How to Lower Testosterone in Women: What Works, Ranked by Evidence

What actually lowers testosterone, ranked by evidence

Seven options have trial evidence behind them, and that evidence ranges from a 36-trial meta-analysis down to a single 42-woman study — a gap worth knowing before you pick one.

What “too much” means, and why the number alone shouldn’t decide your next move

A total testosterone reliably above your lab’s stated ceiling — commonly somewhere around 45–60 ng/dL — is the usual flag, but two women with the identical number can need completely different next steps. What actually counts as high, and what’s behind it, is covered in full here; the exact reference ranges, in both ng/dL and nmol/L, are in this chart. This page picks up from there: once a cause is identified, here is what actually moves the number.

One thing worth knowing before you start: PCOS was renamed polyendocrine metabolic ovarian syndrome, or PMOS, in May 2026, by a global consensus of more than 50 organisations (Teede et al., 2026). Same condition, same mechanism, same treatments below — only the name changed. This article uses PCOS, since that’s still what most readers search.

The first-line option: combined oral contraceptives

A combined oral contraceptive suppresses ovarian androgen production and raises SHBG enough to outperform metformin on every androgen measure tested in a 2024 meta-analysis of 36 randomized trials. Metformin left free androgen index 7.08 points higher, SHBG 118.6 nmol/L lower, and testosterone 0.48 nmol/L higher than the pill did (Melin et al., 2024). That’s why the 2023 international guideline names a combined pill the first-line pharmacologic option for androgen excess, ahead of metformin, ahead of supplements, ahead of lifestyle change alone. Which pill, and what it treats beyond testosterone, is covered separately.

Combining metformin with the pill goes a step further than the pill alone: the same meta-analysis found the combination improved free androgen index by another 0.58 points and SHBG by another 16.6 nmol/L compared with the pill by itself — a real but smaller gain, worth discussing if you’re already managing insulin resistance alongside androgen excess.

Should you combine options, or try them one at a time?

Combining a pill with metformin measurably beats the pill alone — that extra 0.58-point improvement in free androgen index isn’t hypothetical, it’s the same 36-trial meta-analysis measuring women who took both at once (Melin et al., 2024). In practice, guideline-consistent care rarely waits for one option to plateau before starting the next: a pill goes in first, weight loss and exercise run in parallel from day one since neither interacts badly with it, and an anti-androgen gets added around the six-month mark only if hirsutism or acne hasn’t improved enough on the pill alone. What doesn’t stack usefully is redundancy — nothing in the evidence above suggests that adding a second insulin-sensitiser, or doubling up on unrelated supplements, buys anything beyond what the first one already delivered.

Adding an anti-androgen when the pill isn’t enough

Spironolactone, typically 100–200 mg daily, is added after roughly six months if the pill alone hasn’t done enough, and it needs reliable contraception alongside it — anti-androgens carry a feminisation risk to a male fetus, which rules out starting one alone. Because it works by blocking the androgen receptor at the follicle rather than by lowering circulating testosterone itself, the Endocrine Society guideline puts the timeline at roughly six months for an initial effect and about nine months for the full effect — a hair-growth-cycle limit, not a dosing problem. Finasteride is the second-line option; flutamide is specifically not recommended because of its liver toxicity risk relative to its benefit.

Can you lower testosterone naturally? What weight loss and exercise actually do

Losing more than 5% of body weight lowered free testosterone by 31% — from 77 to 53 pmol/L — in a 24-woman trial that has anchored this claim in PCOS literature for three decades (Kiddy et al., 1992). SHBG rose 54% in the same group, and total testosterone didn’t move significantly — the effect ran entirely through binding, not production. Nobody who lost under 5% of their starting weight in that trial saw the same change, which is the honest ceiling on “just lose some weight”: the threshold matters more than the direction.

Exercise works through a similar mechanism on a faster timeline. In a randomized trial of 87 women with PCOS, 16 weeks of treadmill training — either continuous or in intervals, three sessions a week — significantly reduced testosterone in both groups, and the interval group also saw a significant drop in free androgen index (Ribeiro et al., 2021). Neither weight loss nor exercise here matches the pill’s effect size; both work by raising SHBG and improving insulin sensitivity rather than switching off ovarian androgen production directly — the same mechanism that links insulin resistance to androgen excess more broadly.

What about spearmint tea and inositol?

Spearmint tea lowered both free and total testosterone significantly over 30 days in a two-center trial of 42 women — the only randomized trial of its kind at any scale (Grant, 2010). That’s a real result, but it rests on one small study, not the dozens backing the pill.

Myo-inositol’s case is weaker than its reputation. A meta-analysis of nine trials found only a non-significant trend toward lower testosterone (P=0.099), and SHBG rose significantly only in trials that ran 24 weeks or longer (Unfer et al., 2017). The same meta-analysis found real, significant improvements in fasting insulin and HOMA-IR — inositol has a genuine job in PCOS management, it just isn’t reliably lowering testosterone.

Who this will not work for

None of the above touches the small number of cases where testosterone is elevated for a reason other than PCOS — an androgen-secreting tumor, congenital adrenal hyperplasia, or Cushing’s syndrome need a different workup entirely, not a longer trial of the pill or a stricter diet. If hair growth or acne appeared suddenly, over weeks rather than years, that’s the pattern that needs ruling out first.

Anti-androgens are also off the table for anyone trying to conceive or not using reliable contraception, given the feminisation risk to a male fetus. Weight loss and exercise have little to offer someone who is already lean and active, since the mechanism runs through insulin sensitivity and there’s no insulin resistance left to improve — lean PCOS in particular tends to respond better to the pharmacologic options above than to either one, precisely because insulin resistance isn’t the phenotype’s main driver. Spearmint tea and inositol, honestly, won’t get anyone to the effect size a pill or an anti-androgen delivers; treat them as an addition, not a replacement, if hyperandrogenism symptoms are significant. And nothing here reverses hair or acne changes that have already happened — that needs its own treatment, on its own timeline.

Your next step

Confirm the cause before you pick a treatment — a random total testosterone plus SHBG, read against how fast your symptoms actually appeared, is what a clinician uses to rule out the rare causes first. Once PCOS is confirmed, a combined oral contraceptive is the option with the most evidence behind it; weight loss, exercise, and inositol are worth layering in alongside it, not instead of it, if insulin resistance is part of your picture.