Evidence-graded guideReviewed October 31, 2026

Herbs and Supplements for Blood Sugar: What the Trials Actually Show

The herbs and supplements with the best evidence for lowering blood sugar are berberine, psyllium husk and cinnamon, followed by chromium, fenugreek and magnesium. Berberine is the standout: in head-to-head trials it lowered HbA1c by about as much as metformin, and pooled trials show fasting glucose falling roughly 15 mg/dL (0.8 mmol/L). Psyllium taken before meals lowered HbA1c by close to 1 percentage point in people with type 2 diabetes. Everything else on this list produces smaller, less consistent changes, and several popular picks (bitter melon, vitamin D for people who are not deficient) do not hold up.

This guide is for adults with prediabetes, insulin resistance, PCOS or diet-controlled type 2 diabetes who want to add something evidence-based to the things that matter most: weight, fibre, movement and sleep. It is also for people already on metformin who want to know what is safe to combine. It is not a substitute for medication. If you have type 1 diabetes, are pregnant, have a fasting glucose above 250 mg/dL (14 mmol/L), or have symptoms such as thirst, frequent urination, blurred vision or unexplained weight loss, see a clinician first. Supplements do not treat ketoacidosis and do not replace insulin.

One rule above all: never stop or reduce a prescribed diabetes medicine because a supplement is working. Several of these (berberine, gymnema, alpha-lipoic acid, psyllium) genuinely lower glucose, which means they can push you into hypoglycaemia when stacked on insulin, sulfonylureas or glinides. Test more often when you start, keep a log, and let your prescriber adjust the drug.

Quick answer

  • Berberine — lowered HbA1c about as much as metformin in a 3-month trial; fasting glucose down ~15 mg/dL in pooled RCTs
  • Psyllium husk — 10 g/day before meals cut HbA1c by ~1 point in type 2 diabetes (meta-analysis of 35 trials)
  • Cinnamon — fasting glucose down ~24 mg/dL across 10 RCTs, but HbA1c did not change significantly
  • Chromium — HbA1c down ~0.6 points in people with diabetes; no effect in people with normal glucose
  • Fenugreek — seed powder lowered fasting glucose ~17 mg/dL and HbA1c ~0.85 points in diabetics (10 trials)
  • Magnesium — improves fasting glucose in people who are low in it, which is common in type 2 diabetes

The 10 best herbs and supplements for blood sugar

Ranked by the strength of human evidence. Strong = multiple trials or a meta-analysis. Moderate = a few small trials. Preliminary = early or animal data.

  1. Berberine

    Strong evidence

    The strongest plant compound for glucose. In Yin's 2008 pilot trial, 500 mg three times daily for 3 months lowered HbA1c from 7.5% to 6.6%, a drop statistically indistinguishable from metformin in the comparison arm. A 2015 meta-analysis of 27 trials found fasting glucose fell by about 0.8 mmol/L (15 mg/dL) and HbA1c by roughly 0.7 points versus control, with added benefit when combined with standard drugs. It works partly through AMPK activation and partly through the gut, and about a third of people get constipation or loose stools at the start.

    Key evidence
    2015 meta-analysis of 27 RCTs (Lan et al.); 2008 head-to-head pilot vs metformin (Yin et al.)
    Studied dose
    500 mg of berberine HCl 2–3 times daily with meals, 8–12 weeks before judging
    Best for
    Prediabetes, insulin resistance and type 2 diabetes where metformin is not tolerated or not yet prescribed
    Cautions
    Inhibits CYP3A4 and CYP2D6, raising levels of cyclosporine, some statins and many other drugs; adds to the glucose-lowering of insulin and sulfonylureas. Avoid in pregnancy, breastfeeding and newborns (it displaces bilirubin).
    Full Berberine profile

  2. Psyllium husk

    Strong evidencePlantago ovata

    A 2015 meta-analysis of 35 randomised trials found that psyllium lowered glucose in proportion to how high it was: in people with type 2 diabetes, about 10 g/day cut fasting glucose by 37 mg/dL (2.1 mmol/L) and HbA1c by 0.97 points, while people with normal glucose saw no change. The viscous fibre slows carbohydrate absorption, so timing matters: it only helps if taken shortly before meals. It also lowers LDL by about 6%, which is relevant because most people with high glucose have high lipids too.

