Does vitamin K in food interfere with warfarin?

Yes, when the change is large and sudden. In a randomized trial of people on stable warfarin, 9 of 13 (69%, 95% CI 39 to 91%) who ate vitamin K rich vegetables for 7 days, about 1,100 micrograms a day, had clotting activity rise above the target range. Cutting vitamin K sharply works the other way. In 12 stable patients, an 80% cut raised INR from 2.6 to 3.3 within a week. Smaller day to day swings matter less than people assume, and the warfarin label asks for a steady intake of vitamin K. It does not ask anyone to give up greens.

Established. Vitamin K counteracts warfarin, and the FDA approved label says the amount in food may affect therapy. That is the drug's own prescribing information, backed by the diet trials below.

What the trials found

Big changes move INR

The two diet trials point the same way from opposite ends. Eating about 1,100 micrograms of vitamin K a day from vegetables for a week pushed clotting activity above target in 69% of a small group, and vitamin K poor vegetables changed nothing. In a randomized crossover diet study in 12 patients, a fivefold increase lowered INR from 3.1 to 2.8 within 4 days, and an 80% cut raised it from 2.6 to 3.3 by day 7. Both were short hospital or study diets, far more extreme than an ordinary week of eating.

Small changes matter less

In volunteers kept stably anticoagulated for 13 weeks, vitamin K1 lowered INR significantly only from 150 micrograms a day, and a clinically relevant change appeared in 25% of them at that dose. A supplement giving 100 micrograms a day did not interfere. The response to spinach and broccoli meals was short lived. That study was not randomized, and the abstract does not say how many volunteers took part. A systematic review of 2 diet trials and 9 observational studies came to a similar line, with an effect on clotting detectable only above about 150 micrograms a day.

Myth. People on warfarin are often told to avoid leafy greens. The review above found no support for that advice, and the label asks for consistency. A Japanese systematic review of 22 studies named natto as the only ordinary food to avoid, and judged leafy greens acceptable in limited amounts.

Too little vitamin K is also a problem

A 2016 systematic review of 15 studies with 1,838 patients found observational links between low daily vitamin K intake and unstable anticoagulation. Those are associations and cannot show cause, but they fit the diet trial in which cutting vitamin K pushed INR up.

Food as a steering tool

In a randomized trial of 132 outpatients whose INR was out of range, a clinic adjusted the weekly amount of vitamin K rich foods instead of only changing the drug dose. After 90 days, 74% were on target against 58% managed the usual way.

Vitamin K tablets to steady INR are a separate question

Some clinics have tried a small daily vitamin K supplement to make warfarin more stable. A 2014 Cochrane review found 2 randomized trials with 100 people and judged the data insufficient to show an overall benefit. In one trial of 70 people with unstable control, 150 micrograms a day for 6 months raised time in the target range by 28 points against 15 on placebo.

Two other trials found less. A Canadian multicentre trial of 150 micrograms a day found time in range of 65.1% against 66% on placebo. In the vitamin K group the share of extreme INR values fell by 4 points from before the trial to after it (95% CI 2 to 6), which is a change within one group, not a comparison with placebo. In 400 people starting treatment, 100 to 200 micrograms a day changed time in range by 0.9 to 2.7 points, which the authors judged clinically unimportant. A meta-analysis of 3 trials with 626 people put the gain at 3.5 percentage points (95% CI 1.1 to 6.0), on low quality evidence.

Unsettled. Whether a vitamin K supplement prevents bleeding or clots is unknown. The meta-analysis reported a risk ratio of 3.2 for bleeding (95% CI 0.2 to 64.2) and 2.2 for clots (95% CI 0.1 to 47.5). Intervals that wide say the trials were far too small to tell, in either direction.

Who should be careful

Not for everyone. Anyone taking warfarin, and also phenprocoumon, acenocoumarol or tioclomarol, which NIH calls commonly used in some European countries. NIH describes the interaction as serious and potentially dangerous. Sudden changes in either direction are what the trials show moving INR, so a big change in what you eat is worth telling the clinic that checks your INR.

Do not start a vitamin K supplement on your own to steady your INR. The trials gave it under supervision, and none showed fewer bleeds or clots. Natto, the Japanese fermented soybean, is the one ordinary food the Japanese review placed off limits.

