Calcium, vitamin D and K2: what the bone trials support

Calcium and vitamin D pills do very little for fractures in ordinary adults. Calcium alone made no reliable difference. Vitamin D alone made none at all. The two taken together lowered the risk of breaking a bone by around 9%, and that was the best result in the set.¹ Vitamin K2 rests on thinner evidence than either.

What the newest count measured

The review gathered the trials that gave adults calcium, vitamin D or both against a placebo or nothing, in people not already taking drugs for osteoporosis.¹ The question it asked was blunt: did they break a bone.

Calcium alone came out at 0.91, meaning 9% fewer fractures, over a range that still includes no benefit at all. Vitamin D alone came out at exactly 1.00, and the reviewers were confident in that flat line. The combination came out at 0.91 again, this time just clear of no effect.¹ Hip fractures, spine fractures and falls told the same story.¹

Two features of these trials shape everything that follows. They mostly enrolled people who were living in their own homes and were unlikely to break anything. And the review says plainly that its evidence about high-risk patients, and people in residential care, was limited.¹ A null result in people who were not going to fracture is not a null result in people who were.

Why earlier reviews reached different verdicts

The same pile of trials has been added up to opposite conclusions, depending on which trials go in.

A 2016 analysis from the National Osteoporosis Foundation reported 15% fewer fractures overall and 30% fewer hip fractures on calcium plus vitamin D.³ Its own stated limitation is that it used data from a subgroup analysis of a large women's health trial.³

A 2019 review split the question by regimen, which turns out to be the useful cut. Vitamin D on its own, at daily or occasional doses anywhere from 400 to 30,000 IU, did not reduce fractures of any kind.² The pair did, slightly. Taking 400 to 800 IU of vitamin D with 1,000 to 1,200 mg of calcium every day — a dairy-heavy diet's worth of calcium, taken as a supplement — cut any fracture by 6% and hip fracture by 16%.²

Calcium plus vitamin D, any fracture, by review% lower risk

2016 National Osteoporosis Foundation15%

2026 BMJ review (the newest count)9%

2019 JAMA Network Open review6%

The same paper also looked at blood levels rather than pills, and there the slope is much steeper. Every 10 ng/mL more vitamin D in the blood went with 7% fewer fractures overall and 20% fewer hip fractures.²

That contrast is the crux of the whole field. People with more vitamin D in their blood break fewer bones. Giving people vitamin D does not reproduce it. A blood level is the residue of a whole life — sunlight, movement, being well enough to go outdoors — and a capsule supplies only the last link in that chain.

Vitamin D alone, and a signal in women

One analysis narrowed to healthy people aged 60 and over with no bone condition. Fractures overall did not differ between vitamin D and placebo, and neither did falls.⁴

One group moved, in the wrong direction. Among the women, hip fractures were more common on vitamin D than on placebo: 164 against 121.⁴ The authors' response was specific rather than sweeping. Avoid large occasional doses of vitamin D, without calcium, in people over 60 whose blood level, bone status and calcium intake are all unknown.⁴

That is one group within one analysis, not an established harm. It is also close to the opposite of what the capsule is sold to do, which is the reason to keep it in view.

The calcium pill question

Calcium supplements carry a safety argument that calcium in food does not. A 2011 paper added up the supplement trials and reported heart attacks up by 27 to 31%, and strokes up by 12 to 20%.⁵ Its argument was about proportion rather than alarm: the fracture benefit is small, the cardiovascular risk is small, and the two may simply cancel.⁵ The same paper noted that calcium from food appears to do as much for bone density with no such signal attached to it.⁵

A 2016 review reached the opposite verdict on the same question. Its trials found no difference in cardiovascular events or deaths between calcium and placebo, and the population studies showed no consistent pattern by dose.⁶ Its conclusion was that calcium up to the tolerable upper limit, 2,000 to 2,500 mg a day, is not associated with cardiovascular risk in generally healthy adults.⁶ That ceiling sits well above the 1,000 to 1,200 mg a day used in the combined-supplement trials, and the review says data above it were scarce.^2,6

In neither case was a heart attack the outcome any trial was built to measure. That is why the disagreement has lasted: two teams reading heart events that were collected on the side.

K2, and how little there is

Here the evidence is real and much weaker than the marketing. In postmenopausal women and osteoporosis patients, vitamin K went with fewer clinical fractures, an odds ratio of 0.72.⁷ Then the reviewers kept only the trials at low risk of bias, and the figure slid to 0.76, over a range that crosses no effect at all.⁷ The finding depends on the weaker trials staying in.

That review exists at all because the earlier version of it was in doubt. After that version was published, the integrity of some of the evidence underneath it was questioned, and an update was required.⁷ The K2 case looks thinner today than it did a decade ago, and this is why.

A separate analysis tested vitamin K alongside calcium rather than on its own. Bone density at the lumbar spine came out higher than in controls, by a small margin: 0.20 on a standardized scale, which is the conventional cut-off for a small effect.⁸ That is a modest change, at one site in the skeleton.

None of it speaks to the product most people mean by K2: a moderate dose added to a vitamin D capsule and taken by somebody whose bones are fine. These trials recruited postmenopausal women and osteoporosis patients, and the reviewers say there is too little evidence to speak about anyone else.⁷

What remains unknown

The open question is who these supplements are for. Almost everything known comes from adults living at home who were not at high risk of fracture, and the newest review is explicit that its evidence for high-risk and residential-care patients was limited.¹ Whether a supplement repairs a genuine dietary shortfall in a frail person is a different question from whether it helps a well-fed one. The trials have mostly answered the second.

The calcium safety question is unresolved rather than settled in either direction, and both reviews that frame it were built from trials never designed to measure heart attacks.^5,6

For K2, there is still no trial of the thing being sold: a moderate dose taken with vitamin D by someone with healthy bones. What exists was measured in patients who already had bone disease, and the fracture signal does not survive a restriction to the best-run trials.⁷