Vitamin D

One of the cheapest tests in the supplement aisle and the most useful: a meaningful share of adults in higher latitudes runs below the threshold where bone, immune, and possibly cardiovascular function start to wobble, and a low-cost lab plus a cheap pill closes that gap. But the trial evidence for routinely taking vitamin D when you're already replete is much thinner than the marketing implies.

For decades, observational studies linked low 25-hydroxyvitamin D to almost everything — heart disease, cancer, infections, depression, autoimmune conditions. The randomized trials that followed have been more measured: they confirm the importance of correcting genuine deficiency, but do not support routine supplementation in already-replete adults for primary cardiovascular or cancer prevention. A few weaker leads remain — slightly slower telomere shortening, fewer cancer deaths in trials that gave vitamin D daily, and an autoimmune benefit that faded once people stopped — but none is established.

The headline evidence

Evidence: Strong for correcting deficiency; Weak for outcomes in the already-replete. The large randomized trials are null on their primary endpoints; the useful signals are subgroup and long-tail findings.

The largest randomized trial of vitamin D in adults is [1]. The primary endpoints — invasive cancer and major cardiovascular events — were null.

Three analyses are worth knowing about:

  • Cancer mortality. VITAL's total cancer-death result was non-significant overall — a hazard ratio (HR) of 0.83, with a 95% confidence interval of 0.67–1.02. (Throughout this article, the bracketed range is where the true effect most plausibly lies; when it crosses 1.0, as here, the result is statistically compatible with no effect.) Leaving out the first two years, cancer deaths were about a quarter lower (HR 0.75, 95% CI 0.59–0.96) — but that latency analysis was not in the trial protocol, was not adjusted for multiple comparisons, and is best read as hypothesis-generating.[2] A later secondary analysis found advanced (spread or fatal) cancer about 17% lower (HR 0.83, 95% CI 0.69–0.99), with the clearest effect in people of normal weight and none detected in those with overweight or obesity.[3] A 2023 pooled analysis of 14 trials (about 105,000 people) found no clear overall drop in cancer deaths (relative risk, RR 0.94, 95% CI 0.86–1.02). In the 10 trials that gave vitamin D daily, cancer deaths were about 12% lower (RR 0.88, 95% CI 0.78–0.98), while the 4 trials using large intermittent doses showed no benefit (RR 1.07, 95% CI 0.91–1.24) — a promising subgroup finding, not proof. If real, it would suggest steady levels matter more than total dose.[4]
  • Autoimmune disease: a borderline signal that faded. In VITAL, people taking 2000 IU/day had about 22% fewer new autoimmune diagnoses over 5 years (HR 0.78, borderline significant).[5] But once late-confirmed cases were added, the in-trial benefit shrank to about 15% and was no longer statistically clear (HR 0.85, 95% CI 0.70–1.04). Counting two more years after people stopped taking it, there was no difference over the full 7 years of follow-up (HR 0.98, 95% CI 0.83–1.17).[6] Any protection seems to last only while you keep taking it, and may be smaller than first reported.
  • Telomeres shortened a little more slowly. In a VITAL substudy, 2000 IU/day spared about 140 base pairs of telomere over four years, a borderline result. In a separate trial (DO-HEALTH), vitamin D on its own did not change any DNA-methylation aging clock; it only added to omega-3's small effect on one clock. Details in the section below.

The negative findings round out the picture, and for a longevity readership they are the calibration that matters most:

  • All-cause mortality did not budge. VITAL found death from any cause HR 0.99 (95% CI 0.87–1.12).[7] The D-Health trial (21,315 older Australians, 60,000 IU monthly) was likewise null, with the point estimate fractionally above 1.0 (HR 1.04, 95% CI 0.93–1.18).[8] Across trials, supplementation does not lower all-cause mortality in well-nourished, already-replete adults, and most trial participants were replete (D-Health's placebo group averaged 77 nmol/L during the trial),[9] so these trials say less about people who are truly deficient. A 2026 modelling study that re-ran VITAL and D-Health as simulated trials in UK Biobank data reproduced the null results, but projected about 20–25% lower mortality in deficient people (HR 0.75–0.79).[10] That projection rests on observational data, not a trial, and the corrected genetic analysis below found no such effect.
  • The corrected genetic evidence points the other way from that projection. An influential 2021 Mendelian-randomization paper that suggested a causal mortality benefit in deficient people was retracted in 2024 after its statistical methods were found invalid, then republished with a reversed, null conclusion: across 500,962 individuals, genetically-predicted 25-hydroxyvitamin D showed no causal association with all-cause mortality, cardiovascular disease, or stroke at any level — including the deficient stratum.[11] This is the cleanest counterweight to causal claims: low vitamin D is largely a marker of poor health — chronic illness, inactivity, obesity, less time outdoors — rather than a cause of it, which is why observational associations are strong while randomized and genetic evidence are flat.
  • Fractures and falls were unaffected in people not at high risk. In VITAL, 2000 IU/day did not lower fractures (HR 0.98),[12] and a 2026 review of 69 trials (about 154,000 adults, mostly living at home and not at high risk) found vitamin D alone made no difference to fractures (RR 1.00) and little to no difference to falls; adding calcium gave at most a very small benefit.[13] The US Preventive Services Task Force advises against daily low-dose supplements (400 IU or less of vitamin D with 1,000 mg or less of calcium) for preventing fractures in postmenopausal women living at home.[14] Vitamin D matters for bone, but mainly via correcting deficiency, not via "more is better" in already-replete adults.
  • Mood has not moved in major randomized trials in mostly replete adults — including the large VITAL-DEP trial (18,353 adults, 2000 IU/day, no effect on new or recurrent depression).[15] Diabetes is a partial exception: in people who already have prediabetes, pooled data from three trials (using 4000 IU/day, 20,000 IU/week, or the vitamin D analog eldecalcitol) showed about 15% lower risk of progressing to type 2 diabetes (HR 0.85, 95% CI 0.75–0.96), roughly 3 fewer cases per 100 people over 3 years.[16] This has not been shown in people with normal blood sugar.
  • Colds and chest infections. The largest pooled analysis, 40 trials and about 62,000 people, found the odds of a respiratory infection were about 6% lower with vitamin D (odds ratio 0.94, 95% CI 0.88–1.00), a difference small enough to be compatible with no effect, and people who started out low did not benefit more.[17]

Vitamin D and biological aging

Evidence: Weak. One randomized substudy found slightly slower telomere shortening on vitamin D (a borderline result). In the one trial that tested methylation aging clocks, vitamin D by itself changed none of them; omega-3 slowed three, and vitamin D added to omega-3's effect on only one (PhenoAge). Intriguing, not established.

One recent result moved vitamin D, tentatively, toward the territory this site cares about most — the pace of aging itself — and one that is often quoted for it turns out to belong to omega-3.

Telomeres. Telomeres are the protective caps on chromosomes that shorten with each cell division; their attrition is a hallmark of aging. In the four-year VITAL Telomere sub-study (1,031 participants with leukocyte telomere length measured at baseline, year 2, and year 4), adults taking 2000 IU/day of D3 lost significantly less telomere length than those on placebo — attrition was 0.14 kilobases lower over four years (95% CI 0.007–0.27; p = 0.039), i.e. roughly 140 base pairs spared. Marine omega-3 at 1 g/day had no effect on telomere length in the same trial.[18] That fits a large observational literature: across about 185,000 people in 21 studies, higher vitamin D went with slightly longer telomeres, but the link was weak, varied a lot between studies, and was seen mainly in deficient people and women, not in men or in people already above 75 nmol/L (30 ng/mL).[19]

Epigenetic clocks. DO-HEALTH missed all six of its pre-specified primary endpoints; the DNA-methylation-clock result is a post-hoc secondary finding, not a headline. With that caveat: in 777 Swiss participants over 70 (of 2,157 in the trial), vitamin D alone did not change any of the four methylation aging clocks (PhenoAge, GrimAge, GrimAge2, DunedinPACE). Omega-3 alone slowed three of them: about 3 to 4 months less aging over 3 years on PhenoAge and GrimAge2, and about a 1% slower pace on DunedinPACE. Adding vitamin D and exercise to omega-3 helped further on only one clock (PhenoAge).[20] A large US survey found a related pattern on a blood-test-based aging score (phenotypic age, built from routine lab values, not DNA methylation): low vitamin D went with an older score, but the link faded above about 38 nmol/L in men and 63 nmol/L in women, and in men levels above about 92 nmol/L went with a very slightly older score. This is a snapshot study, but it argues against chasing the highest tolerable number.[21]

The honest caveat: telomere length is a surrogate marker, and the vitamin D result comes from a single substudy. The VITAL telomere result is a real randomized finding but only marginally statistically significant (p = 0.039), given the slow overall attrition and the limits of telomere measurement; telomere length itself tracks clinical outcomes only weakly. It is not the same as a demonstrated gain in healthy years — and it sits alongside the null mortality data above, not above it.