What expert bodies say

The FDA approved warfarin label tells patients to eat a normal, balanced diet with a consistent amount of vitamin K. The NIH Office of Dietary Supplements calls the interaction serious and potentially dangerous, and extends it to the related coumarins used in Europe. A Japanese systematic review proposed an acceptable range of 25 to 325 micrograms a day, with about 150 as the optimum and a largest safe daily swing of 292 micrograms. That range comes from one review and is not an official limit.

How we searched

What would change this answer

• A trial large enough to count bleeds and clots against dietary vitamin K. Every trial so far measured INR or time in range, which stand in for the outcomes that matter.
• An updated synthesis. The newest reviews are from 2015 and 2016, and no active trial is registered.
• Trials outside short hospital diets. The diet studies that moved INR used extreme intakes over days, and nobody has tested the ordinary week to week drift of a normal diet in a randomized way.

Sources

1. ev-vkw-01 · Meta-analysis or systematic review · Cochrane systematic review of RCTs · n = 100
   Currently, there are insufficient data to suggest an overall benefit.
   
   Who: people starting or taking warfarin, 2 randomized placebo-controlled trials
   Effect: one trial (70 people, 150 mcg) improved control, one (30 people, 175 mcg, abstract only) did not; no pooled estimate
   Certainty: Main synthesis, but small. Axis 9: there are no large RCTs with clinical outcomes.
   Cochrane Database Syst Rev, 2014 · checked 2026-10-07 · we read the abstract
2. ev-vkw-02 · Randomized controlled trial(s) · Cochrane systematic review of RCTs · n = 70
   This study found that in the group of participants deemed to have poor INR control, the addition of 150 micrograms (mcg) oral vitamin K significantly improved anticoagulation control in those with unexplained instability of response to warfarin.
   
   Who: warfarin users with poor INR control and unexplained instability of response
   Effect: 150 mcg oral vitamin K improved anticoagulation control significantly
   Certainty: Axis 8: benefit, if any, is in people with unstable INR without other explanations.
   Cochrane Database Syst Rev, 2014 · checked 2026-10-07 · we read the abstract
3. ev-vkw-03 · Meta-analysis or systematic review · systematic review and meta-analysis of RCTs · n = 626
   We found low-quality evidence--downgraded for imprecision and risk of bias (i.e., limitation in study design and/or execution)--of no effect of vitamin K use (100 to 200 μg) on hemorrhagic events (relative risk [RR] 3.2, 95% confidence interval [CI] 0.2-64.2) and thromboembolic events (RR 2.2, 95% CI 0.1-47.5) and a significant but clinically unimportant effect on TTR with an absolute increase of 3.5% (95% CI 1.1-6.0).
   
   Who: adults on vitamin K antagonist therapy, most with satisfactory time in range at baseline, 3 RCTs
   Effect: time in therapeutic range +3.5% (95% CI 1.1 to 6.0), judged clinically unimportant; low-quality evidence (GRADE)
   Pharmacotherapy, 2013 · checked 2026-10-07 · we read the abstract
4. ev-vkw-04 · Meta-analysis or systematic review · systematic review and meta-analysis of RCTs · n = 626
   This meta-analysis, despite the few studies and overall low quality, suggests no beneficial role of low-dose (100 to 200 μg) vitamin K supplementation on the reduction of clinically relevant adverse events in patients taking VKAs, despite a small improvement of the TTR.
   
   Who: adults on vitamin K antagonist therapy, 3 RCTs
   Effect: hemorrhage RR 3.2 (0.2 to 64.2); thromboembolism RR 2.2 (0.1 to 47.5); no clinical benefit shown
   Certainty: Axis 6/9: the supplement has not been shown to reduce bleeding or thrombosis; do not take on your own.
   Pharmacotherapy, 2013 · checked 2026-10-07 · we read the abstract
5. ev-vkw-05 · Observational data · systematic review of interventional and observational studies · n = 11 studies
   Median dietary intake of vitamin K1 ranged from 76 to 217 μg/day among studies, and an effect on coagulation may be detected only for high amount of vitamin intake (>150 μg/day).
   
