Vitamin D Doesn't Lower Mortality for Most People. For the Deficient, It May Be a Different Story
Two huge trials found no mortality benefit from vitamin D, yet that may not hold if you start out low. Why baseline status matters, how to test, and a simple decision flow.
Two headlines about vitamin D can both be true, and most of us only ever see one of them. One says that two of the largest trials ever run found vitamin D pills did not reduce deaths. The other, quieter one says that for people who are actually low, the picture may look very different. If you stood in a supplement aisle this month feeling vaguely foolish for owning a bottle, or vaguely guilty for not owning one, the gap between those two sentences is where your answer lives.
A caution first. I am a software engineer, not a clinician, and I am reading published abstracts and summaries here, not treating anyone. Use this to prepare questions for a doctor, not to replace one.
Why two huge trials found "no benefit"
The two trials in question are VITAL, run in the United States, and D-Health, run in Australia. Both were large, randomized, and placebo-controlled, which is about as good as this kind of evidence gets. VITAL gave adults a daily dose of vitamin D3; D-Health gave older Australians a large monthly dose. Both followed people for roughly five to six years, and neither found that supplementing lowered the chance of dying from any cause.
That finding is solid. It is also easy to misread. A trial answers the question it asks, and these asked something specific: if you hand vitamin D to a big group of ordinary adults, do fewer of them die? The answer, in those groups, was no.
The word that matters is "ordinary". Most participants already had enough vitamin D in their blood when they joined. Giving a nutrient to people who are not short of it is a bit like topping up a full tank. You can pour in as much as you like and the car will not drive any further.
Baseline status changes the story
A February 2026 paper in Clinical Nutrition by Wang and colleagues went back to this question using UK Biobank data. Instead of running a new trial, the authors built what is called an emulated trial: they used observational records to estimate what raising blood vitamin D by the same amounts as the two trials might do, over the same follow-up periods.
Two things stand out. When they weighted the population to look like the trial participants, mostly sufficient, they reproduced the null results. The hazard ratios sat essentially at 1.0, about 0.97 and 1.02, which means no detectable difference.
Then they restricted the analysis to people who started with insufficient vitamin D, and again to people who started deficient. The expected mortality reduction became large. The hazard ratios for insufficiency came out around 0.81 to 0.85, and for deficiency around 0.75 to 0.79. In plain terms, that is roughly a fifth to a quarter fewer deaths over the follow-up period in the deficient groups.
Now the honest part. An emulated trial is not a randomized trial. It leans on observational data, so unmeasured differences between people can sneak in. People with low vitamin D often differ from others in ways that also affect survival: less exercise, more illness, less time outdoors. The authors themselves treat this as a pointer, not a proof. I would read it as a good reason to run a proper trial in deficient people, and a strong reason to stop reading "no benefit" as "no benefit for anyone".
A separate paper from July 2026 in Nutrients, drawing on D-Health, found no reduction in hospitalizations for infection among an older group that was not deficient. It fits the same pattern. Replete people do not seem to gain much from extra.
How to actually know your vitamin D status
You cannot feel your vitamin D level. Fatigue, low mood, and aching bones are common to a hundred other conditions, and a deficiency can sit quietly for years. The way to know is a blood test for 25-hydroxyvitamin D, written as 25(OH)D.
Results appear in either ng/mL or nmol/L. To convert, multiply ng/mL by 2.5 to get nmol/L. The thresholds below come from the two most-cited sources I am aware of, and they do not fully agree, which is part of the difficulty.
| Source | Deficient | Insufficient or inadequate | Sufficient |
|---|---|---|---|
| Endocrine Society (older guideline) | below 20 ng/mL (50 nmol/L) | 21 to 29 ng/mL (52 to 72 nmol/L) | 30 ng/mL (75 nmol/L) or above |
| National Academies (Institute of Medicine) | below 12 ng/mL (30 nmol/L) is a deficiency risk | 12 to 19 ng/mL may be inadequate for some | 20 ng/mL (50 nmol/L) or above covers most people |
Because the bodies disagree, a result of 24 ng/mL could be labelled "insufficient" by one and "sufficient" by another. This is exactly why the number should be read by someone who knows your history. Very high levels, above roughly 50 ng/mL (125 nmol/L), are also worth avoiding.