    Key evidence
    2015 meta-analysis of 35 RCTs in euglycaemic, at-risk and type 2 diabetic adults (Gibb et al., Am J Clin Nutr)
    Studied dose
    5 g of psyllium husk in 250 mL water, 10–15 minutes before 2 main meals (10 g/day), ongoing
    Best for
    Type 2 diabetes with high LDL or constipation; anyone who wants a safe add-on to metformin
    Cautions
    Take with at least a full glass of water and space it 2 hours from oral medicines (thyroid hormone, lithium, carbamazepine), because it can reduce their absorption. Not for anyone with swallowing problems or bowel narrowing.
    Full Psyllium husk profile

  3. Cinnamon

    Moderate evidenceCinnamomum verum / Cinnamomum cassia

    The 2013 Annals of Family Medicine meta-analysis of 10 RCTs (543 patients with type 2 diabetes, doses 120 mg to 6 g/day for 4 to 18 weeks) found cinnamon lowered fasting glucose by 24.6 mg/dL, total cholesterol by 15.6 mg/dL and triglycerides by 29.6 mg/dL, but HbA1c did not change significantly. That pattern (fasting glucose moves, the 3-month marker does not) is why it sits below berberine and psyllium. Results vary a lot between trials; the larger effects came from studies using 1 to 3 g of cassia bark.

    Key evidence
    2013 meta-analysis of 10 RCTs, 543 patients (Allen et al., Ann Fam Med)
    Studied dose
    1–3 g/day of Ceylon cinnamon powder, or a water extract standardised to polyphenols, 8–12 weeks
    Best for
    Mildly raised fasting glucose in someone who wants a food-level intervention
    Cautions
    Cassia cinnamon contains coumarin, which can stress the liver above about 1 teaspoon (3 g) a day for long periods; choose Ceylon for daily use. May add to the effect of glucose-lowering drugs.
    Full Cinnamon profile

  4. Chromium picolinate

    Moderate evidence

    Chromium is part of the insulin signalling machinery, and trials split cleanly by who takes it. A 2014 meta-analysis of 25 RCTs in people with diabetes found HbA1c fell by about 0.6 points and fasting glucose by about 1 mmol/L (18 mg/dL), with picolinate at 200 mcg or more outperforming other forms. Trials in people with normal glucose show nothing, which is why the NIH concludes it does not help healthy adults. Responders tend to be people who were marginal in chromium to begin with.

    Key evidence
    2014 meta-analysis of 25 RCTs in diabetes (Suksomboon et al., J Clin Pharm Ther)
    Studied dose
    200–600 mcg/day of chromium picolinate with food, 12–16 weeks
    Best for
    Type 2 diabetes with poor control despite medication; people on long-term low-chromium diets
    Cautions
    Can increase the effect of insulin and sulfonylureas; rare reports of kidney injury at very high doses. Avoid if you have kidney disease without medical advice, and keep below 1,000 mcg/day.
    Full Chromium picolinate profile

  5. Fenugreek seed

    Moderate evidenceTrigonella foenum-graecum

    Fenugreek seed is about half soluble fibre (galactomannan) and also contains 4-hydroxyisoleucine, which stimulates insulin release. A 2014 meta-analysis of 10 trials found fasting glucose fell by 0.96 mmol/L (17 mg/dL), 2-hour glucose by 2.2 mmol/L and HbA1c by 0.85 points in people with diabetes, but not in healthy volunteers. The catch is that most of the effect came from Indian trials using large doses of whole seed powder (5 to 25 g/day baked into food), and extract capsules have less support.