   Who: patients on vitamin K antagonists in 11 clinical studies; median vitamin K1 intake 76 to 217 micrograms a day
   Effect: conflicting evidence; effect on coagulation detected only for intakes above 150 micrograms a day
   Certainty: Mixed design, hence C. This is the main argument against the myth "no greens on warfarin."
   Medicine (Baltimore), 2016 · checked 2026-10-07 · we read the abstract
6. ev-vkw-06 · Observational data · systematic review of observational and interventional studies · n = 1838
   Observational studies suggest an increased risk of unstable anticoagulation control in patients with lower daily vitamin K intake.
   
   Who: patients on vitamin K antagonists in 15 studies
   Effect: lower daily vitamin K intake associated with more unstable control (observational)
   Certainty: Paradox: a small, not a large, K content in the diet is associated with instability.
   Semin Thromb Hemost, 2016 · checked 2026-10-07 · we read the abstract
7. ev-vkw-07 · Randomized controlled trial(s) · randomized controlled trial · n = 37
   Nine patients (69%; 95% CI, 39-91%) who consumed vitamin-K-rich vegetables for 7 d reached activities above the therapeutic level.
   
   Who: patients on stable warfarin treatment, randomized to 1, 2 or 7 days of vitamin K rich vegetables, vitamin K poor vegetables or phytomenadione
   Effect: 7 days of ~1100 mcg/day: 9/13 (69%, 95% CI 39 to 91%) above therapeutic activity; no change with vitamin K poor vegetables
   Certainty: "Above the therapeutic level" means clotting-factor activity above target, that is, a weakened warfarin effect. The dose of greens is large, atypical.
   J Intern Med, 1991 · checked 2026-10-07 · we read the abstract
8. ev-vkw-08 · Randomized controlled trial(s) · randomized crossover dietary trial · n = 12
   In the randomized protocol, the international normalized ratio increased from 2.6 +/- 0.5 at baseline to 3.3 +/- 0.9 at day 7 (P = 0.005) in subjects on the vitamin K-depleted diet and decreased from 3.1 +/- 0.8 at baseline to 2.8 +/- 0.6 at day 4 (P = 0.04) in those on the vitamin K-enriched diet.
   
   Who: 12 patients with stable anticoagulation, 4-day in-hospital diets 1 to 2 weeks apart
   Effect: vitamin K depleted diet INR 2.6 to 3.3 (p=0.005); enriched diet INR 3.1 to 2.8 (p=0.04)
   Certainty: Axis 6: a sharp restriction of vitamin K is as dangerous as a sharp increase (INR up, bleeding risk).
   Am J Med, 2004 · checked 2026-10-07 · we read the abstract
9. ev-vkw-09 · Observational data · non-randomized dose-response challenge study · n = —
   We conclude that short-term variability in intake of K(1) is less important to fluctuations in the international normalized ratio (INR) than has been commonly assumed and that food supplements providing 100 microg/day of vitamin K(1) do not significantly interfere with oral anticoagulant therapy.
   
   Who: healthy volunteers stably anticoagulated for 13 weeks, weekly dose increments 50 to 500 micrograms
   Effect: threshold 150 mcg/day; clinically relevant INR change in 25% of participants at that dose; 100 mcg/day supplements did not interfere
   Certainty: Nonrandomized challenge: C. Axis 2: K1 from spinach and broccoli is absorbed inefficiently.
   Blood, 2004 · checked 2026-10-07 · we read the abstract
10. ev-vkw-10 · Randomized controlled trial(s) · randomized controlled trial · n = 132
    Over time, patients allocated to the vitamin K-guided strategy reached the prespecified INR more frequently so that after 90 days of follow-up, 74% were on target compared with 58% of patients managed conventionally (P=0.04).
    
