Vitamin K and Bone: Past, Present, and Future


  • Caren M Gundberg

    Corresponding author
    1. Department of Orthopaedics and Rehabilitation, Yale University School of Medicine, New Haven, Connecticut, USA
    • Address correspondence to: Caren M. Gundberg, PhD, Department of Orthopaedics and Rehabilitation, Yale University School of Medicine, 333 Cedar Street, New Haven, CT 06510, USA
    Search for more papers by this author

  • The author states that she has no conflicts of interest

Throughout the past 20 yr, substantial epidemiologic evidence has emerged suggesting that suboptimal vitamin K status can have an adverse effect on BMD and increase the risk of bone fracture. However, only recently have prospective clinical trials been conducted to determine whether vitamin K plays a role in the prevention of bone loss. In the current issue of Journal of Bone and Mineral Research, Binkley et al.(1) report on the results of a double-blind placebo-controlled trial of vitamin K supplementation in postmenopausal women.

Vitamin K was discovered in 1935 by the Danish biochemist Henrik Dam, who was studying dietary deficiency diseases in chickens. When fed a diet free of fat and cholesterol, the chickens developed intramuscular hemorrhages and uncontrollable bleeding. Dam, with Edward Doisy, isolated the dietary substance required for blood clotting and called it the “Koagulation Vitamin,” or vitamin K. Soon thereafter, inactive prothrombin was shown to be the cause of the bleeding abnormality observed in vitamin K deficiency.(2,3) Almost simultaneously, dicoumarol was identified as the causative agent of a hemorrhagic disease in cattle fed spoiled sweet clover hay, and its derivative, warfarin, was established as a potent anticoagulant.(4) The mechanism of action of vitamin K remained unknown until 1974 when the modified amino acid, γ-carboxyglutamic acid (Gla), was identified in normal prothrombin.(5,6) The formation of Gla occurs through a unique vitamin K–dependent post-translational carboxylation of specific peptide-bound glutamate residues in a process that is inhibited by warfarin.(7) In the mid-1970s, fetal warfarin syndrome was described. Affected infants manifested bony malformations, stippled epiphyses, and nasal hypoplasia. Mothers of infants with these characteristics were found to have received oral anticoagulation therapy during the first trimester of pregnancy.(8) These findings led to the question of whether vitamin K nutrition or warfarin therapy plays a role in skeletal health and to the discovery of osteocalcin and other vitamin K–dependent proteins in bone.(9–11)

Vitamin K occurs as either phylloquinone (K1) or menaquinones (MK-n). All forms share a common structure, 2-methyl-1,4-napthoquinone, which is necessary for participation in carboxylation, but differ in the number and degree of saturation of isoprenyl side chains.(12) The major dietary form of the vitamin is phylloquinone, which is present in dark green vegetables and certain plant oils. Long-chain menaquinones can be found in fermented foods such as cheeses or in natto, a soybean product popular in Japan. The menaquinones are synthesized by gut bacteria but do not contribute appreciably to vitamin K status in humans.(13) Whereas there is no evidence of change in coagulation times in healthy adults with mild vitamin K deficiency, usual dietary intakes of vitamin K do not support complete carboxylation of osteocalcin.(14) Vitamin K restriction results in an increase in serum uncarboxylated osteocalcin, and supplementation for 2 wk with 400 μg/d K1—equivalent to 2 cups of cooked broccoli—results in a 40% decrease in uncarboxylated osteocalcin.(15) In fact, uncarboxylated osteocalcin (ucOC) in serum is the most sensitive indicator of overall vitamin K nutritional status. It should be pointed out that, because the amount of ucOC is highly correlated to total OC and total osteocalcin reflects osteoblastic activity, the ucOC fraction is expressed as %ucOC.

So, what are the implications of vitamin K status for bone health? Papers regarding direct effects of warfarin therapy on BMD in humans are conflicting. Increased fracture risk has been associated with anticoagulant therapy but primarily in studies of the elderly in a long-term care setting.(16–18) Such studies are problematic because of frequent confounding comorbidities. Within the last 20 yr, a preponderance of studies has advanced the notion that low dietary phylloquinone intake, as assessed by food frequency questionnaires, low circulating vitamin K levels, or high %ucOC is associated with low BMD and increased hip fracture risk.(19) Although most of these studies were relatively small, they have been corroborated by several larger ones. In the Nurses Health Study, risk for hip fracture was significantly greater in the lowest quintile of vitamin K intake; however, a relatively narrow range of vitamin K intake was noted in the entire cohort.(20) In the Framingham Offspring Study, dietary phylloquinone intake was associated with total hip and lumbar spine BMD in women but not in men.(21) All of these studies are hampered by the inability to separate the effects of a single nutrient from those of an overall healthy diet. In fact, high fruit and vegetable consumption was associated with less age-related bone loss in the Framingham Heart Study.(22)

