2,000–10,800 FU/day. For effects on lipids and plaque, the evidence uses 6,500–10,800 FU/day. On an empty stomach or between meals. Effects take months (26 weeks minimum in the trials with plaque reduction).
6,500 FU/day for 26 weeks: −36.6% in carotid plaque area (from 0.25 to 0.16 cm²) in patients with atherosclerosis + hyperlipidemia. Small study in selected patients; replication in broader populations is lacking.
See the study (RCT (Ren 2017, n=82)) →6,500 FU/day for 26 weeks: CIMT from 1.13 to 1.01 mm (−10.6%) in patients with atherosclerosis. Probable mechanism: fibrinolysis + LDL reduction.
See the study (RCT (Ren 2017, n=82)) →Modest reductions in LDL, total cholesterol, and triglycerides; the effect is more consistent when combined with red yeast rice. Small clinical magnitude on its own.
See the study (meta-analysis 2023) →10,800 FU/day for 12 months in 63–85-year-olds: −36.6% in plaque area and −21.7% in CIMT. A large study but without a strict placebo group and with declared conflicts of interest; treat as hypothesis-generating.
See the study (observational study (Chen 2022, n=1,062)) →Drop of ≈−5 mmHg systolic and ≈−2 mmHg diastolic in pre-hypertensive people with 2,000 FU/day for 8 weeks. Small effect but consistent with action on fibrinogen and viscosity.
See the study (small RCT in pre-hypertensive people) →2,000 FU/day for 3 years in adults without cardiovascular disease: NO significant differences in CIMT or arterial stiffness vs placebo. Suggests that the low dose is not enough or that the effect is only seen in patients with established plaque.
See the study (double-blind RCT (NAPS 2021, n=265)) →⏱ 5 min read
Nattokinase is a proteolytic enzyme extracted from natto, a traditional Japanese dish made from soybeans fermented with Bacillus subtilis var. natto. Since Hiroyuki Sumi isolated and described it in the mid-1980s as a serine protease with potent fibrinolytic activity, it has attracted interest as a nutraceutical strategy for modulating coagulation and the vascular system without using drugs. The relevant clinical question is how much of that in vitro effect translates into real changes in arterial plaque in humans, and the current answer is "some, but with many caveats."
Nattokinase does not directly dissolve atheromas or break down calcified plaques: its demonstrated action is on fibrin, the protein that forms the network that stabilizes thrombi. By degrading fibrin and activating the endogenous plasmin pathway, it reduces circulating fibrinogen, platelet aggregation, and blood viscosity. In in vitro studies and animal models, reduced endothelial oxidative stress, improved nitric oxide-mediated vasodilation, and decreased adhesion molecules (VCAM-1, ICAM-1) have also been observed, all steps that take part in the formation and stabilization of atherosclerotic plaque.
The reference study on carotid plaque is the controlled trial by Ren and colleagues published in Zhonghua Yi Xue Za Zhi in 2017 (PMID 28763875). It included 82 Chinese patients with established carotid atherosclerosis and hyperlipidemia, randomized to 6,500 FU/day of nattokinase or to simvastatin 20 mg/day for 26 weeks. In the nattokinase group, the mean carotid plaque area went from 0.25 cm² to 0.16 cm², a reduction of 36.6%, and the common carotid intima-media thickness (CIMT) fell from 1.13 mm to 1.01 mm, a −10.6% change. The magnitudes are notable and are often cited in the popular literature, but it is worth reading the fine print: n=82, a selected population with active disease, a single-center trial, and no independent replication by another group. It is the best available evidence for the effect on plaque and, at the same time, fragile evidence.
In 2022, Chen and colleagues published in Frontiers in Cardiovascular Medicine (PMID 36061575) a prospective study of 1,062 participants aged 63 to 85 treated with 10,800 FU/day for twelve months. They reported a 36.6% reduction in plaque area and a 21.7% reduction in CIMT. It is the largest study to date and its magnitudes strikingly match those of the Ren trial, but it has no strict placebo group: the authors describe the design as "controlled against natural progression," the declared conflicts of interest include ties to nattokinase manufacturers, and for that reason the cardiovascular community treats it as hypothesis-generating rather than definitive evidence.
The Nattokinase Atherothrombotic Prevention Study, NAPS (PMID 33843667), published in 2021, is a randomized, double-blind trial of 265 middle-aged adults with a median age of 65 years without clinical cardiovascular disease, treated with 2,000 FU/day for a median of three years. The result was negative: there were no significant differences in the rate of change of CIMT or in arterial stiffness vs placebo. This trial cools the optimistic reading and suggests two things: first, that at the usual "maintenance" dose (2,000 FU) the arterial effect is indistinguishable from placebo; and second, that nattokinase might not work for primary prevention in people without established plaque but might work in people who already have it. Both interpretations are consistent with the available data.
A meta-analysis published in Reviews in Cardiovascular Medicine in 2023 (PMID 37835658) synthesized the available RCTs and found modest reductions in LDL, total cholesterol, and triglycerides, more consistent when nattokinase is combined with red yeast rice. On its own, its lipid effect is small, on the order of 5 to 10% of LDL. In hypertensive and pre-hypertensive people, a small trial with 2,000 FU/day for eight weeks (PMID 18971533) documented mean decreases of −5 mmHg systolic and −2 mmHg diastolic, magnitudes that are clinically small but consistent with the hemorheological improvement.
The dose used in the trials with documented plaque reduction is 6,500 FU/day, with the best result at 10,800 FU/day in the large observational study. The "maintenance" doses of 2,000 FU/day that are commonly sold match the dose that showed no effect in NAPS, so if the goal is to act on plaque, the relevant dose is much higher. It is taken on an empty stomach or between meals to maximize absorption, and the trials that saw an effect used six months or more. Products should specify FU (fibrinolytic units); those labeled only in milligrams do not allow doses to be compared.
The main risk is bleeding. There have been published cases of intracranial hemorrhage in patients who combined nattokinase with warfarin or antiplatelet drugs, and for that reason the basic rule is not to combine it with anticoagulants or antiplatelet drugs except under active medical supervision. It should be stopped one or two weeks before any surgery or invasive procedure, and it is not recommended in pregnancy, breastfeeding, active peptic ulcer, uncontrolled hypertension, or a history of hemorrhagic stroke. Patients with a soy allergy should verify the product's purification process, although commercial nattokinase is usually free of allergenic soy protein.
The honest reading today is this: there are biologically plausible signals and consistent clinical data that high-dose nattokinase may reduce carotid plaque size and CIMT in patients with established atherosclerosis and hyperlipidemia, with magnitudes that, if confirmed, would be larger than those of conventional dietary interventions. At the same time, in primary prevention and at low doses, the best available randomized trial found no effect. Independent replication and trials with hard endpoints (heart attacks, strokes, cardiovascular mortality) are still lacking. It is not a miraculous "artery cleaner," it should not replace statins when they are indicated, and its main potential clinical usefulness is as an adjunct in patients with documented carotid plaque who are willing to accept the bleeding risk and the cost of a high dose for six or twelve months under medical supervision.
Information, not medical advice. Scientific evidence reflects the quality of the studies.