The role of dietary vitamin K in the management of oral vitamin K antagonists

Article Outline

• Abstract
• 1. Introduction
• 2. Chemistry and nomenclature of K vitamins
• 3. Biochemical function of vitamin K as a cofactor in Gla-protein synthesis
• 4. Mode of action of oral vitamin K antagonists
• 5. Vitamin K requirements for stable oral anticoagulation
• 6. Nutritional aspects of vitamin K: sources, and assessment of dietary intakes
  • 6.1. Nutritional sources of vitamin K (dietary and non-dietary)
  • 6.2. Measurement of vitamin K in foods and establishment of food composition databases for vitamin K
  • 6.3. Intakes of vitamin K in healthy populations
  • 6.4. Intakes of phylloquinone in patients taking oral vitamin K antagonists
  • 6.5. Dietary requirements and recommendations for vitamin K
    • 6.5.1. Healthy populations
    • 6.5.2. Patients taking oral vitamin K antagonists
• 7. Experimental evidence for the influence of vitamin K intakes on the sensitivity and stability to VKA
  • 7.1. Case reports
  • 7.2. Observational studies
    • 7.2.1. Influence of vitamin K intakes on sensitivity of VKA therapy during initiation phase
    • 7.2.2. Influence of vitamin K intakes on sensitivity of VKA therapy after stabilisation
    • 7.2.3. Influence of vitamin K intakes on stability of VKA therapy
• 8. Relationships of biochemical measures of vitamin K status to response to VKA
• 9. Intervention studies: effect of vitamin K supplementation or dietary management on response to VKA
  • 9.1. Short-term dose–response studies in patients
  • 9.2. Short-term dose–response studies in healthy volunteers
  • 9.3. Long-term vitamin K supplementation in patients
  • 9.4. Vitamin K supplementation: relevance of different molecular forms of vitamin K
  • 9.5. Intervention through a dietary vitamin K-guided strategy
• 10. Conclusions
• 11. Practise points
• 12. Research agenda
• Conflict of interest statement
• Acknowledgements
• References
• Copyright

Abstract 

Vitamin K antagonists (VKA) have been the mainstay of oral anticoagulant therapy for over 60years. In this review we critically assess the evidence for the importance of vitamin K nutrition during VKA therapy; the methodologies for measuring dietary intakes; the vitamin K intake data in patients on VKA and healthy people; and the experimental evidence for the influence of vitamin K intakes and biochemical measures of vitamin K status on VKA response. Several studies show that dietary intakes of phylloquinone (vitamin K1) are associated to the sensitivity and stability of anticoagulation during initiation and maintenance dosing with low habitual intakes associated with greater instability of the INR and risk of sub-therapeutic anticoagulation. Preliminary evidence suggests that the stability of anticoagulation therapy may be improved by daily vitamin K supplementation, but further studies are needed to find out whether this, or other dietary interventions, can improve anticoagulant control in routine clinical practice.

Keywords: Dietary vitamin K intakes, Vitamin K status, Phylloquinone (vitamin K₁), Menaquinones (vitamin K₂), Oral vitamin K antagonists, Coumarin anticoagulant drugs, Vitamin K epoxide reductase, International Normalised Ratio

Back to Article Outline

1. Introduction 

In his seminal review in 1959 entitled “The discovery of dicoumarol and its sequels” Karl Link recounted the work of his laboratory that led to the first successful introduction of 4-hydroxycoumarin drugs into clinical practise.¹ Link visualised successful therapy with vitamin K antagonists (VKA) “as being shaped like a triangle with ‘reliable coagulation assays’ at one corner, ‘vitamin K’ at another and ‘sound clinical judgement’ at the third”. He further stated “each corner is vitally dependent on the other two” and that “though the clinical judgement be good and the prothrombin time accurate, the vitamin K corner might still have to be evoked, since each individual patient is essentially an unstandardised biologic entity”. It needs to be remembered that Link wrote his review nearly two decades after the clinical introduction of dicoumarol when virtually nothing was known of the food sources, absorption, and metabolism of vitamin K and there was no molecular explanation for the antidotal effect of vitamin K. Since the 1950s we have seen huge advances in our knowledge of the metabolic, environmental, and genetic factors that influence the dosage or stability of oral anticoagulation with vitamin K antagonists. For example, knowledge of drug interactions with VKA gained from the 1960s onwards changed Link's triangle to a square and from the 1990s to the present day the concept of pharmacogenetics, the heritable variation in response to oral anticoagulants, has gained increasing prominence so that today Link's triangle more resembles a pentagon than a triangle.

It was not until 1974 that the biochemical function of vitamin K was revealed as a cofactor for a unique γ-carboxylation reaction and in the same year the target enzyme for oral anticoagulant drugs was identified as vitamin K epoxide reductase (VKOR).² The identification of the gene encoding for VKOR in 2004 has led to a burst of studies to evaluate the pharmacogenetic influence of this gene. A recent large prospective study that investigated the influence of 29 candidate genes reported that one single nucleotide polymorphism (SNP) in the VKOR gene explained some 30% of the variance of the warfarin maintenance dose.³ In the same study, the only other gene of importance was CYP2C9 which encodes an isoform of cytochrome P450 (CYP) which catabolises the S-enantiomer of warfarin⁴ and explained 12% of the variation in warfarin dose.³ In reviewing the pharmacogenetics of warfarin Kamali and Wynne⁵ estimated that polymorphisms in the VKOR and CYP2C9 genes together account for about 30–60% of the variance in the stabilised warfarin dose distribution. Apart from VKOR, the only other putative vitamin K-related gene shown to affect warfarin requirements is one that encodes for another CYP isoform named CYP4F2. Caldwell et al.⁶ identified a polymorphism of CYP4F2 for which patients homozygous for the variant allele (T) required approximately a 1mg per day higher warfarin dose than patients with wild-type alleles and postulated that this gene takes part in the catabolism of vitamin K. This hypothesis is supported by the finding⁷ that recombinant CYP4F2 hydroxylates the side chain of K vitamins as the first step in the known catabolic pathway that leads to their urinary excretion.^[8], ^[9]

The role of dietary intakes of vitamin K in explaining individual warfarin dose and stability has been more difficult to evaluate so that Link's phrase that each patient represents an ‘unstandardised biologic entity’ still resonates. Some major problems may be listed: (1) The nutritional instruments for measuring nutrient intakes are relatively imprecise and the most accurate methods are cumbersome and expensive to carry out. (2) Even when measured accurately, dietary intakes alone give no information of that fraction of vitamin K that eventually reaches the hepatic microsomes through the sum processes of metabolism that include intestinal digestion and absorption, extracellular and intracellular transport, and catabolism. (3) The bioavailability of vitamin K from meals varies not only with the individual food item (e.g. lower from green leafy vegetables than oil and fat-containing foods) but also with the type of meal consumed and interactions between food components. (4) Although phylloquinone from plant sources is the major dietary form, lower intakes of some bacterial menaquinones with slower turnover rates may also impact on oral anticoagulation.

The aim of this review is to assess our present knowledge of the impact of dietary intakes of K vitamins on VKA dose requirements and stability. First, we briefly outline the biochemical basis of vitamin K function and how VKA interfere with the recycling of vitamin K. Before evaluating the evidence of how dietary vitamin K influences management of anticoagulation therapy with VKA we felt it was important to review the strengths and weaknesses of the various dietary instruments that have been used in human studies to gather vitamin K intake data. This includes an outline of the nutritional sources of vitamin K that have enabled the compilation in several countries of food composition databases for vitamin K. In the final section we critically review the available literature of dose–response studies in healthy subjects and patients that include both observational and interventional studies.

Back to Article Outline

2. Chemistry and nomenclature of K vitamins 

Vitamin K is the family name for a series of fat-soluble compounds which have a common 2-methyl-1, 4-naphthoquinone nucleus but differ in the structures of a side chain at the 3-position. They are synthesised by plants and bacteria. In plants the only important molecular form is phylloquinone (vitamin K₁), which has a phytyl side chain. Bacteria synthesise a family of compounds called menaquinones (vitamin K₂), which have side chains based on repeating unsaturated 5-carbon (prenyl) units. These are designated menaquinone-n (MK-n) according to the number (n) of prenyl units. Some bacteria also synthesise menaquinones in which one or more of the double bonds is saturated. The compound 2-methyl-1,4-naphthoquinone (common name menadione) is widely used in animal husbandry as a cheap provitamin source of dietary vitamin K. Menadione has no cofactor activity per se but animals possess an enzyme that can convert menadione to the biologically active menaquinone-4 (MK-4) by addition of a geranylgeranyl side chain at the 3-position. A novel human enzyme responsible for the biosynthesis of MK-4 has recently been identified as UBIAD1.¹⁰ Menadione is not used as a food supplement in humans because of known cytotoxic effects and possible mutagenic effects on cultured cells.¹¹ Although menadione is licenced as a medicine for the prevention of nutritional deficiency in fat malabsorption syndromes, it has minimal antidotal effect to vitamin K antagonists¹² presumably because of the rate limiting in vivo conversion of this provitamin to active MK-4.

Amongst the naturally occurring K-vitamins, substantial differences may be expected, and have been reported, with respect to physiological processes such as their intestinal absorption, cellular uptake, tissue distribution, and turnover.¹³ These differences are largely the result of the different lipophilicity of K vitamins that is determined by the length and degree of saturation of their side chains. One notable exception to the above rules (especially with respect to tissue distribution) is MK-4 for which there is a body of evidence relating to its unique synthesis from phylloquinone in certain tissues and to biological actions not shared by other K vitamins that seem to be determined by the geranylgeranyl side chain.¹³

Back to Article Outline

3. Biochemical function of vitamin K as a cofactor in Gla-protein synthesis 

Vitamin K is an essential cofactor required for a post-translational carboxylation reaction in which selective glutamic acid (Glu) residues contained in a variety of vitamin K-dependent proteins are transformed into γ-carboxyglutamic acid (abbreviated Gla). The reaction is catalysed by the microsomal enzyme vitamin K-dependent or γ-glutamyl carboxylase (GGCX).¹⁴

An unusual feature of γ-glutamyl carboxylation is that the reaction is intimately linked to a metabolic sequence known as the vitamin K-epoxide cycle. This cycle and the associated enzyme activities [(1) γ-glutamyl carboxylase, (2) vitamin K epoxide reductase and (3) NAD(P)H-dependent quinone reductase] are shown in Fig. 1A. In all tissues and cells that synthesise Gla proteins, the active cofactor form of vitamin K required by the γ-glutamyl carboxylase is not the stable quinone structure found in the diet but the intermittently reduced quinol (or hydroquinone) structure (KH₂). The vitamin K 2,3 epoxide (VKO) produced during γ-glutamyl carboxylation is recycled back to vitamin KH₂ via vitamin K quinone. The only enzyme that can carry out the reduction of VKO to vitamin K quinone is a small 18-kDa protein called vitamin K epoxide reductase (VKOR). Although several enzymes, including the VKOR, can carry out the reduction of vitamin K to KH₂ in vitro, the identity of the enzyme that carries out this reduction in vivo is currently unknown.¹⁵

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 1 

  Scheme showing the vitamin K epoxide cycle in the absence (A) and presence (B) of warfarin. A shows the linkage of the post-translational conversion of peptide-bound glutamic acid (Glu) to γ-carboxy glutamic acid (Gla) residues to the metabolic recycling of vitamin K by a pathway known as the vitamin K-epoxide cycle. Enzyme activities shown are (1) γ-glutamyl carboxylase; (2) vitamin K epoxide reductase (VKOR) and (3) NAD(P)H-dependent quinone reductase(s). The active cofactor form of vitamin K required by the γ-glutamyl carboxylase is the reduced form vitamin K quinol (KH₂). During γ-glutamyl carboxylation KH₂ becomes oxidised to vitamin K epoxide (KO) which in turn undergoes reductive recycling, first to the vitamin K quinone (K) and then to KH₂. Only the VKOR enzyme can carry out the reduction of KO to K and under usual physiological conditions both VKOR and NAD(P)H-dependent quinone reductases can carry out the reduction of K to KH₂. B shows the metabolic inhibition and consequences of a vitamin K antagonist such as warfarin. These drugs block the activity of the VKOR (2) leading to an accumulation of KO in the cell. Given a sufficient supply of vitamin K (e.g. from the diet) an alternative hepatic quinone reductase activity (3) can bypass the warfarin inhibition of the VKOR to provide the KH₂ substrate for the carboxylase enzyme and overcome the inhibitory action of warfarin, even under extreme blockade. The scheme also shows that in the absence of warfarin, the carboxylated substrates (Gla proteins) are secreted into the circulation (A) whereas in the presence of warfarin, species of undercarboxylated forms called PIVKAS are also secreted into the circulation (B). This Figure originally published in “Vitamins in the prevention of human diseases/Wolfgang Herrmann, Rima Obeid/De Gruyter, Berlin 2011” reproduced with kind permission of De Gruyter.

