Vitamin K and multiple diseases: a comparison between vitamin K1 and vitamin K2

ABSTRACT

Background: Vitamin K is a fat-soluble vitamin that may play roles in the prevention of many diseases. Vitamin K is present in the diet in two biologically active forms, vitamin K1 (phylloquinone) and vitamin K2 (menaquinone), that behave differently with regard to absorption, transport and tissue distribution. The main aim of this review was to investigate whether K1 and K2 are differently affecting health and disease processes.

Methods: A systematic literature search was conducted in PubMed from inception to March 7th 2017. Humans studies were included if they examined the association between both K1 and K2 levels (measured by dietary intake, biomarkers) and clinical outcomes such as cardiovascular disease (CVD), bone disease, bone fractures, diabetes or cancer.

Results: Sixteen observational studies met the inclusion criteria. Three studies reported an inverse association between K2 intake and CVD risk. Both K1 and K2 were inversely associated with risk of bone fractures or osteopenia in four studies. In one study, K2 had a linear, inverse association with the risk of type 2 diabetes.

Conclusions: The results show that vitamin K2 may be more effective compared with K1 in the prevention of CVD, but more prospective studies are needed for definitive conclusions.

Abbreviations: BMD, bone mineral density; CHD, coronary heart disease; CVD, cardiovascular disease; FFQ, food frequency questionnaire; GGCX, γ-glutamyl carboxylase; MK, menaquinone; PAD, peripheral arterial disease; VKD, vitamin K-dependent

Keywords: Vitamin K, vitamin K1, vitamin K2, phylloquinone, menaquinone, cardiovascular disease, bone disease, bone fractures

BACKGROUND

The term vitamin K is a general term for 2-methyl-1,4-naphtoquinone (menadione) and its derivatives that exhibit an anti hemorrhagic action in vitamin K-deficient animals (1). Two biologically active forms, vitamin K1 (phylloquinone) and vitamin K2 (menaquinone), are present in the human diet. Structurally, the two forms differ with regard to the length and saturation of the side chain at the 3-position of the backbone ring structure. Phylloquinone has a mostly saturated phytyl side chain, while menaquinone has an unsaturated side chain of repeating prenyl units. Homologues of menaquinone (MK) are designated as MK-n, with n denoting the number of unsaturated prenyl side chains. Some MKs do however contain saturated prenyl units (2).

Phylloquinone is synthesized in plants and cyanobacteria (3). The main dietary sources are dark-green vegetables and vegetable oils. Menaquinones are synthesized by obligate and facultative anaerobic bacteria, with one exception in the form of MK-4 which can be synthesized in animal tissues. The main food sources of menaquinones are meat (MK-4) and fermented foods like natto (MK-7) and cheese (mainly MK-8 and MK-9). The major dietary source of vitamin K in the Western diet is phylloquinone (3).

The most established role of vitamin K is as an essential cofactor for the enzyme γ-glutamyl carboxylase (GGCX), responsible for post-translational modification of vitamin K-dependent (VKD) proteins. Some of these proteins are blood clotting factors of hepatic origin, while others, like osteocalcin and matrix Gla protein, are of extrahepatic origin with different functions (1). Osteocalcin, the second most abundant protein in bone, is involved in insulin production and energy expenditure (4,5). Matrix Gla protein is expressed in almost all soft tissues, and regulates vascular calcification (6,7). Though the functions of many VKD proteins remain uncertain, vitamin K is suggested to play a protective role in cardiovascular disease (CVD), bone health, immune function, energy metabolism and cancer development (3).

K1 and K2 are incorporated into nascent chylomicrons following their intestinal absorption (1). The liver appears to be the main site for uptake and accumulation of K1 whereas the extra-hepatic tissues preferentially accumulate K2 (2,8,9). The differential tissue accretion can partly be explained by how vitamin K is incorporated into lipoproteins (8). Since the lipophilic character (hydrophobicity) increases with increasing saturated chain length, menaquinones like MK-9 is expected to be present in the core of lipoprotein particles, while K1 resides further away from the lipoprotein core and the more water-soluble MK-4 with its unsaturated side chain even further away, close to the edges of the particles. As lipoprotein lipase hydrolyzes the edges of nascent chylomicrons, most of the MK-4 is delivered to extrahepatic tissues during its circulation. 

The longer menaquinones and K1 remain in the chylomicron remnants, and subsequently get internalized along with the remnants mainly by the liver and to a lesser extent by the bone tissue (3,10). In the liver, a sizable fraction of K1 is catabolized while longer menaquinones like MK-7 to MK-9 are either stored or packaged into VLDLs (along with some K1 and shorter menaquinones) (3). During circulation, these particles eventually become LDLs which can survive for several days (8). While every form of vitamin K has the ability to function as a cofactor for GGCX, based on the lipoprotein transport pathways outlined above and the differential tissue accretion, different forms of vitamin K may have unique abilities to support different aspects of health.

The Nordic Nutrition Recommendations 2012 recommends 1 µg/kg body weight per day for total vitamin K intake for both children and adults (11). In the United States the recommendation is based on median consumption levels of healthy individuals and is set as an Adequate Intake (AI) of 120 µg for men and 90 µg for women (12). However, there are currently no official recommendations on K2. Therefore, this review investigated whether K1 and K2 are differently associated with clinical outcomes related to CVD, bone disease, bone fractures, diabetes and cancer, in order to assess whether specific recommendations with regard to K2 intake should be considered.

METHODS

A systematic literature search was conducted in PubMed from inception to March 7th 2017 with the search term “vitamin K1 AND vitamin K2 AND (cardiovascular disease OR bone disease OR fractures OR diabetes OR cancer)”. The search strategy is shown in Figure 1. The inclusion criteria were measurement of both K1 and K2 intake among participants via either Food Frequency Questionnaire (FFQ) or biomarkers (e.g. serum vitamin K levels), or comparison of K1 with K2 supplementation. Additional criteria were measurement of clinical outcomes related to diagnosis of either CVD, bone disease (including osteopenia), bone fractures, diabetes or cancer. Cohort, case-control and cross-sectional studies, in addition to randomized controlled trials, were included. Exclusion criteria were adults with renal or liver disease.

Figure 1. Flow diagram showing results of the search strategy.

RESULTS

Searching the PubMed data base yielded 119 articles, of which 103 were excluded. Sixteen studies met the inclusion criteria, and the details of these studies are shown in Supplementary Table 1, available online at www.ntfe.no. The study design, exposure, exposure measurement and outcome for each of the 16 studies is summarized in Table 1 (13–28). A description of quantiles and multivariate models used in the different studies is presented in Supplementary Table 2 online.

Table 1 Study design, exposure, exposure measurement and outcomes for the studies within this review.

Study

Study design

Exposure

Exposure measurement

Outcomes

Vitamin K1

Vitamin K2a

K2 subtypesb

FFQ

Serum levels

CVD

Bone relatedc

Diabetes

Cancer

Apalset et al. (13)

Cohort

Yes

Yes

Yes

Yes

Beulens et al. (14)

Cohort

Yes

Yes [4-10]

Yes

Gast et al. (15)

Cohort

Yes

Yes [4-9]

Yes

Yes

Geleijnse et al. (16)

Cohort

Yes

Yes [4-10]

Yes

Yes

Hodges et al. (17)

Case-control

Yes

Yes [7,8]

Yes

Yes

Hodges et al. (18)

Case-control

Yes

Yes [7,8]

Yes

Yes

Iwasaki et al. (19)

Cross-sectional

Yes

Yes [4]

Yes

Yes

Juanola-Falgarona et al. (20)

Cohort

Yes

Yes

Yes

Yes

Yes

Kawana et al. (21)

Case-control

Yes

Yes [4,7]

Yes

Yes

Nakano et al. (22)

Case-control

Yes

Yes [7]

Yes

Yes

Nimptsch et al. (23)

Cohort

Yes

Yes [4-14]

Yes

Yes

Nimptsch et al. (24)

Cohort

Yes

Yes [4-14]

Yes

Yes

Tamatani et al. (25)

Cross-sectional

Yes

Yes [7]

Yes

Yes

Tsugawa et al. (26)

Cohort

Yes

Yes [4,7]

Yes

Yes

Vissers et al. (27)

Cohort

Yes

Yes [4-10]

Yes

Yes

Vissers et al. (28)

Cohort

Yes

Yes [4-10]

Yes

Yes

a Total vitamin K2 is measured. Subtypes of menaquinones assayed, when mentioned, are presented in brackets.

b Subtypes of menaquinones assayed. Total vitamin K2 not measured.

c All of the outcomes were related to bone fractures, with the exception of Tamatani et al. who measured bone density in elderly men and Iwasaki et al who measured bone density in young women.

