Vitamin K2 (Menaquinone)

Overview

Vitamin K2 (menaquinone) is the branch of vitamin K that activates the proteins deciding where calcium ends up: osteocalcin, which binds it into bone, and matrix Gla protein, which keeps it out of artery walls. Gut bacteria make some, fermented and animal foods carry a little, and the long-chain MK-7 form stays in the blood for days instead of hours. The catch is that spinach does not cover this job: K1 from greens goes mostly to the liver for clotting.

What is Vitamin K2 (Menaquinone)?

Key takeaways

• MK-7 is still measurable in blood 48 hours after a single 420 mcg dose and climbs to 7 to 8 times its starting level over weeks of daily use; MK-4 at the same dose never appeared in serum at all (Sato 2012, Schurgers 2007).
• The strongest MK-7 bone trial, 244 postmenopausal women taking 180 mcg a day for three years, lost less bone density at the lumbar spine and femoral neck than placebo, but not at the total hip (Knapen 2013, Osteoporosis International).
• Every MK-7 trial that measured it moved osteocalcin carboxylation within 4 to 12 weeks; in healthy adults the dose that did it was around 90 mcg a day or more (Theuwissen 2012, Dalmeijer 2012).
• The artery evidence is thinner than the label copy: one 3-year trial found less arterial stiffness in healthy postmenopausal women, while trials in kidney disease, diabetes, and aortic valve calcification found no effect on the vessels (Knapen 2015; Witham 2020; Zwakenberg 2019; Diederichsen 2022).
• As little as 10 mcg of MK-7 a day changed anticoagulation in healthy volunteers taking acenocoumarol, which is why anyone on warfarin or a related drug does not add K2 without talking to their prescriber (Theuwissen 2013).
• There is no upper intake level for vitamin K because no toxicity has been reported from food or supplements; the U.S. Adequate Intake for total vitamin K is 120 mcg for men and 90 mcg for women (NIH ODS).

Vitamin K2 is the group of vitamin K molecules called menaquinones. Every form of vitamin K shares the same core, a 2-methyl-1,4-naphthoquinone ring, and the forms differ only in the tail hanging off it. Vitamin K1 (phylloquinone) has one partly saturated tail and comes from leafy greens. Vitamin K2 has a tail made of repeating five-carbon isoprene units, and the number of units gives each menaquinone its name: MK-4 has four, MK-7 has seven, and the series runs to MK-13. That quinone ring puts vitamin K in the same chemical family as coenzyme Q10, which is why both are fat-soluble, both travel in lipoproteins, and both need fat in the meal to absorb.

The K1 versus K2 distinction is the whole story, and it is the one most people miss. After a meal, K1 is pulled into the liver, where it keeps the clotting factors working, and very little of it reaches bone or artery wall. In a head-to-head trial in healthy volunteers (Schurgers 2007, Blood), K1 and MK-7 were both absorbed well, with peak blood levels about four hours after the dose. The difference was what happened next. K1 cleared within hours. MK-7 stayed, built up to seven or eight times its starting level during weeks of daily intake, and produced more complete carboxylation of osteocalcin, the bone protein, than K1 did. "I eat spinach" settles the clotting question. It does not settle this one.

Menaquinones come mostly from bacteria. The one food that carries a meaningful amount is natto, soybeans fermented with Bacillus subtilis; the NIH Office of Dietary Supplements lists about 850 mcg of MK-7 in a 3-ounce serving, more than seven times the daily Adequate Intake for total vitamin K. After natto the numbers fall off a cliff. Rotisserie chicken breast has about 13 mcg of MK-4 per 3 ounces, cheddar about 4 mcg per 1.5 ounces, a hard-boiled egg about 4 mcg. Bacteria in the human large bowel make long-chain menaquinones too, but how much the body absorbs from that far down the gut is unclear, and the NIH fact sheet says so plainly. MK-4 is the odd one out: tissues build it from K1 by swapping the tail, without any bacteria involved.

Average American adults get 122 mcg (women) to 138 mcg (men) of vitamin K a day from food, almost all of it K1, and analyses of NHANES data from 2003 to 2010 found only about one in three people above the Adequate Intake. Those figures undercount K2, because the USDA food database barely tracks menaquinones. Clinically visible vitamin K deficiency, meaning bleeding, is rare outside newborns and people with fat malabsorption. Undercarboxylated bone and vessel proteins are a different matter. They are common in healthy adults, as the next section shows.

The body handles K2 like any fat. Bile and pancreatic enzymes fold it into micelles, the small intestine absorbs it, chylomicrons carry it to the liver, and the liver repackages it into lipoproteins that deliver it to bone, heart, pancreas, and brain. Only 30 to 40 percent of an oral dose is retained; the rest leaves in urine and bile within days. That fast turnover is why blood levels of vitamin K are tiny compared with vitamins A, D, and E, and why daily intake matters more than a big occasional dose. The practical consequence is the half-life row in the table above. A single 420 mcg dose of MK-7 was still measurable in healthy women two days later, while the same dose of MK-4 never registered in serum at any time point (Sato 2012, Nutrition Journal).

