Testosterone in women: too high, too low, measured properly

You are sitting there with a printout in your hand. Two lines are flagged. Testosterone, an arrow pointing up. Below it a number that means nothing to you.

Maybe the trigger was hair growth on your chin. Maybe acne that came back at 32. Maybe a menstrual cycle that does whatever it likes. Or the opposite: everything is normal, only desire is gone, and somewhere on the internet it said that could be down to testosterone.

Many women know this pattern. A single value, an arrow, and overnight it becomes an explanation for everything.

So I do not start with the symptoms. I start with the measurement. Because the value most women receive comes from an assay that was built for male levels. In the low female range it can become unreliable. That is not a footnote. That is the core of this text.

Testosterone is a female hormone, only in a different amount

Almost every woman I tell that she has testosterone raises her eyebrows for a moment. Testosterone counts as the male hormone. Beard, muscles, aggression, cars.

That is a misunderstanding that comes from marketing language and not from physiology. Steroid hormones are not gender badges. They are molecules that arise in both bodies from the same construction kit. The difference lies in the amount and the distribution, not in the blueprint.

In your body testosterone is made in three places. In the ovary, in the adrenal cortex and outside the glands, namely by conversion of precursors. These precursors are called DHEA, DHEA sulfate and androstenedione. They are processed further in fat tissue, skin and liver.

Imagine a kitchen with three work stations. Two of them are glands, the third is the whole rest of the body, assembling the finished hormone out of half-finished products. That is also why a single value in the blood says little about where it came from.

What the body uses it for

Androgen receptors do not sit in one place only. They are found in the brain, in muscle, in bone, in breast tissue and in fat tissue, among other places. That is the anatomical reason why androgens show up in so many systems in women.

The areas named most often in the literature are sexual desire, mood and drive, and the metabolism of muscle and bone. Then there is the role as a precursor: part of the testosterone is converted into estradiol via the enzyme aromatase. Without androgens there would therefore be less estrogen.

And here the first bit of sloppiness begins, the one you find everywhere online. From the fact that receptors sit in muscle and bone it does not follow that giving testosterone improves anything there. One is anatomy, the other would be a study result. How that turned out is further down in the treatment section, and it is more sober than the anatomy would suggest.

A word on orders of magnitude, without number games. Male values lie many times higher than female values. That is exactly why the measurement problem this article is about arises. If an assay is designed to detect large amounts, it becomes imprecise at the lower end. A kitchen measuring jug for litres is a poor tool for teaspoons.

If you are looking for the big picture of how the female hormones interact as a whole, that is in the overview article of this cluster: hormonal imbalance in women. This text here deliberately zooms in on a single lab value and on what it can and cannot do.

And now you know why the question about your testosterone is not answered with yes or no, but with a question back.

Two curves that have nothing to do with each other

When the word menopause comes up in the consulting room, most people think of a crash. A curve that breaks off. That is roughly what it looks like for estrogen.

For testosterone it looks different. There are two completely different curves, and they get confused all the time. One is short and runs over a month. The other is long and runs over decades.

And now you know why the question about the cycle day is not a formality but decides whether your report can be read at all.

The core problem: how testosterone is measured in women

Now comes the section this article was written for.

Imagine you have the same blood sample measured in two laboratories and get back two clearly different numbers. No mistake, no oversight. Just two different assays.

That is exactly the normal case in women, and it has been described in the specialist literature for more than twenty years. It just rarely appears on a lab report.

Now you might think: 2003 was a long time ago, surely that has been sorted out by now. So here is a second paper, from 2021, on a present-day instrument.

Why this is not an accusation against laboratories

It quickly sounds like an indictment. It is not.

Immunoassays are fast, inexpensive and widely available. They were developed for the male range and there they do what they were built for. The problem only arises when the same tool is carried into a range it was never designed for.

