Skin makes vitamin D when UVB radiation between roughly 290 and 315 nanometers hits 7-dehydrocholesterol in the epidermis. That reaction is real, it is the main source for most people, and a long list of ordinary things switch it off: winter, latitude, cloud, glass, melanin, age, the hour of the day. What it does not do is guarantee the health outcomes vitamin D has been sold on. Two very large randomized trials tested those directly and came back null on their headline endpoints.
The version of this article that stood here told readers to get ten minutes of daily sun on as much skin as possible, and claimed that antioxidants from superfoods would protect them from sun exposure. Neither of those should have run. The corrections are at the end.
How the reaction works
UVB photons convert 7-dehydrocholesterol in the skin to previtamin D3, which isomerizes thermally to vitamin D3. The liver hydroxylates that to 25-hydroxyvitamin D — the form measured in blood — and the kidney converts it to the active hormone.
Two features matter. The skin self-limits: prolonged exposure degrades previtamin D3 into inert photoproducts, so there is a ceiling and no way to overdose from sunlight. And the reaction depends entirely on UVB, the narrow short-wavelength part of the ultraviolet band and the part most easily filtered out.
Everything that blocks it
The classic demonstration is Webb, Kline and Holick’s 1988 study in the Journal of Clinical Endocrinology and Metabolism. Sunlight in Boston, at 42°N, produced no measurable previtamin D3 in a skin model from November through February. In Edmonton, at 52°N, the dead period ran October through March. At 34°N and 18°N — roughly Los Angeles and Puerto Rico — midwinter sunlight still worked.
Beyond latitude and season, the NIH Office of Dietary Supplements lists time of day, length of day, cloud cover and smog, melanin content and sunscreen. Melanin is an effective UVB filter, so more heavily pigmented skin needs considerably longer exposure for the same yield. Skin concentration of 7-dehydrocholesterol falls with age. And UVB does not pass through window glass, so a sunlit office or car contributes nothing.
Sunscreen is the item most often misreported, including by us. At the 2 mg/cm² thickness used for SPF testing, an SPF 20 product transmits about one-twentieth of erythemally effective UV — but in practice people apply far less, and a study of Danish holidaymakers found a mean of 0.79 mg/cm². Reviewing the field in the British Journal of Dermatology in 2019, Passeron and colleagues concluded that broad-spectrum sunscreens used for daily and recreational photoprotection do not compromise vitamin D status in healthy populations. The exception they flagged was people practicing strict photoprotection for photosensitivity disorders, who should be screened and supplemented.
The numbers
The NIH RDA is 600 IU (15 mcg) a day for ages 1 to 70 and 800 IU (20 mcg) for 71 and over, with a Tolerable Upper Intake Level of 4,000 IU for anyone aged 9 and up. On blood levels, the NIH treats serum 25-hydroxyvitamin D below 30 nmol/L (12 ng/mL) as deficient, 30 to 50 nmol/L (12 to 20 ng/mL) as inadequate for bone health, and 50 nmol/L (20 ng/mL) or above as generally sufficient. Above 125 nmol/L (50 ng/mL) is where adverse effects start being reported.
Hold those cut-points loosely. They were derived for skeletal outcomes, and much of the reported “epidemic of deficiency” came from applying a higher threshold, 30 ng/mL, that other bodies did not adopt.
What the big supplementation trials found
VITAL randomized 25,871 US adults to 2,000 IU of vitamin D3 daily or placebo and followed them for a median of 5.3 years. The primary results, in the New England Journal of Medicine in 2019, showed no reduction in invasive cancer of any type and none in major cardiovascular events. A post hoc analysis excluding the first two years found fewer cancer deaths, which the investigators themselves called hypothesis-generating rather than a result.
D-Health randomized 21,315 Australians aged 60 to 84 to 60,000 IU monthly or placebo for up to five years. In the Lancet Diabetes and Endocrinology in 2022, the hazard ratio for all-cause mortality was 1.04 (95% CI 0.93 to 1.18), cancer mortality 1.15 (0.96 to 1.39), cardiovascular mortality 0.96 (0.72 to 1.28). Nothing reached significance in either direction, and the authors advised caution about monthly bolus dosing in people already replete.
These were well-run, adequately powered trials of exactly the question the supplement market implies has been settled. They settled it, and the answer was no.
Bone and fall evidence, where the case was supposed to be strongest
The VITAL fracture study in the New England Journal of Medicine in 2022 recorded 1,991 incident fractures in 1,551 participants over a median 5.3 years and found no reduction in total, non-vertebral or hip fractures with 2,000 IU daily. D-Health’s fracture analysis in Lancet Diabetes and Endocrinology in 2023 covered 20,326 participants and reported hazard ratios of 0.94 (95% CI 0.84 to 1.06) for total fracture and 1.11 (0.86 to 1.45) for hip fracture.
The US Preventive Services Task Force arrived in the same place: a Grade D recommendation against daily supplementation with 400 IU or less of vitamin D plus 1,000 mg or less of calcium for primary fracture prevention in community-dwelling postmenopausal women, “insufficient evidence” for higher doses and for men, and a separate recommendation against vitamin D supplementation to prevent falls in community-dwelling adults aged 65 and over.
None of that applies to people who are actually deficient, to nursing-home residents, or to osteoporosis treatment, where vitamin D is given alongside other therapy. It applies to the healthy, replete adult buying a bottle because it seemed sensible.