Who should think about testing? The groups usually named are people with little sun exposure, people with darker skin living far from the equator, people with obesity, people with conditions that impair absorption such as coeliac disease or after some weight-loss surgery, people with chronic kidney or liver disease, and some people on long-term medicines that affect vitamin D. As I read recent guidance, routine testing of healthy adults with no risk factors is generally not recommended, which fits the logic above.
Why population advice and personal advice can both be right
This is the part that confuses people, so let me make it concrete. Suppose a hundred people take a pill and the average outcome barely moves. That average can hide two different groups: a large one with no change, and a small one with a real benefit. Both statements are true at once. Population guidance speaks to the average. Your doctor, with your blood result in hand, speaks to you.
Public health bodies set broad intake guidance to protect the people who would otherwise slip into deficiency, because they cannot test everyone. That is a sensible rule for a population, and it is not the same as saying every individual needs a bottle on the shelf.
What this says about "everyone should take vitamin D"
The wellness industry loves a universal recommendation, because a universal recommendation sells to everyone. The trial evidence does not support it. Taking a modest dose is generally low-risk, so this is not a warning against pills. It is a reason to ask what you are paying for.
The picture I come away with is that vitamin D is a deficiency treatment more than a general tonic. It does a real job when the tank is low and very little when it is full. A blood test costs far less than years of guessing, and it turns a vague hope into a decision.
A simple decision flow
Here is the sequence I would bring to an appointment. Treat the dosing as background, not a prescription, because the right amount depends on your result, your age, and your health.
- Test first. Ask whether a 25(OH)D test makes sense for you, particularly if you fit one of the higher-risk groups above.
- Read the number against the table. If you are at or above about 20 ng/mL (50 nmol/L) by the stricter standard, or 30 ng/mL (75 nmol/L) by the Endocrine Society's, you are generally in the sufficient range. If so, skip the extra supplement and revisit if your circumstances change.
- If you are deficient or insufficient, treat it with guidance. General intake recommendations are 600 IU (15 micrograms) a day for most adults up to 70 and 800 IU (20 micrograms) after that, with an upper limit of 4,000 IU (100 micrograms) a day. Correcting a true deficiency can call for more than the everyday amount, for a limited period, and your clinician should set it.
- Retest. Check again after a few months to confirm the level has moved into range, then settle on a maintenance amount.
Food and sunlight still count. Oily fish, eggs, and fortified foods contribute, and sensible sun exposure helps, though how much depends on where you live, your skin, and the season.
Frequently asked questions
If the trials found no benefit, should I stop taking vitamin D?
Not necessarily, and not on the strength of a headline. If you were prescribed it, or know you were low, keep taking it until you have spoken to your doctor. If you take a small dose for general reasons and you have never been tested, a blood test will tell you whether it is doing anything.
Does the new analysis prove vitamin D lowers deaths in deficient people?
No. It used observational data to emulate a trial, so it suggests a benefit without proving one. It is a reason for a proper randomized trial in deficient people.
Which lab range should I trust?
Groups disagree, so the answer is whichever your clinician uses along with your history. Between roughly 20 and 30 ng/mL, the label depends on the source, and your risk factors decide how much weight to give it.
Can I get too much vitamin D?
Yes, though it usually takes high-dose supplements over a long time. Excess can raise blood calcium and cause problems with the kidneys. That is another reason to test before taking large amounts.
Is sunlight a substitute for supplements?
For some people, for part of the year. Latitude, season, skin tone, and time outdoors all change how much you make, which is why a blood test is more reliable than guessing from your habits.
The shelf of bottles will still be there tomorrow. A single test result might be the more useful thing to carry out of the building.