Vitamin D (cholecalciferol)

In brief

What vitamin D is: a prohormone wearing a vitamin's name

Think of the steroid hormones you already know, cortisol, estrogen, testosterone. Each is built from cholesterol, made in one organ, released into the blood, and received by a protein inside the nucleus of distant cells that switches genes on and off. Now take that same cholesterol skeleton of four fused rings, break the second ring open so that the molecule can hinge in the middle, and you have a secosteroid, which is what vitamin D is ^2,3. The break is made by ultraviolet light in skin, the hinged product is carried to the liver and then the kidney for two finishing touches, and the finished molecule, calcitriol, is received by a nuclear receptor from the same gene family as the receptors for cortisol and thyroid hormone ^4,6,7,10. By every criterion an endocrinologist uses, that is a hormone. The word vitamin was attached before any of this was known ².

The formula on this page, C27H44O, describes vitamin D3, cholecalciferol, the form animals make. Its side chain is cholesterol's eight-carbon tail, which branches where it joins the ring and again at its far end; the single oxygen is the hydroxyl on carbon 3 that cholesterol also carries; the broken bond between carbons 9 and 10 is the mark of the light, and the double bond that juts out of the ring from carbon 10 to carbon 19 is the mark of the heat-driven rearrangement of previtamin D3 that follows ^2,4,5. Vitamin D2, ergocalciferol, is the plant and fungal form, derived from ergosterol, and carries an extra methyl group and an extra double bond in the side chain; a historian of the field describes the two as having roughly equal potencies, similar metabolic patterns and identical effects, a judgment the trials comparing them have since qualified ^2,3,42.

What the body actually uses is two steps away. The liver enzyme CYP2R1, a member of the cytochrome P450 family of iron-containing enzymes that add oxygen to steroids, drugs and fats (CYP is the family's prefix; the letters and numbers name the individual enzyme), adds a hydroxyl at carbon 25 to make 25-hydroxyvitamin D, calcifediol, the storage and transport form and the one the blood test measures ⁶. The kidney enzyme CYP27B1 adds a hydroxyl at carbon 1 to make 1,25-dihydroxyvitamin D, calcitriol, the hormone, with the formula C27H44O3 ^1,7. A third enzyme, CYP24A1, breaks both down ³. The receptor, loaded with hormone, binds to sites in DNA called vitamin D response elements and regulates hundreds of genes in a cell-specific fashion ³. None of this is what the word vitamin, a vital amine, was coined to mean. Funk coined it for thiamine, the factor that cures neuritis, which really is an amine; vitamin D is not an amine and, when the sun is available, is not required in the diet at all ².

The confusion is not a quibble about names. A vitamin is something the body cannot make and must eat, in which case more in the diet is the only lever, and the dose that cures deficiency is the dose. A hormone is made on demand, carried by a binding protein, activated and inactivated by regulated enzymes, and held at a set point by feedback. For a hormone, the question is never simply how much is in the bottle but what the system does with it, and for vitamin D the system is designed to make the active hormone's level independent of the precursor's supply over a wide range ^8,9,28. Nearly every reversal on this page follows from that single difference.

Made in skin: the chemistry of a sunburn's useful wavelengths

The precursor is 7-dehydrocholesterol, the last intermediate on the way to cholesterol, and skin keeps a supply of it. When ultraviolet B light of the right energy strikes it, the B ring opens and previtamin D3 forms, throughout the epidermis and into the dermis in lightly pigmented skin ⁴. The previtamin is unstable in a useful way: by heat alone, with no enzyme, it rearranges into vitamin D3 over a period that takes at least three days to complete, and the vitamin D binding protein in blood preferentially picks out the vitamin rather than the previtamin and carries it into the circulation. The skin therefore works as a store that releases vitamin D3 slowly and steadily after a single exposure ⁴.

The wavelengths that do the work are narrow. In pieces of human skin exposed to narrow-band radiation, the optimum for making previtamin D3 lay between 295 and 300 nanometers; at 295 nanometers up to 65 percent of the 7-dehydrocholesterol was converted, whereas under simulated solar radiation the maximum was about 20 percent, because sunlight's broader spectrum pushes the previtamin on into two inert photoproducts, lumisterol and tachysterol ⁵. That photochemical ceiling helps explain why sunlight is not among the causes of vitamin D toxicity that the clinical reviews list, which are overdose of pharmaceutical preparations, diseases that make the hormone outside the kidney and inherited failures to break it down ^5,35, and it is the first piece of evidence that the body treats the molecule as a regulated hormone rather than a nutrient to be accumulated.

Pigment controls the rate. In a 1982 experiment, two lightly pigmented and three heavily pigmented volunteers received a single standard dose of ultraviolet radiation, one minimal erythemal dose. Serum vitamin D rose by up to 60-fold at 24 to 48 hours in the lightly pigmented subjects and did not change significantly in the heavily pigmented ones; when one of the latter was re-exposed to six times the dose, that volunteer's level rose to what the others had reached at the lower dose ³³. This is the physiological basis of the lower 25-hydroxyvitamin D levels measured in people with dark skin at high latitudes, and, as the section on the binding protein explains, also the beginning of an argument about whether those lower levels mean what they seem to ⁴⁵. The melanin star on this site takes the pigment further.

Sunscreen can block it, and in practice mostly does not. In eight volunteers exposed to one minimal erythemal dose, serum vitamin D3 rose from 1.5 to 25.6 ng/mL in the four who were unprotected and stayed at 5.6 and 4.4 ng/mL in the four protected with para-aminobenzoic acid sunscreen of sun protection factor (SPF) 8, a measure of how much longer protected skin takes to redden than unprotected skin, and the same sunscreen blocked the photoconversion in skin slices in vitro ⁴⁶. A 2019 systematic review found that the four experimental studies agreed with this, but the two randomized field trials found no effect of daily sunscreen at an SPF of about 16, and 69 observational studies mostly found no association or higher levels in sunscreen users; the review's authors noted that the very high protection factors now recommended have never been tested ⁴⁷. The vitamin-d-and-sunscreen and sunscreen-filters stars on this site take the sunscreen evidence further. Fat is a sink: a day after whole-body irradiation the rise in blood vitamin D3 was 57 percent smaller in obese volunteers than in lean controls, with no difference in the 7-dehydrocholesterol content of their skin or its conversion in vitro, and peak serum vitamin D2 after an oral 50,000 IU dose was also lower, with body mass index inversely correlated with both responses (r = -0.55 and -0.56) ³⁴.

Why it is a hormone: the receptor, the enzymes and the feedback

The proof that vitamin D is a hormone came in three parts between 1968 and 1988, and each part was a discovery that something the body makes from the vitamin matters more than the vitamin itself. The first was metabolism. Egon Kodicek's group in Cambridge had spent ten years with radioactive vitamin D and concluded that it acted without being metabolized; their tracer was too weakly labeled to show otherwise ². Hector DeLuca at Wisconsin, Harry Steenbock's last graduate student, made vitamin D3 carrying far more radioactivity per molecule (higher specific activity) and in 1968 isolated and identified 25-hydroxycholecalciferol as a biologically active metabolite made in the liver and present in blood at higher concentration than the vitamin ^2,48,49. In 1970 Fraser and Kodicek showed that a still more water-soluble (more polar) metabolite was made uniquely by the kidney, with animals whose kidneys had been removed unable to make it ^2,50. In 1971 three groups converged: Kodicek's group in Cambridge identified the kidney metabolite as 1,25-dihydroxycholecalciferol ⁵¹; Holick, Schnoes and DeLuca isolated it from chick intestine ⁵²; and Norman's group characterized it by mass spectrometry with the formula C27H44O3 and four to thirteen times the activity of the vitamin at stimulating intestinal calcium transport ^1,2.

The second part was the receptor. In 1969 Haussler, working in Norman's laboratory, reported an acidic protein in the chromatin, the DNA-and-protein material of the cell nucleus, of chick intestinal cells that bound a vitamin D metabolite noncovalently and with high affinity, purified it 167-fold, and showed that its binding capacity was saturated after a physiological dose of the parent vitamin ^2,43. That is the signature of a hormone receptor: limited, saturable, high-affinity binding in the nucleus. The receptor's gene, in the form of complementary DNA, a DNA copy made from the messenger RNA that a cell reads to build the protein, was cloned in full in 1988 by a group that included Haussler, Pike and O'Malley, and its sequence placed it in the steroid-receptor gene family, closest to the thyroid hormone receptor ¹⁰. The same year Hughes and colleagues found that children with hereditary vitamin D-resistant rickets, who are resistant to the hormone rather than short of it, carried single point mutations in the receptor's DNA-binding domain, one family near the tip of the first zinc finger and the other at the tip of the second, so that the receptor bound hormone normally but had decreased affinity for DNA ⁴⁴. A receptor that grips DNA poorly produces the deficiency disease although the hormone is present ⁴⁴. Nothing could show more clearly where the action is.

The third part was the enzymes, and the genetic diseases that go with them. Parathyroid hormone was shown in 1972 to act as a tropin for the kidney enzyme: in rats on a low-calcium diet, thyroparathyroidectomy, removal of the thyroid and parathyroid glands, reduced calcitriol production to negligible levels within 40 hours, and parathyroid extract restored it, independent of the changes in serum calcium and phosphorus ⁸. The activating enzyme, CYP27B1, was cloned in 1997 from the kidneys of mice lacking the receptor, in which it was overexpressed because the hormone could no longer suppress it through the receptor; this established the negative feedback loop at the gene level ⁷. The liver enzyme was confirmed in 2004 when a patient with low 25-hydroxyvitamin D and classic deficiency was found to be homozygous for a mutation that replaced leucine 99 of CYP2R1 with proline and abolished its activity ⁶. And in 2004 the bone hormone FGF23 was shown, in rodents, to lower serum calcitriol within 3 hours of injection, to its nadir at 9 hours, by suppressing the activating enzyme's messenger RNA and inducing the degrading enzyme's from the first hour, with calcitriol in turn raising FGF23 within 4 hours ⁹. Calcium, phosphate, parathyroid hormone, FGF23 and the hormone itself: five regulators, two enzymes, one receptor. That is an endocrine system ^3,11.

The receptor turned out to be nearly everywhere. By 2008 it had been found in at least 36 cell types, and at least 10 organs besides the kidney had been shown to make calcitriol locally, in a paracrine fashion, for their own use rather than for the circulation ¹¹. This is the biology behind every non-skeletal hope for vitamin D: immune cells, breast, prostate, colon and vessel wall all carry the receptor and can make the hormone. Whether giving more precursor to a person who already has enough changes what those tissues do is a separate question, and it is the question the trials were built to answer ¹⁸.

Why the blood test measures the precursor and not the hormone

This is the question that confuses almost everyone who has had the test, including many who order it, so it deserves to be stated at length. The blood test for vitamin D status is 25-hydroxyvitamin D. It is not calcitriol. Calcitriol is the hormone, the molecule that acts, and yet measuring it tells you almost nothing about whether a person has enough vitamin D. There are three reasons, and each is a reason that vitamin D is a hormone.

The first is time. 25-hydroxyvitamin D is the body's reservoir and lasts for weeks: in 36 healthy men in The Gambia and the United Kingdom given tracer doses of deuterium-labeled 25-hydroxyvitamin D2 and D3, the plasma half-lives were 13.9 and 15.1 days respectively, longer in men with more binding protein ¹². In a later pilot study, the half-life of 25-hydroxyvitamin D3 was 13.6 days in nonpregnant women and 18.8 days in pregnant women, a difference that was not statistically significant in that small study, with exposure to the metabolite tracking the binding protein ⁵⁶. Calcitriol's plasma half-life is only several hours ¹³. A measurement of the precursor therefore integrates the last month or two of sun and diet; a measurement of the hormone reports the last few hours of the kidney's decisions.

The second is regulation. The kidney does not make calcitriol in proportion to the supply of its precursor; it makes what parathyroid hormone, FGF23, calcium and phosphate tell it to make. Parathyroid hormone drives the enzyme that makes it, within 40 hours in the rat experiments that first showed it ⁸; FGF23 shuts the enzyme down within hours and switches on the one that degrades the hormone ⁹; and calcitriol suppresses its own synthesis through its receptor ⁷. When vitamin D runs short, the parathyroid glands respond with secondary hyperparathyroidism and the kidney is driven to convert a larger fraction of a smaller supply, so that calcitriol and calcium absorption are held normal until 25-hydroxyvitamin D falls to about 10 nmol/L ^28,29. The hormone level is defended; the precursor level is not.

The third follows from the second and is the fact that matters at the bedside: in vitamin D deficiency short of the most severe, calcitriol is normal. A South Australian laboratory examined the records of 3,661 patients and found 319 with 25-hydroxyvitamin D at or below 40 nmol/L who had had calcium absorption, calcium, parathyroid hormone, bone markers and both metabolites measured. Serum calcitriol and calcium absorption were not reduced until 25-hydroxyvitamin D fell to 10 nmol/L or less; above that, secondary hyperparathyroidism held the hormone up. The authors' conclusion is the sentence to keep: vitamin D deficiency does not reduce serum calcitriol, and therefore calcium absorption, until the serum 25-hydroxyvitamin D falls to approximately 10 nmol/L, at which point the substrate concentration seems to be insufficient to maintain the level of the dihydroxy metabolite despite secondary hyperparathyroidism ²⁸. A doctor who orders calcitriol to look for vitamin D deficiency will usually be falsely reassured; only when 25-hydroxyvitamin D falls to about 10 nmol/L does the hormone fall ²⁸. The test for the hormone is useful for the opposite problem, hypercalcemia from a hormone being made where it should not be, in granulomas, the knots of immune cells that form in diseases such as sarcoidosis and tuberculosis, and in some lymphomas, cancers of the lymph glands, or from an enzyme that cannot break it down ^35,41.

The precursor test has its own troubles. A 2017 international conference on controversies in vitamin D concluded that assay variation had contributed to the chaos surrounding the definition of hypovitaminosis D, having too little vitamin D, recommended the Vitamin D Standardization Program's tools for standardizing 25-hydroxyvitamin D measurement in research, and judged that values below 12 ng/mL (30 nmol/L) should be considered associated with increased risk of rickets and osteomalacia, while 20 to 50 ng/mL (50 to 125 nmol/L) appears safe and sufficient for skeletal health in the general population ¹⁴. The editorial that called 25-hydroxyvitamin D a difficult analyte in 2012 was describing immunoassays and mass spectrometry that disagreed with each other and with themselves ⁵⁷. What the test cannot do at all is distinguish cause from consequence. Autier's 2014 systematic review of 290 prospective cohort studies and 172 randomized trials found moderate to strong inverse associations between 25-hydroxyvitamin D and cardiovascular disease, lipids, inflammation, glucose disorders, weight gain, infections, multiple sclerosis, mood, cognition, physical function and mortality, and no effect of supplementation on disease occurrence, apart from an apparent slight reduction in all-cause mortality with 20 micrograms a day in elderly people, mainly women; its explanation is that inflammatory processes involved in disease lower 25-hydroxyvitamin D, so that a low level is a marker of ill health rather than its cause ²³.

The volume of testing rose anyway. Among Medicare Part B laboratory tests between 2000 and 2010, vitamin D was in the group whose reimbursed volume per 10,000 enrollees grew the most, by factors of 83 to 1,510 across that group ²². A 2024 guideline described widespread supplementation and increased laboratory testing driven by associations whose causal basis had not been established ²⁷. The US Preventive Services Task Force concluded in 2021 that the evidence on screening asymptomatic adults was insufficient to assess the balance of benefits and harms, and that no one serum level defines deficiency ⁵⁸. The Endocrine Society's 2024 guideline found no clinical trial evidence to support routine screening in the general population, nor in people with obesity or dark complexion, and suggested against routine testing in all the populations it considered ²⁷.