A Greek review by Kanakis and colleagues puts it soberly: most available immunoassays lack the necessary accuracy at the lower end of the male range and across the entire range in women. As a consequence, mild androgen excess in women often goes undetected. The error therefore also runs in this direction: it is not only that too much is found, something is also missed.

Mass spectrometry counts as the reference method. That it is not used across the board is due, according to the same review, to complexity and cost, not to a lack of professional conviction. That is a question of health service provision, not a matter of dispute.

The Endocrine Society position statement from 2007 has a sentence on this that I quote often. The quality of a laboratory should be judged above all by whether it measures samples of known concentration correctly, and not solely by whether it delivers the same result as other laboratories using the same method. Agreement is not the same as accuracy.

SHBG: the value without which the number is halved

The second reason a single testosterone value says little is called SHBG.

SHBG is a transport protein from the liver. Imagine taxis. A hormone sitting in a taxi cannot get out and cannot do anything. Only the free fraction has access to tissue.

When the number of taxis rises, the free fraction falls, although the ovary is producing exactly as before. When the number of taxis falls, the free fraction rises, again without anything changing in production. That is why a normal total testosterone with a very low SHBG can look clinically quite different from the same number with a high SHBG.

What raises SHBG: a combined pill, estrogens, an overactive thyroid, pregnancy. What lowers SHBG: high insulin, excess weight, an underactive thyroid, glucocorticoids. The connection to glucose metabolism is well described. Low SHBG counts as an early sign of insulin resistance. The common explanation for this is that raised insulin can suppress SHBG production in the liver. The review on this does, however, record that the exact mechanism is not yet clear (PMID: 21178921). More on this in insulin resistance and female hormones.

The free androgen index and its limit

The free androgen index is popular because it is cheap. You only need two values and divide one by the other. Total testosterone divided by SHBG, times one hundred.

As an approximation it is usable. As a number it is a construction. It has no unit and describes no concentration in the body.

And now a conclusion that I label as such. When SHBG is strongly shifted, it sits in the denominator of this calculation and dominates the result. Under a combined pill, SHBG rises on average by around 99 nmol/l (PMID: 24082040). An index whose denominator wanders like that becomes hard to interpret. I did not find a study that quantifies this for pill users. This is a derivation from two findings, not evidence.

On the general methodology of testosterone measurement there is already a separate text in the cluster: understanding testosterone values. And if you are wondering which hormone values make sense in a woman at all and when, that is in testing hormones in women.

And now you know why, with a conspicuous testosterone value, I ask about the assay first and not about the symptoms.

Too much: the causes, ordered by frequency

Suppose the value really is raised after a clean repeat measurement. Then the actual work begins, and it is a search for the source.

There is a short list, and it is sorted by frequency. This order matters, because online the rare things get the most space and the common ones the least.

1. PCOS, by far the most common explanation

Polycystic ovary syndrome is the most common reason for raised androgens in women of childbearing age. It is a medical diagnosis with defined criteria and belongs in a proper work-up, not in self-diagnosis.

Because there is a separate article on it, this stays at a link: understanding PCOS. Notably, the international guideline from 2023, backed by 39 organisations from 71 countries, describes its own evidence base as low to moderate (PMID: 37580861). An invitation to phrase things with humility. [Guideline]

2. Non-classic congenital adrenal hyperplasia, the most commonly missed alternative

There is a variant of congenital adrenal hyperplasia that only shows up in adolescence or adulthood. It is called non-classic CAH and usually arises from a partial deficiency of the enzyme 21-hydroxylase.

Clinically it looks like PCOS. Hair growth, acne, cycle disturbances. That is why it is easily missed. It is looked for via basal 17-hydroxyprogesterone in the blood, measured in the morning and in the follicular phase.

That this form exists and that it is clinically hard to distinguish from PCOS is also in the Endocrine Society guideline on congenital adrenal hyperplasia (PMID: 30272171). I deliberately do not name a threshold value here, because interpretation depends on the laboratory, the timing and the assay. [Guideline]

3. Insulin resistance, the quiet amplifier

Here something interesting happens: the ovary does not necessarily produce more, but more arrives.

