What is the best supplement to lower blood sugar?

Berberine and psyllium fiber have the strongest and most consistent trial evidence. Berberine lowered HbA1c by about 0.73 points across 46 trials in type 2 diabetes; psyllium's drop was larger, about 0.97 points, but only in people with treated, high glucose. Neither matches a structured lifestyle program or metformin for long-term evidence.

Can supplements replace metformin?

No. Metformin lowers HbA1c by about 1.1 points and has decades of outcome and safety data. The FDA warns that products claiming to replace diabetes medicine are marketed illegally, and stopping a prescribed drug for a supplement risks glucose rising unchecked.

Is it safe to take berberine with metformin?

Berberine adds to metformin's glucose-lowering effect; metformin alone rarely causes low glucose, but the combination, especially with other diabetes drugs, can push glucose below 70 mg/dL (3.9 mmol/L). Berberine also slows liver enzymes that clear many other drugs. Only combine the two with your prescriber's agreement.

How long does berberine take to work?

Trials ran from 4 weeks to 6 months. Fasting glucose can shift within a few weeks, but HbA1c needs about three months to show the full change.

Does cinnamon lower A1C?

Probably not by a meaningful amount. A Cochrane review of 10 trials found no significant change in HbA1c and an inconclusive effect on fasting glucose. Cassia cinnamon also contains coumarin, which can harm the liver in large amounts.

Does vitamin D lower blood sugar?

Not by much. In the D2d trial, 4,000 IU a day did not significantly lower diabetes risk in people with prediabetes; pooling three trials found about a 15% lower risk. Correcting a deficiency is still worthwhile for bone and muscle health.

What helps insulin resistance most?

Weight loss and regular activity have the strongest evidence. Among supplements, myo-inositol lowered fasting insulin and HOMA-IR in women with PCOS, and magnesium helped insulin-sensitivity markers in people at high risk, mainly when intake is low.

Supplements to lower blood sugar: what the evidence shows

Q: How fast can a supplement lower blood sugar?

Fasting glucose and fructosamine can change within 4 to 8 weeks, but HbA1c averages about three months of glucose, so its full change shows only after roughly 12 weeks. No supplement lowers HbA1c quickly.

The distinctions

01
A supplement's effect depends on how high your glucose already is
In 35 psyllium trials, glucose fell most in people with treated type 2 diabetes, less in prediabetes, and not significantly when glucose was normal ², and the berberine trials enrolled people who already had type 2 diabetes ¹. Read your own HbA1c, fasting glucose and fasting insulin first, because the pattern tells you whether a supplement has anything to improve.

Berberine and psyllium have the strongest trial results for lowering blood sugar, and both were tested mostly in people whose glucose was already high. In a 2015 meta-analysis of 35 psyllium trials, the higher a person's glucose, the bigger the drop, and with normal glucose the change was not significant ². So start by reading your own numbers, then look at which supplements have earned their reputation.

Key takeaways

01Berberine lowered HbA1c (your three-month average blood sugar) by about 0.7 points across 46 trials in type 2 diabetes ¹, the most-studied supplement here.
02Psyllium fiber lowered HbA1c by about 1 point in people being treated for type 2 diabetes, less in prediabetes, and not significantly when glucose was normal ².
03Myo-inositol lowered fasting insulin and HOMA-IR (a score of how hard your body works to keep glucose normal) in women with PCOS; evidence outside PCOS is thin ³.
04The American Diabetes Association finds no clear evidence that supplements improve outcomes in people with diabetes who are not deficient, and does not generally recommend them for glucose control ⁵.
05A structured lifestyle program cut progression to diabetes by 58% in the Diabetes Prevention Program, a bigger effect than any pill tested ⁶.

Lifestyle and the cause come first

High blood sugar in most adults comes from <a href="/conditions/insulin-resistance">insulin resistance</a>: muscle, liver and fat respond less to insulin, so the pancreas makes more, and glucose drifts up once the pancreas can no longer keep pace. Weight loss, regular movement, more fiber and better sleep act on that cause directly. In the Diabetes Prevention Program, an intensive lifestyle program cut progression from prediabetes to type 2 diabetes by 58%, while metformin cut it by 31%, both against placebo ⁶. No supplement has a prevention result of that size.

