Vitamin K2: Part Three – Rotterdam
In Vitamin K2: Part One and Vitamin K2: Part Two, I reported some weak associations suggesting that vitamin K2 (found in animals foods), but not vitamin K1 (found in plant foods), might play a role in protecting against heart disease.
Part Three is on yet another study from The Netherlands, this time a component of the The Rotterdam Study published in 2004 (1). Unlike the other other two studies, this one had much stronger results.
The study had a prospective component in which 4,807 men and women aged 55 years and older were followed for an average of 7.2 years. At baseline, all participants were given an ECG to determine if they had heart disease and were excluded if they did.
Vitamin K2 intake was positively associated with the intake of total fat, saturated fat, and calcium, as well as body mass index, and diabetes; it was inversely associated with intake of polyunsaturated fatty acids.
For vitamin K1, the results once again showed it not to be associated with a reduced risk of heart disease (nor mortality).
In model 1, that adjusted for age, gender, and total energy intake, when comparing the group with the highest daily intake (> 33 µg) of vitamin K2 to the lowest (< 22 µg), K2 was associated with a reduced risk of heart disease (.71, .51-1.00), death from heart disease (.59, .35-.99), and overall mortality (.81, .67-.98). These findings are borderline statistically significant.
However, in model 2 that adjusted for factors in model 1 and also body mass index, smoking status and history, diabetes, education, and intake of alcohol, saturated fat, polyunsaturated fat, flavonols (a group of antioxidants), and calcium, the associations became much stronger for heart disease (.59, .40-.86), death from heart disease (.43, .24-.77), and overall mortality (.74, .59-.92).
There was also a cross-sectional component of the study in which 4,473 people were given x-rays at baseline to determine if they had aortic artery calcification. Vitamin K2 intake was inversely associated with severe calcification in model 1 (.56, .39-.80) and model 2 (.48, .32-.71).
As for the possibility of reverse causation (in which people with poor health decided to eat fewer foods high in K2), the authors said, “In contrast to phylloquinone [K1], intake of menaquinone [K2] (mainly MK-4 from eggs and meat, and MK-8 and MK-9 from cheese), is not related to a healthy lifestyle or diet, which makes it unlikely that the observed reduction in coronary risk is due to confounding. Subjects with a history of MI were excluded from the analysis to avoid bias that may arise from intentional changes in diet.”
They also said, “We hypothesize that menaquinones in cheese (MK-8 and MK-9) could exert a beneficial effect in the cardiovascular system and that the high cheese consumption in France and the Mediterranean countries may possibly account for lower prevalences of [heart disease].”
Although the results from this one study are fairly strong, it takes a lot more than one cohort study to justify recommendations for preventing chronic disease. You could combine these results with those for the studies reviewed in Vitamin K2: Part One and Part Two, but those studies are not nearly as strong. Finally, all of these studies included only older people from The Netherlands; we need data from other regions.
In a relatively quick search of PubMed, I could not find any other studies looking at the association between vitamin K2 and heart disease in humans. In fairness to me, the search was quick because so few results were returned. But I will be looking around some more and also reviewing more studies on vitamin K2 and other diseases in the upcoming days.
At this point in time, I am not going out to get vitamin K2 supplements. However, I’m also not dismissing the idea that in a few weeks from now, I might be.
Stay tuned.
References
1. Geleijnse JM, Vermeer C, Grobbee DE, Schurgers LJ, Knapen MH, van der Meer IM, Hofman A, Witteman JC. Dietary intake of menaquinone is associated with a reduced risk of coronary heart disease: the Rotterdam Study. J Nutr. 2004 Nov;134(11):3100-5. | link
January 27th, 2014 at 1:35 am
You have missed at least two prospective cohort studies examining the association between vitamin K2 intake and coronary heart disease, the Prospect-EPIC and the EPIC-Heidelberg and cohorts.
https://www.thieme-connect.com/ejournals/abstract/10.1055/s-0030-1266323
http://www.sciencedirect.com/science/article/pii/S0939475308002093
For vitamin K2, the Prospect-EPIC cohort with 16,057 women found a borderline significant reduced risk of CHD (HR 0.91 [95% CI 0.85–1.00]), while the larger EPIC-Heidelberg cohort with 23,464 men and women found a non-significant increased risk of CHD incidence (HR 1.21 [95% CI 0.81–1.80]) and fatal CHD (HR 1.09 [95% CI 0.46–2.62]). It was however found in the EPIC-Heidelberg that vitamin K1 was associated with a significantly lower risk of CHD incidence (HR 0.49 [95% CI 0.25–0.94]).
Regarding the Rotterdam study, that fact that the foods rich in vitamin K2 are typically associated with unhealthy lifestyles or diets does not necessary negate the possibility that the association can be explained by reverse causation. The participants with unfavorable cardiovascular risk factors (markers of undiagnosed CHD) may have reduced foods rich in vitamin K2, such as meat and cheese in order to improve their risk profile. This is similar to the explanation as to why some observational studies have found that saturated fat and dietary cholesterol are associated with reduced levels of serum cholesterol.
http://aje.oxfordjournals.org/content/115/4/506.short
As to an explanation for a lack of published studies examining the association between vitamin K2 and CHD and other diseases, it is possible that there may be more studies that were not published due to publication bias, ie. studies that fail to find significant associations are less likely to be published than those that do. Perhaps the reason why the EPIC-Heidelberg study managed to be published was because at least a statistically significant association was found for vitamin K1 and CHD incidence. The lack of published studies either suggests that there is a lack of interest in vitamin K2 or that few studies have produced statistically significant findings, perhaps because there may not be a causal relationship.