The reason anyone cares about K2 specifically dates to 2004. The Rotterdam Study followed 4,807 Dutch adults aged 55 and older for about seven years and found that people in the top third of menaquinone intake (more than 32.7 mcg a day, mostly from cheese) had 57 percent lower coronary heart disease mortality and about half the odds of severe aortic calcification compared with the bottom third. K1 intake was linked to nothing (Geleijnse 2004, Journal of Nutrition). That is an observational finding in people who ate cheese, not a trial of capsules, and the randomized trials that followed are graded honestly below. But it is why K2 went from a footnote to a category.

K2 is usually sold next to vitamin D3, and the pairing makes mechanistic sense: D3 raises the amount of calcium the gut absorbs, and K2 activates the proteins that decide where that calcium lands. The vitamin D guide covers the D side; whether the pairing has trial evidence of its own is covered in the evidence section and in the D3 and K2 combination article.

How it works

Vitamin K has one job in the body: it is the cofactor for gamma-glutamyl carboxylase, an enzyme that adds a carboxyl group to specific glutamate residues on a short list of proteins. That modification, called carboxylation, gives the protein a pair of negatively charged arms that can grip calcium. Without it, the protein is still made, still circulates, and cannot hold calcium. Researchers call that state undercarboxylated, and the share of a protein that is undercarboxylated is the lab marker nearly every K2 trial uses.

The enzyme burns vitamin K in the process, turning it into an epoxide, and a second enzyme, vitamin K epoxide reductase (VKOR), recycles the epoxide back to active vitamin K. That loop is the vitamin K cycle. Warfarin and its relatives work by jamming VKOR, so the clotting factors stop getting carboxylated and blood clots more slowly. Any vitamin K you swallow bypasses the jam. That single sentence is the entire warfarin interaction, and it is why the safety section leads with it.

Three vitamin K-dependent proteins matter for this page. Prothrombin and the other clotting factors are made in the liver, and the liver gets first call on vitamin K, so they are essentially fully carboxylated in anyone eating a normal diet. Osteocalcin is made by osteoblasts, the bone-building cells, and in its carboxylated form it binds to hydroxyapatite, the mineral crystal of bone. Matrix Gla protein (MGP) is made in vascular smooth muscle and cartilage, and in its carboxylated form it blocks calcium crystals from growing in the artery wall. In a dose-response trial of 42 healthy Dutch adults, 10 to 40 percent of osteocalcin and MGP was undercarboxylated at baseline even though clotting was normal, and MK-7 at doses around the European reference intake raised carboxylation of both without touching thrombin generation (Theuwissen 2012, British Journal of Nutrition). The liver was satisfied. Bone and vessel were not.

The MGP marker deserves its own paragraph because it carries most of the cardiovascular marketing. The inactive form is called dp-ucMGP (desphospho-uncarboxylated MGP), and it drops reliably when people take MK-7. In 60 adults aged 40 to 65, twelve weeks of 180 mcg cut dp-ucMGP by 31 percent and 360 mcg cut it by 46 percent, while the osteocalcin ratio fell 60 and 74 percent (Dalmeijer 2012, Atherosclerosis). Lower dp-ucMGP is a real, reproducible biochemical change. Whether it translates into less calcium in the artery wall is a separate question, and the randomized trials that measured actual calcification mostly say it does not, at least over six months to two years in people who already have it. The table below keeps those two outcomes on separate rows for that reason.

Why MK-7 rather than MK-4 or K1? Delivery. The long isoprene tail makes MK-7 ride in low-density lipoproteins, which circulate for days and drop cargo at bone and vessel wall. MK-4 and K1 are carried mostly in chylomicron remnants that the liver clears within hours. In healthy Japanese women, a 420 mcg dose of MK-7 peaked at six hours and was still measurable at 48, while 420 mcg of MK-4 never appeared in serum; a week of 60 mcg daily raised MK-7 levels and did nothing to MK-4 levels (Sato 2012). The 45 mg MK-4 dose used as a prescription drug in Japan gets around this by brute force, at roughly 375 times a typical supplement.

What a person notices from any of this is nothing. There is no felt effect. The markers move in four to twelve weeks, bone density differences took one to three years in the trials that found them, and arterial stiffness took three years. The mechanism predicts two outcome families, bone and vessel, and the evidence below grades each one on what the trials actually measured.

Evidence at a glance

What the evidence doesn't support yet

The cardiovascular claim is the big one and it is ahead of the data. The observational link from Rotterdam is real, and MK-7 reliably lowers the inactive MGP marker, but the trials that then measured calcium in the vessel wall came up empty: no change on CT in 68 adults with diabetes after six months of 360 mcg (Zwakenberg 2019), no change in valve, aortic, or coronary calcification in 365 older men after two years of 720 mcg plus vitamin D (Diederichsen 2022), and no change in arterial stiffness in 159 people with kidney disease after a year of 400 mcg (Witham 2020). A brand that prints "supports heart health" on a K2 bottle is leaning on the cheese-eaters in Rotterdam and hoping you do not ask. We do not print it. Bone is where the trial evidence is, and even there it is moderate.