• Serum 25-hydroxyvitamin DThe sum of 25-hydroxyvitamin D3 and D2, measured by immunoassay or by liquid chromatography with tandem mass spectrometry; an international conference in 2017 called it the critical measurement for defining vitamin D status and found that assay variation had contributed to the chaos surrounding efforts to define deficiency ¹⁴.Diagnosing deficiency in people at risk: rickets, osteomalacia, malabsorption, kidney disease; monitoring high-dose treatmentGood when standardized to the Vitamin D Standardization Program's reference methods, and the research literature is now asked to report standardized values ^14,39. Unstandardized assays disagree enough that the same sample can fall on either side of a threshold, which an editorial in Clinical Chemistry titled, without irony, a difficult analyte ⁵⁷.It measures the precursor, not the hormone, and it does not say what the hormone is doing. It cannot tell a low level that is causing disease from a low level that illness has caused, which a systematic review of 290 cohorts and 172 trials argued is the usual case outside bone ²³. The 2024 Endocrine Society guideline found no clinical trial evidence to support routine testing in any population it considered, and the US Preventive Services Task Force in 2021 found the evidence on screening asymptomatic adults insufficient ^27,58.
• Serum 1,25-dihydroxyvitamin D (calcitriol)The hormone itself, with a plasma half-life of only several hours ¹³.Hypercalcemia of uncertain cause, where granulomatous disease or lymphoma makes the hormone outside the kidney or an inherited failure of the degrading enzyme leaves it unchecked ^35,41Reliable for what it measures, but it is regulated: parathyroid hormone, FGF23, calcium and phosphate set its level within hours ^8,9.Deficiency. In 319 patients with 25-hydroxyvitamin D at or below 40 nmol/L, the hormone was held normal by secondary hyperparathyroidism until the precursor fell to about 10 nmol/L, so a normal calcitriol in a deficient person is the expected finding, not reassurance ²⁸.
• Parathyroid hormone, calcium, phosphate and alkaline phosphataseThe functional readout of the system: deficiency raises parathyroid hormone and alkaline phosphatase and, when severe, lowers calcium absorption; in one series alkaline phosphatase and urine hydroxyproline tracked calcium absorption more closely than the vitamin D metabolites did ^28,29.Deciding whether a low 25-hydroxyvitamin D is doing anything to bone; the workup of rickets and osteomalacia ²¹Parathyroid hormone rises steadily as total or bioavailable 25-hydroxyvitamin D falls ⁴⁵.Parathyroid hormone has its own assay problems and its own causes for being high, so a raised value is a prompt, not a diagnosis ¹⁴.
• Free and bioavailable 25-hydroxyvitamin DEstimates of the fraction not bound to vitamin D binding protein, calculated from the total, the binding protein and its genotype, or measured directly; proposed to explain why black Americans have low total levels but denser bone ⁴⁵.Research onlyDepends entirely on how the binding protein is measured. A monoclonal immunoassay gave black Americans half the binding protein of white Americans; a mass-spectrometric method found 262 against 266 micrograms per milliliter, no difference, and showed that the immunoassay was reading the protein's isoforms differently ^45,70.It rests on the free-hormone hypothesis, which mice lacking the binding protein both support and complicate: they had low total metabolites but normal bone on an ordinary diet, became deficient faster when starved of the vitamin, and were unexpectedly resistant to toxicity after an overdose ⁵³.

The whole history, including the reversals

Vitamin D was discovered twice, by two camps that did not believe each other, and the man who reconciled them patented the result. The nutrition camp cured rickets with food; the sunlight camp cured it with lamps. Both were right because the food contained what the lamp made. Each later reversal on this timeline, from Kodicek's conclusion that the vitamin acts unchanged to the Endocrine Society's replacement of its own 2011 guideline, was made carefully on the evidence of its day ².

D2 versus D3: the same vitamin, measured differently

Vitamin D2 and D3 differ only in their side chains, and for most of the twentieth century they were treated as interchangeable; the historical review on this page describes them as having roughly equal potencies and identical effects ². The first systematic test of that assumption was a 2012 meta-analysis of randomized trials that had directly compared them. Vitamin D3 raised serum 25-hydroxyvitamin D significantly more than D2 (P = 0.001); the difference was clear when the vitamin was given as a bolus (P = 0.0002) and was lost with daily dosing ⁴². The same group then ran the trial the meta-analysis had asked for: 335 healthy South Asian and white European women in Surrey given 15 micrograms a day of D2 or D3 in juice or a biscuit over 12 winter weeks. D3 in either vehicle raised 25-hydroxyvitamin D by about 15 to 17 nmol/L more than D2 did, and by about 42 nmol/L more than placebo, so at a public-health dose and with daily dosing D3 was still the more effective form ⁸⁹. A 2024 meta-analysis of 33 trial arms found that D2 raises total 25-hydroxyvitamin D, by 11.47 ng/mL on average, while lowering the D3 form, by 4.63 ng/mL (95 percent confidence interval 2.81 to 6.46) ⁹⁰.

Part of the reason is in the half-lives. In the tracer study of 36 men, 25-hydroxyvitamin D2 had a shorter plasma half-life than 25-hydroxyvitamin D3, 13.9 against 15.1 days, and the difference was present in the Gambian men but not the British; both half-lives lengthened with higher binding-protein concentration, and men who carried two copies of Gc1f, one of the common versions of the binding protein, cleared 25-hydroxyvitamin D2 faster ¹². The study did not measure binding affinity; the authors linked the half-lives to binding-protein concentration and genotype ¹².

Whether this matters for anything beyond the blood test is less clear, with one suggestive exception. In the Cochrane review of mortality, only vitamin D3 reduced deaths (risk ratio 0.94 in 38 trials of 75,927 people); vitamin D2, alfacalcidol and calcitriol did not, and a subgroup of trials at high risk of bias suggested that D2 might even increase mortality, which the reviewers attributed to possible random error ⁹¹. A 2019 meta-analysis found the same pattern, with all-cause mortality significantly lower in D3 than D2 trials (P for interaction 0.04), while neither form alone reached significance ⁹². These are comparisons across trials, not within them, and no trial has randomized people to D2 against D3 with death as the outcome. The practical reading is that D3 is the better form for raising and holding the blood level, that D2 remains the usual option for people who avoid animal products, since D3 is mostly of animal origin and a 2017 review judged it unsuitable for vegans ^2,93, although vitamin D3 has also been found in a lichen, which is not an animal, so the objection concerns the source rather than the molecule ⁹⁴, and that the question of whether the two differ in effect, as opposed to blood level, has not been answered ^89,91.

The big trials, in full

The decade from 2017 to 2026 settled, as far as randomized trials can settle anything, the question of whether giving vitamin D to adults who are not deficient prevents the diseases its blood level had been associated with. The answer, with one or two exceptions set out in their own sections, is no. The trials were large, long, well run and mostly funded by governments; they used daily doses of 2000 to 4000 IU or monthly doses of 60,000 to 100,000 IU; and they enrolled people whose average baseline levels were already around or above the thresholds for sufficiency, which is the single most important thing to know about them ^{18,24,25,73,95}.

VITAL was the largest trial of vitamin D without calcium. It randomized 25,871 US adults, men aged 50 or older and women 55 or older, including 5,106 black participants, to 2000 IU of vitamin D3 a day and to marine omega-3 fatty acids in a two-by-two factorial design, and followed them for a median of 5.3 years. Invasive cancer was diagnosed in 793 people on vitamin D and 824 on placebo, a hazard ratio of 0.96 (95 percent confidence interval 0.88 to 1.06); major cardiovascular events occurred in 396 against 409, hazard ratio 0.97 (0.85 to 1.12). Death from cancer, a secondary endpoint, gave a hazard ratio of 0.83 (0.67 to 1.02), and death from any cause 0.99. There was no excess of hypercalcemia or other adverse events ¹⁸. The fracture analysis followed in 2022: 1,991 fractures in 1,551 participants, with total fractures in 769 of 12,927 on vitamin D against 782 of 12,944 on placebo, hazard ratio 0.98 (0.89 to 1.08), and hip fractures 1.01 (0.70 to 1.47), with no modification by baseline 25-hydroxyvitamin D ¹⁹. Depression, in 18,353 participants, gave a hazard ratio of 0.97 (0.87 to 1.09) and no change in mood scores ⁹⁶. The one positive result was autoimmune disease, the primary endpoint of a prespecified ancillary study rather than of VITAL itself: confirmed cases in 123 on vitamin D against 155 on placebo, hazard ratio 0.78 (0.61 to 0.99, P = 0.05), with the omega-3 arm at 0.85 and not significant ⁷⁷. One ancillary endpoint, with a confidence interval that touches 1, is a finding to test, not a finding to act on. It has since weakened. When 65 further cases that had begun during the randomized period were confirmed, the hazard ratio for vitamin D became 0.85 (0.70 to 1.04), no longer significant, and in 21,592 participants followed for two years after the pills stopped the seven-year figure was 0.98 (0.83 to 1.17), while the omega-3 effect held at 0.83 (0.70 to 0.99) ⁷⁹.

D-Health, in Australia, randomized 21,315 adults aged 60 to 84, recruited from the electoral roll, to 60,000 IU of vitamin D3 once a month or placebo for five years. Mean 25-hydroxyvitamin D in blood samples taken from randomly chosen participants during follow-up was 77 nmol/L on placebo and 115 on vitamin D; the trial did not screen anyone's level at entry. Over a median of 5.7 years there were 1,100 deaths, 538 (5.1 percent) on placebo and 562 (5.3 percent) on vitamin D: hazard ratio 1.04 (0.93 to 1.18). Cardiovascular mortality was 0.96, cancer mortality 1.15 (0.96 to 1.39), and in an exploratory analysis that excluded the first two years, cancer mortality was numerically higher on vitamin D, hazard ratio 1.24 (1.01 to 1.54). The investigators wrote that the precautionary principle would suggest this dosing regimen might not be appropriate in people who are vitamin D-replete ²⁴. A later analysis of 21,302 participants found major cardiovascular events in 6.0 against 6.6 percent, hazard ratio 0.91 (0.81 to 1.01), with heart attack at 0.81 (0.67 to 0.98) and stroke at 0.99; the authors called the effect possible, small in absolute terms, 172 people treated for five years to avoid one event, and consistent with no effect ⁹⁷. Cancer incidence in 21,308 participants did not differ, hazard ratio 1.02 (0.95 to 1.10) ⁹⁸, and neither did the start of drug treatment for high blood pressure, high cholesterol or type 2 diabetes, used as a stand-in for new disease (hazard ratios 1.00, 1.05 and 0.97) ⁹⁹. The fracture analysis, in 20,326 participants with linked records, found 568 (5.6 percent) with a fracture on vitamin D against 603 (5.9 percent) on placebo, hazard ratio 0.94 (0.84 to 1.06), and hip fractures 1.11 (0.86 to 1.45), with no evidence that the monthly bolus raised fracture risk ¹⁰⁰.

DO-HEALTH randomized 2,157 Europeans aged 70 and older, without major illness in the previous five years, in a two-by-two-by-two factorial of 2000 IU vitamin D3 a day, 1 gram of omega-3 a day and a strength-training program for three years. None of the six primary outcomes, systolic and diastolic blood pressure, physical performance, cognition, nonvertebral fractures and infections, differed with vitamin D, and 25 deaths were spread evenly ²⁵. The Vitamin D Assessment study in New Zealand gave 5,108 adults aged 50 to 84 an initial 200,000 IU and then 100,000 IU monthly for a median of 3.3 years: cardiovascular disease occurred in 11.8 percent against 11.5 percent, hazard ratio 1.02 (0.87 to 1.20), with no difference in the quarter who were deficient at baseline ⁹⁵; falls gave a hazard ratio of 0.99 and nonvertebral fractures 1.19 (0.94 to 1.50) ¹⁰¹.

D2d, in the United States, randomized 2,423 adults with prediabetes, not selected for low vitamin D, to 4000 IU a day. By month 24 the mean 25-hydroxyvitamin D was 54.3 ng/mL on vitamin D against 28.8 on placebo. Diabetes developed in 293 against 323 participants, 9.39 against 10.66 events per 100 person-years, hazard ratio 0.88 (0.75 to 1.04, P = 0.12) ⁷³. That near miss became one of the four groups for which the 2024 guideline suggests empiric supplementation, alongside children aged 1 to 18, adults 75 and older and pregnancy ²⁷, and the reason is a 2023 meta-analysis of individual data from the three trials built for the question, which found a 15 percent reduction, hazard ratio 0.85 (0.75 to 0.96), an absolute reduction of 3.3 percent over three years, and more regression to normal glucose. Its striking subgroup, a 76 percent reduction in people who maintained a level of 125 nmol/L or more, compares people by a level reached after randomization, which is not a randomized comparison; one of its authors listed an affiliation with Takeda Pharmaceuticals, and kidney stones (rate ratio 1.17) and hypercalcemia (2.34, 0.83 to 6.66) were not significantly increased in trials not powered for safety ¹⁰². A 2026 genetic analysis of 2,098 D2d participants looked at ApaI, a common one-letter variation in the receptor gene, named after the laboratory enzyme used to detect it; everyone carries two copies, written AA, AC or CC. The benefit was confined to people with at least one C (hazard ratio 0.81, 0.66 to 0.99) and absent in AA carriers (1.02), a subgroup finding that will need its own trial ¹⁰³.

Why did trials of this size find so little? The participants were, on average, not deficient: 77 nmol/L in the placebo group during follow-up in D-Health, which did not screen at entry, 28 ng/mL at baseline in D2d, 63 nmol/L at baseline in ViDA ^24,73,101. If vitamin D behaves like a hormone, with a set point and a plateau, then raising a replete person's precursor should do nothing, and that is what was found. The trials cannot say what would happen in people below 25 or 30 nmol/L: ViDA's deficient quarter, defined as below 20 ng/mL, showed no cardiovascular benefit, but severely deficient people were rare in all of them, and the ethical difficulty of leaving such people untreated for five years means that question may never get a trial of this size ^78,81,95.

Bone, fractures and falls: where the harm of the bolus was found

Bone is what vitamin D is for, and the trials of fractures are therefore the fairest test of supplementing people who are not deficient. The belief that it prevents fractures began with a French trial in 1992: 3,270 healthy ambulatory women with a mean age of 84, given 800 IU of vitamin D3 and 1.2 grams of calcium a day for 18 months, had 43 percent fewer hip fractures among those who completed the study (P = 0.043) ⁶³. In 2003, 100,000 IU of vitamin D3 every four months given to 2,686 British adults aged 65 to 85 reduced first fractures by 22 percent (relative risk 0.78, 0.61 to 0.99) ⁶⁴. The test that followed was far larger. In the Women's Health Initiative, 36,282 postmenopausal women aged 50 to 79 received 1000 mg of calcium with 400 IU of vitamin D3 a day or placebo for an average of seven years; hip fracture gave a hazard ratio of 0.88 (0.72 to 1.08) and total fractures 0.96 (0.91 to 1.02), and kidney stones rose, hazard ratio 1.17 (1.02 to 1.34) ²⁰. In 2017 a meta-analysis of 33 randomized trials with 51,145 community-dwelling adults over 50 found no association of vitamin D with hip fracture (risk ratio 1.21, 0.99 to 1.47) or any other fracture, with or without calcium, regardless of dose, sex, fracture history or baseline level ¹⁰⁵. The US Preventive Services Task Force's evidence review in 2018, of 11 trials and 51,419 adults aged 50 and older, reached the same conclusion and added one harm: vitamin D with calcium raised kidney stones by an absolute 0.33 percent (0.06 to 0.60) across three trials ⁴⁰. The Task Force recommended against daily doses of 400 IU or less with 1000 mg or less of calcium for preventing fractures in postmenopausal women and found the evidence insufficient for anything else ¹⁰⁶. VITAL's fracture result in 2022 and D-Health's in 2023 were consistent with this ^19,100. In May 2026 a meta-analysis in the BMJ of 69 trials and 153,902 participants, 87 percent community-dwelling, concluded with high certainty that vitamin D alone has little to no effect on any fracture (risk ratio 1.00, 0.95 to 1.06), that calcium with vitamin D has a small effect (0.91, 0.84 to 0.99) whose absolute size the authors judged below the threshold of clinical importance, and that none of the three has a meaningful effect on falls ⁸⁵. In September 2026 the Finnish FIND trial reported that among 2,495 older adults, with a mean of 75 nmol/L in a measured subgroup, five years of 1600 IU a day was associated with more fractures than placebo (hazard ratio 1.52, 1.06 to 2.18) and 3200 IU a day was not (1.07, 0.72 to 1.58); the excess appeared at the lower dose and not the higher one, and the authors concluded only that the findings do not support high doses for fracture prevention in such populations ⁸⁸.

