GLP-1 Drugs for PCOS: Semaglutide, Tirzepatide and What Happens When You Stop

What do GLP-1 drugs actually do in the body?

GLP-1 receptor agonists mimic a gut hormone your body already makes after eating, and that mimicry — a biochemical signal, not a change in resolve — is the entire mechanism. Semaglutide binds the same receptor as your natural GLP-1, which slows how fast food leaves your stomach, prompts insulin release only when blood sugar is already elevated (which is why it carries a lower hypoglycemia risk than older insulin-boosting drugs), and acts on appetite centers in the hypothalamus to reduce hunger signaling. Tirzepatide does the same thing at the GLP-1 receptor and adds a second action at the GIP receptor. The 2023 review of tirzepatide’s potential in PCOS notes that this second receptor is the proposed reason tirzepatide may produce less severe gastrointestinal side effects than GLP-1-only drugs — a real adherence problem, since nausea and vomiting are the most common reason people stop either drug early.

None of this mechanism is specific to PCOS. It’s the same pathway whether the person taking it has type 2 diabetes, general obesity, or PCOS with insulin resistance — which is exactly why the question of PCOS-specific evidence, rather than borrowed mechanism, matters so much for what follows.

Is there PCOS-specific evidence for semaglutide, or is it all extrapolated?

Semaglutide has real, if limited, PCOS-specific trial data — more than most GLP-1 coverage for PCOS implies, and more than tirzepatide currently has. A 2025 randomized, controlled, open-label trial of 100 overweight or obese women with PCOS compared metformin alone against metformin plus weekly semaglutide for 16 weeks. The combination group lost an average of 6.09 kg versus 2.25 kg with metformin alone, alongside greater improvements in testosterone, visceral adiposity, and inflammatory markers (CRP). In an extension phase, the semaglutide group also had a higher natural pregnancy rate over the following 24 weeks — 35% versus 15% with metformin alone. That is one trial, not a body of evidence, but it’s a real, PCOS-specific, randomized comparison — not an inference borrowed from a different population.

Of the 100 women randomized, 80 completed the full trial — a 20% dropout rate the published results don’t dwell on. The 35% vs. 15% pregnancy-rate comparison is drawn from the 40 women remaining per arm at week 40, not the full 100, which is worth knowing before treating either percentage as precise. The testosterone drop in that trial tracked the weight loss rather than acting independently of it — whether semaglutide does anything for hirsutism, acne or other PCOS symptoms beyond the weight itself is examined symptom by symptom here.

Tirzepatide has no such trial. A 2023 review lays out the mechanistic case for why it might help PCOS — it acts on the same GLP-1 pathway as semaglutide, plus a second receptor (GIP), which may reduce the gastrointestinal side effects that limit adherence to GLP-1-only drugs — but the review is explicit that this is hypothesis, not trial result. Every number attached to tirzepatide’s weight or metabolic effects, including the ones in the next section, comes from general obesity or type 2 diabetes trials that did not screen for or report on PCOS.

What does semaglutide vs. metformin actually look like, side by side?

The 16-week trial above is the clearest head-to-head PCOS data available, and it’s worth reading in full rather than as a soundbite about “6 kg more.”

Two limits matter here. The dose used was 1 mg weekly — well below the 2.4 mg ceiling studied for weight loss in the general population — so this trial doesn’t tell you what a higher dose would do in PCOS specifically. And it ran 16 weeks, which is long enough to show a direction but far short of the multi-year timeline most people actually stay on these drugs. Both are reasons to treat this as a genuine, encouraging first data point — not a settled answer for dosing or duration.

What happens to the weight after you stop a GLP-1 drug?

Roughly two-thirds of lost weight comes back within a year of stopping semaglutide, based on the best long-term data available — an extension of the general-population trial, not a PCOS-specific one. The STEP 1 trial extension followed 327 participants who had lost an average 17.3% of body weight over 68 weeks on semaglutide. STEP 1 was sponsored by Novo Nordisk, semaglutide’s manufacturer, and several authors are listed as Novo Nordisk employees and shareholders — standard practice for a drug’s own registration trial, and worth naming given how much of this section rests on this one dataset. One year after stopping the drug — and the accompanying lifestyle program — participants regained 11.6 percentage points of that loss, landing at a net 5.6% below their starting weight, down from the 17.3% they’d reached on treatment. The comparison group that had been on placebo the whole time showed almost no change either way. Cardiometabolic improvements — blood pressure, lipids, glucose markers — reverted toward baseline on roughly the same timeline as the weight.

No equivalent stopping study exists for tirzepatide or for PCOS specifically, so treat the STEP 1 numbers as the best available estimate of the pattern, not a PCOS-confirmed one.

