Inositol for PCOS: what the studies show and what they do not

It is half past ten in the evening, you are sitting on the sofa and typing the same thing into the search bar for the third time this week. Inositol PCOS dosage. Forty to one.

The first page says it is the gentle alternative to metformin. The second says that without the right ratio it does nothing at all. The third talks about side effects, the fourth calls it well tolerated. Somewhere in between there is a forum where someone writes that after four weeks everything was the same as before.

Many women know this pattern. You read yourself tired, and at the end you know less than before, because everything contradicts everything.

I believe that is not down to you. It is down to an unusual constellation: many small studies, few large ones, a narrow research field, and an international guideline that is worded far more cautiously than the advice pages. This text therefore gives you the effect sizes with their confidence intervals, quotes the 2023 guideline verbatim and names the places where the same studies are read in opposite directions.

One thing first, and it matters: PCOS is a medical diagnosis. This article replaces neither the work-up nor the care provided by your gynaecological practice. It is meant to help you ask more precise questions at your next appointment. And if you are already taking a medicine, whichever one it is, this text is no reason to change anything about it.

What inositol is, and why it is not a vitamin

Let us start with what is printed on almost every tin and still is not right.

Inositol is often called vitamin B8. That is a historical label from a time when it was assumed the body could not build this substance itself. That is exactly the definition of a vitamin: something that has to come from outside.

But inositol does not have to come from outside. The body builds myo-inositol from glucose, and the kidney does a considerable share of that work. Chemically it is a sugar alcohol with six carbon atoms, a ring with six hydroxyl groups. Depending on how those groups sit in space, nine arrangements are possible. Two of them are clinically interesting.

Imagine a hexagonal table with six chairs around it. All chairs are the same. What differs is whether a chair faces up or down. Turn a single chair around and it is the same table and still a different piece of furniture. That is exactly how myo-inositol and D-chiro-inositol relate to each other.

How much of it actually comes from food

That is a question the internet usually answers softly. There is a surprisingly old and surprisingly thorough piece of work on it.

If at this point you think nutrition is therefore settled, that is exactly where a misunderstanding sits. Through food you do not reach the study amounts. Through food you very much do reach your blood sugar curve, and in PCOS that is in many cases the louder factor. What that looks like in practice, I described in detail in avoiding blood sugar spikes.

And now you know why the question "is this not just a vitamin" is the wrong way in.

Myo and chiro: two sisters with different jobs

When you hold a tin in your hand, there are usually two names on it. Myo-inositol and D-chiro-inositol. Sometimes a ratio is printed next to them, sometimes not.

The two are not two ingredients from two different pots. One is formed from the other.

An enzyme called epimerase turns myo-inositol into D-chiro-inositol. This enzyme works in an insulin dependent way. The more insulin acts on a tissue, the more is turned around. And because insulin sensitivity differs from tissue to tissue, the ratio of the two forms differs from tissue to tissue as well.

That is where it gets interesting. Because in PCOS, muscle and ovary do not seem to behave the same way.

The suspicion that the ovary steps out of line in PCOS is not pure theory. There is a cell study that looks closely at exactly this.

Two sentences to take with you. First: these are not two substances, it is one substance in two arrangements, connected by an insulin dependent enzyme. Second: the ovary is not simply another tissue here. And now you know why the question about the ratio gets asked at all.

The 40 to 1 ratio: where it comes from and what it is not

Hardly any pair of numbers is quoted as often in this field. Forty to one. It appears on tins, in forums, in consultations. It sounds like precision.

Its origin is plain: that is how the two forms are present in blood plasma. Roughly forty parts myo-inositol to one part D-chiro-inositol. That is a measurement from blood, not a measurement taken in the ovary.

From this observation a treatment principle was built. An Italian research group formulated the idea in 2012 that the two forms should be given in exactly this physiological ratio.

Seven years later came the comparison of different ratios against each other. This study is often cited as evidence. It is the reason the number sits so firmly today.

The counterpoint: why a lot of D-chiro-inositol at the ovary could be a problem

Now comes the part that is missing from most advice pages, or only brushed with the word "paradox". In my view it is the most important part of this article. I separate three levels here: what has been observed in humans, what the direct comparison of two preparations showed, and what happened in the animal model. This separation is the most common mistake online.

