Nattokinase is an enzyme from natto, a Japanese fermented soy food. It’s sold to “thin the blood”, lower blood pressure and, lately, to clear spike protein and microclots. Here’s what has been tested in people, what hasn’t, and the one risk nobody selling it puts on the front of the tub.
The grades in this post: Strong means good randomised trials agree. Mixed means trials disagree or are small and short. Mechanistic means lab or test-tube work, not tested in people. None means there’s no usable evidence in people.
The short version
- It does something in the blood. One dose changed clotting and clot-breakdown markers for a few hours in healthy young men. The changes stayed within the normal range.
- Over the long term, the best trial found nothing. Three years, 265 people: no difference from placebo in artery thickening, artery stiffness, blood pressure or any blood test, including clotting.
- Blood pressure: smaller, shorter trials pooled together show a drop of about 3/2 mmHg. That conflicts with the long trial.
- Spike protein, long Covid, microclots: lab-dish findings, a proposal that has never been tested, and a case series that has been retracted. No trial in people.
- The risk is bleeding, especially with aspirin, other blood thinners, or before surgery.
What it is
Nattokinase is an enzyme made by the bacterium Bacillus subtilis var. natto when soybeans are fermented into natto. In the lab it breaks down fibrin, the mesh that holds a blood clot together. That’s where the “natural clot-buster” reputation comes from. Supplements are labelled in fibrinolytic units (FU), a measure of that activity in a test tube, not in your body.
What happens in people
One dose: a real but small effect
In a double-blind crossover trial, 12 healthy young men took a single 2,000 FU dose or placebo. Over the next 8 hours, markers of clot breakdown rose (D-dimer at 6 and 8 hours, fibrin degradation products at 4 hours), factor VIII fell, and a standard clotting time (aPTT) lengthened at 2 and 4 hours. All the changes stayed within the normal range (Kurosawa 2015).
Mixed A small trial of short-term markers. It shows the enzyme has some activity after swallowing. It doesn’t show a health benefit.
Three years: nothing measurable
The largest and longest trial, NAPS, randomised 265 people (median age 65, no heart disease) to 2,000 FU a day or placebo for a median of three years. There was no difference in the rate of artery wall thickening or artery stiffness, no effect on blood pressure, and no effect on any laboratory measure, including clotting, clot breakdown, blood flow properties and inflammation (Hodis 2021).
Strong Good evidence of no benefit for these outcomes, in healthy older adults.
The two trials don’t quite contradict each other. Kurosawa measured hours after a dose; NAPS measured at visits months apart. A short-lived effect after each capsule could exist without changing anything you can measure over years.
Blood pressure: the trials disagree
A 2023 meta-analysis pooled six trials (546 people) and found systolic pressure about 3.5 mmHg lower and diastolic about 2.3 mmHg lower than placebo. The same analysis found slightly higher total and LDL cholesterol, and slightly higher blood glucose (Li 2023). NAPS, the longest trial, found no blood pressure effect.
Mixed If there’s an effect, it’s small. For comparison, eating less salt lowered systolic pressure by about 5 mmHg in people with high blood pressure (see Beyond “Your Cholesterol”).
| Nattokinase, 6 short trials pooled | about 3.5 mmHg |
|---|---|
| Nattokinase, 3-year trial (NAPS) | no significant effect |
| Less salt, people with high blood pressure | about 5.4 mmHg |
The brain: one trial, mainly null
In 120 people with narrowed neck or brain arteries but no symptoms, 8,000 FU a day for six months didn’t improve overall thinking scores. One sub-score (visuospatial) did better in an exploratory analysis (Zhang 2025).
Mixed No benefit on the main outcome; one exploratory signal.
