Does Metformin Help PCOS Acne and Facial Hair? What the Trials Found

Does metformin actually clear PCOS acne and facial hair?

In a 2020 Cochrane review pooling 44 randomized trials, the same drug that reliably improves insulin and ovulation showed no consistent benefit for acne or facial hair, and on the one comparison with enough data to pool, metformin was found less effective than the birth control pill for acne. That is not the story most searches on this topic tell. Metformin’s insulin-lowering effect is real and measured — see what metformin actually does for insulin, weight and fertility for that side of it — but insulin resistance is only one input into androgen-driven skin and hair symptoms, and it is not the input metformin reaches most directly. The androgen receptor sitting in a hair follicle or a sebaceous gland responds to local hormone activity that an insulin-sensitizing drug barely touches, which is the mechanical reason the effect on visible symptoms is smaller and less consistent than the effect on a fasting insulin number.

You may see PCOS written as polyendocrine metabolic ovarian syndrome (PMOS) after a May 2026 global consensus of more than 50 organizations renamed it. Nothing about metformin’s mechanism or the trial evidence below changed with the name. This article uses PCOS, since that is still what most readers search.

What do the actual trials show for hirsutism?

Ten randomized trials covering 473 women compared metformin directly against the combined pill on the Ferriman-Gallwey (FG) hirsutism score, and metformin never won a body-size group outright. The Cochrane team split the results by BMI because the trials disagreed too much to pool as one number, and even split apart, the picture is not encouraging for metformin.

That last row comes from a separate 2018 network meta-analysis of hirsutism treatments — not exclusively in PCOS, though PCOS accounts for most cases — which found insulin-sensitizing drugs as a class produced a standardized mean reduction of 0.62 against placebo, compared with 1.29 for antiandrogens and 0.94 for the pill. Metformin was the insulin sensitizer used in most of those trials, and it landed clearly behind the two hormonal options, not just behind them by a small margin.

One older, smaller trial adds a genuine wrinkle. A 14-month placebo-controlled crossover study found a statistically significant FG score improvement on metformin — 15.8 versus 17.5 on placebo (p=0.025) in the 16 women enrolled, 10 of whom completed it. That is a real, p-value-backed result. It is also a 1.7-point shift on a 36-point scale, in a trial too small to be the last word on anything, which is exactly why the guideline below rates this evidence “very low certainty” rather than building a strong recommendation on it.

What does the 2023 international guideline say?

Three numbered recommendations in the 2023 international PCOS guideline name hirsutism specifically when addressing metformin’s place, and every one of them carries the lowest certainty rating the guideline uses. The guideline states plainly that “COCP could be used over metformin for management of hirsutism in irregular menstrual cycles in PCOS” — COCP meaning the combined oral contraceptive pill — a conditional recommendation built on evidence graded “very low certainty.” Where the pill is not an option, the same guideline is explicit about what that leaves you with: metformin may help irregular cycles, but “for hirsutism, other interventions may be needed.”

The guideline does give metformin one comparative win: it says metformin should be preferred over inositol for hirsutism and central adiposity, again on very-low-certainty evidence and noting metformin causes more gastrointestinal side effects. Read together, the guideline’s own hierarchy for hirsutism specifically runs pill, then metformin, then inositol — with every rung resting on weak evidence, and none of it strong enough to be called a first-line hirsutism treatment.

Does metformin help acne specifically?

Only 6 of the 44 trials in that same Cochrane review even measured acne, and the one comparison with enough data to pool found metformin performed worse than the pill — an odds ratio of 0.30 (95% CI 0.11 to 0.79) across 3 trials and 131 women, meaning women on metformin were meaningfully less likely to report acne improvement than women on the pill. A second, smaller comparison using a 0–10 self-rated scale in 34 women found no clear difference either way. One more trial, in 28 women, reported an acne improvement in its discussion section but never published the underlying numbers, so Cochrane could not include it in any pooled result at all. Among adolescents specifically, no randomized trial has ever reported an acne outcome for metformin versus the pill — zero data, not weak data.