Table 2 Dietary intake of vitamin K1 and the risk of CVD (multivariate models controlling for cardiovascular risk factors and dietary variables).

Study

Study size

CVD

eventsa

Effect size by vitamin K1 intake category (95% CI)

P-value

Effect size per 10/50 μg increase vitamin K1 (95% CI)b

1 (low)

2

3

4 (high)

Gast et al. (15)

16057

480 [CHD]

1.00 (1.00, 1.02) p = 0.35

Geleijnse et al. (16)

4807

233 [CHD]

1

1.00 (0.73, 1.37)

0.89 (0.63, 1.25)

0.48

Juanola-Falgarona et al. (20)

7216

81 [mortality]

1

0.89 (0.49, 1.64)

1.04 (0.56, 1.92)

0.63 (0.31, 1.28)

0.20

Vissers et al. (27)

35476

580 [stroke]

1

0.94 (0.74, 1.21)

1.00 (0.78, 1.28)

1.09 (0.85, 1.40)

0.41

1.01 (0.97, 1.06)

Vissers et al. (28)

36629

489 [PAD]

1

0.81 (0.62, 1.08)

1.07 (0.82, 1.38)

1.18 (0.91, 1.53)

0.07

1.04 (0.97, 1.09) p = 0.08

a Relevant outcomes are presented in brackets.

b Effect size was estimated per 10 µg and per 50 µg increase in vitamin K1 in the models of Gast et al. and Vissers et al., respectively.

Eight cohort studies examined the association between the intakes of both K1 and K2 assessed by a Food Frequency Questionnaire (FFQ) and either CVD (15,16,20,27,28), type 2 diabetes (14) or cancer (20,23,24). One cohort study investigated the association between K1 and K2 intake and hip fracture incidence (13), another the association between serum levels of K1 and K2 and vertebral fracture incidence (26). Four case-control studies examined the circulating levels of K1 and K2 in bone fracture patients in comparison to those in healthy controls (17,18,21,22). Two cross-sectional studies examined the association between circulating levels of K1 and K2 and osteopenia (19,25).

Vitamin K and CVD

Dietary K1 intake and risk of CVD

Five cohort studies examined the association between K1 intake (measured by FFQ) and the incidence of CVD (15,16,20,27,28). The results of the multivariate models for each individual study are presented in Table 2. Two studies looked at the incidence of CHD (15,16), one study looked at cardiovascular mortality risk (20), one study looked at the occurrence of stroke (27) and one study looked at the incidence of peripheral artery disease (PAD) (28). The study population in four of these studies were from either the Prospect-EPIC cohort or the EPIC-NL, a Dutch cohort merged from the Prospect-EPIC and MORGEN-EPIC cohorts. Participants in the fifth study were from the PREDIMED cohort. The effect of the amount of K1 intake was expressed in tertiles, quartiles and as a continuous variable (per 10 µg increase in K1 in one study and per 50 µg increase in two studies). All five studies found no significant association between K1 intake and the incidence of CVD.

Dietary K2 intake and risk of CVD

Five cohort studies examined the association between K2 intake (measured by FFQ) and the incidence of CVD (15,16,20,27,28). The results of the multivariate models are presented in Table 3. Two studies looked at the incidence of CHD (15,16), one study looked at cardiovascular mortality risk (20), one study looked at the occurrence of stroke (27) and another study looked at the incidence of PAD (28). The effect size by K2 intake was expressed in tertiles, quartiles and as a continuous variable (per 10 µg increase in K2). Three of the studies found a significant inverse association between K2 intake and incidence of CVD. Gast et al. found a mean K2 intake of 29.1 (SD 12.8) µg/day and a 9% reduction in CHD risk with each 10 µg/day increase in K2 intake (p=0.04) (15). Geleijnse et. al found a mean K2 intake in men/women of 30.8/27.0 (SD 18.0/15.1) µg/day and a 41% reduction in incident CHD in the upper tertile (>32.7 µg/day) compared to the lower tertile (<21.6 µg/day) of K2 intake (ptrend=0.007) (16). K2 intake was also significantly associated with a reduced risk of CHD mortality (ptrend<0.05), while K1 intake was not (data not shown). Vissers et al. found a mean K2 intake of 30.6 (SD 13.7) µg/day and a 29% reduction in incident PAD in the upper quartile compared to the lowest quartile of K2 intake (HR=0.71; 95% CI 0.53-0.95) (28). A dose-response relationship with an 8% reduction in PAD risk per 10 µg/day increase in K2 intake was also found (p=0.03).

Table 3 Dietary intake of vitamin K2 and the risk of CVD (multivariate models controlling for cardiovascular risk factors and dietary variables).

Study

Study size

CVD

eventsa

Effect size by vitamin K2 intake category (95% CI)

P-value

Effect size per 10 μg

increase vitamin K2 (95% CI)

1 (low)

2

3

4 (high)

Gast et al. (15)

16057

480

[CHD]

0.91 (0.85, 1.00)

p = 0.04

Geleijnse et al. (16)

4807

233

[CHD]

1

0.96

(0.70, 1.31)

0.59

(0.40, 0.86)

0.007

Juanola-Falgarona et al. (20)

7216

81 [mortality]

1

1.04

(0.52, 2.06)

0.89

(0.44, 1.83)

1.18

(0.60, 2.34)

0.63

Vissers et al. (27)

35476

580

[stroke]

1

0.94

(0.74, 1.20)

0.86

(0.67, 1.12)

0.99

(0.75, 1.29)

0.82

0.99 (0.99, 1.06)

Vissers et al. (28)

36629

489

[PAD]

1

0.79

(0.61, 1.03)

0.90

(0.68, 1.16)

0.71

(0.53, 0.95)

0.06

0.92 (0.85, 0.99)

p = 0.03

a Relevant outcomes are presented in brackets.

Vitamin K and bone related outcomes

Dietary K1 and K2 intake and risk of hip fractures

Apalset et al. examined the association between dietary intake (measured by FFQ) of K1 and K2 and the incidence of hip fractures (13). The results of the multivariate model for this study is presented in Supplementary Table 3, available online at www.ntfe.no. The study found a significant association between K1 intake and hip fracture incidence, where participants in the lowest quartile had a 63% increased fracture risk compared to those in the highest quartile (ptrend=0.015). No significant association was found between K2 intake and hip fracture incidence.

Serum levels of K1 and risk of bone fractures or osteopenia

Four case-control studies, two cross-sectional studies and one cohort study examined the association between K1 intake (assessed by serum levels) and bone related outcomes (17–19,21,22,25,26). The results are presented in Table 4. Hodges et al. found significantly lower (p<0.01) serum levels of K1 in a group of patients with femoral neck fractures and in another group of patients with spinal crush fractures compared to a control group (17). Another study by Hodges et al. also found significantly lower (p<0.01) serum levels of K1 in patients with femoral neck fractures compared to controls (18). A third case-control study by Kawana et al. found no significant differences between a control group and two separate groups of patients with either vertebral fractures or hip fractures (21). Both male and female patients with hip fractures had significantly lower (p<0.05) serum levels of K1 compared to controls in a study by Nakano and colleagues (22). In a cross-sectional study by Iwasaki et al. no significant difference in serum levels of K1 were found between the low bone mineral density (BMD) group and normal BMD group. In a cross-sectional study by Tamatani et al. the subjects in the decreased BMD group had significantly lower (p<0.05) serum levels of K1 compared to the subjects in the normal BMD group (25). Tsugawa et al. found a significant inverse association between incidence of vertebral fractures and serum levels of K1 (p=0.007).

