What are the four types of PCOS?

The four commonly listed online — adrenal, inflammatory, post-pill and insulin-resistant — appear in no professional guideline and are not part of how PCOS is diagnosed or how treatment is chosen. The recognised classification is the four Rotterdam phenotypes, labelled A to D, defined by which combination of the diagnostic criteria someone meets.

What are the PCOS phenotypes A, B, C and D?

They are the four combinations the two-of-three rule allows. A is hyperandrogenism with ovulatory dysfunction and polycystic ovarian morphology, B is hyperandrogenism with ovulatory dysfunction, C is hyperandrogenism with polycystic ovarian morphology, and D is ovulatory dysfunction with polycystic ovarian morphology.

Is there such a thing as adrenal PCOS?

Adrenal androgens are real and are measured — DHEAS is an adrenal androgen the guideline mentions as a secondary consideration where testosterone is not elevated. What does not exist in the guideline is a distinct subtype called adrenal PCOS with its own diagnostic criteria or its own treatment pathway.

Does my PCOS type change my treatment?

Guideline recommendations are organised by what is being treated — menstrual irregularity, hyperandrogenism, metabolic features, ovulation — rather than by a type assigned to the person. Two people given the same medication may have been given it for different indications, which is closer to how the decision is actually made.

Is lean PCOS a separate type?

It is a description rather than a classification. PCOS occurs across the whole weight range and the associated insulin resistance is found independently of body weight, which is why the guideline asks for metabolic screening at any BMI rather than treating lean PCOS as a different condition.

Why do the four types appear everywhere if they are not official?

They are memorable, they imply a specific cause, and a named cause implies a specific fix — which is a useful structure for selling a protocol. The guideline's version is harder to market because it describes combinations of findings rather than root causes.

PCOS Types: Which Ones Are Recognised

Search for PCOS types and you will be told there are four: adrenal, inflammatory, post-pill and insulin-resistant. The framework is everywhere — quizzes, programmes, supplement ranges organised around it.

There is also a recognised classification of PCOS into four groups. It is not that one, it has duller names, and it is the one the diagnosis and the treatment recommendations are actually built on.

Both things can be worth knowing. Only one of them determines what you will be offered at an appointment.

What the guideline recognises

The 2023 international guideline works from the Rotterdam criteria — two of three findings. Four combinations satisfy that rule, and those combinations are the phenotypes, labelled A to D.

Phenotype	Hyperandrogenism	Ovulatory dysfunction	Polycystic ovarian morphology
A	✓	✓	✓
B	✓	✓	—
C	✓	—	✓
D	—	✓	✓

That is the entire system. It is not a theory about cause; it is a record of which findings were present, which is why it falls out of the diagnostic criteria arithmetically rather than being proposed separately.

The groups do differ in ways that matter. Phenotypes A and B — the two that include both hyperandrogenism and ovulatory dysfunction — are associated with greater metabolic risk than C and D. So the classification carries information. It just is not information about what caused anything.

What the other four are, and are not

Adrenal, inflammatory, post-pill and insulin-resistant PCOS appear in no professional guideline. They are not diagnostic categories, there are no agreed criteria for assigning someone to one, and the guideline’s treatment recommendations are not organised around them.

That is a statement about their status, not about whether the underlying things are real. They mostly are:

Adrenal androgens exist and are measured. DHEAS is one, and the guideline mentions it as a secondary consideration where testosterone is not elevated — noting its poorer specificity and its decline with age. What does not exist is a subtype called adrenal PCOS with its own pathway.
Insulin resistance is real and central, which is why it is the subject the guideline screens everyone for regardless of weight. It is a feature that varies in degree across PCOS rather than a type some people have and others do not.
Inflammation is studied. It is not a diagnostic category.

So the problem is not that the words are invented. It is that a dimension has been converted into a box.

Why the box is the problem

A type implies a cause. A cause implies a specific fix. That chain is the appeal of the four-type framework and it is also where it stops matching the evidence.

The guideline’s recommendations are organised by what is being treated, not by who the person is: combined oral contraceptives for menstrual irregularity and hyperandrogenism, metformin primarily for metabolic features, anti-androgens in a limited role for hirsutism, letrozole first-line for ovulation induction. Two people on the same drug may be on it for different indications. That is a different structure from “you are type X, therefore protocol X”.

It also fails in a specific way worth naming. If you are assigned a type and the matching protocol does not work, the framework has no next step — because the premise was that the type explained you. The guideline’s structure does have one: a different indication, or a different agent for the same indication, with an order already published.

What most of the quizzes are doing

Almost every version of the four types ends somewhere you can spend money, and the structure is doing the selling rather than the claims.

A quiz that assigns a type has, in that moment, told you something specific about your own body that your clinician did not. It has also pre-qualified you for whichever product corresponds to that type. Nothing in the sequence required evidence that the types exist — only that the categories feel like they describe you, which four broad categories applied to a varied condition reliably will.

The honest version is less satisfying. PCOS varies enormously between people, the variation is real, and it has not been carved into four causes by anybody who had to prove it.

What to do with this

If you have been assigned a type, it is not wrong to find it useful. “Mine is mostly metabolic” is a reasonable way to describe where your symptoms sit, and it may map onto something true about you. The error is treating it as a diagnosis that implies a treatment.

If you want to know your phenotype, it is already in your notes. It is simply which of the three criteria were recorded — no extra test, no quiz. It is worth knowing mainly because A and B carry more metabolic risk, which is an argument for not skipping screening.

And if a protocol was built on a type and it has not worked, that is not a failure on your part. It is what happens when a plan is built on a category nobody had to prove, and the published order of options is still there, unused.