Vitamin K2: what it actually does, and why most people are deficient

Vitamin K2: what the evidence shows and why deficiency is common

Vitamin K2 is one of the more quietly compelling nutrients in the literature. A 2015 Rotterdam cohort study found that higher menaquinone intake was associated with a 57% lower risk of aortic calcification and a 52% lower risk of cardiovascular mortality over ten years. Yet most people in the UK get almost none of it from their diet - not because they're eating badly, but because the food sources are genuinely obscure.

What the evidence actually shows

The most cited work on K2 comes from the Rotterdam Study - a large Dutch prospective cohort tracking over 4,800 adults. Geleijnse et al. (2004) found that participants in the highest tertile of menaquinone intake had a 41% lower risk of coronary heart disease mortality compared to the lowest tertile (HR 0.59, 95% CI 0.38-0.93). Dietary phylloquinone - that's K1 - showed no such association. The distinction matters.

On bone, the picture is more nuanced. A Cochrane-adjacent systematic review by Cockayne et al. (2006) pooled 13 randomised controlled trials and found that K2 supplementation (predominantly MK-4 at pharmacological doses) significantly reduced vertebral fracture incidence (RR 0.40, 95% CI 0.25-0.65) and non-vertebral fractures (RR 0.19, 95% CI 0.11-0.35) in Japanese postmenopausal women. The effect sizes are striking. I'd be cautious about direct extrapolation to Western populations - Japanese dietary patterns and baseline K2 status differ substantially - but the mechanism is biologically coherent.

More recently, Knapen et al. (2015) published a three-year RCT in 244 healthy postmenopausal women showing that 180mcg/day of MK-7 significantly improved femoral neck bone mineral density and bone strength indices compared to placebo. This is the dose you'll see most often in European supplement research. The trial was well-controlled and the effect was modest but real.

My honest read: the cardiovascular and bone data are genuinely interesting. They're not conclusive in the way, say, the statin literature is. But K2 has a plausible mechanism, a reasonable safety profile, and a population-level deficiency problem. That combination is worth taking seriously.

The biology: what K2 is actually doing

Vitamin K is a cofactor for a class of enzymes called gamma-glutamyl carboxylases. These enzymes activate a group of proteins known as vitamin K-dependent proteins - or Gla proteins - by adding a carboxyl group to specific glutamic acid residues. Without sufficient K, these proteins remain undercarboxylated and largely non-functional.

Two of the most relevant Gla proteins for health are osteocalcin and matrix Gla protein (MGP).

Osteocalcin is produced by osteoblasts and is essential for incorporating calcium into bone matrix. Undercarboxylated osteocalcin (ucOC) can't bind hydroxyapatite properly - so even if you're taking plenty of calcium and vitamin D, poor K2 status means the signalling for bone mineralisation is impaired. This is the crux of the K2-bone story.

MGP is arguably more interesting from a cardiovascular standpoint. It's the most potent known inhibitor of vascular calcification. It works by binding calcium ions and preventing them from precipitating in arterial walls. When MGP is undercarboxylated - which happens when K2 is insufficient - this inhibition fails. Calcium ends up in the arteries rather than the bones. The Rotterdam data make more sense when you understand this mechanism.

The key distinction between K1 and K2 is tissue distribution. K1 is rapidly cleared by the liver and used primarily for clotting factor activation. K2, particularly the longer-chain menaquinones like MK-7, has a much longer half-life in circulation (roughly 72 hours versus a few hours for K1) and reaches extrahepatic tissues - including bone and the arterial wall - far more effectively. Schurgers et al. (2007) demonstrated this directly, showing MK-7 was significantly more bioavailable and had a greater effect on carboxylation of osteocalcin than MK-4 at equivalent doses.

Why most people are deficient

This is the part that genuinely surprised me when I first looked into it. K2 deficiency isn't about eating badly in the conventional sense. It's about the near-total disappearance of K2 from the modern food supply.

The richest dietary source of K2 is natto - a Japanese fermented soybean product. It contains extraordinary amounts of MK-7, upwards of 900mcg per 100g. Most people in the UK have never eaten it and probably never will. I've tried it. The texture and smell are confrontational.

Beyond natto, meaningful K2 is found in certain hard cheeses (particularly Gouda and Edam), some fermented foods, egg yolks, and organ meats. Grass-fed animal products contain more K2 than grain-fed equivalents - the conversion of K1 from grass to K2 in ruminant fat tissue is the relevant pathway. As livestock farming has shifted toward grain feeding and as organ meat consumption has declined sharply in the UK, dietary K2 has quietly dropped.

Vermeer et al. (2012) estimated that typical Western dietary menaquinone intake is in the range of 10-30mcg/day - well below the amounts used in most intervention studies showing cardiovascular and bone benefits. The Dutch population in the Rotterdam cohort, with relatively higher cheese consumption, was already eating more K2 than the average Briton.

There's also an absorption issue. K2 is fat-soluble. Low-fat diets, fat malabsorption syndromes, and long-term use of certain medications (notably broad-spectrum antibiotics, which disrupt gut bacteria that synthesise some menaquinones, and cholesterol-lowering drugs) all reduce K2 status. If you've been on a low-fat diet for years and eat little fermented food or aged cheese, your K2 intake is probably negligible.

K2 and vitamin D: the pairing that actually matters

You'll see K2 and D3 sold together constantly. Some of that is marketing. But the underlying biology is worth understanding.

Vitamin D3 upregulates the production of osteocalcin and MGP - it essentially increases the demand for K2 as a cofactor. If you're taking meaningful doses of D3 and your K2 status is poor, you may be producing more of these Gla proteins in an undercarboxylated, non-functional state. Masterjohn (2007) laid out this hypothesis clearly, and it has reasonable mechanistic support, though I should be transparent: direct RCT evidence specifically testing D3+K2 co-supplementation versus D3 alone on hard clinical outcomes is limited. The human data on this specific interaction is thin and I'd be overstating it to claim the combination is proven superior.

