Vitamin D Supplements: What the Big Trials Actually Found

Vitamin D is a genuinely essential nutrient, and genuine deficiency causes genuine disease. But the bigger promise of the last fifteen years — that a daily pill would cut cancer, heart attacks, fractures and early death across the general population — has now been tested properly, and it mostly did not happen. VITAL randomised 25,871 US adults to 2,000 IU a day and found no reduction in cancer or major cardiovascular events. D-Health randomised 21,315 older Australians and found no reduction in death. The crucial catch: these trials largely enrolled people whose vitamin D was already adequate. That is a very different question from treating deficiency.

Why vitamin D got so oversold

The hype was not invented from nothing. Observational studies consistently found that people with low blood levels of 25-hydroxyvitamin D — the storage form doctors measure — had more cardiovascular disease, cancer, diabetes and autoimmune disease, and shorter lives.

The problem is that low vitamin D is also an excellent marker of being unwell, indoors, inactive, older or overweight. Sick people go outside less, and inflammation itself lowers circulating 25(OH)D. A correlation with almost every disease is exactly what you would expect from a marker of general ill health — whether or not the vitamin does any causal work. Only randomisation settles that.

The trials that changed the picture

VITAL: 25,871 adults, 2,000 IU a day, five years

VITAL was a nationwide, randomised, placebo-controlled trial with a two-by-two factorial design, published by Manson and colleagues in the New England Journal of Medicine. It enrolled 25,871 participants — men aged 50 and over, women aged 55 and over, including 5,106 Black participants — and gave them vitamin D3 at 2,000 IU per day, marine omega-3s at 1 g per day, both, or neither. The co-primary endpoints were invasive cancer of any type and major cardiovascular events.

Over a median 5.3 years, neither primary endpoint moved. Cancer was diagnosed in 1,617 participants: hazard ratio 0.96 (95% CI 0.88 to 1.06, P=0.47). Major cardiovascular events occurred in 805 participants: HR 0.97 (95% CI 0.85 to 1.12, P=0.69). Death from any cause: HR 0.99 (95% CI 0.87 to 1.12).

One secondary signal deserves naming, because supplement marketing quotes it. Death from cancer showed a hazard ratio of 0.83 (95% CI 0.67 to 1.02) across 341 deaths. That interval crosses 1.0, it was one of many secondary endpoints, and a lone secondary result in a trial that missed both primary endpoints is a hypothesis, not a finding.

The VITAL fracture ancillary: the result that stung

Bone was supposed to be the safe ground. LeBoff and colleagues, also in NEJM (2022), ran the fracture ancillary within the same 25,871 VITAL participants, confirming 1,991 incident fractures in 1,551 people via centralised medical-record adjudication.

Vitamin D3 did not reduce total fractures (769 of 12,927 on vitamin D versus 782 of 12,944 on placebo; HR 0.98, 95% CI 0.89 to 1.08, P=0.70), nonvertebral fractures (HR 0.97, 95% CI 0.87 to 1.07) or hip fractures (HR 1.01, 95% CI 0.70 to 1.47, P=0.96). There was no modification of the effect by age, sex, race, body-mass index or baseline 25(OH)D level.

The paper is explicit that participants were not recruited on the basis of vitamin D deficiency, low bone mass or osteoporosis — which is the whole point of what the result does and does not say.

D-Health: 21,315 Australians, monthly high-dose vitamin D

The D-Health Trial, published in Lancet Diabetes & Endocrinology (2022), took a different approach: 21,315 Australians aged 60 and over, recruited from the Commonwealth electoral roll and randomised to 60,000 IU of vitamin D3 or placebo once a month for five years, with all-cause mortality as the primary endpoint.

The dosing worked biochemically — mean serum 25(OH)D reached 115 nmol/L versus 77 nmol/L on placebo — but it did not translate. Over a median 5.7 years, all-cause mortality had a hazard ratio of 1.04 (95% CI 0.93 to 1.18, P=0.47). Cardiovascular mortality: HR 0.96 (95% CI 0.72 to 1.28). Cancer mortality: HR 1.15 (95% CI 0.96 to 1.39, P=0.13), and in an exploratory analysis excluding the first two years, HR 1.24 (95% CI 1.01 to 1.54, P=0.05).

The investigators were careful about that last number — exploratory and hypothesis-generating, not proof of harm — but concluded that on precautionary grounds this monthly high-dose regimen might not be appropriate in people who are already replete.

D2d: the closest thing to a near miss

Not everything came back flat. The D2d trial (NEJM, 2019) randomised 2,423 adults with prediabetes to 4,000 IU per day or placebo, regardless of baseline level. Over a median 2.5 years, new-onset diabetes occurred in 293 versus 323 participants — HR 0.88 (95% CI 0.75 to 1.04, P=0.12). Formally negative, but the direction and magnitude were enough that the Endocrine Society later flagged prediabetes as one of the few places empiric supplementation may be worth it.

