Keto for PCOS: What the Trials Show and Who Regains It

Keto has more PCOS-specific trial evidence behind it than almost any other named diet on this site — two small studies, both with real numbers, both short. That is a genuinely different starting point from “no evidence,” and it is still a long way from “settled.” PCOS was renamed polyendocrine metabolic ovarian syndrome (PMOS) in 2026 by a consensus of more than 50 organisations (Lancet 2026); the biology below is unchanged, and this article uses PCOS because that is still what people search.

Does Keto Actually Improve PCOS Hormones?

Yes, in the one 12-week trial built to measure it. Fourteen women with PCOS who completed a ketogenic Mediterranean protocol saw total testosterone fall 14% (47.4 to 40.7 ng/dL), the LH/FSH ratio fall 43% (2.00 to 1.15), and sex hormone-binding globulin rise 30% (Paoli et al., 2020). HOMA-IR, the standard insulin resistance marker, also improved significantly, and body weight fell 11.6% over the 12 weeks.

What Does the Evidence Actually Show, Trial by Trial?

Both dedicated PCOS keto trials are small, short, and lost close to half their participants — numbers worth seeing side by side rather than averaged away.

Neither trial had a comparison-diet control group running alongside it, so the changes are measured against each group’s own starting point, not against an alternative diet over the same weeks. That matters for how much weight to put on the size of the effect — real movement, unproven superiority.

The Paoli trial also reported numbers rarely mentioned in keto-for-PCOS content: fasting glucose fell from 5.10 to 4.64 mmol/L, insulin from 12.62 to 11.31 μU/mL, and — notably — both estradiol (139.8 to 191.9 pg/mL) and progesterone (12.2 to 21.1 ng/dL) rose significantly over the 12 weeks (Paoli et al., 2020). Rising estradiol and progesterone alongside a falling LH/FSH ratio is the pattern you would expect if cycling improved, even though the trial did not report cycle data directly to confirm it.

Does Keto Beat Simple Calorie Restriction for PCOS?

The 12-week keto trial and an 8-week very low-calorie diet trial in PCOS land within one percentage point of each other on weight — 11.6% versus 10.9% — despite reaching it through opposite mechanisms. The 24-week keto pilot produced a similar 12% loss (Mavropoulos et al., 2005). None of these three trials ran alongside each other, so this is not head-to-head evidence — it is three separate small studies landing on a similar order of magnitude of weight change, reached through different routes (ketosis versus raw energy deficit), each carrying its own separate hormonal side effects and its own separate risk profile.

Is Low-Carb the Same Thing as Keto for PCOS?

No — the two sit at different points on one spectrum, and only the strict end has PCOS-specific trial data. A ketogenic diet in the trials above meant 20 grams of carbohydrate a day or a comparably strict protocol, producing measurable ketosis. The broader meta-analysis literature defines a “very-low-carbohydrate ketogenic diet” as 50 grams or less (Bueno et al., 2013), while ordinary “low-carb” eating — cutting refined grains and sugar without chasing ketosis — is a milder version with its own, separate evidence base and none of the specific PCOS hormone data above.

How Many People Actually Stick With Keto?

Close to half — even over 12 weeks, the shortest of the two PCOS trials, 10 of 24 women did not finish (Paoli et al., 2020); in the 24-week pilot, 6 of 11 dropped out, citing dietary non-compliance (Mavropoulos et al., 2005). A 2023 scoping review of ketogenic diet interventions found the same pattern across the wider obesity and type 2 diabetes literature: adherence is consistently the weak point, and studies measure it so inconsistently that comparing “how many people actually stay in ketosis” across trials is difficult (Li et al., 2023).

The practical reading: if roughly half the people in a supervised, closely monitored research trial could not sustain 20 grams of carbohydrate a day for 12 to 24 weeks, unsupervised long-term adherence is the real question to answer before starting — not whether the diet can work on paper.

What Happens to Cholesterol on Keto?

It depends heavily on body weight, and this is the part most keto-for-PCOS content skips entirely. Pooled across 13 randomised trials with a year or more of follow-up, a very-low-carbohydrate ketogenic diet raised LDL cholesterol by a modest 0.12 mmol/L on average while also raising HDL and lowering triglycerides (Bueno et al., 2013). That average hides a large split by starting weight: a 2024 meta-analysis of 41 trials found LDL cholesterol rose by 41 mg/dL on average in people with a starting BMI under 25, while it fell in people with a starting BMI of 35 or above.

If your PCOS presentation runs lean rather than insulin-resistant, this table is the single most useful piece of information on this page: the same diet that most reliably helps a higher-BMI metabolic profile is the one most likely to raise LDL cholesterol in a lower-BMI body. A lipid panel before and a few months into any strict low-carb diet is a reasonable ask regardless of which group you fall into.

Neither trial cited above reported fibre intake, which matters because cutting legumes, whole grains and most fruit — the standard route to 20 grams of carbohydrate a day — removes the main sources of it from a typical plate at the same time. Non-starchy vegetables, avocado, chia, flax and nuts are the fibre sources that survive a ketogenic carbohydrate limit, and building a keto plate around them rather than around meat and cheese alone is a reasonable way to protect gut health while the hormone changes above play out.

Who Should Not Try Keto for PCOS?

Anyone with a history of disordered eating should not start keto without a treating clinician. A diet that already loses 42% to 55% of participants to ordinary non-compliance in supervised trials is a considerably higher-risk structure to layer onto disordered eating patterns specifically, given how rules-based and restrictive it is. Pregnancy and breastfeeding are excluded from every trial cited on this page and from the diet in practice: energy and nutrient needs in both states run counter to a carbohydrate-restricted protocol, and no PCOS-specific keto data exists in either population.

A few other situations are worth a conversation with a clinician before starting: a personal or family history of high LDL cholesterol or early cardiovascular disease, given the lipid variability above; gallbladder disease or a history of kidney stones, both more common on very-low-carbohydrate diets generally; and use of any glucose-lowering medication, since a sharp drop in carbohydrate intake changes dosing needs.

What This Means If You Try It

The evidence supports keto as a real option with real trade-offs, not a diet to sell or dismiss. Two small trials showing meaningful hormone shifts is more than most named diets in PCOS content can claim, and a roughly 50% dropout rate even under research supervision is exactly the number to plan around rather than ignore. If the dropout risk or the lipid picture above rules keto out, a more moderate calorie-focused approach has its own trial evidence and its own honest limits, and the broader diet evidence beyond any single named plan moves several of the same markers without the ketosis requirement. For a sense of which individual foods actually have evidence behind avoiding them — keto or not — see the honest list. The mechanism keto is usually reached for is covered directly here, and the rest of the diet section has options that do not require sustained ketosis.