Falls were where high doses turned out to hurt, and the finding was a reversal of what the field expected. In 2010, 2,256 community-dwelling Australian women aged 70 or older, at high risk of fracture, were randomized to a single oral dose of 500,000 IU of vitamin D3 or placebo each autumn or winter for three to five years. The vitamin D group had 171 fractures against 135, and 837 women fell 2,892 times (83.4 per 100 person-years) against 769 women falling 2,512 times (72.7 per 100 person-years), a rate ratio of 1.15 (1.02 to 1.30); the fracture rate ratio was 1.26 (1.00 to 1.59). Falls were most increased, by 31 percent, in the three months after each dose, when 25-hydroxyvitamin D had risen to about 120 nmol/L from a median baseline of 49 ³⁷. In 2016 a Zurich trial randomized 200 adults aged 70 or older with a prior fall to monthly 24,000 IU, monthly 60,000 IU, or monthly 24,000 IU plus 300 micrograms of calcifediol for a year. The two higher regimens reached the 30 ng/mL target more often and did nothing for lower-extremity function, and 66.9 percent and 66.1 percent of their participants fell against 47.9 percent on the lowest dose (P = .048) ³⁸. The Task Force in 2018 found adequate evidence that vitamin D supplementation has no benefit in preventing falls in community-dwelling adults 65 or older and recommended against it, reversing the recommendation it had made in 2012 ^68,69. A 2024 secondary analysis of a four-year trial in 2,303 postmenopausal women, using standardized 25-hydroxyvitamin D, found no association with falls up to 60 ng/mL but twice the odds of two or more falls at 60 ng/mL or above, a group that was only about 2 percent of observations ³⁹.

The mechanism of the bolus harm is not known. What is known is that it is specific to the way the dose is given: the 2024 Endocrine Society systematic review concluded that high-dose intermittent vitamin D may increase falls compared with lower-dose daily dosing, and the guideline suggests daily rather than intermittent administration for nonpregnant people over 50 for whom vitamin D is indicated ^27,78. D-Health's monthly 60,000 IU did not raise fracture risk ¹⁰⁰ and ViDA's monthly 100,000 IU did not raise falls (hazard ratio 0.99) ¹⁰¹, yet monthly 60,000 IU raised falls against 24,000 IU in the Zurich trial ³⁸, and 100,000 IU every four months reduced fractures in one British trial ⁶⁴, so which doses and intervals produce the harm is not known. The sensible reading is that the body's own output is steady, a single exposure to sun being released over days ⁴, and that a bottle can do what the sun cannot.

Where deficiency is real, none of this applies. The 2016 global consensus on nutritional rickets, from 33 experts representing 11 organizations, set out prevention by supplementation and food fortification and treatment of established rickets and osteomalacia, and called for international prevention programs ²¹. The older literature on housebound elderly people found that vitamin D3 lowers parathyroid hormone, reduces bone turnover, raises bone density, and with calcium may reduce hip fractures in nursing-home residents ²⁹. The trials on this page were not run in those people.

Infections and COVID-19: a signal that faded as the trials grew

The immune story had a strong biological foundation, since immune cells carry the receptor and make the hormone, and the most careful evidence for a benefit came in 2017 ¹¹. An individual-participant-data meta-analysis of 25 randomized trials with 11,321 people aged 0 to 95 found that vitamin D supplementation reduced acute respiratory tract infections, adjusted odds ratio 0.88 (0.81 to 0.96). The effect was present with daily or weekly dosing (0.81) and absent with bolus dosing (0.97), and it was strongest in people whose baseline 25-hydroxyvitamin D was below 25 nmol/L (0.30, 0.17 to 0.53), which is the pattern a hormone with a threshold would produce ⁷¹.

The pandemic put the claim under the largest spotlight any supplement has had. The claim had three main supports, and two of them have since been withdrawn. In Córdoba, Spain, 76 hospitalized patients were allocated two to one to calcifediol or no calcifediol in an open-label pilot; 1 of 50 treated patients and 13 of 26 untreated patients went to intensive care, and the authors themselves called for larger trials with properly matched groups ⁷⁴. In Israel, a retrospective study of 253 hospitalized patients whose vitamin D had been measured before infection reported that those below 20 ng/mL were 14 times more likely to have severe or critical disease than those at 40 ng/mL or above ¹⁰⁷. PLOS ONE retracted it in 2025: an independent member of its editorial board judged that selecting only patients who had happened to be tested before infection, with measurements taken 14 to 730 days earlier and possible corrective treatment in between, was a fatal flaw in the design; most of the authors did not agree with the retraction ⁸³. A 2023 meta-analysis that announced definitive evidence of protection against intensive-care admission was retracted in 2025 after the journal confirmed that at least one of its five included studies was not a randomized trial; its authors disagreed ^84,108. The trials then tested the claim. In 240 hospitalized Brazilian patients with moderate to severe COVID-19, a single 200,000 IU dose raised mean 25-hydroxyvitamin D to 44.4 ng/mL, against 19.8 on placebo, from a baseline of 20.9 ng/mL, and did not change the length of stay (7.0 days in both groups), in-hospital mortality (7.6 against 5.1 percent), intensive-care admission or the need for ventilation ⁷⁵. In the United Kingdom, CORONAVIT tested the policy rather than the pill: 6,200 adults were randomized to be offered a postal finger-prick test with six months of 800 or 3200 IU a day if below 75 nmol/L, or to no offer. Of those offered, 86.3 percent qualified and were supplemented. At least one respiratory infection occurred in 5.7 percent of the lower-dose group (odds ratio 1.26, 0.96 to 1.66) and 5.0 percent of the higher-dose group (1.09, 0.82 to 1.46) against 4.6 percent of controls, and COVID-19 in 3.6 and 3.0 percent against 2.6 percent, with no significant differences; the trial was open label ⁷⁶. In Chicago, 923 adults randomized to 4000 or 10,000 IU against 400 IU a day for up to nine months had COVID-19 in 10.9 against 14.7 percent, hazard ratio 0.73 (0.51 to 1.06, P = 0.10), not significant; in a prespecified secondary comparison 4000 IU did better than 400 IU (0.62, 0.39 to 0.99) and 10,000 IU did not (0.99) ¹⁰⁹. In Mongolia, where 95.6 percent of 8,851 schoolchildren were below 20 ng/mL, three years of 14,000 IU weekly left the mean level at 31.0 ng/mL at the end of the trial, against 10.7 ng/mL on placebo, and did not reduce tuberculosis infection (risk ratio 1.10), tuberculosis disease (0.87) or hospitalization for respiratory infection (0.86, 0.52 to 1.40) ¹⁰⁴. That last trial matters most, because it was run in the deficient children the 2017 analysis had predicted would benefit.

In 2025 the same group updated its 2021 meta-analysis of trial-level (aggregate) data, which had found an odds ratio of 0.92 (0.86 to 0.99) in 37 trials. Six new trials (19,337 participants, one of them of 15,804 people) were identified and data obtained from three; the primary comparison of vitamin D against placebo, 40 studies and 61,589 participants, gave an odds ratio of 0.94 (0.88 to 1.00, P = 0.057). The prespecified subgroups no longer showed modification by age, baseline level, dosing frequency or dose size, and a funnel plot showed the asymmetry that suggests small negative studies went unpublished. The authors' own conclusion was that the point estimate was similar to that of the 2021 analysis but the interval now included 1.00, indicating no statistically significant protection ⁸¹. In 2026 Adrian Martineau, who led the 2017 analysis and is the senior author of the 2025 update, concluded in a review that the latest trials and meta-analyses do not support vitamin D supplements to prevent acute respiratory infection or tuberculosis infection, or to improve the response to tuberculosis treatment. He attributed the earlier optimism partly to cell and animal results of limited relevance, to associations that were not causal, and to publication bias and citation bias, the tendency of positive results to be published and cited more often, while listing ways the large trials might have missed a true effect, among them that few participants were deficient at the start ¹¹⁰. A 2026 Cochrane review of 107 trials and 31,521 participants, from pregnancy to age five, found low-certainty evidence that vitamin D may slightly reduce the proportion of children seen for a respiratory infection (risk ratio 0.95, 0.91 to 1.00, in 10 trials of 2,447 children) and moderate-certainty evidence that it probably does not reduce the number of such visits per child ¹¹¹. The 2024 Endocrine Society review nevertheless kept a suggestion of supplementing children aged 1 to 18, on low-certainty evidence of fewer respiratory infections, alongside the prevention of rickets ^27,78.

Mortality and cancer: a small, fragile effect in the old

The cancer hope began with a map. In 1980 two epidemiologists noticed that colon cancer death rates in the United States were highest where people were exposed to the least natural light, in major cities and in rural areas at high latitudes, and proposed vitamin D as a protective factor; they themselves raised the possibility of an ecological fallacy, an association between places that need not hold for the people in them ⁶². In 2007 a Nebraska trial of 1,179 postmenopausal women, designed for fractures, reported cancer as its principal secondary outcome: calcium with 1100 IU of vitamin D3 a day gave an unadjusted relative risk of 0.40 against placebo ⁶⁵. The same group's trial built for the question, 2,303 women with a mean baseline level of 32.8 ng/mL on 2000 IU and 1500 mg of calcium a day for four years, found cancer in 3.89 against 5.58 percent, hazard ratio 0.70 (0.47 to 1.02, P = .06), not significant; one of its authors listed an affiliation with GrassrootsHealth ⁷². In VITAL, a secondary analysis found metastatic or fatal cancer in 1.7 against 2.1 percent, hazard ratio 0.83 (0.69 to 0.99, P = .04), with the reduction confined to people of normal weight (0.62, against 0.89 and 1.05 in overweight and obese participants) ¹¹². D-Health found no effect on cancer incidence in 21,308 people, hazard ratio 1.02 (0.95 to 1.10), and its authors noted that if vitamin D lowers cancer mortality it would have to do so by improving survival ⁹⁸.

If vitamin D does anything for people who are not deficient, the best candidate is a small reduction in deaths among the elderly, and the honest description of that effect is that it has hovered at the edge of significance for a decade. The 2014 Cochrane review of 56 trials and 95,286 participants, mostly women over 70, found that vitamin D reduced mortality from 12.7 to 12.5 percent, risk ratio 0.97 (0.94 to 0.99); the effect belonged to vitamin D3 alone, risk ratio 0.94 (0.91 to 0.98) in 38 trials of 75,927 people, corresponding to 150 people treated for five years to prevent one death, and the reviewers warned that participants lost from the trials (attrition) and the selective reporting of outcomes could have biased the result ⁹¹. Their companion review of cancer, 18 trials and 50,623 people, found no effect on cancer occurrence (risk ratio 1.00) and noted that most trials had a high risk of for-profit bias; vitamin D3 reduced cancer mortality (0.88, 0.78 to 0.98) in four trials, a finding the sequential analysis judged possibly due to random error ¹¹³.

The trials that came after were larger and found less. A 2019 meta-analysis of 52 trials and 75,454 participants found no association with all-cause mortality (0.98, 0.95 to 1.02) or cardiovascular mortality, and a reduction in cancer death (0.84, 0.74 to 0.95) ⁹². A 2023 meta-analysis of 14 trials and 104,727 participants, with individual patient data from 7 of them covering 90 percent of participants, with 2,015 cancer deaths, found a nonsignificant 6 percent reduction in cancer mortality overall (0.94, 0.86 to 1.02), a 12 percent reduction in the ten trials with daily dosing (0.88, 0.78 to 0.98) and none in the four with bolus dosing (1.07, 0.91 to 1.24), with the interaction significant at P = 0.042 ¹¹⁴. D-Health, with monthly bolus dosing, found a numerically higher cancer mortality ²⁴. Set against each other, these results say that if there is an effect it requires daily dosing, is small in absolute terms (in the Cochrane vitamin D3 trials, deaths from any cause 11.0 against 11.4 percent, about one prevented for every 150 people treated for five years, and cancer deaths 2.5 against 2.8 percent) ^91,113, and has never been the primary outcome of a trial that found it.

The 2024 Endocrine Society systematic review rated the evidence for a very small reduction in mortality among adults older than 75 as high certainty, and the guideline suggests empiric supplementation after 75 on that basis ^27,78. A 2026 reanalysis of the same review by a method that admits only double-blind, placebo-controlled trials at low risk of bias kept 8 of its 25 trials, 30,320 participants, and found a risk ratio of 0.96 (0.92 to 1.00, P = 0.04); its authors called this an evidence signal in the population aged 75 and older that remains to be confirmed, with no single conclusive trial ¹¹⁵. Mendelian randomization offers a different kind of evidence: in 307,601 white European UK Biobank participants, genetically predicted 25-hydroxyvitamin D had an L-shaped relation to mortality, falling steeply up to 50 nmol/L and flat above, which fits the trials, since the trials enrolled people above that elbow; its authors noted that a genetic approach is best suited to proof of principle and that the strength of the association is approximate, and a correction to the paper was published in 2025 ^30,31. Nonlinear Mendelian randomization has since been shown to be open to serious bias: a 2024 analysis that tested the two main methods against outcomes the exposure should not affect, with vitamin D among its exposures, found clearly biased estimates in certain situations, and its authors advised that such methods need justification and checks before further use, so the curve's shape should be treated as provisional ³². The message the trials and the genetic study share, with these caveats, is that benefit, if any, lives below 50 nmol/L, in the old, with daily dosing, and that the trials have looked mostly elsewhere.

The guidelines, and the reversal of 2024

Two documents in 2011 set the terms of the argument. The Institute of Medicine's committee reviewed the evidence for skeletal and extraskeletal outcomes and concluded that calcium and vitamin D have a causal role in bone health that justifies intake requirements, while for cancer, cardiovascular disease, diabetes and autoimmune disorders the evidence was inconsistent, inconclusive as to causality and insufficient to set requirements. It set the recommended dietary allowance at 600 IU a day for ages 1 to 70 and 800 IU for 71 and older, corresponding to a serum 25-hydroxyvitamin D of at least 20 ng/mL (50 nmol/L), derived under conditions of minimal sun exposure; it saw no consistent benefit from higher values and U-shaped associations for some outcomes; and it concluded that the prevalence of inadequacy in North America had been overestimated, with laboratory ranges in need of reassessment to avoid both undertreatment and overtreatment ¹⁵. The committee's summary supplies the rest of the scale. A blood level of 16 ng/mL (40 nmol/L) meets the needs of about half the population, the estimated average requirement, and 20 ng/mL those of at least 97.5 percent. Levels above 30 ng/mL (75 nmol/L) were not consistently associated with greater benefit, and levels above 50 ng/mL (125 nmol/L) should raise concern about harm. The tolerable upper intake level, the highest daily intake likely to pose no risk, was set at 4000 IU a day for everyone aged 9 and older, up from 2000 IU in 1997; it was derived from 10,000 IU, an intake below which hypercalcemia and acute toxicity had not been seen, and lowered for uncertainty about chronic disease and mortality. The 1997 values it replaced had been 200 IU a day to age 50, 400 IU to 70 and 600 IU after ¹⁵. The Endocrine Society's clinical practice guideline the same year, with Michael Holick as first author and a task force that received no corporate funding, considered deficiency very common in all age groups, recommended supplementation at suggested daily intakes and tolerable upper limits by age and circumstance, suggested measuring 25-hydroxyvitamin D by a reliable assay as the initial test in anyone at risk, and stopped short of screening people not at risk or of prescribing vitamin D to attain noncalcemic benefit for cardiovascular protection ²⁶.

The two documents disagreed about who was deficient and what to do about it. Testing had already risen steeply before 2011, with vitamin D among the Medicare Part B tests whose reimbursed volume grew most between 2000 and 2010 ²², and a research letter tracked the growth of high-dose supplement use to 2014 ¹¹⁶. In 2012 members of the Institute of Medicine committee published a rebuttal of the Endocrine Society guideline's target levels, its definition of deficiency and its view of who is at risk rather than part of the general population ¹¹⁷. The US Preventive Services Task Force stayed closer to the Institute of Medicine, and its falls advice reversed. A 2009 meta-analysis of eight double-blind trials, 2,426 people, had found that 700 to 1000 IU a day cut the risk of falling by 19 percent (relative risk 0.81, 0.71 to 0.92) and that lower doses did not ⁶⁷, and in 2012 the Task Force recommended vitamin D, with exercise or physical therapy, to prevent falls in adults 65 and older at increased risk, a B grade, its term for a service it recommends ⁶⁸. In 2018 it recommended against low-dose vitamin D with calcium to prevent fractures in postmenopausal women and found the evidence insufficient for other doses and groups ¹⁰⁶; the same year it reversed its falls advice, recommending against vitamin D to prevent falls in community-dwelling adults 65 and older, with adequate evidence of no benefit and of small to moderate harms ⁶⁹; in 2021 it found the evidence on screening asymptomatic adults insufficient and observed that no one serum level defines deficiency ⁵⁸; and in 2024 its new falls statement recommended exercise and an individualized offer of multifactorial programs, with no mention of vitamin D in its abstract ⁸⁰.

