Why we menstruate: the evolutionary logic of a costly system
By Maria Teresa, CEO & co‑founder
This document is part of the Scientific White Paper Reviews series, which explores the biological foundations of uterine function, disease, and therapeutic opportunity.
Menstruation is a rare and expensive biological design. That is exactly why it’s worth understanding what advantage it provides, what it evolved to solve, and what happens when it goes wrong.
Every month, the human uterus builds a complex, highly vascularised tissue in anticipation of pregnancy. Every month, if implantation does not occur, that tissue is actively dismantled and expelled.
Evolution rarely favours processes that involve repeated tissue construction, inflammation, blood loss, and metabolic cost unless they are tightly coupled to a compensating selective advantage. And yet menstruation persists.
Out of more than 5,500 mammalian species, only ~1.5% menstruate: mainly some primates, a few bat species, the elephant shrew, and one known rodent (the spiny mouse). In most mammals, the endometrium is simply reabsorbed if pregnancy does not occur.
Therefore, menstruation is not a universal feature of mammalian reproduction. It is an exception. And a costly one.
Understanding why this exception evolved requires looking beyond menstruation itself and toward the underlying uterine biology that precedes it.
Once menstruation is reframed as the visible outcome of spontaneous decidualisation, things start to make more sense. This is a monthly, hormone-driven transformation of the uterine lining that occurs even when no embryo is present. It appears in only a small number of mammals, and its evolutionary origins are still actively debated. It may have evolved because implantation is invasive and costly, and because embryo quality varies, so the uterus benefits from strong maternal control over what implants and how far invasion proceeds.
I am a physicist by training, and I am drawn to patterns and mechanisms. I specialise in biological systems because nature tends to solve problems in ways that are surprising, and to produce systems that are robust, fault-tolerant and, frankly, quite colourful. In this case, my curiosity began with a simple question: what problem is spontaneous decidualisation solving that makes the high cost of menstruation worth paying?
This article lays out what spontaneous decidualisation is, how it evolved, and why it matters when looking at new drug development in uterine health.
Quick vocabulary (so we can move fast)
Before we get into the biology, here are a few terms you need to be familiar with:
Endometrium: the inner lining of the uterus that changes across the cycle.
Endometrial stroma: the connective-tissue layer of the endometrium (contains fibroblasts, vessels, immune cells, extracellular matrix). This is the part that decidualises.
Decidualisation: stromal fibroblasts differentiating into decidual stromal cells, which ****is a specialised state where they can support implantation but also regulate and limit invasion.
Spontaneous decidualisation: decidualisation that happens cyclically, driven by maternal hormones, even without an embryo.
Placenta: a temporary organ formed during pregnancy that connects fetus and mother, enabling nutrient and oxygen exchange and producing hormones. In humans, fetal trophoblast invades maternal uterine tissue and remodels blood vessels to establish the exchange interface.
Trophoblast: fetal-derived cells that invade the uterus and build much of the placenta.
Hemochorial placentation: a placental type (one of the three main types classified by degree of trophoblast invasion) in which maternal blood comes into direct contact with fetal-derived trophoblast.
With those definitions in place, we can talk about what spontaneous decidualisation actually does and why menstruation is the cost of running this program.
What is spontaneous decidualisation?
Spontaneous decidualisation is the program. Menstruation is what happens when the program is reset.
In most mammals, the uterus only enters a decidual (specialised) state in response to implantation, in other words once an embryo implants in the uterine lining. In humans and a small number of other primates, the uterus runs this transformation in advance, every cycle. That is what “spontaneous” means here: it is driven primarily by cyclical maternal hormonal timing rather than being triggered locally by an embryo.
The process is triggered by the postovulatory rise in progesterone levels and increasing local cyclic AMP production. In simple words, progesterone leads stromal cell specialisation. During this phase, stromal fibroblasts in the endometrial stroma differentiate into decidual stromal cells, a specialised state that supports implantation and helps regulate the early maternal fetal interface.
But spontaneous decidualisation is not only a change in a single cell type. It reorganises the entire local environment. The tissue begins to promote blood vessel growth and remodelling, including changes that prepare uterine arteries for the high flow demands of early pregnancy. The immune landscape shifts toward specialised uterine populations, particularly uterine natural killer cells and macrophages, which help coordinate vascular remodelling and maintain a controlled inflammatory state rather than a general immune suppression. At the same time, the signalling programs of the tissue change, with decidual stromal cells producing factors that regulate trophoblast invasion, immune cell recruitment, and tissue stability.
In other words, spontaneous decidualisation turns the endometrium into a prepared, regulated interface, ready not just to allow implantation, but to control it.
When pregnancy does not occur, progesterone levels fall and the prepared decidualised tissue is dismantled. Because it is highly vascularised and extensively remodelled, a substantial part of it is shed rather than reabsorbed, producing menstruation.
This is where the evolutionary puzzle begins. This process is expensive: it requires running a major tissue program every cycle, even when pregnancy does not happen. What problem is it solving that makes the cost worthwhile?
