Stopping Metformin for PCOS (Now PMOS): Withdrawal Symptoms, What Comes Back, and What Doesn't

The decision usually isn't dramatic. The nausea never fully settled. Or the prescription ran out during a house move. Or you got the result you were on it for, and nobody ever said how long "a while" was meant to be.

Then the question arrives: what happens now?

Most explainers answer with a list of symptoms and a reminder to talk to your doctor. That's not wrong, but it skips the more useful finding — that what returns after you stop is not the same for everyone, and the single best predictor is how long you were taking it.

PCOS is now PMOS — Polyendocrine Metabolic Ovarian Syndrome, renamed by international consensus in May 2026. Both names are still in circulation, so both appear here. Nothing below is medical advice, and stopping or continuing a prescription is a decision for you and your clinician.

There's no metformin withdrawal syndrome — but stopping isn't nothing

Worth separating two things that get filed under the same search term.

Metformin does not cause physical dependence. There is no rebound-craving, none of the pharmacology that makes stopping some medications genuinely risky. If you stop, you will not go into withdrawal in the clinical sense of that word.

What people are describing when they type "metformin withdrawal symptoms" is different, and real: the return of the biology metformin was suppressing. Metformin improves how your body handles insulin. In PMOS, insulin resistance sits upstream of much of the syndrome — higher insulin pushes the ovaries toward more androgen production, which feeds irregular cycles, jawline acne, and unwanted hair growth. Remove the medication and that pathway is no longer being nudged.

So it isn't withdrawal. It's the underlying condition, unmasked. The distinction changes what you should watch for, and over what timescale.

What happens depends on how long you were on it

This is the part almost nobody covers, and it comes from the one study that followed people properly.

A 2021 observational study in Diabetology & Metabolic Syndrome (Kravos et al.) tracked women with PCOS for six months after they stopped metformin. Forty-four enrolled, forty completed, split into two groups: short-term users, averaging about one year on the drug, and long-term users, averaging about five years. Median BMI was around 32.

The two groups came apart almost exactly opposite to how most people would guess.

The long-term users lost cycles. Their menstrual frequency dropped significantly over the six months. The short-term users' cycles held steady.

The short-term users gained weight. Roughly 4 kg — about 9 lb — a statistically significant change, with BMI moving with it. The long-term group's weight was essentially flat, and more of them held stable weight than the short-term group did.

If you were on metformin briefly, the thing most likely to shift is the scale. If you were on it for years, the thing most likely to shift is your cycle. Neither group got both.

What tends to come back: cycles, androgens, weight — and appetite

Androgens moved, but less than the narrative suggests. Androstenedione crept up in both groups and reached only borderline significance in the long-term users. Testosterone did not show a dramatic rebound.

There's an honest reading of that which cuts against metformin marketing: at the end of the study, high androgen levels were present in 100% of the short-term group and 75% of the long-term group. Metformin had not normalized androgens in most of these women while they were taking it. You can't lose much of a benefit you weren't getting. Metformin is a metabolic drug that helps hormones indirectly — it is not an anti-androgen, and spironolactone works on a different pathway entirely.

The most interesting finding was about eating behavior, measured by questionnaire. Short-term users reported a jump in uncontrolled eating. Long-term users reported a drop in cognitive restraint — the deliberate effort of managing what they ate. Both changes were statistically significant, and both point the same direction: some of what metformin does may run through appetite, not only through glucose metabolism. Which is a plausible explanation for that 9 lb.

On timing: this study reported change over six months, not week by week. Claims elsewhere that symptoms return "within two weeks" or "after three months" aren't coming from measured data in PMOS. Give it a couple of cycles before drawing conclusions.

What stayed stable — the part nobody mentions

Every page on this topic is written as a warning. The data is more mixed than that, and the reassuring half deserves saying. Over six months off metformin, in both groups:

• Insulin resistance (HOMA-IR) did not deteriorate. It stayed flat.
• Waist circumference held steady.
• Fasting glucose stayed in range — it actually fell slightly in the short-term group.
• Nobody developed type 2 diabetes, impaired glucose tolerance, or impaired fasting glucose during the follow-up.