Table 4 Serum levels of vitamin K1 in relation to bone fractures and osteopenia.

Study

Study size

Fracture events

Circulating levels of vitamin K1 expressed in pg/ml as median (interquartile range) or mean ± 1 SD

Controls

Patients with femoral neck fractures

Patients with spinal crush fractures

Hodges et al. (17)

46

29

500 (398, 864)

93.5 (41.3, 145)**

129 (37, 85)**

Hodges et al. (18)

89

51

585 ± 490

336 ± 302**

Circulating levels of vitamin K1 expressed in ng/ml as mean ± 1 SD

Controls

Patients with vertebral fractures

Patients with hip fractures

Kawana et al. (21)

74

51

0.48 ± 0.7

0.46 ± 0.54

0.47 ± 0.34

Circulating levels of vitamin K1 expressed in ng/ml as mean ± 1 SD

Male controls

Male patients with hip fractures

Female controls

Female patients with hip fractures

Nakano et al. (22)

147

99

0.55 ± 0.31

0.31 ± 0.24*

0.77 ± 0.36

0.46 ± 0.36**

Circulating levels of vitamin K1 expressed in ng/ml (19) and nnmol/l (25) as mean ± 1 SD

Normal BMD†

Decreased BMD‡

Iwasaki et al. (19)

177

0.51 ± 0.45

0.37 ± 0.18

Tamatani et al. (25)

27

0.85 ± 0.73

0.60 ± 0.73*

Association between incidence of vertebral fractures and circulating levels of vitamin K1 (nmol/l)

β-coefficient

P-value

Tsugawa et al. (26)

379

35

-0.244

0.007

* p<0.05, relative to the control subjects or normal BMD group.

** p<0.01, relative to the control subjects.

† Defined as BMD of at least Young Adult Mean (YAM) - 1.0 SD in Iwasaki et al and at least peak bone density (PBD) - 2.5 SD in Tamatani et al.

‡ Defined as BMD less than YAM - 1.0 SD in Iwasaki et al and less than PBD - 2.5 SD in Tamatani et al.

Serum levels of K2 and risk of bone fractures or osteopenia

Four case-control studies, two cross-sectional studies and one cohort study examined the association between K2 intake (assessed by serum levels) and bone related outcomes (17–19,21,22,25,26). The results are presented in Table 5. Hodges et al. found significantly lower (p<0.05) serum levels of both MK-7 and MK-8 in a group of patients with femoral neck fractures and in another group of patients with spinal crush fractures compared to a control group (17). The second study by Hodges et al. also found significantly lower (p<0.01) serum levels of MK-7 and MK-8 in patients with femoral neck fractures compared to controls (18). Kawana et al. found no significant differences in MK-4 and MK-7 levels between a control group and two separate groups of patients with either vertebral fractures or hip fractures (21). Nakano and colleagues found significantly lower (p<0.05) serum levels of MK-7 in both male and female patients with hip fractures compared to male and female controls (22). In a cross-sectional study by Iwasaki et al. no significant differences in serum levels of MK-4 were found between the low BMD and normal BMD group. In a cross-sectional study by Tamatani et al. the subjects in the decreased BMD group had significantly lower (p<0.05) serum levels of MK-7 compared with the normal BMD group (25). Tsugawa et al. found no significant associations between incidence of vertebral fractures and serum levels of either MK-4 or MK-7 (26).

Table 5 Serum levels of vitamin K2 in relation to bone fractures and osteopenia.

Study

Study

size

Fracture

events

Circulating levels of vitamin K2 expressed in pg/ml as median (interquartile range) or mean ± 1 SD

Controls

Patients with femoral neck fractures

Patients with spinal crush fractures

MK-7

MK-8

MK-7

MK-8

MK-7

MK-8

Hodges et al. (17)

46

29

282

(177.5, 472)

215

(43, 282.5)

17.5

(10, 64.8)**

15

(15, 25.3)**

174

(10, 196)**

15

(15, 127)*

Hodges et al. (18)

89

51

226 ± 178

161 ± 145

120 ± 84**

89 ± 113**

Circulating levels of vitamin K2 expressed in ng/ml as mean ± 1 SD

Controls

Patients with vertebral fractures

Patients with hip fractures

MK-4

MK-7

MK-4

MK-7

MK-4

MK-7

Kawana et al. (21)

74

51

0.05 ± 0.02

1.2 ± 2.2

0.05

1.3 ± 1.9

0.05

1.5 ± 2.2

Circulating levels of MK-7 expressed in ng/ml as mean ± 1 SD

Male controls

Male patients with fractures

Female controls

Female patients with fractures

Nakano et al. (22)

147

99

4.28 ± 3.75

1.60 ± 1.60**

10.8 ± 7.01

2.67 ± 4.13**

Circulating levels of MK-4 expressed in ng/ml (19) and MK-7 expressed in nnmol/l (25) as mean ± 1 SD

Normal BMD†

Decreased BMD‡

Iwasaki et al. (19)

177

0.10 ± 0.07

0.10 ± 0.17

Tamatani et al. (25)

27

1.44 ± 0.85

0.71 ± 0.35*

Association between incidence of vertebral fractures and circulating levels of vitamin K2 (nmol/l)

β-coefficient

P-value

MK-4

MK-7

MK-4

MK-7

Tsugawa et al. (26)

379

35

-0.345

-0.005

0.602

0.672

* p<0.05, relative to the control subjects or normal BMD group.

** p<0.01, relative to the control subjects.

† Defined as BMD of at least Young Adult Mean (YAM) - 1.0 SD in Iwasaki et al and at least peak bone density (PBD) - 2.5 SD in Tamatani et al.

‡ Defined as BMD less than YAM - 1.0 SD in Iwasaki et al and less than PBD - 2.5 SD in Tamatani et al.

Vitamin K and diabetes

Dietary K1 and K2 intake and risk of type 2 diabetes

Beulens et al. investigated the association between vitamin K intake (measured by FFQ) and the incidence of type 2 diabetes (14). The study population was from the EPIC-NL. The multivariate model is presented in Supplementary Table 4, available online at www.ntfe.no. Mean intake of K1 and K2 was 200 (SD 98) µg/day and 31 (SD 7) µg/day, respectively. The study did not find a significant association between K1 intake and type 2 diabetes incidence, but K1 tended to be associated with a reduced risk with a hazard ration of 0.81 (95% CI 0.66-0.99) for the highest versus the lowest quartile. For vitamin K2, participants in the highest quartile had a 20% reduced risk of type 2 diabetes compared to those in the lowest quartile (ptrend=0.04). A 7% reduction in disease risk for each 10 µg/day increase of K2 was also observed (p=0.038).

Vitamin K and cancer

Dietary K1 intake and risk of cancer

Three cohort studies examined the association between K1 intake (measured by FFQ) and risk of cancer outcomes (20,23,24). The results of the multivariate models for each individual study are presented in Table 6. Juanola-Falgarona et al. found a significant inverse association between K1 intake and cancer mortality (20). Subjects in the highest quartile had a 46% reduction in cancer mortality risk compared to the subjects in the lowest quartile (ptrend=0.033). The two other studies found no significant associations between K1 and cancer risk (23,24).

Table 6 Dietary intake of vitamin K1 and the risk of cancer (multivariate models controlling for cancer risk factors and dietary variables).

Study

Study size

Cancer eventsa

Effect size by vitamin K1 intake category (95% CI)

P-value

1 (low)

2

3

4 (high)

Juanola-Falgarona et al. (20)

7216

130 [mortality]

1

0.90 (0.56, 1.45)

1.01 (0.63, 1.62)

0.54 (0.30, 0.96)

0.033

Nimptsch et al. (23)

11319

268 [total]

1

0.88 (0.62, 1.25)

0.91 (0.63, 1.31)

1.02 (0.70, 1.48)

0.74

113 [advanced]

1

1.11 (0.66, 1.88)

0.97 (0.56, 1.71)

0.84 (0.46, 1.56)

0.50

Nimptsch et al. (24)

24340

1755 [incidence]

1

0.95 (0.83, 1.09)

1.02 (0.89, 1.16)

0.96 (0.83, 1.10)

0.71

458 [mortality]

1

0.88 (0.68, 1.14)

0.89 (0.69, 1.16)

0.93 (0.71, 1.22)

0.70

a Relevant outcomes are presented in brackets.