What "deficient" actually means

Evidence: Moderate. The deficiency cutoffs are widely agreed; the mortality plateau comes from observational data, and where to aim above the deficiency threshold is disputed — the 2024 guideline sets no target, while a longevity camp argues for higher levels.

The most widely used cutoffs draw on the US National Academies' 2011 review, which treats 50 nmol/L as enough for almost everyone, finds no consistent extra benefit above 75 nmol/L, and says levels above 125 nmol/L should raise concern about harm:[22]

25(OH)D levelStatusAction
<30 nmol/L (<12 ng/mL)DeficientCorrect urgently
30–50 nmol/L (12–20 ng/mL)InsufficientSupplement
50–125 nmol/L (20–50 ng/mL)SufficientMaintain
>125 nmol/L (>50 ng/mL)HighReduce dose; recheck
>250 nmol/L (>100 ng/mL)Very highStop; check blood calcium

Severe deficiency (<30 nmol/L) is uncommon in healthy adults eating a mixed diet at lower latitudes. Insufficiency — 30–50 nmol/L — is common at higher latitudes, especially through autumn and winter, and is the gap most supplementation actually addresses.

The "vitamin D deficiency epidemic" framing from the 2010s was somewhat overstated; large-scale severe deficiency is rare in adults who eat normally and get any sun exposure. Correcting real deficiency is well supported. Above that, the 2024 Endocrine Society panel found no clear evidence defining an optimal blood level for disease prevention,[23] and the US National Academies' review found 50 nmol/L enough for almost everyone, with no consistent extra benefit above 75 nmol/L.[24] Aiming for roughly 50–125 nmol/L keeps you out of deficiency without chasing high numbers.

Where opinions split

The mortality data anchor the low end. In observational studies, death rates fall steeply as 25-hydroxyvitamin D climbs out of deficiency and then flatten; in one cohort of about 2,900 very old Chinese adults (median age 87), the curve levelled off above roughly 65 nmol/L (26 ng/mL).[25] Most of the survival benefit is in escaping the bottom, not in reaching the top. That leaves a genuine disagreement about how high to aim:

  • The conservative position. The 2024 Endocrine Society guideline advises against routine 25-hydroxyvitamin D testing in people without a medical reason for it, and advises healthy adults under 75 against taking more than the standard 600–800 IU/day recommended dietary allowance (RDA), because trials have not shown extra benefit. It does suggest extra vitamin D, without testing, for adults 75 and over (it may lower death risk), in pregnancy, for ages 1–18, and for people with high-risk prediabetes. For people over 50 who do need vitamin D, it prefers daily doses over large intermittent ones.[26]
  • The optimization position. A longevity-oriented camp argues for targeting 100–150 nmol/L (40–60 ng/mL), pointing to the telomere, autoimmune, and cancer-progression signals that emerge above the deficiency threshold. The TARGET-D trial (630 heart-attack patients, doses adjusted to keep blood levels at about 100–200 nmol/L, 40–80 ng/mL) drew headlines in 2025 for halving repeat heart attacks. The full peer-reviewed paper shows its main outcome — death, heart attack, heart-failure admission or stroke combined — did not improve (15.7% vs 18.4%, HR 0.85, 95% CI 0.58–1.24); the lower heart-attack rate (3.8% vs 7.9%, HR 0.48) was one of several secondary results and is not proof on its own, and it applies to people who have already had a heart attack.[27] That fits the wider randomized record: VITAL found no effect on major cardiovascular events (HR 0.97, 95% CI 0.85–1.12),[28] and no large trial has clearly shown a cardiovascular benefit — the monthly-dose D-Health trial saw a borderline 9% drop in major cardiovascular events that was still compatible with no effect (HR 0.91, 95% CI 0.81–1.01), with fewer heart attacks as a secondary finding (about 19% lower, HR 0.81, 95% CI 0.67–0.98).[29]

The defensible middle: get out of deficiency — that is where the evidence is strongest — and don't chase a high number. The 2024 Endocrine Society panel found no clear evidence defining an optimal target level for disease prevention.[30] The case for climbing into the 100–150 range rests on surrogate markers and subgroup findings rather than hard randomized outcomes, and the slight male upturn in blood-test aging scores above about 92 nmol/L is a reminder that higher is not reliably better.

