The composition and stability of the vaginal microbiota of normal pregnant women is different from that of non-pregnant women

The body during pregnancy becomes more microbially diverse — more open and more variable — to accommodate a new life. That sounds right. It's intuitive. And it's wrong. Pregnancy actually locks the vaginal microbiome into a tighter, more uniform state than it ever occupies outside of pregnancy. Romero and colleagues proved it, and what they found has implications that reach far beyond microbiology. Here's what you need to know first about the vaginal microbiome. It’s genuinely unlike any other microbial community in the body. In the gut, diversity is celebrated. More species, more functions, more resilience. The vagina operates by a different logic entirely. In healthy reproductive-age women, the vaginal ecosystem tends to be dominated by a single genus: Lactobacillus. These bacteria produce lactic acid, keep the vaginal pH below 4.5, and appear to protect against pathogens. In most body sites, a community this simple would look impoverished. Here, it's the signature of health. Researchers in this field organize vaginal communities into what they call community state types, or CSTs. Think of a CST as a stable configuration — a recurring pattern of which bacteria are present and in what proportions. Four of the main CSTs are each dominated by a single Lactobacillus species: L. crispatus, L. iners, L. gasseri, or L. jensenii. Then there are the CST IV states — IV-A and IV-B — which lack substantial Lactobacillus and instead contain a diverse collection of anaerobic bacteria. Atopobium, Prevotella, Sneathia, Gardnerella, Mobiluncus. These are the communities associated with bacterial vaginosis, the most common vaginal condition in reproductive-age women, and one that carries real clinical risk. The question Romero and colleagues set out to answer was simple: when a woman becomes pregnant, which CSTs does she occupy, and do they change? To find out, they ran a retrospective case-control longitudinal study involving thirty-two non-pregnant women and twenty-two pregnant women who delivered at full term, between thirty-eight and forty-two weeks, without complications. Both groups contributed serial vaginal fluid samples collected repeatedly over time. This made it possible to measure not just composition but change. The team used sixteen S ribosomal RNA gene sequencing to identify the bacteria — a culture-independent method that fingerprints microbial communities by a gene all bacteria carry — running samples through a high-throughput pyrosequencing platform. To handle the statistical complexity of repeated measurements from the same individuals, they used linear mixed-effects models and generalized estimating equations. The design is important to hold in mind: this wasn't a single snapshot. It tracked the same women's vaginal microbiomes over time. The results were systematic. Pregnant women's vaginal communities were substantially more dominated by Lactobacillus than those of non-pregnant women — not across one species but across four. L. vaginalis was five and a half times higher in pregnancy. L. jensenii was four point seven times higher. L. gasseri was three point three times higher. L. crispatus was about two times higher. Notably, L. iners — one of the more common Lactobacillus species in non-pregnant women — showed no significant difference between groups. The enrichment wasn't uniform across all Lactobacillus. It was specific. On the other side of the ledger, the bacterial vaginosis-associated CSTs nearly vanished. In the non-pregnant group, twenty-eight and a half percent of samples — more than one in four — were classified as CST IV-B, the state most characterized by high Atopobium. In pregnant women who delivered at term, that number dropped to two point two percent. The odds of observing CST IV-B in a pregnant woman were reduced by roughly ninety-five percent compared to a non-pregnant woman. That's not a statistical nudge. That's a near-elimination. And it wasn't limited to Atopobium: twenty-two phylotypes in total had significantly lower relative abundance in pregnant women. A phylotype here refers to a group of closely related bacterial sequences identified through sixteen S ribosomal RNA gene classification; some are resolved to species, while some are only resolved to genus. But across twenty-two of them — Prevotella, Sneathia, Gardnerella, Ruminococcaceae, Parvimonas, Mobiluncus, and more — the pattern held. Normal pregnancy systematically suppressed the bacteria most associated with dysbiosis. Now comes the second major finding, and it's the one that makes the first make sense. Pregnancy doesn't just change the composition of the vaginal microbiome. It stabilizes it. Romero and colleagues quantified this using Jensen-Shannon divergence — a mathematical measure of how different two community states are, ranging from zero when they're identical to one when they share nothing. When you apply that metric within the same woman across two time points, a lower score means her community stayed more similar to itself. Lower divergence, more stability. The numbers here are telling. The mean within-subject Jensen-Shannon distance was one point six times lower in pregnant women than in non-pregnant women — a difference of minus zero point four seven log units — and that gap was statistically significant. Non-pregnant women's vaginal communities drifted. Pregnant women's held. When shifts did occur during pregnancy, they almost exclusively moved from one Lactobacillus-dominated CST to another. The community stayed within Lactobacillus space. Transitions into CST IV-A or IV-B — the dysbiotic territory — were rare events. Look at the raw CST frequencies to feel the magnitude. In non-pregnant women across seven hundred sixty-one samples: seventeen percent CST I, thirty-five percent CST III, and twenty-eight point five percent CST IV-B. In pregnant women across one hundred thirty-nine samples: thirty-eight percent CST I, fifty-two percent CST III, and two point two percent CST IV-B. Pregnancy didn't shift these numbers at the margin. It restructured the entire distribution, compressing it toward Lactobacillus dominance and away from dysbiosis. The clinical stakes of this are real. Bacterial vaginosis — the condition characterized by exactly the community state types that pregnancy suppresses — is associated in the literature with preterm birth, late miscarriage, and increased susceptibility to genital infection. What Romero and colleagues documented is that in women who carry pregnancies to term without complications, the vaginal microbiome appears to operate in a protected, Lactobacillus-locked state throughout gestation. This was, as they noted, the first longitudinal, sequence-based study of the vaginal microbiome in normal pregnancy. The baseline it establishes — what the healthy pregnant microbiome actually looks like, week by week — didn't exist before this work. That baseline opens a precise and actionable set of questions. The study enrolled women who delivered at term. What about women who deliver preterm? Do their vaginal communities show earlier or more frequent transitions toward CST IV-A or IV-B? Does microbiome instability — a higher within-subject Jensen-Shannon distance — predict adverse outcomes? Could restoration of Lactobacillus dominance in women with dysbiotic communities during early pregnancy reduce those risks? The study's design — twenty-two pregnant women, a predominantly African-American cohort, with specific primer choices for the sixteen S sequencing — means these questions will require larger, outcome-stratified longitudinal studies to answer. But the framework to ask them now exists. What this paper really establishes is that pregnancy isn't a state of microbial loosening. The common intuition — that the pregnant body opens itself and becomes more permissive — doesn't apply here. The vaginal microbiome in a healthy pregnancy holds tighter. Four Lactobacillus species expand their dominance. Twenty-two other phylotypes are suppressed. The dysbiotic CST IV states become rare. And the whole community becomes more stable over time than it ever is outside of pregnancy. In pregnancy, the body doesn't open up. It holds on. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.