Back to Article Outline

4. Mode of action of oral vitamin K antagonists 

An important property of the dithiol-dependent VKOR discussed above is its sensitivity to VKA drugs such as warfarin (Fig. 1B). It is now clear that the therapeutic effect of VKA is based on their ability to inhibit the activity of VKOR activity and block the recycling of VKO. Since it is well known that vitamin K can overcome this inhibition, the question arises as to how vitamin K is reduced to vitamin KH₂ if the VKOR is completely blocked as it can be in anticoagulant overdose. The answer seems to be that the liver contains other NAD(P)H-dependent quinone reductases that are less sensitive to warfarin and which provide an alternative pathway for the reduction of vitamin K to vitamin KH₂. This quinone reductase activity equates to enzyme activity (3) shown in Fig. 1A and B. It follows that stable therapeutic anticoagulation with VKA is dependent on a balance being achieved between the inhibition of the recycling enzymes and the amount of dietary vitamin K that can enter the cycle to support carboxylation at a reduced efficiency. The effect of inhibition of VKOR is to induce a relative vitamin K deficiency at its site of action in the liver, thus decreasing the production of biologically active carboxylated coagulation factors by the liver and lowering their concentration in the circulation. Several different VKA drugs are in clinical use throughout the world belonging to the class of compounds called 4-hydroxycoumarins or 1,3 indandiones. The major 4-hydroxycoumarin anticoagulants in current use are warfarin, acenocoumarol, and phenprocoumon. The major 1,3 indandiones in clinical use are fluindione and phenindione. The usage of individual drugs varies from country to country and these differences can be ascribed to longstanding historical reasons such as the drug supplier and uniformity of prescription in a given country rather than being based on any solid scientific evidence of superiority. There are probable differences in the degree of receptor binding of different VKA drugs to the VKOR and known differences between their dose requirements, plasma half-lives and routes of metabolism.¹⁶ However, it is likely that each VKA has exactly the same mode of action at the molecular level as an inhibitor of VKOR and that the mechanism by which vitamin K acts to counteract their inhibitory effects is also the same.

Back to Article Outline

5. Vitamin K requirements for stable oral anticoagulation 

It follows from the above description of the general mechanism of action of VKA that the objective of oral anticoagulant therapy is to achieve a balance between the degree of inhibition of the VKOR enzyme and the availability of reduced vitamin K that feeds into the vitamin K cycle and drives the synthesis of the clotting factors at a reduced rate. Ideally, to achieve stable anticoagulation with a constant daily dose of a VKA, the daily amounts of vitamin K available to the hepatic site of synthesis of the vitamin K-dependent clotting factors need to be kept constant as well. In the usual situation this is difficult because the major dietary source, phylloquinone (vitamin K₁), is present in different foodstuffs at very variable concentrations. Allied to this, the vitamin K intake of most individuals varies widely from day to day.¹⁷ On the other hand, despite this inevitable dietary variability, many patients achieve remarkable stability on the same, constant dose of anticoagulant. This suggests that there may be adaptive mechanisms that maintain hepatic stores of vitamin K. Such compensatory mechanisms may include increased excretion of higher intakes¹⁸ and mobilisation of extrahepatic stores from organs such as fat and bone when intakes are reduced. There is evidence that in well nourished people extrahepatic stores of phylloquinone exceed those of the liver^[19], ^[20] and that even very low dose warfarin has a pertubatory effect on vitamin K transport causing a rise in fasting plasma phylloquinone concentrations that may represent a mobilisation of vitamin K body stores to maintain hepatic γ-glutamyl carboxylation.²¹ The hypothesis that there are biochemical mechanisms that protect the synthesis of the vitamin K-dependent (Gla) coagulation proteins at the expense of other Gla proteins (an expanding number of Gla proteins are now known to be synthesised in most tissues and cells) has been cited as an example of the triage hypothesis originally proposed by Bruce Ames.²² The triage hypothesis posits that, when the availability of a micronutrient is inadequate, there are biological mechanisms that ensure that functions essential for short-term survival are protected at the expense of functions whose lack has only longer-term consequences. For vitamin K the only life threatening consequence of an acute deficiency of vitamin K is bleeding so it is at least plausible that biological mechanisms may exist to try and protect this function.

Back to Article Outline

6. Nutritional aspects of vitamin K: sources, and assessment of dietary intakes 

6.1. Nutritional sources of vitamin K (dietary and non-dietary) 

The highest concentrations of phylloquinone (in the range of 400–700μg/100g) are found in green vegetables corresponding to its known association with photosynthetic tissues. The next best sources are certain vegetable oils (e.g., soybean, rapeseed, and olive oils) with phylloquinone contents in the range of 50–200μg/100g. Other vegetable oils, such as peanut, corn, coconut, groundnut, sunflower and safflower oils, have much lower contents of phylloquinone in the range of 1–10μg/100g. Many food items in different food categories such as fruits, grains and dairy products also contain phylloquinone in the range of 1–10μg/100g.²³

The list of foods with nutritionally significant concentrations of menaquinones is more restricted than those containing phylloquinone. Apart from animal livers of ruminants that contain long chain menaquinones (MK-7 through to MK-13) at relatively high concentrations (total of 10–20μg/100g), the richest dietary sources of menaquinones are fermented foods, typically represented by cheeses (MK-8, MK-9) in Western diets and natto (MK-7) in Japan. Yeasts do not synthesise menaquinones and the menaquinone content of cheeses and yogurts derive from the starter bacterial culture rather than from any moulds that may be used in their manufacture. The traditional Japanese food natto (fermented soybeans) has a very high content of menaquinone-7 (MK-7) content of about 1000μg/100g in a highly bioavailable form.²⁴ The relative concentrations of MK-8 and MK-9 in cheeses range from 10–20μg/100g for MK-8 to 35–55μg/100g for MK-9.²⁴ Lower contents of MK-8 and MK-9 are found in yogurts. Menaquinone-4 differs from other menaquinones in both not being a common bacterial form and by the fact that MK-4 is uniquely synthesised from phylloquinone in animal tissues or from the precursor menadione that enters the human food chain through its widespread use in animal feeds.¹³ This agricultural use of menadione probably explains the relatively high concentrations of MK-4 in some foods such as egg yolk (~30μg/100g), butter (~15μg/100g), cheeses (~5μg/100g), and meats (~1–10μg/100g).

The human intestinal microflora synthesise large amounts of menaquinones, which in theory could serve as a potential source of vitamin K for human requirements. By far the largest reservoir of intestinal bacteria is found in the large intestine. Quantitatively, the Bacteroides and Bifidobacterium genera together account for over half of the total anaerobic bacterial population of which only bacteroides synthesise menaquinones. The major bacterial menaquinones present in the human large intestine are MK-10 and MK-11 synthesised by Bacteroides, MK-8 by Enterobacteria, MK-7 by Veillonella, and MK-6 by Eubacterium lentum.²⁵

The longstanding question of whether the colonic microbiota provide a quantitatively significant source of menaquinones that can be absorbed and utilised has still not been satisfactorily answered and has been reviewed elsewhere.^[26], ^[27] Overall, the weight of evidence suggests that although gut menaquinones might make some nutritional contribution, their importance is much less than previously thought.^[26], ^[27] Recent evidence that intestinal menaquinones alone cannot maintain Gla protein synthesis comes from studies in healthy human subjects that demonstrate that relatively mild, short term, dietary restriction of vitamin K results in increased concentrations of circulating species of undercarboxylated prothrombin (PIVKA-II) and the bone protein osteocalcin.^[26], ^[28] The reason for their poor bioavailability from bacteria is that menaquinones are highly lipophilic molecules that are tightly bound to the bacterial cytoplasmic membrane, and remain locked in the membrane even after cell death. Even if low concentrations of free menaquinones are present in the colon, the bile salts that are obligatory for their absorption are lacking in this region. Menaquinone-producing bacteria are also present in the distal terminal ileum, albeit at vastly lower concentrations than the colon, and this site represents a more promising site of absorption given sufficient intraluminal concentrations of bile salts.²⁵

6.2. Measurement of vitamin K in foods and establishment of food composition databases for vitamin K 

Although a dietary interaction of vitamin K intakes with VKA therapy has long been recognised, only recently has it been possible to study quantitative aspects of this interaction. The slow progress in the ability to quantitatively assess the effect of diet on oral anticoagulant therapy was originally due to the lack of accurate food composition data for vitamin K which until the mid 1980s relied on chick bioassays that gave only very approximate values.^[2], ^[29] This problem began to be solved during the 1980s after the development of reliable physiochemical methods (based on high performance liquid chromatography) to enable accurate measurements of the vitamin K content of vegetable and animal tissues. With the ability to measure vitamin K in different food types, progress towards the compilation of vitamin K food composition tables proceeded during the 1980s, initially in a rather ad hoc manner, but later in the 1990s using a more systematic approach, particularly in the UK^[23], ^[30] and USA.^[31], ^[32] Nearly all food composition tables have been compiled with respect to the plant form phylloquinone because this was suspected and shown to be the major molecular form of vitamin K present in the majority of foods.

In the UK a usable phylloquinone food table required the direct analysis of the phylloquinone content of around 200 individual food items.³³ Later analyses included the analysis of common composite food dishes. In the absence of data obtained by direct analysis, phylloquinone values were assigned to food items from recipe-based calculations and from food similarities. The UK phylloquinone database was finally incorporated into McCance and Widdowson's “The Composition of Foods”.³⁴ In the USA the phylloquinone content of about 900 foods and drinks is available on line from the US Department of Agriculture.³⁵ It is important to emphasise some limitations of food composition tables for phylloquinone, especially for foods making the highest contribution to human intakes such as green vegetables and oils.³¹ It has been shown that the phylloquinone content of green vegetables is highly correlated to their chlorophyll content so that the phylloquinone content of outer leaves is much higher than the inner leaves.^[23], ^[31] Values for vegetables will also depend on the freshness (water loss), plant maturation, and climatic differences.^[31], ^[36] Another potential source of error is that vitamin K is sensitive to light and it has been shown that the exposure of commercial oils in transparent bottles to sunlight and fluorescent light leads to a considerable loss of their original phylloquinone content although this can be prevented by storage in amber bottles.³⁷

Vegetable fats and oils differ enormously in their natural phylloquinone contents according to the plants from which they are derived.³⁰ This may be compounded by natural variability between different batches of the same oils. In practise, a major problem in calculating the phylloquinone content of many oil or fat-containing foods, especially composite dishes, is the lack of information of the type of fat or oil used in its preparation or industrial manufacture. With the wide (>200-fold) variation in phylloquinone content of oils this can introduce potentially large errors in the calculated food values.

Less attention has been paid to the analysis of menaquinones in foods. Apart from MK-4 that can derive from endogenous synthesis in animals,¹³ dietary sources of menaquinones were expected to be confined to the relatively limited number of foods that employ a bacterial fermentation step in their preparation or manufacture. The exceptions are animal (and human) livers that were shown in the classical studies of Matschiner and co-workers to be a rich source of a wide spectrum of menaquinones, chiefly long-chain lipophilic menaquinones MKs-7 to MK-13.^[38], ^[39], ^[40] Some preliminary values of the menaquinone-content of fermented foods were published by Shearer et al.²³ More recently a more systematic approach to create a food data base for menaquinones has been undertaken by the vitamin K group in Maastricht.²⁴

6.3. Intakes of vitamin K in healthy populations 

Several dietary instruments are used to estimate dietary intakes of nutrients, all of which have advantages and disadvantages.^[41], ^[42] One of the most accurate prospective methods for determining short-term intakes in the home is different versions of the weighed inventory technique in which the subject is provided with a food scale to weigh food items and a diary to record details of the foods consumed. However, obtaining weighed intakes for every meal is laborious and is rarely feasible beyond a few days. More convenient techniques of capturing habitual intakes include the non-weighed food diary in which estimated intakes are entered by the participant over a given number of days (e.g. 4 or 7days) and the food frequency questionnaire (FFQ) that captures the types and quantities of foods consumed over a given period and is often used for large-scale population studies. Very recent intakes can be captured by 24-hour recall of foods consumed.

The major form of vitamin K in most diets is phylloquinone from plant sources and this is the dietary form that is commonly quantified in population surveys. Ideally, we also need to be able to assess the dietary contribution of menaquinones originating either from bacterial synthesis (mainly MKs 5–14) or by endogenous synthesis from menadione or phylloquinone (MK-4). The first estimation of the relative contribution of menaquinones to dietary intakes of vitamin K was carried out by FFQ in an epidemiological study in an aging Dutch population and in whom it was found that about 90% of total vitamin K intakes were provided by phylloquinone, and 10% by menaquinones (7.5% by MKs 5–10, and 2.5% by MK-4).^[43], ^[44] More recently, FFQ-derived estimates of vitamin K intakes in a German population living in the Heidelberg region suggested that the contribution of menaquinones (MKs 4–14) was much higher accounting for 27% of total dietary vitamin K intakes.⁴⁵ Sub analysis showed that 14% of total intakes were provided by MKs 5–10 and 11% by MK-4. Whether this variation between menaquinone intakes in The Netherlands and Germany is due to dietary differences or analytical differences in dietary assessment is not known. However FFQs are known to be prone to greater error compared to other methodologies such as weighed intakes and food diaries and more studies are required to assess the relative contribution of dietary menaquinones to human intakes.