What I can say is that there's no evidence of harm in combining them, the mechanistic rationale is coherent, and if you're supplementing D3 at doses above 1,000 IU regularly, thinking about K2 status seems reasonable rather than paranoid.

Dosing: what the clinical evidence supports

The dose picture splits along form lines: MK-4 versus MK-7.

Most Japanese clinical trials on bone fracture reduction used MK-4 at pharmacological doses - typically 45mg/day (that's milligrams, not micrograms). This is roughly 1,000 times the amount you'd get from diet. These trials produced real results, but the doses are far above what's achievable through food or standard supplementation, and the Japanese regulatory context is different - MK-4 at 45mg is used as a pharmaceutical there.

MK-7 research has generally used lower doses: 90-360mcg/day. The Knapen et al. (2015) trial used 180mcg. A well-designed RCT by Braam et al. (2009) in 325 postmenopausal women used 45mcg MK-7 daily and found significant reductions in undercarboxylated osteocalcin (a marker of K2 sufficiency) compared to placebo. Even 45mcg moved the biomarker needle, which suggests the threshold for biological effect may be relatively modest.

For cardiovascular endpoints specifically, the Rotterdam data were observational - they didn't test a supplement dose. But the intake levels associated with benefit in that cohort were in the range of 32mcg/day of menaquinones, which is achievable through diet with regular cheese consumption or modest supplementation.

If you're looking at a supplement, 100-200mcg of MK-7 is the range most commonly used in European trials. It's worth noting that when I was formulating Kojo, K2 as MK-7 at 100mcg was the dose I landed on - enough to sit within the evidence range without overstating what's known. On the topic of formulation transparency, if you want to understand why dose disclosure matters so much, the piece I wrote on why supplement labels lie covers the mechanics of how brands obscure what's actually in the product.

Who is most at risk of K2 deficiency

Based on the dietary and absorption factors above, certain groups are more likely to have genuinely poor K2 status:

• People eating low-fat diets - fat is required for K2 absorption; if you've been avoiding dietary fat for years, K2 absorption is compromised regardless of intake.
• Postmenopausal women - declining oestrogen accelerates bone turnover, and the demand for carboxylated osteocalcin increases. This population also features most prominently in the fracture-reduction RCT literature. If you're navigating this stage and want a broader look at the evidence base for nutritional support, the piece on perimenopause supplements UK covers this territory honestly.
• People with gut absorption issues - Crohn's, coeliac disease, short bowel syndrome, or any condition affecting fat-soluble vitamin absorption.
• Those on long-term broad-spectrum antibiotics - gut bacteria contribute to menaquinone synthesis; disrupting the microbiome reduces endogenous K2 production.
• Older adults generally - dietary variety tends to narrow with age, and the foods richest in K2 (natto, organ meats, aged cheese) are not typically staples of UK diets at any age.
• People taking warfarin or other vitamin K antagonists - this is a category where supplementation requires medical supervision. K2 supplementation can interfere with anticoagulation therapy. If this applies to you, speak to your GP before doing anything.

What to look for in a supplement

A few practical points, since this is an area where the quality variation is significant.

MK-7 is generally preferable to MK-4 for supplementation purposes at nutritional doses. The longer half-life means more stable tissue saturation from a once-daily dose. Schurgers et al. (2007) demonstrated MK-7's superior bioavailability and carboxylation effect directly in a controlled comparison.

Look for all-trans MK-7 rather than cis-isomers. Some manufacturing processes produce a mixture; the all-trans form is the biologically active one. Brands using MenaQ7 or similar branded MK-7 sources are typically using all-trans material, though it's worth checking.

Dose transparency is non-negotiable for me. If a product lists "vitamin K2" without specifying the form and dose, that's a problem. The all-in-one supplements UK guide I wrote covers how to evaluate combined formulas and what to look for on labels - worth reading if you're trying to make sense of the market.

Finally: take K2 with a meal containing fat. It's fat-soluble. Taking it on an empty stomach meaningfully reduces absorption. This applies to D3 as well.

Frequently asked questions

My honest take

I came to K2 sceptically. It has the profile of a nutrient that gets overclaimed - a neat mechanism, some impressive-sounding observational data, and a convenient deficiency narrative. I've seen enough of those to be wary.

But the more I read the primary literature, the more I think the scepticism is misplaced here. The Rotterdam cardiovascular data is from a large, well-run cohort. The Cockayne bone fracture meta-analysis covers 13 RCTs. The Knapen MK-7 trial is three years long and properly controlled. This isn't a single small study with a p-value of 0.049. There's a body of work, with a coherent mechanism running through it.

The honest caveats: most of the bone fracture RCT data is in Japanese postmenopausal women, which limits direct extrapolation. The cardiovascular data is largely observational. The D3 interaction, while mechanistically sensible, lacks the hard RCT evidence I'd want before making strong claims. And the optimal dose for Western populations hasn't been established with the precision I'd like.

None of that makes K2 uninteresting. It makes it a nutrient that deserves careful attention rather than either dismissal or hype. Given the near-zero dietary intake most people in the UK are getting, the safety profile, and the quality of the available evidence, I think it belongs in a serious conversation about nutritional gaps - alongside the more familiar ones like D3, magnesium, and omega-3s.

What I'd push back on is anyone selling K2 as a solution to anything specific. It's not. It's a cofactor for proteins that matter for long-term cardiovascular and skeletal health. That's worth something. It's just not a dramatic claim, and I'd rather say that clearly than dress it up.