In 2010 a randomised trial gave 500,000 IU of vitamin D in a single dose, once a year, to 2,256 women over 70 at high risk of fracture. Double blind, against placebo. Falls rose by 15%, fractures by 26%, and the excess risk was highest in the three months following each dose.
The trial was looking for the opposite, and the idea held together: an annual dose solves adherence, and 500,000 IU spread over a year works out at roughly 1,400 IU per day. An unremarkable dose, except that it was not taken that way.
That is where reasoning in total dose breaks. Between what you swallow and what it produces sits a blood level, and that level depends on more than the number printed on the bottle.
The same dose does not produce the same level
The marker that gets measured is 25(OH)D. The relationship between the dose swallowed and that level is not a straight line, and the gap is striking.
Across 17,614 adults followed in a preventive health programme, the gain was about 12 nmol/L per 1,000 IU in the 0 to 1,000 range, against 1.1 nmol/L per 1,000 IU between 15,000 and 20,000. The same 1,000 IU step returns eleven times less at the top of the curve than at the bottom.
Body weight shifts the whole curve. At the same reported intake, participants with obesity averaged 19.8 nmol/L less than those of normal weight, and overweight participants 8.0 nmol/L less. This is an observational study, not a trial: it describes what happens, it does not demonstrate it. The order of magnitude is still enough to strip the meaning from a single dose meant for everyone.
Schedule matters as much as total
A second trial compared three monthly schedules in 200 people over 70 with a prior fall: 24,000 IU monthly, 60,000 IU monthly, or 24,000 IU alongside calcifediol.
The higher doses did exactly what was asked of them, reaching the threshold of 30 ng/mL of 25(OH)D, that is 75 nmol/L. They brought no benefit to lower extremity function, and they were associated with more falls than the 24,000 IU schedule.
Hitting the target and being better off are therefore not the same thing. It is the most uncomfortable result in this literature, because it attacks the way nearly everyone reasons: aim at a number, reach it, consider the matter settled.
What the large trials did not find
VITAL is the largest trial in the field: 25,871 participants, 2,000 IU daily, a median follow-up of 5.3 years. No reduction in cancer. No reduction in major cardiovascular events. The fracture arm finds nothing either, not on total fractures, not on nonvertebral fractures, not on hip fractures.
One methodological detail changes the entire meaning of those results: participants were recruited neither on vitamin D deficiency nor on low bone mass. VITAL therefore answers the question «should a broadly replete population be supplemented?», and the answer is no. It does not answer «should a deficiency be corrected?», which is a different question with different trials.
Cofactors, sorted by what was tested
| What you read | What was tested | What survives |
|---|---|---|
| Magnesium improves your vitamin D level | Randomised trial in 180 adults, dose matched to each person's intake | It moves the level both ways: up when the starting point is low, down when it is already high |
| Without K2, vitamin D calcifies your arteries | Two years of 360 µg MK-7 in symptomatic coronary patients | No clear gain on the calcium score; an effect on noncalcified plaque of undetermined clinical significance |
| You have to reach 30 ng/mL | Three monthly schedules in 200 older people with a prior fall | The target was indeed reached, and falls went up |
The K2 case deserves to be stated precisely, because it is the most repeated advice in this world. The mechanism is real: vitamin D increases calcium absorption, and vitamin K activates a protein that slows its deposition in arterial walls. The missing step is always the same, the one where somebody tests the question actually being asked, namely whether supplementing vitamin D without K2 damages anything in anyone.
The closest available evidence remains VITAL: 25,871 people on 2,000 IU daily, without K2, for more than five years, with no excess of major cardiovascular events. It is not a direct answer. It is the best there is, and it does not point towards the fear.
The ceiling, and who goes past it
EFSA sets the upper limit at 100 µg per day, that is 4,000 IU. The endpoint is not outright toxicity but persistent hypercalciuria, a lasting excess of calcium in urine considered an earlier sign. It comes from a dose of 250 µg per day at which that effect was observed, divided by an uncertainty factor of 2.5.
The agency adds a sentence aimed rather precisely at the readers of this blog: European populations do not exceed that limit, with the exception of regular users of high-dose supplements.
What this changes in practice
- 1
Measure your 25(OH)D before choosing a dose
It is the only number that is about you. Without it you are choosing for an average you may not belong to.
- 2
Daily rather than bolus
Both trials that found harm had large, widely spaced doses in common.
- 3
Adjust for body weight
The average gap between normal weight and obesity approaches 20 nmol/L at equal intake. One dose cannot suit both.
- 4
Stay under 4,000 IU without medical advice
Past that you leave the European limit, and the gain per extra unit has already become marginal.
- 5
Re-test, then stop thinking about it
One measurement three months after a change is enough to know whether the schedule works for you.
What this article does not say
- That vitamin D is useless. Correcting a documented deficiency and supplementing an already replete population are different questions. VITAL answers the second.
- That K2 is useless. It has not been tested on the question attributed to it, which is not the same as failing.
- That these trials apply to you as they stand. Two of them involve older people at risk of falls, and the dose to level curve comes from an observational study.
- That you can skip medical advice. With kidney disease, sarcoidosis or an ongoing treatment, the dosing question is a different one.
« A dose is an intention. A blood level is a result. »
What Helix does
Vitamin D is the textbook case of the problem Helix solves. The dose sits in your stack, 25(OH)D sits in your blood panel, your weight sits somewhere else, and the useful information only appears when the three are crossed over time. Kept apart, those numbers say nothing.
The engine flags doses approaching the European upper limit, accounts for the curve flattening instead of rewarding the highest dose, and always separates what comes from a trial from what comes from an assumed mechanism.
See what your dose actually produces in you, panel after panel.
Join the beta