4. Medications and substances

Some medications can influence androgen values or androgen signs. These include certain progestins, glucocorticoids via the detour of SHBG, anabolic substances from the sports and fitness world and valproate. Testosterone-containing gels from the environment of another person can also be transferred.

Valproate needs its context. Valproic acid (valproate) is a prescription medicine for epilepsy and for bipolar disorder. It can be associated with cycle disturbances and raised androgens in women. In women of childbearing age it is subject in Germany to special regulatory conditions, among them a pregnancy prevention programme, because it can seriously harm an unborn child. That is exactly why it is a medicine to be discussed with the prescribing practice, and never one to be reduced or stopped on your own. Stopping it abruptly can be dangerous.

The important sentence on this: an existing medication is not changed, reduced or stopped on your own initiative because of it. If a preparation comes into question as a contributing cause, that is a reason for a conversation with the prescribing practice, not for going it alone.

5. Other endocrine causes

Rarer, but kept in mind during the work-up: Cushing syndrome with increased cortisol production, a prolactinoma of the pituitary, a thyroid disorder that plays a part via SHBG. On the thyroid there is a separate text in the cluster: the thyroid and female hormones.

6. Androgen-producing tumours, rare and therefore the reason for the box above

Tumours of the ovary or the adrenal gland that produce androgens are rare. They are, however, the reason why the pace of change matters so much.

The case report from Hangzhou shows why laboratory logic alone is not enough. Testosterone was clearly raised, while DHEA sulfate, androstenedione, 17-hydroxyprogesterone and cortisol were within the normal range. The source was nonetheless the adrenal gland. The authors' key statement: the origin of a hyperandrogenaemia cannot be read off these values alone, and with virilisation both ovaries and adrenal glands belong in the examination (PMID: 30674304). [Case Series, n=1]

And not every patch of hair needs a laboratory

An Endocrine Society guideline on hirsutism says something that surprises many people. In women with an abnormal hirsutism score, androgen measurement is suggested. In women with a regular menstrual cycle and only locally unwanted hair growth, that is, without an abnormal score, androgen measurement is explicitly suggested against (PMID: 29522147). [Guideline]

Measuring more does not automatically mean knowing more. Sometimes it only means more numbers that have to be placed in context.

And now you know why, with a raised value, the next question is not how to lower it but where it comes from.

Too little: where testosterone in women really does fall

The other direction is even messier online. There, a low value quickly becomes a diagnosis, and the diagnosis becomes a product.

So here is the honest sentence right at the start of this section: there is no recognised clinical entity of androgen deficiency in women.

There are nonetheless situations in which androgen values in women really do fall. The most important ones are well described.

After removal of both ovaries

Here a source drops away, and that is measurable. In the Australian cross-sectional paper, total testosterone and free testosterone in women aged 55 and over after removal of both ovaries were significantly lower than in the reference group of the same age band (PMID: 15827095). This is one of the few points at which the connection between situation and value is clear.

Under combined hormonal contraception

This is the most common explicable reason for low values in young women, and the data on it are unusually solid.

Now comes the part that regularly gets overstated online: the time after stopping.

Important and explicit: no recommendation to stop or switch a contraceptive follows from this section. That is a medical decision in which contraceptive reliability, cycle complaints, skin findings, migraine, thrombosis risk and your life planning are weighed against each other. What happens physically in the time after stopping is described in detail in stopping the pill: what happens afterwards.

Under glucocorticoids and in pituitary disorders

Cortisone therapy can dampen androgen production in the adrenal gland. In disorders of the pituitary, the control signals from above are missing. In the study by Miller, women with severe androgen deficiency due to hypopituitarism formed a separate study group, precisely because values there are particularly low.