The interpretive point almost everyone skips

Here is the sentence that reconciles the headlines. In VITAL, mean baseline 25(OH)D was 30.8 ng/mL; in the fracture ancillary, just 2.4% of participants were below 12 ng/mL and 12.9% below 20 ng/mL. In D-Health, the placebo group averaged 77 nmol/L, comfortably adequate. In D2d, mean baseline was 28.2 ng/mL.

These trials therefore tested one specific question extremely well: does adding more vitamin D to people who already have enough produce broad health benefits? The answer is a well-powered no. They did not test whether correcting real deficiency helps, because they mostly did not enrol deficient people. Applying a blanket “vitamin D does not work” to someone with a level of 9 ng/mL is a straightforward misreading.

Where vitamin D unambiguously works

Vitamin D regulates calcium and phosphate absorption and is required for normal bone mineralisation. When it is genuinely absent, the consequences are neither subtle nor controversial:

• Rickets in children — soft, deforming bones. Preventing it is why vitamin D is added to milk and infant formula in the first place.
• Osteomalacia in adults — defective mineralisation causing bone pain, muscle weakness and fractures. It reverses with treatment.
• Documented deficiency, where repletion corrects secondary hyperparathyroidism and the bone loss that comes with it.

None of that is in dispute. The dispute is only about whether more is better once you are already sufficient — and there, the trials have answered.

Who is actually at risk of deficiency

The NIH Office of Dietary Supplements lists the groups where inadequacy is genuinely common:

• People with little sun exposure — housebound or institutionalised people, night-shift workers, those who cover most of their skin, and anyone at high latitude through winter, when UVB is too weak for skin synthesis.
• People with higher skin melanin at high latitude. Melanin is effective UV protection, so substantially more sun exposure is needed for the same skin synthesis — a real disadvantage in a northern winter.
• Older adults. Skin synthesises less vitamin D with age, and time outdoors typically falls too.
• Fat malabsorption conditions — coeliac disease, Crohn’s disease, cystic fibrosis, pancreatic insufficiency. Vitamin D is fat-soluble, so absorption tracks fat absorption.
• People with obesity, or after gastric bypass. Vitamin D is sequestered in adipose tissue, and bypass takes part of the upper small intestine out of the loop.
• Exclusively breastfed infants, which is why supplementation is routinely recommended.
• People on certain medications. ODS flags orlistat (reduced absorption), corticosteroids (impaired metabolism), statins, and thiazide diuretics, where combining with vitamin D raises hypercalcaemia risk.

This article is educational and not medical advice. Whether to test, whether to treat, and at what dose are decisions for a clinician who knows your history and your medications.

What a vitamin D test does and does not tell you

The standard test measures serum 25-hydroxyvitamin D. The NIH ODS thresholds are: deficiency below 30 nmol/L (12 ng/mL); inadequacy from 30 to under 50 nmol/L (12 to under 20 ng/mL); adequacy at or above 50 nmol/L (20 ng/mL); and possible adverse effects above 125 nmol/L (50 ng/mL).

What the number does not do is give you a target to chase. The 2024 Endocrine Society clinical practice guideline found no clinical trial evidence supporting routine 25(OH)D screening in the general population — including in people with obesity or darker skin — and no clear evidence defining an optimal target level for disease prevention. On that basis the panel suggested against routine testing. That is a recommendation about screening healthy people, not about investigating someone with symptoms, malabsorption or unexplained bone disease, where testing has an actual indication. Assay variability is a further reason not to treat one number as gospel: results for the same sample can differ meaningfully between laboratory methods.

How much you actually need

The US Recommended Dietary Allowances, set by the Institute of Medicine and reported by NIH ODS, are modest:

• Infants 0 to 12 months: 10 mcg (400 IU) — an adequate intake rather than an RDA.
• Ages 1 to 70, including pregnancy and lactation: 15 mcg (600 IU).
• Over 70: 20 mcg (800 IU).

The Endocrine Society guideline layers nuance on top rather than replacing those figures. It suggests empiric supplementation for ages 1 to 18, for adults 75 and over, for pregnancy, and for high-risk prediabetes — and suggests against empiric supplementation above the DRI in healthy adults under 75. For people over 50 for whom vitamin D is indicated, it favours daily dosing over intermittent high doses.

The upper limit, and what mega-dosing actually does

The tolerable upper intake level for adults is 100 mcg (4,000 IU) per day. Frank toxicity is rare and usually requires sustained very high intakes: ODS describes marked hypercalcaemia, with total calcium above 11.1 mg/dL, typically at 25(OH)D above 375 nmol/L (150 ng/mL), with severe cases involving renal failure, soft tissue calcification, arrhythmias and death. It is a supplement problem, not a sun problem — skin synthesis is self-limiting.

