GLP-1 and the PCOS Metabolic Phenotype

PCOS is diagnosed on reproductive criteria — irregular cycles, high androgens, ovaries with a particular appearance on a scan. That is how it is found; for most patients it is not what the condition is. Underneath sits a metabolic picture: insulin resistance, compensatory hyperinsulinaemia, fat in the visceral compartment. The reproductive features are largely downstream of that, which is why treating the metabolic layer moves the reproductive layer and treating the reproductive layer alone moves nothing else.

What the diagnosis captures, and what it leaves out

The Rotterdam criteria require any two of three: anovulation, clinical or biochemical hyperandrogenism, polycystic ovaries on imaging (Rotterdam ESHRE/ASRM Consensus Workshop Group, Fertil Steril 2004). They were built to be usable in a gynaecology clinic and they do that job. Their weakness is that nothing in them mentions insulin.

Most patients who meet those criteria also have insulin resistance, hyperinsulinaemia and central adiposity, and the relationship runs in a direction: high insulin drives ovarian androgen production, and the androgens produce the features the criteria describe. The dysfunction is causal rather than merely associated (Diamanti-Kandarakis & Dunaif, Endocr Rev 2012). Read the diagnosis as a description of what is visible rather than an account of what is happening, and everything below follows.

The phenotypes, and why lean PCOS confuses everyone

Four presentations are usually described:

• Classic PCOS — central adiposity, insulin resistance, irregular cycles, hyperandrogenism, hirsutism and acne. The easiest to recognise.
• Lean PCOS — a normal BMI alongside visceral adiposity, insulin resistance and androgen excess.
• Ovulatory PCOS — regular cycles, but androgens and metabolic markers still elevated.
• Non-hyperandrogenic PCOS — anovulation and polycystic ovaries without androgen excess.

Lean PCOS is where the reasoning breaks down, because BMI says nothing about where fat is stored: a woman can carry a normal total fat mass with a disproportionate share packed around the organs, and the consequences follow the compartment rather than the total. In practice the label matters less than one question — is fasting insulin elevated? That single measurement separates patients who will respond to metabolic treatment from those who will not, far more reliably than which of the four boxes they sit in.

The chain, one step at a time

The mechanistic sequence in classic PCOS is worth holding in order, because you can intervene at several points and only one of them is upstream of the rest:

1. Insulin resistance develops, usually with a genetic component, amplified by body composition and inactivity.
2. The pancreas compensates, so glucose stays normal while insulin climbs. This is the stage most often missed, because glucose testing alone looks reassuring.
3. High insulin acts on ovarian theca cells and stimulates androgen production directly.
4. High insulin also suppresses hepatic production of sex hormone binding globulin (insulin and hepatic SHBG regulation). Less SHBG means more of the testosterone already present circulates unbound, so free androgen rises faster than total.
5. Elevated androgen disrupts follicle maturation, so follicles stall rather than ovulating.
6. Anovulation produces irregular or absent cycles, and infertility.
7. Androgens acting on skin and hair follicles produce hirsutism, acne and scalp thinning.

The chain also loops: androgen excess favours visceral fat deposition, visceral fat worsens insulin resistance, and insulin climbs further. That self-reinforcing quality is why the condition intensifies rather than stays put, and why an intervention that lowers insulin can produce changes out of proportion to its size — it breaks the loop rather than treating one of its outputs. SHBG and fasting insulin are the markers that let you see the chain rather than infer it.

Where a GLP-1 medication acts on that chain

GLP-1 is a hormone the gut already releases when you eat, and medications in this class are long-acting versions of that signal. They act at three points: appetite and satiety signalling in the hypothalamus and brainstem, gastric emptying, and glucose-dependent insulin release.

The relevant consequence for PCOS is that circulating insulin falls — partly because energy intake falls and fat mass with it, partly because insulin sensitivity improves independently of that. Both sit upstream of ovarian androgen production and of hepatic SHBG suppression. Nothing is aimed at the ovary, and the reproductive features move anyway. That is the whole argument for treating PCOS metabolically.

What it does not do is remove the predisposition. Medication changes the environment, not the genotype: if treatment stops and body composition returns, the metabolic environment tends to return with it, and the reproductive features follow. Worth knowing before starting.

What the PCOS-specific research shows

Studies run specifically in PCOS populations report the same pattern: weight loss, falling total and free testosterone, rising SHBG, reduced fasting insulin and HOMA-IR, more regular cycles, resumption of ovulation, improved acne and hirsutism scores. Guidance now places metabolic treatment alongside lifestyle intervention rather than after it (Teede et al., Fertil Steril 2023).

The honest caveat is that much of this literature is small, short and open-label. Cycle regularity is a soft endpoint both patients and investigators want to see improve, and hirsutism moves slowly because hair follicle cycles are slow. The direction of effect is consistent and mechanistically expected; the precision of the effect size is not.

What to expect, and in what order

The sequence patients describe is fairly reproducible, and knowing it prevents a lot of premature discouragement:

• Months 0-3. Appetite changes first and is usually obvious within days. Weight begins moving. Some women notice cycles shifting almost immediately; many notice nothing reproductive at all yet.
• Months 3-6. Testosterone falls and SHBG rises, so free androgen drops more than the total figure suggests. Acne typically improves. Scalp shedding sometimes worsens briefly first — a known response to rapid weight change, not a sign of failure.
• Months 6-12. Ovulation commonly resumes and cycles normalise. Hirsutism starts to change, slowly.
• Beyond a year. The metabolic gains — insulin, lipids, liver enzymes — are the ones that compound, and the reason this condition matters beyond fertility.

Fertility is the part that catches people out

Restored ovulation means restored fertility, and it arrives without announcing itself. Women who have been anovulatory for years, often told conception would require assistance, can become pregnant on a cycle they did not know they were having. Anyone not seeking pregnancy needs reliable contraception in place from the start of treatment, not added later. Anyone who is seeking pregnancy needs the opposite conversation: a planned, physician-coordinated stop before active attempts, with time built in for the medication to clear. GLP-1 and fertility covers the timing. These are the two most consequential practical decisions in the whole protocol.

What to measure

A panel that tracks the chain rather than the symptom:

• Total and free testosterone, plus SHBG — free is what tracks symptoms
• Fasting insulin, glucose and HbA1c — insulin is the upstream number, and the one most often not ordered
• Lipid panel and hs-CRP
• Liver enzymes — PCOS and fatty liver travel together (GLP-1 and NAFLD/MASH)
• Vitamin D, frequently low here
• Cycle tracking, free, and the most direct readout of whether the reproductive layer is responding

Resistance training belongs in the same list, because muscle is the largest site of insulin-mediated glucose disposal in the body — building it improves the exact variable this condition turns on. A three-day full-body plan does more for insulin sensitivity than additional cardio. The 60-second assessment routes the medical side to a physician.

Bottom line

For most patients PCOS is a metabolic condition wearing reproductive clothing. Insulin resistance drives hyperinsulinaemia, which drives ovarian androgen production and suppresses SHBG, and the cycles, the acne and the hair follow from there. A GLP-1 medication acts on the metabolic end of that chain, which is why the reproductive end moves without being targeted — and why the effect fades if the metabolic environment returns. The measurements that matter are fasting insulin, free testosterone and SHBG, not weight. Whether medication is appropriate is a physician's decision after individual evaluation.

Educational content, not medical advice. Laboratory interpretation and any treatment decision are made by a licensed physician after individual evaluation. Individual results vary.