Nattokinase is having a moment. It is the fibrinolytic enzyme isolated from natto, the sticky fermented soybean dish eaten at Japanese breakfast tables, and it is sold on the promise that it does something no ordinary supplement claims to do: dissolve the clots inside your blood vessels. Search for it and you will find pages describing it as a natural clot buster, an artery cleaner, and a gentler alternative to the blood thinner your doctor prescribed.
The underlying biology is real. Nattokinase genuinely cleaves fibrin, and in a test tube it does so impressively. But the distance between what an enzyme does in a dish and what a capsule does in a person is the whole story here, and it is a distance most of the internet skips. This guide walks the actual human evidence, including the largest trial ever run on it, which almost nobody quotes because it found nothing.
Nattokinase affects clotting, and the documented harms are concentrated in a few specific situations.
- Never use it to replace a prescribed blood thinner. A patient with a mechanical aortic valve who swapped warfarin for nattokinase developed thrombus on the valve after about a year and needed a second valve replacement.
- If you take aspirin or another antiplatelet, treat the bleeding risk as real. The one published brain bleed was a patient on aspirin for stroke prevention who had a cerebellar hemorrhage after seven days of nattokinase at 400 mg a day, roughly four times the usual dose, with cerebral microbleeds visible on MRI.
- Serious bleeding has been reported without any blood thinner involved. A fatal bleed into the abdomen was reported in a 92-year-old woman taking nattokinase but no anticoagulant. She had fallen and had a ruptured liver cyst, and her clotting tests were normal, so the authors offer nattokinase as a contributing hypothesis rather than a proven cause. She could not say how many capsules she took, describing it as sometimes a handful.
- Stop it before procedures and say so by name. That includes dental extractions, colonoscopy, spinal or epidural injections and eye surgery, not just major operations. There is no established washout period, so ask your surgeon how far in advance to stop.
- Avoid it entirely if you have a bleeding or clotting disorder, a history of hemorrhagic stroke or known cerebral microbleeds, an active ulcer or recent gastrointestinal bleed, or uncontrolled high blood pressure.
- Symptoms of a clot or a bleed are an emergency. Sudden chest pain, breathlessness, one-sided leg swelling, a severe headache, one-sided weakness, or blood in your stool or urine mean seek urgent care, not a change of supplement.
- Do not take it in pregnancy or while breastfeeding. There is no safety data whatsoever, and the European authorization excludes pregnant and lactating women.
- Avoid it if you react to natto or soy. Anaphylaxis to the isolated enzyme has been reported in a patient with fermented soybean allergy.
Nattokinase is a real enzyme with real fibrinolytic activity, and the most defensible human finding is a modest blood pressure reduction of roughly 3 mmHg systolic in small, mostly short trials. The largest and longest trial, 265 people treated for a median of three years at a university atherosclerosis unit, found no effect on carotid artery thickness, arterial stiffness, blood pressure, or any laboratory measure. Pooled results actually show slightly worse cholesterol, not better. No trial has shown it dissolving a clot in a person, and there are no outcome trials for stroke, heart attack or death. The spike protein claims rest on test-tube work. It interacts with blood thinners, and the clearest harms in the literature involve people who combined it with one or substituted it for one.
What nattokinase actually is
Natto has been eaten in Japan for centuries. It is made by fermenting soybeans with the bacterium Bacillus subtilis var. natto, and the result is stringy, pungent and, to most people raised outside Japan, an acquired taste. The Japanese researcher Hiroyuki Sumi identified a potent fibrin-digesting enzyme in it and named it nattokinase, first reporting it in 1987. Chemically it is subtilisin NAT, a serine protease in the subtilisin family, about 275 amino acids long.
What it does mechanically. Blood clots are held together by fibrin, a mesh of protein strands that traps platelets and red cells. Your body dissolves that mesh with an enzyme called plasmin. Nattokinase cleaves fibrin directly, and it also nudges the body's own system, with reported effects on tissue plasminogen activator and on the inhibitor that holds fibrinolysis back. Either way it is a genuinely different mechanism from anticoagulant drugs, which mostly prevent clots from forming rather than breaking down clots that already exist. You can read more on our nattokinase ingredient page.
Natto, nattokinase and vitamin K2 are three different things
This confusion causes real problems, so it is worth being precise. Natto is the whole food. Nattokinase is one enzyme isolated from it. Vitamin K2 as MK-7 is a different compound entirely, and natto happens to be one of the richest dietary sources of it.