Fracture prevention at supplement doses is the second claim that outruns the evidence. The fracture numbers people quote come from Japanese trials of 45 mg MK-4 a day, a prescription dose about 375 times a 120 mcg capsule (Cockayne 2006), and a 2022 meta-analysis of 16 K2 trials found no overall reduction in fractures (Ma 2022).

Two smaller gaps. No MK-7 trial has shown a bone density benefit in men (Zhang 2020 tested them and found none) or in women within five years of menopause (Emaus 2009, one year, nothing). And adding calcium and vitamin D on top of K2 added nothing measurable in the one trial that tested the stack directly (Zhang 2020).

Clinical evidence

• Three-year low-dose menaquinone-7 supplementation helps decrease bone loss in healthy postmenopausal women — 244 healthy postmenopausal women in the Netherlands, 3 years, 180 mcg/day MK-7 (MenaQ7) capsules versus placebo. Measured DXA bone mineral density and content at lumbar spine, total hip, and femoral neck, femoral neck strength indices, vertebral height, and the ucOC/cOC ratio. MK-7 improved vitamin K status and reduced the age-related decline in BMC and BMD at the lumbar spine and femoral neck but not at the total hip, improved bone strength indices, and reduced loss of vertebral height in the lower thoracic spine. Limitation: one population of healthy Dutch women, and the trial was not powered for fractures.
• Menaquinone-7 supplementation improves arterial stiffness in healthy postmenopausal women. A double-blind randomised clinical trial — The same 244 postmenopausal women, 3 years, 180 mcg/day MK-7 versus placebo. Measured carotid-femoral pulse wave velocity, carotid stiffness index, intima-media thickness, dp-ucMGP, and inflammatory and endothelial markers. Pulse wave velocity and the stiffness index fell in the MK-7 group overall, with larger local carotid improvements in women whose baseline stiffness index was above the median of 10.8; dp-ucMGP fell 50 percent versus placebo; inflammation and endothelial markers did not change. Limitation: a vascular analysis of a bone trial in healthy women, with the effect concentrated in the stiffer half.
• Effect of Low-Dose Vitamin K2 Supplementation on Bone Mineral Density in Middle-Aged and Elderly Chinese: A Randomized Controlled Study — 311 community-dwelling Chinese men and postmenopausal women aged 50 to 75, 1 year, randomized to placebo, 50 mcg/day MK-7, 90 mcg/day MK-7, or 90 mcg MK-7 plus 500 mg calcium and 10 mcg vitamin D3. Measured DXA BMD and the cOC/ucOC ratio. Femoral neck bone loss was lower in postmenopausal women in both 90 mcg groups than on placebo; 50 mcg did not differ; men showed no effect; adding calcium and D brought no additional benefit. Limitation: one year, one site, and only the femoral neck in women moved.
• Vitamin K2 (menaquinone-7) prevents age-related deterioration of trabecular bone microarchitecture at the tibia in postmenopausal women — 148 postmenopausal women with osteopenia in Denmark, 12 months, 375 mcg/day MK-7 versus placebo, all on calcium and vitamin D. Measured DXA BMD, HR-pQCT bone microarchitecture at tibia and radius, and ucOC. ucOC fell 65.6 percent on MK-7 versus 6.4 percent on placebo by month 3; at the tibia, trabecular number and spacing were preserved on MK-7 while the placebo group lost trabeculae; no difference in BMD at any site or in radius microarchitecture. Limitation: one year, a structural surrogate rather than a fracture outcome, and the radius did not replicate the tibia.
• Vitamin K2 supplementation does not influence bone loss in early menopausal women: a randomised double-blind placebo-controlled trial — 334 healthy Norwegian women aged 50 to 60, 1 to 5 years past menopause, 12 months, 360 mcg/day MK-7 as natto capsules versus olive-oil placebo. Measured DXA BMD at total hip, femoral neck, lumbar spine, and total body, plus bone turnover markers and osteocalcin species. Carboxylated osteocalcin rose and undercarboxylated fell on MK-7 (p < 0.001), but bone loss rates did not differ between groups at any site. Limitation: one year may be too short for a BMD endpoint in early menopause, when bone loss is fastest and hormonally driven.
• Low-dose menaquinone-7 supplementation improved extra-hepatic vitamin K status, but had no effect on thrombin generation in healthy subjects — 42 healthy Dutch men and women aged 18 to 45, 12 weeks, randomized to placebo or 10, 20, 45, 90, 180, or 360 mcg/day MK-7. Measured circulating ucOC, cOC, dp-ucMGP, and thrombin generation. Doses around the European reference intake and above raised carboxylation of both osteocalcin and MGP; lower doses did not reliably; thrombin generation was unchanged at every dose. Limitation: six participants per arm, so groups were collapsed into low-dose and high-dose for analysis.