In 2024 the Endocrine Society replaced its 2011 guideline with one written for a different purpose, the use of vitamin D to lower the risk of disease in people without an established indication for treatment or testing. Its panel prioritized placebo-controlled trials, defined empiric supplementation as intake above the dietary reference intake without testing, and worked through 14 questions with a systematic review of 151 studies, rating the certainty of the evidence for each with the GRADE system ^27,78. It suggests empiric supplementation for children and adolescents aged 1 to 18 to prevent rickets and possibly respiratory infections, for adults 75 and older for a possible reduction in mortality, in pregnancy for possible reductions in preeclampsia (a dangerous rise in blood pressure in pregnancy), in death in the womb, in preterm birth, in the birth of babies small for the length of pregnancy and in death in the first four weeks of life, and for people with high-risk prediabetes, blood sugar above normal with features that make progression to diabetes likely. For healthy adults younger than 75 it suggests against empiric supplementation above the dietary reference intake. It found no trial evidence to support routine 25-hydroxyvitamin D testing in the general population, nor in people with obesity or dark complexion, and no clear evidence of an optimal target level, and suggests against routine testing in all the populations it considered. For nonpregnant people over 50 for whom vitamin D is indicated it suggests daily rather than intermittent high doses ²⁷. The guideline states that it is not meant to replace the dietary reference intakes and does not apply to people with established indications ²⁷.

That is a reversal in all but name. The guideline's own text says that the benefit-risk ratio of the increase in vitamin D use was not clear, that the associations that drove it had not been shown to be causal, and that the optimal intake and the role of testing remained uncertain ²⁷. The International Osteoporosis Foundation's 2024 position paper draws the same distinction between overt deficiency, which warrants prompt replacement, and population supplementation, where the aim is to keep the average person out of the range where harm begins ¹¹⁸. A 2025 commentary in JAMA on the guideline's use in primary care, and a letter in the BMJ in 2026 on what health systems spend on vitamin D supplementation, mark how far the argument has traveled ^119,120.

The binding protein and the free-hormone hypothesis

Like the other steroid hormones, vitamin D and its metabolites travel bound to a carrier, the vitamin D binding protein, first recognized in 1975 as the same molecule as the group-specific component (Gc), a plasma protein that comes in common inherited variants and whose function had been unknown ⁵⁴, and like them they are thought to act only in the small free fraction that can leave the blood and enter cells ². What the carrier does was shown in 1999 in mice engineered to lack it. On a vitamin D-replete diet they had low total serum metabolites and were otherwise normal; on a deficient diet they developed secondary hyperparathyroidism and the bone changes of deficiency faster than normal mice; the protein markedly prolonged the half-life of 25-hydroxyvitamin D by slowing its uptake into the liver; and after an overload of vitamin D the mice without it were unexpectedly less susceptible to hypercalcemia. The authors concluded that the protein's role is to maintain a stable serum store and to modulate how fast the vitamin is activated and delivered, properties that may have evolved to buffer an erratic supply ⁵³. In people, the half-lives of both 25-hydroxyvitamin D2 and D3 rose with the concentration of the binding protein, and men with two copies of its Gc1f variant cleared 25-hydroxyvitamin D2 faster ¹². The first person known to lack the protein entirely, a 58-year-old woman whose parents were related, was described in 2019: she had a homozygous deletion of the gene, so that neither copy worked, an apparently severe vitamin D deficiency that did not respond to supplementation, normal blood calcium and only a relatively mild disturbance of bone metabolism, complicated by severe autoimmune disease ⁵⁵.

The free-hormone hypothesis became a clinical argument in 2013. Black Americans have lower total 25-hydroxyvitamin D than white Americans and yet higher bone density, which the total level cannot explain. In 2,085 black and white participants of a US cohort, total 25-hydroxyvitamin D was 15.6 ng/mL in black and 25.8 in white participants, the binding protein 168 against 337 micrograms per milliliter, and bone density 1.05 against 0.94 grams per square centimeter; two common variants in the binding-protein gene explained 79.4 percent of the variation in the protein, and among participants with two identical copies of the gene the estimated bioavailable 25-hydroxyvitamin D was the same, 2.9 against 3.1 ng/mL. The conclusion was that low totals in black Americans reflect less carrier rather than less vitamin ⁴⁵. The paper was widely read as settling the matter.

It did not survive a better assay. The binding protein in that study was measured by a monoclonal immunoassay, one built on a single antibody, and the protein exists in several isoforms whose frequencies differ by ancestry. In 2016 a method using liquid chromatography and tandem mass spectrometry, which identifies the isoforms directly and was 97 percent concordant with genotyping, found mean binding-protein concentrations of 262 micrograms per milliliter in white and 266 in black participants of a large cardiovascular cohort, no difference, and showed that the immunoassay had been reading the isoforms differently ⁷⁰. The large racial gap in binding protein was an artifact; the gap in total 25-hydroxyvitamin D, and the puzzle of denser bone despite it, remain. The honest state of the hypothesis is that the free fraction is probably what matters, that it cannot yet be measured well enough to use, and that an international conference in 2017 listed free and bioavailable 25-hydroxyvitamin D among the candidate markers that need standardization before anyone can act on them ¹⁴. The 2024 guideline, asked whether people with dark complexion should be screened, found no trial evidence that they should ²⁷.

Dose, and where it turns: toxicity

How much vitamin D a person needs depends on how much sun they get, and the quantitative answer was worked out in 67 men in Omaha given 0, 25, 125 or 250 micrograms of vitamin D3 a day for about 20 winter weeks. Serum 25-hydroxyvitamin D changed in direct proportion to the dose, by about 0.70 nmol/L for each microgram, which is 40 IU, of daily input; sustaining the autumn level through winter needed 12.5 micrograms, 500 IU, from a supplement, but the total from all sources needed to hold it was about 96 micrograms, roughly 3800 IU a day, the difference being drawn from what the skin had stored in summer ¹²¹. The arithmetic matters for reading the trials: 2000 IU a day should raise 25-hydroxyvitamin D by about 35 nmol/L, which is what the trials found, and a person at 77 nmol/L taken to 115 has been moved from one part of the plateau to another ^24,121.

Toxicity is hypercalcemia. Its symptoms are confusion, apathy, recurrent vomiting, abdominal pain, excessive urination and thirst, and dehydration; exogenous toxicity follows excessive long-term intake of pharmacological preparations, and 25-hydroxyvitamin D above 150 ng/mL (375 nmol/L) is its hallmark ³⁵. In 13 published reports of intoxication, serum 25-hydroxyvitamin D ranged from 150 to 1,220 ng/mL and calcium from 11.1 to 23.1 mg/dL, with vomiting, dehydration, pain and loss of appetite; the causes were manufacturing errors, overdosing by patients or prescribers, and combinations of these ³⁶. Endogenous toxicity comes from the hormone being made where it should not be, in granulomatous disease and some lymphomas, or from failure to degrade it: children with recessive loss-of-function mutations in CYP24A1 developed severe hypercalcemia, failure to thrive, vomiting, dehydration and nephrocalcinosis, some of them after bolus prophylaxis, a susceptibility that is a plausible but unproven explanation for some of the increase in idiopathic infantile hypercalcemia in Britain in the 1950s, when milk products were heavily fortified ⁴¹. Some apparently healthy people have hypercalcemia on doses considered safe for the general population, a phenomenon the toxicologists call hypersensitivity to vitamin D ³⁵.

Below frank toxicity sit the harms of the trials: more falls and fractures after annual or large monthly boluses, kidney stones when vitamin D is combined with calcium, and the numerically higher cancer mortality in the bolus-dosed D-Health trial ^24,37,38,40. In the United States, the proportion of adults taking more than 1000 and more than 4000 IU a day rose between 1999 and 2014 ¹¹⁶. The Institute of Medicine in 2011 warned of overtreatment as well as undertreatment ¹⁵. Sunlight is not among the causes: the skin's photochemistry caps what one exposure makes, and skin releases what it has made over days rather than all at once ^4,5,35.

Magnesium and vitamin K2: the companion claims at their strength

Two supplements are commonly sold alongside vitamin D with the claim that it cannot work without them. The magnesium claim rests on biochemistry, since the enzymes that make and break down the vitamin D metabolites use magnesium, and on one randomized trial. In 180 adults aged 40 to 85 nested within a colorectal-cancer prevention trial, magnesium supplementation changed plasma 25-hydroxyvitamin D3 in a direction that depended on where the person started: it raised the level when baseline 25-hydroxyvitamin D was near 30 ng/mL and lowered it when baseline was higher, from about 30 to 50 ng/mL, with the interaction surviving correction for multiple testing; the authors concluded that optimal magnesium status may be important for optimizing vitamin D status. Two of the authors were affiliated with a center for magnesium education and research ¹²². That is a single trial of a biochemical interaction with no clinical outcome; it supports taking magnesium deficiency seriously, and nothing more.

The vitamin K2 claim is that vitamin D raises calcium and K2 is needed to direct it into bone and away from arteries. The best test was a Danish trial in 365 men aged about 71 with a calcium score above 300 units on a computed tomography (CT) scan of the aortic valve, a measure of how much calcium the scan shows on it, randomized to 720 micrograms of menaquinone-7 with 25 micrograms of vitamin D daily or placebo for 24 months. Valve calcification progressed by 275 units on treatment and 292 on placebo, a difference of -17 units (95 percent confidence interval -86 to 53, P = 0.64); valve area, the speed of blood through the narrowed valve (jet velocity), aortic and coronary calcification, valve surgery, deaths and cardiovascular events did not differ, even though the treatment did what it was meant to do biochemically, lowering uncarboxylated matrix Gla protein by 212 picomoles per liter against a rise of 45 ¹²³. For bone, a 2020 meta-analysis of eight randomized trials with 971 participants found that vitamin K with vitamin D raised total bone mineral density by a pooled standardized mean difference of 0.316 (0.031 to 0.601), a small-to-moderate effect expressed in units of the spread between people, and lowered undercarboxylated osteocalcin, a bone protein that is incompletely activated when vitamin K is short; the trials were small and no fracture outcome was reported ¹²⁴. The combination is therefore not shown to protect arteries, and its effect on bone is a change in a density measurement in small trials, not a prevented fracture. The vitamin-k2 and magnesium stars on this site take each claim further.

The 2025 to 2026 literature, verified

The newest results have not changed the shape of the story, but they have added one positive trial in a disease where vitamin D had long been suspected to matter. D-Lay MS, run in 36 French centers, randomized 316 untreated adults with a clinically isolated syndrome typical of multiple sclerosis, less than 90 days old, with serum vitamin D below 100 nmol/L, to 100,000 IU of cholecalciferol or placebo every two weeks for 24 months. Disease activity, defined as a relapse or new MRI activity, occurred in 60.3 percent on vitamin D and 74.1 percent on placebo, hazard ratio 0.66 (0.50 to 0.87, P = .004), with the median time to activity 432 against 224 days. All three MRI outcomes favored vitamin D; none of the ten clinical outcomes did, with relapses in 17.9 against 21.8 percent, hazard ratio 0.69 (0.42 to 1.16). The authors called for further investigation of pulse high-dose vitamin D as add-on therapy ⁸². It is a placebo-controlled trial that met its primary endpoint in a disease where, as its authors note, the data on supplementation had been conflicting, and the endpoint it met is radiological ⁸².

Two 2026 results went the other way or stayed ambiguous. In 24 Chinese fertility centers, 876 women with polycystic ovary syndrome, a hormonal disorder that disrupts ovulation, were randomized to 4000 IU a day or placebo for up to 90 days before in vitro fertilization, in which eggs are fertilized in the laboratory and the embryo is transferred to the uterus. Serum 25-hydroxyvitamin D rose from 16.5 to 32.3 ng/mL, and live births after the first embryo transfer were 52.0 percent against 50.2 percent, risk ratio 1.03 (0.91 to 1.18) ¹²⁵. In Copenhagen, a post hoc analysis, one decided after the data were seen, of 498 ten-year-olds whose mothers had been randomized to 2800 or 400 IU a day from pregnancy week 24 found better verbal memory (0.17 standard deviations, about a sixth of the usual spread of scores) and visual memory (0.24) in the high-dose group, with a third difference lost after correction for multiple testing; the trial was unblinded when the children were three, so that families then knew which dose the mother had received, and the analysis was not prespecified ¹²⁶. In prediabetes, the D2d genetic substudy of 2,098 participants reported benefit confined to carriers of the ApaI AC and CC receptor genotypes ¹⁰³.

Two 2025 results concern markers of aging rather than disease. In 1,031 VITAL participants, 2000 IU a day slowed the shortening of telomeres, the protective caps on the ends of chromosomes, measured in white blood cells, by 0.14 thousand base pairs over four years (95 percent confidence interval 0.007 to 0.27, P = 0.039) ¹²⁷. In a post hoc analysis of 777 DO-HEALTH participants, omega-3 alone slowed three of four DNA-methylation clocks, which estimate biological age from chemical tags on DNA, and the three treatments together, vitamin D, omega-3 and exercise, had an additive effect on one; the abstract reports no effect of vitamin D on its own, and one author lists an affiliation with Altos Labs Limited, a company ¹²⁸. Neither is a clinical outcome. In Finland, the FIND trial gave 2,434 mostly replete older adults 1600 or 3200 IU a day or placebo for five years: major depression was diagnosed in 11, 8 and 14 people, hazard ratios 0.78 (0.35 to 1.71) and 0.57 (0.24 to 1.35), with a borderline result only for the higher dose over an extended 7.8 years of follow-up (0.54, 0.30 to 1.00, P = 0.05) ¹²⁹; the same trial's fracture analysis, published in September 2026, found more fractures on 1600 IU a day than on placebo but not on 3200 IU ⁸⁸.

The syntheses moved toward the null. The updated respiratory-infection meta-analysis of 61,589 participants lost its statistical significance ⁸¹. The BMJ fracture and falls meta-analysis of 153,902 participants found, with high certainty, little to no effect of vitamin D alone ⁸⁵. The reanalysis of the Endocrine Society's mortality evidence kept a risk ratio of 0.96 in people over 75 and called it a signal to be confirmed ¹¹⁵. A 2024 analysis with standardized assays found no excess falls below 60 ng/mL ³⁹. Searches of PubMed for 2025 and 2026 work, run for this page in October 2026, found new primary-prevention reports from D-Health, VITAL, DO-HEALTH and the Finnish FIND trial, set out above and in the sections on the big trials and on bone, and no new guideline that departs from the 2024 Endocrine Society position ^{88,98,99,127,128,129}. The Task Force's 2021 screening statement and 2018 fracture statement are still its current ones; its falls statement was replaced in 2024 by one that recommends exercise, a B grade, and an individualized offer of multifactorial programs, a C grade, its term for a service to offer selectively, and whose abstract does not mention vitamin D ^58,80,106. A vitamin D trial after heart attack was among the highlights of the American Heart Association's 2025 meeting ¹³⁰. TARGET-D, published in full in September 2026 by investigators at Intermountain Medical Center in Utah, randomized 630 patients who had had a heart attack to usual care or to vitamin D3 adjusted to keep the blood level above 40 and up to 80 ng/mL; their median baseline level was 25 ng/mL, and about half started on 5000 IU a day. Over an average of 4.2 years, major adverse cardiovascular events, a composite of death, heart attack, hospitalization for heart failure and stroke, occurred in 15.7 against 18.4 percent, hazard ratio 0.85 (0.58 to 1.24, P = .40), which is null. Repeat heart attack, one of four secondary endpoints, was 3.8 against 7.9 percent, hazard ratio 0.48 (P = .03), in a trial compared with usual care rather than placebo, a finding to test rather than to act on ⁸⁷.

What was believed, and is not

That vitamin D is a vitamin. It is the prohormone of a steroid hormone, made in skin from a cholesterol precursor by light, activated by two regulated enzymes, carried by a binding protein, received by a nuclear receptor of the steroid family and switched off by feedback; a historian of the field writes that strictly speaking it has been misnamed ^2,7,9,10.

That the vitamin acts unchanged. Kodicek's group concluded this after ten years of tracer work; DeLuca's better-labeled vitamin found the metabolites within a few years, and by 1971 the hormone had been identified by three laboratories ^1,2,49,51,52.

That the active hormone is the thing to measure. Calcitriol is held normal by secondary hyperparathyroidism until 25-hydroxyvitamin D falls to about 10 nmol/L, so a normal hormone level is the expected finding in deficiency ²⁸.

That a bigger, rarer dose is a convenient way to give it. The annual 500,000 IU dose raised falls and fractures; monthly 60,000 IU raised the proportion of fallers to 66.9 percent against 47.9 percent on 24,000 IU; and the 2024 guideline suggests daily rather than intermittent dosing for nonpregnant people over 50 for whom vitamin D is indicated ^27,37,38.