Does semaglutide or tirzepatide cause muscle loss, and does that matter?

Roughly a quarter of the weight lost while on tirzepatide is lean mass, not fat — but that split turns out to be a property of weight loss itself, not something specific to the drug. The SURMOUNT-1 body-composition substudy — sponsored by Eli Lilly, tirzepatide’s manufacturer, with six co-authors listed as Lilly employees and shareholders, standard for a registration-adjacent trial — scanned 160 participants and found that of the total weight lost, about 75% was fat mass and 25% was lean mass on tirzepatide. The placebo group showed the identical 75/25 split: fat mass fell 8.2% and lean mass fell 2.6% on placebo, versus 33.9% and 10.9% on tirzepatide. The ratio held up whether someone lost a little weight or a lot, and it held up on placebo too.

What tirzepatide changes isn’t that ratio — it’s the total amount of weight lost. A much bigger overall loss means a much bigger absolute amount of muscle lost alongside it, even though the fraction is no different from what placebo, or any other route to that much weight loss, would produce. In absolute terms that’s still a meaningful amount of muscle, which is exactly why resistance training and protein intake earn a place in this conversation — not because the drug does something unusual to muscle tissue, but because moving more total weight moves more total muscle along with it. A small case series tracking patients who paired GLP-1 therapy with structured resistance training and protein intake above roughly 1.6 g per kilogram of fat-free mass found much better preservation of lean tissue than the population-level SURMOUNT-1 proportions above — a promising early signal, not proof, since case series lack a comparison group. The direction is consistent with basic physiology: muscle responds to mechanical loading and protein availability regardless of what’s driving the calorie deficit, and a GLP-1-driven deficit doesn’t exempt anyone from that. See strength training for PCOS for a starting program built around the same insulin mechanism these drugs are working on. Hair shedding is the other well-documented consequence of losing weight this quickly — telogen effluvium versus androgenic hair loss separates the two patterns and their different timelines.

Who should not take a GLP-1 drug — the contraindications

A personal or family history of medullary thyroid carcinoma or Multiple Endocrine Neoplasia type 2 rules out GLP-1 receptor agonists entirely — this is an absolute contraindication, not a relative one requiring a judgment call. A systematic literature review of semaglutide’s thyroid safety profile found that while trials to date have not confirmed a causal link to thyroid cancer in the general population, the rodent-model signal behind the original warning is the reason this contraindication has not been lifted, and it applies regardless of how compelling the metabolic case looks otherwise.

Beyond that, a personal history of pancreatitis is a relative contraindication requiring a specific risk conversation with a prescriber before starting, and pregnancy is a reason to stop rather than start — GLP-1 drugs are not studied for safety in pregnancy, which matters specifically in PCOS given how many people start one of these drugs while also actively trying to conceive. The trial in Table 1 illustrates that point rather than contradicting it: semaglutide ran for 16 weeks and was then stopped in both arms, with metformin alone carrying every participant through the pregnancy window that followed. The higher pregnancy rate in the group that had received semaglutide is evidence for what stopping the drug before conception can look like, not for continuing it once pregnant. The sequencing — and the point at which to stop — is a conversation for a prescriber, not a self-directed decision.

Several further contraindications trace back to the same mechanism that opened this article: slowed gastric emptying. Gastroparesis is a labelled risk in its own right and existing gastroparesis generally rules these drugs out. The same slowed emptying is behind the elevated gallbladder risk — gallstones and gallbladder disease are more common on GLP-1 drugs, and the rapid weight loss this article documents compounds that risk further. It is also the reason tirzepatide’s label specifically advises backup contraception for four weeks after starting and after every dose increase: delayed gastric emptying can reduce how completely an oral contraceptive is absorbed, which matters directly here since the combined pill is a routine PCOS prescription and, as above, these drugs are not studied for safety in pregnancy. And because delayed emptying raises the risk of stomach contents being aspirated under sedation, current anaesthesiology guidance calls for holding these drugs before elective procedures — relevant for anyone in this audience facing a laparoscopy or an egg retrieval. Breastfeeding sits alongside pregnancy as a further reason to avoid starting, since neither has safety data behind it.

What symptoms mean same-day medical attention while you’re taking a GLP-1 drug?