The distinction matters because vitamin K2 and nattokinase pull in opposite directions on clotting. Vitamin K is what warfarin works against; it helps activate clotting factors. So natto the food can interfere with warfarin through its K2 content, while purified nattokinase does not, because manufacturers deliberately strip the K2 out. The material used in much of the published human research is described as a soy extract from which vitamin K2 has been removed, and the European specification caps vitamin K2 in authorized nattokinase at no more than 0.1 mg per kilogram. If you want the K2, that is a separate supplement with its own separate evidence.
What FU means on the label
Nattokinase is measured in fibrinolytic units, abbreviated FU. This is an activity unit, not a weight, which is why milligrams and FU are not interchangeable between products. The anchor most labels use is that 100 mg of standardized extract delivers about 2,000 FU. That is consistent with the European specification, which puts activity at 20,000 to 28,000 FU per gram.
One detail worth knowing: the authorized material is not pure enzyme. The European specification describes it as roughly 30 percent fermented soybean extract and 70 percent resistant dextrin used as a carrier. So most of what is in a nattokinase capsule, by weight, is not nattokinase.
What the evidence actually shows
There is more human research here than for most supplements, which is a point in nattokinase's favor. The problem is that the research does not agree with itself, and the disagreement runs along a line that should make any careful reader uncomfortable.
Blood pressure: the best supported claim, and still modest
A 2023 systematic review included six randomized trials totalling 546 participants, but only three of them, 223 people, contributed blood pressure data. Across those three, nattokinase reduced systolic blood pressure by 3.45 mmHg and diastolic by 2.32 mmHg, both statistically robust. (The published systolic confidence interval, given as 4.37 to 2.18 lower, is not symmetric about its own point estimate, which looks like a typo in the paper rather than in our reading.)
The review's own authors concluded that nattokinase can serve as an effective adjunctive option for hypertension, which is a fair reading of their pooled numbers.
The best single trial behind that figure is a 2008 Korean study, and it is the most directly on-point positive evidence nattokinase has. It was randomized, double-blind and placebo-controlled, enrolled 86 adults with pre-hypertension or stage 1 hypertension, and used the standard 2,000 FU a day for eight weeks. Net of placebo, systolic pressure fell 5.55 mmHg and diastolic 2.84 mmHg. A 2016 North American trial in 79 people found the same direction, with average diastolic pressure falling from 87 to 84 mmHg while placebo stayed at 87.
So the blood pressure signal is genuine, replicated in more than one country, and obtained at an ordinary dose. Size it correctly, though. Roughly 3 to 5 mmHg is a fraction of what a single blood pressure medication delivers, and it is comparable to what you would expect from modest changes in salt intake or exercise. It is useful, not transformative.
The trial that complicates everything
In 2021 the Atherosclerosis Research Unit at the University of Southern California published the Nattokinase Atherothrombotic Prevention Study. It is by a wide margin the most rigorous test this ingredient has ever received: 265 people, double-blinded, 2,000 FU a day, treated for a median of three years, with carotid ultrasound every six months.
It found nothing. The rate of change in carotid artery intima-media thickness and arterial stiffness did not differ from placebo. And in the authors' own words, there was no significant effect of nattokinase supplementation on blood pressure or any laboratory determination. Their conclusion was that nattokinase has a null effect on subclinical atherosclerosis progression.
One fair caveat: this trial enrolled people without clinical cardiovascular disease, median age 65, and the authors themselves described the population as healthy individuals at low risk. A null result in a low-risk group does not prove a null result in everyone. But it remains the longest and largest test available, and in the meta-analysis's own risk-of-bias assessment it is one of only three trials rated low risk across every domain. In fairness, the other two of that three, a 2009 lipid trial and a 2019 coagulation trial, both reported positive findings.
Cholesterol: the results run backwards
Here the pooled results run the other way from the marketing. Across four trials in 399 people, total cholesterol came out higher on nattokinase than placebo, and in the shorter-exposure subgroup LDL was higher by 6.49 mg/dL and HDL, the one you want high, was lower by 2.76 mg/dL. Blood glucose was slightly higher across two trials in 341 people. Triglycerides were unaffected.
Three honest caveats belong with that. First, the heterogeneity is severe: the reviewers reported I-squared values of 52, 95 and 93 percent for total cholesterol, HDL and LDL, and a 95 percent figure means the HDL result is close to uninterpretable rather than a solid finding. Second, the paper's "low total dosage" subgroup is defined by short duration, not by a small daily dose: those two trials actually ran at 6,000 and 8,000 FU a day for eight weeks. Third, the long USC trial sits in the other subgroup, so it is not what produced the unfavorable numbers.
