Hormone Science

Somewhere between her late thirties and late forties, a woman opens her cycle tracking app, logs her period start, and waits for the prediction. The app predicts her ovulation around day 14, when the Luteal phase begins, and her next period will start in 28 days.

Except that’s not what her body will do.

The next period comes on day 19. Then day 47. Then twice in one month, neither of them resembling what the app called a cycle. Hot flashes arrive on what should have been a calm follicular week. A mood crash lands on a day the app is color-coded green. She’s exhausted in ways that her app cannot chart.

This is not a failure of the app. It’s a category error. Cycle tracking is a tool designed for a system that operates predictably. Perimenopause is defined, at the physiological level, by the collapse of that predictability. Applying cycle-tracking tools to perimenopause is like using a tide chart to navigate a river in flood.

What cycle tracking actually tracks

Menstrual cycle tracking apps were built on a solid premise: the reproductive cycle follows a predictable hormonal arc. Estrogen rises in the follicular phase, peaks to trigger an LH surge, ovulation occurs, progesterone rises in the luteal phase, and if no pregnancy occurs, both hormones fall and the cycle resets. The whole system is elegantly recursive. The apps learned to predict small variations within a tight range. For a woman in her reproductive prime, with regular ovulatory cycles, this works reasonably well.

The underlying assumption is ovulation. The apps predict forward from the last period, estimate ovulation, and extrapolate the luteal phase and next bleed. Everything downstream of that logic depends on the cycle being ovulatory and roughly consistent in length.

Perimenopause breaks both assumptions.

What actually happens hormonally in perimenopause

The textbook description of perimenopause is “declining estrogen.” This is partially true but can be misleading.

Estrogen does not decline steadily throughout perimenopause. In early perimenopause, peak estrogen levels are often higher than they’ve been before, so the drops are steeper. Think rollercoaster, not a gentle slope.

Simultaneously, and less discussed, progesterone begins its own decline through a different mechanism. Progesterone levels are lower in perimenopause than in premenopause because progesterone production is triggered by ovulation. No ovulation (aka anovulatory cycles), no progesterone. The balance between estrogen and progesterone shifts toward estrogen even when estrogen hasn’t fallen in absolute terms.

Anovulatory cycles occur at increased frequency in the last 30 months before the final menstrual period, but they begin appearing much earlier, interspersed unpredictably with ovulatory cycles. A woman might ovulate in January, not in February, ovulate again in March. She has no way of knowing which month was which without hormone testing. Her cycle tracker, working from period dates alone, cannot handle this unpredictability.

In the late menopausal transition, more anovulatory cycles appear, characterized by low progesterone and erratic estrogen. Fluctuating FSH causes periods of both hypo- and hyper-estrogenism, meaning most women in the menopausal transition are exposed to erratic hormonal flux in both directions.

The result is not simply “low hormones.” It’s two hormones behaving unpredictably, declining at different rates, through different mechanisms, with different downstream effects on the brain and body.

It’s not just low estrogen. It’s two partially independent mechanisms.

This is where science has moved in ways the public conversation hasn’t fully caught up with yet. The symptom picture in perimenopause is driven by two overlapping but distinct processes.

The first is estrogen variability.

It was previously thought that estrogen deficiency caused hot flashes and night sweats. However, vasomotor symptoms also present in women in very early perimenopause who still have regular menstrual cycles and whose estrogen levels have not decreased. The evidence now suggests that downward swings of estradiol cause the dramatic neuroendocrine release, with elevated central norepinephrine levels leading to thermoneutral zone narrowing and vasomotor symptoms. High brain estrogen exposure followed by estrogen withdrawal, rather than low estrogen per se, is the underlying cause.

This reframes the hot flash entirely. It is not a symptom of estrogen deficiency. It is a symptom of estrogen withdrawal from a previously higher state. Two women with identical estradiol readings on a Tuesday can have completely different symptom experiences, because what drove the symptoms in one of them wasn’t the level she arrived at but the speed and magnitude of the drop that got her there.

The same mechanism operates for mood. Dr. Hadine Joffe’s research linked an increase in depression symptoms at perimenopause specifically with fluctuations of progesterone and estradiol, not simply low levels of either. The rate of change is the signal. The absolute number is not.

The second is progesterone withdrawal, which operates through different pathways.

Progesterone is not simply estrogen’s counterbalance. It has its own direct routes to symptoms that are independent of what estrogen is doing.

Progesterone plays a key role in stabilizing the brain’s thermoregulatory system. When progesterone drops, the brain becomes hypersensitive to small temperature changes, which can trigger hot flashes even when estrogen is relatively normal. Progesterone also has well-characterized GABAergic sedating effects that directly support sleep architecture, separate from whether hot flashes are disrupting sleep. And it modulates the stress response through the HPA axis in ways that estrogen does not fully substitute for.