Where this evidence is weak
The null trials share one real limitation: they mostly enrolled people who were not deficient at baseline. Mean 25-hydroxyvitamin D in VITAL was around 30 ng/mL. Topping up someone already sufficient is a different intervention from correcting a genuine shortfall, and these trials tested the first, not the second.
D-Health used monthly bolus dosing, which some researchers argue is physiologically different from daily dosing and has performed worse in several trials. The 2024 Endocrine Society guideline prefers daily lower-dose regimens over intermittent high-dose ones for adults over 50 who need treatment.
And the observational literature linking low vitamin D to almost everything investigated remains large and consistent. The most likely explanation is reverse causation and confounding — ill, sedentary, indoor, heavier people have lower 25-hydroxyvitamin D for reasons unrelated to the vitamin causing anything — but “most likely” is not “certainly.”
Who should get tested
Fewer people than currently do. The 2024 Endocrine Society guideline suggested against routine 25-hydroxyvitamin D testing in healthy adults, including people with dark skin or obesity, because outcome-specific benefits tied to particular levels have not been identified. It also advised against healthy adults under 75 taking doses beyond the reference intakes.
Testing makes sense when there is a reason: malabsorption such as celiac disease, Crohn’s or bariatric surgery; chronic kidney or liver disease; medications that accelerate vitamin D metabolism; osteoporosis under treatment; strict photoprotection. That is a clinical decision, not a shopping decision.
On toxicity, the NIH puts the serum level associated with toxicity above about 375 nmol/L (150 ng/mL) and notes that signs are unlikely below daily intakes of 250 mcg (10,000 IU), though lower intakes may cause problems over long periods. It presents as hypercalcemia — nausea, vomiting, weakness, excessive thirst and urination, kidney stones — and in severe cases renal failure and soft-tissue calcification. Cases almost always trace to high-dose supplements or dosing errors, never to sunlight. Dosing and product quality more generally are covered in the piece on supplement safety.
The honest tension with skin cancer risk
UV radiation is classified by IARC as a Group 1 human carcinogen. In a 2025 IARC analysis, an estimated 83 percent of the 331,700 cutaneous melanoma cases diagnosed worldwide in 2022 were attributed to ultraviolet radiation, rising above 95 percent in North America, northern Europe and Australasia. The same wavelengths that make vitamin D also damage DNA, and there is no version of sun exposure that gives one without the other.
The American Academy of Dermatology resolves the tradeoff by declining it: it does not recommend getting vitamin D from sun exposure or indoor tanning, and points to food and fortified products instead. That position is defensible precisely because the oral route is cheap, reliably raises blood levels and carries no carcinogenic exposure.
A reasonable person can land somewhere slightly different — incidental exposure during ordinary outdoor activity, without seeking burns — and still be acting on the evidence. What nobody can honestly do is present deliberate unprotected midday sun as a health practice with no cost attached.
What the previous version of this post claimed
Four claims stood here and were wrong. It said sunscreen makes vitamin D synthesis “near impossible.” The review evidence above says otherwise for normal use. It said food and supplements provide “very minimal amounts” compared with sun exposure, which is not true: oral vitamin D reliably raises serum 25-hydroxyvitamin D, which is why every trial above could separate its arms.
It advised ten minutes of daily sun on as much skin as possible for everyone, children included. That prescription ignores skin type, latitude, season and cumulative UV dose, and we should not have issued it as a blanket rule.
The worst claim was that antioxidants from superfoods provide protection against sun exposure “passed on to humans who consume them.” There is no adequate evidence that eating antioxidant-rich foods protects skin from ultraviolet damage, and presenting food as a substitute for shade, clothing or sunscreen is the kind of thing that gets people hurt. We retract it without qualification. When nutrient source genuinely matters is covered in natural versus synthetic nutrients; diet and normal immune function are handled separately, as is vitamin C, the other nutrient most often oversold on antioxidant grounds.
Common questions
How much sun do I need for vitamin D?
There is no single answer. The variables that drive it — latitude, season, hour, skin pigmentation, age, how much skin is uncovered — differ by an order of magnitude between people, which is why dermatology bodies point to diet and supplements instead of issuing an exposure prescription.
Can I make vitamin D through a window?
No. UVB does not penetrate window glass, so sunlight indoors through a window produces none. UVA passes through, which is why sun through glass can still contribute to skin aging without contributing any vitamin D.
Should I take 5,000 IU a day?
That exceeds the 4,000 IU upper limit for adults and sits above what the trials above tested. The 2024 Endocrine Society guideline advised healthy adults under 75 against doses beyond the reference intakes. Higher doses belong to a clinician correcting a documented deficiency.
Does vitamin D matter for immune function?
Vitamin D receptors are present on immune cells, so the biological rationale exists, but the large trials above were not designed around respiratory outcomes and the trial literature that does exist is mixed. Adequacy is the defensible goal; anything past that is not established.
The bottom line
Sunlight is how most people make most of their vitamin D, and adequacy is a sensible thing to want. What the largest trials did not show is that supplementing an already-sufficient adult changes anything measurable. And nothing supports treating deliberate sun exposure as free.
Greens Plus makes whole-food blends that support a diet already built on plants, and our labels state exactly what each serving contains. The range is at greensplus.com.
About the Author