The spike protein, long Covid and microclot claims
You’ve probably seen nattokinase recommended for spike protein “detox”, long Covid or microclots. Here’s each claim and what’s behind it.
| Claim | What the evidence is | Grade |
|---|---|---|
| Nattokinase breaks down spike protein | In a lab study, nattokinase degraded spike protein in cell extracts and on the surface of cells grown in a dish, depending on dose and time. Four of the ten authors work for commercial companies rather than the university (Tanikawa 2022). Nobody has shown this happens in a person, or that enough enzyme reaches the right place. | Mechanistic untested in people |
| A “spike detox protocol” with nattokinase, bromelain and curcumin | Proposed in a 2023 review. The authors themselves call for randomised trials to find out whether it helps or harms (Hulscher 2023). None has been published. | None a proposal |
| Case reports show it works | The case series most often quoted (five patients treated with a combination including nattokinase) is marked as retracted on PubMed (Zeballos 2025). | None |
| It helps long Covid | A 2026 review of 14 treatments prioritised by clinicians and patients found long-Covid trials for six of them. Nattokinase wasn’t one; its evidence was indirect or very low certainty (Baptista 2026). | None in people |
| It dissolves “microclots” | In a lab-on-a-chip study of microclots from plasma, ultrasound did the work. Adding nattokinase or other enzymes made no significant difference at the frequencies that worked best (Rasouli 2025). | Mechanistic and not encouraging |
The bigger problem with the “clears the blood” idea: the one long trial that measured clotting and clot breakdown, for three years, found no change (Hodis 2021). Whatever the reason someone takes it, the effect the reasoning depends on hasn’t shown up in long-term use.
What this doesn’t mean. It doesn’t mean people with lasting symptoms after an infection aren’t unwell, or that clotting and blood-vessel problems aren’t worth investigating. They are, and some of that research is serious. It means nattokinase hasn’t been shown to be the answer, and it carries a real risk.
The risk: bleeding
An enzyme sold to break down clots can make you bleed. That isn’t a side effect; it’s the intended action, in the wrong place. The published cases are few, and single cases can’t prove cause, but they point the same way.
- A bleed in the brain: a woman taking aspirin after a mini-stroke had a bleed in the cerebellum after seven days of nattokinase. Her scans also showed small old bleeds in the brain (microbleeds), which she wouldn’t have known about (Chang 2008).
- A fatal bleed: an elderly woman taking nattokinase developed bleeding into the abdomen and died. The doctors who reported it linked it to the supplement (Ramachandran 2021).
- Swapping a prescription for it: a patient with a mechanical heart valve replaced their warfarin with nattokinase. After nearly a year, a clot formed on the valve and it had to be replaced again (Elahi 2015).
- Allergy: a man with a known natto allergy had anaphylaxis after taking nattokinase (Awatani-Yoshidome 2021).
Mechanistic Case reports only, but the mechanism is plausible and it’s the same mechanism the product is sold for.
you take aspirin, clopidogrel, warfarin or another blood thinner, or other supplements that affect bleeding (ask your pharmacist); you have a bleeding disorder, a stomach ulcer or a history of stroke; you have surgery or a dental extraction coming up; you’re pregnant or breastfeeding (there are no safety data); or you’re allergic to natto or soy.
Don’t swap a prescribed blood thinner for it. A supplement with no long-term evidence isn’t a replacement for a drug with trial evidence, and the valve case above shows what can happen.
Natto the food is different again. It’s very high in vitamin K2, which works against warfarin; people on warfarin are told to avoid it (Kudo 1990; Nakajima 2021). If you’re on warfarin, check with your anticoagulation clinic before eating natto or taking any natto product.
How much?
I’m not giving a dose. Trials used 2,000 FU a day (NAPS, three years) and 8,000 FU a day (the brain trial, six months). Neither showed a benefit on its main outcome. If a dose did “clean the blood” enough to matter, it would carry the bleeding risk of a drug, and it should be managed like one.
Where I land
Nattokinase is a good example of a supplement where the mechanism is real and the outcomes aren’t there. It does something to clotting markers for a few hours; over three years it did nothing measurable. The spike and microclot claims rest on lab work and a retracted case series. And unlike most supplements, its intended action is also its main risk.
If you’re dealing with lasting symptoms after an infection, or you’re worried about clotting, the useful first step is testing, not guessing: a proper history, and the blood tests that actually measure what you’re worried about.