Finally, pharmacologic doses of MK4 (menatetrenone) have been studied in Japan as a treatment for osteoporosis. This is derived from the observation that the incidence of osteoporotic fracture is relatively lower in regions of the country where dietary intake of natto is high.(23) There are also some suggestions that the anti-osteoporotic effects of MK4 are independent of γ-carboxylation.(24) In a recent meta-analysis of 13 small randomized controlled trials with data on bone loss, the authors concluded that supplementation with either MK4 or K1 has a positive effect on BMD. Further analysis, including only the trials that provided both BMD and fracture data, concluded that supplementation with 45 mg/d of menatetrenone was associated with increased BMD and reduced fracture incidence.(25) However, the authors of the meta-analysis raised concerns that the trials were often poorly designed. Published commentary have also questioned the findings.(1,19,26–28)

Results of several placebo-controlled clinical trials of vitamin K1 are now being reported. One(29) showed a modest positive effect of 200 μg/d of K1 for 2 yr on BMD at the distal radius but not at any other skeletal site. Another provided 5000 μg/d of K1 for 2 yr but did not find a beneficial effect on bone turnover markers or BMD at any site, including the distal radius.(30)

In the well-designed study in this issue of JBMR, Binkley et al. supplemented 381 postmenopausal women with either 1 mg K1 daily, 15 mg menatetrenone three times a day, or corresponding placebos for 1 yr. All were provided calcium and vitamin D. The dosing reflects the previous study of Binkley et al.,(31) in which 1 mg/d of phylloquinone was required to fully carboxylate osteocalcin. A particular strength of this study is the head-to head comparison of optimal doses of K1 and MK4. The vast majority of studies of MK4 have been in Japanese subjects, and its potential efficacy in Western societies has not been explored. To maximize the potential benefit of vitamin K, Binkley et al. selected only women who were not vitamin K replete as defined by an elevated %ucOC. Outcome measures included BMD of the lumbar spine and proximal femur as well as hip geometry and bone turnover markers. In both treatment groups, %ucOC decreased by 60% after 1 mo of supplementation and remained stable thereafter, verifying compliance, effectiveness of dose, and inadequate vitamin K nutritional status at baseline. Compared with placebo, there were no differences in bone-specific alkaline phosphatase or N-telopeptide (NTX), nor were there differences in BMD of the lumbar spine or total femur in either treatment group. These results are similar to those observed by Booth et al.(26) in another, recently published, 3-yr duration placebo-controlled trial of 500 μg/d of K1.

It has been suggested that vitamin K may improve bone quality and microarchitectural bone strength. For example, femoral neck width was significantly greater in women receiving MK4, 45 mg/d for 3 yr, than in a placebo group.(32) Addressing these considerations, this study observed similar speed of sound or broadband ultrasound attenuation, femoral neck BMC, BMD, area, diameter, cross-sectional area, cross-sectional moment of inertia, length, and strength index across all study groups.

The authors acknowledge limitations in their study such as the relatively short 1-yr duration, the inclusion of only healthy women, and the lack of fracture incidence or reduction as outcome measures. It would seem from the lack of an effect that extremely large numbers of subjects would be necessary to detect differences in fracture rate. In an as yet unpublished trial carried out by Eisai in 3000 elderly subjects, there was no benefit of 3 yr of 45 mg/d MK4 plus calcium in subjects without a history of fracture, and only a small effect in those with a history of more than five vertebral fractures at baseline.(33)

In summary, the absence of change in bone turnover markers or BMD in this and other studies suggests that it is unlikely that vitamin K supplementation will serve to prevent fractures in individuals who are receiving calcium and vitamin D supplementation.

Nevertheless, investigators in the bone field cannot forget about vitamin K yet. Last year, the group led by Gerard Karsenty published evidence that osteocalcin is involved in the endocrine regulation of energy metabolism in mice.(34) This is a novel and exciting concept that challenges much of what we know about osteocalcin. Even more provocative are their data suggesting that, with respect to regulation of energy homeostasis, the active form of osteocalcin is uncarboxylated and that the carboxylated form is inactive. Animals with higher levels of uncarboxylated osteocalcin were protected from obesity and glucose intolerance. To determine whether this effect is evident in humans as well, Yoshida et al.(35) recently re-examined the available data in their study of the effect of vitamin K on bone loss, finding that supplementation with 500 g of phylloquinone daily for 36 mo had a protective effect on the progression of insulin resistance in older men. Because increased vitamin K intake decreases uncarboxylated osteocalcin, these results suggest a possible species difference between humans and mice in this regard. These data clearly indicate the need for further research and better understanding of the relationship between osteocalcin, its carboxylation, and vitamin K.