That last one is the finding worth carrying with you. Stopping metformin did not tip anyone into a worse metabolic category within six months, in a group whose median BMI was over 30.

The caveats are real and we'd rather state them than bury them. Forty completers is small. There was no control group, so nothing here proves metformin caused the differences. Six months is short — it says nothing about year three. And a group averaging BMI 32 doesn't necessarily describe lean PMOS. This is the best evidence available on the question, and it is still one modest study.

If you're trying to conceive, or just got a positive test

Different question, higher stakes — and this is where "just stop when you're pregnant" turns out to be shakier than it sounds.

A 2025 systematic review and meta-analysis in the American Journal of Obstetrics & Gynecology pooled 12 trials covering 1,708 women with PCOS, comparing metformin started before conception against placebo or no treatment. Women who continued metformin through the first trimester had higher clinical pregnancy rates (odds ratio 1.57) and a possible reduction in miscarriage. Women who stopped at the positive test also conceived more often than controls — but showed a signal pointing toward more miscarriage, not less. Every indirect comparison leaned the same way.

The confidence intervals are wide and the trials were graded low-to-moderate quality, so this is a trend, not a settled answer; the authors say plainly that better research is needed. What it does mean: if you're on metformin and trying to conceive, what happens at a positive test is a conversation to have in advance, not a default to assume.

Tapering, and what to have in place before you stop

Do you need to taper? There is no trial evidence either way. No study has compared stopping abruptly against stepping down. Advice to taper is reasonable clinical caution rather than a proven benefit — and if you're coming off because of gastrointestinal side effects, reducing the dose is often the thing that's tried before stopping altogether. Extended-release formulations are the other common move for that specific problem.

What's actually worth doing is deciding what you're monitoring before you stop, so you're reading signal instead of anxiety:

• Track your cycles from the month you stop. If you were a long-term user, this is your main variable — and cycle length drifting past 35 days is the change to bring to your clinician.
• Weigh yourself on a fixed schedule if you were a short-term user. Same reason, different variable.
• Know your starting numbers. A fasting glucose, HbA1c, and lipid panel before or around stopping give you something to compare against later. Our guide to what PCOS blood test numbers actually mean covers how to read them.
• Have the replacement in place first, not after. If metformin was doing insulin-sensitivity work, something needs to take that over — most often nutrition and resistance training, sometimes inositol, which we've compared against metformin directly.

One broader point. This is the same shape we found with stopping GLP-1 medications like Ozempic: these drugs manage PMOS while you take them, and they don't retrain the underlying biology. Which is an argument for knowing your own driver — whether insulin, androgens, ovulation, or SHBG is doing the most work in your case — before you change what you're taking. Our free assessment reads your pattern across those drivers in about five minutes. It takes no lab work and it isn't a diagnosis, but it points at which lever is likely to matter most for you.

The 2023 International Evidence-based Guideline sets no fixed duration for metformin in PMOS — no standard length of course, no point at which you're supposed to be finished. That makes stopping a judgment call rather than a milestone, which is exactly why it's worth making deliberately, with your clinician, and with a plan for what you'll watch afterward.

Sources

• Kravos NA, Janež A, Goričar K, Dolžan V, Jensterle M. Effects of metformin withdrawal after long and short term treatment in PCOS: observational longitudinal study. Diabetology & Metabolic Syndrome, 2021 Apr 12. PMID: 33845893. DOI: 10.1186/s13098-021-00660-5
• Cheshire J, Garg A, Smith P, Devall AJ, Coomarasamy A, Dhillon-Smith RK. Preconception and first-trimester metformin on pregnancy outcomes in women with polycystic ovary syndrome: a systematic review and meta-analysis. American Journal of Obstetrics & Gynecology, 2025 Dec;233(6):530–547. PMID: 40473092. DOI: 10.1016/j.ajog.2025.05.038
• Teede HJ, et al. 2023 International Evidence-based Guideline for the Assessment and Management of Polycystic Ovary Syndrome. Monash University / international consensus, 2023.

This article is for information only and is not medical advice. Do not start, stop, or change a prescription based on it — those decisions belong with you and your clinician.