Vitamin K2 (MK-7 and MK-4)

What vitamin K2 is

Vitamin K is not a single substance but a group of substances. Vitamin K1, phylloquinone, is found in green leaves. Vitamin K2, the menaquinones, are formed during fermentation and are found in cheese, egg yolk and above all in natto. They are numbered by the length of their side chain, from MK-4 to MK-13. Natto is the richest known source: the NIH gives 850 µg of vitamin K per 3-ounce serving, almost exclusively as MK-7.

The task is the same everywhere. Vitamin K is a cofactor in gamma-carboxylation — in the liver for the clotting factors, in bone for osteocalcin, in the vessel wall for matrix Gla protein. Without vitamin K, these proteins remain undercarboxylated and non-functional. The inactive forms are measurable in the blood and are the currency in which almost all K2 studies calculate.

MK-4 and MK-7 are not the same

Advertising lumps both together under the K2 label; pharmacologically they are worlds apart. In a comparative study, a single dose of 420 µg MK-7 reached its peak serum level after 6 hours and was still detectable after 48 hours. MK-4 in the same amount was not measurable in any participant, and even 60 µg daily for 7 days did not raise the level.

MK-7 also has a very long half-life, accumulates 7- to 8-fold with continued intake and carboxylates osteocalcin more completely than vitamin K1. That is why it dominates the shelf, and that is why the studies work with micrograms: 180, 360 or 720 µg daily.

MK-4 comes from another world. In Japan, as menatetrenone, it is a medicinal product for osteoporosis, dosed in milligrams — the Japanese phase IV study used 15 mg capsules, a North American comparative study 45 mg daily. Anyone citing Japanese fracture data for an MK-7 product is transferring milligrams to micrograms. This is the most common mistake in communication about vitamin K2.

The story of the traffic cop for calcium

The story goes like this: vitamin D increases calcium absorption, without K2 the calcium ends up in the arteries instead of in bone, and K2 redirects it. The biochemical core is correct — matrix Gla protein strongly inhibits vascular calcification, and its activation depends on vitamin K. The leap to the claim that a capsule redirects calcium in humans is a hypothesis. And it has been tested in randomized trials.

The result is mostly sobering. In men with a calcified aortic valve, in type 2 diabetes with cardiovascular disease and on dialysis, dp-ucMGP fell markedly every time, but calcification progressed just as quickly as on placebo. The chain of action does not break at the beginning but at the end.

There is a counterweight. A study published in 2026 in patients with coronary heart disease found a slower increase in coronary calcium after 2 years of MK-7; the authors themselves consider the clinical significance open. It remains untested whether vitamin K2 lowers heart attacks or mortality: a Cochrane review in 2015 found a single suitable study with 60 participants over 3 months, and it did not record these endpoints.

What the studies show

3 years of MK-7 in 244 postmenopausal women

Double-blind and placebo-controlled: 180 µg MK-7 daily for 3 years in 244 healthy postmenopausal women, 124 on placebo and 120 on MK-7. The age-related decline in bone density was smaller at the lumbar spine and femoral neck, but not at the total hip. In the vascular analysis, pulse wave velocity and stiffness index fell. Fractures were not an endpoint.

AVADEC: 720 µg MK-7 plus vitamin D against aortic valve calcification

Randomized, double-blind, multicenter: 365 men with an aortic valve calcium score above 300 units, average age 71.0 years, took 720 µg MK-7 plus 25 µg vitamin D or placebo daily for 24 months. The calcium score rose by 275 units on the active treatment and by 292 on placebo; the difference of 17 units was not significant, nor were valve opening area and coronary calcium. The primary endpoint was missed, although dp-ucMGP fell markedly.

VitaK-CAC: 2 years of MK-7 in coronary heart disease

Placebo-controlled at two Dutch hospitals: 180 randomized patients with symptomatic coronary heart disease and a coronary calcium score between 50 and 400 units, of whom 85 on 360 µg MK-7 daily and 82 on placebo were analyzed. After 2 years, the median calcium score rose from 145 to 214 units on placebo and from 135 to 184 on MK-7. The difference was significant; the authors consider the clinical significance open.