Practical dosing

  • Test first — with a caveat. A single morning 25-hydroxyvitamin D draw in autumn or winter (the seasonal trough) gives the most useful read. Note that the Endocrine Society advises against routine testing,[31] and the US Preventive Services Task Force says there is not enough evidence to judge whether screening people without symptoms helps,[32] since correcting an asymptomatic low level hasn't changed hard outcomes in trials. Testing earns its keep mainly when you want to dose deliberately rather than guess — not as a population screen.
  • Already in range (50 nmol/L or more): trials have not shown that taking more than the standard 600–800 IU/day prevents disease in healthy adults under 75, and the 2024 guideline advises against it, though it does suggest extra vitamin D from age 75.[33] People with overweight or obesity get a noticeably smaller rise from the same dose,[34] so if they are below range they may need longer on the correction doses below before retesting.
  • Insufficiency (30–50 nmol/L): 2000 IU/day for 12 weeks, then retest.
  • Deficiency (<30 nmol/L): 4000 IU/day for 8–12 weeks, retest, then drop to the 600–800 IU/day standard intake (more from age 75).
  • Take it with a meal if that's easy. In one trial, taking it with a meal that contained fat raised peak absorption by about a third compared with a fat-free meal.[35] In another, blood levels one and three months later were the same whether the dose was taken with no meal, a low-fat meal or a high-fat meal.[36]
  • Re-test annually at the end of winter once you have settled on a dose.
  • Magnesium is a minor note. The enzymes that convert vitamin D into its active hormone are magnesium-dependent, so in principle low magnesium could blunt the response — but the clinical evidence that magnesium repletion changes 25(OH)D response in non-deficient people is thin. Worth a glance only if your level isn't rising on a sensible dose. See Magnesium.

Forms and timing

  • D3 (cholecalciferol) is the right form for almost everyone. Beyond raising blood levels more reliably than D2 (ergocalciferol), D2 appears to accelerate the clearance of the body's own D3 — a meta-analysis of D2 trials found circulating 25-hydroxyvitamin D3 dropping even as total vitamin D rose.[37] Choose D3.
  • Daily dosing beats weekly or monthly megadosing — the body handles steady-state vitamin D more cleanly than large boluses, and some intermittent-dose trials hint at harm. An annual 500,000 IU dose raised falls by 15% and fractures by 26% in women aged 70 and over at high fracture risk; in a post hoc analysis, the excess in falls was largest in the three months after each dose (31% higher, versus 13% higher afterwards).[38] The monthly-bolus D-Health trial saw about 15% more cancer deaths (HR 1.15, 95% CI 0.96–1.39, compatible with no effect; HR 1.24, 95% CI 1.01–1.54 after excluding the first two years, an exploratory analysis),[39] though the same monthly dose did not raise fractures (HR 0.94, 95% CI 0.84–1.06)[40] or new cancers (HR 1.02, 95% CI 0.95–1.10).[41] Steady daily dosing is the safe default; large intermittent boluses are at best useless and at worst harmful.
  • A drop or softgel is fine. The vehicle (olive oil, medium-chain triglyceride oil) doesn't matter much.
  • Calcifediol (25-OH-D3) raises serum 25-hydroxyvitamin D faster than ordinary D3 and is worth knowing about for malabsorption or when rapid correction is needed; for routine use, plain D3 is still the default.

Sun is not a substitute in winter

Beyond about 40° latitude, north or south, winter sunlight is too weak for the skin to make vitamin D for part of the year — roughly November to February at 42°, stretching to October to March at 52°.[42] [43] The half-life of 25-hydroxyvitamin D is several weeks; a summer of good sun cannot cover a high-latitude winter. Use oral D3 across the dark months. See Sun exposure for the broader picture, including a 20-year Swedish study of about 30,000 women in which those who avoided the sun had roughly twice the death rate of the most sun-exposed women — an observational link that may partly reflect healthier, more active people spending more time outdoors.[44]

Pair with K2 if you're targeting bone

Evidence: Weak (mechanistically plausible, no hard-outcome trial).