• The effect of menaquinone-7 supplementation on circulating species of matrix Gla protein — 60 healthy adults aged 40 to 65 in the Netherlands, 12 weeks, 180 mcg/day or 360 mcg/day MK-7 versus placebo. Measured dp-ucMGP, dp-cMGP, t-ucMGP, the ucOC/cOC ratio, and cardiovascular risk factors at baseline, 4, and 12 weeks. dp-ucMGP fell 31 percent on 180 mcg and 46 percent on 360 mcg; the osteocalcin ratio fell 60 and 74 percent; other MGP species and risk factors did not change. Limitation: a biomarker study with no structural or clinical endpoint.
• Comparison of menaquinone-4 and menaquinone-7 bioavailability in healthy women — Healthy Japanese women given single 420 mcg doses of MK-4 or MK-7 with a standard breakfast, then 60 mcg/day of each for 7 days. Measured serum MK-4 and MK-7 by HPLC. MK-7 peaked at 6 hours and remained detectable at 48 hours; MK-4 was undetectable in every subject at every time point; a week of 60 mcg MK-7 raised serum MK-7 in all subjects while 60 mcg MK-4 did not raise serum MK-4. Limitation: a small pharmacokinetic study in young women with serum levels only and no functional outcome.
• The effect of menaquinone-7 supplementation on vascular calcification in patients with diabetes: a randomized, double-blind, placebo-controlled trial — 68 adults with type 2 diabetes and known cardiovascular disease in the Netherlands, 6 months, 360 mcg/day MK-7 versus placebo. Measured femoral arterial calcification activity by 18F-sodium fluoride PET and calcification mass by CT, plus dp-ucMGP. PET activity tended to increase on MK-7 (p = 0.06), CT calcification mass did not differ, dp-ucMGP fell by about 206 pmol/L, and no adverse events were reported. Limitation: small, short, and the meaning of the PET signal is uncertain.
• Vitamin K2 and D in Patients With Aortic Valve Calcification: A Randomized Double-Blinded Clinical Trial — 365 Danish men, mean age 71, with aortic valve calcification scores above 300 AU, 24 months, 720 mcg/day MK-7 plus 25 mcg vitamin D versus placebo. Measured change in aortic valve calcification score, valve area, jet velocity, aortic and coronary calcification, dp-ucMGP, and clinical events. Valve calcification progressed 275 AU on treatment and 292 AU on placebo, a non-significant 17 AU difference; aortic and coronary calcification progression did not differ; dp-ucMGP fell 212 pmol/L versus a 45 pmol/L rise on placebo; deaths and cardiovascular events did not differ. Limitation: men only, with established calcification, and vitamin D was co-administered.
• Vitamin K Supplementation to Improve Vascular Stiffness in CKD: The K4Kidneys Randomized Controlled Trial — 159 adults with stage 3b or 4 chronic kidney disease in the United Kingdom, mean age 66, 12 months, 400 mcg/day oral vitamin K2 versus placebo. Measured carotid-femoral pulse wave velocity, augmentation index, abdominal aortic calcification, blood pressure, physical function, and mineral markers. No difference in pulse wave velocity, augmentation index, blood pressure, or physical function at 12 months; an updated meta-analysis including this trial found no effect of vitamin K on stiffness or calcification. Limitation: a kidney disease population, where calcification biology differs from the general population.
• Vitamin K and the prevention of fractures: systematic review and meta-analysis of randomized controlled trials — 13 randomized trials of oral phytonadione or menaquinone lasting over 6 months in adults, 7 of them reporting fractures, all Japanese and all using MK-4 at 45 mg/day (six trials) or 15 mg/day (one). Measured bone loss and vertebral, hip, and nonvertebral fractures. Twelve of 13 trials showed less bone loss on vitamin K; pooled fracture odds ratios favored MK-4 at 0.40 for vertebral, 0.23 for hip, and 0.19 for all nonvertebral fractures. Limitation: pharmacological MK-4 doses in Japanese patients, not supplement-dose MK-7, and later trials have been less consistent.