    Key evidence
    2014 meta-analysis of 10 trials, mostly in type 2 diabetes (Neelakantan et al., Nutrition Journal)
    Studied dose
    5–10 g/day of ground seed with meals, or 1 g/day of a hydroalcoholic extract, 8–12 weeks
    Best for
    Type 2 diabetes in someone willing to cook with the seed; high triglycerides as well as high glucose
    Cautions
    Can enhance warfarin and other anticoagulants, and causes a maple-syrup body odour that can be mistaken for a metabolic disorder in children. Avoid in pregnancy (uterine stimulant); it is related to peanuts and chickpeas, so beware of allergy.
    Full Fenugreek seed profile

  6. Magnesium

    Moderate evidence

    Low magnesium impairs insulin signalling, and up to a third of people with type 2 diabetes are deficient because high glucose makes the kidneys waste it. A 2016 meta-analysis of 18 double-blind trials in people with or at risk of diabetes found supplementation lowered fasting glucose by about 0.3 mmol/L (5 mg/dL) after 3 months and improved insulin resistance in those at risk. The effect is modest and concentrated in people who start low, so it is worth checking a serum magnesium level, remembering that serum underestimates true deficiency.

    Key evidence
    2016 meta-analysis of 18 RCTs in diabetics and at-risk adults (Veronese et al., Eur J Clin Nutr)
    Studied dose
    250–400 mg/day of elemental magnesium as glycinate or citrate, with food, for at least 12 weeks
    Best for
    Type 2 diabetes or prediabetes with cramps, poor sleep, or a low or low-normal serum magnesium
    Cautions
    Reduce the dose if you have kidney disease, because magnesium accumulates. Separate by 2 hours from quinolone and tetracycline antibiotics, bisphosphonates and levothyroxine; high doses cause diarrhoea.
    Full Magnesium profile

  7. Alpha-lipoic acid

    Moderate evidence

    Alpha-lipoic acid has its firmest evidence for diabetic nerve pain: in the SYDNEY 2 trial, 600 mg/day orally for 5 weeks halved neuropathy symptom scores versus placebo in 181 patients. On glucose itself the picture is weaker: a 2018 meta-analysis of 20 trials found fasting glucose and HbA1c fell by small but significant amounts, and insulin sensitivity improved by around 20% in intravenous studies. It belongs here mainly for people whose diabetes has already reached their nerves.

    Key evidence
    SYDNEY 2 RCT (181 patients, Diabetes Care 2006); 2018 meta-analysis of 20 trials on glycaemic markers
    Studied dose
    600 mg/day of R- or racemic alpha-lipoic acid on an empty stomach, 4–12 weeks
    Best for
    Type 2 diabetes with burning, tingling or numb feet
    Cautions
    Can cause hypoglycaemia alongside insulin or sulfonylureas; may reduce thyroid hormone conversion, so time it away from levothyroxine. Rare cases of insulin autoimmune syndrome, mostly in East Asian patients.
    Full Alpha-lipoic acid profile

  8. Gymnema sylvestre

    Moderate evidenceGymnema sylvestre

    Gymnema's gymnemic acids block sweet-taste receptors on the tongue and gut and appear to support insulin secretion. The best-known trial gave 400 mg/day of a GS4 extract to 22 people with type 2 diabetes for 18 to 20 months alongside their drugs: HbA1c fell from 11.9% to 8.5% and five patients were able to stop their sulfonylurea, but there was no placebo arm. A 2021 pooled analysis of about 10 small trials found meaningful reductions in fasting glucose and HbA1c, with the caveat that trial quality was low.