The most reliable way to lower <a href="/biomarkers/hba1c">HbA1c</a> naturally is diet and weight. The program aimed for a loss of about 7% of body weight and 150 minutes of activity a week ⁶. In practice that means more fiber from oats, beans, lentils and vegetables, fewer refined carbohydrates and sugary drinks, and moving most days.

The ADA's 2025 Standards of Care say this about supplements:

There is no clear evidence that dietary supplementation with vitamins, minerals (such as chromium and vitamin D), herbs, or spices (such as cinnamon or aloe vera) can improve outcomes in people with diabetes who do not have underlying deficiencies, and they are not generally recommended for glycemic management.American Diabetes Association, Standards of Care in Diabetes 2025, Section 5 ⁵

Some supplements still lower glucose and HbA1c in trials, but the ADA does not consider that evidence strong enough to recommend them for glucose control ⁵. The rest of this page ranks them by how well the trial results hold up, so you can discuss them with your clinician alongside proven treatment.

01Start from your own numbers

Which number is high?

The supplements in this guide were studied on different parts of the problem, so the pattern in your results matters more than the brand on the bottle. Four patterns cover most people.

Raised HbA1c or <a href="/biomarkers/glucose-fasting/high">fasting glucose</a>: HbA1c of 5.7% to 6.4% or fasting glucose of 100 to 125 mg/dL (5.6 to 6.9 mmol/L) is the ADA range for <a href="/conditions/prediabetes">prediabetes</a>. Berberine and psyllium were studied mainly in people at or above this range.
High fasting insulin or HOMA-IR with normal glucose: this is insulin resistance before glucose rises. Weight loss, magnesium and, in PCOS, myo-inositol are what was studied here. If your panel has no fasting insulin, ask your clinician to add it to the same morning draw as glucose so you can work out HOMA-IR below.
Spikes after meals: soluble fiber such as psyllium, taken before meals, blunts the rise, and so does fiber from food.
Everything normal: trials found little or no effect on normal glucose, so a supplement has no number to improve.

Where an example HbA1c of 5.9% sits%

5.9 %in the prediabetes range

Normal4–5.7Prediabetes5.7–6.5Diabetes range6.5–8

An HbA1c of 5.9% is in the prediabetes range. A repeat test and a talk with your doctor are the usual next steps; lifestyle change has the strongest evidence, and whether a supplement adds anything is a question for that conversation.

Cutoffs from the American Diabetes Association Standards of Care 2025, Section 2: normal below 5.7%, prediabetes 5.7% to 6.4%, diabetes 6.5% or higher. The International Expert Committee starts its at-risk range at 6.0%, which is why the middle band is shaded as contested.

Markers to read before choosing anything

Hemoglobin A1cYour average glucose over about three months; the number most trials measured.Glucose, FastingA single morning reading; shows whether glucose is already raised.Insulin, FastingPicks up insulin resistance years before glucose climbs.HOMA-IRCombines fasting glucose and insulin into one insulin-resistance score.

CalculatorHOMA-IR calculatorHave a fasting glucose and a fasting insulin from the same draw? Turn them into a HOMA-IR score.Work it out Free, no accountRead your blood testHave a panel with HbA1c, glucose or insulin on it? Read each marker in plain words.Read my results