The fair conclusion is narrower than "nattokinase worsens cholesterol," and it is simply this: there is no pooled human evidence that nattokinase on its own improves your lipids, and some signals point mildly the other way. The trials showing real lipid improvement tested it combined with red yeast rice. One 2009 trial made the point in its own title, reporting that combined nattokinase with red yeast rice, but not nattokinase alone, had potent effects on blood lipids. Red yeast rice contains a compound chemically identical to a statin, and we cover that combination separately in our red yeast rice and nattokinase guide.
The 1,062-participant study everyone cites
Search for nattokinase and plaque and you will land on a 2022 paper reporting reduced carotid thickness and plaque size in 1,062 participants at 10,800 FU per day over 12 months, with improvement rates quoted between 66.5 and 95.4 percent. It is the source behind most high-dose products on the market.
It is not a trial. The authors describe it as a retrospective analysis, list "retrospective study" among the paper's own keywords, and state in their limitations that it represents a retrospective study and not a randomized controlled trial. There was no randomization, no placebo and no control group; participants had chosen to take nattokinase as a supplement. Those headline improvement rates are the proportion of a single group whose numbers moved from their own baseline, with nothing to compare against.
Two further things belong next to it. First, five of its authors were employed by the company that manufactured the tablet, and a sixth by a pharmaceutical distributor; all of this is openly disclosed in the paper's own conflict of interest statement. Second, its explicit conclusion is that 3,600 FU per day is ineffective and that the standard 2,000 FU recommendation should be challenged, which is to say a manufacturer-authored analysis concluded you need roughly five times more product. It also reports that co-administering aspirin improved outcomes, which is a notable thing to recommend in a supplement whose main documented harm is bleeding on aspirin.
None of that makes its observations false, and industry funds a great deal of legitimate research. But this is not two trials disagreeing. One is a randomized, double-blind, placebo-controlled trial and the other has no control group at all, and that difference matters more than the funding does.
Cognition and stenosis
A 2026 randomized, double-blind trial gave 8,000 FU a day for six months to 120 people with asymptomatic narrowing of the carotid or intracranial arteries. The primary outcome, global cognition on the Montreal Cognitive Assessment, showed no difference whatsoever: a mean difference of 0.038 points, p = 0.948 in the per-protocol analysis, with the intention-to-treat result consistent. An exploratory analysis found a small improvement in visuospatial scores, which the authors appropriately framed as a signal worth investigating rather than a result. When the primary endpoint misses and a subdomain hits, the subdomain is a hypothesis.
Does it reach your bloodstream?
Everything above depends on an unanswered question. Nattokinase is a protein, roughly 28,000 daltons. Your digestive system is specifically built to break proteins into amino acids. So how does a large enzyme survive stomach acid and reach the bloodstream intact and still working?
What the human data show. There is exactly one human pharmacokinetic study, a pilot in 11 healthy volunteers given a single 2,000 FU dose. Using an ELISA built on a polyclonal anti-nattokinase antibody, the researchers detected a rise in signal from two hours out, peaking around 13.3 hours after the dose. That is genuine evidence that something reaches the blood.
But read what the authors say about their own result. They state that the bioavailability of nattokinase is currently unknown, that their assay detected nattokinase and/or its metabolites, and that looking for intact enzyme and bioactive peptides should be a consideration for future studies. In other words, an antibody-based test cannot distinguish active enzyme from broken-down fragments that still bind the antibody. Whether functional nattokinase circulates in humans has not been established.
What the animal and cell data add. A 2023 study using rat gut sac, ligated intestinal loop and Caco-2 cell monolayer models concluded that nattokinase is a moderately absorbed molecule taken up through active, energy-dependent endocytosis. That gives a plausible route. It does not tell you the concentration achieved in human blood.
What does change measurably. The most informative human study here is a double-blind, placebo-controlled crossover trial in 12 healthy young men given a single 2,000 FU dose. It found real movement: D-dimer rose at six and eight hours, fibrin degradation products rose at four hours, factor VIII activity fell, antithrombin rose, and clotting time lengthened. Something is clearly happening. The authors then add the sentence that matters: all of the changes were within the normal range.
And it is not only single doses in young men. A 2019 trial randomized 100 people with raised cholesterol to nattokinase or placebo for eight weeks and measured clotting directly. The nattokinase group showed significantly greater increases than placebo in collagen-epinephrine closure time and in activated partial thromboplastin time. That is a sustained, between-group, adequately sized effect on human clotting, and it is stronger evidence of a real hemostatic action than the single-dose work.
One caveat cuts against the "it is probably just fragments" reading. In spontaneously hypertensive rats, nattokinase fragments that possessed no protease activity at all still lowered blood pressure, apparently by suppressing plasma angiotensin II rather than by digesting fibrin. So the fragments may not be inert; they may simply do something different from the intact enzyme.
