Natto, Longevity and Heart Health: What the Studies Show

Table of Contents

1. Why Researchers Study Natto in Japan
2. The JPHC Study: 92,915 Adults, 14.8 Years
3. The Takayama Study: Natto and Stroke
4. All the Hazard Ratios Side by Side
5. Natto and Bone: From Hip-Fracture Maps to a 15-Year Cohort
6. Possible Mechanisms, Graded by Evidence
7. Association Is Not Proof: The Confounding Problem
8. What It Means at the Table
9. Safety: Warfarin, Allergy and Bleeding
10. Key Research Papers
11. Connections
12. Featured Videos

Why Researchers Study Natto in Japan

Natto is whole soybeans that have been steamed and fermented by the bacterium Bacillus subtilis (the natto strain). The result is sticky, stringy and strongly flavoured — and, nutritionally, quite different from the plain bean. Fermentation produces three things researchers care about: very large amounts of vitamin K2 in its MK-7 form, an enzyme called nattokinase that can digest fibrin (the protein mesh of a blood clot), and a rich supply of polyamines such as spermidine.

Japan is a natural laboratory for studying it, for a simple reason: within one country, some people eat natto almost every day and others never touch it. Natto has traditionally been far more popular in eastern Japan (around Tokyo) than in the west of the country. That gives epidemiologists a wide spread of intake inside a population that otherwise shares much of its diet, health system and record keeping — exactly what you need to look for a signal.

It also matters that most of these studies separate fermented soy (natto and miso) from unfermented soy foods such as tofu. As you will see, that distinction turns out to be the whole story: the benefit signal sits with the fermented foods, and especially natto, while total soy intake on its own shows nothing.

Back to Table of Contents

The JPHC Study: 92,915 Adults, 14.8 Years

Design: prospective cohort (evidence tier: observational association). The Japan Public Health Center-based Prospective Study enrolled residents of 11 public health centre areas across Japan. Katagiri and colleagues, writing in the BMJ in 2020, analysed 92,915 adults (42,750 men and 50,165 women) aged 45 to 74 who filled in a detailed food questionnaire in the mid-to-late 1990s. People who already had cancer, stroke or a heart attack were excluded. Over an average of 14.8 years, 13,303 participants died.

What was compared: people were split into five equal groups (fifths) by how much of each soy food they ate, and the highest fifth was compared with the lowest. For natto, the lowest group was people who ate none at all, and the top group ate more than 26.2 g a day — a little over half a typical pack. The results were adjusted for age, smoking, alcohol, body weight, exercise, other foods and more.

The results, in plain numbers. A hazard ratio (HR) of 1.00 means “same risk”; 0.80 means 20% lower risk in the high-intake group during follow-up. The number range in brackets is the 95% confidence interval — if it crosses 1.00, the result could be chance.

• Total soy, all forms together — no association. Highest versus lowest fifth: HR 0.98 (0.91–1.06) in men and 0.98 (0.89–1.08) in women. Eating more soy in general did not track with living longer.
• Fermented soy (natto plus miso) — about 10% lower death from all causes. HR 0.90 (0.83–0.97) in men and 0.89 (0.80–0.98) in women.
• Natto and death from any cause. In women, HR 0.84 (0.76–0.93) — a clear association. In men, HR 0.94 (0.87–1.02) — the interval crosses 1.00, so no firm association.
• Natto and cardiovascular death — the most consistent finding. HR 0.76 (0.65–0.90) in men and 0.79 (0.65–0.95) in women. The authors called this “a significant association commonly observed” in both sexes.

Interestingly, miso — also fermented soy — did not show the cardiovascular association that natto did. The authors suggested one possible reason: miso is salty, and sodium raises blood pressure, while natto carries far less salt. That is a hypothesis, not a measured result.

The authors were candid about the limits. They wrote that the fermented-soy findings “should be interpreted with caution because unadjusted residual confounding might remain,” and that the associations “do not translate directly to clinical benefit.” When they excluded people who died in the first three years (a check against already-ill people eating less), the fermented-soy and all-cause association became marginal.

A companion JPHC analysis by Nozue and colleagues (2021) looked at new cases of heart disease and stroke rather than deaths, in 79,648 participants. In women, the highest quarter of fermented soy intake had an HR of 0.80 (0.68–0.95) for cardiovascular disease, with natto itself also inversely associated; there was a similar link with stroke in women. In men, no soy food was associated with cardiovascular disease, and no soy food was linked to total cancer in either sex.