    Key evidence
    Baskaran 1990 open-label 18–20 month trial (n=22); 2021 pooled analysis of small trials
    Studied dose
    200–400 mg/day of a leaf extract standardised to 25% gymnemic acids, 12 weeks minimum
    Best for
    Type 2 diabetes with strong sugar cravings (it dulls sweet taste for about an hour)
    Cautions
    Lowers glucose additively with insulin and sulfonylureas, so monitor closely. No safety data in pregnancy; stop 2 weeks before surgery.
    Full Gymnema sylvestre profile

  9. Myo-inositol

    Moderate evidence

    Myo-inositol is a second messenger in insulin signalling, and its evidence is specific to PCOS and gestational diabetes. In PCOS trials, 2 g twice daily for 12 to 24 weeks lowered fasting insulin and HOMA-IR, restored ovulation in roughly half of women, and performed similarly to metformin with fewer side effects. In pregnant women with a family history of diabetes, 4 g/day from the first trimester cut the rate of gestational diabetes from 15.3% to 6% in one 220-woman trial. Evidence in ordinary type 2 diabetes is thin.

    Key evidence
    RCT of 220 pregnant women (D'Anna et al., Diabetes Care 2013); multiple PCOS RCTs vs metformin
    Studied dose
    2 g of myo-inositol twice daily (often with 50 mg D-chiro-inositol per 2 g), 12–24 weeks
    Best for
    PCOS with insulin resistance; women at high risk of gestational diabetes, under obstetric care
    Cautions
    Mild nausea and loose stools at doses above 12 g/day. Use in pregnancy only with your obstetrician's knowledge; it can enhance the effect of other insulin sensitisers.
    Full Myo-inositol profile

  10. White mulberry leaf

    Preliminary evidenceMorus alba

    Mulberry leaf contains 1-deoxynojirimycin (DNJ), which blocks the alpha-glucosidase enzymes that break starch and sucrose into glucose, the same target as the drug acarbose. In a 2017 double-blind crossover trial, 250 mg of a standardised extract taken with 50 g of maltodextrin reduced the post-meal glucose peak by about 40% and the insulin peak by around 40% in 37 healthy adults. Longer trials in type 2 diabetes are small and short, so it is a tool for post-meal spikes rather than a proven HbA1c reducer.

    Key evidence
    2017 double-blind crossover RCT of a 5% DNJ extract, n=37 (Lown et al., PLoS One)
    Studied dose
    250 mg of a leaf extract standardised to 5% DNJ immediately before carbohydrate-rich meals
    Best for
    People who see large glucose spikes after rice, bread or pasta on a continuous glucose monitor
    Cautions
    Causes gas and bloating in some people because undigested carbohydrate reaches the colon. Additive with acarbose and other glucose-lowering drugs; no pregnancy data.
    Full White mulberry leaf profile

How to choose

  • If you have type 2 diabetes or prediabetes and want the biggest effect from one supplement, start with berberine 500 mg two to three times daily with meals and re-check fasting glucose after 8 weeks.
  • If you are already on metformin or do not tolerate berberine's gut effects, psyllium 5 g before two meals is the safest add-on and also lowers LDL.
  • If your fasting glucose is only slightly raised (100–125 mg/dL) and you want a food-level change, cinnamon, fenugreek and extra fibre are reasonable; do not expect HbA1c to move much.
  • If your main problem is spikes after meals, mulberry leaf or psyllium before meals targets that directly; a continuous glucose monitor will tell you within a week whether it works for you.
  • If you have PCOS, myo-inositol is the first choice; if you have diabetic nerve pain, alpha-lipoic acid is.
  • Bitter melon, banaba and vitamin D are in our library but did not earn a place above: see the skip list for why, and ask AskMN if you want the details on any of them.

What to skip

Bitter melon (Momordica charantia)

The 2012 Cochrane review of four trials found no significant effect on fasting glucose or HbA1c versus placebo, and a 4-week trial found 2 g/day much weaker than metformin. Later meta-analyses show a small fasting-glucose drop from low-quality trials. It can also cause hypoglycaemia and stomach upset.

