Vitamin K2: Part Two – Artery Calcification

In Vitamin K2: Part One, I reported on a study from EPIC that followed 16,000 women for 8 years and found an arguably borderline statistically significant, beneficial association for vitamin K2 and the prevention of heart disease.

The same group of researchers, from The Netherlands, published another study of 564 post-menopausal women, comparing their vitamin K intakes to coronary artery calcification (1).

The theory is that vitamin K activates proteins that sequester calcium. Vitamin K1 is cleared from the bloodstream for use in the liver while vitamin K2 is much more likely to be found in the vessel walls where it can prevent calcium from being deposited in the vessel walls.

As predicted, the results showed that vitamin K1 was not associated with a reduced amount of coronary artery calcification. Vitamin K2, however, was associated with reduced risk in a borderline statistically significant manner.

In comparing the highest one-fourth of intake to the lowest, in the model adjusted only for age, the risk was barely statistically significant at 0.82 (0.68–0.99).

In the model adjusted further for smoking, diabetes, BMI, hypertension, educational attainment, HDL and LDL cholesterol, the risk was 0.85 (0.72–1.02); not statistically significant.

And in the model adjusted even further for alcohol consumption and energy-adjusted intake of protein, calcium and fiber the risk was back to statistically significant at 0.80 (0.65–0.98).

Interestingly, the women in the highest intake of vitamin K2 also had the highest intake of calcium at 1,317 mg per day, and calcium and vitamin K2 intakes were significantly associated with each other. Ditto for protein (but not fiber).

Why would women with the highest calcium intake have the lowest amount of artery calcification? Well, as I have written about before, dietary calcium does not appear to cause calcification of the arteries until calcium intakes reach at least 1,400 mg per day, if at all (see Calcium Supplements – The Final Word?).

MK4 was the only vitamin K2 subtype that showed an individual trend towards less artery calcification.

People should keep in mind that this study was cross-sectional, making it a less reliable form of evidence than a prospective study (other things being equal).

In summary, vitamin K2 intake was weakly associated with a lower risk for artery calcification.


If you like my posts, please like my posts! Or share them. Thanks!

I greatly appreciate donations of any amount and it allows me to spend more time on nutrition (click here).

Purchase anything through these links and JackNorrisRD gets a percentage: Gift Cards – E-mail Delivery

Vegan for Life: Everything You Need to Know to Be Healthy and Fit on a Plant-Based Diet from


1. Beulens JW, Bots ML, Atsma F, Bartelink ML, Prokop M, Geleijnse JM, Witteman JC, Grobbee DE, van der Schouw YT. High dietary menaquinone intake is associated with reduced coronary calcification. Atherosclerosis. 2009 Apr;203(2):489-93. | link

5 Responses to “Vitamin K2: Part Two – Artery Calcification”

  1. Ayla Says:

    Thanks Jack. I was wondering if these new findings change any of the recommendations on the site. From your page on vitamin K, I got the idea that humans absorb K2 from bacteria in their own intestines. How does the amount that we absorb naturally relate to the amounts that provide borderline significant benefits in these studies?

    Also, did your quest result in any (new) research on vitamin K2 and bone/teeth health? Especially, do we know whether vegans produce enough vitamin K2 to maintain healthy teeth? Although I hate to admit it, WAPF made me somewhat worried on that topic.

  2. Dan Says:

    Hi Jack,

    I don’t know but I am not seeing enough here – in this research article and the previous – to start taking vitamin K2 for health promotion/disease prevention. Cross-sectional data relating one-time dietary intake to a vascular marker? Some cardiologists have made the point that this vascular marker — arterial calcification — only reflects late established atherosclerosis, and not the earlier, non-calcified lesions that can erode or disrupt and lead to acute thrombosis (MI, atherothrombotic stroke, etc). And of course, K2 intake could be a marker of something else entirely – since it is meat-derived (except for Natto), it could reflect something else in meat such as B12, creatine, carnitine, taurine, or it could reflect the lifestyles of the people who eat a lot of K2.