Back to Table of Contents

The Takayama Study: Natto and Stroke

Design: prospective cohort (observational association). In 1992, residents of Takayama city in Gifu Prefecture aged 35 and over — 13,355 men and 15,724 women — completed a validated food-frequency questionnaire. Nagata and colleagues then followed deaths for 16 years and reported in the American Journal of Clinical Nutrition in 2017. There were 1,678 cardiovascular deaths, including 677 from stroke and 308 from ischaemic heart disease.

This study was designed with natto’s clot-dissolving enzyme in mind: the authors noted that natto’s relationship to cardiovascular disease “has not been studied” despite its potent fibrinolytic enzyme. Intake was split into quarters, highest versus lowest compared, adjusted for the usual risk factors.

• Natto and total cardiovascular death: HR 0.75 (0.64–0.88) — a quarter lower risk.
• Natto and stroke death: HR 0.68 (0.52–0.88).
• Natto and ischaemic stroke death (the clot-type stroke): HR 0.67 (0.47–0.95).
• Total soy protein and total soy isoflavones and cardiovascular death: no significant association. Soy protein was linked to fewer stroke deaths (HR 0.75, 0.57–0.99), but soy protein and isoflavones from soy foods other than natto showed nothing for total cardiovascular death.

Think of it like finding that a neighbourhood’s gardeners live longer, then discovering it is only the ones who grow one particular plant. Whatever the effect is, it does not look like a general “soy” effect — it looks specific to natto. That fits with the JPHC result, where total soy showed nothing and natto stood out.

The authors’ own conclusion was measured: the data “suggest that natto intake may contribute to the reduction of CVD mortality.” “May” is the right word for a cohort study.

Back to Table of Contents

All the Hazard Ratios Side by Side

The chart below puts every verified number from this page on one scale. Each square is a hazard ratio for the highest-intake group compared with the lowest; the line through it is the 95% confidence interval. Anything to the left of the 1.0 line means lower risk; the scale runs from 0.3 to 1.2 and is spaced by ratio, so 0.5 (half the risk) sits as far from 1.0 as a doubling would on the other side. Grey rows are the ones whose interval crosses 1.0 — no clear difference.

Three patterns stand out. First, the cardiovascular rows for natto cluster tightly between 0.67 and 0.79 across two separate cohorts run by different research teams — consistency is one of the things that makes an association more believable. Second, the stroke rows sit furthest left among the heart-and-circulation results, which fits the idea of a clot-related mechanism (though it does not prove one). Third, the bottom grey row — total soy in all forms — sits right on the line. Whatever is going on, it is not “eat more soy.”

Back to Table of Contents

Natto and Bone: From Hip-Fracture Maps to a 15-Year Cohort

Bone is part of the longevity story: a hip fracture in older age is one of the strongest predictors of losing independence. The natto-and-bone evidence comes in three layers of increasing strength.

Layer 1: the map (ecological correlation)

Kaneki and colleagues (2001) measured blood levels of MK-7 in postmenopausal women. The average was 5.26 ng/mL in women in Tokyo, where natto is popular, 1.22 ng/mL in Hiroshima, where it is not, and 0.37 ng/mL in British women. Eating natto produced a marked, sustained rise in blood MK-7. Across Japan’s prefectures, higher natto consumption went with lower hip-fracture rates in women. Tier: ecological plus a small feeding observation. Comparing regions cannot tell you whether the individual women who ate natto were the ones who avoided fractures — but it was the clue that launched the later studies.

Layer 2: bone density over three years

In the Japanese Population-Based Osteoporosis (JPOS) Study, Ikeda and colleagues (2006) measured bone mineral density in 944 women aged 20 to 79 and again three years later. In postmenopausal women, more habitual natto went with higher hip bone density at baseline and slower bone loss at the femoral neck (the top of the thigh bone, where hips break); the femoral-neck link held after adjustment. Tofu and other soybean products showed no such association. Tier: prospective cohort.

Layer 3: actual fractures over 15 years

Kojima and colleagues (2020) followed 1,417 postmenopausal women from the same JPOS cohort for a median of 15.2 years; 172 had an osteoporotic fracture. Compared with women eating less than one pack (about 40 g) a week:

• 1–6 packs a week: fully adjusted HR 0.79 (0.56–1.10) — the interval crosses 1.00, so not a firm result.
• 7 or more packs a week (roughly a pack a day): fully adjusted HR 0.56 (0.32–0.99) — about 44% lower fracture risk, independent of baseline bone density.