And now you know why the number on the tin carries less weight than its presence suggests.

The evidence, sorted: what moves and how strongly

I am going to give you numbers now, and with every number I will write the confidence interval next to it. That is unfamiliar to read, and it is worth it anyway.

A confidence interval is the range in which the true effect probably lies. Imagine you throw a ball at a target. The interval is the area where your hits would land if you were allowed a hundred throws. Narrow means: we know fairly precisely where we are. Wide means: small or large, we do not know.

And if the range includes the number 1, it means there might be no difference at all.

First: menstrual cycle and ovulation

Second: androgens

This is the area that means the most to many women, because it touches the skin, the hair and the feeling of being at home in one's body. And it is the area where I word things most carefully.

The 2023 guideline words it soberly at this point: it explicitly calls the clinical benefit for ovulation, hirsutism and weight limited. Anyone expecting a clear change in skin and hair is not carried by the data.

Third: insulin and metabolism

This is where the most stable part of the evidence sits. When I name a reason in the consultation why inositol can play a role at all, this is it.

The founding study of this whole field belongs here, because it appeared in the New England Journal of Medicine in 1999 and set everything else in motion.

Fourth: trying to conceive, and what Cochrane says about it

This section matters to me particularly, because the greatest hope sits here and because the data are thinnest here.

And now you know why it depends on which question you are actually asking.

Why the studies disagree so much

Above you read two analyses that sound friendly and two that sound reserved. Now comes the question that occupied me most: how can that be, when they all draw on the same pool of studies?

The answer has little to do with inositol and a lot to do with the way research and reporting have worked in this field.

The most uncomfortable finding: what was not defined in advance

A letter appeared in BJOG that examined the ovulation analysis quoted above from a methodological angle. It is short, and in my view it is the most honest document in this whole field.

Imagine throwing twenty darts at a wall and then painting the target around the densest cluster. The hit looks impressive, and still you have hit nothing that was a target beforehand. A study registration published in advance is exactly what protects against this. In this field it was missing for a long time.

The conflict of interest question, calmly named

One more thing belongs here, and I word it deliberately without accusation.

A considerable share of the positive work on inositol comes from a closely connected research field, among others around the groups of Unfer, Facchinetti and Nestler as well as the circle of the Experts Group on Inositol. The same names appear as authors of the primary studies, of the reviews and of the conceptual papers.

That does not devalue this work. Specialisation is normal and often the precondition for depth. But it explains why an independent group assesses the same studies differently from one that has been working on them for twenty years.

There is a second point that I would rather not leave out. Several of the key papers quoted here appeared in the same journal, which was removed from two large literature databases in 2023. That does not make the individual results wrong. It is one more reason why I treat them as signals and not as proof.

And now you know why two serious sources can say opposite things about the same substance.

Inositol and metformin: the direct comparison

This is the most common question in my consultation. It usually sounds like this: can I not simply take inositol instead of metformin.

I start with the sentence that stands above everything else. Metformin is not stopped, reduced or replaced on your own. The same applies to the pill, to antiandrogens and to thyroid hormones.

Metformin is available on prescription only. And here is a point that almost never appears in advice texts: in Germany the substance is licensed for type 2 diabetes, not for PCOS. When it is used in PCOS, that is so-called off-label use, meaning use outside the marketing authorisation. This is permitted and anchored in international guidelines. It does mean, though, that your doctor has to inform you separately, that liability sits differently and that statutory health insurers do not automatically cover the cost. That is exactly why this decision belongs in the conversation with the practice that writes your prescription, and not in an article.

And now to the numbers.

Now the other direction, and this time from the guideline itself.

In plain words: for increased hair growth and for central body fat, metformin should be considered over inositol, while metformin has more gastrointestinal side effects than inositol. The guideline says both in one sentence, and both belong quoted together. Fairness requires the qualifier from the same systematic review: the complaints under metformin are described there as typically mild and self-limited (Fitz 2024, PMID 38163998).

The most impressive number in the field, and why it needs careful reading

And now you know why the answer to the most common question in the consultation is neither yes nor no.

What was taken in studies, for how long, and what that is not

Now comes the section you were probably looking for. It starts with a limitation that I mean seriously.

