A daily 1,000–2,000 IU softgel costs under fifteen dollars a year and takes the willpower of a toothbrush. The honest payoff is modest and slow — a small drop in respiratory infections, a small bend in long-term mortality, real relief if you turn out to be deficient — not the transformation a podcast might have promised. Take it daily, with food that has some fat, and skip the megadoses and the boutique combos.
Strictly speaking, vitamin D3 isn't a vitamin. It's a hormone precursor your skin makes from cholesterol when ultraviolet light hits it — the same way a leaf makes sugar from sunlight. Cod liver oil and salmon contribute a little; everything else is fortification, supplements, or your own skin. The liver converts what you absorb or make into a storage form called 25(OH)D — that's the number a vitamin D blood test reports — and the kidney converts that into the active hormone, which then walks into the nucleus of nearly every cell in your body and tells specific genes to switch on Holick 2007.
Which is why the same molecule turns up in conversations about bones, immunity, mood, cancer, and dying-of-old-age — and why none of those conversations is simple. The hormone is everywhere; the question is how much it actually does at each address. The body's setup, evolutionarily, was: stand outside half-naked, get plenty. The modern setup — desks, latitudes, sunscreen, winter — is one your physiology was never asked to handle, and a chunk of the adult population walks around with stored levels lower than any human in the species' history had until about two centuries ago.
What it actually does — and what it doesn't
The clearest case is the bottom of the range. Adults running below 20 ng/mL — about a third of US adults at any given moment, more in winter, more if you have darker skin or spend your days indoors Forrest & Stuhldreher 2011 — get measurable wins from replacement: less of the proximal-muscle fatigue that comes with running on empty, fewer respiratory infections, in older adults a real reduction in falls. Severe deficiency causes a disease called osteomalacia — soft bones, deep ache, weak hips and shoulders. Childhood version is rickets. These were epidemic in industrial Northern cities a century ago. Fortified milk and cod liver oil ended them.
What the last decade of large trials showed — and this is the part wellness culture has not fully absorbed — is that the picture above the deficient range gets a lot more modest, fast. The biggest single trial, VITAL, dosed 25,871 healthy adults at 2,000 IU/day for five years and watched for cancer, heart attacks, strokes, and depression. None of those endpoints moved.
What did survive the big-trial era: daily D3 nudges all-cause mortality down a little. The Cochrane review of 56 trials found D3 specifically — not the prescription form, D2 — cut mortality by about 6% Bjelakovic et al. 2014; the BMJ umbrella analysis put the number closer to 11% Chowdhury et al. 2014. Cancer-specific mortality moved in the same direction. The signal hinges on daily dosing — bolus doses (50,000+ IU monthly, or worse, annual mega-doses) don't show this benefit, and one famous annual-bolus trial actually increased falls and fractures Sanders et al. 2010. The body wants this hormone trickled in, not flooded.
The infection signal is real and worth understanding. A pooled analysis of every individual participant from 25 RCTs — not just trial-level numbers, individual data — found daily or weekly D3 reduced acute respiratory infections by about 12% across the whole population, and by a much larger margin in the people who started genuinely deficient Martineau et al. 2017. The dose-response is the giveaway: the worse off you were, the more replacement gave you back.
The cleanest read on everything else — mood, cognition, blood pressure, type 2 diabetes prevention, fractures in community-dwelling adults — is null or close to it. Smaller trials in deficient or seasonally-affected subgroups sometimes find something; the big trials in unselected populations rarely do. The lesson is consistent: D3 is mostly a deficiency-replacement story, not a more-is-better one.
The cost of not bothering
Skip this and the most likely outcome is: nothing visible, ever. That has to be said honestly. D3 isn't the entry where the missing version of you walks around with a face people stop recognising. The stakes are quieter than that, and they are real anyway.
The version of you who never sets the floor lives at a 25(OH)D somewhere in the teens through every winter, and the trade is paid in small, replicable ways. One more bad respiratory infection a year than the version who supplemented Martineau et al. 2017. A few more days of the deep proximal-muscle fatigue — the legs-feel-heavy-getting-up-from-a-chair kind — through February, dismissed as the season. Across decades, an actuarial drift in the wrong direction on all-cause mortality, small enough that nobody could point at a single year and call it the one you bought back, large enough to show up consistently in meta-analyses of tens of thousands of people Bjelakovic et al. 2014. In your seventies, when the falls start, the version who supplemented from middle age has slightly better odds of staying off the floor — that signal weakened in recent trials but never disappeared in the older, deficient populations the 700–1,000 IU trials were built on Bischoff-Ferrari et al. 2009.
None of this rises to fear-mongering. The honest forecast is the small one: a quietly elevated background risk through every winter you do not bother, paid in colds and fatigue and an old-age curve that bends slightly the wrong way.
How to take it
The whole protocol fits in a sentence: one D3 softgel with a meal that has some fat, every day, at a dose somewhere between 1,000 and 2,000 IU/day. Anything fancier is the supplement industry talking.
That is the whole thing. There is no protocol where you take 10,000 IU because of a paper you half-read, no need to retest every six months, no boutique combination product worth four times the price.