Here too the guideline sentence applies: routine prescription of testosterone for low androgen levels due to hypopituitarism, adrenal insufficiency, surgical menopause or glucocorticoid administration is explicitly recommended against (PMID: 25279570). [Guideline]

And now you know why a low value in you is a question and not a diagnosis.

Reading symptoms without attributing them to the value

There is a moment in the conversation that I experience almost every time. A woman points at a number and says: so that is why.

I understand that moment well. A value gives something to hold on to. It turns a diffuse feeling into something visible. It takes away the suspicion that you are imagining it.

And still the sentence is usually not true. Not because the complaints are not real, but because the connection between number and complaint is loose with androgens.

A sentence that must not be missing from any text about the cycle and hormones: if your period is absent, what comes first is not the hormone report but the pregnancy test. Even when you are sure that it cannot be that. Only after that does thinking about androgens, thyroid, weight, stress and the pituitary begin.

The word hirsutism means something specific here. Not every hair that bothers you is hirsutism. What is meant is a pattern of coarse, dark hair in places where it typically grows in men. There are scores for this, and the guideline makes androgen measurement depend on exactly that.

With skin and hair, something else plays a part that does not show up in the blood. How sensitively a hair follicle responds to DHT appears to be genetically co-determined as well. How large that share is has not been cleanly quantified. Two women with the same value can therefore look different. More on this: DHT and hair loss. On hormonal acne likewise: hormonal acne from the inside.

What else the body throws into the same pot

Fatigue, lack of drive and little desire are symptoms with a great many possible sources. Lack of sleep, chronic stress, iron deficiency, thyroid function, depression, pain, medications, relationship dynamics, stage of life. The list is long and it is not comfortable, because it offers no quick solution.

Chronic stress deserves its own mention here, because it reaches into this field along several routes. Cortisol and the control axis in the brain can influence cycle regulation, and glucocorticoids can dampen androgen production in the adrenal gland. On the interplay of cortisol and female hormones there is a separate article: cortisol, stress and female hormones.

And now you know why I rarely look at a report on its own, but always place beside it what was happening in your life at the same time.

Treatment: what is established, what is not, and what applies in Germany

This section gets uncomfortable, and I am writing it the way the data stand anyway. Testosterone for women is a field with a lot of marketing and little licensing. If someone offers you a preparation as a clearly established solution, it is worth asking what that rests on.

In 2019, eleven international professional societies sat down together and formulated a joint position. Among them the International Menopause Society, the Endocrine Society, the European and the North American menopause societies, the RCOG and the International Society of Endocrinology.

What was not found in the same meta-analysis

No effects on body composition. None on musculoskeletal variables. None on cognitive measures.

And now the fair part that has to be said alongside: the author group itself notes that only few women contributed data for exactly these endpoints. That is a not-established, not a refuted. The difference matters. Turning it into proof that testosterone does nothing for muscle and bone goes just as far beyond the data as someone claiming the opposite.

There is, incidentally, an older review from 2015 whose first author is also the senior author of this meta-analysis, and it sounded more optimistic on cognition and musculoskeletal health. I follow the newer and methodologically stricter paper here. That two good sources sound different is part of science and is not tidied away here.

How things stand in Germany

Two things belong to this topic and often get lost in the discussion.

First: no testosterone preparation is licensed for women in Germany. Every use in a woman is therefore off-label, with everything that implies for information, documentation and questions of cost.

Second: the German S3 guideline on peri- and postmenopause formulates a may-option, not a recommendation. It says literally that peri- and postmenopausal women with loss of libido may be offered testosterone therapy, after psychosexual exploration and if hormone replacement therapy has not been sufficient, and that the off-label use should be explained. Evidence level 1b, recommendation grade 0. [Guideline]

In the accompanying text of the guideline there is another sentence that fits the rest of this article: low estrogen and testosterone levels can go along with loss of libido, but there is no strict connection between sex hormone levels and libido, because the reasons for loss of libido are complex (PMID: 32661753).