• Efficacy of vitamin K2 in the prevention and treatment of postmenopausal osteoporosis: A systematic review and meta-analysis of randomized controlled trials — 16 randomized controlled trials with 6,425 postmenopausal women, published before March 2022, of vitamin K2 alone or combined with other therapy. Measured lumbar spine BMD, fracture incidence, ucOC, cOC, and adverse reactions. Lumbar spine BMD improved with K2 (p = 0.006); fracture incidence did not differ overall (RR 0.96, p = 0.65) and reached significance only after excluding one heterogeneous trial (RR 0.43); ucOC and the ucOC/cOC ratio fell; adverse reactions matched control (RR 1.03). Limitation: pooled MK-4 drug doses with MK-7 supplement doses, and the fracture result depended on a post-hoc exclusion.
• Efficacy and safety of vitamin K2 for postmenopausal women with osteoporosis at a long-term follow-up: meta-analysis and systematic review — 9 randomized controlled trials with 6,853 postmenopausal women with osteoporosis followed long term, searched to November 2021. Measured percent change in lumbar and forearm BMD, ucOC, osteocalcin, and adverse reactions. Lumbar BMD rose 2.17 percent and forearm BMD 1.57 percent more than control; ucOC fell and osteocalcin rose; adverse reaction rates were higher on K2 (RR 1.33) with no serious events. Limitation: dominated by Japanese 45 mg MK-4 trials, so it says little about 100 to 180 mcg MK-7.
• Vitamin K status, supplementation and vascular disease: a systematic review and meta-analysis — 13 controlled trials (n = 2,162) of vitamin K supplementation measuring vascular calcification, vascular stiffness, or vitamin K-dependent protein levels, plus 14 longitudinal studies (n = 10,726). Measured pooled change in calcification, stiffness, dp-ucMGP, and ucOC, and the association of inactive vitamin K-dependent proteins with cardiovascular events. Supplementation reduced vascular calcification by 9.1 percent and dp-ucMGP by 44.7 percent versus control, with a non-significant change in stiffness; vitamin K-dependent protein levels were associated with the combined endpoint of cardiovascular disease or death (HR 0.45). Limitation: few trials with substantial heterogeneity, mixing K1 and K2 across diverse populations.
• Vitamin K-containing dietary supplements: comparison of synthetic vitamin K1 and natto-derived menaquinone-7 — Healthy volunteers given synthetic vitamin K1 or natto-derived MK-7 over prolonged daily intake in a multicenter comparative study. Measured serum vitamin K species as the absorption marker and osteocalcin carboxylation as the activity marker. Both forms were absorbed well with peak serum levels at 4 hours; MK-7 had a far longer half-life, more stable serum levels, and accumulated to 7 to 8 times higher levels with prolonged intake, producing more complete osteocalcin carboxylation; the authors warned that 50 mcg/day or more of MK-7 may interfere with oral anticoagulant treatment. Limitation: a pharmacokinetic and marker study without a placebo arm or clinical endpoint.
• Effect of low-dose supplements of menaquinone-7 (vitamin K2) on the stability of oral anticoagulant treatment: dose-response relationship in healthy volunteers — 18 healthy men and women anticoagulated with acenocoumarol for 4 weeks, 15 of whom reached a target INR of 2.0, then given 10, 20, and 45 mcg/day MK-7 in sequence over 6 weeks while continuing their established drug dose. Measured INR, uncarboxylated factor II, ucOC, dp-ucMGP, and endogenous thrombin generation. 45 mcg/day lowered mean INR and uncarboxylated factor II by about 40 percent; 10 and 20 mcg/day produced a clinically relevant INR drop in at least 40 and 60 percent of subjects and raised thrombin generation by about 20 and 30 percent; ucOC and dp-ucMGP did not change at these doses. Limitation: small, sequential dose escalation with no placebo, in volunteers rather than patients.
• Dietary intake of menaquinone is associated with a reduced risk of coronary heart disease: the Rotterdam Study — 4,807 Dutch adults aged 55 and older in the Rotterdam Study with dietary data and no prior heart attack, followed from a 1990 to 1993 baseline until 2000. Measured incident coronary heart disease, coronary mortality, all-cause mortality, and aortic calcification by tertile of energy-adjusted phylloquinone and menaquinone intake. The top tertile of menaquinone intake (above 32.7 mcg/day) had a relative risk of 0.43 for coronary mortality, 0.74 for all-cause mortality, and an odds ratio of 0.48 for severe aortic calcification versus the bottom tertile; phylloquinone intake was unrelated to any outcome. Limitation: food-frequency estimates of dietary menaquinone, largely from cheese, with no supplement arm and the usual confounding of observational data.