So the fair summary is this. Oral nattokinase produces measurable, replicated shifts in coagulation chemistry in humans, the enzyme or something antigenically like it appears in blood, and there is a plausible mechanism by which even its breakdown products could affect blood pressure. What remains unresolved is whether enough intact active enzyme reaches anywhere to dissolve an actual clot. Every clot-dissolving claim inherits that uncertainty; the blood pressure claim depends on it less.
Clots, DVT and arterial plaque
This is the reason most people buy it, so it deserves the bluntest treatment.
No human trial has ever shown nattokinase dissolving a clot in a person. The clot-busting claim traces to laboratory fibrin plate work, where purified enzyme is applied directly to fibrin, and to the original 1990 study in which capsules were given to dogs with experimentally induced thrombosis and lysis was observed on angiography. Sumi's own paper described the effect in human plasma as mild. There are also no trials reporting hard clinical outcomes: not stroke, not heart attack, not mortality, not recurrence of venous thromboembolism.
The long-haul flight study
The DVT prevention claim rests almost entirely on a 2003 trial in high-risk travelers on 7 to 8 hour flights. In the treated group there were no deep vein thromboses; in controls there were five, plus two superficial thromboses. That sounds decisive until you read the methods.
The product tested was Flite Tabs, identified in the published report only as 150 mg of pinokinase, from a company in Tempe, Arizona. The paper does not say what pinokinase contains, and its abstract and PubMed record never use the word nattokinase. The manufacturer describes pinokinase as a proprietary blend of nattokinase with pine bark extract, but that comes from outside the study, so the result belongs to a combination product rather than to nattokinase on its own. Beyond that: the thrombotic events were asymptomatic and detected only because everyone was scanned by ultrasound before and after the flight; both groups also followed an exercise program; and the headline intention-to-treat comparison counted 18 control-group failures, of which 11 were people lost to follow-up rather than people with clots. It is a suggestive small study of a proprietary blend over a single flight, not proof that a nattokinase capsule prevents travel thrombosis.
Arterial plaque
The apparent contradiction here dissolves once you look at the designs. The manufacturer-linked retrospective analysis at 10,800 FU reports reduced carotid thickness and plaque, with no control group. The independent university randomized trial at 2,000 FU, running three times as long and measuring the same artery by the same method, found no difference from placebo. And it is not alone: a separate randomized, double-blind, placebo-controlled trial of a nattokinase and red yeast rice product in 113 people over four months improved lipids but likewise found no significant effect on carotid intima-media thickness. Every randomized test of this endpoint has come back null.
Stroke recovery
A 2020 trial in Vietnam gave a nattokinase product to 61 people recovering from ischemic stroke alongside standard care. It reported significantly greater improvement in the nattokinase arm than in the control arm on three validated scales: the modified Rankin, the Orgogozo and the Barthel. That is a real between-group result and it should be reported as one.
The reasons to hold it lightly are about size and design rather than direction. It was single-blinded, ran at a single center in 61 people, used ordinal rating scales, and tested a proprietary Vietnamese product whose composition the paper does not specify. It is a promising small signal in a setting where nobody has repeated it, not a demonstration that nattokinase aids stroke recovery.
The spike protein question
Since 2022, nattokinase has been marketed heavily for so-called spike protein detoxification after COVID-19 infection or vaccination. Because this drives a large share of current interest, it deserves a precise answer rather than a dismissal.
What the actual paper found. A 2022 laboratory study in Molecules showed that nattokinase degraded SARS-CoV-2 spike protein in a dose- and time-dependent way. The researchers incubated the enzyme with cell lysates containing spike protein, and separately added nattokinase to culture medium, which degraded spike protein on the cell surface. The finding is real. The authors phrased their conclusion as suggesting potential. Three of the co-authors work for a life science company.
Why that does not transfer to a capsule. Pipetting a purified enzyme directly onto cells bypasses every obstacle a swallowed capsule faces: stomach acid, digestive proteases, intestinal absorption, dilution in blood, and reaching whichever tissue is supposed to hold the spike protein. Given that human data cannot yet confirm intact active enzyme circulates, the leap from this experiment to a clinical effect is very large.
There are no human trials. None for COVID, none for long COVID, none for post-vaccine symptoms. The widely circulated protocol combining nattokinase, bromelain and curcumin is a narrative review that proposes a protocol, not a trial, assembled from a literature search that explicitly included gray literature. Its conclusions call for large randomized, double-blind, placebo-controlled trials to determine risks and benefits, meaning the authors are not claiming it has been tested. Its senior author discloses partial salary support from and an equity position in The Wellness Company, which the paper states had no role in the work.