Dietary K2 intake and risk of cancer

Three cohort studies examined the association between K2 intake (measured by FFQ) and risk of cancer outcomes (20,23,24). The results of the multivariate models for each individual study are presented in Table 7. Two of the studies found a significant inverse association between K2 intake and incidence of cancer related outcomes. Both studies were conducted by the same authors, and the study population in both were from the EPIC-Heidelberg cohort. The first study by Nimptsch et al. found a median (25-75th percentile) K2 intake of 34.7 (25.7-45.7) µg/day and a 63% reduction in the risk of advanced prostate cancer incidence in the highest quartile compared to the lowest quartile (ptrend=0.03) (23). The second study by Nimptsch et al. found a 28% reduction in incident total cancer mortality in men and women in the highest quartile compared to the lowest quartile (ptrend=0.03) (24). In sex-specific analyses they found that this association was significant only for men, and mainly driven by significant inverse associations with prostate and lung cancer (data not shown).

Table 7 Dietary intake of vitamin K2 and the risk of cancer (multivariate models controlling for cancer risk factors and dietary variables).

Study

Study size

Cancer eventsa

Effect size by vitamin K2 intake category (95% CI)

P-value

1 (low)

2

3

4 (high)

Juanola-Falgarona et al. (20)

7216

130 [mortality]

1

0.88 (0.55, 1.43)

1.00 (0.60, 1.67)

0.62 (0.32, 1.21)

0.45

Nimptsch et al. (23)

11319

268 [total]

1

0.76 (0.53, 1.08)

0.71 (0.47, 1.06)

0.65 (0.39, 1.06)

0.10

113 [advanced]

1

0.79 (0.47, 1.34)

0.67 (0.36, 1.25)

0.37 (0.16, 0.88)

0.03

Nimptsch et al. (24)

24340

1755 [incidence]

1

0.91 (0.80, 1.04)

0.91 (0.79, 1.04)

0.86 (0.73, 1.01)

0.08

458 [mortality]

1

0.77 (0.60, 0.99)

0.64 (0.49, 0.85)

0.72 (0.53, 0.98)

0.03

a Relevant outcomes are presented in brackets.

DISCUSSION

A total of 16 studies met the inclusion criteria and were included in this review. Exposure measurements included FFQ and serum levels of vitamin K. Five cohort studies examined the association between dietary vitamin K intake and CVDs. No associations were found between K1 and risk of CVD in any of the studies. Three studies reported significant associations between K2 and a reduction of CVD events (15,16, 28). Eight of the included studies examined the association between vitamin K intake and bone related outcomes, defined as bone fractures or osteopenia. Apalset et al. used an FFQ to measure vitamin K intake and found a significant association between K1 intake and reduced risk of hip fracture incidence (13). No association was found for K2. The remaining seven studies used serum levels of vitamin K as an exposure measurement. Both K1 and subtypes of K2 were significantly lower in patients with fractures compared with control subjects in three case-control studies (17,18,22). Kawana et al. found no significant difference between either K1 or K2 in patients compared with controls (21). Only K1 was significantly associated with a lower vertebral fracture incidence in the study by Tsugawa at al. (26). Iwasaki et al. found no significant differences between K1 or K2 levels in the low BMD group compared with the normal BMD group (19). Tamatani et al. however found significantly lower serum levels of both K1 and K2 in the decreased BMD group (25). 

Only one study examined the association between vitamin K intake and diabetes (14). This cohort study found a significant association between K2 intake and reduced risk of type 2 diabetes, while K1 only tended to be associated with a reduced risk. Three cohort studies examined the association between vitamin K intake and cancer. Juanola-Falgarona et al. found a significant association between K1 intake and reduced risk of cancer mortality (20), while Nimptsch et al. found a significant association between K2 intake and reduced risk of advanced cancer incidence (23) and between K2 intake and total cancer mortality (24).

Limitations of the included studies

All of the studies discussed in this review are observational studies, which by design are not able to detect a cause and effect relationship between vitamin K and the outcomes examined. It is however possible to detect correlations, but these may be subject to potential biases and confounding factors. Five of the six cohort studies used FFQs to measure dietary intake of vitamin K, and while FFQs are valid tools to rank participants according to nutritional intake their ability to determine absolute intakes are limited. The study by Apalset et al. used an FFQ that had been validated for several nutrients, but not specifically for vitamin K (13). In addition, median intake of K1 and K2 in this study was 57.9-78.4 µg/day (depending on gender and morbidity) and 10.8-12.6 µg/day, respectively. The intake of K2, when compared to the K2 intake in other studies included in this review (mean intake of about 30 µg/day), may have been too low in order to exert a protective effect. This could explain why no significant association was found for K2 intake and reduced risk of hip fractures in this study. On the other hand, in the studies where only K2 was inversely associated with CVD, the relative validity of the administered FFQs was better for K2 than for K1 (15,16,28).

Most of the studies examining bone related outcomes measured serum levels of vitamin K, which may not be the best method with regard to determining vitamin K exposure (29). Both the method and subtype of vitamin K measured is going to influence the results. Iwasaki et al. found no significant difference between serum levels of MK-4 in the low BMD group compared to the normal BMD group (19). This is not unexpected, considering humans have a very low fasting blood concentration of MK-4 and a rapid tissue uptake following dietary ingestion (8). MK-9 however has been shown to have a very long half-life compared to MK-4 and K1, but since menaquinones in general are not detected unless supplements are taken or K2-rich foods are consumed, correlating K2 intake with serum levels is problematic (29). Nakano et al. also noted that earlier HPLC methods (used in several studies included in this review) appear to be less sensitive and more imprecise compared to newly developed assay methods (22).

The time aspect also needs to be considered in relation to the FFQs, as they were administered only at baseline in almost all the included studies and thus may not fully capture the vitamin K exposure since no information about dietary changes during the study period would be detected (supplementary Table 1). A combination of biomarkers (e.g. serum levels) with dietary intake, both subjected to repeated measurements, would likely be the best way to measure vitamin K exposure in prospective studies.

As noted by many of the studies included in this review, K1 and K2 are usually associated with opposite lifestyles. K1 is associated with a healthy food pattern, as it is mainly found in vegetables. Residual confounding could therefore lead to or strengthen an inverse association with K1 and disease outcomes. In contrast, K2 is found mainly in cheese and meat; foods associated with an unhealthy lifestyle. Residual confounding could in this case attenuate the results for K2. Another limitation is the fact that vitamin K content of foods may vary according to growth and production conditions, and that the bioavailability is dependent on food preparation, fat content of meals and subtype of vitamin K (3,30,31).

Limitations of the review

This review was not intended to be a comprehensive review of the associations between vitamin K and disease, therefore only studies that measured both K1 and K2 intake in the subjects were included. This may be regarded as both a weakness and a strength. On the one hand, this excludes many studies examining only one of them in relation to relevant disease outcomes. On the other hand, measuring both K1 and K2 intake in the same study population can be valuable when trying to determine if the two forms are differently associated with health and disease in a population.

How efficient is the conversion of K1 to K2?

The human body accumulates both K1 and K2 (mainly in the form of MK-4) in a tissue-specific distribution pattern (3,9). MK-4 has an unusual distribution pattern exceeding the levels of K1 in some organs, suggesting important roles for this subtype of K2. It has recently been confirmed that K1 can be converted to MK-4 in vivo, starting with intestinal cell cleavage of ingested K1 to menadione (32). Menadione is then transported into the blood via the lymphatic system and subsequently taken up by different tissues, where it gets converted to MK-4 by the K2 biosynthetic enzyme UBIAD1 that adds a prenyl side chain to the ring structure (33). How efficient this conversion is remains to be determined, but experiments in rats seem to suggest that Wistar rats (34) make the conversion better than Lewis rats (35), indicating variation within species.

