset. Earlier preterm birth (<32 weeks, <34 weeks) was more predictable from the vaginal microbiome than late preterm birth (34 - 37 weeks), both within and across datasets. Integrated differential abundance analysis revealed a highly significant negative association between L. crispatus and PTB that was consistent across almost all studies. The presence of the majority (18 out of 25) of genera was associated with a higher risk of PTB, with L. iners, Prevotella, and Gardnerella showing particularly consistent and significant associations. Some example discrepancies between studies could be attributed to specific methodological differences, but not most study-to-study variations in the relationship between the vaginal microbiome and preterm birth. We believe future studies of the vaginal microbiome and PTB will benefit from a focus on earlier preterm births, and improved reporting of specific patient metadata shown to influence the vaginal microbiome and/or birth outcomes.

Microbiome Preterm Birth DREAM Challenge: Crowdsourcing Machine Learning Approaches to Advance Preterm Birth Research

“…Metaanalysis as well as rigorous evaluation of models on independent validation data is a robust approach to contend with these biological challenges with microbiome data. However there are significant technical challenges in aggregating and combining microbiome data across studies, therefore there have been few studies taking on this task [30][31][32] . In previous work, we have shown that by aggregating microbiome data across several studies we can gain significant statistical power to show that higher diversity is associated with PTB especially in the first trimester of pregnancy and to identify several novel microbial associations 33 .…”

“…While ML approaches have been applied to the vaginal microbiome, most have involved a single dataset with limited sample size [34][35][36] . One recent work explored the application of ML to 12 vaginal microbiome datasets to predict PTB; however, while they leveraged public data extensively to ensure their findings were robust across studies, their work did not include an independent validation dataset 30 . Moreover, their work involved a single approach -a random forest ML model -with predictive accuracy for PTB ranging from 0.28 to 0.79.…”

GardnerellaDiversity and Ecology in Pregnancy and Preterm Birth

“…A common bioinformatics workflow was used to define ASVs in each of these studies, yielding a total of 85 unique Gardnerella ASVs across all samples and all cohorts. 57 However, many of these variants were found in very few samples and at very low abundances (Figure S12). 99% of all Gardnerella reads were contained in just the top 5 ASVs, which correspond to G1-G5 described above.…”

“…13,14,[53][54][55][56] The sequencing data were processed uniformly as part of a larger meta-analysis effort. 57 Seven cohorts from six publications were assessed. Exact sequencing variants were resolved using DADA2 58 version 1.12.1.…”

Advances in mass spectrometry technologies to characterize cervicovaginal microbiome functions that impact spontaneous preterm birth

“…The bacteria associated with sPTB differ between studies and appear to vary significantly at the individual level, dependent on many factors, including specific combinations of bacteria, strains of bacteria, bacterial load, sites of bacterial localization, kinetics and timing of infection, the level of inflammation induced, as well as ethnicity and geographical region. 40,48 These differences hinder the development of accurate, universal diagnostics and effective strategies to treat women at risk of sPTB.…”