The Japanese fracture data and what is left of them

The much-cited 2006 meta-analysis found very large effects for menaquinone-4: 7 studies, all from Japan, odds ratios of 0.40 for vertebral fractures, 0.23 for hip fractures and 0.19 for all non-vertebral fractures. The largest Japanese study itself, open-label and in 4,378 women over 36 months, missed its primary endpoint. A 2019 review arrived at an odds ratio of 0.72, which rose to 0.76 and lost significance when restricted to studies with a low risk of bias. In 2020, all estimates that had included the Yamaguchi study (YOPS) were corrected, because its integrity is in question.

Where the data stop

The statement that vitamin K2 prevents vascular calcification does not hold up against the randomized trials. In three out of four studies, calcification remained unaffected, although the marker responded clearly. Moreover, what is measured are calcium scores on CT, that is, surrogate markers and not heart attacks.

The observational data, too, are less consistent than their reputation. The Rotterdam Study found a relative risk of 0.43 for death from coronary heart disease and an odds ratio of 0.48 for severe aortic calcification for the top third of menaquinone intake. A later cohort with 33,289 participants over an average of 16.8 years, by contrast, found no association with mortality.

The bone data are weaker than the story suggests. Two studies outside Japan were completely negative: 381 women on 45 mg MK-4 or 1 mg vitamin K1 and 334 women on 360 µg MK-7, each over 12 months and with no effect on bone density, although the marker responded.

The synergy with vitamin D3 is also an assumption. The only large trial with MK-7 and vitamin D against placebo was negative; an analysis in 200 participants found no indication that vitamin K status changes the effect of vitamin D.

Status, approval and legal

Vitamin K is an authorized nutrient for food supplements in the EU. Exactly two health claims are permitted: that vitamin K contributes to normal blood clotting and to the maintenance of normal bones. A claim relating to the heart and blood vessels is missing from the list in Regulation (EU) No 432/2012 and may not be used. A tolerable upper intake level was never set, because the data were insufficient. For food supplements, the BfR proposes maximum levels of 80.0 µg vitamin K1 and 25.0 µg vitamin K2 per recommended daily dose, plus a warning for people on anticoagulants; commercially available MK-7 products are well above this.

Safety

The most important point is the interaction with vitamin K antagonists. In a dose-response study in 18 healthy people who had been set to an INR of 2.0 with acenocoumarol, 45 µg MK-7 daily lowered INR and uncarboxylated factor II by around 40 percent. Even 10 µg and 20 µg lowered the INR to a clinically relevant extent in at least 40 and 60 percent of participants respectively, as judged by two hematologists; the authors conclude that MK-7 products should be avoided under this therapy. The BfR puts the potency of K2 against coumarins at around 3.5 times that of K1. Anyone taking phenprocoumon or warfarin should clarify vitamin K2 with a physician beforehand. Apart from that, vitamin K is considered to have low toxicity: in the studies with 180, 360 and 720 µg MK-7 over 2 to 3 years, no notable safety problems occurred. In the Japanese fracture study with menatetrenone in the milligram range, adverse events were more frequent in the combination group. Data on pregnancy, breastfeeding and children are lacking.

Sources

• Knapen et al., Osteoporosis International 2013
• Knapen et al., Thrombosis and Haemostasis 2015
• Diederichsen et al., Circulation 2022 (AVADEC)
• Vossen et al., JAMA Cardiology 2026 (VitaK-CAC)
• Mott et al., Osteoporosis International 2019
• Inoue et al., Journal of Bone and Mineral Metabolism 2009 (OF study)
• Sato et al., Nutrition Journal 2012
• Theuwissen et al., Journal of Thrombosis and Haemostasis 2013
• BfR, proposed maximum levels for vitamin K in foods including food supplements
• Regulation (EU) No 432/2012, list of permitted health claims