Tofu and other soybean products again showed no association. The upper end of that confidence interval sits just under 1.00, and the study is small, so treat it as encouraging rather than settled. Tier: prospective cohort. The randomised-trial evidence for MK-7 and bone is covered in the next section. For the bone story in more depth, see Natto for Bone Density.

Back to Table of Contents

Possible Mechanisms, Graded by Evidence

Cohort studies tell us that natto eaters do well; they cannot tell us why. Several components of natto have plausible mechanisms, and they differ a great deal in how well they have been tested.

MK-7 and matrix Gla protein — the strongest mechanism

Vitamin K is the switch that activates a handful of proteins. Two matter here. Matrix Gla protein (MGP) sits in artery walls and, once activated by vitamin K, helps stop calcium from depositing there — think of it as a security guard that only works when it has its badge. Osteocalcin, made by bone-building cells, needs the same activation to bind calcium into bone. Natto is the richest common food source of MK-7, a long-lasting form of K2; in a feeding experiment, Schurgers and Vermeer (2000) found blood K2 after a natto meal reached levels about 10 times higher than blood K1 after a spinach meal.

Two Dutch randomised, placebo-controlled trials in the same group of 244 healthy postmenopausal women tested 180 µg of MK-7 a day for 3 years (Knapen 2013 and 2015). MK-7 cut inactive MGP by 50% compared with placebo, improved arterial stiffness (especially in women whose arteries were stiffest at the start), and slowed age-related loss of bone density at the lumbar spine and femoral neck, though not at the total hip. Tier: RCT — but of a supplement, not of natto, and on markers (stiffness, density), not on heart attacks or fractures.

On calcification of the heart’s own arteries, the evidence is weaker. A cross-sectional study of 564 Dutch postmenopausal women (Beulens 2009) found that the highest quarter of dietary menaquinone (K2) intake went with less coronary calcification (relative risk 0.80, 0.65–0.98), while K1 showed no link. That diet contained little or no natto, and a single-time-point snapshot cannot show cause and effect. Tier: cross-sectional. We did not find a study measuring coronary calcification directly against natto intake. More on the K2 side: Natto’s Vitamin K2 (MK-7) and Vitamin K and arterial calcification.

Nattokinase — real in the test tube, small and mixed in people

Nattokinase digests fibrin, and in a double-blind crossover trial of 12 healthy young men (Kurosawa 2015), a single 2,000 FU dose of a nattokinase supplement raised D-dimer and fibrin-degradation products (signs of clot breakdown) and modestly prolonged a clotting-time test over the next 2–8 hours — all within the normal range. Tier: small RCT, short-term markers.

The largest long-term test points the other way. In the Nattokinase Atherothrombotic Prevention Study (Hodis 2021), 265 adults without heart disease (median age 65) took 2,000 FU or placebo for a median of 3 years. Nattokinase did not slow thickening or stiffening of the carotid arteries and had no effect on blood pressure or any laboratory measure. Tier: RCT, null result. Whether nattokinase eaten as part of the food survives digestion in a useful amount is not established. The Takayama authors raised it as a possible explanation; it remains a hypothesis. See Nattokinase and Fibrinolysis and the Nattokinase hub.

Polyamines (spermidine and spermine) — promising, early

Polyamines are small molecules involved in cell growth and in autophagy, the cell’s recycling system. Kobayashi and colleagues (2017) measured 1,055–2,306 nmol of spermidine per gram across 16 raw soybean samples and showed that natto’s polyamine content tracks the bean it was made from very closely (correlation about 0.95), so natto made from high-polyamine soybeans is a good source. In a 12-month study (Soda 2021), 30 men aged 40–69 ate polyamine-rich natto daily while 27 kept their usual diet; blood spermine rose modestly, blood spermidine did not, and an immune-cell marker linked to age-related inflammation (LFA-1) fell in the natto group. The group was not described as randomised. Tier: small controlled human study, surrogate marker; the lifespan extension is from mice. See Spermidine.

Isoflavones and fibre — probably not the main story

Natto contains soy isoflavones and fibre like other whole-soy foods. But in the Takayama Study, total soy isoflavone intake was not associated with cardiovascular death, and isoflavones from non-natto soy foods showed nothing. The JPOS bone authors suggested isoflavones might add to MK-7’s effect on bone, but that is untested. Tier: association / hypothesis. If isoflavones alone explained the benefit, tofu should have shown it too — and in these studies it did not.