Vitamin D if you are not deficient

The D2d trial gave 4,000 IU/day to 2,423 adults with prediabetes for 2.5 years and did not significantly reduce progression to diabetes overall. A 2023 pooled analysis of three trials found a 15% relative reduction, which works out to about 3 fewer cases per 100 people over 3 years. Worth correcting a low level, not worth taking for glucose on its own.

Banaba leaf (corosolic acid)

The trials are tiny, short and mostly from a single manufacturer; a typical result is a 10% fall in fasting glucose over 2 weeks at 10 mg corosolic acid, with no HbA1c data. Not worth paying extra for in a blend.

Proprietary 'sugar blocker' blends

Most combine six to ten of the ingredients above at a fraction of the studied dose of each. If a label lists 50 mg of berberine, it is a tenth of a working dose. Buy single ingredients at the dose the trials used.

Safety and when to see a clinician

  • Never stop or lower a prescribed diabetes medicine on your own because a supplement is working. Bring your glucose log to your prescriber and let them adjust it.
  • Hypoglycaemia risk is real when berberine, gymnema, alpha-lipoic acid, psyllium or fenugreek are combined with insulin, sulfonylureas (gliclazide, glipizide, glimepiride) or glinides. Test before meals and at bedtime for the first 2 weeks, and know the symptoms: shaking, sweating, confusion.
  • Berberine interacts with many drugs through CYP3A4 and CYP2D6, including cyclosporine, some statins, calcium-channel blockers and certain antidepressants; check with a pharmacist before combining.
  • Psyllium and magnesium reduce the absorption of levothyroxine, some antibiotics and lithium; separate doses by at least 2 hours.
  • Pregnancy: skip berberine, fenugreek, gymnema and bitter melon. Myo-inositol and magnesium are the only items here with pregnancy data, and only under obstetric care.
  • See a clinician urgently if you have fasting glucose above 250 mg/dL (14 mmol/L), vomiting, abdominal pain, fruity breath or fast breathing (possible ketoacidosis), or if you are newly diagnosed and have not yet had a full assessment.

What matters as much as any herb

  • Weight loss of 5–7% of body weight cut the rate of progression from prediabetes to diabetes by 58% in the Diabetes Prevention Program, far more than any supplement here.
  • A 10–15 minute walk after each meal lowers the post-meal glucose rise by around 20–30% because contracting muscle takes up glucose without needing insulin.
  • Eating vegetables and protein before the carbohydrate in a meal blunts the glucose peak by up to 40% in small crossover studies: a free, drug-free version of what mulberry leaf and psyllium do.
  • Sleeping under 6 hours for even a few nights raises insulin resistance measurably; fixing sleep often improves fasting glucose more than a new supplement.
  • Resistance training twice a week builds the muscle that stores glucose; it improves HbA1c by about 0.5 points in meta-analyses of people with type 2 diabetes.

Frequently asked questions

What is the best herb to lower blood sugar?

Berberine has the most and strongest human trial data: around 0.7 to 0.9 points off HbA1c and 15 mg/dL off fasting glucose, comparable to metformin in small head-to-head trials. It is a drug-like compound with real interactions, so treat it like one. Psyllium is the safest second choice.

Can berberine replace metformin?

No. The head-to-head trials are small, short and did not measure what matters most, which is long-term complications. Metformin has decades of outcome data, including fewer heart attacks and deaths. Berberine is reasonable as an add-on, or for people who cannot take metformin, but switching should be a decision made with your prescriber.

How long do herbs take to lower blood sugar?

Fasting glucose usually responds within 2 to 4 weeks if it is going to. HbA1c reflects the previous 3 months, so re-test it 12 weeks after starting. If nothing has moved by then at a studied dose, stop and try something else rather than stacking more products.

Does cinnamon really lower blood sugar?

It lowers fasting glucose by about 24 mg/dL on average across 10 trials, but the 3-month marker, HbA1c, did not change significantly. So it helps a little, mostly in people with mildly raised glucose. Use Ceylon rather than cassia for daily doses above a teaspoon because of coumarin.