In humans, the conversion of other nutrients from the diet has been shown to be highly variable. Conversion of β-carotene to retinol depends on factors like vitamin A status, gut integrity and genetic polymorphisms (36,37). Another example is alpha-linolenic acid (18:3n−3) conversion to longer-chain PUFAs, where the conversion seems to be greater in women than in men (38). In addition, many people are taking medications (like lipophilic statins) that inhibits the enzymatic activity of UBIAD1 (33). Participants administered FFQs in the studies included in this review had K1 intakes that were many times higher than K2 intakes. If everyone converted enough K1 to K2 to meet their need, it should not matter which form is consumed through the diet. However, only K2 intake seemed to be correlated with CVD (15,16,28) and to some extent type 2 diabetes (14). As Schurgers et al. have suggested, applying RDA values defined on the basis of blood coagulation data and hepatic vitamin K requirement may ultimately be inadequate to allow adequate carboxylation of all extrahepatic VKD proteins (39).

Conclusions

The identified studies suggest protective roles for both K1 and K2. Only K2 was associated with a reduced risk of CVD in three studies (15,16,28). Both K1 and K2 were inversely associated with reduced risk of bone fractures or low bone density in four studies (17,18,22,25), but K2 was not associated with a greater reduced risk compared with K1. Two studies found significant associations for only K1 (13,26), but one of them had limitations with regard to its ability to detect a protective effect for K2 (13). Compared to K1, K2 was associated with a greater reduced risk of type 2 diabetes in one study (14). Two studies found an inverse association between K2 intake and cancer risk (23,24), while one study found an inverse association for K1 intake and cancer risk (20). Looking at CVD, the results look promising for K2 compared to K1. Moreover, K2 may exert a more protective effect in the prevention of type 2 diabetes. However, there is not sufficient data available to draw any definitive conclusions. More prospective studies are therefore needed in order to determine whether separate recommendations should be made for the intake of vitamin K2.

Competing interests

Both authors declare no competing interests.

Author’s contributions

This review is a shortened version of a thesis submitted by Thomas K. Reinan (TKR) as part of the course ERN3120 in spring 2017 at the Department of Nutrition, University of Oslo. TKR wrote the paper with contributions from Prof. Asim K. Duttaroy. Both authors read and approved the final manuscript.

Supplementary Table 1 Characteristics of the included studies.

Study (design)

Title (year)

Participants

Exposure

Outcomes

Apalset et al. (cohort study) (13)

Intake of vitamin K1 and K2 and risk of hip fractures: The Hordaland Health Study (2011)

Participants were from the community-based Hordaland Health Study (HUSK) in Norway conducted in 1997-1999, and consisted of 2807 men and women aged 71-75 years

At baseline participants completed a validated (but not specifically for vitamin K) FFQ, and intakes of vitamin K1 and total vitamin K2 were computed from a food database developed at the Department of Nutrition, University of Oslo

The primary outcome was incidence of hip fractures. Between baseline and 2009, computerized discharge diagnoses records (coded according to ICD-9 and ICD-10) of the hospitals serving Hordaland County were searched, and hip fractures confirmed by a concurrent code of an adequate surgical procedure were included in the calculations

Beulens et al. (cohort study) (14)

Dietary phylloquinone and menaquinones intakes and risk of type 2 diabetes (2010)

Participants were from the EPIC-NL, a Dutch cohort merged from the Prospect-EPIC (17357 women aged 49-70 years) and MORGEN-EPIC (22654 adults aged 20-59 years) cohorts set up simultaneously in 1993-1997. 38094 participants were included in the analytic cohort

Intake of vitamin K1 and vitamin K2 (MK-4 to MK-10) was obtained from a validated FFQ administered during the year preceding enrollment. Information on vitamin K contents of foods came from data on a large series of Dutch foods assessed at the Biochemistry Laboratory, Maastricht University. The relative validity of the FFQ had a Spearman’s correlation coefficient of 0.24 for vitamin K1 and 0.51–0.72 for vitamin K2 subtypes (MK-4 to MK-9)

The primary outcome was diagnosis of diabetes, coded according to the ICD-9-CM. Occurrence of diabetes during follow-up was self-reported in two follow-up questionnaires. In addition, in the Prospect study incident cases of diabetes were detected via a urinary glucose strip test included with the first follow-up questionnaire. Diagnoses of diabetes were also obtained from the Dutch Centre for Health Care Information. Median follow-up was 10.3 years

Gast et al. (cohort study) (15)

A high menaquinone intake reduces the incidence of coronary heart disease (2008)

Participants were from the Prospect-EPIC cohort, where 17357 women aged 49-70 years living in Utrecht and surroundings were enrolled between 1993 and 1997. Recruitment was done through a regional breast cancer screening program. The final study population consisted of 16057 women

A validated FFQ was used at baseline to assess the intake of vitamin K1 and vitamin K2 (MK-4 to MK-9). Information on vitamin K contents of foods came from data on a large series of Dutch foods assessed at the Biochemistry Laboratory, Maastricht University. For some foods, published data by others were used to update the dietary database for vitamin K. The relative validity of the FFQ had a Spearman’s correlation coefficient of 0.24 for vitamin K1 and 0.51–0.72 for vitamin K2 subtypes (MK-4 to MK-9)

The cohort was linked to a hospital discharge database, and diagnoses of CHD were coded according to ICD-9 and ICD-10. Whenever multiple events occurred, the first diagnosis was taken as endpoint. Mean follow-up was 8.1 (1.6 years)

Geleijnse et al. (cohort study) (16)

Dietary intake of menaquinone is associated with a reduced risk of coronary heart disease: The Rotterdam Study (2004)

Participants were from the Rotterdam Study, a prospective population-based study designed to assess the occurrence of diseases in the elderly and clarify their determinants. Baseline data was collected from the cohort, consisting of 7983 men and women aged 55 years and over, from 1990-1993. 4807 subjects were included in the final study population

The diet was assessed by a trained dietician at the study center, using a validated semiquantitative FFQ. Concentrations of vitamin K1 and vitamin K2 (MK-4 to MK-10) in a large series of Dutch foods had previously been assessed in a laboratory by one of the study authors

Fatal and nonfatal CHD events were reported by general practitioners (GPs) in the research area (85% of the cohort) by a computerized system. Paper forms was obtained from GPs outside the research area. GP records and hospital discharge letters were verified by research physicians, and events were coded by two physicians according to ICD-10. Events were considered fatal if death occurred within 28 days after the onset of symptoms. Mean follow-up was 7.2 (1.9) years

Hodges et al. (case-control study) (17)

Depressed levels of circulating menaquinones in patients with osteoporotic fractures of the spine and femoral neck (1991)

Participants were one group of 18 patients (aged 69-90 years) with fractures of the neck of the femur, one group of 11 patients (aged 62-87 years) with spinal fractures and one group consisting of 17 healthy control subjects (aged 56-86 years). All groups contained both men and women, and all patients had senile or postmenopausal osteoporosis

Blood samples were taken from all three groups and circulating levels of vitamin K1 and two forms of vitamin K2 (MK-7 and MK-8) were measured

The primary outcome was osteoporotic fractures of either the spine, in patients under long-term clinical follow-up, or the femoral neck in newly diagnosed patients

Hodges et al. (case-control study) (18)

Circulating levels of vitamins K1 and K2 decreased in elderly women with hip fracture (1993)

Participants were 51 women (mean age 80.8 ± 6.6 years) who in the last few hours had sustained a femoral neck fracture, and a control group of 38 healthy women (mean age 79.6 ± 4.8 years) from the same population

Blood samples for serum vitamin K1 and vitamin K2 (MK-7 and MK-8) were obtained before surgical treatment within a mean 5.8 h after fracture from the cases and during the morning from the control group

The primary outcome was fracture of the hip

Iwasaki et al. (cross-sectional study) (19)

Bone mineral density and bone turnover among young women in Chiang Mai, Thailand (2013)

Participants were 177 women (mean age 23.4 ± 2.5 years) affiliated with Chiang Mai University hospital. Subjects either worked as nurses and administrative employees, or were nursing or medical students undergoing practical training

