Accepted answer
Insulin resistance is upstream of the androgens in most cases, and the chain is specific enough to be worth writing out. It also answers your lean-PCOS question, because the defect is in insulin signalling rather than adiposity, which is an amplifier rather than the cause.
The chain
Four steps, each independently documented:
- Insulin resistance produces compensatory hyperinsulinaemia. Peripheral tissues resist insulin's metabolic actions, beta cells compensate by secreting more, and circulating insulin rises. Crucially the resistance is selective: it affects glucose disposal in muscle and adipose tissue while leaving other insulin-responsive pathways intact.
- Insulin acts on the ovarian theca cell to augment androgen production. Theca cells retain insulin sensitivity for steroidogenesis, and insulin acts synergistically with luteinising hormone to increase androgen output. So hyperinsulinaemia is a direct androgen-driving signal at the ovary, not merely a correlate.
- Insulin suppresses hepatic sex hormone binding globulin synthesis. This is the step most people miss and it is arithmetically the largest. SHBG binds testosterone with high affinity; the unbound fraction is what acts on tissue. Suppress SHBG and free testosterone rises even at unchanged total testosterone. So hyperinsulinaemia raises androgen production and increases the biologically active fraction of what is produced, simultaneously.
- Elevated intra-ovarian androgen disrupts follicular development. Follicles are recruited and then arrest rather than selecting a dominant follicle, producing the polycystic morphology and anovulation. Disordered GnRH pulse frequency, with a raised LH-to-FSH ratio, both contributes to and results from the androgen excess, so the loop is self-reinforcing.
Adiposity enters at step one as an amplifier — visceral adiposity worsens insulin resistance — and again via adipose aromatase activity altering the oestrogen environment. That is why weight loss improves the picture and why it is not the fundamental lesion.
Why lean women get PCOS
Because the insulin resistance in PCOS is not solely a consequence of fat mass. There is a body of evidence for an intrinsic, post-receptor defect in insulin signalling in PCOS tissue — abnormal serine phosphorylation of the insulin receptor and downstream substrates — present in lean women with the syndrome and independent of BMI. Lean women with PCOS are, on average, more insulin resistant than BMI-matched women without it.
So: PCOS involves an intrinsic insulin-signalling defect, and adiposity adds an acquired one on top. Both feed step one, which is why the phenotype spans the BMI range, why "just lose weight" is partially correct and insufficient, and why weight loss helps a great deal in some women and disappointingly little in others — it removes the acquired component and leaves the intrinsic one.
Where a GLP-1 receptor agonist acts
Mostly, and possibly entirely, at step one — and mostly through weight rather than through a direct ovarian action. The honest breakdown:
- Weight loss reduces insulin resistance, lowering circulating insulin, which relieves steps two and three. The dominant pathway, and well established for weight loss by any means.
- Glucose-dependent effects on insulin secretion and on glucagon improve glycaemia somewhat independently of weight, which may reduce hyperinsulinaemia modestly beyond the weight effect.
- Direct ovarian GLP-1 receptor signalling is speculative. Receptor expression has been reported in some reproductive tissues in animal models; there is no persuasive human evidence of a meaningful direct ovarian effect. Anyone presenting this as established is over-reaching.
The implication of "mostly through weight" is that expected benefit scales with weight lost, so a woman with lean PCOS and a strong intrinsic defect has less to gain by this route than one with BMI 34 and marked visceral adiposity. Useful to know before starting rather than after.
On the evidence, since you asked to be told rather than sold
The evidence base for GLP-1 receptor agonists specifically in PCOS is genuinely thin, and here is the shape of it.
- No large dedicated randomised outcome trial exists. The PCOS-specific randomised evidence is small studies — dozens to a couple of hundred participants, mostly 12 to 32 weeks, mostly liraglutide or exenatide — reporting weight, insulin indices, androgens and menstrual frequency rather than pregnancy or live birth.
- The large obesity programmes did not enrol PCOS as a stratum. STEP and SURMOUNT recruited on BMI and comorbidity criteria; PCOS was neither an inclusion criterion nor a prespecified subgroup with reproductive endpoints. Their weight-loss numbers transfer as an expectation about weight, and not at all as evidence about ovulation, hirsutism or fertility.
- The international guideline is explicit. The 2023 international evidence-based guideline positions anti-obesity pharmacotherapy including GLP-1 receptor agonists as an option for weight management in PCOS while noting the PCOS-specific evidence is limited, and retains metformin as the insulin-sensitising agent with the largest PCOS-specific evidence base [1].
- The lifestyle-weight-loss literature is older and better. Studies from the 1990s established that 5 to 10% weight loss restores ovulatory cycles in a substantial minority of women with PCOS and obesity. That supports "weight loss helps PCOS"; the inference to a specific agent is a further step.
The defensible summary: strong mechanistic rationale, consistent small-trial signals on surrogates, robust evidence that these agents produce the weight loss the older literature says helps, and no adequately powered PCOS-specific trial with reproductive outcomes. A clinic describing this as a PCOS treatment is stating a plausible extrapolation as a finding.
Nothing here is medical advice, and PCOS management benefits from a clinician who addresses your individual priority — cycles, hirsutism, metabolic risk or fertility — because the sensible first line differs by which one it is.
edited 9 Sept 2024 by tri_gly_ala — expanded the table to cover the lower concentration
Setting out the four steps and then showing that SHBG suppression is arithmetically the largest is the clearest version of this I have read. – claudia_ferrante 5 months ago 2The selective nature of the insulin resistance — glucose disposal impaired, ovarian steroidogenesis not — is the detail that makes the whole chain make sense. – Dr_Bram_Verhoeven 6 months ago 3Thank you for stating plainly that the obesity trials did not enrol PCOS as a stratum. Every clinic page implies otherwise. – plate_count_9k 41 days ago add a comment