Dosage & forms

• MK-7 (menaquinone-7), natto-fermented:
• MK-7 (menaquinone-7), synthetic all-trans:
• MK-4 (menatetrenone) at supplement doses:
• Vitamin K1 (phylloquinone or phytonadione):
• D3 plus K2 combination (liquid drops or capsule):
• Menadione (vitamin K3), synthetic:

Safety & interactions

How much to take

The research dose for MK-7 is 90 to 180 mcg a day, and the most common supplement dose, 100 to 120 mcg, sits at the bottom of that range. The three-year bone and artery trials used 180 mcg (Knapen 2013, 2015). The one trial that tested lower doses head to head found 90 mcg slowed femoral neck bone loss in postmenopausal women and 50 mcg did not (Zhang 2020). In the dose-response study, carboxylation of osteocalcin and MGP moved at doses around the European reference value and above, and not reliably below it (Theuwissen 2012). So a 120 mcg product matches the research at the low end, and a 180 mcg product matches the strongest trial exactly. Below 90 mcg you are buying a label, not a studied dose.

More is not better past about 180 mcg. Going from 180 to 360 mcg improved MGP activation from 31 to 46 percent at 12 weeks (Dalmeijer 2012), a real but shrinking return, and the trials that used 360 to 375 mcg for bone density (Emaus 2009, Rønn 2016) found nothing the 180 mcg trial had not already found. Nobody has shown a bone or vessel benefit from 500 mcg or 1,000 mcg capsules. They are safe. They are also unsupported.

Take it with the meal that has the most fat in it. K2 absorbs the way dietary fat does, through bile and micelles, and an oil-based liquid or softgel does part of that work for you. Once a day is enough for MK-7 because of its multi-day half-life; MK-4 would need several doses a day to stay in the blood, which is one more reason it is the wrong supplement form. If you take a bile acid sequestrant or orlistat, put K2 at a different meal, since both block fat absorption (NIH ODS).

Pairing with vitamin D3 and calcium is standard, and the mechanism supports it, but the trial evidence does not show that adding them to K2 improves bone outcomes beyond K2 alone (Zhang 2020). Pair them because you need vitamin D anyway; how much vitamin D to take is its own question with its own answer.

Expect nothing you can feel. The blood markers shift within 4 to 12 weeks. Bone density differences took one to three years to show on a DXA scan. There is no established upper limit: the Food and Nutrition Board declined to set one because no adverse effects from vitamin K in food or supplements have been reported in humans or animals (NIH ODS).

Safety profile

The long-term safety data for MK-7 are good. The Dutch trials followed 244 postmenopausal women on 180 mcg for three years with no safety signal (Knapen 2013, 2015). The Danish valve trial gave 365 men aged about 71 a dose of 720 mcg MK-7 plus 25 mcg vitamin D for two years and found no difference from placebo in all-cause death (1 versus 4), cardiovascular events (10 versus 10), or valve surgery (Diederichsen 2022). The diabetes trial at 360 mcg reported no adverse events (Zwakenberg 2019). A 2022 meta-analysis of 16 K2 trials with 6,425 participants found adverse reaction rates identical to control (RR 1.03), and a second meta-analysis of 9 longer trials found a higher rate of minor complaints (RR 1.33) but no serious events; that second pool is dominated by Japanese 45 mg MK-4 trials, not supplement-dose MK-7 (Ma 2022; Zhou 2022).

By frequency: common side effects above placebo, none reported in the MK-7 trials. Uncommon, mild stomach upset, reported mainly in the high-dose MK-4 trials pooled by Zhou 2022. Rare, nothing documented. There is no known overdose syndrome and no sign of taking too much to watch for, which is unusual for a fat-soluble vitamin and is the reason there is no upper limit.

Drug interactions

Who should ask a clinician first

Anyone taking warfarin or another vitamin K antagonist, before the first dose, full stop. Pregnant and nursing women, because the MK-7 trials excluded them and the Adequate Intake of 90 mcg total vitamin K is the only number that applies. People with fat malabsorption from celiac disease, cystic fibrosis, short bowel syndrome, ulcerative colitis, or bariatric surgery, who may actually need vitamin K and should have status checked rather than guessed (NIH ODS). People with chronic kidney disease, where calcification is common, trials are still running, and the one completed stiffness trial found no benefit (Witham 2020). Anyone with a history of blood clots or a clotting disorder.

If you take a blood thinner, this is the one supplement that can change your lab result, so the conversation happens before the first dose, not after.

Buying guide

The dose: 90 to 180 mcg of MK-7

Match the number on the label to the trials. The three-year bone trial used 180 mcg; the lowest dose that slowed bone loss in a head-to-head comparison was 90 mcg; 50 mcg did nothing. A product at 100 to 120 mcg is inside that window. A product at 45 mcg is below it, and a product at 500 mcg or more is paying for a dose no trial has shown to do more than 180 mcg. For the D side of a combination product, the vitamin D buying guide covers what to look for.

The form: all-trans MK-7, named as such

The label should say "menaquinone-7" or "MK-7," not just "vitamin K2." Walk past MK-4 at microgram doses: it does not reach the blood at supplement doses, and the Japanese fracture data came from 45 mg prescriptions, not 100 mcg capsules. Walk past vitamin K1 if bone or vessel proteins are the goal; it is a fine clotting vitamin and a poor osteocalcin activator by comparison. Only the all-trans geometric isomer of MK-7 is biologically active, and the share of inactive cis isomer in a finished product depends on how it was made and stored (Lal 2020, Applied Microbiology and Biotechnology). Fermented (natto-derived) and synthetic MK-7 are the same molecule when both are all-trans; a brand that states "all-trans" on the label or the certificate of analysis has answered the question, and one that does not should be asked.