The hypothalamic-pituitary-gonadal axis (HPG) and the hypothalamic-pituitary-adrenal (HPA) axis are intertwined such that changes in estrogen and progesterone influence cortisol directly. As progesterone disappears from anovulatory cycles, this buffering effect on the stress axis disappears with it. The woman who finds herself disproportionately reactive to stressors that wouldn’t have touched her five years ago is not imagining it, and it is not purely estrogen-driven.

The ratio matters, but not as a simple number.

When progesterone falls while estrogen remains relatively elevated or fluctuating, the physiological context is different from when both are low simultaneously. Progesterone counterbalanced varying estrogen levels in the pre-menopausal years, providing a stabilizing buffer on thermoregulation, sleep, and mood. Its decline in perimenopause removes that buffer.

However, be aware that the use of the progesterone-to-estradiol ratio as a clinical diagnostic tool has not been extensively validated in RCTs. Much of the ratio research has centered on IVF outcomes rather than perimenopausal symptom prediction, and results in that context are mixed. Many clinicians find the ratio concept useful as a framework. Definitive trial evidence for it as a precision diagnostic is not yet there.

What is clear is that neither hormone tells the full story alone. A cycle tracker that records period dates captures neither.

What would actually help

If cycle day is an unreliable proxy for hormonal state in perimenopause, the question is what better proxies exist.

Symptom patterns over time, not cycle position. Track symptoms longitudinally without anchoring them to cycle day. Patterns begin to emerge that are more meaningful than “day 18.” You might notice that your worst hot flash clusters come after several days of elevated energy, reflecting a post-estrogen-peak drop. You might notice that mood crashes correlate with sleep disruption that preceded them by 48 hours. These temporal cause-and-effect patterns are more informative than where you are in a cycle whose phases may no longer exist as labeled.

Basal body temperature, used differently. BBT is a significant metric in perimenopause. Temperature data can reveal whether ovulation actually occurred, which matters enormously because it tells you whether progesterone rose at all in a given cycle. A cycle with a sustained temperature shift after mid-cycle was ovulatory. One without wasn’t. This is a real and useful distinction that calendar-based tracking cannot make..

Lifestyle inputs as leading indicators. Sleep quality, dietary patterns, stress load, and alcohol intake all interact with estrogen metabolism, progesterone’s buffering effects, and cortisol. They’re modifiable, they’re tractable, and they precede symptoms rather than following them. A woman who notices that two consecutive nights of disrupted sleep reliably precede a hot flash cluster has more actionable information than one who knows she’s on day 22 of an irregular cycle.

The symptom-input pair. The most useful data in perimenopause is the relationship between what you did or experienced and what followed. Not “I had a hot flash on day 18” but “I had three glasses of wine Wednesday, poor sleep Thursday, and a significant hot flash cluster Friday.” The signal lives in the cause-and-effect pair, not in the calendar position of the effect. What your body is doing to you is downstream of what was done to it, often by 24 to 72 hours. Tracking the downstream event alone, without the upstream input, is half the data at best.

The limit of all of this

None of these alternatives, BBT, symptom logging, input tracking, are as clean or predictive as a well-behaved 28-day cycle. Perimenopause is, by definition, a period of reduced predictability. The goal isn’t to find a tool that restores the certainty of regular cycles. That certainty isn’t coming back until the transition is complete.

The more achievable goal is pattern recognition in a noisy system. Not “your period will come in 12 days” but “your worst symptom windows tend to follow these conditions, and here’s what you’ve found that shortens them.”

That’s a different relationship with your body than cycle tracking offers. It requires more patience, more data over time, and a willingness to treat your own physiology as a hypothesis to test rather than a schedule to predict. But it’s grounded in what’s actually happening hormonally, in both hormones, and the relationship between them.

That’s what Waves Women is building.

Waves does not provide medical diagnosis or treatment. Always consult your healthcare provider for personalized medical advice.

Sources

Santoro N, et al. “Characterization of Reproductive Hormonal Dynamics in the Perimenopause.” Journal of Clinical Endocrinology and Metabolism 81, no. 4 (1996).

Joffe H, et al. “Impact of Estradiol Variability and Progesterone on Mood in Perimenopausal Women.” Journal of Clinical Endocrinology and Metabolism 105 (2020): e642-e650.

Burger HG, et al. “Cycle and Hormone Changes During Perimenopause: The Key Role of Ovarian Function.” Menopause(2008). PubMed: 18574431.

Prior JC. “Paradigm Shift in Pathophysiology of Vasomotor Symptoms: Effects of Estradiol Withdrawal and Progesterone Therapy.” Maturitas (2020).

Yin J, et al. “Perimenopausal State Oestradiol to Progesterone Imbalance Drives Alzheimer’s Risk via ERRα Dysregulation.” Nature Communications (2025).

Rubinow DR, et al. “The Menopause Transition: Estrogen Variability, Stress Reactivity and Mood.” Clinical Trials NCT03003949.

 

Waves Women does not provide medical diagnosis or treatment. Always consult your healthcare provider for personalized medical advice.