A five-patient case series published in 2025 and often shared as evidence has since been retracted, with a retraction notice published by the journal in December 2025.
The one independent test on real microclots. In fairness to the hypothesis, a 2026 study did test nattokinase against the thing that actually matters here: amyloid microclots derived from human plasma, of the kind implicated in long COVID. Researchers compared three fibrinolytic enzymes, including nattokinase, alongside ultrasound on a lab-on-chip platform. The result is instructive rather than encouraging. Low-frequency ultrasound did most of the work, reducing mean clot diameter by over 60 percent, while the enzymes did not produce statistically significant effects at 150 to 300 kHz. At 500 kHz the enzymes moderately improved things, and the authors summarized enzyme co-treatment as offering a limited but measurable effect. It is still a dish rather than a person, but it is the closest anyone has come to testing the actual premise, and nattokinase was not the active ingredient.
Dosage: FU, mg and what trials used
Dosing is where the gap between the research and the retail shelf is widest.
What trials used. Most human research used 2,000 FU per day, typically 100 mg of standardized extract. That includes the three-year atherosclerosis trial, the eight-week North American blood pressure trial, and the single-dose crossover work. The Vietnamese stroke study used 1,200 FU. The cognition trial used 8,000 FU. The manufacturer-linked plaque study used 10,800 FU.
What regulators allow. The European Union authorized nattokinase as a novel food ingredient at a maximum of 100 mg per day, for adults only, excluding pregnant and lactating women, and it requires the label to carry a statement that people taking medication should use the product only under medical supervision. Note that this is a ceiling on the weight of the authorized extract, not on FU. Since that material runs 20,000 to 28,000 FU per gram, 100 mg carries roughly 2,000 to 2,800 FU, which makes a 10,800 FU product about five times the authorized daily amount by weight. It is the only such ceiling we are aware of any regulator setting.
The uncomfortable pattern. The studies producing impressive results used high doses and were run or co-authored by manufacturers. The study using the standard 2,000 FU dose, run independently over three years, found nothing. Those two facts are often combined into a marketing argument that you simply need a higher dose. That is one possible reading. Another is that higher-dose findings have not yet been independently replicated. Both readings are live, and a product page that only tells you the first one is selling rather than informing.
On timing. You will see confident advice about taking nattokinase on an empty stomach or at night. There is no human trial comparing dosing schedules, so treat all of it as reasoning rather than evidence.
Side effects and interactions
In controlled trials nattokinase is generally well tolerated. The 2023 meta-analysis noted no notable adverse events attributable to it across the included studies, and a real-world observational report following 153 vascular surgery patients on 100 mg a day recorded no adverse drug reactions, although those patients were monitored, including INR checks, and had their doses adjusted by clinicians.
The published harms are rare but serious, and they cluster in recognizable situations.
- Bleeding on top of an antiplatelet. The cerebellar hemorrhage case involved aspirin, seven days of nattokinase at 400 mg daily, and pre-existing cerebral microbleeds. The authors concluded nattokinase may increase intracerebral hemorrhage risk in people with bleeding-prone small vessel disease who are already on an antithrombotic.
- Bleeding without any anticoagulant. A fatal hemoperitoneum was reported in a 92-year-old woman taking nattokinase and no blood thinner. Read carefully, though: she had had a mechanical fall, the bleeding source was a ruptured hepatic cyst, her INR and clotting times were normal, and her intake was uncontrolled and unmeasured. The authors present nattokinase as a possible contributor, not an established cause.
- Clotting caused by substitution. The mechanical valve case is the clearest harm in the entire literature, and it came from stopping a real drug rather than from the supplement's own activity.
- Allergy. Anaphylaxis to nattokinase has been reported in a patient with fermented soybean allergy, so if natto or soy causes you any reaction, the isolated enzyme can too.
Interactions worth naming
- Warfarin. Two opposite problems. Natto the food supplies vitamin K2 and opposes warfarin; purified nattokinase does not, but may add to bleeding risk. And substitution has caused documented thrombosis.
- Direct oral anticoagulants such as apixaban, rivaroxaban and dabigatran. There is no interaction data. That is an evidence gap, not a clean bill of health.
- Antiplatelets including aspirin and clopidogrel. Additive bleeding risk, and the one published brain bleed sits here.
- Blood pressure medication. Effects may be additive, which is mostly a monitoring issue rather than a danger.