Who actually needs more
Vitamin D status splits the adult population sharper than almost any other nutrient. A few honest groupings, in descending order of "you should probably just take it":
- Take it without testing first. Dark-skinned adults living above the latitude of Atlanta (roughly anywhere in Northern Europe, the upper US, Canada). Adults with a BMI over 30 — the molecule partitions into fat tissue, so you need two to three times the dose to reach the same blood level. Adults over 70, whose skin makes about half the vitamin D it did at 20. Anyone who is institutionalised, homebound, or veiled. Pregnant or breastfeeding women.
- Test first, then decide. Indoor knowledge workers in temperate climates. Vegans without fortified-food intake. Adults aged 50–70 with no obvious sun exposure. The single test is cheap; the answer is usually "yes, take some," but knowing the starting number is worth $40.
- Probably skip it. Outdoor workers in sunny climates. People who get an honest 15–30 minutes of midday sun on bare arms most days in summer and live somewhere the sun keeps working in winter. Adults who eat fatty fish more than twice a week. Your skin and your dinner plate are already doing the job.
The third group is smaller in the modern world than it sounds. If you commute, work indoors, and live anywhere north of Atlanta or south of Sydney, default to the first or second group.
When not to take it
For a supplement with a daily dose of 1,000–2,000 IU, the safety profile is among the cleanest in this catalogue. The conditions where caution is warranted are narrow and most readers will not have them — but they are real.
What the wellness world gets wrong
Vitamin D sits at the intersection of cheap supplement, dense literature, and a decade of confident podcasters; the misinformation has had a lot of room to grow. Four worth flagging:
"Most people are deficient." Depends entirely on where you draw the line. The Institute of Medicine calls 20 ng/mL sufficient; the Endocrine Society calls 30 ng/mL sufficient. Move the line ten units and you reclassify roughly half the country Forrest & Stuhldreher 2011, IOM 2011, Holick et al. 2011. The genuine clinical deficiency rate (below 20 ng/mL) is real but more like a third of adults, concentrated in identifiable groups. Calling 70% of the population "insufficient" is mostly a threshold artefact.
"Aim for 70, 80, 100 ng/mL." Above 50 ng/mL there is essentially no trial evidence of additional benefit, and for several outcomes the dose-response curve flattens or bends back down. The high targets came from observational data and from extrapolation, not from trials showing the high group does better than the medium group.
"D3 needs K2 or it'll calcify your arteries." A favourite of supplement marketing, especially at the price-doubled D3+K2 combination. The mechanistic story (a vitamin-K-dependent protein directs calcium to bone rather than vessels) is genuinely plausible. The human trial evidence that K2 changes anything when you take ordinary doses of D3 is essentially absent. Not clearly wrong, not clearly necessary. Plain D3 is fine.
"Sunscreen is making everyone deficient." In the lab, SPF 30 blocks 95% of UVB and stops the synthesis pathway cold. In real life, people who use sunscreen regularly don't have meaningfully lower vitamin D than people who don't. The reason is the obvious one: nobody applies sunscreen the way the lab tests do, and the incidental sun on your hands and face walking to lunch turns out to be plenty.
What to buy
D3 is one of the cheapest things in the supplement aisle. A year's supply of 1,000- or 2,000-IU softgels runs five to fifteen dollars from the bulk-supplement section of any pharmacy or online retailer. Liquid drops for children or anyone who hates pills cost a little more. Stick to plain D3 — combination products that bury an underdose alongside calcium, K2, magnesium, and a price markup are not the move; if you want those other things, buy them separately and at honest doses.
The 25(OH)D blood test runs $30–60 cash through direct-to-consumer labs in the US and is reimbursed by most insurance when ordered for documented risk factors. The LC-MS/MS method is the gold standard at low concentrations; standard immunoassay is usually fine. One test, once, to know where you stand on your usual dose — then stop thinking about it unless something changes (you move, you lose a lot of weight, you start a new medication that affects it).
Sunlight and food, for what they're worth
Vitamin D came from sunlight long before it came from softgels, and in the right conditions sunlight is still by far the most efficient delivery system. Ten to thirty minutes of midday summer sun on bare arms and face will produce several thousand IU of D3 in fair-skinned adults — vastly more than any reasonable supplement dose, and bundled with the unrelated mood and circadian benefits of bright light per se. The catch is that this works only when the UV index is high enough, which is roughly April through September above 35° latitude (most of the US north of Atlanta, all of the UK, most of continental Europe). October through March in those places, your skin makes essentially zero D3 no matter how clear the sky.
Food can carry meaningful amounts if you eat the right things deliberately. Wild salmon and other fatty fish run 600–1,000 IU per 100 g serving; canned mackerel and sardines, 200–500 IU; cod liver oil, around 1,000 IU per teaspoon; egg yolks about 40 IU each; UV-exposed mushrooms carry D2 (not D3). Fortified milk and orange juice add ~100 IU per cup. Reaching 1,000 IU/day from food alone is achievable but requires intent.
For most readers the honest answer is "all of the above": get sun when the season offers it, eat the fish, take the pill in winter, stop worrying about it.
What you actually get
The honest forecast depends mostly on where you started. Three rough cases.
If you started genuinely deficient (under 15 ng/mL): six to eight weeks in, the heavy-legs fatigue lifts. The bone ache, if you had it, eases. Through the next winter, the colds that used to flatten you are shorter and milder. Your blood test comes back somewhere readable. The change is felt — modest, real — because there was a deficit, and you closed it.