Set against that, a 2021 meta-analysis pooling 51 studies and 2,405 patients found metformin associated with improved acne scores — a modest but statistically significant effect (SMD −0.26 as an add-on therapy, SMD −0.71 comparing before and after treatment). The gap between that positive-looking number and Cochrane’s negative head-to-head result is not a contradiction so much as a difference in what was pooled: most of those 51 studies were open-label or non-randomized, exactly the design that inflates apparent benefit when neither patients nor assessors are blinded. The few actual randomized trials that compared metformin against a real alternative, blinded or not, did not confirm the larger meta-analysis’s optimism — and on the one outcome with a clear statistical result, they found the opposite direction.

Why doesn’t a testosterone drop show up on your skin?

A testosterone drop on a lab report and a testosterone drop a follicle can feel are two different things, and one 2003 trial showed exactly that gap in 52 women. That 12-month trial compared metformin against a hormonal pill combining an antiandrogen, and reported that metformin produced “negligible” change in androgen markers while the hormonal pill caused “profound suppression” of androgen activity — yet on the visible hirsutism measures, metformin performed comparably to, and by some measures better than, the hormonal treatment. Skin and hair follicles carry their own local enzyme machinery that converts and responds to androgens somewhat independently of what is circulating in blood, which is part of why a treatment’s effect on a hormone panel and its effect on what you see in the mirror do not always move together, in either direction. The full mechanism behind PCOS acne and what actually slows PCOS-pattern chin hair both cover this local-conversion detail in more depth.

How long would you actually have to wait to know?

Every metformin-hirsutism trial with a defined endpoint ran 6 months or longer — Kelly and Gordon used two 6-month treatment phases, and the 2003 metformin-versus-pill trial assessed at 6 and 12 months — because a hair follicle needs a full growth cycle before a hormonal change becomes visible. Nobody has shown a faster timeline for metformin’s effect on hirsutism or acne, and there is no version of “give it six weeks” that any of this evidence supports. If six months pass with no change in the specific symptom you are tracking, that absence of change is itself a data point worth bringing back to whoever prescribed it — not a sign you stopped too soon.

Who this will not help

None of the 10 hirsutism trials or 6 acne trials above enrolled women by insulin-resistance status, so metformin cannot be assumed to work better for these specific symptoms in the insulin-resistant phenotype — that reasoning holds for the metabolic picture, not for skin and hair. It will not help someone expecting a result inside weeks: no trial has ever shown one. It will not help someone whose main goal is clearing acne or facial hair and who is not also getting a metabolic benefit worth the gastrointestinal trade-off, since the one acne comparison with real numbers behind it favored a different option. And no randomized trial comparing metformin against the pill in adolescents has ever reported an acne outcome, which means a teenager and her prescriber are working from the adult data above, extrapolated, not from a trial in her own age group.

What has stronger trial evidence, if skin and hair are the main goal

Two prescription options outperformed metformin on these exact outcomes in the trials above: the combined pill, and — when added to it — an antiandrogen such as spironolactone. Both carry their own cautions that travel with the recommendation, not just with the drug’s own page: the combined pill is avoided with a history of blood clots, migraine with aura, or smoking past 35, and spironolactone requires a baseline and follow-up potassium check and is not safe in pregnancy, since it carries a teratogenicity risk to a male fetus. Neither of those cautions makes either option automatically right for you — that is a conversation for whoever prescribes it, weighed against your own history — but the guideline and the head-to-head trials above agree that these are the better-evidenced choices when hirsutism or acne, specifically, is the symptom you are treating. The full range of hirsutism treatment options and timelines covers the pill-plus-antiandrogen combination, laser, and eflornithine in the detail this article does not repeat, and the wider supplements and medications section covers where metformin fits alongside them.

Your next step

If acne or facial hair is the symptom that actually brought you here, track it for the same 6 months any hirsutism trial would use before judging whether metformin is doing anything for it — a dated photo under consistent lighting is more reliable than memory. Bring that record, along with your most recent insulin and cycle data, to your next appointment, since those are the outcomes the trial evidence actually supports metformin moving.