- Other supplements that affect bleeding. Nattokinase is very often sold inside systemic enzyme or circulation blends alongside serrapeptase, lumbrokinase, fish oil, garlic or ginkgo. The risk is what you are taking in total, not what is on the front of one bottle.
- Laboratory tests. It can prolong clotting times, which may confuse test results if your care team does not know you are taking it.
What the FDA has actually done
Nattokinase is sold in the US as a dietary supplement, which means no pre-market approval, no efficacy review, and no verification that the FU number on the label is accurate. Structure and function claims are permitted; disease claims are not, and the agency has enforced that line here. A 2019 warning letter cited a company for marketing copy saying nattokinase is used for heart disease, high blood pressure, stroke, deep vein thrombosis and atherosclerosis. A 2020 warning letter cited another firm for claiming its product may also help dissolve existing blood clots and supports natural blood thinning, and concluded the products were new drugs that could not lawfully be sold without approval. If a seller uses that language, they are not describing a gray area.
How to choose a product
If you and your doctor decide it is reasonable to try, these are the distinctions that actually mean something.
- Look for FU, not just milligrams. FU measures enzyme activity. Milligrams tell you almost nothing, particularly since the authorized material is around 70 percent carrier by weight.
- Prefer the extract the trials used. Most published human research used one specific standardized commercial material. A product naming the branded extract it contains is traceable to the literature; one that just says nattokinase is not.
- Check whether vitamin K2 was removed. Most products remove it, but few say so clearly. If you take warfarin, this is the single most important thing on the label.
- Read the whole ingredient list. If nattokinase sits in a blend with other enzymes or blood-thinning botanicals, your total exposure is what matters.
- Understand what a certification seal covers. The industry seal most often shown on these products is administered by a trade association whose members are the companies selling the ingredient. Its four published criteria are that the nattokinase is derived from Bacillus subtilis natto, that the product supplies at least 2,000 FU a day, that FU is declared on the supplement facts panel, and that the product's safety has been demonstrated. What the mark does not claim to establish is that the product works. The association also does not publish a searchable list of certified products, directing enquiries to its secretariat instead, so a shopper cannot independently check a seal against a registry. It is worth knowing that the same association's public FAQ answers yes to the question of whether someone prescribed warfarin can take nattokinase. The published case reports in this article are a good reason to treat that as a question for your own prescriber rather than a settled point.
- Nobody verifies the FU number before sale. There is no compendial standard for nattokinase potency in the US, so the activity claim is the manufacturer's assertion.
For context on where this fits among circulation-focused options, see our circulation supplement guide and cardiovascular top picks.
Claims versus evidence, side by side
| The claim | How it is sold | What the evidence actually is |
|---|---|---|
| Lowers blood pressure | A natural way to bring your numbers down | The most defensible claim on this list, and still modest. A meta-analysis of six trials in 546 people found systolic pressure about 3.5 mmHg lower and diastolic about 2.3 mmHg lower. But the largest and longest trial, 265 people over a median of three years, reported no effect on blood pressure at all, and it did not contribute to that pooled blood-pressure figure. |
| Lowers cholesterol | Clears fats from the blood alongside plaque | The pooled data point the other way. In the same meta-analysis, lower-dose nattokinase was associated with higher total cholesterol, higher LDL, lower HDL and slightly higher blood glucose than placebo. Trials reporting lipid improvements generally used combination products or much higher doses. |
| Dissolves existing blood clots | A natural clot buster that clears thrombosis | No human trial has ever measured a clot dissolving in a person taking it orally. The evidence is test-tube fibrin digestion plus a 1990 study in dogs. The FDA has cited this exact wording in a warning letter as an unapproved drug claim. |
| Cleans out arterial plaque | Reverses atherosclerosis and unclogs arteries | Directly contradicted between studies. The trial reporting plaque reduction ran at 10,800 FU a day and had five authors employed by the manufacturer of the tablet. The independent three-year trial measuring carotid thickness by ultrasound every six months found no difference from placebo. |
| Prevents clots on long flights | Proven to prevent DVT on long-haul flights | The trial everyone cites tested a product its report identifies only as 150 mg of pinokinase, and never names nattokinase or states what the product contains. The clots it counted were asymptomatic and found by screening ultrasound, both groups also followed an exercise program, and the headline intention-to-treat gap counts people lost to follow-up as failures. |
| Degrades the COVID spike protein | Clears spike protein after infection or vaccination | Laboratory work only, with enzyme applied directly to cells. There are no human trials for COVID or long COVID. The best-known protocol is a published proposal that calls for trials rather than reporting them, and a five-case series circulated as proof has been retracted. The one independent test on real plasma-derived microclots found the enzymes added no significant effect over ultrasound alone. |
| A safer alternative to blood thinners | Natural anticoagulation without the prescription | The most dangerous claim here. A patient with a mechanical heart valve who replaced warfarin with nattokinase developed a clot on the valve and needed repeat surgery. Serious bleeding has also been reported in people not taking any anticoagulant, though the one fatal case involved a 92-year-old who had fallen and had a ruptured liver cyst, so causation there is the authors' hypothesis rather than a settled finding. |
Frequently asked questions
Is nattokinase a blood thinner?
