Which foods are highest in vitamin K?

1 cup, chopped of Collards, frozen, chopped, cooked gives 1059.4 µg of vitamin K. That is 883% of the Daily Value, and it is the highest of the 1000 foods on this site that has not been fortified.

This ranking is per typical portion rather than per 100 g. Lists built per 100 g end up full of dried herbs and fortified bars, which is technically correct and practically useless, because nobody sits down to 100 g of thyme. Items marked fortified rank where they do because nutrients were added during manufacture, not because the food itself is rich in them.

What this means for you

Vitamin K is the easiest requirement on this site to meet by accident. Men need 120 micrograms a day and women 90. One cup of cooked collards carries 1,059 micrograms, roughly nine days' worth. Kale, chard, spinach, beet greens and broccoli are all in the same league.

Because the numbers are so far above the requirement, the percentages in this table are close to meaningless for a healthy person. Eat green leaves sometimes and vitamin K is handled for life.

The exception that makes this page worth reading

If you take warfarin or a similar anticoagulant, vitamin K stops being a nutrient and becomes part of your dose.

Not for everyone. The rule is consistency rather than avoidance. Sudden changes in vitamin K intake, in either direction, make warfarin work more or less than intended. A big salad after weeks without one is exactly the kind of change that shifts an INR, and quietly cutting greens out is just as disruptive as adding them. Tell the clinic when your eating changes, rather than editing your diet to match the drug.

Unsettled. Vitamin K2 is sold hard for bones and arteries, and the bone half of that has now been tested enough to answer. Pooling six randomised trials in postmenopausal women, fractures came out the same on K2 as on placebo, and the result only turned positive after one study was removed from the analysis. In 123 dialysis patients given 360 micrograms a day for two years, bone density fell faster at the wrist and slower at the spine, and the authors concluded it does not preserve bone.

The artery half is messier and more interesting. The longest trial is the newest one. In people with symptomatic coronary disease and a calcium score between 50 and 400, two years of 360 micrograms of MK-7 a day left the score at 184 Agatston units against 214 on placebo. Pooling 14 earlier trials pointed the same way.

Then the rest of the evidence arrives. Six months of the same dose in people with type 2 diabetes and existing heart disease changed arterial calcification by nothing and bone density by nothing. A year of high-dose K1 in dialysis patients cut the vitamin K blood marker by 86 per cent and left their coronary calcification exactly where it was. So the marker moves every time, the scans move sometimes, and every one of those trials measured a picture rather than a heart attack. None of it is a reason for a healthy person to buy K2.

Deficiency is rare and mostly belongs to people who cannot absorb fat properly, such as those with cystic fibrosis, coeliac disease or short bowel syndrome. Newborns are the other exception, which is why the vitamin K injection at birth exists.

How little of it you actually absorb

The figures in this table are how much vitamin K is in the food. How much reaches you is a different and much smaller number.

Unsettled. People were fed a large serving of kale with oil, which is the most favourable arrangement anyone has tested, and absorbed 4.7 per cent of its phylloquinone. Under one twentieth. Blood levels peaked between 6 and 10 hours at about 2.1 nmol/L and fell quickly afterwards, with a half-life of 8.8 hours in plasma against 215 hours in tissue. So a cup of raw spinach really does carry 120.7 micrograms, the entire daily reference intake, and what crosses into you is a small fraction of that. Since nobody has established what the body actually needs, this is not a reason to eat more greens or to buy anything. It is a reason to stop treating the table as a delivery note.

Unsettled. Adding oil is the usual advice and the evidence for it is split inside a single experiment. Comparing a fortified oil against broccoli, the 24-hour area under the curve was clearly greater on the oil. The 24-hour fasting plasma phylloquinone showed no difference between them at all. One study, two defensible readings. Eating greens with some fat is sensible and it is not a settled fact.

Two things our own data cannot tell you

The tables disagree with each other more than you would expect for a leafy green. The British table puts raw curly kale at 623 micrograms per 100 g and the American one puts raw kale at 390, a difference of about 60 per cent for the same vegetable. On spinach they swap places, 394 against 482.9. The British table has no vitamin K figure at all for baby spinach or for boiled kale.

And what cooking does to vitamin K cannot be answered from our data. The USDA retention table covers sixteen nutrients, and vitamin K is not among them. It appears neither in the list of nutrients mapped nor in the list of nutrients skipped. If you read somewhere that boiling destroys the vitamin K in kale, that figure did not come from the retention table, because the row does not exist.

Vitamin K2 is on weaker ground again. It reaches the diet as at least ten different menaquinones of different chain lengths, made by bacteria except menaquinone-4, which the body makes itself from phylloquinone. How much of any of them a food contains is much less well established than the K1 figures in this table, so treat a K2 number with more doubt than a K1 number.

Written from the evidence cards and the data on this page, then checked back against the sources. Bioma Learn has no human medical reviewer at this time and we say so rather than invent one. Information, not medical advice. How we verify.

Full ranking →

Full guide: Vitamin K: the nutrient whose most useful fact is about a medicine

Data: USDA FoodData Central. % Daily Value against the FDA DV of 120 µg. Checked 2026-09-15.