All studies
Metabolism·October 16, 2026·7 min read

Metformin Inositol or GLP-1 for PCOS Insulin Resistance

Metformin, inositol, and GLP-1s each help PCOS insulin resistance, but the right choice depends on whether the driver is fasting insulin, post-meal insulin, or weight.

There is no single best medication for PCOS insulin resistance. Metformin may be preferred when the dominant signal is high fasting insulin and androgen excess. Inositol may be preferred when the main issue is high post-meal insulin and excess weight is not the main driver. GLP-1s move the most when excess weight is driving the insulin resistance. The useful question is not which one is strongest, but which pattern you are actually trying to fix.

The short version: each option matches a different signal
MetforminInositolGLP-1
Best fitHigh fasting insulin, high androgensHigh post-meal insulin, weight not main driverExcess weight driving insulin resistance
Lab clueFasting insulin high, fasting glucose normalPost-load insulin high on OGTTWeight high, fasting insulin high
CGM clueBaseline creeps up overnightMeal rise above 140 mg/dL may last longer than 2 hoursGeneral high variability from larger meals
Typical effectLower fasting insulin, more cyclesFlatter post-meal insulin, better cycle regularityWeight loss, lower insulin, lower androgens

Scroll the table sideways for the rest →

These are patterns seen in practice; individual responses vary.

What medication actually helps PCOS insulin resistance: metformin, inositol, or GLP-1?

All three can work, but they work through different mechanisms. Metformin mainly lowers basal insulin and hepatic glucose output. Inositol improves insulin signalling after meals, particularly in ovarian cells. GLP-1s reduce weight, delay gastric emptying, and increase glucose-dependent insulin release. Metformin and inositol may be preferred based on the clinical pattern, but this is not a strict split; some people respond better to one than the other, and some use both.

Choose the option that matches your dominant driver: high fasting insulin, high post-meal insulin, or excess weight.

What insulin resistance actually looks like on labs and a CGM

PCOS insulin resistance often hides behind normal fasting glucose and normal A1c. The body compensates by producing more insulin, so glucose can look fine while insulin stays high. On a lab report, some clinicians use illustrative patterns such as fasting insulin above roughly 10 to 12 mIU/L, or an oral glucose tolerance test where insulin at 60 or 120 minutes rises far above fasting. HOMA-IR uses fasting glucose and insulin together; in PCOS, some clinicians use an illustrative pattern above 2.0 to 2.5. These are not established diagnostic cutoffs, and no single source is cited here. Post-load insulin can be more informative when fasting values are normal.

On a CGM, you cannot see insulin, only glucose. Some people with insulin resistance do not see a dramatic spike. Instead, they notice a slow return: a meal pushes glucose above 140 mg/dL and it stays there for longer than two hours, or the overnight baseline runs higher than an illustrative typical baseline of 80 to 90 mg/dL, for example 95 to 105 mg/dL. These are clinical patterns, not diagnostic cutoffs. Healthy adults in a 2024 CGM study by Spartano and colleagues (Spartano 2024) spent about 3 hours a day above 140 mg/dL, so the goal is to find your own pattern, not to chase a flat line.

Metformin is the better pick for high fasting insulin and androgen excess

Metformin activates AMPK in the liver, which reduces gluconeogenesis and lowers basal insulin. In PCOS, that basal insulin reduction has a downstream effect on the ovary: lower insulin means less stimulation of androgen production, and SHBG can rise. That is why metformin tends to help with high fasting insulin, high free testosterone, and irregular cycles together.

Dose is usually started low, 500 mg once daily, and increased by 500 mg every 1 to 2 weeks to 1500-2000 mg daily, often split. Extended-release metformin is easier on the gut. The change you see first is usually a lower morning glucose or a lower fasting insulin, not weight loss. Metformin is not a meaningful weight-loss drug; a few kilograms are possible, but not guaranteed.

On CGM, metformin often flattens the overnight baseline more than the post-meal rise. If your main problem is a large meal spike, metformin alone may disappoint, but it can still be combined with other options.