What follows are literature notes: the amounts and time frames from the studies cited. They are not recommendations for you. Which amount fits in an individual case depends on your findings, your medicines, your kidney function, your stage of life and your goals. That belongs in a medical decision and not in an article.

In plain words: specific types, amounts or combinations of inositol cannot currently be recommended in adults and adolescents with PCOS, because quality evidence is lacking. This one sentence explains why I give you no number here that you could adopt. It does not exist.

Tolerability: two statements that stay side by side

These two statements only appear to contradict each other. One says: in the studies carried out, hardly any side effects occurred. The other says: for the fertility context this has not been studied sufficiently. I leave both standing, because smoothing this over would be dishonest.

The guideline also has a point that is often overlooked and that matters in practice.

Put plainly: what is printed on a tin is not as strictly controlled as for a medicine. That is no reason for panic. It is a reason to bring the product to the practice rather than only naming it.

And now you know why there is no dose in this article that you could adopt.

The honest frame: why I look elsewhere first

When a patient tells me she has ordered inositol, I usually do not ask about the dose. I ask what was measured beforehand.

That is not scepticism towards the substance. It is a matter of sequence. First the diagnostics, then the basics, then the supplement. In that order, because otherwise they cover each other up.

By diagnostics I mean: is the diagnosis secure, and has what looks similar been ruled out. That includes the thyroid, prolactin and 17-OH-progesterone, which is used to rule out a mild form of congenital adrenal hyperplasia. Where there are corresponding signs, the question of Cushing syndrome or an androgen producing tumour comes on top. And, very importantly because it is the most frequent mix-up: hypothalamic amenorrhoea, meaning a cycle shut down by too little energy, too much sport or too much stress. There an approach centred on insulin and carbohydrates is the wrong path, and it can do harm. Then come iron, vitamin D and inflammatory markers. The practical way in is described in testing hormones in women, and the map of the diagnosis in understanding PCOS.

What I mean by basics: sleep, movement, the rhythm of blood sugar across the day, and the stress axis. These four sound banal and are not. There is a review on this that gives surprisingly much for placing inositol in context.

How that translates onto the plate is described in nutrition, hormones and blood sugar. And if stress is the louder factor for you rather than blood sugar, the path runs through cortisol and female hormones.

And then the question that comes on top in my practice

Now comes a question that regularly comes on top in my consultation.

When the insulin axis is out of rhythm in a woman, I do not only ask what she eats and how she sleeps. I also ask what she is surrounded by. Home, workplace, moisture damage, materials, cosmetics, packaging. I ask this question in addition because it belongs to the way I work, not because it would be missing elsewhere. The evidence in this field is harder to handle than a laboratory value, and there are good reasons to be reserved about it. The gynaecological work-up remains the foundation. The environmental perspective comes alongside it, not in its place.

What that means concretely and where endocrine disruptors turn up in everyday life at all, I described in xenoestrogens in everyday life. I deliberately do not repeat it here, because this text should stay with inositol.

What I observe clinically, explicitly as an observation

Here ends what studies can carry. What follows is my experience from the consultation, and I label it as what it is: an observation without a study basis.

What I describe are individual cases from my consultation, not an analysis and not a success rate. I have seen courses in which a supplement played a smaller role than expected after longer work on sleep, movement and blood sugar rhythm. And I have seen courses in which an exposure from the living situation came up in conversation. Whether the one is connected with the other I cannot say, and I do not claim it either. I am only describing which questions I therefore ask in addition.

I cannot derive causality from this, and I do not try to. I describe temporal connections in individual cases. That is less than a study would deliver, and it is more than nothing if you turn it into the right question rather than a claim.

If you decide to try it, do it with a baseline, with a period long enough to judge, and with the practice that cares for you in the picture. And if nothing moves after several months, that is information and not a failure. Then the question of what has not yet been looked at is worth asking.

And now you know why on this topic I talk more about sequence than about milligrams.

Frequently asked questions

Where this topic connects to the rest

Inositol does not stand on its own. It hangs on the insulin axis, on the blood sugar rhythm, on the stress axis and on the question of what was measured beforehand. Here are the texts that lead further, depending on your situation.