Back to Table of Contents

Association Is Not Proof: The Confounding Problem

A cohort study is like noticing that people who carry umbrellas are more often wet-shoed: the umbrella is not causing the wet shoes — rain is causing both. In nutrition, the “rain” is everything else that travels with a food choice.

• Natto eaters differ. In the JPHC data, current smoking was less common in the highest natto group than the lowest. Researchers adjust for smoking, but they can only adjust for what they measured, and only as well as they measured it.
• Natto travels with a traditional diet. A daily natto habit usually comes with rice, fish, vegetables, pickles and green tea. Some of the benefit attributed to natto may belong to that pattern.
• Regional differences. Natto is concentrated in eastern Japan, so natto intake is partly a marker of where someone lives, with its own healthcare, climate and habits.
• Reverse causation. People who are already unwell may eat less of a strongly flavoured food. The JPHC sensitivity check that dropped deaths in the first three years weakened the fermented-soy and all-cause signal, a reminder this matters.
• One food questionnaire. Intake was asked once (or at a few points), then people were followed for 15 years. Diets change.

What makes the natto findings more credible than a typical single-food headline is the combination: two independent cohorts agreeing on cardiovascular death, a specific food standing out while total soy does not, a stroke signal that fits a plausible clot mechanism, and supplement trials showing MK-7 changes arterial and bone markers. What keeps it at “association” is that no randomised trial has given people natto and counted heart attacks, strokes or deaths. That trial may never be run.

Back to Table of Contents

What It Means at the Table

The honest summary: eating natto regularly, as part of a traditional whole-food diet built around fermented foods, is a reasonable choice that is consistent with some of the best long-term population data we have — and it carries a real nutritional payoff (vitamin K2, protein, fibre, polyamines) whether or not the mortality associations turn out to be fully causal.

• How much? A typical pack is about 40–50 g. The top JPHC group ate more than 26.2 g a day, and the JPOS fracture benefit appeared at 7 or more packs a week — roughly one pack a day. Nobody has tested whether more than that adds anything.
• Make it part of a fermented-food pattern. The benefit signal in these studies lives with fermented soy, not unfermented soy foods such as tofu or soy milk, which showed no association. Natto sits naturally alongside miso, traditional soy sauce, pickled vegetables, kimchi and sauerkraut.
• Serve it the traditional way. Over warm brown rice with sliced green onion, a little mustard, a raw egg yolk or a splash of traditional soy sauce. Whole foods, full fat, nothing “lite”.
• If the taste is a hurdle, start small and mix it in. Our acquired-taste tips help, and you can make it yourself: How to Make Natto.
• Food first. The cohort data are about the food. Isolated nattokinase capsules are a different product: the largest long-term trial found no benefit, and they add bleeding risk with blood thinners.

Back to Table of Contents

Safety: Warfarin, Allergy and Bleeding

• Warfarin (and other vitamin K antagonists) — the critical caution. Warfarin works by blocking vitamin K. Natto delivers so much K2, and keeps blood levels high for so long, that it can blunt warfarin’s effect and push the INR down, raising clot risk. Schurgers and Vermeer flagged natto’s potential to interfere with oral anticoagulants back in 2000. People on warfarin should not start, stop or change natto intake without their anticoagulation clinic. The newer direct oral anticoagulants do not work through vitamin K, but nattokinase supplements still add bleeding risk to any blood thinner. Details: Natto Safety and Interactions and PT/INR.
• Bleeding and surgery. Concentrated nattokinase supplements shifted clotting markers in the Kurosawa trial. Combined with aspirin, antiplatelet drugs or anticoagulants, or before surgery, that matters.
• Soy allergy. Natto is soy. People with soy allergy should avoid it.
• Kidney disease. Natto is high in potassium and phosphorus; people on restricted diets should check their limits.