A few symptoms on this drug class are urgent rather than routine, and they map directly onto the risks above. Severe, persistent abdominal pain — especially pain radiating to the back, with or without vomiting — is the hallmark of pancreatitis and needs same-day medical attention regardless of whether pancreatitis history was screened for at the start. Pain in the upper right abdomen, especially alongside fever, yellowing skin or eyes, or pale stools, points to gallbladder disease and needs the same urgent response. A lump or swelling in the neck, hoarseness, difficulty swallowing, or unexplained shortness of breath are the symptoms tied to the thyroid-tumor warning and should be reported to a prescriber promptly rather than tracked until the next routine visit. And while GLP-1 drugs carry a lower hypoglycemia risk on their own, that changes when one is combined with insulin or a sulfonylurea — shakiness, sweating, confusion, or a racing heartbeat in that combination is low blood sugar until proven otherwise and needs the same immediate response it would on either drug alone.

What is the compounded-semaglutide risk, specifically?

A case series reported to a regional poison control center described three patients who suffered adverse drug events after obtaining semaglutide from compounding pharmacies or a med spa, and two of the three had self-administered doses roughly 10 times higher than intended — the clearest documented evidence that compounded semaglutide, sold outside the FDA-approved supply chain and often without the safety mechanisms built into manufactured pens, carries a real, quantified overdose risk. This happens because compounded vials require the patient to draw up their own dose with a syringe rather than using a pre-set pen, and dosing units (milligrams, milliliters, “units”) got confused in the process. All three patients experienced days of nausea, vomiting and abdominal pain; one required IV fluids.

This isn’t a case against GLP-1 drugs — it’s a case against the unregulated supply chain some people use to access them more cheaply or without a prescription. A prescription for the manufactured, FDA-approved product, dispensed with the calibrated pen it was designed for, doesn’t carry this specific risk. Microdosing — deliberately taking less than a standard dose — runs through this same self-measured supply chain in practice, and what the evidence for microdosing GLP-1 actually shows is worth reading before assuming it’s a safer middle ground.

Ozempic vs. metformin for PCOS — which comes first?

The only head-to-head PCOS trial available tested semaglutide added to metformin over 16 weeks, not semaglutide instead of it, and found the combination outperformed metformin alone on weight, hormones and pregnancy rate. Metformin remains the more established option for PCOS’s underlying insulin resistance, with a trial history that runs decades longer — see metformin for PCOS for the mechanism and dosing conversation. That’s a meaningfully different question from “which one should I start with” — the trial doesn’t establish semaglutide as a metformin replacement, only as something that added benefit on top of it in this specific 16-week window. For most people, metformin’s much longer track record makes it the reasonable starting conversation. Beyond that, this article can’t hand you a sequencing rule: the 2023 international PCOS guideline this site checks against doesn’t name GLP-1 drugs as PCOS-specific therapy at all — its pharmacological recommendations centre on metformin and combined hormonal contraception, and every GLP-1 access pathway runs through general obesity or diabetes guidance instead (see GLP-1 access and cost, which covers this in full). Whether and when a GLP-1 drug makes sense alongside or after metformin is a question for a prescriber working from your specific weight, metabolic markers and history — not a step-therapy sequence this page can supply. A different, older option sometimes comes up in the same conversation: phentermine, a short-term appetite suppressant that works through neither of the mechanisms above. Naltrexone-bupropion is a third appetite-focused option, working through yet another mechanism, worth knowing about alongside these two.

You may see PCOS referred to as polyendocrine metabolic ovarian syndrome (PMOS), after a 2026 global consensus of more than 50 medical organisations renamed it. Nothing about the drug mechanisms or trial data above changes under either name — this article uses PCOS because that’s still what most readers search.

This is one entry in the full weight-loss guide, which treats weight as a metabolic marker tied to insulin, androgens and cardiometabolic risk — not a number a drug is meant to force down for its own sake.

Where this will not work for you

If your PCOS presents lean, with normal or low body weight and androgen excess as the dominant feature, none of the trials above apply to you — every one of them enrolled women who were overweight or obese, and GLP-1 drugs are approved and studied for weight-related metabolic indications, not as a treatment for androgen excess on its own. If your main goal is fertility and you are not also managing a weight-related metabolic picture, the trial in Table 1 doesn’t tell you whether semaglutide would help your specific case; it tested a population selected for excess weight. And if you have any of the contraindications above, this entire drug class is off the table regardless of how well-suited the metabolic profile otherwise looks.

Your next step

If you’re overweight or obese with PCOS and weighing whether a GLP-1 drug fits into your care, bring three things to your next appointment: your personal and family thyroid and pancreatitis history, your current fertility plans and timeline, and a direct question — “Is a GLP-1 medication appropriate for me, and if I start one, what does stopping it eventually look like?” That last question, given the STEP 1 data above, deserves an answer before you start, not after. If a GLP-1 turns out not to be appropriate, or the regain after stopping one is the deciding factor, when surgery outranks a GLP-1 prescription is worth understanding at the same appointment.