02The supplements, ranked by evidence

How the blood sugar supplements compare

Supplement	What the evidence shows	Quality of evidence	Amount used in trials	Who should avoid it or check first
Berberine	HbA1c down about 0.73 points, fasting glucose down about 15 mg/dL (0.86 mmol/L) in type 2 diabetes ¹	Moderate: 46 trials, mostly small and in China	Usually 0.6 to 1.5 g a day, split with meals	Pregnancy, breastfeeding, anyone on diabetes drugs, statins, ciclosporin or other drugs cleared by the same liver enzymes (ask your pharmacist)
Psyllium fiber	Fasting glucose down about 37 mg/dL (2.1 mmol/L) and HbA1c about 0.97 points in treated type 2 diabetes; smaller in prediabetes; no significant effect with normal glucose ²	Moderate: 35 trials	Commonly about 10 g a day, split before meals	Swallowing problems or bowel narrowing; take apart from medicines
Myo-inositol	Lower fasting insulin and HOMA-IR in women with PCOS ³	Moderate in PCOS, thin elsewhere	Commonly 2 to 4 g a day	Pregnancy without medical advice; anyone on diabetes drugs
Magnesium	Lower fasting glucose in diabetes, better insulin sensitivity in people at high risk ⁴	Moderate, largest benefit when intake or levels are low	Varied; the NIH upper limit from supplements is 350 mg a day ¹⁵	Kidney disease; check medicine timing with your pharmacist
Cinnamon	No significant HbA1c change; fasting glucose effect inconclusive ⁸	Low: a Cochrane review found the evidence insufficient	About 2 g a day on average	Liver disease; cassia cinnamon carries coumarin
Chromium	Few of 20 trials reached meaningful glucose targets ⁹	Low to mixed	Mostly chromium picolinate, in micrograms	Kidney or liver disease; anyone on insulin
Alpha-lipoic acid	Studied for diabetic nerve pain, not for lowering glucose ¹⁰	Moderate for nerve symptoms, weak for glucose	600 mg a day in the SYDNEY 2 trial	Rare insulin autoimmune syndrome causing low blood sugar
Vitamin D	No significant drop in diabetes risk in the largest trial; a modest drop when three trials were pooled ¹¹ ¹²	High quality; small or no benefit on its own	4,000 IU a day in the D2d trial	Correct a deficiency on its own merits

Berberine: the strongest evidence, and the most drug-like

Berberine is a plant compound from goldenseal, barberry and Coptis. A 2021 meta-analysis of 46 randomized trials in type 2 diabetes found berberine, alone or added to standard treatment, lowered HbA1c by about 0.73 percentage points and fasting glucose by about 0.86 mmol/L (about 15 mg/dL) compared with control ¹. The usual amounts were 0.6 to 1.5 g a day and the trials ran 4 weeks to 6 months, so any effect shows on an HbA1c at about three months ¹. Most of those trials were small and run in China, which makes the effect size less certain than the headline suggests.

Berberine behaves like a drug in the body. Common side effects are constipation, diarrhea and stomach cramps. In a study of healthy men, two weeks of berberine slowed the liver enzymes that clear many medicines (named CYP3A4, CYP2D6 and CYP2C9 on drug labels) ¹³, so it may raise blood levels of drugs cleared this way, such as ciclosporin and some statins; ask your pharmacist whether any of yours are cleared this way. Berberine should not be used in pregnancy or breastfeeding because it can harm newborns ¹⁷. The full comparison with the prescription drug is on <a href="/compare/berberine-vs-metformin">berberine vs metformin</a>.

The supplement with the most trials

Strongest evidenceBerberineLowered HbA1c and fasting glucose across 46 randomized trials in type 2 diabetes.In our shop

Psyllium and soluble fiber: best for after-meal spikes

Psyllium is a soluble fiber that forms a gel in the gut and slows how fast glucose from a meal reaches the blood. In a 2015 meta-analysis of 35 trials, psyllium taken before meals lowered fasting glucose by about 37 mg/dL (2.1 mmol/L) and HbA1c by about 0.97 points in people being treated for type 2 diabetes ². The effect was smaller in prediabetes and not significant in people with normal glucose, the clearest example on this page of a supplement working in proportion to how high glucose is.

Fiber from oats, beans, lentils and vegetables does the same job. Psyllium needs plenty of water, is easier on the gut when built up slowly, and is best taken a couple of hours apart from medicines because it can slow their absorption.

Myo-inositol: insulin resistance, especially with PCOS

Myo-inositol is a sugar alcohol the body uses in insulin signaling. A meta-analysis of 9 randomized trials covering 496 women with <a href="/conditions/pcos">PCOS</a> found myo-inositol supplementation lowered fasting insulin and HOMA-IR compared with control ³. Those trials studied the second pattern above: high insulin with glucose still normal. Trials outside PCOS are few, so the case for myo-inositol in other people with <a href="/conditions/insulin-resistance">insulin resistance</a> is unproven.