More accurate and consistent data is available for phylloquinone intakes. Several studies in Europe and the USA have shown that some 60–70% of total phylloquinone intakes come from vegetables and vegetable products, of which some 25% to 40% comprises green leafy vegetables.^[45], ^[46], ^[47], ^[48], ^[49]

National surveys in the UK and USA reveal considerable variations in phylloquinone intakes between individuals of these populations.^[47], ^[48], ^[49] In the UK, retrospective analyses of nationally representative vitamin K intakes has been facilitated by the National Diet and Nutrition Surveys (NDNS) that are periodically carried out for different age groups. Such surveys afford a valuable gold standard for patient comparisons because they are based on 4-day or 7-day weighed records and the same UK food database for phylloquinone.³³ In the NDNS survey of British elderly for 1994–1995 the absolute range of daily phylloquinone intakes varied from 9 to 585μg/day with an inner 95% range of 17–244μg/day.⁴⁸ Similar lower 5th and upper 95th percentile values of 23–237 and 22–217μg/day were found in two NDNS surveys of adults aged 16–64years during the years 1986–7 and 2000–1 respectively.⁴⁹ The distribution of phylloquinone intakes is positively skewed. For example in healthy British elderly the geometric mean intake was 65μg/day compared to an arithmetic mean of 82μg/day.⁴⁸ In this study elderly men had significantly higher intakes than women (geometric means of 70 versus 61μg/day respectively).⁴⁸ In healthy young adults geometric (arithmetic) mean phylloquinone intakes were 72 (86) μg/day and 67 (80) μg/day for the 1986–7 and 2000–1 surveys respectively.⁴⁹ Average intakes in the USA (using similar methodology) are comparable but with a trend to higher intakes in older adults.⁴⁷

A longitudinal study in Scottish adults examined the daily and seasonal variation in phylloquinone intakes.¹⁷ Sixty-five participants (aged 22–54years) completed four separate 7-day weighed food records over a period of one year (one for each season). The arithmetic mean intake of phylloquinone over the 4 seasons was 72 (±65) and 64 (±33) μg/day in men and women respectively with no discernible seasonal variation.¹⁷ The high daily variation in phylloquinone intakes within the same individual is highlighted by the finding that the intra-individual coefficient of variation for the 28 records completed by each participant was 103%.

6.4. Intakes of phylloquinone in patients taking oral vitamin K antagonists 

A summary of the dietary intakes of phylloquinone found in patients taking VKA is shown in Table 1. Interpretation needs to be tempered by the fact that the data were obtained in different countries and by different methodologies, including non-standardised food tables for phylloquinone contents.

Table 1. Dietary intakes of phylloquinone (vitamin K₁) in patients at the onset or during treatment with oral vitamin K antagonists.

Author, year and reference	Dietary instrument	Country	Number of patients	Age of patients (SD or range)	Dietary intakes (µg/day) of phylloquinone (vitamin K1)
					Mean (SD)	Median	Range	Lower-upper quartiles
Lubetsky, 199958	FFQ (two 1-week recalls) during treatment a	Israel	50	Men 56 (16) Women 64 (10)	248 (205)	179	17–974
Cushman, 200150	FFQ (6-month recall at onset of treatment) b 4-day weighed intakes for first 4days of treatmentc	USA	40	62 (13)	141 (87) 33 (24)
Khan, 200451	Mean of four separate 4-day weighed intakes during treatment and prior to INR determination	UK	53	67 (24–87)	47 (44)	47	7–377	30–79
Franco, 200463	3-day food diary during treatment	Brazil	12	57 (14)	118 (51)		18–211
Custodio das Dores, 200798	24-hour recall during treatment FFQ (30days to 1year recall) during treatment d	Brazil	115	59 (15)		76 128		47–120 77–177
Sobczynska-Malefora, 200952	7-day food diary at onset of treatment 7-day food diary after 6months of treatment e	UK	45	59 (17) (23–85)	63 (38) 76 (48)		17–180 12–273
Kim, 201067	3-day food diary during treatment	Korea	66	60 (11)		161	31–617

A few features of the data in Table 1 may be noted. Dietary intakes of phylloquinone measured by FFQ tended to be higher than intakes assessed by other methods, especially compared to weighed intakes as has been found in healthy populations.⁴⁷ The problem of overestimation of frequency is well known in questionnaires. This is illustrated by the findings in a comparative study of dietary assessment methods which showed that the vegetable consumption measured by FFQ was almost double that measured by the weighed inventory method resulting in a significant overestimation of micronutrients such as carotene and vitamin C.⁴² The same may well apply to the measurement of phylloquinone intakes by FFQ, especially as vegetables generally account for well over half of phylloquinone intakes. As shown in Table 1, assessments of phylloquinone intakes in orthopaedic patients starting warfarin in a perioperative setting⁵⁰ gave highly disparate values for their usual intakes that were estimated by FFQ (141μg/day) as compared to those measured post-operatively over the first 4days in hospital using a weighed inventory technique (33μg/day). It is difficult to know whether this large reduction in intakes post-surgery is genuine or a methodological problem with the FFQ. Whilst an overestimation of the intakes by FFQ seems likely, some reduction in dietary intakes following hip replacement might also be expected, especially on a hospital diet.

A study from Tyneside, UK in non-surgical, non-resident patients represents the most rigorous assessment to date of dietary intakes of phylloquinone in patients taking oral vitamin K antagonists.⁵¹ It was undertaken in 53 patients (mean age 67years) who were stably anticoagulated (all taking a constant individualised warfarin dose) and who completed four sequential 7-day diaries of all foods consumed in their homes. In analysing the data the authors found that the mean 4-day intakes prior to an INR determination provided the best fit in their regression models and these are the intake values reported in their paper and in Table 1.⁵¹ As found nationally in the UK,^[48], ^[49] the mean daily intakes of phylloquinone in this Tyneside cohort ranged widely from 7 to 377μg/day. It is notable that the arithmetic mean phylloquinone intake of 47μg/day phylloquinone in this mainly elderly warfarin patient group was much lower than the arithmetic mean of 82μg/day reported in a nationally representative cohort of healthy elderly in the UK using comparable weighed inventory techniques⁴⁸ (see also Section 6.3 above).

Table 1 also shows phylloquinone intake data by validated 7-day food diaries collected during a study from our hospital in which the effect of a 6-month course of warfarin therapy on folate status was examined.⁵² Although this study did detect a significant reduction in red-cell folate concentrations which suggested that patients might have inadvertently reduced their intake of vegetables this was not mirrored by the dietary intake data for either folate or phylloquinone. This failure to detect dietary changes might have reflected the measurement accuracy of the folate and phylloquinone intakes and the smaller subset of patients in whom dietary intakes were captured. Nevertheless the mean values of phylloquinone intakes of 60–70μg/day were consistent with studies in both healthy^[48], ^[49] and warfarin-treated patients⁵¹ in the UK population.

6.5. Dietary requirements and recommendations for vitamin K 

The Adequate Intake (AI) for vitamin K in the USA is presently 120μg/day for men and 90μg/day for women.⁵³ The vitamin K Reference Labelling Value for food labelling purposes varies between countries but for adults is about 75 to 80μg of vitamin K/day. Most surveys have shown that actual intakes in the USA and European populations vary widely between individuals but that mean intakes are in the range of 60–200μg/day.⁴⁷

Johnson⁵⁴ discussed the dietary information available to patients taking warfarin in the USA and identified advice that was often vague, inappropriate and conflicting. In particular some sources suggested patients should avoid and/or limit foods high in vitamin K as well as inappropriately proscribing foods that were not high contributors, such as milk.⁵⁴

As pointed out by Johnson⁵⁴ the advice given by various “expert” bodies in the USA is often misleading and confusing to the patient. Some sources⁵⁵ advise patients to “limit” intakes of vitamin K-rich foods but the definitions of “limit” are often not explained and assume knowledge of vitamin K dietary sources and vitamin K contents that few patients would possess.⁵⁴ Allied to this is the lack of knowledge of the bioavailability of phylloquinone from different foods. The most universal (and sensible) advice is to continue normal dietary patterns and avoid gross daily fluctuations in intakes of foods known to have very high vitamin K contents (i.e. certain green leafy vegetables such as spinach). Constancy in dietary intakes of vitamin K is the key. It is the authors' experience that this is poorly understood by many anticoagulation clinics.

Back to Article Outline

7. Experimental evidence for the influence of vitamin K intakes on the sensitivity and stability to VKA 

7.1. Case reports 

There are many case-reports linking dietary intakes of vitamin K as a cause of unstable anticoagulant control. Most describe the association of an identifiable high source of vitamin K resulting in subtherapeutic control rather than low intakes causing supratherapeutic anticoagulant control. Specific examples have been discussed elsewhere and will not be considered further here.^[56], ^[57]

7.2. Observational studies 

One of the earliest observational studies to investigate the relationship of habitual dietary vitamin K intakes to VKA sensitivity in the setting of the anticoagulant clinic followed 50 patients who were commencing warfarin.⁵⁸ The effect of diet on warfarin response was studied over the first eight-week period of warfarin treatment during which the patients consumed their regular diet and with the help of a dietician (blinded to the warfarin treatment) completed a one-week recall dietary questionnaire at weeks 2 and 8. For each patient a “warfarin sensitivity index” (WSI) was calculated as the ratio of the final INR to the final warfarin dose (mg/day/m² of body surface area) at 8weeks. Dietary intakes of phylloquinone ranged from 17 to 974μg/day (Table 1). To study the relationship of dietary intakes on warfarin dose requirements Lubetsky et al.⁵⁸ divided patients into those with “high” and “low” intakes based on a cut off for a “high” intake of ≥250μg/day. Eighteen patients (36%) had an intake of ≥250μg/day that mainly resulted from their higher consumption of green leafy vegetables (lettuce, cabbage, broccoli and spinach). The average daily vitamin K consumption of this “high-intake” cohort was about fourfold higher than their “low-intake” counterparts (505±181 vs. 133±50μg/day) and they required a higher mean warfarin maintenance dose (5.8 vs. 4.4mg/day, P=0.033).

An examination of the relationship between dietary vitamin K intakes and warfarin sensitivity revealed that 16/18 patients with vitamin K intakes of ≥250μg/day had a WSI of 1.1 or below.⁵⁸ Further analysis showed that this WSI value of 1.1 to be optimal (sensitivity 0.89, specificity 0.53) for distinguishing (P<0.001) patients with “high” and “low” vitamin K intakes and was also predictive of both the INR and warfarin dosage.⁵⁸ Thus after 5days of warfarin (loading phase) patients subsequently found to have a WSI value of ≤1.1 at 8weeks had a lower median INR (1.9 vs. 3.0, P<0.001), and required a higher warfarin maintenance dose (5.7 vs. 3.5mg/day, P<0.001). Limitations of this study relate to the unspecified food data base used to calculate vitamin K intakes and accuracy of the one-week FFQ.

A later study by Cushman et al.⁵⁰ was mainly aimed at investigating the association of biochemical markers of vitamin K status with warfarin sensitivity during the initiation phase (see Section 8) but also included an assessment of the influence of habitual dietary intakes of vitamin K on warfarin response. This study followed 40 patients after admission to hospital for elective hip surgery and who began warfarin therapy (5mg warfarin) the evening prior to admission. Whilst in hospital the patients completed an FFQ to assess their usual daily vitamin K intake in the previous 6months. During the 4-day study period, the patients were offered a diet of known vitamin K content designed to provide a daily intake of 65–80μg of phylloquinone based on the US daily recommendation for vitamin K intakes at that time (it has since changed) of 1μg per kg body weight.⁵⁹ No dietary supplements were allowed and any food not consumed was weighed to calculate actual intakes. The FFQ data showed that higher intakes of phylloquinone were significantly associated with a slower rise in PIVKA-II (Protein Induced by Vitamin K Absence) but not with the rise in INR.⁵⁰ This association with PIVKA-II but not with INR is likely to reflect the higher sensitivity and specificity of PIVKA-II as a biochemical marker of γ-glutamyl carboxylation and hence of warfarin response.^[60], ^[61]

Using their meticulous dietary intake data (see Section 6.4 and Table 1) Kamali and co-workers examined how vitamin K intakes measured for either 7 or 4days prior to an INR influenced the observed anticoagulant response in patients with stable warfarin dose requirements.⁵¹ Stepwise regression showed that both 4-day and 7-day intakes correlated negatively with the INR but that the 4-day record was a better predictor of INR than the 7-day record (P<0.002 vs. P<0.02 respectively). This finding that the more recent intake data had a greater influence on INR is in line with knowledge of the rapid metabolic turnover of phylloquinone.^[13], ^[23] The regression plot of the change in INR from baseline against the change in the 4-day mean phylloquinone intakes showed that an increase of 100μg in phylloquinone intakes caused a reduction of 0.2 in the INR value.⁵¹ The extrapolation of this regression equation to higher intakes is limited by the fact that the majority of the data points fell within the intake range of 10 to 100μg/day. This rate of reduction in INR of 0.2 per 100μg increase of phylloquinone intake in foods is slightly lower than that found for equivalent doses of synthetic phylloquinone in the dose–response intervention study of Schurgers et al.⁶² which is discussed in Section 9.2. The lesser effect on INR from phylloquinone contained in food rather than supplements would be expected from the lower bioavailability of phylloquinone from food, especially vegetables.

