High-Dose Vitamin D Did Not Prevent Falls, and Higher Doses Looked Worse

Vitamin D and falls has been one of the more confusing areas of nutrition research, partly because
trials have used wildly different doses and partly because “more must be better” is such a durable
instinct. STURDY was designed to test the dose question directly.

What the trial did

688 participants, all aged 70 or over, all with serum 25(OH)D
levels between 10 and 29 ng/mL, which is the range where supplementation is most
often considered worthwhile.

They were randomized to one of four daily doses:

• 200 IU/day (control)
• 1000 IU/day
• 2000 IU/day
• 4000 IU/day

The primary outcome was time to first fall or death over two years. Design and main
results were published in Innovation in Aging in 2020.

Note what this design makes possible. Most supplement trials compare one dose against placebo, which
can only answer whether the supplement works. Four active arms can answer a different and more useful
question: whether the amount matters, and in which direction.

The headline result

High-dose vitamin D supplementation at 1000 IU/day or above did not prevent falls.

That is a negative result, and negative results in supplement research deserve more attention than
they get, because they are the ones least likely to be promoted by anyone.

The result that is harder to explain away

The higher-dose groups did not merely fail to do better. They did worse.

Compared with the 1000 IU/day group, event rates were higher in the 2000 IU/day
group (hazard ratio 1.86) and the 4000 IU/day group (hazard ratio
1.68).

Then something informative happened. When participants in the higher-dose arms were switched down to
1000 IU/day, the difference against the control group went away: HR 1.02, P = 0.84.

That pattern, worse on high dose and equivalent once the dose came down, is more suggestive than a
single comparison would be. It is also why the trial’s own language is careful: it notes uncertainty
about whether doses exceeding 2000 IU/day increase fall risk, rather than declaring that they do.

Why researchers hedge on a finding like this

Because harm is a serious claim and the evidence has to carry it.

The higher-dose comparisons were not the trial’s primary question, the number of participants in each
arm was modest, and a hazard ratio derived from a secondary comparison sits on weaker ground than the
primary outcome does. Uncertainty is the honest description of that situation.

What the trial does establish firmly is the negative: at these doses, in this population, more vitamin
D did not produce fewer falls. Anyone taking a high-dose supplement specifically to reduce fall risk is
acting on an expectation this trial tested and did not support.

Who this applies to

Adults aged 70 and over with 25(OH)D between 10 and 29 ng/mL. That is a specific population, and it
is the population most likely to be advised to supplement.

It does not describe people with severe deficiency, who are a different clinical situation, and it
does not describe younger adults. It also says nothing about vitamin D’s other roles, including bone
health, which is a separate question with its own literature.

The general pattern worth noticing

This is the second finding of the same shape to appear in nutrition research recently: a trial
comparing two doses of a supplement, finding that the higher one produced no additional benefit. It
happened with vitamin B12 at 1000 against 2000 micrograms, where both doses improved outcomes and
neither beat the other.

The retail assumption runs the other way. Doses on shelves have drifted upward for years on the logic
that if some is good, more is better, and dose-comparison trials keep failing to find the extra benefit
that logic predicts. STURDY goes further than most by finding a signal pointing the other way, however
tentatively.

Frequently Asked Questions

Does vitamin D prevent falls?

Not at these doses in this trial. STURDY randomized 688 adults aged 70 and over to 200, 1000, 2000 or
4000 IU per day and concluded that high-dose vitamin D supplementation at 1000 IU/day or above did not
prevent falls, with time to first fall or death over two years as the primary outcome. That is a direct
test of the proposition rather than an inference from observational data.

Did higher doses cause more falls?

The data point that way and the trial does not claim it as established. Event rates were higher in the
2000 IU/day group (HR 1.86) and 4000 IU/day group (HR 1.68) compared with 1000 IU/day, and when those
participants were moved to 1000 IU/day the difference against control disappeared (HR 1.02, P = 0.84).
The trial notes uncertainty about whether doses above 2000 IU/day increase fall risk, which is the
appropriate way to describe a signal from a secondary comparison.

Who was in the trial?

688 people aged 70 or over with serum 25(OH)D between 10 and 29 ng/mL, followed for two years. That
range matters: these were people with low-to-marginal vitamin D levels, which is the group most likely
to be told to supplement. It excludes both people with normal levels and people with severe deficiency,
so the result should not be stretched to either.

Should I stop taking vitamin D?

Nothing here supports stopping a supplement your doctor recommended, and vitamin D is prescribed for
reasons beyond falls, including bone health and correcting documented deficiency. What this trial
addresses is one specific expectation: that taking more of it reduces falls in older adults. If that is
why you are taking a high dose, this is a reason to raise it at your next appointment rather than to
change anything unilaterally.

What dose does the trial support?

It does not identify a dose that prevented falls, which is the point. The comparison that looked
cleanest was 1000 IU/day performing better than 2000 and 4000, but performing better in this context
means fewer events than the higher doses rather than fewer than control. Reading it as an endorsement of
1000 IU/day would be reading more into it than the design supports.

Why does the design matter so much here?

Because four active dose arms answer a question a single dose against placebo cannot. A conventional
trial can tell you whether a supplement beats nothing. This one can tell you whether the amount changes
the answer, and in which direction. That is why a null result from STURDY carries more information than
a null result from a two-arm trial would.

How does this fit with other vitamin D research?

It sits alongside a broader pattern of large vitamin D trials producing more modest results than
observational studies predicted. Observational research repeatedly finds that people with higher vitamin
D levels have better outcomes, and randomized trials keep struggling to reproduce those benefits by
raising levels with supplements. That gap usually means the low level was a marker of something else
rather than the cause of it.

Is high-dose vitamin D dangerous?

This trial reports a signal it is careful not to overstate, and that is different from a safety
warning. Vitamin D is fat soluble and does accumulate, and toxicity at very high sustained intakes is
well documented in the clinical literature, which is why upper intake levels exist. The practical point
from STURDY is narrower: taking more of it did not buy the benefit people take it for, and it may not
have been neutral.

Sources

• Design and Main Results of STURDY: A Randomized Clinical Trial of Four Vitamin D3 Doses to Prevent
  Falls in Older Adults. Innovation in Aging, 2020. 688 participants aged 70 and over with serum
  25(OH)D of 10 to 29 ng/mL; doses of 200 (control), 1000, 2000 and 4000 IU/day; primary outcome time to
  first fall or death over two years; hazard ratios of 1.86 and 1.68 for the 2000 and 4000 IU/day groups
  against 1000 IU/day; HR 1.02, P = 0.84 after participants moved to 1000 IU/day; conclusion that
  high-dose vitamin D supplementation at 1000 IU/day or above did not prevent falls —
  Full text
• NIH Office of Dietary Supplements, Vitamin D Fact Sheet for Health Professionals: recommended
  intakes, upper intake levels and toxicity —
  ods.od.nih.gov

Participant numbers, dose arms, primary outcome, hazard ratios
and the trial’s conclusion above are taken from the published report of the STURDY trial, checked August
2026.