And if the value is too high

Here too: treatment follows the cause and not the number.

For hirsutism, the Endocrine Society guideline suggests a combined estrogen-progestin preparation for the majority of women, supplemented if needed after six months by an antiandrogen. Antiandrogen monotherapy without reliable contraception is explicitly recommended against, because antiandrogens pose a risk in pregnancy (PMID: 29522147). That is exactly why these are not substances for self-experimentation. [Guideline]

The context belongs with it. Both are prescription medicines and both belong in medical hands. Combined estrogen-progestin preparations can raise the risk of thrombosis and embolism, and this risk can be higher with smoking, with excess weight, with migraine with aura and with a corresponding family history. Antiandrogens can harm an unborn child, which is why the guideline presumes reliable contraception. Whether either of them suits you is decided in the examination and in conversation, not from a lab value. Doses are deliberately not given here.

And a point that explicitly belongs here: if you want a child, that is a topic of its own with its own urgency. Raised androgens and an irregular menstrual cycle can make planning harder, and time is a real factor in this. Nothing in this article is a reason to postpone a gynaecological or reproductive medicine work-up. I cannot hold out any outcome to you, and nobody can. What I can tell you: this work-up belongs early and in parallel, not only once everything else has been tried.

What you can influence independently of that lies with metabolism rather than with the hormone itself. Sleep, movement, muscle mass, the sugar load of meals and stress burden can act on the insulin and SHBG axis. That is mechanistically plausible and anchored in guidelines as a basis in PCOS. It does not replace a work-up or a medically prescribed therapy.

And now you know why I am slower on this topic than you might expect from an integrative practice.

Why I also look at environmental factors

Now the part some patients come to me for and others become sceptical about. Both are justified.

When a woman's androgens run outside the usual range with no recognisable reason, I additionally look at insulin, stress, the thyroid and at the surroundings in which this body lives. Not instead of the gynaecological work-up. After it, and on top of it.

And I tell you in advance how strong these data are, so that you can place them yourself. Because this is where the internet exaggerates most, in both directions.

What is well established: the mechanism

Exactly this double statement is the reason routine care is cautious. And it comes from the same professional society that takes the field seriously. So this is not a fight between camps but an honest description of a difficult evidence base.

What human observational data show

Where hormonally active substances turn up in everyday life is described in xenoestrogens in everyday life. Zearalenone, a mould toxin with an estrogen-like structure, has its own text: zearalenone and hormones. For clarity: it is a mycoestrogen, not an androgen analogue. Whether measuring a metal exposure makes sense is covered in measuring heavy metals.

A word about gynaecology

Sometimes patients tell me their gynaecologist has never mentioned xenoestrogens. That quickly becomes an accusation, and I do not share it.

A gynaecology consultation has a different task and a different time budget. Cancer screening, contraception, pregnancy, bleeding disorders, endometriosis, menopause. That is a full specialty with clear priorities, and environmental medicine is simply not a focus in the training of that specialty. That is an observation about structures, not about people.

The gynaecological work-up remains the foundation. Ultrasound, cycle history, physical examination, the differential diagnoses that have been taught cleanly there for decades. Nothing in this article replaces that, and nothing in it is a reason to postpone a recommended examination or an indicated operation.

What a functional and environmental medicine view can add is an additional direction of questioning. Not only: which diagnosis fits this pattern. But also: what in the surroundings of this body could be co-writing the pattern. This second question comes on top, not in place of the first.

And now you know why I ask these questions and why at the same time I attach no promises to them.

Where this topic connects to the rest of your body

An androgen value never stands alone. It hangs on glucose metabolism, on the thyroid, on the stage of life, on contraception and on the question of which symptom actually brought you here. Here are the texts that lead onward from this page.

Frequently asked questions about testosterone in women

The questions I am asked most often on this topic, with short answers and the sources behind them.