An observational study from Brazil also provided evidence of the importance of recent dietary vitamin K intake as a significant determinant of INR during chronic oral anticoagulation.⁶³ In this study, 39 patients who made 230 visits were interviewed by a trained nutritionist to assess whether in the week before their visit they had eaten more, less or the usual amount of vitamin K-rich foods. Vitamin K intake was also scored semi-quantitatively by assigning a score of 0, 1 or −1 according to whether the patients' intakes had been unchanged, greater or less than their usual consumption respectively. The food analysis showed that patients with INR values <2 had consumed significantly greater amounts of vitamin K-rich foods (particularly green vegetables) whilst patients with INR values >4 had significantly decreased their consumption of the same foods during the week before testing. Univariate analysis revealed that the vitamin K intake score was inversely associated with the level of anticoagulation. Multivariate regression analysis showed that the vitamin K intake score was the only variable independently associated with an INR>4 (odds ratio 1.15; 95% CI 1.01 to 1.3; P=0.04) and was one of three variables independently associated with an INR<2 (odds ratio 0.89; 95% CI 0.79 to 0.99; P=0.04).

A follow up study by Kamali's group⁶⁴ was aimed at determining dietary intakes in a group of patients defined as having unstable control anticoagulation to examine whether their intakes differed from those in the stable patients from their first study.⁵¹ In this study a patient was classified as unstable if the standard deviation of his/her INR values was >0.5 and had required three or more dose changes within the previous six months. Each patient completed two consecutive 7-day weighed intake food diaries with analysable data being obtained for 26 of 30 eligible patients. For comparison, equivalent intake data was obtained from 26 of 53 patients with stable anticoagulation taken from their earlier study and matched for demographics against the unstable group. The results showed that the mean (SD) daily intake in the unstable group was only 29 (±17) μg and significantly lower than the mean daily intake of 76 (±40) μg in patients with stable anticoagulation.⁶⁴ This study suggested that patients with unpredictable and unstable anticoagulant control consume significantly lower amounts of vitamin K than their matched stable counterparts. One possible limitation of this study was that the intake data from the control patients with stable anticoagulation was derived retrospectively from a separate study. However this criticism is obviated by the large difference in the mean dietary intake of phylloquinone in the unstable patients, which was less than half that in the stable group, and by the fact that all the intake data was obtained using the same weighed inventory technique, which is regarded as one of the most accurate methods of determining nutrient intakes.

A prospective cohort study from Leiden which investigated the relationship between vitamin K intakes and subtherapeutic INR values has strengthened the hypothesis that low dietary intakes of vitamin K are a risk factor for unstable anticoagulation.⁶⁵ The cohort comprised a total of 1157 patients who had reached stable anticoagulation and who had been sent an FFQ by mail to determine their usual vitamin K intakes. Based on the data from 840 (73%) valid FFQs returned, patients were categorised as having normal (100–300μg/day; 63% of patients), high (>300μg/day; 30%) or low (<100μg/day; 7%) vitamin K intakes. Patients were then followed until their first subtherapeutic INR was recorded when they were deemed eligible as a ‘case’ for inclusion into a nested case–control study against two control patients from the same cohort who had not as yet experienced a subtherapeutic INR.⁶⁵ Of the 1157 patients, 335 had a subtherapeutic INR during follow-up. Analysis of usual vitamin K intakes showed that compared to patients with a normal usual vitamin K intake, those with a high usual K intake had a slightly lower risk of a subtherapeutic INR (Hazard Ratio [HR] 0.80; 95% CI 0.56–1.16) and those with a low usual intake had a higher risk (HR 1.33; 95% CI 0.79–2.25). When patients with high and low usual vitamin K intakes were compared directly, those with a low usual intake were found to have a 1.66-fold increased risk for a subtherapeutic INR (95% CI 0.93–2.96). For the nested case–control study, recent vitamin K intake over the last 48-hours was collected from a subgroup (63 cases; 188 controls) by dietary recall. Analysis of the three dietary groups revealed that amongst those with a low usual vitamin K intake, the patient cases who had developed a subtherapeutic INR had consumed twice as much vitamin K in the previous 48h as the controls (164 vs. 83μg/day). No such case–control differences were observed in patients with a normal or high usual vitamin K intake. The authors concluded that a high vitamin K intake reduces the risk of a low INR by lessening the influence of incidental consumption of vitamin K-rich foods on the INR.⁶⁵ Evidence that this beneficial effect could be due to the lower relative variability in intakes of individuals with high vitamin K intakes is provided by findings showing that the proportion of healthy elderly individuals with a high relative variability (>20%) in their usual intake decreased as the geometric mean of their intakes increased.⁶⁶

Further support for the importance of the level of dietary phylloquinone to the stability of anticoagulant control comes from a recent study from South Korea.⁶⁷ In this study the phylloquinone intakes of 66 patients determined by 3-day food diaries were negatively and independently correlated to anticoagulant stability as expressed by the coefficient of variation of both INR and warfarin dose.⁶⁷ In addition, there was a progressive lowering of the coefficient of variation of the INR as the dietary intakes increased through the dividing tertiles.⁶⁷ The authors concluded that long-term anticoagulation with warfarin is more stable in patients with phylloquinone intakes greater than 195μg/day. However, phylloquinone intakes in this Korean study⁶⁷ were 2–3 fold higher than the values found in the UK study that first reported an association of low intakes with less stable anticoagulant control⁶⁴ (Table 1). These differences could reflect methodological differences in dietary assessments or to genuinely higher phylloquinone intakes in South Korea.

Back to Article Outline

8. Relationships of biochemical measures of vitamin K status to response to VKA 

In addition to assessing the effects of dietary vitamin K intakes to anticoagulant response, a few studies have also included measures of vitamin K status and metabolism. The two biomarkers used to date are circulating concentrations of phylloquinone, the major form of vitamin K in the blood, together with its metabolite phylloquinone epoxide. Evidence from many studies suggests that circulating phylloquinone reflects tissue stores^[2], ^[27] including the liver⁶⁸ although there are caveats. A major limitation is that because phylloquinone is mainly transported in blood with triglyceride-rich lipoproteins (TRL), circulating phylloquinone concentrations are readily influenced by lipid and lipoprotein concentrations and recent dietary intake. Other limitations are that measurement of phylloquinone does not reflect tissue stores of menaquinones that contribute to overall vitamin K status²⁷whilst VKA antagonists may themselves perturb vitamin K transport.²¹ As previously outlined in Section 4, phylloquinone epoxide is an intermediate of the vitamin K cycle and accumulates in the liver and other tissues when the VKOR is inhibited by vitamin K antagonists.^[2], ^[27], ^[69] In humans treated with vitamin K antagonists, increased concentrations of phylloquinone epoxide are released into the circulation.^[70], ^[71] The interpretation of circulating phylloquinone epoxide concentrations during VKA therapy is not well understood. In controlled dose–response studies in which healthy volunteers were given increasing single doses of warfarin, the plasma kinetics of entry and clearance of phylloquinone epoxide in response to a 45μg tritium-labelled intravenous bolus of phylloquinone were shown to resemble a bell shaped curve over 8h with the accumulation closely reflecting warfarin dosage.⁷¹ Importantly, the maximum plasma concentrations of labelled phylloquinone epoxide were shown to be related to both dose and plasma concentration of warfarin by typical log-dose response curves that were linear over the therapeutic range.⁷¹ In this and other studies^[72], ^[73] the dose–response curves of phylloquinone epoxide accumulation have been interpreted as closely reflecting the degree of inhibition of the VKOR, the receptor for VKA. In the absence of VKA, plasma concentrations of phylloquinone epoxide are virtually undetectable.

The relationship of plasma concentrations of phylloquinone and phylloquinone epoxide to the sensitivity to warfarin anticoagulation was studied in a cross-sectional study in 73 patients with stable anticoagulation.⁷⁴ Multivariate analysis revealed significant relationships between INR and both plasma phylloquinone concentrations (P=0.034) and phylloquinone epoxide (P=0.028). As expected the association between INR and plasma phylloquinone was negative reflecting the greater anticoagulant response in patients with a lower vitamin K status. The finding of an opposite positive association of INR with phylloquinone epoxide confirms that raised concentrations of this metabolite are directly reflecting anticoagulant response, and as predicted by early biochemical studies, the inhibition of the cyclic interconversion of vitamin K and vitamin K epoxide.⁷¹

Similar associations of vitamin K status with warfarin sensitivity were obtained by Cushman et al.⁵⁰ but at the onset of treatment rather than when patients had been stabilised (see Section 7.2.1. for study protocol). In this study, baseline fasting phylloquinone and phylloquinone epoxide measurements were made on the morning (day 1) that the patients had had been admitted for hip surgery having all taken the same dose of 5mg warfarin the previous evening. Warfarin sensitivity was assessed from the increase in INR on the day after the operation (day 2 of hospitalisation) with the patients receiving a second 5mg warfarin dose after the operation. A second measure of warfarin sensitivity was obtained by measuring PIVKA-II on day 4 of hospitalisation. Of the three vitamin K status variables measured (habitual dietary intake, plasma phylloquinone and phylloquinone epoxide), only plasma phylloquinone measured on admission was significantly associated with the 2-day change in INR (P=0.003). However, in the same multivariate model, the 4-day change in PIVKA-II was a more informative predictor of warfarin sensitivity being significantly associated with all three status indicators (P=≤0.05 for all). As found in the earlier study of Kamali et al.⁷⁴ the directions of association were negative for plasma phylloquinone and positive for its epoxide metabolite.⁵⁰ Thus, lower 4-day increases in PIVKA-II were associated with a higher plasma phylloquinone (P=0.03) but a lower phylloquinone epoxide (P=0.01) as measured on hospital admission, about 12h post the first 5mg dose of warfarin. These associations support the concept that plasma phylloquinone concentrations were reflecting endogenous vitamin K stores whilst phylloquinone epoxide concentrations were reflecting the pharmacodynamic response at the level of VKOR inhibition.

The above studies demonstrate that warfarin sensitivity is related to initial vitamin K status as assessed by circulating phylloquinone concentrations but did not provide information as to whether this biochemical marker of status is predictive for the stability of anticoagulation as would be expected from the association of low dietary intakes with unstable anticoagulant control.⁶⁴ Some affirmation that instability of VKA therapy is associated with an underlying poor vitamin K status has been obtained by an Israeli group.^[75], ^[76] First clues came from observations that 3 patients in whom initiation or cessation of a daily multivitamin preparation containing 25μg of phylloquinone resulted in sub-or supra-therapeutic INR values respectively, with severe complications in two of them.⁷⁵ They hypothesised that an unsuspected low vitamin K status might have rendered these patients exceptionally responsive to low-dose vitamin K supplements and went on to show that about 12% of the patients attending their anticoagulant clinic had “low” plasma phylloquinone concentrations defined as ≤0.1μg/l.⁷⁵ This report is limited by being applicable to 3 individual patients in whom vitamin K status was not assessed, and to unrelated observations of plasma vitamin K levels in another cohort of patients. In a second study, the same group investigated the effect of giving a small extra vitamin K supplement (25μg/day) for 4weeks to a small group of patients taking warfarin who had been stratified according to whether they had “normal” (n=7) or “low” (n=9) plasma phylloquinone concentrations.⁷⁶ During this 4-week period, sub-therapeutic INR values were recorded in 9/9 patients with an initially “low” vitamin K status but only in 1/7 patients with a “normal” vitamin K status. The data suggested that patients who have low vitamin K stores are more sensitive to dietary vitamin K supplements than patients whose stores are replete. The limitations of this second study⁷⁶ are the small sample size and reliance on a single measurement of non-fasting total circulating phylloquinone (plus its epoxide metabolite) to assess vitamin K status.

Back to Article Outline

9. Intervention studies: effect of vitamin K supplementation or dietary management on response to VKA 

Several dose–response studies have been carried out to test the influence of dietary intervention on anticoagulant stability during VKA therapy. Such studies can be divided into two general designs according to whether the investigators have used a dietary approach to manipulate vitamin K intakes or whether they have used synthetic vitamin K supplements. To date most studies have examined the effects of short-term intervention over days or weeks with relatively few long-term studies that have been carried out over several months. Earlier studies carried out before the introduction of the INR used either the Thrombotest⁷⁷ or Normotest^[78], ^[79] to monitor anticoagulant response. Both these tests are modifications of Quick's one stage prothrombin time; main differences are that the Thrombotest uses bovine thromboplastin and is highly insensitive to PIVKA whereas the Normotest uses rabbit thromboplastin and is minimally influenced by PIVKA.⁸⁰

9.1. Short-term dose–response studies in patients 

The first systematic experimental study to evaluate the influence of both vitamin K supplements and vitamin K-rich foods on the pharmacodynamic response to warfarin was carried out by Karlson et al.⁸¹ in Sweden. This study was carried out in 21 patients who had been stabilised on warfarin as monitored by the Thrombotest for at least 3months. The patients had been advised to keep to their usual diet and avoid excessive green vegetables, cabbage, liver and alcohol. In phase I of the study the patients were given either a single oral dose of 250μg of synthetic phylloquinone (Konakion) or single meals of 250g of spinach and broccoli that from the food data available at that time²⁹ were calculated to be equivalent to a range of 500 to 800μg phylloquinone. Over the 4 next days it was found that although both the single dose of synthetic phylloquinone and the single portions of vegetables tended to elicit a slight rise of the Thrombotest (peaking on day 2 after the doses), the values remained within the anticoagulation target range.