To answer that, we need to look at the type of pregnancy it supports. Especially, what happens when fetal tissue invades maternal tissue to build a placenta.
Why might spontaneous decidualisation have evolved?
Understanding mechanisms is one thing, understanding evolutionary causes is another. We cannot experiment on ancestral mammals. Evolutionary explanations therefore rely on comparative phylogenetics, lineage distribution (which species show spontaneous decidualisation and menstruation), evolutionary developmental inferences about how uterine cell states evolved, and theoretical models that fit known constraints.
That said, modern evolutionary models converge on a simple idea: spontaneous decidualisation is costly and rare, so it likely evolved because it helped solve a difficult problem in a small number of lineages.
Two hypotheses are most often discussed.
Reason 1: implantation and placentation
Decidualising the endometrium in advance allows the uterus to be ready to regulate invasion of the trophoblast by the time the embryo arrives.
Early pregnancy in humans is very invasive.
Fetal trophoblast cells enter maternal uterine tissue and remodel maternal arteries to establish blood flow to the developing placenta (I highly encourage you to read about the viral origin of this remarkable organ). This is called hemochorial placentation, where maternal blood comes into direct contact with fetal derived trophoblast. This interface is powerful, but it is biologically risky. Too little invasion limits placental blood supply, too much threatens uterine integrity. Maternal immune and inflammatory responses must also be tightly regulated at the implantation site.
Mice also have hemochorial placentation, but they do not menstruate. The reason is that decidualisation in mice is triggered by implantation (the embryo), not spontaneous decidualisation (maternal hormones). This raises a key evolutionary question: why did this system evolve in humans but not in mice?
Reason 2: embryo screening
Decidualising the endometrium in advance also allows the cells to be able to screen for early embryo defects that could compromise pregnancy outcomes.
Humans and higher primates experience unusually high variability in embryo developmental potential. Chromosomal abnormalities are common in preimplantation embryos, and early embryo loss is common. That means the uterus is repeatedly exposed to embryos with very different chances of progressing.
Implantation also represents a major commitment. Once an embryo implants, maternal investment ramps up quickly through more tissue remodelling, vascular adaptation, and immune coordination. If many embryos are developmentally compromised, then a mechanism that reduces investment in those embryos could confer a selective advantage.
This is the basis of the embryo screening hypothesis. Spontaneous decidualisation may have evolved as a form of maternal gatekeeping. By transforming the endometrium in advance, the uterus creates a decidualised tissue state that is not simply receptive, but responsive. These cells can sense embryo-derived signals and adjust the environment accordingly, either supporting implantation or facilitating early rejection.
This model is often described as a “choosy uterus.” In this framing, menstruation is not selected because bleeding is useful. It is the predictable cost of withdrawing a decidualised tissue state when no viable pregnancy is established.
The cost, the strategy, and the open questions
What makes menstruation so interesting is that it is not the purpose, but the visible consequence of spontaneous decidualisation, and, crucially, a non-invasive window into it.
We understand relatively well the cellular transition of stromal fibroblasts into decidual stromal cells, the coordinated changes in immune and vascular environments, and the withdrawal and repair processes that follow. What remains more uncertain is the evolutionary driver. We cannot test selection pressures directly in ancestral mammals, so evolutionary explanations rely on comparative patterns, lineage distribution, and evolutionary developmental inferences.
Within those constraints, the most widely discussed models converge on a consistent idea. Spontaneous decidualisation is costly and rare, so it likely evolved because it provides a meaningful advantage in a narrow set of lineages.
This matters not only for evolutionary biology, but also for disease. Many of the same tissue behaviours that make spontaneous decidualisation effective, such as cyclical inflammation, tissue breakdown and repair, immune repatterning, and tightly regulated tissue invasion and remodelling, are also central themes in disorders of the endometrium.
Endometriosis, for example, seems to occur naturally only in humans and other menstruating primates. That association raises the question: what can evolutionary biology tell us about why this disease exists, and what it reveals about the system behind menstruation?
References (and very good reads!)
Brosens JJ, et al. Cyclic decidualisation of the human endometrium in reproductive health and failure. Endocrine Reviews (2014).
Emera D, Romero R, Wagner GP. The evolution of menstruation: a new model for genetic assimilation. BioEssays (2012).
Wagner GP, Kin K, Muglia L, Pavličev M. Evolution of mammalian pregnancy and the origin of the decidual stromal cell. International Journal of Developmental Biology (2014).
Alvergne A, Högqvist Tabor V. Is Female Health Cyclical? Evolutionary Perspectives on Menstruation. Trends in Ecology & Evolution (2018).
Carter AM. Hemochorial placentation: development, function and adaptations. Biology of Reproduction (2018).
Bellofiore N, Cousins F, Temple-Smith P, Dickinson H, Evans J. A missing piece: the spiny mouse and the puzzle of menstruating species. Journal of Molecular Endocrinology (2018)