If you started mildly insufficient (15–25 ng/mL, the modal case for indoor adults at northern latitudes): nothing dramatic happens day to day. Your winter goes a little better — one fewer cough that turns into a chest thing, a flatter energy floor through February — but it is the kind of change you notice only if you compare to last year. Across decades, the actuarial weather bends in your direction: a small reduction in all-cause mortality, a small reduction in cancer mortality, a small reduction in fall risk when you hit your seventies Bjelakovic et al. 2014, Bischoff-Ferrari et al. 2009. Nobody can point at the single year you bought back; it shows up across the population, not on any one chart.
If you started already replete (above 30 ng/mL year-round, somehow): you will not feel anything change. The big trial in this population — VITAL — found basically nothing on cancer, heart disease, depression, or fractures Manson et al. 2019, LeBoff et al. 2022, Okereke et al. 2020. The pill costs you nothing to keep taking, but the payoff is essentially insurance against ever sliding into the deficient group, not a transformation.
The reason most adults should still take it is that the cost of being wrong about which group you are in is asymmetric. A daily pill the price of nothing, on the off chance you were the indoor winter person who would have spent the next decade at 18 ng/mL. The pill bought you out of that. Quietly. In the background. The way most useful health decisions do their work.
Adjacent topics worth knowing about, each their own entry: morning sunlight (the natural pathway plus its independent circadian and mood payoff), magnesium (a cofactor for the hydroxylation steps; severe magnesium deficiency can blunt the response to D3), calcium intake (the classical bone partner; dose carefully and prefer food over supplements), and the 25(OH)D blood test itself as a standalone screening question.
Substance and claimed effects
Vitamin D3 (cholecalciferol) is a fat-soluble secosteroid that the body synthesises in the skin from 7-dehydrocholesterol under UVB radiation (290–315 nm), or absorbs from oily fish, egg yolks, fortified dairy, and supplements. It is biologically inert until the liver hydroxylates it to 25-hydroxyvitamin D (25(OH)D, calcidiol) — the circulating storage form, half-life ~3 weeks, the metric measured on serum vitamin D tests — and the kidney (and many extrarenal tissues) hydroxylates it again to 1,25-dihydroxyvitamin D (1,25(OH)2D, calcitriol), the active hormone. Calcitriol binds the vitamin D receptor (VDR), a ligand-activated nuclear transcription factor expressed in nearly every nucleated cell type, and modulates transcription of hundreds of genes Holick 2007.
The entry covers supplementation of D3 in adults — daily oral doses typically in the 1,000–2,000 IU range, weekly equivalent ~10,000–14,000 IU, intermittent loading doses (50,000 IU/week × 8 weeks) for documented deficiency — at the standard target of pushing serum 25(OH)D into the 30–50 ng/mL (75–125 nmol/L) range. Claimed effects, holistically: bone mineralisation and fracture/fall reduction in older adults; modest reduction in acute respiratory infections; small but replicated reduction in all-cause and cancer mortality with daily (not bolus) dosing; muscle strength preservation; mood, particularly in the seasonally-affected and deficient; immune modulation broadly. Null or contested for: incident cardiovascular events, incident cancer, type 2 diabetes prevention in normoglycemic populations, fracture prevention in generally healthy community-dwelling adults, COVID-19 outcomes.
Evidence by addressing question
mechanism
Cholecalciferol is converted to 25(OH)D in hepatocytes by CYP2R1 (the principal 25-hydroxylase) and CYP27A1. 25(OH)D is the assay target on standard "vitamin D" blood tests because it is the most abundant circulating metabolite and reflects integrated intake plus cutaneous synthesis over the prior 2–3 months. Renal 1α-hydroxylase (CYP27B1) — tightly regulated by parathyroid hormone, calcium, phosphate, and FGF-23 — produces 1,25(OH)2D, which has a half-life of only 4–8 hours. Calcitriol's classical action is to upregulate intestinal calcium and phosphate absorption (via TRPV6 and other transporters), maintain serum calcium homeostasis in concert with PTH, and direct osteoblast and osteoclast activity. The VDR is also expressed extrarenally — in T and B lymphocytes, monocytes, dendritic cells, vascular smooth muscle, pancreatic β-cells, skeletal myocytes, keratinocytes, prostate, breast, and colonic epithelium — and many of these tissues express their own CYP27B1, enabling local autocrine/paracrine generation of calcitriol independent of renal supply. This non-skeletal mechanism is the conceptual basis for the entire wave of trials testing vitamin D for cancer, cardiovascular disease, infection, autoimmunity, and mood Holick 2007, Holick et al. 2011.
Pharmacokinetically, oral D3 raises serum 25(OH)D roughly 1 ng/mL per 100 IU/day at steady state in lean adults, but the dose-response is nonlinear: gains flatten above ~40 ng/mL, and obese adults (BMI >30) achieve roughly half the rise per IU because D3 is sequestered in adipose. Cutaneous synthesis can produce 10,000–20,000 IU of D3 in a single full-body, fair-skinned, midday-summer, midlatitude exposure of 10–30 minutes — but is effectively zero from October to March above ~35° latitude, in heavily-pigmented skin, in the elderly (whose 7-dehydrocholesterol concentration in skin falls by ~50% from age 20 to age 70), through clothing or window glass, or under sunscreen Holick 2007.
evidence
Mortality. The two most-cited meta-analyses converge on a small but consistent all-cause mortality reduction. The Cochrane review of 56 randomised trials (n≈95,000) found vitamin D3 specifically — but not D2 — reduced all-cause mortality (RR 0.94, 95% CI 0.91–0.98) and cancer mortality (RR 0.88, 0.78–0.98); D2 and active analogues showed no benefit Bjelakovic et al. 2014. The BMJ umbrella meta-analysis found D3 cut all-cause mortality by ~11% (RR 0.89, 0.80–0.99) Chowdhury et al. 2014. The signal is modest, hinges on daily (not bolus) dosing, and is most visible in older institutionalised populations; the absolute risk reduction is small.