Back to Table of Contents

Key Research Papers

1. Katagiri R, Sawada N, Goto A, Yamaji T, Iwasaki M, Noda M, Iso H, Tsugane S (2020). Association of soy and fermented soy product intake with total and cause specific mortality: prospective cohort study. BMJ. — PubMed PMID: 31996350
2. Nagata C, Wada K, Tamura T, Konishi K, Goto Y, Koda S, Kawachi T, Tsuji M, Nakamura K (2017). Dietary soy and natto intake and cardiovascular disease mortality in Japanese adults: the Takayama study. American Journal of Clinical Nutrition. — PubMed PMID: 27927636
3. Nozue M, Shimazu T, Charvat H, Mori N, Mutoh M, Sawada N, et al. (2021). Fermented soy products intake and risk of cardiovascular disease and total cancer incidence: The Japan Public Health Center-based Prospective study. European Journal of Clinical Nutrition. — PubMed PMID: 32887936
4. Kaneki M, Hodges SJ, Hosoi T, Fujiwara S, Lyons A, Crean SJ, et al. (2001). Japanese fermented soybean food as the major determinant of the large geographic difference in circulating levels of vitamin K2: possible implications for hip-fracture risk. Nutrition. — PubMed PMID: 11369171
5. Ikeda Y, Iki M, Morita A, Kajita E, Kagamimori S, Kagawa Y, Yoneshima H (2006). Intake of fermented soybeans, natto, is associated with reduced bone loss in postmenopausal women: Japanese Population-Based Osteoporosis (JPOS) Study. Journal of Nutrition. — PubMed PMID: 16614424
6. Kojima A, Ikehara S, Kamiya K, Kajita E, Sato Y, Kouda K, Tamaki J, Kagamimori S, Iki M (2020). Natto Intake is Inversely Associated with Osteoporotic Fracture Risk in Postmenopausal Japanese Women. Journal of Nutrition. — PubMed PMID: 31825069
7. Knapen MH, Drummen NE, Smit E, Vermeer C, Theuwissen E (2013). Three-year low-dose menaquinone-7 supplementation helps decrease bone loss in healthy postmenopausal women. Osteoporosis International. — PubMed PMID: 23525894
8. Knapen MH, Braam LA, Drummen NE, Bekers O, Hoeks AP, Vermeer C (2015). Menaquinone-7 supplementation improves arterial stiffness in healthy postmenopausal women. A double-blind randomised clinical trial. Thrombosis and Haemostasis. — PubMed PMID: 25694037
9. Beulens JW, Bots ML, Atsma F, Bartelink ML, Prokop M, Geleijnse JM, et al. (2009). High dietary menaquinone intake is associated with reduced coronary calcification. Atherosclerosis. — PubMed PMID: 18722618
10. Schurgers LJ, Vermeer C (2000). Determination of phylloquinone and menaquinones in food. Effect of food matrix on circulating vitamin K concentrations. Haemostasis. — PubMed PMID: 11356998
11. Kurosawa Y, Nirengi S, Homma T, Esaki K, Ohta M, Clark JF, Hamaoka T (2015). A single-dose of oral nattokinase potentiates thrombolysis and anti-coagulation profiles. Scientific Reports. — PubMed PMID: 26109079
12. Hodis HN, Mack WJ, Meiselman HJ, Kalra V, Liebman H, Hwang-Levine J, et al. (2021). Nattokinase atherothrombotic prevention study: A randomized controlled trial. Clinical Hemorheology and Microcirculation. — PubMed PMID: 33843667
13. Kobayashi K, Horii Y, Watanabe S, Kubo Y, Koguchi K, Hoshi Y, Matsumoto KI, Soda K (2017). Comparison of soybean cultivars for enhancement of the polyamine contents in the fermented soybean natto using Bacillus subtilis (natto). Bioscience, Biotechnology, and Biochemistry. — PubMed PMID: 28052719
14. Soda K, Uemura T, Sanayama H, Igarashi K, Fukui T (2021). Polyamine-Rich Diet Elevates Blood Spermine Levels and Inhibits Pro-Inflammatory Status: An Interventional Study. Medical Sciences (Basel). — PubMed PMID: 33805535

PubMed Topic Searches

1. PubMed: Natto intake and mortality cohorts
2. PubMed: Fermented soy and cardiovascular disease in Japan
3. PubMed: Natto, menaquinone-7 and fracture

Back to Table of Contents

Connections

• All Food
• Natto
• Fermented Foods
• Natto Benefits
• Natto Vitamin K2 (MK-7)
• Nattokinase and Fibrinolysis
• Natto for Bone Density
• Natto Safety and Interactions
• How to Make Natto
• Bacillus subtilis and Gut Health
• Nattokinase
• Nattokinase and Stroke Prevention
• Vitamin K2
• Vitamin K and Arterial Calcification
• Spermidine
• Stroke
• Atherosclerosis
• Osteoporosis
• PT/INR
• Longevity

Back to Table of Contents

Featured Videos