In phase II of the study, the investigators studied the effects of the administration of 100, 250 and 500μg of synthetic phylloquinone, or meals containing 250g spinach or 250g broccoli, given daily to 11 patients over one week. After the daily administration of 100μg of pure phylloquinone there was a slight but non-significant rise in the Thrombotest values over the week. In contrast, supplementation with the higher doses of 250 or 500μg of phylloquinone caused the Thrombotest values to significantly increase above baseline within 3–4days and to exceed the therapeutic limit of the Thrombotest, requiring warfarin dose adjustment.⁸¹ The corresponding rise in the Thrombotest from the daily ingestion of 250g of broccoli was similar to that observed with a daily dose of 250μg of synthetic phylloquinone whilst the response to 250μg spinach was even more marked.

An experimental study with a similar design to that of Karlson et al.⁸¹ was carried out in Denmark using the Normotest to monitor anticoagulant response to warfarin.⁸² In this study, patients on stable anticoagulation were randomised to take either of three dietary regimens comprising either a high daily intake of “vitamin K-rich” vegetables, a high intake of “vitamin K-poor” vegetables or their habitual diet supplemented with 1000μg of synthetic phylloquinone. Using the then available food tables the authors calculated that the phylloquinone-rich vegetables contained 1100μg phylloquinone and the vitamin K-poor vegetables 135μg of phylloquinone. The patient groups were further subdivided to take the regimens for a varying number of days, namely 1day (n=5), 2days (n=7) or 7days (n=13). In the group who took the phylloquinone-rich diet for 7days, plasma coagulant activity had increased in all patients by day 5 and in 9/13 patients (69%) the activity (% Normotest) was above the therapeutic range of 10–25% (corresponding to INR values of 2.0–3.6).⁸² These disturbances in warfarin stability lasted for several days after resumption of the patients' habitual diet. In the smaller groups who took the diets for one and two days, the intake was sufficient to increase coagulant activity beyond the therapeutic range in 2/5 and 3/7 patients respectively. No changes in plasma coagulant activity were observed in the patients eating the diet rich in “vitamin K-poor” vegetables for 6days but supplementation with 1000μg of synthetic phylloquinone caused a much sharper and sustained increase in coagulant activity.⁸²

It is pertinent to note that in attempting to reconcile the greater neutralising effect on warfarin activity of synthetic phylloquinone compared to similar amounts of phylloquinone provided by green vegetables, both Karlson et al.⁸¹ and Pederson et al.⁸² had speculated that the vitamin K content of the vegetables may have been overestimated. They argued that this overestimation might have arisen because of inaccuracies in the vitamin K food data base available at that time²⁹ or to changes in vitamin K content brought about by cooking or storage. In the light of more recent knowledge the reason for the lack of comparability between the effectiveness of synthetic phylloquinone compared to plant phylloquinone can be explained by the inefficient intestinal absorption of phylloquinone from green vegetables. Thus, in two independent absorption studies the relative bioavailability of phylloquinone from spinach (as judged from the areas under the plasma-time curves) was estimated to be about 5% to 15% of that from the same dose of pure vitamin K preparations.^[83], ^[84] We also know that the calculations of the vitamin K contents were derived from inaccurate food tables based on bioassays. From modern food tables based on physiochemical analyses we can recalculate that the 250g of spinach and broccoli used in the study by Karlson et al.⁸¹ would have contained ~1000μg and ~500μg of phylloquinone respectively rather than the authors' estimate of ~500 to 800μg phylloquinone for both these vegetables. It seems likely that this approximately 2-fold difference in phylloquinone contents between spinach and broccoli accounted for the greater effect of spinach over broccoli in their study.⁸¹

In a second phase to their observational study of the role of dietary vitamin K intake on anticoagulant stability, Franco et al.⁶³ carried out a small randomised crossover trial to assess the effects of extreme, short-term increases and decreases in vitamin K intake. Participating patients first completed a 3-day food dairy to assess their habitual vitamin K intake (mean 118±51μg/day) and were then randomly assigned to either a vitamin K-rich or vitamin K-depleted diet for 4 consecutive days with a 1- to 2-week washout period between diets. Intakes of phylloquinone increased to an average of 591±257μg/day during the enriched phase and decreased to 26±8μg/day during the depleted-phase. Of the 11/13 patients who completed both crossover phases, a statistically significant effect on the INR on day 4 was seen only after the vitamin K-rich diet with the INR decreasing from 3.1±0.8 at baseline to 2.8±0.6 (P=0.04). The authors attributed the lack of effect of the reduced intake on day 4 to the lower percentage decrease in intakes (80%) during this depletion phase than the increase during the repletion phase (500%). However a delayed effect on INR of the vitamin K-depleted diet was seen on day 7, some 3days after the patients had returned to their normal diet, with the INR increasing from 2.6±0.5 at baseline to 3.3±0.9 (P=0.005).

9.2. Short-term dose–response studies in healthy volunteers 

Studies that update the findings of the dose–response studies of Karlson et al.⁸¹ and Pedersen et al.⁸² to the modern context of INR monitoring have been carried out by the Maastricht Vitamin K group.^[62], ^[85] As with the earlier Scandinavian studies, the objective of the Dutch studies was to determine the dose–response relationships of different daily vitamin K supplements (and individual meals) on the stability of oral anticoagulant therapy with the difference that the studies were carried out in young, healthy volunteers rather than in patients. In addition to the detailed coagulation assessments these updated dose–response studies included measurements of background dietary intakes and serum levels of phylloquinone together with biochemical assays designed to evaluate the effects of supplementation on the γ-carboxylation status of hepatic coagulation factors II and VII together with the vitamin K-dependent bone protein, osteocalcin.^[62], ^[85]

The first study⁶² comprised two phases; phase I was a dose escalation study designed to assess the response to weekly incremental doses of dietary phylloquinone on anticoagulant stability whilst phase II was designed to test the response to vitamin K-rich food items. For phase I, 12 healthy volunteers were first stably anticoagulated with acenocoumarol (INR target range of 2.0) and maintained on the same dose so that the response to weekly incremental doses of dietary phylloquinone given as a daily supplement could be measured.⁶² During the study, subjects were instructed to refrain from eating foods known to be rich in phylloquinone (e.g. spinach, kale, broccoli, Brussels sprouts) as well as those rich in menaquinones (e.g. curd cheese). The average background intake of phylloquinone from food sources was accurately assessed from 7-day food records to be 55μg/day. The dose range of phylloquinone supplements was 50–500μg/day with each dose being taken daily for 7days. A summary of the response of the INR and factor IIc values to the increasing weekly supplements of phylloquinone are shown in Table 2. The threshold phylloquinone dose causing a statistically significant lowering of the mean INR value from baseline (end of adjustment phase week) was 150μg/day in women (and all patients) and 200μg/day in men.⁶² When the change in INR from baseline was analysed individually, a response was found for one woman after the dose had reached 100μg/day. However, based on the clinical criteria for adjusting the VKA dose in this study to maintain the target INR, none of the participants would have required any dose adjustment whilst taking supplemental phylloquinone at a level of 100μg/day but 3/12 participants would have required a higher acenocoumarol dose at a supplemental phylloquinone intake of 150μg/day. Using more sensitive functional markers (PIVKA-II and undercarboxylated osteocalcin) it was shown that whereas hepatic carboxylation of FII was significantly increased at a dose of 100μg, a dose of 300μg phylloquinone was needed to affect the carboxylation status of osteocalcin in the bone.⁶² These findings are consistent with previous evidence that the uptake of dietary phylloquinone by the liver is more efficient than by bone.^[13], ^[86]

Table 2. Effect of weekly increases in vitamin K supplements on sensitivity to oral vitamin K antagonists: dose–response studies in healthy volunteers.

	Study reference (Schurgers et al. 2004)62			Study reference (Schurgers et al. 2007)85			Study reference (Schurgers et al. 2007)85
	Phylloquinone (vitamin K1)			Phylloquinone (vitamin K1)			Menaquinone-7 (vitamin K2)
Week	Dose (μg)	INR (SD) % change	% F II (SD) % change	Dose (μg)	INR (SD) % change	% F II (SD) % change	Dose (μg)	INR (SD) % change	% FII (SD)
1	0	1.93 (0.26)	45.7 (7.1)	0	2.02 (0.34)	46.3 (9.0)	0	1.98	45.7 (12.5)
2	50μg	1.85 (0.31) 4.1%	59.0 (8.4) 9.4%	50μg	1.89 (0.26) 6.4%	45.7 (7.1) 1.2%	–	–	–
3	100μg	1.75 (0.26) 9.3%	48.3 (17.2) 5.7%	100μg	1.79 (0.31) 11.3%	50.0 (8.4) 8.0%	95μg	1.66 (0.29) 16.2%	53.7 (13.6) 17.5%
4	150μg	1.58 (0.20) 18.1%	55.0 (8.0) 20.4%	150μg	1.74 (0.26) 13.9%	48.3 (14.2) 4.3%	–	–	–
5	200μg	1.53 (0.18) 20.7%	59.3 (7.5) 29.8%	200μg	1.60 (0.20) 20.8%	55.0 (8.0) 18.8%	190μg	1.37 (0.24) 30.8%	63.8 (10.5) 39.6%
6	250μg	1.47 (0.24) 23.8%	61.9 (6.7) 35.4%	250μg	1.53 (0.18) 24.3%	59.3 (7.5) 28.1%	–	–	–
7	300μg	1.42 (0.17) 26.4%	64.5 (9.5) 41.1%	300μg	1.47 (0.24) 27.2%	61.9 (6.7) 33.7%	285μg	1.27 (0.13) 35.9%	70.8 (11.3) 54.9%
8	500μg	1.37 (0.12) 29.9%	–	500μg	1.35 (0.17) 33.2%	64.5 (9.5) 18.2%	–	–	–

In phase II of this study⁶² the same subjects undertook a 2-week vitamin K-washout period in which they continued to take their normal meals (avoiding vitamin K-rich items as in phase I with a mean daily intake of 55μgK₁/day) whilst maintaining their previously individualised daily dose of acenocoumarol. After re-establishing stable anticoagulation (INR 1.97±0.3), the effects of a single serving of four different vitamin K-rich food items were tested. The food items tested contained either phylloquinone from 400g portions of cooked broccoli and spinach (providing 700μg and 1500μg phylloquinone respectively) or menaquinones from either 500g of curd cheese (providing 103μg as MK-9 plus 47μg as MK-8) or 100g of natto (providing 1000μg of MK-7). The response to the meals of spinach and broccoli were far less than those predicted from their phylloquinone contents and although both vegetables brought about a similar and significant statistical reduction in the INR, the changes were not clinically relevant. The negligible and short-lived effects on anticoagulation from these single large servings of phylloquinone-rich vegetables in young, healthy volunteers are in agreement with earlier studies^[81], ^[82] in patients taking VKA (see Section 9.1). The lack of clinical effect from single high-dose phylloquinone contained in green vegetables can be reasonably explained by the inefficient bioavailability from these sources. In contrast a much smaller portion of 100g of natto caused a reduction in INR from 2.01 (±0.57 to 1.50 (±0.21) within 24h which declined further to 1.44 (±0.21) at 4days. With its very high MK-7 content, natto has previously been reported to cause resistance in patients taking VKA therapy.⁸⁷ The single large serving of curd cheese had no significant effect on anticoagulation; this is likely to be due to its relatively low vitamin K content even though cheese products comprise one of the richest sources of menaquinones in Western style diets.