    Even if there was a trial showing a reduction in arterial calcification with K2, I would have caveats. For example, strong evidence from randomized trials suggested that estrogen improved endothelial dysfunction (which was taken at the time as very predictive of CV events), but when definitively tested in large subsequent hard outcomes based studies, that same molecule, premarin (horse estrogen) actually increased the risk of cardiovascular events – in multiple trials. One problem is that there are so many unknown mechanistic markers and pathways; hence, we can rarely predict what is going to happen on the basis of a single marker/single pathway.

    K2 appears to be the latest poster child for health promotion, perhaps as we see many failed recent trials of omega-3 supplements in omnivores. I wish people would make the real lifestyle changes necessary to improve their long-term health, and not simply take false security and comfort from taking the latest promoted panacea. In fact, I believe many of these supplements (not B12, mind you) are an impediment to achieving real lifestyle change, since they make people feel better about themselves without actually doing the requisite work. How sanctimonious I am!!

  3. Jack Norris RD Says:


    > I was wondering if these new findings change any of the recommendations on the site.

    I’m not done with my examination of vitamin K2, I still have to read the 2004 Rotterdam Study and I’m not sure if there is other, more recent research out there yet, I’m going to look for it soon. I haven’t even started on bone health, yet.

    So far, I agree with what Dan says.

  4. Dan Says:

    Another possibility for accounting for the relationship in the study by Beulens et al in Atherosclerosis is reverse causality.

    I am not sure which exclusions the authors made at the beginning of the study to the patient sample, but it is possible that patients who had heavier vascular risk factor burdens, or known overt vascular disease, modified their diet to reduce their meat intake (in fact, this is often what such persons do).

    Even if the investigators went to great lengths to measure things like hypertension and family history of cardiovascular disease, such measurements are often imperfect. For example, they may have used the standard questionnaire item relating to baseline diabetes status which runs: “Have you ever been told you have diabetes by a physician or other health care provider and/or do you currently use medications or diet to treat high blood sugar?” If a person was told they were pre-diabetic or borderline diabetic, but not yet overtly diabetic, this item could well be rated as a “no” by the interviewer. That patient would then be misclassified, yet may still reduce their saturated fat intake (and inadvertently their K2 intake).

    Measurement of comorbidities, risk factors and confounders is fraught with error when these items are dichotomized (e.g. diabetes, yes/no; e.g. hypertension, yes/no). People with even a weak family history of cardiovascular disease which does not conform to the standards of “premature family history in first degree relative – males less than 55, females less than 65” may still modify their diet to reduce meat intake. So even adjustment for measured risk factors and exclusion of patients with known disease does not necessarily exclude robust reverse causality. Having said that, I’d have to read the paper in detail to see how careful they were about these things, although authors rarely include their CRFs (case report forms — i.e. baseline questionnaires and data entry) with their manuscripts. The best thing is if they actually adjusted for physiological data variables like 24 hour blood pressure values, insulin levels, hemoglobin A1c, visceral adiposity, etc – but studies rarely go to these lengths when they do their standard risk factor adjustment. It’s too expensive, timely and resource-intensive. Hence measured confounding adjustment is rarely sufficient to exclude true confounding!

  5. Dan Says:

    I should have read the paper before I opened my mouth. But I think I may be on to something. Patients in the lowest intake of vitamin K1 were more likely to have high systolic blood pressure and a history of hypertension, had slightly lower HDL cholesterol and were substantially less well-educated. Given all these differences (and some of them conflict — e.g. diabetes, which was significantly higher in Q4 than Q1), there is very likely unmeasured confounding, or even imperfectly measured confounding, that is biasing this association. Which is why randomized trials are so critical, as people clearly do not select their dietary habits completely at random (as this table from the paper shows). if a large randomized trial showed that menaquinone intake reduced the risk of cardiovascular events, then as a vegan, I would feel very deprived unless I decided to take menaquinone supplements. However, as far as I know, such a trial does not exist, though I am eagerly looking forward to Jack’s review on bone health and K2.

Leave a Reply