That vitamin D prevents falls. A 2009 meta-analysis found a 19 percent reduction; the Task Force recommended it in 2012 and recommended against it in 2018, and its 2024 statement does not include it ^67,68,69,80.

That calcium with vitamin D prevents hip fractures. A 43 percent reduction among very old French women in 1992 was not reproduced in 36,282 postmenopausal American women aged 50 to 79 in 2006 ^20,63.

That low levels in black Americans are an artifact of a low binding protein. The binding-protein difference was an artifact of the immunoassay; measured by mass spectrometry it is 262 against 266 micrograms per milliliter ^45,70.

That supplementing replete adults prevents cancer, cardiovascular disease, fractures, diabetes or depression. VITAL, D-Health, DO-HEALTH, ViDA and D2d, with more than 56,000 participants between them, found no such effect on their primary outcomes ^{18,24,25,73,95}.

That vitamin D prevents autoimmune disease. VITAL's hazard ratio of 0.78 became 0.85, not significant, once late-confirmed cases were added, and 0.98 two years after supplementation ended ^77,79.

That it prevents respiratory infections. The 2017 individual-participant analysis found a 12 percent reduction; the group's 2025 update, now pooling stratified aggregate data from 40 studies and 61,589 participants, found an interval that includes no effect, and a trial of 8,851 deficient Mongolian children found nothing ^71,81,104.

That it treats COVID-19. A 200,000 IU dose in hospital and a test-and-treat program in 6,200 adults both failed ^75,76. The observational study that put the risk of severe disease at fourteen-fold and the meta-analysis that called the protective evidence definitive were both retracted in 2025 ^83,84.

That vitamin D with calcium cuts cancer risk by 60 percent. The 2007 secondary finding was not reproduced by the trial built to test it ^65,72.

That healthy adults should be tested and supplemented beyond the dietary intake. The Endocrine Society recommended measurement in people at risk in 2011 and in 2024 suggested against routine testing in every population it considered and against empiric supplementation above the reference intake in healthy adults under 75 ^26,27.

That one serum threshold defines deficiency. The Institute of Medicine's 16 ng/mL average requirement and its 20 ng/mL, the 30 ng/mL target used in trials such as the 2016 Zurich trial, the 12 ng/mL rickets threshold and the 20 to 50 ng/mL safe range of the 2017 conference are all in use, and the US Preventive Services Task Force concluded in 2021 that no one cutpoint defines deficiency ^14,15,38,58.

What is strange about it

The skin stores the product of light for days and releases it slowly, because the first thing light makes is not vitamin D but a previtamin that converts by heat alone over at least three days, and the binding protein takes only the finished product ⁴. A person who sunbathes on Monday is still exporting vitamin D on Thursday.

Sunlight polices its own dose. Beyond a point the same light that makes previtamin D3 turns it into two inert cousins, lumisterol and tachysterol, which is why under simulated sunlight the yield of previtamin peaked at about 20 percent of the precursor ⁵, and why sun exposure is not among the causes of toxicity the clinical reviews list ³⁵.

A patent on sunshine built a research foundation. Steenbock's irradiation of food, and the fortification of milk with irradiated yeast that followed, was patented by the University of Wisconsin, and the proceeds founded an organization that went on to fund several Nobel laureates ^2,61.

The hormone's receptor was found before the hormone was. Haussler and Norman purified a saturable nuclear binding protein for a vitamin D metabolite in 1969; the metabolite it bound was not identified until 1971 ^1,43.

The molecule that cures rickets can cause a serious disease of infancy. Children lacking the enzyme that degrades the hormone can develop severe hypercalcemia, vomiting, dehydration and nephrocalcinosis after routine prophylaxis, a susceptibility that may explain part of the rise in infantile hypercalcemia that followed heavy fortification of milk in 1950s Britain ⁴¹.

Mice without the carrier protein are harder to poison. Animals lacking the vitamin D binding protein were unexpectedly less susceptible to hypercalcemia after an overload; the authors concluded that the carrier keeps a stable serum store and modulates how fast the vitamin is activated and delivered ⁵³.

What is still unknown

Whether supplementing people who are truly deficient, below 25 or 30 nmol/L, prevents anything beyond rickets and osteomalacia. The trials enrolled replete people; in 34 intervention studies of 2,805 people whose mean level was below 50 nmol/L, doses of 50 micrograms a day or more did not show better results ²³; a nonlinear Mendelian randomization study placed the excess mortality associated with low vitamin D below 50 nmol/L, by nonlinear methods since shown to be prone to bias ^30,32; and the 2025 respiratory meta-analysis no longer finds the baseline-level interaction it once did ^78,81. The strongest recent argument that treating deficiency would help comes from observational data, not trials. In 2026 a group in Heidelberg emulated VITAL and D-Health inside the UK Biobank. When participants were weighted to the trials' baseline levels, the expected effect on mortality was null, as in the trials (hazard ratios 0.97 and 1.02); restricted to people with vitamin D insufficiency the emulations predicted hazard ratios of 0.85 and 0.81, and restricted to people with deficiency 0.79 and 0.75 ⁸⁶, with similar predictions for cardiovascular death ¹³¹. These are associations between measured blood levels and outcomes, open to the reverse causation described in the section on the blood test ²³, and they are a reason to run a trial in deficient people, not a substitute for one.

Whether the small mortality reduction in people over 75 is real. Pooled estimates have sat at a risk ratio of 0.94 to 0.97 for twelve years; the one large trial with all-cause mortality as its primary outcome, D-Health in adults aged 60 to 84, found a hazard ratio of 1.04 (0.93 to 1.18) with monthly dosing; and a 2026 reanalysis calls the benefit a signal to be confirmed ^24,78,91,115.

Why bolus dosing increases falls, and at what size the harm begins; D-Health's monthly 60,000 IU did not raise fractures, while annual 500,000 and monthly 60,000 against 24,000 did raise falls ^37,38,100; and whether the excess of fractures on 1600 IU a day, but not 3200 IU, in the Finnish trial is chance ⁸⁸.

Whether the autoimmune signal in VITAL, an ancillary-study endpoint that lost significance when late cases were added and vanished after the pills stopped, and the MRI benefit in early multiple sclerosis, with no significant effect on relapses, mark a real immune action of pharmacological doses ^77,79,82.

Whether the free fraction of 25-hydroxyvitamin D can be measured well enough to explain why people with dark skin have lower totals and denser bone, now that the binding-protein explanation has failed ^14,70.

Why D-Health's cancer mortality went the wrong way after two years of monthly dosing, and whether that is chance, bolus dosing, or the first sign of a harm in replete people that daily-dose trials did not produce ^24,114.

Which 25-hydroxyvitamin D level, if any, should define sufficiency for the general population now that the assays are standardized; the 2024 guideline found no clear evidence of an optimal target ^14,27. In a Norwegian cohort with a mean level of 34 nmol/L, none of 423 participants had the full biochemical pattern of osteomalacia, high parathyroid hormone and alkaline phosphatase with low calcium, although the authors cautioned that without bone biopsies this cannot set a threshold ¹³².

Where it connects

Sources

132 sources, numbered as they are cited. Every one was checked against PubMed or its publisher before it was cited here; the note under each says what it shows and what it does not.

1. 1
   Norman AW, Myrtle JF, Midgett RJ, Nowicki HG, Williams V, Popják G. 1,25-dihydroxycholecalciferol: identification of the proposed active form of vitamin D3 in the intestine.Science · 1971 · 173(3991):51-4doi:10.1126/science.173.3991.51 · PMID 4325863

   Mass-spectrometric characterization of the hormone, formula C27H44O3, four to thirteen times as active as the vitamin at stimulating intestinal calcium transport.

2. 2
   Jones G. 100 YEARS OF VITAMIN D: Historical aspects of vitamin D.Endocr Connect · 2022 · 11(4):e210594doi:10.1530/EC-21-0594 · PMID 35245207

   The historical review read in full for this page: Mellanby's dogs, McCollum's oxygenated cod-liver oil, Steenbock's irradiation and patent, Windaus's Nobel Prize, Kodicek's reversal and the identification of the hormone. It states plainly that vitamin D is, strictly speaking, misnamed.

3. 3
   Bikle DD. Vitamin D metabolism, mechanism of action, and clinical applications.Chem Biol · 2014 · 21(3):319-29doi:10.1016/j.chembiol.2013.12.016 · PMID 24529992

   The reference review of the enzymes CYP2R1, CYP27B1 and CYP24A1 and of the receptor's action on vitamin D response elements.

4. 4
   Holick MF, MacLaughlin JA, Clark MB, et al. Photosynthesis of previtamin D3 in human skin and the physiologic consequences.Science · 1980 · 210(4466):203-5doi:10.1126/science.6251551 · PMID 6251551

   Previtamin D3 forms throughout the epidermis, converts by heat over at least three days, and is preferentially carried off by the binding protein.

5. 5
   MacLaughlin JA, Anderson RR, Holick MF. Spectral character of sunlight modulates photosynthesis of previtamin D3 and its photoisomers in human skin.Science · 1982 · 216(4549):1001-3doi:10.1126/science.6281884 · PMID 6281884

   The optimum wavelengths are 295 to 300 nanometers; narrow-band light converted up to 65 percent of the precursor, simulated sunlight about 20 percent, the rest going to lumisterol and tachysterol.

6. 6
   Cheng JB, Levine MA, Bell NH, Mangelsdorf DJ, Russell DW. Genetic evidence that the human CYP2R1 enzyme is a key vitamin D 25-hydroxylase.Proc Natl Acad Sci U S A · 2004 · 101(20):7711-5doi:10.1073/pnas.0402490101 · PMID 15128933

   A patient homozygous for a leucine-to-proline mutation in CYP2R1 had low 25-hydroxyvitamin D and classic deficiency, settling the identity of the liver enzyme after thirty years.

7. 7
   Takeyama K, Kitanaka S, Sato T, Kobori M, Yanagisawa J, Kato S. 25-Hydroxyvitamin D3 1alpha-hydroxylase and vitamin D synthesis.Science · 1997 · 277(5333):1827-30doi:10.1126/science.277.5333.1827 · PMID 9295274

   The kidney enzyme cloned from mice lacking the receptor, in which it ran unchecked; the negative feedback of the hormone on its own synthesis runs through the receptor.

8. 8
   Garabedian M, Holick MF, Deluca HF, Boyle IT. Control of 25-hydroxycholecalciferol metabolism by parathyroid glands.Proc Natl Acad Sci U S A · 1972 · 69(7):1673-6doi:10.1073/pnas.69.7.1673 · PMID 4340153

   Rat experiments: removing the parathyroids abolished calcitriol production within 40 hours and parathyroid extract restored it. The authors call parathyroid hormone a tropin for the hormonal form of vitamin D.

9. 9
   Shimada T, Hasegawa H, Yamazaki Y, et al. FGF-23 is a potent regulator of vitamin D metabolism and phosphate homeostasis.J Bone Miner Res · 2004 · 19(3):429-35doi:10.1359/JBMR.0301264 · PMID 15040831

   Rodent experiments showing FGF23 lowers calcitriol within hours by switching the activating enzyme off and the degrading enzyme on. The authors worked in the pharmaceutical research laboratories of Kirin Brewery.

10. 10
    Baker AR, McDonnell DP, Hughes M, et al. Cloning and expression of full-length cDNA encoding human vitamin D receptor.Proc Natl Acad Sci U S A · 1988 · 85(10):3294-8doi:10.1073/pnas.85.10.3294 · PMID 2835767

    The receptor's sequence places it in the steroid-receptor gene family, closest to the thyroid hormone receptor.

11. 11
    Norman AW. From vitamin D to hormone D: fundamentals of the vitamin D endocrine system essential for good health.Am J Clin Nutr · 2008 · 88(2):491S-499Sdoi:10.1093/ajcn/88.2.491S · PMID 18689389

    The receptor in at least 36 cell types and paracrine synthesis of the hormone in at least 10 extrarenal organs, by one of the discoverers of calcitriol; the title is the argument of this page.

12. 12
    Jones KS, Assar S, Harnpanich D, et al. 25(OH)D2 half-life is shorter than 25(OH)D3 half-life and is influenced by DBP concentration and genotype.J Clin Endocrinol Metab · 2014 · 99(9):3373-81doi:10.1210/jc.2014-1714 · PMID 24885631

    Tracer study in 36 men: plasma half-lives of 13.9 days for 25-hydroxyvitamin D2 and 15.1 for D3, both longer with more binding protein.

13. 13
    Reichel H. Optimum mode of administration of vitamin D in renal failure. Insights from experimental data.Nephrol Dial Transplant · 1995 · 10 Suppl 4:33-6doi:10.1093/ndt/10.supp4.33 · PMID 7478149

    A review of calcitriol pharmacology stating that its plasma half-life is only several hours, while its effect on the parathyroid glands persists for days; cited for the half-life only.

14. 14
    Sempos CT, Heijboer AC, Bikle DD, et al. Vitamin D assays and the definition of hypovitaminosis D: results from the First International Conference on Controversies in Vitamin D.Br J Clin Pharmacol · 2018 · 84(10):2194-2207doi:10.1111/bcp.13652 · PMID 29851137

    Assay standardization, the 12 ng/mL rickets threshold and the 20 to 50 ng/mL range judged safe and sufficient for skeletal health.

15. 15
    Ross AC, Manson JE, Abrams SA, et al. The 2011 report on dietary reference intakes for calcium and vitamin D from the Institute of Medicine: what clinicians need to know.J Clin Endocrinol Metab · 2011 · 96(1):53-8doi:10.1210/jc.2010-2704 · PMID 21118827

    The committee's own summary: 600 and 800 IU a day, 20 ng/mL, extraskeletal evidence inconclusive, inadequacy overestimated, U-shaped associations.

16. 16
    McCollum EV, Simmonds N, Becker JE, Shipley PG. Studies on experimental rickets. XXI. An experimental demonstration of the existence of a vitamin which promotes calcium deposition. 1922.J Biol Chem · 2002 · 277(19):e8 (JBC Classics reprint of the 1922 paper)PMID 11991957

    The paper that named vitamin D, reprinted as a classic; the PubMed record carries no abstract, so it is cited for the fact and date of the discovery, with the experimental detail taken from the historical review.

17. 17
    Wolf G. The discovery of vitamin D: the contribution of Adolf Windaus.J Nutr · 2004 · 134(6):1299-302doi:10.1093/jn/134.6.1299 · PMID 15173387

    A biographical article on Windaus and his 1928 Nobel Prize; the record carries no abstract and is cited with the historical review for the prize and the structures.

18. 18
    Manson JE, Cook NR, Lee IM, et al. Vitamin D Supplements and Prevention of Cancer and Cardiovascular Disease.N Engl J Med · 2019 · 380(1):33-44doi:10.1056/NEJMoa1809944 · PMID 30415629

    VITAL: 25,871 US adults, 2000 IU a day, median 5.3 years; invasive cancer hazard ratio 0.96 and major cardiovascular events 0.97.

19. 19
    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):299-309doi:10.1056/NEJMoa2202106 · PMID 35939577

    VITAL's fracture analysis: 1,991 fractures, hazard ratio 0.98 for total and 1.01 for hip, with no modification by baseline level.

20. 20
    Jackson RD, LaCroix AZ, Gass M, et al. Calcium plus vitamin D supplementation and the risk of fractures.N Engl J Med · 2006 · 354(7):669-83doi:10.1056/NEJMoa055218 · PMID 16481635

    The Women's Health Initiative: 36,282 postmenopausal women, 1000 mg calcium with 400 IU vitamin D3 for an average of seven years; hip fracture hazard ratio 0.88 (0.72 to 1.08), kidney stones 1.17 (1.02 to 1.34).

21. 21
    Munns CF, Shaw N, Kiely M, et al. Global Consensus Recommendations on Prevention and Management of Nutritional Rickets.J Clin Endocrinol Metab · 2016 · 101(2):394-415doi:10.1210/jc.2015-2175 · PMID 26745253

    33 experts from 11 organizations on the definition, diagnosis, treatment and prevention of nutritional rickets and osteomalacia.

22. 22
    Shahangian S, Alspach TD, Astles JR, Yesupriya A, Dettwyler WK. Trends in laboratory test volumes for Medicare Part B reimbursements, 2000-2010.Arch Pathol Lab Med · 2014 · 138(2):189-203doi:10.5858/arpa.2013-0149-OA · PMID 23738761

    Vitamin D was among the tests with the greatest increase in reimbursed volume per 10,000 Medicare enrollees over the decade, in a group with volume ratios from 83 to 1510.