The Maastricht group was also the first to compare the relative efficacies of phylloquinone with the bacterial form MK-7 in counteracting VKA therapy.⁸⁵ The rationale for testing MK-7 is the growing availability of this form as an over-the-counter supplement usually marketed with bone health claims. The design of the oral anticoagulant interaction was the same as in their earlier study⁶² with 12 healthy subjects being first stably anticoagulated with acenocoumarol to a target INR of 2.0, maintained on their individualised doses, and then given increasing doses of either phylloquinone or MK-7 pure supplements, with each dose being taken as a tablet for one week before the next higher dose was introduced. As shown in Table 2 the dose–response results for phylloquinone supplements essentially mirrored their earlier study.⁶² Following the phylloquinone dose–response study and a two-week washout period during which anticoagulation was re-established, the same participants were given increasing weekly doses of MK-7.⁸⁵ The antidotal efficacy of MK-7 in reversing anticoagulation was found to be much greater than for phylloquinone (Table 2). After taking a daily MK-7 supplement of 95μg for one week the mean group INR was reduced from the baseline value of 1.98 to a value of 1.66. Overall, the results indicated that on a weight basis MK-7 is approximately 2.5 times more potent in reversing oral anticoagulation than phylloquinone and on a molar basis 3 to 4 times more potent.⁸⁵ The reason for the greater antidotal activity of MK-7 than phylloquinone seems to lie with differences in their metabolism and plasma transport whereby phylloquinone is mainly transported by TRL that have fast clearance kinetics and MK-7 becomes mainly associated with low-density lipoproteins (LDL) that have a much longer residence time in the circulation.^[13], ^[85] Presumably transport of MK-7 in LDL allows greater availability of MK-7 to the hepatic γ-glutamyl carboxylase for coagulation factor synthesis. Schurgers et al.⁸⁵ also showed that MK-7 was more effective than phylloquinone in increasing the degree of γ-carboxylation of circulating osteocalcin in subjects not taking a VKA. Osteocalcin is synthesised by osteoblasts and is known to incompletely carboxylated even in healthy individuals.⁸⁶ The effect on both hepatic and non-hepatic Gla proteins shows that MK-7 had a far greater whole body efficacy on γ-carboxylation than phylloquinone. Of relevance to the question of continuous intakes of different K vitamins is the finding that during administration for 40days of the same molar amounts of phylloquinone and MK-7 (0.22μmol; equivalent to 100μgK₁ and 142μg MK-7), the plasma concentration of phylloquinone attained a steady state concentration after 3days whereas plasma concentrations of MK-7 did not plateau until about 20days.⁸⁵

9.3. Long-term vitamin K supplementation in patients 

As already discussed, Sconce et al.⁶⁴ provided evidence that patients taking warfarin who have a low dietary intake of vitamin K are more likely to suffer from unstable anticoagulant control. Based on their findings the authors suggested that: “daily supplementation with vitamin K could be an alternative method in stabilising anticoagulation control, lessening the impact of variable dietary vitamin K intake”.⁶⁴ Subsequently the same group carried out the first randomised, controlled trial of vitamin K supplementation in Newcastle upon Tyne in the UK.⁸⁸ In this trial, patients taking warfarin for atrial fibrillation with a target INR of 2.0 to 3.0 were first selected as having unstable anticoagulant control based on the variability of their INR values over 6months that showed a standard deviation of >0.5 together with the criterion that this had resulted in at least 3 warfarin dose adjustments. Seventy of these unstable patients were then randomly assigned to receive either a daily dose of 150μg phylloquinone or a placebo (both taken as a 5ml aqueous-ethanolic solution) orally for 6months. The patients randomised to the vitamin K arm showed a greater improvement in the stability of anticoagulation as evidenced by (i) a greater than two-fold decrease in the standard deviation of the INR (−0.24 vs. –0.11, P<0.001) and (ii) a greater increase in time spent within the target INR range (28% vs. 15%, P<0.01). In the group given vitamin K supplements, 19 of 35 (54%) now fulfilled the “stable anticoagulation” criteria vs. 7/33 (21%) in the placebo group.⁸⁸

In another randomised, controlled trial from Leiden in the Netherlands, 200 patients taking phenprocoumon were randomly selected to receive either a daily dose of 100μg phylloquinone or placebo in capsule formulations for 24weeks.⁸⁹ Both groups showed an improvement in the time spent in the therapeutic range from baseline increasing from 80% to 85.5% in the placebo group and from 79% to 89.5% in the vitamin K group. The modest 4% improvement in the vitamin K group was not significant. However patients taking the daily vitamin K supplement were almost twice as likely to have all their INR values within the therapeutic range (43% vs. 24%; RR 1.8, 95% CI 1.1–2.7). It seems likely that this Dutch study was underpowered and that other factors other than lower vitamin K dose or the different anticoagulant (warfarin in UK, phenprocoumon in The Netherlands) accounted for the modest effect. Factors which may have contributed to the different outcomes in the two studies could have been (i) population differences such that the UK patients were pre-selected, unstable patients, all with atrial fibrillation whereas the Leiden patients were unselected and had been anticoagulated for 1year (ii) baseline differences such that 59–63% of the UK patients were in range compared to 79–80% of the Dutch patients and (iii) the INR target range was narrower in the UK (all 2.0–3.0) than in The Netherlands (either 2.0–3.5 or 2.5–4.0).

9.4. Vitamin K supplementation: relevance of different molecular forms of vitamin K 

To date, all the dose–response data in patients has been obtained for phylloquinone, the major dietary form of vitamin K. However, as already discussed (see Section 9.2 and Table 2), one study carried out in healthy human volunteers showed that dietary supplements of MK-7 exert a greater neutralising effect on anticoagulation than phylloquinone.⁸⁵ The strength of this study was that the comparison of these K vitamins was carried out in the same volunteers. Although this data gives some indication of the expected response if patients on VKA therapy were to take MK-7 in the diet, one limitation is that even the lowest dose of 95μg MK-7 was found to have a marked impact on anticoagulation, reducing the INR by 16% after one week of supplementation. Thus further studies are needed to evaluate the effects of lower doses. The greater antidotal potency of MK-7 can be explained by its different metabolic disposition and transport which results in a longer circulating half-life compared to phylloquinone.^[13], ^[85] With the exception of natto, most human diets provide relatively low concentrations of menaquinones, and MK-rich foods are unlikely to present a problem to VKA stability.⁶² However, dietary supplements containing vitamin K do present a potential problem, particularly as their use has increased over the last few years. This follows research evidence that vitamin K and Gla proteins have roles outside of haemostasis, with a major focus on the essentiality of vitamin K for the integrity of bone⁹⁰ and vascular tissues.⁹¹ With evidence that greater daily amounts of vitamin K are required to maintain γ-carboxylation of extrahepatic Gla proteins than the coagulation hepatic proteins, there has been a trend to recommend increased dietary intakes of vitamin K (both phylloquinone and menaquinones). Target populations for bone health include elderly populations at risk of osteoporosis⁹² and specific patient groups such as those with cystic fibrosis.⁹³ Natto-derived MK-7 is readily available from several manufacturers as over-the-counter dietary supplements. These supplements can contain doses as high as 100μg of MK-7 which would detrimentally inhibit VKA therapy.⁸⁵

Recently the European Food Safety Authority has ratified a health claim “that a cause and effect relationship has been established between the dietary intake of vitamin K (both phylloquinone and K₂ forms) and the maintenance of normal bone”.⁹⁴ Food products are only allowed to carry this bone health claim if they contain a minimum of 15% of the Reference Labelling Value for vitamin K (75μg for adults and 12μg for children 6months to 4years). To meet this condition, foods or beverages would need to contain a minimum of 11μg per 100g or 100ml of either phylloquinone or MK-7. Although vitamin K-containing functional foods aimed at the bone health market are not as yet widely available, their future introduction under this legislation will have obvious consequences to patients taking coumarin and indandione anticoagulants.

9.5. Intervention through a dietary vitamin K-guided strategy 

A novel approach to the control of oral vitamin K antagonist therapy has recently been proposed whereby instead of the traditional approach of adjusting the anticoagulant dose, the patient is asked to make adjustments to their dietary intake of vitamin K.⁹⁵ Patients were initially interviewed by a trained nurse who made dietary recommendations such that if under-anticoagulated, the patient would be asked to reduce their consumption of vitamin K-rich food items by half and if over-anticoagulated to double their consumption of the same items until re-evaluation 2weeks later. To maintain simplicity, the dietary instrument was restricted to 16 vitamin K-rich foods^[63], ^[96] and was based on the investigators' previous work demonstrating that a dietary vitamin K score was inversely correlated with the level of anticoagulation⁶³ (see also 7.2, 9.1). Dietary intervention was based exclusively on changes made to the number of times each food was consumed each week with no account of their quantities. Thus if a patient's vitamin K intake was based on the consumption of 3 food items (e.g. lettuce, broccoli and liver) 4 times per week for each item and the INR was below target, the patient was asked to decrease the ingestion of the same food items to 2 times per week.⁹⁵ To test the utility of this dietary approach, 132 patients were enrolled to a randomised control trial in which the dietary vitamin K-guided strategy was directly compared to the conventional approach of anticoagulant dose adjustment using an internationally accepted algorithm. The criteria for entry to the trial were that patients had been taking an oral vitamin K antagonist for at least 3months with the last INR having been outside the pre-specified therapeutic range without a definite cause for instability. All patients were re-evaluated at 15, 30, 60 and 90days with assessments of dietary intakes of vitamin K being made in both groups. The study results showed that over time, the patients allocated to the vitamin K-guided strategy reached their pre-specified INR progressively more frequently, so that after 90days of follow-up, 74% had attained their target INR compared to 58% of patients managed conventionally (P=0.04). Minor bleeding events or the need for parenteral vitamin K were also lower in the dietary managed group (1.5% vs. 11%) without quite attaining significance (P=0.06). This is a unique prospective study and provides evidence that modulation of dietary vitamin K is at least a feasible approach to attain anticoagulation stability. This dietary approach needs to be further investigated in studies of similar design. It will be important to address questions such as the long-term adherence to dietary counselling as well as the cost-effectiveness of this novel strategy.

Back to Article Outline

10. Conclusions 

This review demonstrates that despite the 60years or so in which oral VKAs have been in clinical use there are still many gaps in our knowledge of the quantitative influences of dietary intakes of vitamin K on anticoagulation control. Furthermore, there are few dietary guidelines for patients on VKA therapy and none internationally agreed.

The main technological obstacles to the compilation of accurate food composition tables for vitamin K have now been overcome although accurate data for menaquinones is somewhat lacking. In practise, the estimation of intakes of phylloquinone provides a reasonably robust instrument for vitamin K status assessments for most populations. However, as with other nutrients, the validity of dietary intake data for vitamin K needs to take into account the likely bias and uncertainties known to be associated with the use of different nutrient databases and dietary intake instruments. The wide range of dietary intakes of vitamin K reported in patients taking VKA (Table 1) are broadly in line with an expected bias towards higher intakes using FFQ methodologies compared to the more accurate weighed intake techniques but may also reflect genuine differences in the diets of the representative countries or indeed the demographics of the patient populations taking VKA.

Much remains to be learned about the bioavailability of vitamin K and known and suspected differences in bioavailability from different sources have not yet been factored into dietary vitamin K status assessments. A recent study using stable-isotope (¹³C)-labelled phylloquinone showed that the absorption of a physiological dose (20μg) of phylloquinone in human volunteers is affected by the type of meal (in this study defined as convenience, cosmopolitan and animal-orientated) with which the free phylloquinone was given.⁹⁷ In addition, the bioavailability of the phylloquinone that was contained within the meal matrix differed between the three types of meals ingested (matrix effect).⁹⁷ This might suggest that consideration of dietary intakes alone might offer a rather imprecise guide to the dietary management of VKA therapy. Despite such caveats, several studies discussed in this review have demonstrated clear associations of total phylloquinone intakes to both the sensitivity and stability of the anticoagulant response. The sensitivity towards VKA has also been shown to be associated with biochemical markers including serum phylloquinone as a marker of vitamin K status and serum VKO as a marker of the pharmacological response to VKA. Such assays are under-utilised because they are only available in a few centres but are certainly useful in identifying patients who are over or under sensitive to VKA.

There is at present limited quantitative information of dose–response relationships that are predictive of how changing dietary intakes of phylloquinone affect the pharmacodynamic response to warfarin. In patients, the most informative study to date suggests that, on average, for every 100μg increase in phylloquinone intake in the 4days before the INR is measured the INR will fall by 0.2 units.⁵¹ By comparison, in the two carefully designed short-term dose–response studies carried out in healthy volunteers (Table 2) an extra daily intake of 100μg phylloquinone given in tablet form over one week lowered the INR by an average of 0.2–0.3 units and an extra 200μg phylloquinone by 0.4–0.5 units.^[62], ^[85] A caveat to the supplement studies is that synthetic phylloquinone contains about 10–15% of the essentially inactive cis isomer that would lower the effective antidotal effect for a given dose. By comparing the anticoagulant response generated by single meals with the dose–response curves for pure supplements it was calculated that the relative bioavailabilities of phylloquinone from spinach and broccoli were 13% and 29% respectively.⁶² The much lower bioavailability from green-leafy vegetables than from pure phylloquinone supplements is in agreement with direct absorption studies that measured the areas under the plasma curves following oral doses.^[83], ^[84] It needs to be emphasised that the average response to a given dietary intake of vitamin K in the studies described above encompasses a wide individual variability that is the inevitable consequence of a number of individual and interacting factors that include the type of food matrix, the efficiency of intestinal digestion and absorption, aspects of intermediary metabolism and differences in the pharmacodynamic response of the enzymes of the vitamin K-epoxide cycle.