VITAL trial (n=25,871). The single largest test of D3 for hard endpoints. Healthy U.S. adults (≥50 men, ≥55 women, ~20% Black) received 2,000 IU/day cholecalciferol vs placebo, mean follow-up 5.3 years. Primary endpoints — invasive cancer and major cardiovascular events — were null (cancer HR 0.96 [0.88–1.06], CVD HR 0.97 [0.85–1.12]) Manson et al. 2019. Pre-specified secondary endpoint of cancer mortality showed a non-significant 17% reduction; in a sensitivity analysis excluding the first two years of follow-up (to address latency), cancer mortality fell 25%. The VITAL ancillary on incident fractures (n=25,871) was also null (HR 0.98 for total fractures) LeBoff et al. 2022, as was the VITAL-Depression ancillary (n=18,353; HR 0.97 for incident depression or clinically relevant depressive symptoms) Okereke et al. 2020.
D2d trial (n=2,423). 4,000 IU/day cholecalciferol vs placebo in adults with prediabetes. Progression to diabetes: HR 0.88 (0.75–1.04), p=0.12 — a directionally consistent but statistically null result Pittas et al. 2019. Subgroup analyses of those with baseline 25(OH)D <12 ng/mL suggested benefit, but the trial enrolled few such participants.
DO-HEALTH (n=2,157). 2,000 IU/day cholecalciferol (with or without omega-3 and a strength-training program, factorial design) in community-dwelling adults ≥70 in Europe. Null for non-vertebral fractures, cognitive decline, systolic blood pressure, and incident infection Bischoff-Ferrari et al. 2020.
Falls. Older randomised data, including the influential 2009 BMJ meta-analysis of 8 RCTs (n=2,426 older adults), found vitamin D 700–1,000 IU/day reduced falls by ~19% (RR 0.81, 0.71–0.92), with no benefit below 700 IU/day Bischoff-Ferrari et al. 2009. The picture darkened in the 2010s: the D-Health trial (n=21,315 Australians ≥60, 60,000 IU/month for up to 5 years) found no reduction in falls overall Waterhouse et al. 2023, and the Bolland meta-analysis of 81 RCTs concluded that the totality of evidence does not support vitamin D supplementation for fracture or fall prevention in community-dwelling adults Bolland et al. 2018. The Sanders trial of annual 500,000 IU oral D3 in older women paradoxically increased falls and fractures (RR 1.15) — strong evidence against bolus dosing Sanders et al. 2010.
Bone density and fracture. Combined calcium + vitamin D in institutionalised older adults remains the cleanest positive case: the Women's Health Initiative (n=36,282) found 1,000 mg calcium + 400 IU D3 reduced hip fracture by 12% (HR 0.88) in intention-to-treat, and by 29% in adherent participants Wactawski-Wende et al. 2006. In ambulatory, generally healthy adults the benefit largely disappears.
Respiratory infections. The Martineau IPD meta-analysis (25 RCTs, n=10,933) found vitamin D supplementation modestly reduced acute respiratory tract infection (adjusted OR 0.88, 0.81–0.96), driven almost entirely by daily or weekly dosing — not boluses — and concentrated in those with baseline 25(OH)D <25 nmol/L (10 ng/mL), where the OR was 0.30 Martineau et al. 2017. An updated 2021 aggregate-data meta-analysis (43 trials, n=48,488) attenuated the effect (OR 0.92, 0.86–0.99) but preserved the pattern Jolliffe et al. 2021.
COVID-19. Observational signal was strong (low 25(OH)D associated with worse outcomes), but RCTs of D3 supplementation showed no benefit on hospitalisation or severity Jorgensen et al. 2024. Reverse causation (acute illness lowers 25(OH)D) likely explains much of the observational signal.
Mood and depression. VITAL-Depression definitively rejected the hypothesis that D3 prevents incident depression in generally-replete adults Okereke et al. 2020. Smaller trials in deficient or seasonally-affected populations show modest improvements; the overall evidence is mixed.
protocol
Dose ranges and targets. The Institute of Medicine's 2011 DRI sets the RDA at 600 IU/day (ages 1–70) and 800 IU/day (71+), with a target serum 25(OH)D of ≥20 ng/mL (50 nmol/L) for skeletal health in 97.5% of the population, and a tolerable upper intake level of 4,000 IU/day for adults IOM 2011. The Endocrine Society's 2011 Clinical Practice Guideline takes a more permissive line: target ≥30 ng/mL (75 nmol/L) — calling 20–30 "insufficiency" — and recommends 1,500–2,000 IU/day in adults to reach it, with up to 10,000 IU/day considered safe long-term Holick et al. 2011. The empirical pharmacokinetics: oral D3 raises serum 25(OH)D by roughly 0.7–1.0 ng/mL per 100 IU/day at steady state in lean adults; obese adults require ~2–3× that dose. Steady state is reached in ~3 months.