Serum levels of vitamin K1 and vitamin K2 (MK-4) were measured by HPLC

The primary outcome was decreased BMD. Measurement of BMD of the calcaneus was performed by Quantitative Ultrasound. Subjects were divided into two groups based on the measurement; “Normal BMD” with BMD of at least Young Adult Mean (YAM) – 1.0 SD T-score, and “Low BMD” with BMD less than YAM – 1.0 SD T-score

Juanola-Falgarona et al. (cohort study) (20)

Dietary intake of vitamin K is inversely associated with mortality risk (2014)

Participants were from the PREDIMED study, a parallel-group, multicenter, randomized controlled clinical trial aiming to assess the effect of Mediterranean diets on the primary prevention of cardiovascular disease in elderly individuals at high cardiovascular disease risk. A prospective cohort analysis was conducted in 7216 participants from the cohort, consisting of men and women aged 55–80 and 60-80 years

The diet was assessed by trained dieticians who administered a validated FFQ at baseline and annually thereafter. Dietary vitamin K1 intake was calculated using the USDA nutrient database, and total vitamin K2 intake was calculated using previously published composition data sources. Reproducibility for dietary vitamin K1 and vitamin K2 intake estimated by the Pearson correlation coefficient (r) was 0.755 and 0.655, respectively

The primary outcomes of the analysis were cardiovascular mortality, cancer mortality and all-cause mortality. Identification of outcomes was done by repeated contact with participants, contact with family physicians, annual review of medical records and consultation of the National Death Index. All medical records related to end points were evaluated by an end-point adjudication committee. Median follow-up was 4.8 years

Kawana et al. (case-control study) (21)

Circulating levels of vitamin K, menaquinone-4, and menaquinone-7 in healthy elderly Japanese women and patients with vertebral fractures and patients with hip fractures (2001)

Participants were one group of 13 patients (mean age 80.3 ± 7.8 years) with vertebral fracture, one group of 38 patients with hip fracture (mean age 79.8 ± 9.2 years) and one control group consisting of 23 healthy women (mean age 80.1 ± 3.5 years)

Blood samples were collected from all subjects and circulating levels of vitamin K1 and vitamin K2 (MK-4, MK-7) were measured by HPLC. The samples were collected from the hip fracture patients immediately after they were taken to an emergency room

The primary outcomes were fracture of the vertebrae (diagnosed by X-ray of the thoracic and lumbar spine) or hip fracture

Nakano et al. (case-control study) (22)

High prevalence of hypo-vitaminosis D and K in patients with hip fracture (2011)

Participants were 99 patients with hip fracture transferred to Tamana Central Hospital, and 48 age-matched nursing home residents in close proximity to the hospital. Duration of the enrollment was 6 months

Blood was drawn within 24 hours following fracture in the patients. Plasma vitamin K1 and K2 (MK-7) levels were determined by high-performance liquid chromatography-tandem mass-mass spectrometry with atmospheric pressure chemical ionization (LC-APCI-MS/MS)

The primary outcome was fracture of the hip

Nimptsch et al. (cohort study) (23)

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) (2008)

The EPIC-Heidelberg cohort includes 25540 participants, of whom 11928 are men aged 40-65 years. Recruitment from the general population from Heidelberg and surrounding communities took place between 1994-1998, and information on dietary and nondietary factors was assessed at baseline. 11319 men were included in the analytic cohort

Habitual dietary intake during the previous 12 months preceding recruitment was assessed at baseline by a validated self-administered semi-quantitative FFQ. Intake of vitamin K1 and vitamin K2 (MK-4 to MK-14) was estimated by using previously published HPLC-based food-content data

Identification of prostate cancer cases was based on self-reported primary prostate cancer during follow-up or on death certificates that were coded for prostate cancer as the underlying cause of death. Verification was done by medical records, death certificates or both. Information on stage and grade of prostate cancer was extracted by the study physician from pathology reports. Mean follow-up was 8.6 years

Nimptsch et al. (cohort study) (24)

Dietary vitamin K intake in relation to cancer incidence and mortality: results from the Heidelberg cohort of the European Prospective Investigation into Cancer and Nutrition (EPIC-Heidelberg) (2010)

The EPIC-Heidelberg cohort includes 25540 participants aged 35-64 years. Recruitment from the general population from Heidelberg and surrounding communities took place between 1994-1998. A total of 24340 participants were included in the final study population

Habitual dietary intake during the previous 12 months preceding recruitment was assessed at baseline by a validated self-administered semi-quantitative FFQ. Intake of vitamin K1 and vitamin K2 (MK-4 to MK-14) was estimated by using previously published HPLC-based food-content data

The primary outcome was incidence (ICD-O-2 codes C00-C77 and C80, excluding nonmelanoma skin cancer) and mortality (ICD-10 codes C00-C97 and B21) of total invasive cancer. Incident cancer cases were self-reported by means of follow-up questionnaires mailed out at regular time intervals, and verified by pathology reports, medical records and/or death certificates on the basis of the ICD-O-2. Death certificates were obtained from the mortality registries for all deceased participants

Tamatani et al. (cross-sectional study) (25)

Decreased circulating levels of vitamin K and 25-hydroxyvitamin D in osteopenic elderly men (1998)

Participants were 27 ambulatory men aged 60-90 years who were residents of a nursing home

Blood samples were obtained in the morning after an overnight fast, and plasma levels of vitamin K1 and vitamin K2 (MK-7) were measured by HPLC

The primary outcome was decreased BMD. Measurement of BMD of L2-L4 was performed by DXA. Subjects were divided into two groups based on the measurement; “Normal BMD” with BMD of at least peak bone density (PBD) - 2.5 SD, and “Decreased BMD” with BMD less than PBD - 2.5 SD, calculated using values from a Japanese study

Tsugawa et al. (cohort study) (26)

Low plasma phylloquinone concentration is associated with high incidence of vertebral fracture in Japanese women (2007)

Participants were 379 women (mean age 63.0 ± 10.8. years) living in a rural area of Nagano enrolled in 2002-2003 and followed up until 2006

Plasma was collected in the morning and subsequently measured for vitamin K1 and vitamin K2 (MK-4 and MK-7) by high-performance liquid chromatography-tandem mass spectrometry at the beginning of the study

Incidence of vertebral fractures was measured by DXA of the lumbar spine (L2-4) and femoral neck. Incident fractures with apparent major trauma were excluded from the study

Vissers et al. (cohort study) (27)

Intake of dietary phylloquinone and menaquinones and risk of stroke (2013)

Participants were from the EPIC-NL, a Dutch cohort merged from the Prospect-EPIC (17357 women aged 49-70 years) and MORGEN-EPIC (22654 adults aged 20-59 years) cohorts set up simultaneously in 1993-1997. The final study population consisted of 35476 participants, on average 49 ± 12 years at baseline

Intake of vitamin K1 and vitamin K2 (MK-4 to MK-10) was obtained from a validated FFQ administered during the year preceding enrollment. Information on vitamin K contents of foods came from data on a large series of Dutch foods assessed at the Biochemistry Laboratory, Maastricht University. The relative validity of the FFQ had a Spearman’s correlation coefficient of 0.24 for vitamin K1 and 0.51–0.72 for vitamin K2 subtypes (MK-4 to MK-9)

The primary outcome was incidence of stroke. Information was obtained during follow-up from causes of death and from the Dutch Centre for Health Care Information, which holds a standardized computerized register of hospital discharge diagnoses. The cohort was linked to the database with a validated probabilistic method. All diagnoses were coded according to the ICD-9-CM. Mean follow-up was 12.1 ± 2.1 years

Vissers et al. (cohort study) (28)

The relationship between vitamin K and peripheral arterial disease (2016)

Participants were from the EPIC-NL, a Dutch cohort merged from the Prospect-EPIC (17357 women aged 49-70 years) and MORGEN-EPIC (22654 adults aged 20-59 years) cohorts set up simultaneously in 1993-1997. 36629 participants were included in the analytic cohort. At baseline, the study population was on average 50 ± 12. years