Evidence that low dietary intakes of vitamin K (and hence poor vitamin K status) are associated with greater instability of the INR was obtained in a major UK study⁶⁴ and supported by a South Korean study⁶⁷ as well as a case–control study from The Netherlands that specifically linked low usual intakes to an increased risk of sub-therapeutic anticoagulant therapy.⁶⁵ The finding in the Dutch study⁶⁵ that low intakes of vitamin K increase the risk of developing a low INR is somewhat counterintuitive but could be explained by the hypothesis that the impact of occasional intakes of vitamin K-rich foods is proportionally higher in patients with low tissues stores than in individuals who have replete tissue stores that act as a buffer to dietary variation.⁶⁶ This concept is supported by the findings that patients with an impaired vitamin K status indicated by low circulating phylloquinone are more likely to display a hypersensitive response to the initiation or cessation of small daily supplements of 25μg phylloquinone.⁷⁶ The corollary of the concept that poor vitamin K status equates to greater instability of VKA therapy is the concept that stability might be improved by increasing the intake of vitamin K-rich foods⁶⁶ or by regular supplementation with appropriate amounts of phylloquinone. Dose-escalation studies carried out in healthy volunteers^[62], ^[85] have helped to define what daily doses of vitamin K may be safely taken during VKA therapy. Although there is now some evidence that daily supplementation with phylloquinone does indeed improve stability^[88], ^[89] further trials are needed to show the efficacy and safety of supplementation strategies. However, the trials to date demonstrate that that there is no a priori reason why, if properly explained to the patient, taking an extra daily phylloquinone supplement of 100–150μg should not be compatible with VKA therapy. Consideration needs to be given to patients with an initially poor vitamin K status who may be more responsive to low doses of vitamin K supplements. Also once a patient has started taking a vitamin K supplement, the main danger to stability would derive from erratic compliance or worst, sudden cessation of the supplement which would present a danger of over-anticoagulation and possible bleeding.

Multi-vitamin supplements containing vitamin K, usually as phylloquinone, are readily available to the general public from conventional and internet shopping sources. However, pure supplements of phylloquinone or MK-7 (often with vitamin D and calcium) intended for bone health are also available and this may be soon followed by functional foods containing vitamin K. The consequences to patients on VKA taking MK-7 supplements that are marketed for bone health are incompletely understood but based on the available evidence a daily supplement containing as little as 25μg of MK-7 may interfere with VKA stability. The majority of healthcare professionals in anticoagulation clinics will already discuss the importance of dietary intakes of vitamin K with their patients and this would normally include discussion of any vitamin supplements that may or may not include vitamin K. In view of the increasing availability and far greater antidotal potency of MK-7, healthcare professionals and patients need to be aware that the effects of MK-7 are likely to more problematic than phylloquinone.

Back to Article Outline

11. Practise points 

Back to Article Outline

12. Research agenda 

Back to Article Outline

Conflict of interest statement 

No conflicts of interest to declare.

Back to Article Outline

Acknowledgements 

MVH is supported by a Medical Research Council Population Health Scientist Fellowship (G0802432).