January 27th, 2014 at 9:14 am
Health,
I covered the second study you list, Gast et a. (2009) in Vitamin K2: Part One. I had not seen the study by Nimptsch et al, it appears not to be in PubMed – thanks for pointing it out.
> This is similar to the explanation as to why some observational studies have found that saturated fat and dietary cholesterol are associated with reduced levels of serum cholesterol.
Cross-sectionally, that makes complete sense. I think it’s a lot less likely to be the case in Rotterdam. Also, the Rotterdam study didn’t show that animal products reduce heart disease, and it wasn’t until they adjusted for saturated fat that the K2 finding became more than barely statistically significant.
> it is possible that there may be more studies that were not published due to publication bias,
Agreed.
January 27th, 2014 at 4:03 am
I would also like to add that there much data that casts doubt on the suggestion that France truly has had a low prevalence of coronary heart disease. Data from the World Health Organization MONICA Project suggests that the official mortality statistics for France significantly underreport deaths from cardiovascular disease compared to other countries, with deaths from coronary heart disease being underestimated by 75%. Other reports suggest that this is likely explained by a much higher rate of French doctors classifying deaths as due to ‘other causes’ than in other countries.
http://eurheartj.oxfordjournals.org/content/27/13/1610.full
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1115846/
January 27th, 2014 at 9:22 am
Health,
Very interesting, re: France.
January 27th, 2014 at 7:12 am
An ECG is a very poor way to exclude patients with heart disease; both a resting ECG and a standard treadmill ECG. Even if they focused on relatively healthy people, it appears that patients with risk factors segregated according to the quartiles of K2 intake (BMI, diabetes), which suggests major confounding influences.
Without a large randomized trial powered to detect patient-relevant endpoints, I would be reluctant to recommend K2 intake for cardiovascular prevention — what if it ends up having some untoward effect and no benefit when given as an isolated high-dose supplement, as we have seen for magnesium, vitamin C, vitamin E, folic acid (non-periconceptually), hormone replacement therapy (both testosterone and estrogen), etc? A balanced diet buffers both the benefits and the harms of single isolated high dose supplements.
January 27th, 2014 at 9:28 am
Dan,
In fairness to nutrition supplements, testosterone and estrogen are not nutrients. And folic acid has only been implicated in (some, though not all) cohort studies with which you are regularly pointing out the flaws; a Cochrane analysis showed no evidence of folic acid causing cancer in clinical trials: http://www.ncbi.nlm.nih.gov/pubmed/23440809
January 27th, 2014 at 10:30 am
Jack,
It is more precise to say that folic acid has been implicated in some, though not all randomized trials as increasing the risk of cancer. I went through these a while back, and they included both prostate cancer and colorectal cancer, as well as lung cancer in others. The problem with large meta-analyses is that they can often mask effects seen in specific trials and patient populations by diluting them together with other trials and other patient populations where no risk would be seen or expected.
The point I was making about HRT and testosterone is that cohort studies suggested very dramatic benefit (e.g. the Nurses Health Study for estrogen) but when trials were done, there was actual harm for the same endpoints (e.g. MI). It’s a general pattern that some nutritional supplements (micronutrients) have also fallen into – e.g. beta carotene (positive cohort studies, with RCTs revealed significant harms). You are right and I did not mean to conflate estrogen/testosterone with supplements, but the general principle holds, and this is why large, precise RCTs are considered to be at the pinnacle of the evidence-based medicine pyramid.
January 27th, 2014 at 10:44 am
Dan,
> It is more precise to say that folic acid has been implicated in some, though not all randomized trials as increasing the risk of cancer.
They were RCTs? The 2013 meta-analysis published in Lancet (2013; 381: 1029–36), didn’t show statistical significance for any of the RCTs they included.
January 27th, 2014 at 11:04 am
The trials in the Lancet meta-analysis are only those that participated in the consortium and provided their individual patient data. There were a number of other trials such as SELECT, some of the colorectal adenoma trials, and a couple of the Norwegian trials, which found and separately reported harms (prostate cancer, colorectal cancer, lung cancer). I’d have to go back through these carefully to locate and ascertain where this was reported. The editorialists accompanying the Lancet meta-analysis also noted that by lumping together early effects of folic acid with later effects, they were diluting any signal. For example, folic acid appears to act as a tumor promoter rather than tumor initiator, and therefore in people with hidden polyps, may accelerate the onset of cancer. This would tend to be seen very early in follow-up, whereas tumor promoter effects usually have a pronounced lag phase.
January 28th, 2014 at 8:00 am
Please read this article. You will find studies (in vivo) showing thatt suppliments of vitamin K2, MK-7 increases levels in serium, it actiavates both MGP and Osteocalcin. And there is studies to show this, one very interesting that is still pending. You will find the name of the studies in the below document. It would be interesting to see how you valuate these.
http://menaq7.com/uploads/Nutracos_Maresz.pdf
January 28th, 2014 at 8:25 am
And here is some information about the data in the pending article regarding reduction of arterial stiffness when takin MK-7 supplement.
http://www.nattopharma.com/groundbreaking-data-on-menaq7-presented-today-at-vitafoods-in-geneva.html.
The article about bone health (same study) was published in Osteoperosis International in March 2013.
January 28th, 2014 at 5:57 pm
menne,
Interesting. I’ll take a closer look at this stuff as soon as I am able.