Intake of vitamin K1 and vitamin K2 (MK-4 to MK-10) was obtained from a validated FFQ administered during the year preceding enrollment. Information on vitamin K contents of foods came from data on a large series of Dutch foods assessed at the Biochemistry Laboratory, Maastricht University The validity of the FFQ was estimated against twelve monthly 24-h recalls in 58 women and 63 men. The reliability of vitamin K1 intake had a Spearman’s correlation coefficient of 0.26 for men and 0.16 for women. The reliability of vitamin K2 intake had a Spearman’s correlation coefficient of 0.60 for men and 0.48 for women

The primary outcome was PAD. Information on mortality and morbidity was obtained during follow-up from the causes of death registry of ‘Statistics Netherlands’ and from the Dutch Centre for Health Care Information, which holds a standardized computerized register of hospital discharge diagnoses. The cohort was linked to the database with a validated probabilistic method. All diagnoses were coded according to the ICD-9-CM, except for causes of death which were coded according to the ICD-10. Mean follow-up was 12.1 ± 2.1 years

Abbreviations: BMD, bone mineral density; BMI, body mass index; CHD, coronary heart disease; DXA, dual-energy X-ray absorptiometry; FFQ, Food Frequency Questionnaire; HPLC, high-performance liquid chromatography; ICD, International Classification of Diseases; MI, myocardial infarction; MK-n, menaquinone-n (where n refers to the number of sidechains in the molecule); PAD, peripheral arterial disease; USDA, United States Department of Agriculture.

Supplementary Table 2 Description of quantiles and multivariate models used in included studies measuring dietary vitamin K intake.

Study (outcome)

Expression of effect size

Quantiles of vitamin K intake (µg/d)a

Adjustmentsb

Apalset et al. (bone fractures) (13)

Quartiles

K1

Q1: <42.2 [women]

<52.9 [men]

Q2: 42.2-66.7 [women]

52.9-77.4 [men]Q3: 66.7-108.6 [women]

77.4-113.9 [men] Q4: >108.7 [women]

>113.9 [men]

K1: Sex, total energy intake, smoking, BMI, vitamin D- and calcium intake

K2

Q1: <7.2 [women]

<8.5 [men]

Q2: 7.2-10.7 [women]

8.5-11.9 [men] Q3: 10.7-14.5 [women]

11.9-16.2 [men] Q4: >14.5 [women]

>16.2 [men]

K2: The same variables as for K1

Beulens et al. (diabetes) (14)

Quartiles

Continuous model for vitamin K increments (per 50 µg K1 and 10 µg K2)

K1

Q1: 100.1

Q2: 155.7 Q3: 211.4 Q4: 308.1

K1: Age, sex, waist circumference, smoking status, physical activity, hypertension, education, alcohol consumption, total energy intake and energy-adjusted intake of saturated, polyunsaturated, and monounsaturated fat, protein, fiber, calcium, vitamin C, and vitamin E

K2

Q1: 16.0

Q2: 24.5 Q3: 32.9 Q4: 46.1

K2: The same variables as for K1

Gast et al. (CVD) (15)

Continuous model for vitamin K increments (per 10 µg K1 and K2)

K1: Age, BMI, smoking, hypertension, diabetes, hypercholesterolemia and energy-adjusted dietary factors (alcohol, fiber, protein from plant origin and folic acid)

K2: Age, BMI, smoking, hypertension, diabetes, hypercholesterolemia and energy-adjusted dietary factors (alcohol, saturated fat, polyunsaturated fat and energy)

Geleijnse et al. (CVD) (16)

Tertiles

K1

T1: <200

T2: 200-278 T3: >278

K1: Age, gender, total energy intake, BMI, smoking status, pack-years of cigarette smoking, diabetes, education and intake of alcohol, saturated fat, polyunsaturated fat, flavonols and calcium

K2

T1: <21.6

T2: 21.6-32.7 T3: >32.7

K2: The same variables as for K1

Juanola-Falgarona et al. (CVD and cancer) (20)

Quartiles

K1

Q1: 170.5

Q2: 276.1 Q3: 349.7 Q4: 626.4

K1: Sex, age, BMI, recruiting center, intervention group, smoking, leisure time activity, education, history of diabetes, hypertension, and hypercholesterolemia, oral antidiabetic medication, antihypertensive medication, statin medication and energy-adjusted dietary factors (fruits, legumes, cereals, dairy products, meat, fish, olive oil, nuts, alcohol)

K2

Q1: 18.4

Q2: 29.9 Q3: 39.0 Q4: 57.5

K2: The same variables as for K1, except the energy-adjusted dietary factor dairy products was substituted with vegetables

Supplementary Table 2 (Continued)

Study (outcome)

Expression of effect size

Quantiles of vitamin K intake (µg/d)a

Adjustmentsb

Nimptsch et al. (cancer) (23)

Quartiles

K1

Q1: <71

Q2: 71-94 Q3: 94-124 Q4: >124

K1: Smoking, education, vigorous physical activity, energy from fat, alcohol, nonfat-nonalcohol energy, calcium, vitamin D, tomato or tomato products, BMI, history of diabetes, family history of prostate cancer, and intake of vegetables, dairy products, and meat or meat products

K2

Q1: <26

Q2: 26-35 Q3: 35-46 Q4: >46

K2: The same variables as for K1

Nimptsch et al. (cancer) (24)

Quartiles

K1

Not specified

K1: Total energy, alcohol, BMI, waist-to-hip ratio, smoking, smoking duration, physical activity, and educational level

K2

Not specified

K2: The same variables as for K1

Vissers et al. (CVD) (27)

Quartiles

Continuous model for vitamin K increments (per 50 µg K1 and 10 µg K2)

K1

Q1: 101

Q2: 157 Q3: 213 Q4: 308

K1: Age, sex, waist circumference, smoking, physical activity, use of oral anti-contraceptives, energy intake and energy adjusted intake of protein, fat, glycemic index, alcohol, vitamin C, and fiber

K2

Q1: 16

Q2: 25 Q3: 33 Q4: 46

K2: The same variables as for K1

Vissers et al. (CVD) (28)

Quartiles

Continuous model for vitamin K increments (per 50 µg K1 and 10 µg K2)

K1

Q1: 97

Q2: 157 Q3: 214 Q4: 333

K1: Age, sex, waist-hip ratio, BMI smoking, diabetes, physical activity, level of education, energy intake and energy-adjusted intake of protein, vitamin C and fiber and glycemic load

K2

Q1: 16

Q2: 25 Q3: 33 Q4: 49

K2: The same variables as for K1

a Data are in either quantile ranges or means.

b Cox proportional hazards models.

Abbreviations: BMI, body mass index; CVD, cardiovascular disease.

Supplementary Table 3 Dietary intake of vitamin K1 and K2 and the risk of hip fracture (multivariate model controlling for bone fracture risk factors and dietary variables).

Study

Study size

Fracture events

Effect size by vitamin K1 intake category (95% CI)

P-value

1 (low)

2

3

4 (high)

Apalset et al. (13)

2807

225

1.63 (1.06, 2.49)

1.21 (0.82, 1.80)

0.96 (0.64, 1.44)

1

0.015

Effect size by vitamin K2 intake category (95% CI)

1 (low)

2

3

4 (high)

1.03 (0.64, 1.67)

1.09 (0.71, 1.66)

1.19 (0.80, 1.75)

1

0.983

Supplementary Table 4 Dietary intake of vitamin K1 and K2 and the risk of type 2 diabetes (multivariate model controlling for type 2 diabetes risk factors and dietary variables).

Study

Study size

Diabetes events

Effect size by vitamin K1 intake category (95% CI)

P-value

Effect size per 50 µg increase vitamin K1 (95% CI)

P-value

1 (low)

2

3

4 (high)

Beulens et al. (14)

38094

918

1

0.87

(0.71, 1.06)

0.90

(0.74, 1.09)

0.81

(0.66, 0.99)

0.08

0.98 (0.95, 1.02) p = 0.31

Effect size by vitamin K2 intake category (95% CI)

Effect size per 10µg increase vitamin K2 (95% CI)

1 (low)

2

3

4 (high)

1

0.99

(0.82, 1.21)

0.89

(0.72, 1.10)

0.80

(0.62, 1.02)

0.04

0.93 (0.87, 1.00) p = 0.038

REFERENCES

  1. Suttie JW. Vitamin K. In: Ross AC, Caballero B, Cousins RJ et al., editors. Modern nutrition in health and disease. 11th ed. Philadelphia: Lippincott Williams & Wilkins; 2012.