What supplements help insulin resistance and PCOS?

Myo-inositol at 2 g twice daily has the most PCOS-specific evidence, improving insulin resistance and ovulation about as well as metformin. Berberine has also been trialled in PCOS with similar results on insulin and lipids. Magnesium is worth adding if your level is low.

Can I take blood sugar herbs with metformin?

Psyllium, cinnamon, magnesium and chromium combine safely with metformin. Berberine can be combined but both cause gut side effects, so start berberine at one dose a day. The dangerous combinations are with insulin and sulfonylureas, where low blood sugar becomes likely.

Are there herbs that lower blood sugar quickly?

Nothing here is a rescue treatment. Mulberry leaf and psyllium taken immediately before a meal blunt the spike from that meal within the hour, which is the closest thing to a fast effect. Very high glucose with symptoms needs medical care, not a supplement.

References

  1. Yin J, Xing H, Ye J. Efficacy of berberine in patients with type 2 diabetes mellitus. Metabolism (2008)
  2. Lan J, Zhao Y, Dong F, et al.. Meta-analysis of the effect and safety of berberine in the treatment of type 2 diabetes mellitus, hyperlipemia and hypertension. Journal of Ethnopharmacology (2015)
  3. Gibb RD, McRorie JW, Russell DA, et al.. Psyllium fiber improves glycemic control proportional to loss of glycemic control: a meta-analysis of data in euglycemic subjects, patients at risk of type 2 diabetes mellitus, and patients being treated for type 2 diabetes mellitus. American Journal of Clinical Nutrition (2015)
  4. Allen RW, Schwartzman E, Baker WL, et al.. Cinnamon use in type 2 diabetes: an updated systematic review and meta-analysis. Annals of Family Medicine (2013)
  5. Suksomboon N, Poolsup N, Yuwanakorn A. Systematic review and meta-analysis of the efficacy and safety of chromium supplementation in diabetes. Journal of Clinical Pharmacy and Therapeutics (2014)
  6. NIH Office of Dietary Supplements. Chromium: Fact Sheet for Health Professionals (2022)View source
  7. Neelakantan N, Narayanan M, de Souza RJ, van Dam RM. Effect of fenugreek (Trigonella foenum-graecum L.) intake on glycemia: a meta-analysis of clinical trials. Nutrition Journal (2014)
  8. Veronese N, Watutantrige-Fernando S, Luchini C, et al.. Effect of magnesium supplementation on glucose metabolism in people with or at risk of diabetes: a systematic review and meta-analysis of double-blind randomized controlled trials. European Journal of Clinical Nutrition (2016)
  9. Ziegler D, Ametov A, Barinov A, et al.. Oral treatment with alpha-lipoic acid improves symptomatic diabetic polyneuropathy: the SYDNEY 2 trial. Diabetes Care (2006)
  10. Ooi CP, Yassin Z, Hamid TA. Momordica charantia for type 2 diabetes mellitus. Cochrane Database of Systematic Reviews (2012)
  11. D'Anna R, Scilipoti A, Giordano D, et al.. Myo-inositol supplementation and onset of gestational diabetes mellitus in pregnant women with a family history of type 2 diabetes: a prospective, randomized, placebo-controlled study. Diabetes Care (2013)
  12. Pittas AG, Kawahara T, Jorde R, et al.. Vitamin D and risk for type 2 diabetes in people with prediabetes: a systematic review and meta-analysis of individual participant data from 3 randomized clinical trials. Annals of Internal Medicine (2023)

Related conditions

More guides

Not sure which one fits you?

Ask AskMN. It knows these studies, checks interactions with the medications you tell it about, and it is free and unlimited, no account needed.

Ask about blood sugar

This guide is educational and is not medical advice. Herbs and supplements can interact with prescription medicines and are not regulated like drugs. Talk with a clinician or pharmacist before starting anything new, especially if you are pregnant, breastfeeding, have a chronic condition, or take medication.