    Who: outpatients on long-term oral anticoagulation with a recent INR outside target; 58% mechanical heart valves, 35% atrial fibrillation
    Effect: on target at 90 days 74% vs 58% (P=0.04); minor bleeding or parenteral vitamin K 1.5% vs 11% (P=0.06)
    Certainty: Vitamin K food was used as a tool for managing INR under clinic supervision, not as a ban.
    Circulation, 2009 · checked 2026-10-07 · we read the abstract
11. ev-vkw-11 · Randomized controlled trial(s) · double-blind randomized placebo-controlled trial · n = 70
    Vitamin K supplementation resulted in a significantly greater decrease in standard deviation of international normalized ratio (INR) compared with placebo (-0.24 +/- 0.14 vs -0.11 +/- 0.18; P < .001) and a significantly greater increase in percentage time within target INR range (28% +/- 20% vs 15% +/- 20%; P < .01).
    
    Who: 70 warfarin-treated patients with unstable anticoagulation control, double-blind, 6 months
    Effect: time in target range +28% vs +15% (P<.01); INR SD -0.24 vs -0.11 (P<.001)
    Blood, 2007 · checked 2026-10-07 · we read the abstract
12. ev-vkw-12 · Randomized controlled trial(s) · multicentre randomized placebo-controlled trial · n = —
    We conclude that LDVK administration did not increase mean TTR, but did decrease the number of INR excursions.
    
    Who: patients on chronic warfarin therapy at four Canadian hospitals, 6 months after a 1-month run-in
    Effect: TTR 65.1% vs 66% (p=0.8); INR excursions below 1.5 or above 4.5 fell in the vitamin K group, pre minus post absolute difference 4% (95% CI 2 to 6%), not a between-group difference
    Certainty: The newest RCT (OVWAC VII, NCT00990158). Diverges from 17110451: there only unstable ones were selected.
    Thromb Haemost, 2016 · checked 2026-10-07 · we read the abstract
13. ev-vkw-13 · Randomized controlled trial(s) · randomized placebo-controlled dose-finding trial · n = 400
    Differences between doses were, however, small and the improvement is unlikely to be of clinical relevance.
    
    Who: 400 patients starting vitamin K antagonists at two Dutch anticoagulation clinics, 6 to 12 months
    Effect: TTR difference vs placebo 2.1% (100 mcg), 2.7% (150 mcg), 0.9% (200 mcg), all CIs crossing zero; no difference in complications
    Haematologica, 2011 · checked 2026-10-07 · we read the abstract
14. ev-vkw-14 · Position of an expert body · drug label (manufacturer text approved by FDA) · n = —
    The amount of vitamin K in food may affect therapy with warfarin sodium. Advise patients taking warfarin sodium to eat a normal, balanced diet maintaining a consistent amount of vitamin K.
    
    Who: patients taking warfarin sodium
    Effect: competitive antagonism by vitamin K; consistent intake advised, not avoidance
    Certainty: Regulator position: constancy, not a restriction.
    DailyMed, Warfarin Sodium label, set b8c33971-54b3-4971-bf7f-1bec66999a29, effective 2025-06-17 · checked 2026-10-07 · we read the section
15. ev-vkw-15 · Position of an expert body · fact sheet · n = —
    Vitamin K can have a serious and potentially dangerous interaction with anticoagulants such as warfarin (Coumadin) as well as phenprocoumon, acenocoumarol, and tioclomarol, which are commonly used in some European countries [7,8].
    
    Who: people taking warfarin, phenprocoumon, acenocoumarol or tioclomarol
    Effect: vitamin K antagonizes these drugs; sudden intake changes raise or lower their effect
    NIH Office of Dietary Supplements, Vitamin K Fact Sheet for Health Professionals · checked 2026-10-07 · we read the section
16. ev-vkw-16 · Observational data · systematic review of mixed study designs · n = 22 studies
    When these results were applied to usual foods, except for dietary supplements or health foods, the only prohibited foods were fermented soybean (natto) and foods containing it, while green leafy vegetables could be acceptable if their intake is limited.
    
    Who: warfarin users in 22 studies on adverse events from the warfarin and vitamin K interaction
    Effect: acceptable 25 to 325 mcg/day, maximum daily variation 292 mcg; natto prohibited; leafy greens acceptable in limited amounts
    Certainty: Japanese context (natto: up to ~1000 mcg K2 per serving). Natto is already in natto-r1-14.
    Shokuhin Eiseigaku Zasshi, 2015 · checked 2026-10-07 · we read the abstract