Vitamin K2 for Osteoporosis

The case for vitamin K2 in osteoporosis rests mostly on menatetrenone (MK-4) at 45 mg a day, a dose several hundred times the 90-120 mcg adequate intake for adults and used as a treatment for osteoporosis in Japan (NIH ODS). A 2006 meta-analysis of seven Japanese MK-4 trials found much lower odds of vertebral fractures (OR 0.40) and hip fractures (OR 0.23) (Cockayne 2006). A UK review judged the menatetrenone trials, including the large OF study, open-label and poorly reported (Stevenson 2009), and the OF study itself, in 4,378 women, found no reduction in new vertebral fractures over 36 months (Inoue 2009).

An updated 2019 systematic review, done partly because the integrity of some earlier evidence had been questioned, tested its results against risk of bias and found no evidence that vitamin K affects bone density or vertebral fractures in postmenopausal or osteoporotic patients, and a possible reduction in clinical fractures that the evidence could not confirm (Mott 2019). Two 2022 reviews reported small lumbar spine bone density gains (Ma 2022; Zhou 2022), with no clear fracture benefit in Ma 2022. The over-the-counter form, MK-7 at microgram doses, slowed bone loss in one 3-year trial of healthy women (Knapen 2013) but not in two other placebo-controlled trials (Emaus 2010; Rønn 2021). For comparison, see vitamin D for osteoporosis and calcium for osteoporosis.

Vitamin K is the cofactor that carboxylates osteocalcin (OC), a protein that can bind hydroxyapatite (bone mineral) to promote mineralisation (Ma 2022). When vitamin K is short, more osteocalcin circulates in its undercarboxylated form (ucOC), so the ucOC to carboxylated osteocalcin ratio is used as a marker of vitamin K status (Knapen 2013).

Vitamin K2 is a family of menaquinones. MK-4 is also made in the body from vitamin K1, and MK-7, the form in natto and a common supplement form, has a longer half-life than vitamin K1 (NIH ODS). Its longer half-life and greater potency are why MK-7 was tested at low microgram doses (Knapen 2013).

In the bone trials reviewed here, vitamin K2 lowered undercarboxylated osteocalcin, including in trials where bones did not change. MK-7 at 360 mcg lowered ucOC without changing bone loss (Emaus 2010); MK-7 at 375 mcg cut ucOC by 65.2% at 1 year, yet bone density fell in both groups over 3 years with no significant difference (Rønn 2021); and MK-4 at 45 mg lowered ucOC without changing bone density or turnover (Binkley 2009).

Adding vitamin K2 to risedronate dropped ucOC from 5.81 to 2.59 ng/mL, yet fracture rates were the same as with risedronate alone (Tanaka 2017). A better ucOC result shows the vitamin is being absorbed; it is not evidence of stronger bones.

The fracture trials used MK-4 (menatetrenone) at 45 mg a day (Shiraki 2000; Tanaka 2017); one trial in the 2006 meta-analysis used 15 mg a day (NIH ODS). That is a pharmacological dose, not a nutritional one.

Trials of MK-7, a common supplement form, used 180 mcg a day (Knapen 2013), 360 mcg as natto capsules (Emaus 2010) and 375 mcg (Rønn 2016; Rønn 2021). Most trials gave vitamin K2 on top of calcium and vitamin D. Vitamin K is fat-soluble and is absorbed along with dietary fats (NIH ODS), so a meal that contains some fat is a reasonable time to take it.

In an open-label trial of 241 patients with osteoporosis, 45 mg a day of MK-4 for 2 years led to fewer clinical fractures than no treatment (p = 0.0273), and lumbar bone density fell 0.5% versus 3.3% (p = 0.0339) (Shiraki 2000). Pooling seven such trials, a 2006 meta-analysis found odds ratios of 0.40 for vertebral fractures, 0.23 for hip fractures and 0.19 for nonvertebral fractures (Cockayne 2006).

A UK health technology assessment judged that all four menatetrenone trials it included were poorly reported and three were very small, with fewer than 100 people in each group (Stevenson 2009). None of the four used a placebo: the comparators were no treatment, etidronate or calcium.