23. 23
    Autier P, Boniol M, Pizot C, Mullie P. Vitamin D status and ill health: a systematic review.Lancet Diabetes Endocrinol · 2014 · 2(1):76-89doi:10.1016/S2213-8587(13)70165-7 · PMID 24622671

    290 cohort studies and 172 trials: strong associations, no effects of supplementation, and the proposal that low 25-hydroxyvitamin D is a marker of ill health rather than its cause.

24. 24
    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):120-128doi:10.1016/S2213-8587(21)00345-4 · PMID 35026158

    21,315 older Australians, 60,000 IU monthly for five years: all-cause mortality hazard ratio 1.04, and cancer mortality numerically higher after year two.

25. 25
    Bischoff-Ferrari HA, Vellas B, Rizzoli R, et al. Effect of Vitamin D Supplementation, Omega-3 Fatty Acid Supplementation, or a Strength-Training Exercise Program on Clinical Outcomes in Older Adults: The DO-HEALTH Randomized Clinical Trial.JAMA · 2020 · 324(18):1855-1868doi:10.1001/jama.2020.16909 · PMID 33170239

    2,157 healthy Europeans aged 70 and over: no effect of 2000 IU a day on any of six outcomes over three years.

26. 26
    Holick MF, Binkley NC, Bischoff-Ferrari HA, et al. Evaluation, treatment, and prevention of vitamin D deficiency: an Endocrine Society clinical practice guideline.J Clin Endocrinol Metab · 2011 · 96(7):1911-30doi:10.1210/jc.2011-0385 · PMID 21646368

    The 2011 guideline that recommended measuring 25-hydroxyvitamin D in people at risk and supplementing to higher intakes; its task force reported no corporate funding. Replaced in 2024.

27. 27
    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):1907-1947doi:10.1210/clinem/dgae290 · PMID 38828931

    The 2024 guideline: against empiric supplementation above the reference intake in healthy adults under 75 and against routine 25-hydroxyvitamin D testing in every population considered; daily rather than intermittent dosing.

28. 28
    Need AG, O'Loughlin PD, Morris HA, Coates PS, Horowitz M, Nordin BE. Vitamin D metabolites and calcium absorption in severe vitamin D deficiency.J Bone Miner Res · 2008 · 23(11):1859-63doi:10.1359/jbmr.080607 · PMID 18597633

    Records of 3,661 patients: calcitriol and calcium absorption are held normal by secondary hyperparathyroidism until 25-hydroxyvitamin D falls to about 10 nmol/L. The key source for why the hormone is not the test.

29. 29
    Lips P. Vitamin D deficiency and secondary hyperparathyroidism in the elderly: consequences for bone loss and fractures and therapeutic implications.Endocr Rev · 2001 · 22(4):477-501doi:10.1210/edrv.22.4.0437 · PMID 11493580

    The physiology of deficiency in the elderly and the older evidence that vitamin D3 with calcium may reduce hip fractures in nursing-home residents.

30. 30
    Sutherland JP, Zhou A, Hyppönen E. Vitamin D Deficiency Increases Mortality Risk in the UK Biobank : A Nonlinear Mendelian Randomization Study.Ann Intern Med · 2022 · 175(11):1552-1559doi:10.7326/M21-3324 · PMID 36279545

    307,601 white European participants: an L-shaped genetic relation between 25-hydroxyvitamin D and mortality, with risk falling up to 50 nmol/L and flat above. A correction was published in 2025.

31. 31
    Correction: Vitamin D Deficiency Increases Mortality Risk in the UK Biobank.Ann Intern Med · 2025 · 178(8):1215-1216doi:10.7326/ANNALS-25-02471 · PMID 40523290

    The published erratum to the 2022 Mendelian randomization paper; the PubMed record carries no text, so this page cannot say what was corrected.

32. 32
    Hamilton FW, Hughes DA, Spiller W, Tilling K, Davey Smith G. Non-linear Mendelian randomization: detection of biases using negative controls with a focus on BMI, Vitamin D and LDL cholesterol.Eur J Epidemiol · 2024 · 39(5):451-465doi:10.1007/s10654-024-01113-9 · PMID 38789826

    Negative-control tests of the residual and doubly-ranked nonlinear Mendelian randomization methods, with vitamin D among the exposures: clearly biased estimates in certain situations, and a call for justification and sanity checks before further use.

33. 33
    Clemens TL, Adams JS, Henderson SL, Holick MF. Increased skin pigment reduces the capacity of skin to synthesise vitamin D3.Lancet · 1982 · 1(8263):74-6doi:10.1016/s0140-6736(82)90214-8 · PMID 6119494

    Five volunteers: one standard dose raised serum vitamin D up to 60-fold in the two lightly pigmented and not significantly in the three heavily pigmented; one of the latter, re-exposed to six times the dose, reached the lightly pigmented level.

34. 34
    Wortsman J, Matsuoka LY, Chen TC, Lu Z, Holick MF. Decreased bioavailability of vitamin D in obesity.Am J Clin Nutr · 2000 · 72(3):690-3doi:10.1093/ajcn/72.3.690 · PMID 10966885

    Obese volunteers had a 57 percent smaller rise in blood vitamin D3 after whole-body irradiation and lower peak D2 after an oral dose, with the same precursor content and conversion in their skin; the authors attribute it to deposition in fat.

35. 35
    Marcinowska-Suchowierska E, Kupisz-Urbańska M, Łukaszkiewicz J, Płudowski P, Jones G. Vitamin D Toxicity-A Clinical Perspective.Front Endocrinol (Lausanne) · 2018 · 9:550doi:10.3389/fendo.2018.00550 · PMID 30294301

    The clinical review of toxicity: symptoms, the 150 ng/mL hallmark, exogenous and endogenous causes, and hypersensitivity to vitamin D.

36. 36
    Galior K, Grebe S, Singh R. Development of Vitamin D Toxicity from Overcorrection of Vitamin D Deficiency: A Review of Case Reports.Nutrients · 2018 · 10(8):953doi:10.3390/nu10080953 · PMID 30042334

    13 reports: 25-hydroxyvitamin D of 150 to 1,220 ng/mL and calcium of 11.1 to 23.1 mg/dL, from manufacturing errors and overdosing; from a Mayo Clinic laboratory.

37. 37
    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):1815-22doi:10.1001/jama.2010.594 · PMID 20460620

    2,256 women aged 70 and over: a single annual 500,000 IU dose increased falls (rate ratio 1.15) and fractures (1.26), most in the three months after dosing.

38. 38
    Bischoff-Ferrari HA, Dawson-Hughes B, Orav EJ, et al. Monthly High-Dose Vitamin D Treatment for the Prevention of Functional Decline: A Randomized Clinical Trial.JAMA Intern Med · 2016 · 176(2):175-83doi:10.1001/jamainternmed.2015.7148 · PMID 26747333

    200 adults aged 70 and over with a prior fall: monthly 60,000 IU or 24,000 IU plus calcifediol reached 30 ng/mL more often, did nothing for function, and produced more fallers than 24,000 IU.

39. 39
    Binkley N, Sempos CT, Borchardt G, et al. Association of standardized serum 25-hydroxyvitamin D with falls in post-menopausal women.Osteoporos Int · 2025 · 36(2):255-264doi:10.1007/s00198-024-07310-2 · PMID 39585370

    2,303 women with standardized assays: no excess falls up to 60 ng/mL and twice the odds of two or more falls at 60 or above, a group of about 2 percent of observations.

40. 40
    Kahwati LC, Weber RP, Pan H, et al. Vitamin D, Calcium, or Combined Supplementation for the Primary Prevention of Fractures in Community-Dwelling Adults: Evidence Report and Systematic Review for the US Preventive Services Task Force.JAMA · 2018 · 319(15):1600-1612doi:10.1001/jama.2017.21640 · PMID 29677308

    Eleven trials, 51,419 adults: no reduction in fractures; vitamin D with calcium raised kidney stones by an absolute 0.33 percent.

41. 41
    Schlingmann KP, Kaufmann M, Weber S, et al. Mutations in CYP24A1 and idiopathic infantile hypercalcemia.N Engl J Med · 2011 · 365(5):410-21doi:10.1056/NEJMoa1103864 · PMID 21675912

    Recessive loss-of-function mutations in the hormone-degrading enzyme in children with idiopathic infantile hypercalcemia and in infants with severe hypercalcemia after bolus prophylaxis; the 1950s British increase is cited as the background to the dose debate.

42. 42
    Tripkovic L, Lambert H, Hart K, et al. Comparison of vitamin D2 and vitamin D3 supplementation in raising serum 25-hydroxyvitamin D status: a systematic review and meta-analysis.Am J Clin Nutr · 2012 · 95(6):1357-64doi:10.3945/ajcn.111.031070 · PMID 22552031

    D3 raised 25-hydroxyvitamin D more than D2, clearly with bolus dosing and not significantly with daily dosing.

43. 43
    Haussler MR, Norman AW. Chromosomal receptor for a vitamin D metabolite.Proc Natl Acad Sci U S A · 1969 · 62(1):155-62doi:10.1073/pnas.62.1.155 · PMID 5253652

    The first evidence of the receptor: a saturable, high-affinity binding protein in chick intestinal chromatin, purified 167-fold.

44. 44
    Hughes MR, Malloy PJ, Kieback DG, et al. Point mutations in the human vitamin D receptor gene associated with hypocalcemic rickets.Science · 1988 · 242(4886):1702-5doi:10.1126/science.2849209 · PMID 2849209

    Two families with hereditary vitamin D-resistant rickets carried single-nucleotide mutations in the receptor's DNA-binding domain, one near the tip of the first zinc finger and one at the tip of the second; the receptor bound hormone normally but had decreased affinity for DNA.

45. 45
    Powe CE, Evans MK, Wenger J, et al. Vitamin D-binding protein and vitamin D status of black Americans and white Americans.N Engl J Med · 2013 · 369(21):1991-2000doi:10.1056/NEJMoa1306357 · PMID 24256378

    The study that proposed that black Americans' low total 25-hydroxyvitamin D reflects low binding protein, measured by immunoassay; the binding-protein result was overturned in 2016.

46. 46
    Matsuoka LY, Ide L, Wortsman J, MacLaughlin JA, Holick MF. Sunscreens suppress cutaneous vitamin D3 synthesis.J Clin Endocrinol Metab · 1987 · 64(6):1165-8doi:10.1210/jcem-64-6-1165 · PMID 3033008

    Eight subjects; a para-aminobenzoic acid sunscreen of factor 8 blocked the rise in serum vitamin D3 and the photoconversion in skin slices.

47. 47
    Neale RE, Khan SR, Lucas RM, Waterhouse M, Whiteman DC, Olsen CM. The effect of sunscreen on vitamin D: a review.Br J Dermatol · 2019 · 181(5):907-915doi:10.1111/bjd.17980 · PMID 30945275

    Four experimental studies, three field trials and 69 observational studies: sunscreen blocks synthesis under a lamp and makes little difference in real life, with no trials of the high-factor sunscreens now recommended.

48. 48
    DeLuca HF. History of the discovery of vitamin D and its active metabolites.Bonekey Rep · 2014 · 3:479doi:10.1038/bonekey.2013.213 · PMID 24466410

    DeLuca's own account of the discovery of the metabolites, from the laboratory that found 25-hydroxyvitamin D; the abstract states that vitamin D is converted to a hormone that acts through a nuclear receptor.

49. 49
    Blunt JW, DeLuca HF, Schnoes HK. 25-hydroxycholecalciferol. A biologically active metabolite of vitamin D3.Biochemistry · 1968 · 7(10):3317-22doi:10.1021/bi00850a001 · PMID 4300699

    The identification of the first metabolite, the one the blood test measures; the record carries no abstract.

50. 50
    Fraser DR, Kodicek E. Unique biosynthesis by kidney of a biological active vitamin D metabolite.Nature · 1970 · 228(5273):764-6doi:10.1038/228764a0 · PMID 4319631

    The kidney as the site of the hormone's synthesis, shown with nephrectomized animals; indexed under kidney and nephrectomy, no abstract.

51. 51
    Lawson DE, Fraser DR, Kodicek E, Morris HR, Williams DH. Identification of 1,25-dihydroxycholecalciferol, a new kidney hormone controlling calcium metabolism.Nature · 1971 · 230(5291):228-30doi:10.1038/230228a0 · PMID 4323313

    Kodicek's group's identification of the kidney hormone, published in March 1971; the record carries no abstract, so it is cited for the identification and its date.

52. 52
    Holick MF, Schnoes HK, DeLuca HF, Suda T, Cousins RJ. Isolation and identification of 1,25-dihydroxycholecalciferol. A metabolite of vitamin D active in intestine.Biochemistry · 1971 · 10(14):2799-804doi:10.1021/bi00790a023 · PMID 4326883

    The isolation of calcitriol from chick intestine by DeLuca's group; no abstract on the record.

53. 53
    Safadi FF, Thornton P, Magiera H, et al. Osteopathy and resistance to vitamin D toxicity in mice null for vitamin D binding protein.J Clin Invest · 1999 · 103(2):239-51doi:10.1172/JCI5244 · PMID 9916136

    Mice without the binding protein: normal when fed, deficient quickly when starved, and unexpectedly resistant to toxicity after an overload.

54. 54
    Daiger SP, Schanfield MS, Cavalli-Sforza LL. Group-specific component (Gc) proteins bind vitamin D and 25-hydroxyvitamin D.Proc Natl Acad Sci U S A · 1975 · 72(6):2076-80doi:10.1073/pnas.72.6.2076 · PMID 49052

    Showed that Gc, a plasma protein with a worldwide polymorphism and no known function, is the vitamin D-binding globulin.

55. 55
    Henderson CM, Fink SL, Bassyouni H, et al. Vitamin D-Binding Protein Deficiency and Homozygous Deletion of the GC Gene.N Engl J Med · 2019 · 380(12):1150-1157doi:10.1056/NEJMoa1807841 · PMID 30893535

    A single case report: a 58-year-old woman with no circulating binding protein from a homozygous GC deletion, apparent severe deficiency unresponsive to supplementation, normal calcium and mild disruption of bone metabolism, with severe autoimmune disease.

56. 56
    Best CM, Sherwood R, Novotny JA, Zhang S, Pressman EK, O'Brien KO. Vitamin D kinetics in nonpregnant and pregnant women after a single oral dose of trideuterated vitamin D3.J Steroid Biochem Mol Biol · 2022 · 216:106034doi:10.1016/j.jsbmb.2021.106034 · PMID 34843870

    A pilot tracer study: 25-hydroxyvitamin D3 half-life 13.6 days in nonpregnant and 18.8 in pregnant women, tracking the binding protein.

57. 57
    Carter GD. 25-hydroxyvitamin D: a difficult analyte.Clin Chem · 2012 · 58(3):486-8doi:10.1373/clinchem.2011.180562 · PMID 22238254

    An editorial on the disagreement between immunoassays and mass spectrometry; no abstract on the record.

58. 58
    US Preventive Services Task Force; Krist AH, Davidson KW, et al. Screening for Vitamin D Deficiency in Adults: US Preventive Services Task Force Recommendation Statement.JAMA · 2021 · 325(14):1436-1442doi:10.1001/jama.2021.3069 · PMID 33847711

    An I statement: evidence insufficient on screening asymptomatic adults, and no single serum cutpoint defines deficiency.

59. 59
    Rajakumar K. Vitamin D, cod-liver oil, sunlight, and rickets: a historical perspective.Pediatrics · 2003 · 112(2):e132-5doi:10.1542/peds.112.2.e132 · PMID 12897318

    Glisson's 1650 treatise, the folklore of cod-liver oil and the geography of sunshine, from a pediatric historian.

60. 60
    Mellanby E. An experimental investigation on rickets. 1919.Nutrition · 1989 · 5(2):81-6 (reprint of the 1919 Lancet paper)PMID 2520279

    Mellanby's dog experiments, reprinted; the record carries no abstract and is indexed under dogs, diet and rickets. Cited for the experiment's existence and date.

61. 61
    Schneider HA. Harry Steenbock (1886-1967)--a biographical sketch.J Nutr · 1973 · 103(9):1233-47doi:10.1093/jn/103.9.1233 · PMID 4579907

    The biography of Steenbock, indexed under food irradiation and foundations; no abstract. Cited with the historical review for the irradiation work and the Wisconsin Alumni Research Foundation.

62. 62
    Garland CF, Garland FC. Do sunlight and vitamin D reduce the likelihood of colon cancer?Int J Epidemiol · 1980 · 9(3):227-31doi:10.1093/ije/9.3.227 · PMID 7440046

    The ecological paper that started the cancer hypothesis: colon cancer mortality highest where natural light was least; the authors discuss the possibility of an ecological fallacy.