Back to Article Outline

References 

1. Link KP. The discovery of dicumarol and its sequels. Circulation. 1959;19:97–107
   • View In Article
   • MEDLINE
2. Suttie JW. Vitamin K in health and disease. First ed.. Boca Raton: CRC Press (Taylor & Francis Group; 2009;
   • View In Article
3. Wadelius M, Chen LY, Lindh JD, Eriksson N, Ghori MJ, Bumpstead S, et al. The largest prospective warfarin-treated cohort supports genetic forecasting. Blood. 2009;113:784–792
   • View In Article
   • CrossRef
4. Rettie AE, Wienkers LC, Gonzalez FJ, Trager WF, Korzekwa KR. Impaired (S)-warfarin metabolism catalysed by the R144C allelic variant of CYP2C9. Pharmacogenetics. 1994;4:39–42
   • View In Article
   • CrossRef
5. Kamali F, Wynne H. Pharmacogenetics of warfarin. Annu Rev Med. 2010;61:63–75
   • View In Article
   • CrossRef
6. Caldwell MD, Awad T, Johnson JA, Gage BF, Falkowski M, Gardina P, et al. CYP4F2 genetic variant alters required warfarin dose. Blood. 2008;111:4106–4112
   • View In Article
   • CrossRef
7. McDonald MG, Rieder MJ, Nakano M, Hsia CK, Rettie AE. CYP4F2 is a vitamin K1 oxidase: an explanation for altered warfarin dose in carriers of the V433M variant. Mol Pharmacol. 2009;75:1337–1346
   • View In Article
   • CrossRef
8. Shearer MJ, McBurney A, Barkhan P. Studies on the absorption and metabolism of phylloquinone (vitamin K1) in man. Vitam Horm. 1974;32:513–542
   • View In Article
   • MEDLINE
   • CrossRef
9. Harrington DJ, Soper R, Edwards C, Savidge GF, Hodges SJ, Shearer MJ. Determination of the urinary aglycone metabolites of vitamin K by HPLC with redox-mode electrochemical detection. J Lipid Res. 2005;46:1053–1060
   • View In Article
   • MEDLINE
   • CrossRef
10. Nakagawa K, Hirota Y, Sawada N, Yuge N, Watanabe M, Uchino Y, et al. Identification of UBIAD1 as a novel human menaquinone-4 biosynthetic enzyme. Nature. 2010;468:117–121
    • View In Article
    • CrossRef
11. World Health Organization International Agency for Research on Cancer . IARC monographs on the evaluation of carcinogenic risks to humans. In: Some antiviral and antineoplastic drugs, and other pharmaceutical agents. 76:Lyon: IARC Press; 2000;p. 417–486
    • View In Article
12. Hanson HH, Barker NW, Mann FD. Clinical experiences with 4-hydroxycoumarin anticoagulant no. 63 and the antagonistic effect of menadione and vitamin K. Circulation. 1951;4:844–863
    • View In Article
    • MEDLINE
13. Shearer MJ, Newman P. Metabolism and cell biology of vitamin K. Thromb Haemost. 2008;100:530–547
    • View In Article
14. Furie B, Bouchard BA, Furie BC. Vitamin K-dependent biosynthesis of gamma-carboxyglutamic acid. Blood. 1999;93:1798–1808
    • View In Article
    • MEDLINE
15. Wu S, Liu S, Davis CH, Stafford DW, Kulman JD, Pedersen LG. A hetero-dimer model for concerted action of vitamin K carboxylase and vitamin K reductase in vitamin K cycle. J Theor Biol. 2011;279:143–149
    • View In Article
    • CrossRef
16. Park BK. Warfarin: metabolism and mode of action. Biochem Pharmacol. 1988;37:19–27
    • View In Article
    • MEDLINE
    • CrossRef
17. Price R, Fenton S, Shearer MJ, Bolton-Smith C. Daily and seasonal variation in phylloquinone (vitamin K1) intake in Scotland. Proc Nutr Soc. 1996;55:244A
    • View In Article
18. Harrington DJ, Booth SL, Card DJ, Shearer MJ. Excretion of the urinary 5C- and 7C-aglycone metabolites of vitamin K by young adults responds to changes in dietary phylloquinone and dihydrophylloquinone intakes. J Nutr. 2007;137:1763–1768
    • View In Article
    • MEDLINE
19. Olson RE, Chao J, Graham D, Bates MW, Lewis JH. Total body phylloquinone and its turnover in human subjects at two levels of vitamin K intake. Br J Nutr. 2002;87:543–553
    • View In Article
    • MEDLINE
    • CrossRef
20. Shea MK, Booth SL, Gundberg CM, Peterson JW, Waddell C, Dawson-Hughes B, et al. Adulthood obesity is positively associated with adipose tissue concentrations of vitamin K and inversely associated with circulating indicators of vitamin K status in men and women. J Nutr. 2010;140:1029–1034
    • View In Article
    • CrossRef
21. Shearer MJ, Harrington DJ, Bolton-Smith C, Clark KGA, Savidge GF. A very low, fixed dose of warfarin has a minimal effect on the gamma-carboxylation of the vitamin K dependent coagulation factors but perturbs vitamin K metabolism. J Thromb Haemost. 1997;(Supplement:708):(Abstract)
    • View In Article
22. McCann JC, Ames BN. Vitamin K, an example of triage theory: is micronutrient inadequacy linked to diseases of aging?. Am J Clin Nutr. 2009;90:889–907
    • View In Article
    • CrossRef
23. Shearer MJ, Bach A, Kohlmeier M. Chemistry, nutritional sources, tissue distribution and metabolism of vitamin K with special reference to bone health. J Nutr. 1996;126:1181S–1186S
    • View In Article
    • MEDLINE
24. Schurgers LJ, Vermeer C. Determination of phylloquinone and menaquinones in food. Effect of food matrix on circulating vitamin K concentrations. Haemostasis. 2000;30:298–307
    • View In Article
    • MEDLINE
25. Conly JM, Stein K. Quantitative and qualitative measurements of K vitamins in human intestinal contents. Am J Gastroenterol. 1992;87:311–316
    • View In Article
    • MEDLINE
26. Suttie JW. The importance of menaquinones in human nutrition. Annu Rev Nutr. 1995;15:399–417
    • View In Article
    • MEDLINE
    • CrossRef
27. Shearer MJ, Vitamin K. Metabolism and nutriture. Blood Rev. 1992;6:92–104
    • View In Article
    • CrossRef
28. Booth SL, Al Rajabi A. Determinants of vitamin K status in humans. Vitam Horm. 2008;78:1–22
    • View In Article
    • CrossRef
29. Parrish DB. Determination of vitamin K in foods: a review. Crit Rev Food Sci Nutr. 1980;13:337–352
    • View In Article
    • MEDLINE
30. Shearer MJ, Bolton-Smith C. The UK food data-base for vitamin K and why we need it. Food Chem. 2000;68:213–218
    • View In Article
31. Booth SL, Sadowski JA, Weihrauch JL, Ferland G. Vitamin K1 (Phylloquinone) content of foods: a provisional table. J Food Compos Anal. 1993;6:109–120
    • View In Article
32. Booth SL, Sadowski JA, Pennington JAT. Phylloquinone (Vitamin K1) content of foods in the U.S. food and drug administration's total diet study. J Agric Food Chem. 1995;43:1574–1579
    • View In Article
    • CrossRef
33. Bolton-Smith C, Price RJ, Fenton ST, Harrington DJ, Shearer MJ. Compilation of a provisional UK database for the phylloquinone (vitamin K1) content of foods. Br J Nutr. 2000;83:389–399
    • View In Article
    • MEDLINE
34. McCance , Widdowson's . The Composition of foods. 6th Revised ed.. United Kingdom: Royal Society of Chemistry; 2002;
    • View In Article
35. Agricultural Research Service . USDA National Nutrient Database for standard reference. In: Release 17 vitamin K (phylloquinone) (μg) content of selected foods per common measure, sorted alphabetically. US Department of Agriculture; 2004;Available at http://www.nal.usda.gov/fnic/foodcomp/Data/SR17/wtrank/sr17a430.pdf
    • View In Article
36. Ferland G, Sadowski JA. Vitamin K1 (phylloquinone) content of green vegetables: effects of plant maturation and geographical growth location. J Agric Food Chem. 1992;40:1874–1877
    • View In Article
    • CrossRef
37. Ferland G, Sadowski JA. Vitamin K1 (phylloquinone) content of edible oils: effects of heating and light exposure. J Agric Food Chem. 1992;40:1869–1873
    • View In Article
    • CrossRef
38. Matschiner JT, Amelotti JM. Characterization of vitamin K from bovine liver. J Lipid Res. 1968;9:176–179
    • View In Article
    • MEDLINE
39. Duello TJ, Matschiner JT. Characterization of vitamin K from pig liver and dog liver. Arch Biochem Biophys. 1971;144:330–338
    • View In Article
    • MEDLINE
    • CrossRef
40. Duello TJ, Matschiner JT. Characterization of vitamin K from human liver. J Nutr. 1972;102:331–335
    • View In Article
    • MEDLINE
41. Nelson M. Methods and validity of dietary assessment. In:  Garrow JS,  James WPT,  Ralph A editor. Human nutrition and dietetics. 10th ed.. Edinburgh: Churchill Livingstone; 1999;
    • View In Article
42. Bingham SA, Gill C, Welch A, Day K, Cassidy A, Khaw KT, et al. Comparison of dietary assessment methods in nutritional epidemiology: weighed records v. 24h recalls, food-frequency questionnaires and estimated-diet records. Br J Nutr. 1994;72:619–643
    • View In Article
    • MEDLINE
    • CrossRef
43. Schurgers LJ, Geleijnse JM, Grobbee DE, Pols HAP, Hofman A, Witteman JCM, et al. Nutritional intake of vitamins K1 (Phylloquinone) and K2 (Menaquinone) in The Netherlands. J Nutr Environ Med. 1999;9:115–122
    • View In Article
44. Geleijnse JM, Vermeer C, Grobbee DE, Schurgers LJ, Knapen MH, van der Meer IM, et al. Dietary intake of menaquinone is associated with a reduced risk of coronary heart disease: the Rotterdam Study. J Nutr. 2004;134:3100–3105
    • View In Article
    • MEDLINE
45. Nimptsch K, Rohrmann S, Linseisen J. Dietary intake of vitamin K and risk of prostate cancer in the Heidelberg cohort of the European Prospective Investigation into Cancer and Nutrition (EPIC-Heidelberg). Am J Clin Nutr. 2008;87:985–992
    • View In Article
46. Booth SL, Sokoll LJ, O'Brien ME, Tucker K, Dawson-Hughes B, Sadowski JA. Assessment of dietary phylloquinone intake and vitamin K status in postmenopausal women. Eur J Clin Nutr. 1995;49:832–841
    • View In Article
    • MEDLINE
47. Booth SL, Suttie JW. Dietary intake and adequacy of vitamin K. J Nutr. 1998;128:785–788
    • View In Article
    • MEDLINE
48. Thane CW, Paul AA, Bates CJ, Bolton-Smith C, Prentice A, Shearer MJ. Intake and sources of phylloquinone (vitamin K1): variation with socio-demographic and lifestyle factors in a national sample of British elderly people. Br J Nutr. 2002;87:605–613
    • View In Article
    • MEDLINE
    • CrossRef
49. Thane CW, Bolton-Smith C, Coward WA. Comparative dietary intake and sources of phylloquinone (vitamin K1) among British adults in 1986–7 and 2000-1. Br J Nutr. 2006;96:1105–1115
    • View In Article
    • MEDLINE
    • CrossRef
50. Cushman M, Booth SL, Possidente CJ, Davidson KW, Sadowski JA, Bovill EG. The association of vitamin K status with warfarin sensitivity at the onset of treatment. Br J Haematol. 2001;112:572–577
    • View In Article
    • MEDLINE
    • CrossRef
51. Khan T, Wynne H, Wood P, Torrance A, Hankey C, Avery P, et al. Dietary vitamin K influences intra-individual variability in anticoagulant response to warfarin. Br J Haematol. 2004;124:348–354
    • View In Article
    • MEDLINE
    • CrossRef
52. Sobczyńska-Malefora A, Harrington DJ, Lomer MC, Pettitt C, Hamilton S, Rangarajan S, et al. Erythrocyte folate and 5-methyltetrahydrofolate levels decline during 6months of oral anticoagulation with warfarin. Blood Coagul Fibrinolysis. 2009;20:297–302
    • View In Article
    • CrossRef
53. Food and Nutrition Board, Institute of Medicine . Dietary reference intakes for vitamin A, vitamin K, arsenic, boron, chromium, copper, iodine, manganese, molybdenum, nickel, silicon, vanadium and zinc. Washington, DC: National Academy Press; 2001;
    • View In Article
54. Johnson MA. Influence of vitamin K on anticoagulant therapy depends on vitamin K status and the source and chemical forms of vitamin K. Nutr Rev. 2005;63:91–97
    • View In Article
    • MEDLINE
    • CrossRef
55. Spratto GR, Woods AL. PDR nurse's drug handbook. Revised ed: Thomson PDR; 2002;
    • View In Article
56. Wells PS, Holbrook AM, Crowther NR, Hirsh J. Interactions of warfarin with drugs and food. Ann Intern Med. 1994;121:676–683
    • View In Article
    • MEDLINE
57. Harris JE. Interaction of dietary factors with oral anticoagulants: review and applications. J Am Diet Assoc. 1995;95:580–584
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (611 KB)
    • CrossRef
58. Lubetsky A, Dekel-Stern E, Chetrit A, Lubin F, Halkin H. Vitamin K intake and sensitivity to warfarin in patients consuming regular diets. Thromb Haemost. 1999;81:396–399
    • View In Article
    • MEDLINE
59. Food and Nutrition Board . Recommended dietary allowances (dietary reference intakes). 10th Revised ed.. Washington, DC: National Academies Press; 1989;
    • View In Article
60. Suttie JW. Vitamin K and human nutrition. J Am Diet Assoc. 1992;92:585–590
    • View In Article
    • MEDLINE
61. Lee W, Chung HJ, Kim S, Jang S, Park CJ, Chi HS, et al. PIVKA-II is a candidate marker for monitoring the effects of the oral anticoagulant warfarin. Clin Biochem. 2010;43:1177–1179
    • View In Article
    • CrossRef
62. Schurgers LJ, Shearer MJ, Hamulyák K, Stöcklin E, Vermeer C. Effect of vitamin K intake on the stability of oral anticoagulant treatment: dose-response relationships in healthy subjects. Blood. 2004;104:2682–2689
    • View In Article
    • MEDLINE
    • CrossRef
63. Franco V, Polanczyk CA, Clausell N, Rohde LE. Role of dietary vitamin K intake in chronic oral anticoagulation: prospective evidence from observational and randomized protocols. Am J Med. 2004;116:651–656
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (122 KB)
    • CrossRef
64. Sconce E, Khan T, Mason J, Noble F, Wynne H, Kamali F. Patients with unstable control have a poorer dietary intake of vitamin K compared to patients with stable control of anticoagulation. Thromb Haemost. 2005;93:872–875
    • View In Article
    • MEDLINE
65. Rombouts EK, Rosendaal FR, van der Meer FJ. Influence of dietary vitamin K intake on subtherapeutic oral anticoagulant therapy. Br J Haematol. 2010;149:598–605
    • View In Article
    • CrossRef
66. Presse N, Kergoat MJ, Ferland G. High usual dietary vitamin K intake is associated with low relative variability in vitamin K intake: implications for anticoagulant therapy. Br J Haematol. 2011;153:129–130
    • View In Article
    • CrossRef
67. Kim KH, Choi WS, Lee JH, Lee H, Yang DH, Chae SC. Relationship between dietary vitamin K intake and the stability of anticoagulation effect in patients taking long-term warfarin. Thromb Haemost. 2010;104:755–759
    • View In Article
    • CrossRef
68. Usui Y, Tanimura H, Nishimura N, Kobayashi N, Okanoue T, Ozawa K. Vitamin K concentrations in the plasma and liver of surgical patients. Am J Clin Nutr. 1990;51:846–852
    • View In Article
    • MEDLINE
69. Caldwell PT, Ren P, Bell RG. Warfarin and metabolism of vitamin K1. Biochem Pharmacol. 1974;23:3353–3362
    • View In Article
    • MEDLINE
    • CrossRef
70. Shearer MJ, McBurney A, Barkhan P. Effect of warfarin anticoagulation on vitamin-K 1 metabolism in man. Br J Haematol. 1973;24:471–479
    • View In Article
    • MEDLINE
    • CrossRef
71. Shearer MJ, McBurney A, Breckenridge AM, Barkhan P. Effect of warfarin on the metabolism of phylloquinone (vitamin K1):dose-response relationships in man. Clin Sci Mol Med. 1977;52:621–630
    • View In Article
    • MEDLINE
72. Choonara IA, Scott AK, Haynes BP, Cholerton S, Breckenridge AM, Park BK. Vitamin K1 metabolism in relation to pharmacodynamic response in anticoagulated patients. Br J Clin Pharmacol. 1985;20:643–648
    • View In Article
    • MEDLINE
73. Park BK, Choonara IA, Haynes BP, Breckenridge AM, Malia RG, Preston FE. Abnormal vitamin K metabolism in the presence of normal clotting factor activity in factory workers exposed to 4-hydroxycoumarins. Br J Clin Pharmacol. 1986;21:289–293
    • View In Article
    • MEDLINE
74. Kamali F, Edwards C, Butler TJ, Wynne HA. The influence of (R)- and (S)-warfarin, vitamin K and vitamin K epoxide upon warfarin anticoagulation. Thromb Haemost. 2000;84:39–42
    • View In Article
    • MEDLINE
75. Kurnik D, Lubetsky A, Loebstein R, Almog S, Halkin H. Multivitamin supplements may affect warfarin anticoagulation in susceptible patients. Ann Pharmacother. 2003;37:1603–1606
    • View In Article
    • MEDLINE
    • CrossRef
76. Kurnik D, Loebstein R, Rabinovitz H, Austerweil N, Halkin H, Almog S. Over-the-counter vitamin K1-containing multivitamin supplements disrupt warfarin anticoagulation in vitamin K1-depleted patients. A prospective, controlled trial. Thromb Haemost. 2004;92:1018–1024
    • View In Article
    • MEDLINE
77. Owren PA. Thrombotest. A new method for controlling anticoagulant therapy. Lancet. 1959;2:754–758
    • View In Article
    • MEDLINE
78. Owren PA. The interrelationship between Normotest and Thrombotest. Farmakoterapi. 1969;25:1–13
    • View In Article
79. Multicentre evaluation of Normotest performance characteristics. Interlab variation, comparison of manual and instrumental techniques, reproducibility, normal range. Italian CISMEL study group. Scand J Haematol. 1984;32:501–506
    • View In Article
    • CrossRef
80. Arnesen H, Smith P. The predictability of bleeding by prothrombin times sensitive or insensitive to PIVKA during intensive oral anticoagulation. Thromb Res. 1991;61:311–314
    • View In Article
    • Abstract
    • Full-Text PDF (260 KB)
    • CrossRef
81. Karlson B, Leijd B, Hellstrom K. On the influence of vitamin K-rich vegetables and wine on the effectiveness of warfarin treatment. Acta Med Scand. 1986;220:347–350
    • View In Article
    • MEDLINE
    • CrossRef
82. Pedersen FM, Hamberg O, Hess K, Ovesen L. The effect of dietary vitamin K on warfarin-induced anticoagulation. J Intern Med. 1991;229:517–520
    • View In Article
    • MEDLINE
    • CrossRef
83. Gijsbers BL, Jie KS, Vermeer C. Effect of food composition on vitamin K absorption in human volunteers. Br J Nutr. 1996;76:223–229
    • View In Article
    • MEDLINE
    • CrossRef
84. Garber AK, Binkley NC, Krueger DC, Suttie JW. Comparison of phylloquinone bioavailability from food sources or a supplement in human subjects. J Nutr. 1999;129:1201–1203
    • View In Article
    • MEDLINE
85. Schurgers LJ, Teunissen KJ, Hamulyák K, Knapen MH, Vik H, Vermeer C. Vitamin K-containing dietary supplements: comparison of synthetic vitamin K1 and natto-derived menaquinone-7. Blood. 2007;109:3279–3283
    • View In Article
    • MEDLINE
    • CrossRef
86. Vermeer C, Shearer MJ, Zittermann A, Bolton-Smith C, Szulc P, Hodges S, et al. Beyond deficiency: potential benefits of increased intakes of vitamin K for bone and vascular health. Eur J Nutr. 2004;43:325–335
    • View In Article
    • MEDLINE
    • CrossRef
87. Kudo T. Warfarin antagonism of natto and increase in serum vitamin K by intake of natto. Artery. 1990;17:189–201
    • View In Article
    • MEDLINE
88. Sconce E, Avery P, Wynne H, Kamali F. Vitamin K supplementation can improve stability of anticoagulation for patients with unexplained variability in response to warfarin. Blood. 2007;109:2419–2423
    • View In Article
    • MEDLINE
    • CrossRef
89. Rombouts EK, Rosendaal FR, Van Der Meer FJ. Daily vitamin K supplementation improves anticoagulant stability. J Thromb Haemost. 2007;5:2043–2048
    • View In Article
    • CrossRef
90. Bonjour JP, Gueguen L, Palacios C, Shearer MJ, Weaver CM. Minerals and vitamins in bone health: the potential value of dietary enhancement. Br J Nutr. 2009;101:1581–1596
    • View In Article
    • CrossRef
91. Schurgers LJ, Cranenburg EC, Vermeer C. Matrix Gla-protein: the calcification inhibitor in need of vitamin K. Thromb Haemost. 2008;100:593–603
    • View In Article
92. Cockayne S, Adamson J, Lanham-New S, Shearer MJ, Gilbody S, Torgerson DJ. Vitamin K and the prevention of fractures: systematic review and meta-analysis of randomized controlled trials. Arch Intern Med. 2006;166:1256–1261
    • View In Article
    • MEDLINE
    • CrossRef
93. Conway SP. Vitamin K in cystic fibrosis. J R Soc Med. 2004;97(Suppl 44):48–51
    • View In Article
    • MEDLINE
    • CrossRef
94. EFSA panel on dietetic products Nutrition and Allergies (NDA). In: Scientific opinion on the substantiation of health claims related to vitamin K and maintenance of bone (ID 123, 127, 128 and 2879), blood coagulation (ID 124 and 126) and function of the heart and blood vessels (ID 124, 125 and 2880) pursuant to Article 13(1) of Regulation (EC) No 1924/2006 on request from the European Commission. 7:EFSA; 2009;p. 1228;Available online www.efsa.europa.eu
    • View In Article
95. de Assis MC, Rabelo ER, Avila CW, Polanczyk CA, Rohde LE. Improved oral anticoagulation after a dietary vitamin k-guided strategy: a randomized controlled trial. Circulation. 2009;120:1115–11223 p following 22
    • View In Article
    • CrossRef
96. Hylek EM, Heiman H, Skates SJ, Sheehan MA, Singer DE. Acetaminophen and other risk factors for excessive warfarin anticoagulation. Jama. 1998;279:657–662
    • View In Article
    • MEDLINE
    • CrossRef
97. Jones KS, Bluck LJ, Wang LY, Stephen AM, Prynne CJ, Coward WA. The effect of different meals on the absorption of stable isotope-labelled phylloquinone. Br J Nutr. 2009;102:1195–1202
    • View In Article
    • CrossRef
98. Custodio das Dores SM, Booth SL, Martini LA, de Carvalho Gouvea VH, Padovani CR, de Abreu Maffei FH, et al. Relationship between diet and anticoagulant response to warfarin: a factor analysis. Eur J Nutr. 2007;46:147–154
    • View In Article
    • MEDLINE
    • CrossRef