Skip to main content
Research Areas

Community State Types: The Five Vaginal Microbiome Profiles

By November 18, 2026No Comments

When researchers first sequenced vaginal microbiomes in large numbers, the expectation was a spectrum — a continuous range of compositions with no natural divisions. That is not what the data showed. The communities clustered into a small number of recurring patterns, distinct enough to be named.

The framework that emerged, published in 2011 in the Proceedings of the National Academy of Sciences from a study of nearly 400 women, defines five community state types. It is still the standard vocabulary in this field, and understanding it makes a great deal of the research literature legible.

The five types

CST-I — Lactobacillus crispatus dominant. The most stable configuration, and the one most consistently associated with low rates of bacterial vaginosis and reduced susceptibility to infection. If the literature has a reference state, this is it.

CST-II — Lactobacillus gasseri dominant. Less common. Protective, though less studied than CST-I.

CST-III — Lactobacillus iners dominant. Common, and the most interesting of the lactobacillus-dominated types. L. iners is taxonomically a lactobacillus but behaves unlike the others: communities dominated by it transition to CST-IV considerably more readily, and L. iners is frequently present during and after episodes of bacterial vaginosis rather than being displaced by them. It has an unusually small genome and appears less metabolically self-sufficient.

CST-IV — diverse, low in lactobacilli. Characterised by anaerobes including Gardnerella vaginalis, Prevotella, Atopobium vaginae and others, usually with raised pH. This is the state associated with bacterial vaginosis, with increased susceptibility to sexually transmitted infection, and with elevated rates of recurrent urinary tract infection. It is sometimes subdivided — CST-IV-A, IV-B and so on — because the diverse state is itself heterogeneous.

CST-V — Lactobacillus jensenii dominant. The least common of the five. Protective.

The two counterintuitive parts

Low diversity is good here. In almost every other body site, microbial diversity is a marker of health — a diverse gut microbiome is a healthy one. The vagina inverts this. A community that is 90% one species is the healthy state; a diverse community is the one associated with problems. People who have read about the gut microbiome routinely get this backwards, and it is an easy mistake to make.

CST-IV is not a diagnosis. It is a community composition. A substantial proportion of women with CST-IV have no symptoms whatsoever and would never present to a doctor. Bacterial vaginosis — the clinical condition, with discharge, odour and raised pH — occurs against a CST-IV background, but the two are not the same thing. Having CST-IV means being in a state associated with elevated risk, not being ill.

This distinction gets lost constantly in consumer-facing material, usually in the direction of alarm.

Stability, and why it varies

The types differ not only in composition but in how firmly they hold.

Longitudinal studies — sampling the same women repeatedly over weeks and months — show that some women maintain one state almost invariantly, while others shift between states, sometimes week to week. CST-I tends to be the most stable. CST-III tends to be the least, with frequent transitions to and from CST-IV.

Menstruation is a recognised destabiliser: menstrual blood is near-neutral in pH and transiently raises vaginal pH, and communities commonly shift during menses and return afterwards. Whether they return to the same state varies.

The practical implication is that a single sample is a snapshot, not a verdict. One measurement tells you the state on one day.

Why the protective types protect

The mechanism is primarily lactic acid.

Lactobacillus species ferment glycogen from vaginal epithelial cells and produce lactic acid, holding pH in the range of roughly 3.8–4.5. Testing under anaerobic conditions approximating the real environment found that lactic acid at physiological concentrations inactivated all seventeen BV-associated species tested, while leaving vaginal lactobacilli unaffected.

The same work found that hydrogen peroxide — long cited as a protective mechanism, and still widely used in marketing — failed to inactivate any BV-associated species at physiological concentrations, and was more toxic to lactobacilli than to its supposed targets at high ones. A subsequent analysis in Microbiome concluded that an in vivo antimicrobial role for hydrogen peroxide here is implausible. Lactic acid is the supported mechanism; hydrogen peroxide is not.

There is also an isomer distinction. L. crispatus produces both D- and L-lactic acid, while L. iners produces essentially only the L form, and research in mBio has found the two isomers are not biologically equivalent in their effects on extracellular matrix remodelling. This is one plausible reason CST-I behaves differently from CST-III.

What shifts the state

Well established: broad-spectrum antibiotics, which deplete lactobacilli along with the target organism; falling oestrogen at menopause or postpartum, which reduces the epithelial glycogen the community feeds on; douching, which removes the community mechanically and raises pH.

Reasonably supported: menstruation; semen, which is alkaline; some hormonal contraceptives.

Emerging, and particularly relevant in India: household water and sanitation. A 2024 prospective cohort in BMC Infectious Diseases followed 436 adolescent girls in western Kenya over 18 months with 16S rRNA sequencing. Odds of CST-I were higher among those using piped water versus pond water, and lower among those using traditional pit latrines versus flush toilets. Coliform relative abundance was higher with stream water and non-flush latrines, and increased rainfall was associated with lower odds of CST-I.

That is the most direct evidence currently available that water source and sanitation type track with vaginal microbiome composition — and it is a Kenyan cohort, not an Indian one. Indian data on CST distribution is thin generally. Most foundational work was done in North American and European populations, CST distribution is known to vary by population, and what Indian women’s communities actually look like at scale is an open question.

What the classification is useful for, and what it is not

Useful for: reading the literature, since studies routinely report findings by CST; understanding that “has lactobacilli” is an incomplete description, because which lactobacillus matters; interpreting risk at a population level.

Not useful for: self-diagnosis. There is no reliable consumer test that returns a CST, and a home pH strip tells you about pH rather than composition. More fundamentally, a CST result would not change what anyone should do, because the interventions that reliably shift community state are limited and the evidence for most consumer products is weak.

The classification is a research tool that has escaped into marketing. It is worth understanding because it makes the science readable. It is not worth paying for a test to learn your own.


This article is general information and is not medical advice.

Sources

  • Ravel J, et al. Proc Natl Acad Sci USA. 2011;108(Suppl 1):4680-4687. doi:10.1073/pnas.1002611107
  • O’Hanlon DE, Moench TR, Cone RA. BMC Infect Dis. 2011;11:200. doi:10.1186/1471-2334-11-200
  • Tachedjian G, O’Hanlon DE, Ravel J. Microbiome. 2018;6:29. doi:10.1186/s40168-018-0418-3
  • Witkin SS, et al. mBio. 2013;4(4):e00460-13. doi:10.1128/mBio.00460-13
  • Czapar AE, et al. BMC Infect Dis. 2024;24(1):1419. doi:10.1186/s12879-024-10313-3