  2. Shearer MJ, Newman P. Metabolism and cell biology of vitamin K. Thromb Haemost. 2008;100:530–47.

  3. Shearer MJ, Newman P. Recent trends in the metabolism and cell biology of vitamin K with special reference to vitamin K cycling and MK-4 biosynthesis. J Lipid Res. 2014;55:345–62.

  4. Mizokami A, Kawakubo-Yasukochi T, Hirata M. Osteocalcin and its endocrine functions. Biochem Pharmacol. 2017;132:1–8.

  5. Lee NK, Sowa H, Hinoi E, et al. Endocrine Regulation of Energy Metabolism by the Skeleton. Cell. 2007;130:456–69.

  6. Oldenburg J, Marinova M, Müller-Reible C, Watzka M. The Vitamin K Cycle. Vitam Horm. 2008;78:35–62. Luo G, Ducy P, McKee MD, et al. Spontaneous calcification of arteries and cartilage in mice lacking matrix GLA protein. Nature. 1997;386:78–81.

  7. Luo G, Ducy P, McKee MD, et al. Spontaneous calcification of arteries and cartilage in mice lacking matrix GLA protein. Nature. 1997;386:78–81.

  8. Schurgers LJ, Vermeer C. Differential lipoprotein transport pathways of K-vitamins in healthy subjects. Biochim Biophys Acta. 2002;1570:27–32.

  9. Thijssen HH, Drittij-Reijnders MJ. Vitamin K status in human tissues: tissue-specific accumulation of phylloquinone and menaquinone-4. Br J Nutr. 1996;75:121–7.

  10. Niemeier A, Niedzielska D, Secer R, et al. Uptake of postprandial lipoproteins into bone in vivo: Impact on osteoblast function. Bone. 2008;43:230–7.

  11. Nordic nutrition recommendations 2012 : integrating nutrition and physical activity. 5th ed. Copenhagen: Nordic Council of Ministers; 2014.

  12. Food and Nutrition Board, Institute of Medicine. Dietary Reference Intakes for Vitamin A, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc. Washington, D.C.: National Academies Press; 2001.

  13. Apalset EM, Gjesdal CG, Eide GE, Tell GS. Intake of vitamin K1 and K2 and risk of hip fractures: The Hordaland Health Study. Bone. 2011;49:990–5.

  14. Beulens JWJ, van der A DL, Grobbee DE, et al. Dietary phylloquinone and menaquinones intakes and risk of type 2 diabetes. Diabetes Care. United States; 2010;33:1699–705.

  15. Gast GCM, de Roos NM, Sluijs I, et al. A high menaquinone intake reduces the incidence of coronary heart disease. Nutr Metab Cardiovasc Dis. 2009;19:504–10.

  16. Geleijnse JM, Vermeer C, Grobbee DE, et al. Dietary intake of menaquinone is associated with a reduced risk of coronary heart disease: the Rotterdam Study. J Nutr. American Society for Nutrition; 2004;134:3100–5.

  17. Hodges SJ, Pilkington MJ, Stamp TC, et al. Depressed levels of circulating menaquinones in patients with osteoporotic fractures of the spine and femoral neck. Bone. United States; 1991;12:387–9. Hodges SJ, Akesson K, Vergnaud P, et al. Circulating levels of vitamins K1 and K2 decreased in elderly women with hip fracture. J Bone Miner Res. United States; 1993;8:1241–5.

  18. Hodges SJ, Akesson K, Vergnaud P, et al. Circulating levels of vitamins K1 and K2 decreased in elderly women with hip fracture. J Bone Miner Res. United States; 1993;8:1241–5.

  19. Iwasaki E, Morakote N, Chaovistsaree S, Matsuo H. Bone mineral density and bone turnover among young women in Chiang Mai, Thailand. Kobe J Med Sci. 2013;59:149–56.

  20. Juanola-Falgarona M, Salas-Salvado J, Martinez-Gonzalez MA, et al. Dietary intake of vitamin K is inversely associated with mortality risk. J Nutr. 2014;144:743–50.

  21. Kawana K, Takahashi M, Hoshino H, Kushida K. Circulating levels of vitamin K1, menaquinone-4, and menaquinone-7 in healthy elderly Japanese women and patients with vertebral fractures and patients with hip fractures. Endocr Res. England; 2001;27:337–43.

  22. Nakano T, Tsugawa N, Kuwabara A, et al. High prevalence of hypovitaminosis D and K in patients with hip fracture. Asia Pac J Clin Nutr. Australia; 2011;20:56–61.

  23. 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;87:985–92.

  24. Nimptsch K, Rohrmann S, Kaaks R, Linseisen J. Dietary vitamin K intake in relation to cancer incidence and mortality: results from the Heidelberg cohort of the European Prospective Investigation into Cancer and Nutrition (EPIC-Heidelberg). Am J Clin Nutr. 2010;91:1348–58.

  25. Tamatani M, Morimoto S, Nakajima M, et al. Decreased circulating levels of vitamin K and 25-hydroxyvitamin D in osteopenic elderly men. Metabolism. 1998;47:195–9.

  26. Tsugawa N, Shiraki M, Suhara Y, et al. Low plasma phylloquinone concentration is associated with high incidence of vertebral fracture in Japanese women. J Bone Miner Metab. 2008;26:79–85.

  27. Vissers LET, Dalmeijer GW, Boer JMA, et al. Intake of dietary phylloquinone and menaquinones and risk of stroke. J Am Heart Assoc. 013;2:e000455.

  28. Vissers LET, Dalmeijer GW, Boer JMA, et al. The relationship between vitamin K and peripheral arterial disease. Atherosclerosis. 2016;252:15–20.

  29. Kyla Shea M, Booth SL. Concepts and controversies in evaluating vitamin K status in population-based studies. Nutrients. 2016;8.

  30. Schurgers LJ, Vermeer C. Determination of phylloquinone and menaquinones in food. Effect of food matrix on circulating vitamin K concentrations. Haemostasis. 2001;30:298–307.

  31. Schurgers LJ, Teunissen KJF, Hamulyak K, et al. Vitamin K-containing dietary supplements: comparison of synthetic vitamin K1 and natto-derived menaquinone-7. Blood. 2007;109:3279–83.

  32. Hirota Y, Tsugawa N, Nakagawa K, et al. Menadione (Vitamin K3) Is a Catabolic Product of Oral Phylloquinone (Vitamin K1) in the Intestine and a Circulating Precursor of Tissue Menaquinone-4 (Vitamin K2) in Rats. J Biol Chem. 2013;288:33071–80.

  33. Hirota Y, Nakagawa K, Sawada N, et al. Functional Characterization of the Vitamin K2 Biosynthetic Enzyme UBIAD1. PLoS One. 2015;10:e0125737.

  34. Thijssen HH, Drittij-Reijnders MJ. Vitamin K distribution in rat tissues: dietary phylloquinone is a source of tissue menaquinone-4. Br J Nutr. 1994;72:415–25.

  35. Ronden JE, Thijssen HH, Vermeer C. Tissue distribution of K-vitamers under different nutritional regimens in the rat. Biochim Biophys Acta. 1998;1379:16–22.

  36. Haskell MJ. The challenge to reach nutritional adequacy for vitamin A: β-carotene bioavailability and conversion--evidence in humans. Am J Clin Nutr. 2012;96:1193S–1203S.

  37. Borel P, Desmarchelier C. Genetic Variations Associated with Vitamin A Status and Vitamin A Bioavailability. Nutrients. 2017;9:246.

  38. Burdge GC. Metabolism of α-linolenic acid in humans. Prostaglandins, Leukot Essent Fat Acids. 2006;75:161–8.

  39. Schurgers LJ, Dissel PE, Spronk HM, et al. Role of vitamin K and vitamin K-dependent proteins in vascular calcification. Z Kardiol. 2001;90 Suppl 3:57–63.