Magnesium: helps most if you are low

Low magnesium is common in people with diabetes, partly because high glucose increases magnesium loss in urine. A 2021 meta-analysis of double-blind randomized trials found oral magnesium lowered fasting glucose in people with diabetes and improved insulin-sensitivity markers in people at high risk of diabetes ⁴. The benefit is largest when intake or blood level is low, and a normal serum magnesium does not rule out low body stores; a <a href="/biomarkers/magnesium-rbc">red-cell magnesium</a> test is sometimes used, though it is not standardized, and our <a href="/conditions/magnesium-deficiency">magnesium deficiency</a> page cover that gap.

The NIH sets the upper limit for magnesium from supplements at 350 mg a day for adults, because higher amounts cause diarrhea, and with poor kidney function magnesium can build up to unsafe levels ¹⁵. Anyone with kidney disease should not take magnesium without medical advice. Magnesium also binds some antibiotics and osteoporosis drugs, so timing matters ¹⁵.

If your intake runs low

Helps most if lowMagnesiumLowered fasting glucose in diabetes and helped insulin sensitivity in people at high risk.In our shop

Cinnamon and chromium: popular, mixed results

A Cochrane review (an independent systematic review group) of 10 trials with 577 people found no significant change in HbA1c with cinnamon, an inconclusive effect on fasting glucose, and insufficient evidence to support it for diabetes ⁸. Cassia cinnamon, the common supermarket kind, contains coumarin, which can harm the liver in large amounts; Ceylon cinnamon contains far less.

Chromium has been tested for decades with mixed results. A 2016 review of 20 trials in type 2 diabetes found that fasting glucose reached a clinically meaningful target in only 5 of 20 trials and HbA1c in only 3 of 14, and concluded chromium has limited effectiveness ⁹. Chromium picolinate is the most studied form.

Mixed evidence

Mixed evidenceChromiumFew of 20 trials in type 2 diabetes reached clinically meaningful glucose targets.In our shop

What does not work, despite the marketing

Alpha-lipoic acid is sold as a blood sugar supplement, but its trials are about nerve pain. The SYDNEY 2 trial gave 181 people with diabetic nerve damage 600, 1,200 or 1,800 mg a day for five weeks; it eased nerve-pain scores, with no extra benefit from the higher amounts, only more side effects ¹⁰. Alpha-lipoic acid has also been linked to insulin autoimmune syndrome (Hirata disease), a rare condition in which the body makes antibodies to its own insulin and causes episodes of low blood sugar ¹⁴.

0.88hazard ratio for diabetes with vitamin D3 in the D2d trial, not statistically significant2,423 adults with prediabetes, 4,000 IU a day, median 2.5 years ¹¹

Vitamin D did not significantly lower diabetes risk in people with prediabetes in the largest trial, D2d ¹¹. When the three main trials were pooled in 2023, vitamin D cut the risk by about 15%, a 3-year absolute drop of 3.3 percentage points ¹². That is a small effect next to lifestyle change, and a vitamin D deficiency is worth correcting for bone and muscle health either way.

Sold for blood sugar, studied for something else

Nerve symptoms, not glucoseAlpha-lipoic acidEased diabetic nerve-pain scores in trials; does little for glucose itself.In our shop

Small or no glucose benefitVitamin DNo significant drop in diabetes risk in the largest prediabetes trial; a modest drop when trials were pooled.In our shop

Trials of <a href="/compounds/probiotics">probiotics</a>, <a href="/compounds/resveratrol">resveratrol</a> and ginseng used different strains, products and populations and found small, inconsistent changes in glucose.

03Scale, safety and retesting

How big is a supplement's effect, compared with the alternatives?

