PCOS and insulin resistance: how biochemistry restores hormonal balance

Polycystic ovary syndrome (PCOS) affects 10–13% of women of reproductive age and is under-treated in Estonia. Insulin resistance underlies more than 70% of cases — which is why lifestyle intervention often outperforms hormonal medication alone. This guide takes a biochemical view: why excess insulin turns the ovary into an androgen factory, which labs are clinically justified, and what steps restore ovulation in 12–16 weeks.

What PCOS is — and why it is not just "cysts"

PCOS is a hormonal and metabolic syndrome. Diagnosis uses the Rotterdam criteria: two of three features — 1) irregular or absent ovulation, 2) clinical or lab hyperandrogenism (acne, hirsutism, elevated testosterone), 3) polycystic ovarian morphology on ultrasound. Visible cysts are not required — the name is misleading. The 2023 international guideline (Teede et al., Fertility and Sterility 2023) is explicit: PCOS is primarily a metabolic disease, with reproductive symptoms as a consequence.

How insulin resistance feeds the androgens

The key mechanism: elevated insulin (a) stimulates ovarian theca cells to produce more testosterone and androstenedione, (b) suppresses hepatic SHBG production, raising free biologically active testosterone even higher, (c) arrests follicular maturation — ovulation fails. Nestler et al. (JCEM 2001) showed that lowering insulin with metformin reduced androgens and restored ovulation in about half of women with PCOS. Lifestyle intervention delivers a comparable or stronger effect through the same biochemistry (Legro et al., JCEM 2016).

Which labs actually matter

The standard PCOS panel I order:

• Fasting insulin + glucose → HOMA-IR (most sensitive for hidden hyperinsulinaemia)
• HbA1c → long-term glycemic picture
• Total testosterone, SHBG → free-testosterone index (FAI)
• DHEA-S → adrenal-origin androgens
• LH and FSH (cycle days 2–5) → LH/FSH > 2 is classic but not required
• 17-OH-progesterone → to rule out congenital adrenal hyperplasia
• Prolactin, TSH, free T4, TPO-Ab → to distinguish look-alikes
• Extended lipids, ALT/AST → the metabolic side

A pelvic ultrasound on cycle days 2–5 is useful but not mandatory — the 2023 guideline permits a diagnosis without ultrasound if the other two criteria are clear. Ovarian appearance is a consequence, not a cause.

PCOS phenotypes — why one plan does not fit all

PCOS is divided into four phenotypes (Azziz et al., Nat Rev Dis Primers 2016). The commonest is phenotype A: hyperandrogenism + oligo-anovulation + polycystic ovarian morphology — and it carries the strongest insulin resistance (~80% of cases). Phenotype D is subtler: irregular cycles + polycystic ovaries without overt hyperandrogenism — insulin is less often the driver, and disrupted hypothalamic rhythm more so. That distinction changes the plan: phenotype A needs insulin lowering first; phenotype D needs sleep, training and stress regulation.

What lifestyle changes are clinically proven

Three lifestyle interventions carry high-quality evidence: 5–10% weight loss (if BMI > 25), carbohydrate quality and quantity, and strength training. The Moran et al. Cochrane meta-analysis (2011) showed that a 5% weight loss restores ovulation in about 30% of women, and 10% in about 60% — without medication. On carbohydrates, what matters is not low-carb per se but glycemic load: refined sugar, white flour and sweet drinks raise insulin the most and longest. Fiber, adequate protein and mono-saturated fats (Mediterranean or low-glycemic pattern) are the first choice. I do not require strict keto if BMI is near normal — it works, but long-term social sustainability matters more than strict biochemistry.

Where metformin and GLP-1 come in

Metformin is not first-line for PCOS, but a useful add-on: if HOMA-IR is above 3, if there is a gestational diabetes risk, or if ovulation has not returned in 12–24 weeks of lifestyle work. GLP-1 agonists (semaglutide, tirzepatide) have shown weight and insulin reductions in PCOS in recent studies. But — per Estonian Advertising Act § 21 — specific-medication recommendations happen only in a personal consultation with full labs, not in an article.

What to expect in 12–16 weeks

With correct lifestyle work, in practice I see: cycles shorten and regularise (ovulation often appears by weeks 8–12), acne calms (weeks 12–16 — the follicle needs a cycle to catch up), post-meal fatigue disappears (weeks 2–4 — the earliest marker), and HOMA-IR falls on average 30–40% by week 12. That internal order — metabolism first, reproductive symptoms after — confirms the direction is right.

PCOS and long-term risk

A woman with PCOS carries a lifelong elevated risk of type 2 diabetes (4×), gestational diabetes (3×), fatty liver disease (2×), endometrial cancer (2–6×, from prolonged anovulation), and depression (3×). That is why the metabolic side deserves serious attention immediately, even when pregnancy is not currently on the horizon: I screen every 1–3 years.

PCOS is not only an ovarian problem — it is a metabolic diagnosis

Up to 70% of PCOS patients are insulin resistant regardless of body weight (Diamanti-Kandarakis & Dunaif, Endocr Rev 2012). Insulin drives ovarian androgen production and lowers hepatic SHBG — a loop the only effective break of which is restoring insulin sensitivity, not just hormonal contraception.

Four treatment pillars often skipped

• Inositol (myo + D-chiro 40:1) 4 g/day — restores ovulation in ~50% of women (Nordio & Proietti, 2012)
• Resistance training 3×/week — raises muscle GLUT4 and lowers fasting insulin
• Sleep as priority — 1 h of debt raises morning testosterone and androstenedione
• Vitamin D 2000–4000 IU when serum < 75 nmol/l

Pregnancy and PCOS — a realistic timeline

Metabolic intervention (−5–10% weight, HOMA-IR below 2.0) restores ovulation in ~60% of women within 3–6 months (Legro et al., NEJM 2007). Often more effective than jumping straight to clomiphene or IVF, and the effect lasts longer.