Proof of testing

Vitamin K2 is dosed in micrograms, which makes it easy to underdose by accident and easy to skip testing on purpose. Ask any brand, including us, for a certificate of analysis (a lab report for a specific production lot showing identity, potency against label claim, and contaminant screens). Our products are made in our own cGMP facility in Mt. Angel, Oregon; every batch is tested in our in-house lab for identity, potency, and purity; and certificates of analysis are published by lot on the lab-testing page. How to read a supplement certificate of analysis explains which lines on that report matter. Demand the same document from any brand.

Fillers and the oil base

K2 needs fat to absorb, so the carrier matters more than it does for most vitamins. In a liquid, look for a named oil (fractionated coconut, MCT, sunflower, olive) and nothing else. In a capsule, look for a short "other ingredients" line; rice extract and cellulose are fine, and a long list of flow agents and colors signals a formula built around the capsule rather than the dose. Liquid, softgel, or capsule is mostly a preference question once the oil base is there.

Price per day for the real dose

Price the dose, not the bottle. A 120 mcg MK-7 serving in a bone-and-D stack should land between about 30 cents and a dollar a day depending on what else is in it.

Everyday Liquid Sunshine 2 carries 120 mcg of all-trans MK-7 with 5,000 IU of vitamin D3 per dropper in a coconut and sunflower oil base, about 70 servings for $25.99, or roughly $0.37 a day. It supports normal calcium handling and healthy bones as part of a daily routine.* The 5,000 IU of D is a high daily dose for someone already taking vitamin D elsewhere; read the D dosage guide before stacking.

Bones + K2 carries 120 mcg of MK-7 with 1,000 IU of D3, 500 mg of calcium and 40 mg of magnesium from Aquamin marine minerals, plus manganese and boron, in two vegetarian capsules; 30 servings for $27.99, about $0.93 a day. It is the one to pick when the goal is a single bone-support capsule with its own calcium rather than K2 alone.* Prices and serving counts checked on the live product pages 2026-10-13.

Red flags

• A "bone blend" or "K complex" that lists K2 inside a proprietary blend without a microgram amount.
• "Clinically proven" with no study named. The studies on this page are the ones a label can honestly point to.
• MK-4 at 100 to 500 mcg sold on the strength of the 45 mg trials.
• "Vitamin K2" with no MK number.
• Doses of 1,000 mcg or more with no reason given.
• A K2 product that does not warn blood-thinner users on the label.

The bottom line

If you are a postmenopausal woman building a bone routine, 90 to 180 mcg of all-trans MK-7 a day, with fat, alongside the vitamin D and calcium you already need, has moderate trial support and a clean safety record over three years. If you are buying it for your arteries, know that the trials have not yet delivered what the observational data promised. If you take warfarin or a related anticoagulant, do not buy it until your prescriber says so. If you eat natto regularly, you are already past every dose in this article.

Related reading

• Best D3 + K2 supplements compared on MK-7 dose, carrier, and price per day

From Mt. Angel Vitamins

• Bones + K2 – Comprehensive Bone Support With Aquamin®, Vitamin D3 And K2

  $27.99

  Shop Now

Frequently Asked Questions

What is the difference between vitamin K1 and K2?

K1 (phylloquinone) comes from leafy greens and goes mostly to the liver, where it keeps clotting factors active. K2 (the menaquinones, MK-4 through MK-13) comes from bacteria, fermented foods, and animal fat, and the long-chain forms stay in the blood long enough to reach bone and artery wall. In a head-to-head trial, both absorbed well, but MK-7 built to 7 to 8 times its starting level with daily use and activated more osteocalcin than K1 did (Schurgers 2007, Blood). They share the same ring and the same enzyme. They differ in the tail, and the tail decides where they go.

MK-7 vs MK-4: which lasts longer in the body and why?

MK-7, by a wide margin. In healthy women, a 420 mcg dose of MK-7 peaked at six hours and was still measurable 48 hours later, while the same dose of MK-4 never showed up in serum at any time point; a week of 60 mcg a day raised MK-7 levels and did nothing to MK-4 (Sato 2012, Nutrition Journal). The reason is the seven-unit side chain, which keeps MK-7 in low-density lipoproteins that circulate for days. MK-4's four-unit chain is cleared by the liver within hours. One MK-7 dose a day holds a steady level; MK-4 would need several.

How much K2 did the bone and arterial studies use?

The three-year Dutch trials in 244 postmenopausal women used 180 mcg of MK-7 a day for both the bone density and the arterial stiffness results (Knapen 2013, 2015). The Chinese bone trial found 90 mcg slowed femoral neck loss and 50 mcg did not (Zhang 2020). The null bone trials used 360 and 375 mcg (Emaus 2009, Rønn 2016), and the null calcification trials used 360 mcg (Zwakenberg 2019) and 720 mcg with vitamin D (Diederichsen 2022). Supplement products at 100 to 120 mcg sit at the low end of the effective range; nothing above 180 mcg has shown extra benefit.

Does K2 work without vitamin D?