Effect sizes only make sense side by side. Metformin used alone lowers HbA1c by about 1.1 points against placebo, according to a 2012 meta-analysis of 35 trials ⁷. Berberine's 0.73 points comes from weaker trials. Psyllium's near 1-point drop applied to people with treated diabetes and high glucose, not to people with prediabetes.

Option	Typical effect	Evidence base
Intensive lifestyle program	58% fewer people progressed to diabetes over about 3 years ⁶	Diabetes Prevention Program, 3,234 adults with prediabetes
Metformin	31% fewer progressed to diabetes ⁶; about 1.1 points lower HbA1c in diabetes ⁷	Large, long-term randomized trials
Berberine	About 0.73 points lower HbA1c in type 2 diabetes ¹	46 trials, mostly small
Psyllium	About 0.97 points lower HbA1c in treated type 2 diabetes; smaller in prediabetes ²	35 trials
Cinnamon	No significant HbA1c change ⁸	Cochrane review of 10 trials

Supplements at best add to the weight loss and activity that outperformed metformin in the Diabetes Prevention Program ⁶. The <a href="/conditions/prediabetes">prediabetes</a> and <a href="/compounds/metformin">metformin</a> pages cover the next steps and the drug's lab checks.

The prescription benchmark

Decades of outcome dataMetforminLowered HbA1c by about 1.1 points alone and cut progression to diabetes by 31% in prediabetes.

Who should be careful

Anyone on metformin, a sulfonylurea such as glipizide, insulin or a GLP-1 drug such as semaglutide: glucose-lowering supplements can add to the drug and cause <a href="/conditions/hypoglycemia">hypoglycemia</a>, a glucose below 70 mg/dL (3.9 mmol/L), most often with insulin or a sulfonylurea. Shakiness, sweating, a racing heart, confusion and hunger are warning signs.
Anyone on statins, ciclosporin, some blood thinners or other drugs cleared by the liver enzymes berberine slows: it may raise their levels ¹³.
Pregnant or breastfeeding women: avoid berberine ¹⁷, and take nothing for glucose without medical advice.
People with kidney disease: magnesium and chromium can build up.
People with liver disease: be careful with cassia cinnamon and chromium.
Anyone already taking a 'glucose support' blend: check its ingredients before adding a single supplement that duplicates one.

How to tell if a supplement is working

If you and your doctor decide to try a supplement, check HbA1c, fasting glucose and fasting insulin before starting, so you have a baseline; <a href="/learn/fasting-before-a-blood-test">fasting before the draw</a> keeps the readings comparable. Change one thing at a time, otherwise you cannot tell what moved the number. Fasting glucose and fructosamine (a blood test that averages glucose over the past 2 to 3 weeks) respond within weeks, so a retest at 4 to 8 weeks gives an early read. HbA1c reflects glucose over the life of red blood cells, roughly three months, so an HbA1c retest makes sense at about 12 weeks and not sooner.

HbA1c lags behind fasting glucose after a change

Fasting glucose (mg/dL)HbA1c (%)

Illustrative numbers, not study results. The shape reflects that fasting glucose responds within weeks while HbA1c averages about three months of glucose, so its full change shows only after roughly 12 weeks.

A meaningful change is one larger than normal day-to-day variation: a few mg/dL of fasting glucose is noise, while a sustained drop in HbA1c of several tenths of a point is real. If nothing has moved after three months, the supplement is probably not doing much for you. The two tests can disagree, which <a href="/compare/a1c-vs-fasting-glucose">HbA1c vs fasting glucose</a> explains.

For the retest

FructosamineAverages glucose over 2 to 3 weeks, so it moves before HbA1c does.Hemoglobin A1cThe 12-week check of whether anything changed.

CalculatorA1C calculatorTurn an HbA1c into the average glucose it stands for, in mg/dL.Work it out

CalculatorGlucose converterReading results in mmol/L? Convert them to mg/dL to compare with the trial numbers on this page.Work it out

Free, no accountDrop in your lab report and see every marker explained in plain words, placed on the body, against your lab's range. It's read on your device: nothing is uploaded and nothing is kept.

Read my results Download the app

Citations

Educational context only, not medical advice or a diagnosis. Always discuss your results with a clinician.