GLP-1s for PCOS and insulin resistance: what TikTok gets wrong

What did @amyinhalf actually say?

She made a straightforward analogy: if we accept blood pressure medication as preventative care, why not use GLP-1 medications to prevent diabetes when someone already has PCOS, insulin resistance, or prediabetes? Her closing line, "if I can take a medication that prevents diabetes, why wouldn't I," frames GLP-1 use as logical preventative medicine, not just a weight loss tool.

This is a more sophisticated argument than most GLP-1 content on TikTok. She's not claiming these drugs cure anything. She's invoking the logic of established preventative medicine and applying it to metabolic conditions that carry real downstream risk. That framing deserves an honest evaluation, not reflexive skepticism.

Does the science back this up?

Largely, yes, though the evidence is stronger for some conditions than others. The clearest data comes from the SCALE Obesity and Prediabetes trial (le Roux et al., 2017, The Lancet), which found liraglutide reduced progression from prediabetes to type 2 diabetes by 80% over three years compared to placebo. That is a striking number.

For PCOS specifically, the picture is more complex. Women with PCOS have significantly elevated lifetime diabetes risk due to chronic insulin resistance and hyperandrogenism. A 2023 meta-analysis (Guo et al., Frontiers in Endocrinology) found GLP-1 receptor agonists improved insulin sensitivity, reduced androgen levels, and supported weight loss in PCOS populations. Whether that translates to long-term diabetes prevention in PCOS specifically has not been definitively established in randomized trials yet, but the mechanistic rationale is solid.

The blood pressure analogy holds up. Treating elevated blood pressure before a cardiac event is standard care. Treating metabolic dysfunction before overt diabetes is not a fringe idea.

What did they get wrong (or right)?

She got the core logic right. PCOS, insulin resistance, and prediabetes are genuine diabetes risk factors. The American Diabetes Association's 2024 Standards of Care explicitly lists prediabetes as a target for intervention, including pharmacotherapy when lifestyle changes are insufficient. That directly supports her argument.

Where she was imprecise: calling obesity a "disease" is medically accepted (the AMA recognized it as such in 2013), but lumping it with PCOS and prediabetes as equivalent diabetes precursors oversimplifies the relationship. Obesity raises diabetes risk substantially, but the pathway differs from PCOS-driven insulin resistance or true prediabetes defined by fasting glucose or A1C thresholds.

She also says these conditions "often lead to diabetes," which is defensible but could use calibration. Prediabetes progression to diabetes is roughly 5-10% per year without intervention (Tabak et al., 2012, The Lancet), not inevitable. That distinction matters for informed decision-making.

What should you actually know?

GLP-1 receptor agonists are not currently FDA-approved specifically for diabetes prevention or for PCOS. They are approved for type 2 diabetes management and chronic weight management. Using them in prediabetes or PCOS is off-label, which does not make it wrong, but it means insurance coverage is inconsistent and prescribing requires clinical judgment about individual risk.

The preventative care analogy has real limits too. Blood pressure medications have decades of safety data at scale. Long-term GLP-1 safety data beyond five years is still accumulating. That is not a reason to avoid them if your metabolic risk is high, but it is a reason to have the conversation with an actual clinician rather than a TikTok comment section.

• Prediabetes affects roughly 96 million American adults, per the CDC, and the majority are undiagnosed.
• GLP-1 drugs are not the only intervention. Metformin, lifestyle modification, and weight management all have evidence in prediabetes.
• If you have PCOS with insulin resistance, your risk profile is genuinely elevated and worth discussing with a provider.