63. 63
    Chapuy MC, Arlot ME, Duboeuf F, et al. Vitamin D3 and calcium to prevent hip fractures in elderly women.N Engl J Med · 1992 · 327(23):1637-42doi:10.1056/NEJM199212033272305 · PMID 1331788

    3,270 ambulatory French women with a mean age of 84: 800 IU of vitamin D3 and 1.2 g of calcium a day for 18 months, 43 percent fewer hip fractures among completers.

64. 64
    Trivedi DP, Doll R, Khaw KT. Effect of four monthly oral vitamin D3 (cholecalciferol) supplementation on fractures and mortality in men and women living in the community: randomised double blind controlled trial.BMJ · 2003 · 326(7387):469doi:10.1136/bmj.326.7387.469 · PMID 12609940

    2,686 British adults aged 65 to 85: 100,000 IU every four months for five years, first fracture relative risk 0.78 (0.61 to 0.99).

65. 65
    Lappe JM, Travers-Gustafson D, Davies KM, Recker RR, Heaney RP. Vitamin D and calcium supplementation reduces cancer risk: results of a randomized trial.Am J Clin Nutr · 2007 · 85(6):1586-91doi:10.1093/ajcn/85.6.1586 · PMID 17556697

    1,179 Nebraska women in a fracture trial: cancer, the principal secondary outcome, at an unadjusted relative risk of 0.402 with calcium plus 1100 IU vitamin D3.

66. 66
    Holick MF. Vitamin D deficiency.N Engl J Med · 2007 · 357(3):266-81doi:10.1056/NEJMra070553 · PMID 17634462

    Holick's 2007 review of vitamin D deficiency; the PubMed record carries no abstract, so it is cited as a historical document only, for its existence and date.

67. 67
    Bischoff-Ferrari HA, Dawson-Hughes B, Staehelin HB, et al. Fall prevention with supplemental and active forms of vitamin D: a meta-analysis of randomised controlled trials.BMJ · 2009 · 339:b3692doi:10.1136/bmj.b3692 · PMID 19797342

    Eight double-blind trials, 2,426 people: 700 to 1000 IU a day reduced the risk of falling by 19 percent (relative risk 0.81); lower doses did not.

68. 68
    Moyer VA; US Preventive Services Task Force. Prevention of falls in community-dwelling older adults: U.S. Preventive Services Task Force recommendation statement.Ann Intern Med · 2012 · 157(3):197-204doi:10.7326/0003-4819-157-3-201208070-00462 · PMID 22868837

    The 2012 B recommendation for exercise or physical therapy and vitamin D to prevent falls in adults 65 and older at increased risk, reversed in 2018.

69. 69
    US Preventive Services Task Force; Grossman DC, Curry SJ, et al. Interventions to Prevent Falls in Community-Dwelling Older Adults: US Preventive Services Task Force Recommendation Statement.JAMA · 2018 · 319(16):1696-1704doi:10.1001/jama.2018.3097 · PMID 29710141

    A D recommendation against vitamin D to prevent falls, with adequate evidence of no benefit and of small to moderate harms.

70. 70
    Henderson CM, Lutsey PL, Misialek JR, et al. Measurement by a Novel LC-MS/MS Methodology Reveals Similar Serum Concentrations of Vitamin D-Binding Protein in Blacks and Whites.Clin Chem · 2016 · 62(1):179-87doi:10.1373/clinchem.2015.244541 · PMID 26453697

    Mass spectrometry found binding protein of 262 against 266 micrograms per milliliter in white and black participants and showed the immunoassay was isoform-sensitive.

71. 71
    Martineau AR, Jolliffe DA, Hooper RL, et al. Vitamin D supplementation to prevent acute respiratory tract infections: systematic review and meta-analysis of individual participant data.BMJ · 2017 · 356:i6583doi:10.1136/bmj.i6583 · PMID 28202713

    25 trials, 11,321 people: adjusted odds ratio 0.88, with the effect in daily or weekly dosing and strongest below 25 nmol/L. Superseded by the 2025 update.

72. 72
    Lappe J, Watson P, Travers-Gustafson D, et al. Effect of Vitamin D and Calcium Supplementation on Cancer Incidence in Older Women: A Randomized Clinical Trial.JAMA · 2017 · 317(12):1234-1243doi:10.1001/jama.2017.2115 · PMID 28350929

    The trial built for cancer: 2,303 women, 2000 IU vitamin D3 and 1500 mg calcium for four years; hazard ratio 0.70 (0.47 to 1.02), not significant. One author lists an affiliation with GrassrootsHealth.

73. 73
    Pittas AG, Dawson-Hughes B, Sheehan P, et al. Vitamin D Supplementation and Prevention of Type 2 Diabetes.N Engl J Med · 2019 · 381(6):520-530doi:10.1056/NEJMoa1900906 · PMID 31173679

    D2d: 2,423 adults with prediabetes, 4000 IU a day; diabetes hazard ratio 0.88 (0.75 to 1.04), not significant.

74. 74
    Entrenas Castillo M, Entrenas Costa LM, Vaquero Barrios JM, et al. Effect of calcifediol treatment and best available therapy versus best available therapy on intensive care unit admission and mortality among patients hospitalized for COVID-19: A pilot randomized clinical study.J Steroid Biochem Mol Biol · 2020 · 203:105751doi:10.1016/j.jsbmb.2020.105751 · PMID 32871238

    An open-label pilot in 76 patients in Córdoba, allocated two to one: intensive care in 1 of 50 on calcifediol and 13 of 26 without it; the authors call for larger trials with properly matched groups.

75. 75
    Murai IH, Fernandes AL, Sales LP, et al. Effect of a Single High Dose of Vitamin D3 on Hospital Length of Stay in Patients With Moderate to Severe COVID-19: A Randomized Clinical Trial.JAMA · 2021 · 325(11):1053-1060doi:10.1001/jama.2020.26848 · PMID 33595634

    240 hospitalized Brazilian patients: 200,000 IU raised the mean level to 44.4 ng/mL against 19.8 on placebo and did not shorten the stay or change mortality, intensive care or ventilation.

76. 76
    Jolliffe DA, Holt H, Greenig M, et al. Effect of a test-and-treat approach to vitamin D supplementation on risk of all cause acute respiratory tract infection and covid-19: phase 3 randomised controlled trial (CORONAVIT).BMJ · 2022 · 378:e071230doi:10.1136/bmj-2022-071230 · PMID 36215226

    6,200 British adults, open label: offering testing and six months of 800 or 3200 IU a day did not reduce respiratory infections or COVID-19.

77. 77
    Hahn J, Cook NR, Alexander EK, et al. Vitamin D and marine omega 3 fatty acid supplementation and incident autoimmune disease: VITAL randomized controlled trial.BMJ · 2022 · 376:e066452doi:10.1136/bmj-2021-066452 · PMID 35082139

    The primary endpoint of a prespecified ancillary study of VITAL: confirmed autoimmune disease in 123 against 155, hazard ratio 0.78 with the upper limit at 0.99. The effect faded in the 2024 follow-up.

78. 78
    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):1961-1974doi:10.1210/clinem/dgae312 · PMID 38828942

    151 studies: a very small, high-certainty mortality reduction over 75, moderate-certainty benefit in prediabetes, low-certainty fewer respiratory infections in children, and possible excess falls with intermittent high doses.

79. 79
    Costenbader KH, Cook NR, Lee IM, et al. Vitamin D and Marine n-3 Fatty Acids for Autoimmune Disease Prevention: Outcomes Two Years After Completion of a Double-Blind, Placebo-Controlled Trial.Arthritis Rheumatol · 2024 · 76(6):973-983doi:10.1002/art.42811 · PMID 38272846

    VITAL's autoimmune follow-up: with 65 late-confirmed cases the randomized hazard ratio for vitamin D became 0.85 (0.70 to 1.04), and at seven years 0.98 (0.83 to 1.17); the omega-3 effect persisted.

80. 80
    US Preventive Services Task Force; Nicholson WK, Silverstein M, Wong JB, et al. Interventions to Prevent Falls in Community-Dwelling Older Adults: US Preventive Services Task Force Recommendation Statement.JAMA · 2024 · 332(1):51-57doi:10.1001/jama.2024.8481 · PMID 38833246

    The 2024 falls statement that replaced the 2018 one: exercise (B) and individualized multifactorial interventions (C); the abstract does not address vitamin D.

81. 81
    Jolliffe DA, Camargo CA, Sluyter JD, et al. Vitamin D supplementation to prevent acute respiratory infections: systematic review and meta-analysis of stratified aggregate data.Lancet Diabetes Endocrinol · 2025 · 13(4):307-320doi:10.1016/S2213-8587(24)00348-6 · PMID 39993397

    An update of the group's 2021 aggregate-data meta-analysis: 40 studies, 61,589 participants, odds ratio 0.94 (0.88 to 1.00), no subgroup modification, funnel-plot asymmetry; the authors conclude there is no statistically significant protection.

82. 82
    Thouvenot E, Laplaud D, Lebrun-Frenay C, et al. High-Dose Vitamin D in Clinically Isolated Syndrome Typical of Multiple Sclerosis: The D-Lay MS Randomized Clinical Trial.JAMA · 2025 · 333(16):1413-1422doi:10.1001/jama.2025.1604 · PMID 40063041

    316 French patients: 100,000 IU every two weeks reduced disease activity (hazard ratio 0.66), driven by MRI outcomes, with no significant effect on relapses or any clinical outcome.

83. 83
    Retraction: Pre-infection 25-hydroxyvitamin D3 levels and association with severity of COVID-19 illness.PLoS One · 2025 · 20(9):e0331693doi:10.1371/journal.pone.0331693 · PMID 40920624

    The retraction notice: an independent editorial board member judged the design fatally flawed, because only patients tested before infection were included, with measurements 14 to 730 days earlier and possible corrective treatment in between; most authors did not agree.

84. 84
    RETRACTED: Argano et al. Protective Effect of Vitamin D Supplementation on COVID-19-Related Intensive Care Hospitalization and Mortality: Definitive Evidence from Meta-Analysis and Trial Sequential Analysis. Pharmaceuticals 2023, 16, 130.Pharmaceuticals (Basel) · 2025 · 18(11):1645doi:10.3390/ph18111645 · PMID 41172117

    The retraction notice: the journal's investigation confirmed that at least one included study was not a randomized trial; the authors disagree with the retraction.

85. 85
    Massé O, Mercurio CM, Dupuis S, et al. Calcium, vitamin D, or combined supplementation to prevent fractures and falls: systematic review and meta-analysis.BMJ · 2026 · 393:e088050doi:10.1136/bmj-2025-088050 · PMID 42161415

    69 trials, 153,902 participants: vitamin D alone has little to no effect on any fracture (risk ratio 1.00, high certainty) or falls; combined supplementation 0.91, judged below the threshold of clinical importance.

86. 86
    Wang Y, Sha S, Gwenzi T, Schöttker B, Brenner H. Impact of vitamin D supplementation on all-cause mortality: Randomized trials revisited.Clin Nutr · 2026 · 58:106597doi:10.1016/j.clnu.2026.106597 · PMID 41719624

    Observational emulations of VITAL and D-Health in the UK Biobank: null when weighted to the trials' levels, lower predicted mortality in insufficient and deficient people. Not a randomized comparison.

87. 87
    May HT, Le VT, Anderson JL, et al. Targeted vitamin D supplementation and major adverse cardiovascular events after myocardial infarction: the TARGET-D trial.Eur Heart J · 2026doi:10.1093/eurheartj/ehag611 · PMID 42702500

    630 patients after heart attack randomized to vitamin D3 titrated to 40 to 80 ng/mL or usual care, not placebo: major adverse cardiovascular events hazard ratio 0.85 (0.58 to 1.24), not significant; repeat heart attack 0.48 as one of four secondary endpoints.

88. 88
    Rikkonen T, Hantunen S, Kröger H, et al. The effect of vitamin D3 supplementation on fracture incidence in an aging population - The Finnish vitamin D trial.Bone · 2026 · 213:118081doi:10.1016/j.bone.2026.118081 · PMID 42704968

    2,495 older Finns: fractures hazard ratio 1.52 (1.06 to 2.18) on 1600 IU a day and 1.07 on 3200 IU against placebo over 4.1 years; the authors conclude high doses are not supported for fracture prevention in largely replete people.

89. 89
    Tripkovic L, Wilson LR, Hart K, et al. Daily supplementation with 15 μg vitamin D2 compared with vitamin D3 to increase wintertime 25-hydroxyvitamin D status in healthy South Asian and white European women: a 12-wk randomized, placebo-controlled food-fortification trial.Am J Clin Nutr · 2017 · 106(2):481-490doi:10.3945/ajcn.116.138693 · PMID 28679555

    335 women: at a daily public-health dose, D3 still raised the level by about 15 to 17 nmol/L more than D2.

90. 90
    Zhou F, Jamilian A, Prabahar K, Hernández-Wolters B, Kord-Varkaneh H, Bai D. The effect of vitamin D2 supplementation on vitamin D levels in humans: A time and dose-response meta-analysis of randomized controlled trials.Steroids · 2024 · 205:109394doi:10.1016/j.steroids.2024.109394 · PMID 38458370

    33 study arms: D2 raised total 25-hydroxyvitamin D by 11.47 ng/mL and lowered 25-hydroxyvitamin D3 by 4.63 ng/mL.

91. 91
    Bjelakovic G, Gluud LL, Nikolova D, et al. Vitamin D supplementation for prevention of mortality in adults.Cochrane Database Syst Rev · 2014 · 2014(1):CD007470doi:10.1002/14651858.CD007470.pub3 · PMID 24414552

    56 trials, 95,286 people: mortality risk ratio 0.97, with the effect confined to vitamin D3 (0.94; 150 treated for five years to prevent one death) and warnings about attrition and reporting bias.

92. 92
    Zhang Y, Fang F, Tang J, et al. Association between vitamin D supplementation and mortality: systematic review and meta-analysis.BMJ · 2019 · 366:l4673doi:10.1136/bmj.l4673 · PMID 31405892

    52 trials, 75,454 people: all-cause mortality risk ratio 0.98, cancer death 0.84, and D3 trials better than D2 trials.

93. 93
    Wilson LR, Tripkovic L, Hart KH, Lanham-New SA. Vitamin D deficiency as a public health issue: using vitamin D2 or vitamin D3 in future fortification strategies.Proc Nutr Soc · 2017 · 76(3):392-399doi:10.1017/S0029665117000349 · PMID 28347378

    A review of the D2-versus-D3 evidence for fortification policy, noting that most trials favor D3 and, in a parenthesis, that D3 is not suitable for vegans.

94. 94
    Wang T, Bengtsson G, Kärnefelt I, Björn LO. Provitamins and vitamins D2 and D3 in Cladina spp. over a latitudinal gradient: possible correlation with UV levels.J Photochem Photobiol B · 2001 · 62(1-2):118-22doi:10.1016/s1011-1344(01)00160-9 · PMID 11693362

    Vitamin D3 identified in the lichen Cladina arbuscula, a non-animal source, at 0.67 to 2.04 micrograms per gram dry matter.

95. 95
    Scragg R, Stewart AW, Waayer D, et al. Effect of Monthly High-Dose Vitamin D Supplementation on Cardiovascular Disease in the Vitamin D Assessment Study : A Randomized Clinical Trial.JAMA Cardiol · 2017 · 2(6):608-616doi:10.1001/jamacardio.2017.0175 · PMID 28384800

    ViDA: 5,108 New Zealanders, 100,000 IU monthly; cardiovascular disease hazard ratio 1.02, including in the deficient quarter.

96. 96
    Okereke OI, Reynolds CF, Mischoulon D, et al. Effect of Long-term Vitamin D3 Supplementation vs Placebo on Risk of Depression or Clinically Relevant Depressive Symptoms and on Change in Mood Scores: A Randomized Clinical Trial.JAMA · 2020 · 324(5):471-480doi:10.1001/jama.2020.10224 · PMID 32749491

    VITAL-DEP, 18,353 adults: hazard ratio 0.97 for depression and no change in mood scores.

97. 97
    Thompson B, Waterhouse M, English DR, et al. Vitamin D supplementation and major cardiovascular events: D-Health randomised controlled trial.BMJ · 2023 · 381:e075230doi:10.1136/bmj-2023-075230 · PMID 37380191

    21,302 D-Health participants: major cardiovascular events hazard ratio 0.91 (0.81 to 1.01), heart attack 0.81, stroke 0.99; number needed to treat 172 over five years; the interval is consistent with no effect.