Not in the way a prescription anticoagulant is, but it is not inert either. Nattokinase is a fibrinolytic enzyme, meaning that in a test tube it cleaves fibrin, the protein mesh that holds a clot together. That is a different mechanism from warfarin, which suppresses production of clotting factors, and from apixaban or rivaroxaban, which block a specific factor. In a double-blind crossover study in 12 healthy young men, a single 2,000 FU dose measurably raised D-dimer and fibrin degradation products, lowered factor VIII activity, and prolonged the activated partial thromboplastin time. The authors noted plainly that all of those changes stayed within the normal reference range. So the honest description is that nattokinase nudges clotting chemistry in an anticoagulant direction without reliably producing a clinically anticoagulated state. That is exactly why it is risky to treat it as either a real blood thinner or as harmless. It is not strong enough to replace a prescription, and not weak enough to ignore when it is stacked on top of one.
Does nattokinase dissolve existing blood clots?
No human trial has ever shown a clot dissolving in a person who took nattokinase by mouth. That claim comes from laboratory work, where the purified enzyme is applied directly to fibrin, and from a 1990 study in which capsules were given to dogs with experimentally induced thrombosis and lysis was seen on angiography. Neither is the same as swallowing a capsule and clearing a clot in your leg or lung. There are no trials reporting hard clinical outcomes such as stroke, heart attack or death for nattokinase at all. The US Food and Drug Administration has treated this specific claim as a drug claim: a 2020 warning letter to World Nutrition, Inc. cited marketing copy saying the product may help dissolve existing blood clots, and concluded the products were unapproved new drugs. If you have or suspect a clot, that is an emergency for a hospital, not a supplement.
How much nattokinase should I take per day?
There is no established effective dose, and the dose question is genuinely unsettled rather than merely unclear. Most human trials used 2,000 fibrinolytic units per day, which is typically 100 mg of the standardized extract. That is also the only ceiling any regulator has set: the European Union authorized nattokinase at a maximum of 100 mg per day for adults, excluding pregnant and lactating women. Many US products sell 4,000, 6,000 or 10,800 FU per day, which is up to five times that ceiling. The 10,800 FU figure traces to a single retrospective analysis with no control group, whose authors included five employees of the company that made the tablet, and which concluded that its findings challenge the recommended 2,000 FU dose. Meanwhile the largest and longest independent trial used 2,000 FU a day for a median of three years and found no effect on anything it measured. Taking more is a product conclusion, not an evidence-based one. This is a dose to settle with a clinician who knows your medications.
Can you take nattokinase with aspirin, warfarin or Eliquis?
This is the question to take to your doctor or pharmacist rather than to a supplement label, and there are three separate problems. With aspirin and other antiplatelets, the risk is additive bleeding: the one published brain bleed involved a patient taking aspirin for stroke prevention who developed a cerebellar hemorrhage after seven days of nattokinase, and who turned out to have cerebral microbleeds on MRI. With warfarin, there are two opposite problems. Natto the food is one of the richest sources of vitamin K2, which opposes warfarin, while purified nattokinase supplements usually have the K2 removed, so the two are not interchangeable. And substituting nattokinase for warfarin has caused documented harm: a patient with a mechanical aortic valve who made that swap developed thrombus on the valve after about a year and needed a second valve replacement. With the newer direct oral anticoagulants such as apixaban and rivaroxaban, there is simply no interaction data at all, which is a gap rather than reassurance. Never substitute nattokinase for a prescribed anticoagulant.
Does nattokinase contain vitamin K2?
Usually not, and this trips up a lot of people because natto, nattokinase and vitamin K2 are three different things that get used interchangeably. Natto is the whole fermented soybean food, and it is one of the richest known dietary sources of vitamin K2 as MK-7. Nattokinase is a single enzyme isolated from that food. Most commercial nattokinase is deliberately stripped of vitamin K2 during manufacture, and the European specification caps K2 in authorized nattokinase at no more than 0.1 mg per kilogram. The trial material used in much of the published human research is described as an extract from which vitamin K2 has been removed. This matters mainly if you take warfarin, because vitamin K works against warfarin while the enzyme does not. It also means you cannot assume a nattokinase capsule gives you the bone and vascular benefits studied for vitamin K2, or vice versa. If a product does contain K2, that should be on the label, and if you are on warfarin it is worth checking rather than assuming.