On its own markers, yes. Every MK-7 trial raised osteocalcin and MGP carboxylation whether or not vitamin D was given, and the one trial that compared K2 alone with K2 plus calcium and vitamin D found no added bone density benefit from the extras (Zhang 2020). Vitamin D does a different job, raising the amount of calcium absorbed from food, and most adults need it for its own reasons. The pairing is logical, not proven to be more than the sum of its parts; [the D3 and K2 combination article](/health/vitamin-d-complete-guide/vitamin-d3-k2-combination-benefits) walks through what the combination trials did and did not show.

Can K2 interfere with warfarin or other anticoagulants?

Yes, at doses far below what any supplement carries. Warfarin, acenocoumarol, and phenprocoumon work by blocking the enzyme that recycles vitamin K; any vitamin K you take bypasses the block and the INR drops. In 18 healthy volunteers stabilized on acenocoumarol, 10 mcg of MK-7 a day lowered the INR to a clinically relevant degree in at least 40 percent of them, and 45 mcg cut the group INR by about 40 percent (Theuwissen 2013, Journal of Thrombosis and Haemostasis). Direct oral anticoagulants such as apixaban and rivaroxaban do not use the vitamin K pathway, but tell the prescriber either way.

Which foods contain meaningful K2?

Natto, and then not much. The NIH Office of Dietary Supplements lists about 850 mcg of MK-7 in a 3-ounce serving of natto (fermented soybeans), more than any supplement. After that come small amounts of MK-4: about 13 mcg in 3 ounces of rotisserie chicken breast, 4 mcg in 1.5 ounces of cheddar, 4 mcg in a hard-boiled egg, 6 mcg in 3 ounces of chicken liver. If you eat natto a few times a week you do not need a K2 supplement; if you do not, food alone will not reach the 90 to 180 mcg the trials used. [Vitamin D food sources](/health/vitamin-d-complete-guide/vitamin-d-food-sources) is a separate list with its own gaps.

Is there an upper limit for K2?

No. The Food and Nutrition Board did not set a Tolerable Upper Intake Level for vitamin K because no adverse effects from food or supplement intake have been reported in humans or animals (NIH ODS). The trial record backs that up: 720 mcg of MK-7 plus vitamin D for two years in 365 older men produced no difference in deaths or cardiovascular events versus placebo (Diederichsen 2022), and 180 mcg for three years in 244 women raised no safety flags (Knapen 2013). The one real ceiling is practical, not toxic: nothing above 180 mcg has shown more benefit, and every microgram matters if you take a vitamin K antagonist.

How long does it take for K2 to affect osteocalcin carboxylation?

Weeks, not months. In 60 adults taking 180 or 360 mcg of MK-7, the ratio of undercarboxylated to carboxylated osteocalcin had already shifted at the four-week measurement and was down 60 to 74 percent by week 12 (Dalmeijer 2012, Atherosclerosis). In 148 postmenopausal women on 375 mcg, undercarboxylated osteocalcin fell about 66 percent by month three (Rønn 2016). The marker is fast because MK-7 reaches a steady blood level within about a week of daily dosing. Bone density is slow for a different reason: bone remodels over years, and the trials that saw a difference needed one to three of them.

Is synthetic MK-7 the same as natto-derived MK-7?

Chemically, yes, with one caveat. MK-7 exists as geometric isomers, and only the all-trans form is biologically active; the cis forms are inert. Fermented MK-7 from Bacillus subtilis and synthetic MK-7 are the same molecule when both are all-trans, but the isomer mix in a finished product depends on the production method, the purification steps, and storage conditions, and it is not always on the label ([Lal 2020, Applied Microbiology and Biotechnology](https://pubmed.ncbi.nlm.nih.gov/32009201/)). The trials on this page used fermented MK-7 (MenaQ7 in the Dutch trials, natto capsules in the Norwegian one). Buy on the all-trans specification, which either source can meet, rather than on the origin story.

Who should ask a doctor before taking K2?

Anyone on warfarin or another vitamin K antagonist, before the first dose; 10 mcg a day moved the INR in volunteers. Pregnant and nursing women, since the MK-7 trials excluded them. People with fat malabsorption (celiac disease, cystic fibrosis, short bowel syndrome, bariatric surgery), who may genuinely be short on vitamin K and should be tested rather than guessed at. People with chronic kidney disease, where the stiffness trial was negative and the calcification trials are ongoing. Anyone with a clotting disorder or a history of clots. For everyone else, the three-year trial record at 180 mcg a day is clean, and the safety section above has the full interaction table.

Related Mt. Angel Vitamins Products

• Recommended for Vitamin K2 (Menaquinone)

  Everyday Liquid Sunshine 2 – D3 + K2

  $25.99

  Shop Now

• Recommended for Vitamin K2 (Menaquinone)

  Bones + K2 – Comprehensive Bone Support With Aquamin®, Vitamin D3 And K2

  $27.99

  Shop Now