Well short of poisoning, very large intermittent doses have produced measurable harm. In a JAMA trial by Sanders and colleagues (2010), 2,256 community-dwelling women aged 70 and over received a single annual 500,000 IU dose or placebo. The vitamin D group fell more (incidence rate ratio 1.15, 95% CI 1.02 to 1.30, P=0.03) and fractured more (IRR 1.26, 95% CI 1.00 to 1.59, P=0.047), with the excess concentrated in the three months after each dose. That is a large part of why modern guidance prefers daily dosing.

Limitations of this evidence

The null trials deserve the same scepticism as the hype. Several caveats genuinely constrain what they prove:

• Replete populations. This is the big one. With mean baseline levels around 30 ng/mL and only about 2% frankly deficient in VITAL, these trials had very little room to demonstrate a repletion benefit even if one exists.
• Background supplement use. Participants in many of these trials could continue their own vitamin D-containing supplements, which narrows the gap between arms and biases results toward the null. The Endocrine Society systematic review names this as a reason optimal doses remain unclear.
• Adherence. VITAL adherence was good but not perfect — about 87% at two years and 85% at five in the fracture ancillary, and 80 to 82% taking at least two-thirds of capsules in the parent trial. Intention-to-treat analysis, which is the right analysis, dilutes any true effect.
• Dose and schedule. 2,000 IU daily, 4,000 IU daily and 60,000 IU monthly are three different interventions. Monthly bolus dosing produces peak-and-trough kinetics that daily dosing does not, and may not be biologically equivalent.
• Event rates and duration. Cancer and cardiovascular disease develop over decades, so five years in mostly healthy adults may be too short a window. The confidence intervals leave room for effects the trials could not rule out — VITAL’s hip fracture HR of 1.01 spans 0.70 to 1.47.
• Generalisability. VITAL enrolled US adults aged 50 and over; D-Health, Australians aged 60 and over. Neither speaks directly to children, pregnancy, or people with malabsorption or chronic kidney disease.

Funding and disclosures

Worth knowing, because it cuts against the usual suspicion. VITAL was NIH-funded, with Pharmavite LLC donating the vitamin D capsules and matching placebos; its fracture ancillary was funded by the National Institute of Arthritis and Musculoskeletal and Skin Diseases. D-Health was funded by Australia’s National Health and Medical Research Council, with two authors declaring unrelated industry grants or honoraria (AstraZeneca; and Amgen, Sanofi, Novartis, Eli Lilly and Alexion) and the rest declaring none. D2d was funded by the National Institute of Diabetes and Digestive and Kidney Diseases. So the trials that found no benefit were largely publicly funded — and one had its pills donated by a supplement maker that presumably hoped for a different answer.

The honest bottom line

Vitamin D is an essential nutrient. Deficiency is real, concentrated in identifiable groups, and cheaply treated. What has collapsed is the far broader claim — that topping up already-adequate people with high-dose vitamin D would prevent cancer, heart disease, fractures and death. Across tens of thousands of randomised participants, it did not.

The reasonable position sits between the extremes. If you are in a higher-risk group, a modest daily dose around the official recommendations is cheap, safe and sensible. If you are not, a mega-dose is unlikely to buy you anything, and very large intermittent doses have a track record of causing harm. And if you suspect actual deficiency, that is a conversation with your doctor — not a bigger bottle.

Sources

• Manson JE, Cook NR, Lee IM, et al.; VITAL Research Group. Vitamin D Supplements and Prevention of Cancer and Cardiovascular Disease. N Engl J Med. 2019;380(1):33-44. PubMed 30415629
• LeBoff MS, Chou SH, Ratliff KA, et al. Supplemental Vitamin D and Incident Fractures in Midlife and Older Adults. N Engl J Med. 2022;387(4):299-309. PubMed 35939577
• Neale RE, Baxter C, Romero BD, et al. The D-Health Trial: a randomised controlled trial of the effect of vitamin D on mortality. Lancet Diabetes Endocrinol. 2022;10(2):120-128. PubMed 35026158
• Pittas AG, Dawson-Hughes B, Sheehan P, et al.; D2d Research Group. Vitamin D Supplementation and Prevention of Type 2 Diabetes. N Engl J Med. 2019;381(6):520-530. PubMed 31173679
• Sanders KM, Stuart AL, Williamson EJ, et al. Annual high-dose oral vitamin D and falls and fractures in older women: a randomized controlled trial. JAMA. 2010;303(18):1815-1822. PubMed 20460620
• Demay MB, Pittas AG, Bikle DD, et al. Vitamin D for the Prevention of Disease: An Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab. 2024;109(8):1907-1947. PubMed 38828931
• Shah VP, Nayfeh T, Alsawaf Y, et al. A Systematic Review Supporting the Endocrine Society Clinical Practice Guidelines on Vitamin D. J Clin Endocrinol Metab. 2024;109(8):1961-1974. PubMed 38828942
• National Institutes of Health, Office of Dietary Supplements. Vitamin D: Fact Sheet for Health Professionals. ods.od.nih.gov