Does nattokinase break down the COVID spike protein?
In a dish, the enzyme degraded spike protein. In a person, nobody knows, and the marketing has run far ahead of the science. The underlying paper is a 2022 laboratory study in which nattokinase was incubated with cell lysates containing spike protein, and added directly to culture medium to degrade spike protein on cell surfaces. Three of its authors work for a life science company. Adding a purified enzyme straight onto cells is not the same as swallowing a capsule and having active enzyme reach the relevant tissue, and the human absorption data do not establish that it does. There are no human trials of nattokinase for COVID or long COVID. The widely shared spike detoxification protocol is a proposal that is a narrative review proposing a protocol, not a trial; its conclusions call for randomized placebo-controlled trials, and its senior author discloses partial salary support from and an equity position in The Wellness Company, which the paper states had no role in the work. A five-patient case series often cited as proof has since been retracted, with a retraction notice published in December 2025. The one independent study to test nattokinase against real plasma-derived microclots found that ultrasound did most of the work and the enzymes added no statistically significant effect at the frequencies where ultrasound worked best.
Who should not take nattokinase?
Anyone taking a prescribed anticoagulant or antiplatelet should not add it without their prescriber's involvement, and nobody should ever use it as a replacement for one. Beyond that, the published harms point to a fairly clear list: people with a bleeding or clotting disorder, a history of hemorrhagic stroke or known cerebral microbleeds, an active ulcer or recent gastrointestinal bleed, or uncontrolled high blood pressure. A fatal spontaneous abdominal bleed has been reported in an elderly woman taking nattokinase without any anticoagulant, so age and frailty matter too. There is no safety data at all in pregnancy or breastfeeding, and the European authorization explicitly excludes pregnant and lactating women and is for adults only. Anyone with a soy or natto allergy should avoid it, since anaphylaxis to the isolated enzyme has been reported in a patient allergic to fermented soybean. And it should be stopped before surgery and other invasive procedures.
How long before surgery should you stop taking nattokinase?
There is no established washout period, which is itself worth knowing. No study has determined how long the fibrinolytic effect persists after the last capsule, so any specific number you see online is an extrapolation rather than a finding. What is known is that a single dose measurably shifts coagulation markers for several hours, and that human pharmacokinetic work found levels peaking around 13 hours after a dose. Most surgical teams ask patients to stop supplements that affect bleeding one to two weeks before a procedure, which is the same guidance commonly applied to fish oil, garlic and ginkgo. The practical answer is to tell your surgeon, anesthetist and dentist that you take it, name it specifically rather than saying you take vitamins, and follow the window they give you. That applies to dental extractions, colonoscopy with planned biopsy, spinal or epidural injections and eye surgery, not just major operations.
The bottom line
Nattokinase is one of the more interesting supplements on the market, and it is not snake oil. The enzyme is real, the fibrin-cleaving mechanism is real, and there is a defensible signal for a small reduction in blood pressure. Compared with most of the supplement aisle, that is a respectable position.
But the marketing has run a long way past it. The strongest test ever conducted, three years and 265 people at a university atherosclerosis unit, found no effect on artery thickness, arterial stiffness, blood pressure or any laboratory measure. Pooled lipid results run in the wrong direction. No study has shown a clot dissolving in a person, no trial has measured whether anyone lives longer or has fewer strokes, the flight study tested a combination product, and the spike protein story is a petri dish plus a retracted case series.
Meanwhile the risks, though rare, are concrete and documented: a brain bleed in a patient already on aspirin who was taking four times the usual dose, a fatal abdominal bleed in a frail 92-year-old where nattokinase was one plausible contributor among several, and a clotted heart valve in a man who trusted it enough to stop his warfarin. That last case is the one to remember, because the harm did not come from nattokinase being dangerous. It came from someone believing it was strong enough to replace a medicine.
If you are considering nattokinase, the conversation to have is with a clinician who knows your medications and your cardiovascular risk. Elevated blood pressure is a medical condition, and it is managed with a doctor rather than with a supplement chosen from an article. If you and that clinician decide a trial is reasonable, 2,000 FU a day is the dose most of the human research used, and home blood pressure readings are the only effect on this list you could actually observe. If you are on a blood thinner, heading for surgery, pregnant, or hoping to clear a clot or a plaque, the evidence does not support it, and in some of those cases it argues actively against.