Vitamin K2 benefits

This page takes up five commonly heard claims about K2 one by one: clotting, fracture, bone density, arterial stiffness and vascular-valve calcification. General information, food sources and reference intakes are in the vitamin K2 guide.

How is a role in the body separated from the benefit of a supplement?

Vitamin K is the cofactor of the carboxylation enzyme needed for some proteins to become functional. Besides the clotting proteins, osteocalcin in bone and matrix Gla protein (MGP), found in vessels, bone and cartilage, also depend on it. ^[1] This mechanism is the basis of marketing sentences such as “K2 directs calcium to the bone and away from the vessel.” However, a protein depending on vitamin K does not show that extra supplements prevent a disease. The steps of evidence are separate: a change in a blood marker (for example dp-ucMGP or uncarboxylated osteocalcin), a change in intermediate measures such as bone density, and a change in real outcomes such as fracture or a cardiovascular event are different questions.

What does vitamin K do for clotting?

Clotting factors such as prothrombin cannot work without vitamin K; for this reason, the classic sign of deficiency is bleeding. NIH reports that clinically significant deficiency is very rare in adults, mostly seen in people with malabsorption or those using drugs that affect vitamin K metabolism. In a person eating a healthy and varied diet, reaching an intake low enough to change clotting measurements is almost impossible. ^[1] For this reason, being “necessary for clotting” does not mean that a healthy adult needs supplements.

Does K2 prevent fractures?

Japanese studies and the updated meta-analysis. NIH’s 2021 fact sheet reports that in a 2006 meta-analysis all seven studies with fracture data were conducted in Japan, six used 45 mg per day and one 15 mg of MK-4, and fractures decreased significantly. These are pharmacological doses far above the amounts found in dietary supplements (µg level) and are used for osteoporosis treatment in Japan and some regions of Asia. ^[1]

This meta-analysis was later corrected: the integrity of some of the included studies was questioned and the authors updated the analysis. In the 2019 update covering 36 studies (Mott et al.), there was no difference between vitamin K and control for vertebral fracture in postmenopausal or osteoporotic patients (OR 0.98; 95% CI 0.85–1.14). For any clinical fracture the OR was 0.74 (0.56–0.97); when only studies at low risk of bias were taken, the OR was 0.76 (0.58–1.01) and significance disappeared. Small differences in lumbar spine density also ceased to be significant when studies at high risk of bias were removed. The authors’ conclusion: there is no evidence in this group that vitamin K affects bone density or vertebral fracture; it may reduce clinical fracture, but the evidence is insufficient to confirm this. For other patient groups, the number of studies is insufficient to draw a conclusion. ^{[2] [9]}

The ECKO trial. In this single-center trial, 440 postmenopausal women with osteopenia took 5 mg of K1 per day or placebo; participants’ vitamin D levels were adequate (mean 25(OH)D 77 nmol/L). Over two years, the change in lumbar spine and total hip bone density did not differ between groups. The number of clinical fractures was lower in the K group (nine versus twenty, p=0.04); however, the authors state that the trial was not powered for fracture and that the numbers are small. ^[3] This trial was done with K1; it gives no result for MK-4 or MK-7.

Therefore, the sentence “K2 prevents fractures” is not supported by the current evidence: the strong effect in the older Japanese studies was not confirmed when the problematic studies were screened out, and the other studies gave small and mixed results. The answer to the same question with doses at the dietary supplement level was not measured in these studies.

Does MK-7 affect bone density?

In the study by Knapen and colleagues, 244 healthy postmenopausal women took 180 µg of MK-7 per day or placebo for 3 years. In the MK-7 group, vitamin K status (osteocalcin marker) improved; the age-related decline in bone mineral content and density at the lumbar spine and femoral neck was smaller, and no difference was seen at the total hip. An effect in one region and not in another suggests that the result is small and region-specific. The summary results reported are indicators of density, bone strength and vertebral height; the authors note that whether the results can be generalized to men and children is a question to be investigated separately. ^[4]

NIH also reports that in a 12-month study 45 mg of MK-4 or 1 mg of phylloquinone improved osteocalcin carboxylation but did not make a difference in lumbar spine or proximal femur bone density in 381 postmenopausal women; NIH points out that giving vitamin D and calcium supplements together with K can confound the results. ^[1]

Does K2 affect arterial stiffness and calcification?

Arterial stiffness (intermediate measure). In the 2015 double-blind study by Knapen and colleagues, in 244 healthy postmenopausal women 3 years of 180 µg MK-7 significantly lowered carotid-femoral pulse wave velocity and the “stiffness index β” in the whole group; local carotid indicators improved only in women whose baseline stiffness index β was above the median (10.8). MK-7 reduced dp-ucMGP by about half compared with placebo. ^[5] This is a vascular stiffness indicator, not a cardiovascular event; the outcomes the study measured are vascular stiffness and blood markers.

Aortic valve calcification. In the AVADEC trial, 365 men with a calcium score above 300 AU and a mean age of 71 took 720 µg of MK-7 plus 25 µg of vitamin D per day or placebo for 24 months. The increase in valve calcium score was 275 AU versus 292 AU (difference −17 AU; 95% CI −86 to 53; P=0.64). Progression of aortic and coronary calcification, valve area and peak flow velocity were also no different; the numbers of heart valve surgeries, deaths and cardiovascular events were similar. dp-ucMGP fell, that is, the biochemical marker changed but the disease marker did not. ^[6]

Dialysis patients. In the Valkyrie trial, 132 hemodialysis patients with atrial fibrillation took a vitamin K antagonist, rivaroxaban, or rivaroxaban plus 2,000 µg of K2 three times a week for 18 months. Vitamin K status improved in the K2 group but still remained high; coronary, thoracic aortic and valve calcium scores and pulse wave velocity did not differ between groups. ^[7]

These three studies are in different populations and do not answer the same question: while the arterial stiffness indicator improved in healthy women, valve and vascular calcification did not change in two disease populations. These earlier results do not establish a general preventive effect; the sentence “it may improve a specific intermediate measure” rests on a single study and has not been translated into clinical events.

Evidence summary: result, population and limit

Safety: expectation of benefit does not remove drug interaction

In people using warfarin and similar vitamin K antagonists, MK-7 supplements can disrupt treatment: Schurgers and colleagues reported that preparations providing 50 µg per day or more of MK-7 can interact with oral anticoagulant therapy in a clinically significant way. ^[8] NIH states that no upper limit has been set for vitamin K, but that people using these drugs should keep their K intake consistent. ^[1] This page does not recommend doses.

Why the 2020 correction matters

Mott and colleagues removed the YOPS trial after concerns about data integrity. The corrected odds ratio was 0.74 (95% CI 0.56–0.97) for clinical fractures and 0.98 (0.85–1.14) for vertebral fractures. Low-risk-of-bias estimates and the authors’ overall interpretation were unchanged. These results do not establish that K2 prevents fractures in everyone. The numerical summary above uses the corrected values. ^[9]

The 2026 VitaK-CAC trial: positive imaging, uncertain clinical benefit

In symptomatic coronary artery disease, 180 people were randomized; 85 received MK-7 and 82 received placebo. Coronary calcium-score progression was slower with MK-7 over two years (between-group p=0.02). Scores increased in both groups, so this is not evidence of “cleaning arteries.” The trial measured imaging; implications for plaque stability, heart attacks and survival remain uncertain. It studied a different population from earlier valve and dialysis trials, so those findings should not be collapsed into one universal efficacy claim. ^[10]