Daily vs weekly vs bolus. Daily and weekly oral D3 produce equivalent and stable 25(OH)D in most adults; monthly or larger bolus dosing produces sawtooth pharmacokinetics and, in the case of annual mega-doses, has caused harm Sanders et al. 2010. The non-skeletal benefits (mortality, infection) appear specifically in the daily/weekly arm of meta-analyses, not the bolus arm Martineau et al. 2017, Bjelakovic et al. 2014. Daily is the safer default.
Loading for documented deficiency. The standard protocol for serum 25(OH)D <20 ng/mL is 50,000 IU/week × 6–8 weeks (300,000–400,000 IU total), then maintenance 1,500–2,000 IU/day, with retest at 3 months Holick et al. 2011.
Form. D3 (cholecalciferol, animal/lichen-derived) raises serum 25(OH)D more efficiently and durably than D2 (ergocalciferol, prescription-only in the US). The mortality benefit in Cochrane appears with D3 only Bjelakovic et al. 2014. Buy D3.
Absorption. D3 is fat-soluble; take with a meal containing fat to improve absorption (~30–50% improvement over fasted state). Magnesium is a cofactor for both hydroxylation steps; severe magnesium deficiency can blunt response.
contraindications
Hypercalcemia and granulomatous disease. Sarcoidosis, tuberculosis, and other granulomatous conditions express extrarenal 1α-hydroxylase activity in macrophages that is not feedback-regulated by calcium — calcitriol production runs unchecked, and standard D3 supplementation can precipitate hypercalcemia. Hyperparathyroidism (primary), Williams syndrome, and any history of vitamin D toxicity are contraindications. Patients on high-dose thiazide diuretics have impaired urinary calcium excretion and a higher risk of hypercalcemia from D3.
Toxicity. Genuine D3 toxicity (hypercalcemia, hypercalciuria, nephrocalcinosis, acute kidney injury) requires sustained intake above ~10,000 IU/day for months or a single mega-dose. The toxicity floor is conservatively placed at 25(OH)D >150 ng/mL (375 nmol/L). Cases in the literature are rare and almost universally involve dispensing errors or extreme self-supplementation (100,000+ IU/day).
Kidney stones. Calcium + D3 supplementation increased kidney stones in the WHI by 17% (HR 1.17) Wactawski-Wende et al. 2006. D3 alone at standard doses does not appear to raise stone risk.
misconceptions
"Most people are deficient." In NHANES 2001–2006, ~32% of US adults had 25(OH)D <20 ng/mL (deficient by IOM threshold) and ~71% had <30 ng/mL (insufficient by Endocrine Society threshold) Forrest & Stuhldreher 2011. Deficiency is concentrated in dark-skinned, obese, indoor, and elderly subgroups. Calling 71% of adults "insufficient" is an artifact of where the threshold is set — there is no RCT evidence that pushing a 25 ng/mL person to 35 ng/mL changes any clinical outcome.
"Higher is better." Above 50 ng/mL there is no RCT evidence of additional benefit, and the dose-response for several outcomes (falls in some trials, all-cause mortality in some analyses) appears U-shaped or J-shaped.
"Vitamin D prevents cancer/CVD/diabetes." The major RCTs (VITAL, D2d, DO-HEALTH) in generally healthy adults are null for incident cancer, CVD, and diabetes. The mortality signal from meta-analyses is real but modest, and may operate through preventing severe deficiency rather than through any "more is more" mechanism.
"D3 needs K2 to work / to avoid arterial calcification." Widely repeated in supplement-industry marketing. The mechanistic hypothesis (vitamin K2-dependent matrix Gla-protein directs calcium to bone rather than vessels) is plausible; the RCT evidence in humans taking standard D3 doses is essentially absent. Not clearly wrong, not clearly necessary.
"Sunscreen causes vitamin D deficiency." In laboratory studies, SPF 30 blocks ~95% of UVB, and theoretically synthesis. In real-world studies, regular sunscreen users do not have meaningfully lower 25(OH)D — likely because incidental sun exposure outside sunscreened windows (hands, face) is sufficient, and because nobody applies sunscreen the way the laboratory tests do.
audience
Stratify supplementation decisions on prior probability of deficiency:
- High-prior groups (worth supplementing without testing, default 2,000 IU/day): adults with dark skin living above 35° latitude; institutionalised or homebound older adults; obese adults (BMI >30); pregnant or lactating women; exclusively breastfed infants (universally recommended 400 IU/day in the first year); patients with malabsorption (Crohn's, celiac, gastric bypass, cystic fibrosis); patients on enzyme-inducing medications (phenytoin, phenobarbital, rifampin).
- Moderate-prior groups (test first, then decide): indoor knowledge workers in temperate climates; vegans with limited fortified-food intake; adults >65.
- Low-prior groups (probably no benefit, no need to test): outdoor workers in low-latitude/sunny climates; people with daily significant unprotected sun exposure on at least arms and face; adults with consistently fatty-fish-heavy diets.
alternatives
Sunlight. Ten to thirty minutes of midday summer exposure on arms and face produces several thousand IU of cutaneous D3 in fair-skinned adults — meaningful dose, with the added benefits of circadian entrainment and mood from light per se. Useless October–March above 35° latitude.