98. 98
    Neale RE, English DR, McLeod DSA, et al. The effect of vitamin D supplementation on cancer incidence in the randomised controlled D-Health Trial: Implications for policy and practice.J Steroid Biochem Mol Biol · 2025 · 250:106738doi:10.1016/j.jsbmb.2025.106738 · PMID 40096917

    21,308 D-Health participants: no difference in cancer incidence, hazard ratio 1.02 (0.95 to 1.10).

99. 99
    Duarte Romero BL, Armstrong BK, Baxter C, et al. Effect of Vitamin D Supplementation on Cardiometabolic Outcomes in Older Australian Adults-Results from the Randomized Controlled D-Health Trial.Nutrients · 2026 · 18(2):357doi:10.3390/nu18020357 · PMID 41599971

    Starting drugs for high blood pressure, high cholesterol or type 2 diabetes, used as surrogates for new disease: hazard ratios 1.00, 1.05 and 0.97.

100. 100
     Waterhouse M, Ebeling PR, McLeod DSA, et al. The effect of monthly vitamin D supplementation on fractures: a tertiary outcome from the population-based, double-blind, randomised, placebo-controlled D-Health trial.Lancet Diabetes Endocrinol · 2023 · 11(5):324-332doi:10.1016/S2213-8587(23)00063-3 · PMID 37011645

     No effect on fractures overall (hazard ratio 0.94) and no sign that the monthly bolus raised fracture risk.

101. 101
     Khaw KT, Stewart AW, Waayer D, et al. Effect of monthly high-dose vitamin D supplementation on falls and non-vertebral fractures: secondary and post-hoc outcomes from the randomised, double-blind, placebo-controlled ViDA trial.Lancet Diabetes Endocrinol · 2017 · 5(6):438-447doi:10.1016/S2213-8587(17)30103-1 · PMID 28461159

     Falls hazard ratio 0.99 and nonvertebral fractures 1.19 (0.94 to 1.50) with monthly bolus dosing.

102. 102
     Pittas AG, Kawahara T, Jorde R, et al. Vitamin D and Risk for Type 2 Diabetes in People With Prediabetes : A Systematic Review and Meta-analysis of Individual Participant Data From 3 Randomized Clinical Trials.Ann Intern Med · 2023 · 176(3):355-363doi:10.7326/M22-3018 · PMID 36745886

     Three trials pooled: hazard ratio 0.85, absolute reduction 3.3 percent over three years. The 76 percent reduction in people reaching 125 nmol/L is a comparison by achieved level, not a randomized one; one author lists an affiliation with Takeda Pharmaceuticals.

103. 103
     Dawson-Hughes B, Huggins GS, Nelson J, Vickery E, Powers SN, Pittas AG. Vitamin D Receptor Polymorphisms and the Effect of Vitamin D Supplementation on Diabetes Risk Among Adults With Prediabetes.JAMA Netw Open · 2026 · 9(4):e267332doi:10.1001/jamanetworkopen.2026.7332 · PMID 42024385

     D2d genetic substudy of 2,098 participants: benefit confined to ApaI AC and CC carriers (hazard ratio 0.81) and absent in AA carriers (1.02). A hypothesis-generating subgroup analysis.

104. 104
     Ganmaa D, Uyanga B, Zhou X, et al. Vitamin D Supplements for Prevention of Tuberculosis Infection and Disease.N Engl J Med · 2020 · 383(4):359-368doi:10.1056/NEJMoa1915176 · PMID 32706534

     8,851 Mongolian schoolchildren, 95.6 percent deficient: three years of 14,000 IU weekly did not reduce tuberculosis infection, disease or respiratory infection.

105. 105
     Zhao JG, Zeng XT, Wang J, Liu L. Association Between Calcium or Vitamin D Supplementation and Fracture Incidence in Community-Dwelling Older Adults: A Systematic Review and Meta-analysis.JAMA · 2017 · 318(24):2466-2482doi:10.1001/jama.2017.19344 · PMID 29279934

     33 trials, 51,145 people: no reduction in hip or any fracture with calcium, vitamin D or both.

106. 106
     US Preventive Services Task Force; Grossman DC, Curry SJ, et al. Vitamin D, Calcium, or Combined Supplementation for the Primary Prevention of Fractures in Community-Dwelling Adults: US Preventive Services Task Force Recommendation Statement.JAMA · 2018 · 319(15):1592-1599doi:10.1001/jama.2018.3185 · PMID 29677309

     Against low-dose vitamin D with calcium for fractures in postmenopausal women; insufficient evidence otherwise; increased kidney stones.

107. 107
     Dror AA, Morozov N, Daoud A, et al. Pre-infection 25-hydroxyvitamin D3 levels and association with severity of COVID-19 illness.PLoS One · 2022 · 17(2):e0263069 (retracted 2025)doi:10.1371/journal.pone.0263069 · PMID 35113901

     RETRACTED. A retrospective study of 253 patients with pre-infection vitamin D measurements: an odds ratio of 14 for severe or critical disease below 20 ng/mL. Cited only for the claim and its retraction.

108. 108
     Argano C, Mallaci Bocchio R, Natoli G, Scibetta S, Lo Monaco M, Corrao S. Protective Effect of Vitamin D Supplementation on COVID-19-Related Intensive Care Hospitalization and Mortality: Definitive Evidence from Meta-Analysis and Trial Sequential Analysis.Pharmaceuticals (Basel) · 2023 · 16(1):130 (retracted 2025)doi:10.3390/ph16010130 · PMID 36678627

     RETRACTED. A meta-analysis of five studies that claimed definitive evidence of protection against intensive-care admission. Cited only for the claim and its retraction.

109. 109
     Meltzer DO, Moy JN, Schram AW, et al. Effect of moderate or high versus low dose daily vitamin D on COVID-19 incidence: A double-blind, parallel-group randomized clinical trial.Clin Nutr · 2026 · 64:106722doi:10.1016/j.clnu.2026.106722 · PMID 42480239

     923 Chicago adults: 4000 or 10,000 IU against 400 IU a day did not significantly lower COVID-19 (hazard ratio 0.73, 0.51 to 1.06); 4000 IU alone 0.62 in a secondary analysis.

110. 110
     Martineau AR. Vitamin D for tuberculosis or acute respiratory infections: lost in translation?Curr Opin Infect Dis · 2026 · 39(2):147-152doi:10.1097/QCO.0000000000001181 · PMID 41549871

     A 2026 review by the senior author of the 2025 meta-analysis: the latest trials do not support vitamin D to prevent respiratory infection or tuberculosis; publication and citation bias among the reasons for earlier optimism.

111. 111
     van Arragon M, Grant CC, Scragg RK, Jordan VM. Vitamin D for preventing acute respiratory infections in children up to five years of age.Cochrane Database Syst Rev · 2026 · 4:CD015111doi:10.1002/14651858.CD015111.pub2 · PMID 42037591

     107 trials, 31,521 participants from pregnancy to age five: low-certainty evidence of a slight reduction in the proportion of children with respiratory-infection visits (risk ratio 0.95), and probably no reduction in visits per child.

112. 112
     Chandler PD, Chen WY, Ajala ON, et al. Effect of Vitamin D3 Supplements on Development of Advanced Cancer: A Secondary Analysis of the VITAL Randomized Clinical Trial.JAMA Netw Open · 2020 · 3(11):e2025850doi:10.1001/jamanetworkopen.2020.25850 · PMID 33206192

     A secondary analysis of VITAL: metastatic or fatal cancer 1.7 against 2.1 percent, hazard ratio 0.83 (0.69 to 0.99), confined to people of normal weight.

113. 113
     Bjelakovic G, Gluud LL, Nikolova D, et al. Vitamin D supplementation for prevention of cancer in adults.Cochrane Database Syst Rev · 2014 · 2014(6):CD007469doi:10.1002/14651858.CD007469.pub2 · PMID 24953955

     18 trials, 50,623 people: no effect on cancer occurrence, most trials at high risk of for-profit bias, and a cancer-mortality reduction with D3 that sequential analysis judged possibly random.

114. 114
     Kuznia S, Zhu A, Akutsu T, et al. Efficacy of vitamin D3 supplementation on cancer mortality: Systematic review and individual patient data meta-analysis of randomised controlled trials.Ageing Res Rev · 2023 · 87:101923doi:10.1016/j.arr.2023.101923 · PMID 37004841

     14 trials, 104,727 people: cancer mortality risk ratio 0.94 overall, 0.88 with daily dosing and 1.07 with bolus dosing.

115. 115
     Kostakis N, Lega JC, Boussageon R. Vitamin D Supplementation and Mortality in the Elderly: A Review of a Meta-Analysis of Randomized Controlled Trials Using the REB Method.Fundam Clin Pharmacol · 2026 · 40(6):e70112doi:10.1111/fcp.70112 · PMID 42823826

     A reanalysis of the Endocrine Society's mortality evidence admitting only low-risk-of-bias trials: 8 trials, 30,320 people, risk ratio 0.96 (0.92 to 1.00), described as a signal to be confirmed.

116. 116
     Rooney MR, Harnack L, Michos ED, Ogilvie RP, Sempos CT, Lutsey PL. Trends in Use of High-Dose Vitamin D Supplements Exceeding 1000 or 4000 International Units Daily, 1999-2014.JAMA · 2017 · 317(23):2448-2450doi:10.1001/jama.2017.4392 · PMID 28632857

     A research letter on the rise in high-dose supplement use in US adults; the PubMed record carries no abstract, so no figures are quoted from it.

117. 117
     Rosen CJ, Abrams SA, Aloia JF, et al. IOM committee members respond to Endocrine Society vitamin D guideline.J Clin Endocrinol Metab · 2012 · 97(4):1146-52doi:10.1210/jc.2011-2218 · PMID 22442278

     Institute of Medicine committee members dispute the 2011 Endocrine Society guideline's target levels, definition of deficiency and definition of the population at risk.

118. 118
     Harvey NC, Ward KA, Agnusdei D, et al. Optimisation of vitamin D status in global populations.Osteoporos Int · 2024 · 35(8):1313-1322doi:10.1007/s00198-024-07127-z · PMID 38836946

     The International Osteoporosis Foundation's position paper separating treatment of overt deficiency from population supplementation.

119. 119
     Gillman MW, Fu LY. Vitamin D Guidelines in Primary Care.JAMA · 2025 · 333(20):1779-1780doi:10.1001/jama.2025.5544 · PMID 40266826

     A 2025 commentary on applying the 2024 guideline in primary care; the PubMed record carries no abstract and it is cited only as a marker of the debate.

120. 120
     Avenell A, Grey A, Macleod AD, Bolland MJ. Healthcare expenditure on vitamin D supplementation.BMJ · 2026 · 394:e100477doi:10.1136/bmj-2026-100477 · PMID 42556872

     A 2026 letter on what health systems spend on vitamin D; no abstract on the record, cited only as a marker of the debate.

121. 121
     Heaney RP, Davies KM, Chen TC, Holick MF, Barger-Lux MJ. Human serum 25-hydroxycholecalciferol response to extended oral dosing with cholecalciferol.Am J Clin Nutr · 2003 · 77(1):204-10doi:10.1093/ajcn/77.1.204 · PMID 12499343

     67 men over a winter: about 0.70 nmol/L of 25-hydroxyvitamin D per microgram of daily intake, and a total requirement of roughly 3800 IU a day met mostly from summer skin stores.

122. 122
     Dai Q, Zhu X, Manson JE, et al. Magnesium status and supplementation influence vitamin D status and metabolism: results from a randomized trial.Am J Clin Nutr · 2018 · 108(6):1249-1258doi:10.1093/ajcn/nqy274 · PMID 30541089

     180 adults: magnesium raised 25-hydroxyvitamin D3 when baseline was near 30 ng/mL and lowered it when higher; no clinical outcome. Two authors were affiliated with a center for magnesium education and research.

123. 123
     Diederichsen ACP, Lindholt JS, Möller S, et al. Vitamin K2 and D in Patients With Aortic Valve Calcification: A Randomized Double-Blinded Clinical Trial.Circulation · 2022 · 145(18):1387-1397doi:10.1161/CIRCULATIONAHA.121.057008 · PMID 35465686

     365 Danish men: two years of menaquinone-7 with vitamin D did not slow aortic valve, aortic or coronary calcification despite lowering uncarboxylated matrix Gla protein.

124. 124
     Kuang X, Liu C, Guo X, Li K, Deng Q, Li D. The combination effect of vitamin K and vitamin D on human bone quality: a meta-analysis of randomized controlled trials.Food Funct · 2020 · 11(4):3280-3297doi:10.1039/c9fo03063h · PMID 32219282

     Eight small trials, 971 people: a pooled increase in bone mineral density and a fall in undercarboxylated osteocalcin; no fracture outcome.

125. 125
     Hu KL, Liao T, Wu Q, et al. Vitamin D supplementation before in vitro fertilisation in women with polycystic ovary syndrome: multicentre, double blind, placebo controlled, randomised clinical trial.BMJ · 2026 · 392:e087438doi:10.1136/bmj-2025-087438 · PMID 41702641

     876 women in 24 Chinese centers: 4000 IU a day doubled the blood level and did not change live births (52.0 against 50.2 percent).

126. 126
     Frederiksen OF, Jepsen JRM, Brustad N, et al. High-Dose Vitamin D3 Supplementation During Pregnancy and Test-Based Cognitive Performance at Age 10 Years: A Post Hoc Secondary Analysis of a Randomized Clinical Trial.JAMA Netw Open · 2026 · 9(5):e2611464doi:10.1001/jamanetworkopen.2026.11464 · PMID 42149595

     498 Danish ten-year-olds: small advantages in verbal and visual memory after high-dose prenatal vitamin D, in a post hoc analysis of a trial unblinded at age three.

127. 127
     Zhu H, Manson JE, Cook NR, et al. Vitamin D3 and marine ω-3 fatty acids supplementation and leukocyte telomere length: 4-year findings from the VITamin D and OmegA-3 TriaL (VITAL) randomized controlled trial.Am J Clin Nutr · 2025 · 122(1):39-47doi:10.1016/j.ajcnut.2025.05.003 · PMID 40409468

     1,031 VITAL participants: 2000 IU a day reduced leukocyte telomere attrition by 0.14 kilobase pairs over four years (P = 0.039). A marker, not a clinical outcome.

128. 128
     Bischoff-Ferrari HA, Gängler S, Wieczorek M, et al. Individual and additive effects of vitamin D, omega-3 and exercise on DNA methylation clocks of biological aging in older adults from the DO-HEALTH trial.Nat Aging · 2025 · 5(3):376-385doi:10.1038/s43587-024-00793-y · PMID 39900648

     A post hoc analysis of 777 DO-HEALTH participants: omega-3 slowed three of four clocks and the three treatments together had an additive effect on one. One author lists an affiliation with Altos Labs Limited.

129. 129
     Ahl NM, Hantunen S, Tuomainen TP, et al. Vitamin D supplementation and incidence of major depressive disorder - A randomized clinical trial.Clin Nutr · 2025 · 57:106570doi:10.1016/j.clnu.2025.106570 · PMID 41506003

     The Finnish FIND trial, 2,434 mostly replete older adults: no significant effect of 1600 or 3200 IU a day on major depression over five years; borderline for 3200 IU in extended follow-up.

130. 130
     Abbasi J. Coffee and AFib, Oral PCSK9 Drugs, an HFrEF Polypill, and Vitamin D Post-MI-Highlights From AHA 2025.JAMA · 2026 · 335(1):10-11doi:10.1001/jama.2025.22800 · PMID 41348343

     A medical-news summary of the 2025 American Heart Association meeting whose title lists a vitamin D trial after myocardial infarction; the record carries no abstract and no results, so it is cited only for the presentation at that meeting. The full TARGET-D report is cited separately.

131. 131
     Wang Y, Sha S, Gwenzi T, Schöttker B, Brenner H. Effect of vitamin D supplementation on cardiovascular outcomes: randomized trials revisited.Eur J Epidemiol · 2026doi:10.1007/s10654-026-01438-7 · PMID 42484779

     The same emulations for cardiovascular outcomes: largely null at the trials' levels, lower predicted cardiovascular mortality (0.74 to 0.79) in insufficient or deficient people. Not a randomized comparison.

132. 132
     Jorde R, Thorsby PM, Grimnes G. Low Prevalence of Exploratory Biochemical Patterns Compatible with Osteomalacia in a Large Cohort of Participants with Low Serum 25-Hydroxyvitamin D Levels.Calcif Tissue Int · 2026 · 117(1)doi:10.1007/s00223-026-01612-9 · PMID 42776204

     423 Tromsø participants with a mean level of 34 nmol/L: none had the full biochemical pattern of osteomalacia; without bone biopsies the findings cannot set a threshold.