Insulin Resistance and GLP-1s: The Test That Isn't, the PCOS Link, and What Ozempic and Zepbound Change

Does Ozempic help insulin resistance?

Yes, in trials. Semaglutide lowered fasting insulin and HOMA-IR, the usual insulin-resistance estimate, more than placebo in adults without diabetes, and the gains faded after people stopped. Most of the effect tracks weight loss. In a clamp study in type 2 diabetes, tirzepatide (Zepbound, Mounjaro) improved insulin sensitivity more than semaglutide.

That answer is the easy part. The harder part, and the one most of us are actually stuck on, is the word itself. "Insulin resistant" gets said in exam rooms, on reels and in PCOS forums as if it were a diagnosis with a number attached. It mostly is not.

What insulin resistance is, and why it is not a lab diagnosis

Insulin moves sugar out of the blood and into muscle, liver and fat. When those tissues answer it less readily, the pancreas makes more to get the same job done. For years, sometimes decades, that works: blood sugar stays normal because insulin is running high. Insulin resistance is the name for that hidden extra effort. Prediabetes and type 2 diabetes are what happens when the pancreas cannot keep up.

The trouble is measuring the effort. The gold standard, the euglycemic clamp, means hours on two IV lines in a research unit. Everything a regular clinic can order is an estimate, and the 2023 international PCOS guideline is blunt about it: insulin resistance is a key feature of PCOS, "yet routinely available measures of insulin resistance are inaccurate and clinical measurement is not currently recommended."¹

The tests, and what each one can and cannot tell you

HOMA-IR is the number most likely to show up in your patient portal, so here is how it works. It was published in 1985 as a model of the feedback loop between fasting glucose and insulin. Its estimate tracked the clamp closely (a correlation of 0.88), but the authors warned that its low precision, a 31% coefficient of variation, "limits its use"².

Is 2.8 high? It depends on whose cut-off you use. In one Spanish population study of 2,459 adults, the threshold fell from 3.46, using the top 10% of the population, to 2.05 when it was tied to metabolic syndrome risk. In non-diabetic women, the best cut-off also shifted with age³. A single number on a portal, without that context, does not tell you much. It is a prompt for a conversation, not a verdict.

PCOS: where insulin resistance is the rule

If any group has reason to care, it is women with PCOS. In a small clamp study, insulin resistance was present in 75% of lean women with PCOS and 95% of overweight women with PCOS, against 62% of overweight women without it⁴. The authors called it intrinsic: it is there even without extra weight, and weight makes it worse⁴. The ADA lists PCOS, along with signs like acanthosis nigricans (darkened skin folds), among the reasons to be screened for diabetes⁵.

What women with PCOS say about getting there is its own finding. In a survey of 1,385 women with PCOS, a third (33.6%) reported more than two years and nearly half (47.1%) reported three or more health professionals before a diagnosis. Only 15.6% were satisfied with the information they were given. Their single most common concern was difficulty losing weight, at 53.6% of respondents⁶. If you have spent years hearing "just lose weight" about a condition that makes losing weight harder, those numbers are you.

What GLP-1s actually do to insulin sensitivity

Semaglutide. In STEP 1, adults with overweight or obesity and no diabetes saw bigger falls in fasting insulin and HOMA-IR on semaglutide 2.4 mg than on placebo⁷. In STEP 4, those improvements held while people stayed on the drug. They got worse after the switch to placebo⁷. The benefit is real, and it lasts only as long as you keep taking the drug.

Tirzepatide. The cleanest mechanism data come from a clamp study in 117 adults with type 2 diabetes, not from women with PCOS. After 28 weeks, tirzepatide 15 mg improved clamp-measured insulin sensitivity more than semaglutide 1 mg did⁸. A follow-up analysis of the same trial found tirzepatide improved insulin sensitivity more per kilogram lost, and concluded the gain was "not simply attributable to greater weight or fat loss"⁹. The GIP half of the molecule is the leading explanation, but that is still a hypothesis.

Two honest caveats. Neither drug is approved to treat insulin resistance, and tirzepatide has no PCOS-specific trials, a gap we cover in GLP-1s and PCOS: the evidence and the hype. The PCOS guideline itself places GLP-1s as an option for higher weight, alongside lifestyle care, and asks for reliable contraception while taking them because pregnancy safety data are lacking¹.

Metformin, inositol and lifestyle first

The same guideline says metformin should be considered in adults with PCOS and a BMI of 25 or more, for metabolic outcomes including insulin resistance¹. On inositol, the supplement half the internet recommends, it is careful. Specific types, doses or combinations cannot currently be recommended because quality evidence is lacking, and supplement quality can vary¹. Weight loss of any kind, and regular activity, improve insulin sensitivity. That is why every trial above ran a lifestyle program in both arms.

If your blood sugar is already in the prediabetes range, our prediabetes and GLP-1s feature covers the numbers and the prevention trials. If you want a practice that treats insulin resistance as the main subject rather than a checkbox, Allara Health and PCOS Sisters are built around it. Our board for women lists more.

Questions worth taking to your clinician

What are my A1c and fasting glucose, and has anyone looked at a 2-hour glucose test? If you ordered fasting insulin or HOMA-IR, what cut-off are you using, and would the result change what we do? Is metformin a reasonable first step for me? If we try a GLP-1, what is the plan for contraception, and what happens to my insulin sensitivity if I stop? And if your cycle changes once you start, which is common, our piece on heavier periods and PMS on a GLP-1 explains why.