Diet. Fatty fish (wild salmon: ~600–1,000 IU per 100 g; canned salmon, mackerel, sardines: ~200–500 IU), cod liver oil (~400–1,200 IU/tsp), egg yolks (~40 IU each), UV-exposed mushrooms (D2, not D3), and fortified milk/orange juice (~100 IU/cup) are the available food sources. Reaching 1,000 IU/day from food alone is achievable but requires deliberate intake.
D2 (ergocalciferol). Prescription, plant-derived, less efficient at raising 25(OH)D, lacks the mortality signal. Used clinically because high-dose D2 (50,000 IU) is available by prescription; otherwise no advantage.
Active analogues (calcitriol, paricalcitol). Used in chronic kidney disease and renal osteodystrophy when the kidney's 1α-hydroxylation step fails. Not a substitute for D3 in the general population.
practicalities
D3 supplements are among the cheapest in the supplement aisle: 1,000-IU softgels run roughly $5–15 for a year's supply; 5,000-IU capsules similar. Liquid drops (useful for infants and the pill-averse) cost slightly more. Standalone D3 is preferable to many combination products that bury an inadequate dose alongside other ingredients. Serum 25(OH)D testing costs $30–60 cash through direct-to-consumer labs in the US, or is reimbursed when ordered by a clinician for documented risk factors. Both LC-MS/MS and immunoassay are accepted; LC-MS/MS is more accurate at low concentrations.
stakes
Stakes for the typical reader of this catalogue are not the textbook horror cases (rickets in children, osteomalacia in adults, severe proximal myopathy) but the diffuse, dimmer cost of a chronic low-grade deficit. The honest picture: a healthy adult living indoors in a temperate climate, never supplementing, hits winter at 15–20 ng/mL; falls more often than they otherwise would in older age; catches one more respiratory infection a year than they would at 30 ng/mL; over a lifetime, sits at a small but real elevated all-cause mortality risk. None of it is dramatic. Most of it is invisible. The reader who skips D3 forever is overwhelmingly likely to be fine — there is no "you will get cancer" projection the evidence supports. The honest stake is the small, replicable, unflashy mortality and infection cost from leaving an easy floor unset.
payoff
For the typical reader starting from "probably mildly insufficient" (25(OH)D 20–28 ng/mL), supplementing 1,000–2,000 IU/day delivers: a steady winter 25(OH)D in the 30–50 ng/mL range within 8–12 weeks; a small reduction in the probability of acute respiratory infection (NNT roughly 33 to prevent one ARI over a year in the population; NNT ~4 in those starting severely deficient) Martineau et al. 2017; modest reduction in all-cause mortality across years Bjelakovic et al. 2014; in older adults specifically, plausible reduction in fall risk that has weakened in recent trials Bolland et al. 2018. For the deficient (25(OH)D <15 ng/mL) — which is real for a non-trivial slice of the population — replacement can shift mood, energy, and proximal muscle weakness within weeks. For the already-replete, the payoff is essentially insurance: small, possibly real, definitely not transformative.
history
Vitamin D's history is the rickets story: epidemic in industrial Northern Europe and the urban tenement US by the late 19th century; identified as preventable by sunlight (Huldschinsky, 1919) and cod liver oil (Mellanby, 1919); fortification of milk (1930s) effectively eliminated the disease in the developed world. The renaissance of interest dates to the late 1990s with the recognition that the VDR is expressed in nearly every tissue and that 25(OH)D status correlates observationally with dozens of non-skeletal outcomes. The 2010s were the era of the large RCTs designed to test those observational signals; the results — largely null for hard endpoints, modestly positive for infection and mortality — reset enthusiasm to a more sober baseline.
out-of-scope
Magnesium status (a cofactor for vitamin D hydroxylation), calcium intake (the classical partner of D3 for bone), morning sunlight (the alternative natural pathway plus its own circadian payoff), and serum 25(OH)D testing as a standalone screening question all warrant separate treatment.
The credibility range
Optimist case
Vitamin D is one of the safest, cheapest interventions in the catalogue. The Cochrane review of 56 trials and the BMJ umbrella meta-analysis both find that daily D3 reduces all-cause mortality by roughly 6–11% Bjelakovic et al. 2014, Chowdhury et al. 2014; the Martineau IPD meta-analysis shows a real reduction in acute respiratory infection, especially in the deficient Martineau et al. 2017; older meta-analyses of falls show ~19% reduction in deficient older adults at therapeutic doses Bischoff-Ferrari et al. 2009. Population deficiency is widespread — roughly a third of US adults below 20 ng/mL — and disproportionately affects dark-skinned, obese, elderly, and indoor subgroups whose risk-benefit calculus is the most favourable. Severe deficiency causes unambiguous disease (osteomalacia, proximal myopathy, rickets in children). The cost of supplementing is essentially zero, the toxicity floor is far above standard doses, and the dose-response curve is flat enough that imprecision is forgiven. The right read is: a high-evidence safety profile and a modest but durable benefit; an obvious thing to take.