Inositol is the better pick for high post-meal insulin when weight is not the main driver

Inositol is not a classic drug in every country, but it is one of the most studied supplements for PCOS. Myo-inositol and D-chiro-inositol act as second messengers in the insulin signalling pathway, downstream of the insulin receptor. In the ovary, this can improve glucose uptake and reduce insulin-driven androgen production. The typical studied dose is 2000 mg myo-inositol plus 50 mg D-chiro-inositol twice daily, in the 40:1 ratio found in plasma.

Inositol may be considered when the main problem is high post-meal insulin and excess weight is not the main driver. Cycle regularity can improve, often over 3 to 6 months. On a CGM, you might see lower post-meal excursions and a faster return to baseline if the problem was signalling rather than weight. It is not a weight-loss treatment.

GLP-1s are the strongest lever when excess weight is the main driver

For many people with PCOS, the insulin resistance is weight-driven. Extra adipose tissue, especially visceral fat, raises baseline insulin and worsens ovarian androgen production. A GLP-1 receptor agonist such as semaglutide, or a dual GIP/GLP-1 like tirzepatide, reduces appetite and slows gastric emptying, and the resulting weight loss lowers fasting insulin, androgens, and often improves cycles.

Weight loss with GLP-1s varies by agent and dose, but 5 to 15 percent of body weight over several months is common, with the dual agents usually at the higher end. Some metabolic improvements follow weight loss, but GLP-1s may also have weight-independent effects on insulin release and glucose handling. If weight is not the main driver, GLP-1s can still help, but they may be more medicine than the problem requires.

On labs, changes vary by person; weight, fasting insulin, and testosterone can improve at different times. On CGM, post-meal rises often shrink because meals are smaller and stomach emptying slows, not because the insulin resistance itself disappeared overnight.

How to choose: blood tests and CGM patterns

Use the pattern, not the A1c alone. Ask for fasting glucose, fasting insulin, HOMA-IR, and ideally an oral glucose tolerance test with insulin at 60 and 120 minutes. If the fasting insulin is high, metformin is the usual starting point. If the fasting insulin is unremarkable but post-load insulin is high, inositol is the better fit. If excess weight is large and fasting insulin is high, GLP-1 is often the strongest option.

Match the driver to the option
High fasting insulin
Normal fasting glucose
High post-meal insulin
Big OGTT insulin at 60/120 min
Excess weight
Most of the insulin resistance
Metformin
First-line for high basal insulin
Inositol
Often considered for meal-driven insulin
GLP-1
Strongest for weight-driven

The common mistake is treating PCOS insulin resistance as if a normal fasting glucose or normal A1c rules it out. It does not. A normal glucose with high insulin is still insulin resistance, and it can still drive anovulation, androgens, and long-term metabolic risk. The second mistake is expecting metformin or inositol to do what a GLP-1 does for weight. These are not interchangeable.

What if my fasting glucose and A1c are normal?

This is the most common PCOS pattern. Ask for a fasting insulin and a HOMA-IR next. If you can, get an OGTT with insulin at 0, 60, and 120 minutes, because many women with PCOS only show the problem after a glucose load. A CGM can help you see whether meals are returning to baseline slowly, but it cannot measure the insulin your pancreas is using to keep glucose there. If glucose looks normal but your CGM shows a slow return or a high baseline, that is enough reason to investigate further.

Limits and what the evidence does not settle

Inositol has good mechanistic data and many positive trials, but trial quality is variable and it is not regulated as a drug in most places. GLP-1s have strong weight-loss data, but long-term PCOS fertility data are still accumulating. There is no single head-to-head trial that proves one option superior across all PCOS phenotypes, because the condition is not one phenotype. This also does not apply to people with type 1 diabetes, pregnancy, or anyone being treated for another condition without their clinician's input.

Once you can see your own glucose pattern, the choice usually gets clearer. Upload a LibreView, Dexcom Clarity, Stelo, Lingo, or SiBionics export at https://healthosx.com/glucose and your browser will show your AGP, time in range, and biggest rises, no account. If you have no sensor, HealthOS Sensor is a wellness product, not a medical device, and is not intended for diagnosis or treatment. It is sold in supported countries, not in the US, for $70 for a 30-day supply with no membership; see https://healthosx.com/glucose/sensor.

Written by

HealthOS Research

Keep reading