The OF study randomised 4,378 postmenopausal women with osteoporosis to calcium alone or calcium plus menatetrenone. New vertebral fractures over 36 months, the primary endpoint, did not differ, and clinical fractures over 48 months were not significantly lower. Only in a subgroup with five or more vertebral fractures at the start were new vertebral fractures lower (Inoue 2009).

In JOINT-03, 1,874 women aged 65 or older took risedronate with or without 45 mg a day of vitamin K2 for 2 years. Fracture rates were similar (incidence rate ratio 1.074, 95% CI 0.811-1.422) and bone density rose equally in both groups (Tanaka 2017). In a smaller placebo-controlled trial, adding MK-4 to bisphosphonate treatment for 18 months did not change bone density or bone turnover (Moore 2023).

The 2019 update, covering 36 trials and testing its results against risk of bias, found that for postmenopausal or osteoporotic patients the odds of any clinical fracture were lower with vitamin K (OR 0.72, 95% CI 0.55-0.95), but the result lost significance when limited to trials at low risk of bias (OR 0.76, 95% CI 0.58-1.01). Vertebral fractures did not differ (OR 0.96), and small bone density gains shrank when high-risk trials were removed (Mott 2019).

Reviews limited to menatetrenone or vitamin K2 found improved lumbar spine bone density but no clear fracture benefit: a non-significant vertebral fracture RR of 0.87 (Su 2019), and overall fracture RRs of 0.63 (p = 0.08) and 0.96 (p = 0.65) that became significant only after one heterogeneous study was dropped (Huang 2015; Ma 2022). A review of long-term trials reported a 2.17-point greater percentage change in lumbar bone density (Zhou 2022). The bone density benefit in Huang 2015 appeared only in women with osteoporosis.

In 244 healthy postmenopausal women, MK-7 at 180 mcg a day for 3 years slowed the decline in bone density at the lumbar spine and femoral neck (not the total hip) and reduced loss of vertebral height in the lower thoracic spine (Knapen 2013). In 334 healthy Norwegian women early after menopause, 360 mcg a day for a year had no effect on bone loss (Emaus 2010).

In 142 women with osteopenia, 375 mcg a day on top of calcium and vitamin D did not change bone density, turnover or microarchitecture over 3 years (Rønn 2021), although a tibia microarchitecture benefit had been seen at 1 year (Rønn 2016). In a 381-woman trial comparing vitamin K1, MK-4 at 45 mg a day and placebo in healthy North American women, neither form changed spine or hip bone density over a year (Binkley 2009), while a 3-year MK-4 trial found no change in bone density but better femoral neck width and strength indices (Knapen 2007).

No tolerable upper intake level has been set for vitamin K because of its low potential for toxicity (NIH ODS). At the 45 mg menatetrenone dose, though, side effects were more common than with placebo: adverse events RR 1.47 and adverse drug reactions RR 1.29, with no difference in serious adverse events (Su 2019). In the OF study, adverse events were more frequent with menatetrenone (Inoue 2009), including more skin and skin appendage lesions (Stevenson 2009), and in JOINT-03 more people stopped treatment (10.0% vs 6.7%) (Tanaka 2017).

Not every review agrees: one found no difference in adverse reactions (RR 1.03) (Ma 2022). The main real-world risk is not toxicity but interference with warfarin-type anticoagulants.

The main interaction is with vitamin K antagonists such as warfarin. In healthy volunteers on acenocoumarol, 45 mcg of MK-7 a day cut the INR by about 40%, and 10 mcg a day caused a clinically relevant drop in at least 40% of participants, so the authors advise avoiding MK-7 supplements during vitamin K antagonist therapy (Theuwissen 2013). People on these drugs should keep vitamin K intake consistent (NIH ODS).

Antibiotics, bile acid sequestrants such as cholestyramine, and orlistat can lower vitamin K status, although orlistat does not usually have a clinically significant effect on it (NIH ODS). Adding vitamin K2 to risedronate brought no unknown adverse drug reactions but also no extra fracture protection (Tanaka 2017).