Insulin resistance is the mechanism invoked in almost every explanation of PCOS and explained properly in almost none of them. It usually arrives as a diagram with arrows, immediately followed by something to buy.
Here it is without either.
What insulin does, and what “resistance” means
Insulin is a signal. After you eat, glucose enters the blood, and insulin is released to tell cells to take it up. That is the whole job: it is a message, not a substance being used up.
Insulin resistance means the cells respond less to that message. The system’s compensation is to send more of it — so for a long time the glucose reading stays normal because insulin levels are higher. That compensation is the important part, and it is why the story is easy to miss on a standard test.
Two consequences follow, and they explain most of what gets said about this:
Glucose can look fine for years. A normal fasting glucose does not mean insulin is normal; it may mean insulin is working harder to produce that result. This is why screening is repeated over time rather than done once.
Insulin is not only about glucose. It has effects elsewhere, including on ovarian androgen production, which is the connection to the rest of PCOS. Higher circulating insulin is associated with higher androgen output, and androgens are one of the three diagnostic criteria.
That is the mechanism in full. Everything beyond it in most explanations is decoration.
It is not a diagnostic criterion
Worth stating plainly, because a lot of writing treats PCOS and insulin resistance as the same thing.
They are not. Insulin resistance is not one of the criteria, not everyone with PCOS has it, and having it does not give you PCOS. What is true is that it is common enough, and associated with PCOS independently of weight, that metabolic assessment is part of ongoing care rather than an optional extra.
The independence from weight is the part that gets lost. It is why screening is recommended across the weight range, and why being slim does not mean the metabolic side has been ruled out. It is also the fact that most complicates the weight-centred account of PCOS.
What the tests actually measure
This is where the panel confuses people, because the routine tests measure glycaemic status, not insulin resistance itself.
HbA1c reflects average blood glucose over roughly the previous two to three months. It needs no fasting and does not care what day of your cycle it is, which is why it appears so often. Being an average, it is stable — useful for trends, slow to move.
Fasting glucose is a single moment, first thing, before eating. Cheap, quick, and the least sensitive of the three for picking things up early.
An oral glucose tolerance test measures how you handle a glucose load rather than where you sit at rest. It is the most sensitive of the three and the most unpleasant — the one where you drink a glucose solution and get bled again afterwards. It picks up changes the other two miss, which is why it is sometimes asked for despite a normal fasting result.
Fasting insulin measures insulin directly and is not recommended for routine clinical use. The assays are not standardised well enough between laboratories for the number to mean the same thing everywhere. If you have seen HOMA-IR mentioned online, it is calculated from fasting insulin and glucose and inherits the same problem — it is a research tool, not a routine one.
So: nothing on a normal panel measures insulin resistance directly. What is being tracked is the consequence, over time. Which is the argument for keeping your own results with their dates and reference ranges — three HbA1c values across two years say something no single one can.
What the guideline says about doing something
Two points, and they are less prescriptive than most of what you will read.
Lifestyle is recommended, with no specific regimen preferred. The guideline is explicit that no one diet or physical activity approach has demonstrated benefit over others in PCOS, and it warns against unduly restrictive or nutritionally unbalanced diets. There is no evidence-endorsed PCOS diet, and any product claiming one is claiming something the guideline declines to.
Metformin is positioned for metabolic features, specifically recommended in adults with a BMI at or above 25 primarily for that purpose. Whether it applies to you is a prescribing conversation — what each of the common medications is aimed at is the context for having it, and this page names no doses.
Which leaves this article in a slightly unsatisfying place, honestly: it can explain the mechanism, and it cannot tell you what to eat, because the guideline does not.
Why “insulin resistance” gets attached to everything
Because it is a real mechanism with genuine explanatory power, which makes it an excellent thing to attach a product to.
Once a mechanism is established in the audience’s mind, any supplement, protocol or eating plan can be positioned as addressing it, and the mechanism does the persuading. The claim never has to be tested — it inherits the credibility of the biology.
The tell is straightforward: does the claim come with a measurement? Real management of this has numbers attached, and they repeat over years. A claim with no measurement and no timescale is not a treatment, whatever the arrows in its diagram are doing.
What is worth tracking
The measurements are the point, and they are slow. HbA1c moves over months. A lipid panel is annual. The guideline treats glycaemic assessment as something repeated over years rather than settled once.
So the useful unit here is not this week. It is having the last four results with their dates, in one place, at the appointment — because a value inside the reference range and rising is a conversation, and the same value seen once is not.