Skeptic case
For everything beyond severe deficiency, the largest, best-designed RCTs are null. VITAL — 25,871 generally healthy adults followed for 5+ years on 2,000 IU/day — was null for cancer, CVD, fractures, depression Manson et al. 2019, LeBoff et al. 2022, Okereke et al. 2020. D2d was null for diabetes prevention Pittas et al. 2019. DO-HEALTH was null for fractures, falls, cognition, BP Bischoff-Ferrari et al. 2020. The Bolland trial-sequential meta-analysis concluded that the evidence for fall and fracture prevention in community-dwelling adults does not survive cumulative meta-analysis Bolland et al. 2018. The USPSTF found insufficient evidence to recommend either screening or supplementing asymptomatic adults USPSTF 2021. Observational signals are dense with reverse causation (acute illness, inflammation, obesity, and lack of outdoor activity all lower 25(OH)D independently of vitamin D's biological effect). The mortality signal from meta-analyses is small, fragile to which trials are included, and may be explained by selecting for trials with high-risk populations. There is also a real industry-and-guideline-body interest in setting thresholds high — the Endocrine Society's 30 ng/mL threshold reclassifies a majority of the population as "insufficient" and drives both supplement sales and clinical testing.
Author's call
Both cases hold partial ground. The honest position: supplementation has a strong evidence base for safety and a modest evidence base for benefit in deficient or high-risk populations; for the already-replete, benefits are small or absent. The IOM 2011 threshold of 20 ng/mL is the defensible floor; the Endocrine Society's 30 ng/mL is a reasonable aim but should not be sold as a clinical necessity. A daily 1,000–2,000 IU D3 supplement is a low-cost, low-effort, modest-benefit move appropriate for most indoor adults, especially in winter and at higher latitudes. Testing 25(OH)D once is worth doing; chasing high targets >50 ng/mL is not. The article should land confidently on "take it; here's why; here's where the hype outruns the data." Evidence rates 4 (large RCT base, consistent on safety and mechanism, mixed on hard endpoints). Controversy rates 3 (active disagreement on optimal targets and routine supplementation).
Stakeholder and incentive map
- Supplement industry. Vitamin D is a multi-billion-dollar category; manufacturers benefit from high-threshold framing and from K2/D3 combo upsells.
- Endocrine Society. Tends toward the permissive end on targets (≥30 ng/mL) and dosing (up to 10,000 IU/day safe).
- Institute of Medicine / NAM. Conservative on RDA (600–800 IU/day) and target (≥20 ng/mL); explicitly states evidence beyond skeletal health is insufficient.
- USPSTF. Most conservative; insufficient evidence to recommend either screening asymptomatic adults or routine supplementation in community-dwelling adults USPSTF 2021.
- Pop-wellness culture (Holick, Patrick, Attia, Huberman, etc.). Generally advocates 2,000–5,000 IU/day and higher 25(OH)D targets, often pairing with K2.
- Clinical labs. Profit on 25(OH)D testing volume; benefit from low thresholds that flag more "abnormal" results.
- Skeptics (Bolland, Reid, Avenell). Have published trial-sequential and meta-analytic critiques arguing that the evidence for non-deficient supplementation is weaker than the field believes Bolland et al. 2018.
Population variability
The dose-response and the baseline 25(OH)D distribution both vary widely across populations:
- Skin pigmentation. Eumelanin absorbs UVB; producing equivalent cutaneous D3 requires roughly 3–6× the sun exposure in Fitzpatrick VI vs Fitzpatrick II skin. US Black adults have a substantially lower mean 25(OH)D than US white adults (in NHANES, ~16 vs ~26 ng/mL), though debate continues on whether the relevant threshold is the same — Black adults have higher bone density at lower 25(OH)D, possibly due to higher bioavailable fraction (lower vitamin D binding protein).
- Latitude and season. Above ~35° latitude, cutaneous synthesis is effectively zero from late October to early March (the UVB index is too low). Below ~35°, year-round synthesis is possible.
- Age. Cutaneous 7-dehydrocholesterol concentration falls ~50% from age 20 to 70; renal 1α-hydroxylation also declines. Older adults need higher intakes to reach the same serum 25(OH)D.
- Adiposity. D3 partitions into adipose tissue; obese adults (BMI >30) achieve roughly half the rise per IU and need 2–3× the maintenance dose.
- Malabsorption. Crohn's, celiac, cystic fibrosis, gastric bypass impair fat-soluble vitamin absorption — higher doses or parenteral routes may be needed.
- Pregnancy and lactation. Demand rises; deficient pregnancies are associated with neonatal hypocalcemia, low birthweight, and possibly preeclampsia. Routine 400–1,000 IU/day is standard prenatal practice.
- Genetics. Polymorphisms in CYP2R1, GC (vitamin D binding protein), and VDR account for a meaningful share of inter-individual variance in serum 25(OH)D and possibly in response.
Knowledge gaps
- Whether the modest mortality benefit in meta-analyses survives in populations with higher baseline 25(OH)D — VITAL says no, the meta-analyses say possibly yes; the discrepancy is unresolved.
- Whether any benefit exists from pushing 25(OH)D above 50 ng/mL. RCT data essentially absent.
- Whether daily vs weekly dosing differs for non-skeletal outcomes. The bolus-is-worse signal is now established; the daily-vs-weekly question is less settled.
- Whether K2 co-supplementation matters for arterial health when D3 is taken at standard doses. Mechanism plausible, human RCT data absent.
- Whether the 25(OH)D threshold for clinical sufficiency is the same across ethnic backgrounds, given differences in vitamin D binding protein and bone density at low 25(OH)D.