How strong is the evidence?
| Claim | Grade | Main source |
|---|---|---|
| Short-term changes in clotting markers after one dose | Mixed | Kurosawa 2015 (12 men, within normal range) |
| Long-term benefit for arteries, blood pressure or blood tests | Strong no benefit | Hodis 2021 (NAPS, 265 people, 3 years) |
| Lowers blood pressure | Mixed | Li 2023 (6 trials) vs Hodis 2021 |
| Improves thinking | Mixed | Zhang 2025 (main outcome null) |
| Breaks down spike protein or microclots in people | Mechanistic | Tanikawa 2022; Rasouli 2025 (lab only) |
| Helps long Covid | None | Baptista 2026; Zeballos 2025 retracted |
| Can cause bleeding, especially with blood thinners | Mechanistic plus case reports | Chang 2008; Ramachandran 2021; Elahi 2015 |
Sources
- Awatani-Yoshidome K et al. Anaphylaxis from nattokinase in a patient with fermented soybean (natto) allergy. Allergol Int 2022;71:153–54. PMID 34452810. doi:10.1016/j.alit.2021.08.004
- Baptista SN et al. Candidate treatments for long COVID: a narrative review of expert and patient-driven priorities. Front Med 2026;13:1734600. PMID 41836927. doi:10.3389/fmed.2026.1734600
- Chang YY et al. Cerebellar hemorrhage provoked by combined use of nattokinase and aspirin in a patient with cerebral microbleeds. Intern Med 2008;47:467–69. PMID 18310985. doi:10.2169/internalmedicine.47.0620
- Elahi MM et al. Consequence of patient substitution of nattokinase for warfarin after aortic valve replacement with a mechanical prosthesis. Proc (Bayl Univ Med Cent) 2015;28:81–82. PMID 25552810. doi:10.1080/08998280.2015.11929198
- Hodis HN et al. Nattokinase atherothrombotic prevention study: a randomized controlled trial. Clin Hemorheol Microcirc 2021;78:339–53. PMID 33843667. doi:10.3233/CH-211147
- Hulscher N et al. Clinical approach to post-acute sequelae after COVID-19 infection and vaccination. Cureus 2023;15:e49204. PMID 38024037. doi:10.7759/cureus.49204
- Kudo T. Warfarin antagonism of natto and increase in serum vitamin K by intake of natto. Artery 1990;17:189–201. PMID 2360879. (No DOI.)
- Kurosawa Y et al. A single-dose of oral nattokinase potentiates thrombolysis and anti-coagulation profiles. Sci Rep 2015;5:11601. PMID 26109079. doi:10.1038/srep11601
- Li X et al. Nattokinase supplementation and cardiovascular risk factors: a systematic review and meta-analysis of randomized controlled trials. Rev Cardiovasc Med 2023;24:234. PMID 39076715. doi:10.31083/j.rcm2408234
- Nakajima T et al. First clinical trial of a newly developed, low menaquinone-7, fermented soybean natto in warfarin-dependent patients. J Med Food 2021;24:860–65. PMID 34406879. doi:10.1089/jmf.2021.0036
- Ramachandran L et al. Nattokinase-associated hemoperitoneum in an elderly woman. Cureus 2021;13:e20074. PMID 35003944. doi:10.7759/cureus.20074
- Rasouli R et al. Investigation of the synergistic effect of enzymatic and ultrasound-induced amyloid microclot degradation. J Thromb Thrombolysis 2026;59:764–72. PMID 41405757. doi:10.1007/s11239-025-03220-0
- Tanikawa T et al. Degradative effect of nattokinase on spike protein of SARS-CoV-2. Molecules 2022;27:5405. PMID 36080170. doi:10.3390/molecules27175405
- Zeballos RS et al. Post spike syndrome (PSS): simple solution leading to resolving results, five cases report. IDCases 2025;41:e02278. PMID 40671933. doi:10.1016/j.idcr.2025.e02278 (retracted)
- Zhang K et al. Nattokinase supplementation for cognitive enhancement in asymptomatic intracranial/carotid stenosis: a randomized controlled trial. J Stroke Cerebrovasc Dis 2026;35:108511. PMID 41325794. doi:10.1016/j.jstrokecerebrovasdis.2025.108511