Vitamin D raises calcium absorption from the gut. Vitamin K2 (menaquinone-7, MK-7) activates the proteins that direct that calcium into bone matrix rather than soft tissue and arterial walls. The combination has more biological coherence than either alone, but the trial record is thin. In people with heavy coronary calcification, a high-dose combination (720 µg MK-7 plus 1000 IU D3 daily) slowed the build-up of coronary artery calcium over two years by about a fifth compared with placebo.[45] An earlier trial of the same regimen, in older men with aortic valve calcification, found no effect on valve or coronary calcification.[46] Both trials used 720 µg/day, far above the 90–180 µg suggested below. No large trial has yet shown fewer fractures, heart attacks or deaths, and both trials compared D3+K2 with placebo rather than with D3 alone, so treat it as plausible rather than proven.

If you want it, 90–180 µg/day of MK-7 alongside D3 is reasonable. Coordinate with a prescriber if you're on warfarin — vitamin K antagonises it. See Vitamin K2 for the full evidence picture and Bone density for the loading side.

Cautions

  • 4000 IU/day is the official adult upper limit — set as a cautious ceiling below the 10,000 IU/day level that has not been linked to high blood calcium, not a toxic dose.[47] In a 3-year trial in healthy 55–70-year-olds, mildly high blood calcium showed up in 3% of people on 4000 IU/day and 9% on 10,000 IU/day (none on 400 IU), all of it temporary, while high urine calcium was common across groups (17–31%).[48] Both higher doses also left forearm bone density slightly lower than 400 IU/day, not higher (shin bone only at 10,000 IU), with no difference in bone strength.[49] If you take more than 4000 IU/day for months, check blood calcium and 25(OH)D.
  • Granuloma-forming diseases (sarcoidosis, some lymphomas, tuberculosis) can dysregulate vitamin D metabolism — work with a clinician.
  • High-dose calcium supplementation has its own risk profile; dietary calcium does not show the same cardiovascular signal as high-dose calcium supplements. See Supplements to avoid.
  • Clinically meaningful vitamin D toxicity is rare from supplementation alone but real with cumulative megadosing over years.

Further reading

  • Manson JE et al. Vitamin D Supplements and Prevention of Cancer and Cardiovascular Disease (VITAL). NEJM 2019.[50]
  • Chandler PD et al. Effect of Vitamin D3 Supplements on Development of Advanced Cancer: A Secondary Analysis of the VITAL Randomized Clinical Trial. JAMA Network Open 2020.[51]
  • Hahn J et al. Vitamin D and Marine Omega-3 Fatty Acid Supplementation and Incident Autoimmune Disease. BMJ 2022.[52]
  • LeBoff MS et al. Supplemental Vitamin D and Incident Fractures in Midlife and Older Adults. NEJM 2022.[53]
  • Bischoff-Ferrari HA et al. Individual and additive effects of vitamin D, omega-3 and exercise on DNA methylation clocks of biological aging (DO-HEALTH). Nature Aging 2025.[54]
  • Zhu H 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). Am J Clin Nutr 2025.[55]
  • Kuznia S et al. Efficacy of vitamin D3 supplementation on cancer mortality: systematic review and individual patient data meta-analysis. Ageing Research Reviews 2023.[56]
  • Demay MB et al. Vitamin D for the Prevention of Disease: An Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab 2024.[57]
  • Neale RE et al. The D-Health Trial: a randomised controlled trial of the effect of vitamin D on mortality. Lancet Diabetes & Endocrinology 2022.[58]
  • Emerging Risk Factors Collaboration / EPIC-CVD / Vitamin D Studies Collaboration. Retraction and republication — Estimating dose-response relationships for vitamin D with coronary heart disease, stroke, and all-cause mortality. Lancet Diabetes & Endocrinology 2024.[59]
  • Okereke OI 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.[60]
  • Costenbader KH 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 & Rheumatology 2024.[61]
  • Pittas AG 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. Annals of Internal Medicine 2023.[62]
  • May HT et al. Targeted vitamin D supplementation and major adverse cardiovascular events after myocardial infarction: the TARGET-D trial. European Heart Journal 2026.[63]
  • Sanders KM et al. Annual High-Dose Oral Vitamin D and Falls and Fractures in Older Women. JAMA 2010.[64]

— § —