- Whether long-term high-dose supplementation (5,000+ IU/day for decades) has subtle harms that would only emerge in very long follow-up — none demonstrated, none ruled out.
Scope vs brief. The brief named bone density, immune function, muscle strength, mood, and mortality. Each of those gets a home in the body and a meta dimension: bone via beauty_cumulative and the fall/fracture evidence in evidence/payoff; immune via health_short_term and the Martineau IPD analysis; mood at score 2 with the honest "subset only" framing; mortality as the headline longevity score of 3; muscle strength absorbed into energy (proximal myopathy in deficient adults) rather than getting its own dimension. No silent narrowing.
Score 39 — dream narrative was by choice. Overall lands at 39, just below the obligatory-40 line. Wrote a dream narrative anyway in the relief / not-being-conned register, since the honest hook for D3 is exactly that: a cheap floor-setter and a calm read of the literature, not aspiration. Dek and tagline pulled from the narrative carry the floor-setting / handled-it move; no false transformation.
Hardest scoring call: longevity at 3 vs 2. VITAL's null primary endpoints would push toward 2; the Cochrane and BMJ mortality meta-analyses (RR 0.94 and 0.89 for daily D3) push toward 3. Landed at 3 because the mortality signal is replicated across multiple meta-analyses, hinges specifically on daily dosing (matching the recommended protocol), and is more credible than VITAL's incident-cancer endpoint for the long-horizon question. A reviewer who downgrades to 2 on the strength of VITAL is defensible.
Hardest call #2: evidence at 4 vs 5. By raw study count and longevity of the literature, vitamin D earns 5. The reason I held at 4 is the headline failure mode: the biggest, best-designed trials of the past decade (VITAL, D2d, DO-HEALTH, D-Health) were null on the non-skeletal endpoints the field had been promising. That gap between "extensively studied" and "convincingly effective" is exactly what the 4-vs-5 distinction is for.
Contraindications limited to kidney-disease from the closed vocabulary. The genuinely important contraindications — sarcoidosis and other granulomatous diseases, primary hyperparathyroidism, baseline hypercalcemia — are not in the schema's closed vocabulary. Flagged them in the body's contraindications section instead, with the rationale spelled out. kidney-disease is the closest schema match and is genuine (kidney stones, the activation site is the kidney, calcium-D3 combo elevated stone risk in WHI).
Skipped: applicability lift for "avoid / never-start" — this is a do entry, so the standard prevalence reading applies. Landed at 4 on the basis that ~70% of US adults sit below the Endocrine Society's 30 ng/mL threshold and the decision is universal in temperate latitudes.
Future-link candidates. Several entries are mentioned in out-of-scope and should be wired up once they exist:
morning-sunlight(the natural pathway + circadian benefit)magnesium(hydroxylation cofactor)calcium(the classical bone partner; the WHI combination-product caveat lives here)vitamin-d-blood-testas a standalonetest-action entry, if 25(OH)D testing warrants its own treatment
Separate-entry candidates. The "what is osteomalacia and how to recognise it" topic could warrant its own know entry — it sits at the deficient edge of this one and is genuinely under-recognised in adults. Flagging for the backlog.
K2 / D3-K2 combo. Treated as a misconception rather than a separate entry. If the supplement world keeps marketing the combination aggressively enough that readers ask, it could justify its own short know entry.
Vitamin D3 (cholecalciferol)
1,000–5,000 IU softgels run $5–15 for a year's supply. Optional one-time 25(OH)D test adds $30–60.
One pill daily, ideally with a fat-containing meal. No routine retesting required after the first stable result.
One of the most-studied micronutrients: dozens of large RCTs (VITAL n=25,871, D-Health, DO-HEALTH, D2d, VITAL-Bone, VITAL-Depression), Cochrane and IPD meta-analyses, century-old mechanism understanding. Universal agreement on deficiency biology and bone effects; large trials in healthy adults largely null for non-skeletal hard endpoints. Not 5 because the headline non-skeletal claims (CVD, cancer prevention, fracture in healthy adults) failed in the biggest trials.
Cochrane meta-analysis of 56 trials (Bjelakovic 2014) shows daily D3 specifically reduces all-cause mortality (RR 0.94) and cancer mortality (RR 0.88); BMJ umbrella meta (Chowdhury 2014) concurs. Effect modest, dependent on daily (not bolus) dosing, and absent in the largest single trial (VITAL) on incident endpoints.
Modest reduction in acute respiratory infection rate (Martineau IPD meta-analysis, adjusted OR 0.88), most pronounced in those starting deficient. For the already-replete, no felt short-term change.
Replacement in severely deficient adults (<15 ng/mL) clearly resolves the fatigue and proximal myopathy of osteomalacia within weeks. In replete adults, no detectable energy effect (VITAL, DO-HEALTH).
Smaller trials in deficient or seasonally-affected adults show modest mood improvement; VITAL-Depression (Okereke 2020) definitively rejected D3 as prevention of incident depression in generally-replete adults. Net effect: real but limited to a subset.
Indirect long-term contribution through preserved bone density and muscle function in older age; no direct cutaneous or aesthetic effect.
Cognitive benefit limited to replacement of severe deficiency; DO-HEALTH found no cognitive effect in 2,000 IU/day for 70+ adults over 3 years.
Some observational links between low 25(OH)D and poor sleep quality, but no convincing RCT evidence that supplementation improves sleep in adults.