What do the trials say — and why did the billion-person supplement lose its miracle billing?
Vitamin D's evidence arc is the supplement era's defining story: strong deficiency-correction biology, blockbuster observational associations, and a decade of general-population trials that came back modest. The field teaches the surrogate-versus-outcome lesson at grand scale — blood levels associate with everything, correction helps the deficient with specific things, and megadoses harm the elderly in the one trial that tested them. The honest summary this page grades: partnership effects for bone, deficient-subgroup effects for immunity, association-with-modest-correction-benefit for mood, and dose discipline everywhere.
The field's rise, correction, and discipline teach every reading skill at once:
The industrial revolution's sunless cities produce the rickets epidemic; cod liver oil enters medicine as the folk cure — a century before the vitamin is isolated.
The antirachitic factor is identified — the fourth vitamin, named in sequence. The ultraviolet-light cure is demonstrated the same era: sun or oil, either works.
Vitamin D added to milk at population scale — the rickets epidemic in fortified countries ends within a generation. One of nutrition's greatest public-health wins.
The two-step liver-kidney hydroxylation is discovered, and the field realizes: vitamin D is a prohormone. Receptors are found across tissues — the story expands beyond bone.
VITAL and the megadose trials deliver the discipline: general-population supplementation shows modest results, annual boluses cause harm, and the benefit map concentrates in the deficient.
The 25-OH test becomes routine, dosing individualizes to bloodwork, and the field's honest consensus: correct deficiency reliably, support bone with the partnership, promise nothing beyond it.
POPULATION · 2,256 community-dwelling women aged 70+, single annual 500,000 IU dose, ~3 years
FINDING · The megadose group experienced MORE falls and MORE fractures than placebo — approximately 15% higher rates, concentrated in the first months after dosing.
LIMITATION · Single trial ending a practice; the mechanism (receptor dysregulation from supraphysiologic spikes) is hypothesized, not proven — but the harm signal was decisive enough to end annual-bolus protocols worldwide.
POPULATION · 36,282 postmenopausal women, 1,000 mg calcium + 400 IU D3 daily, 7 years
FINDING · A modest hip-fracture reduction overall, statistically clearest in the women who actually took the supplements — with a small stone-risk increase in the supplement arm.
LIMITATION · The 400 IU dose is below modern targeting and adherence diluted the intent-to-treat effect — the adherent-subgroup signal and the dose limitation are both part of the honest reading. Our osteoporosis page carries the full analysis.
POPULATION · 25,871 generally healthy US adults, 2,000 IU D3 daily, median 5.3 years
FINDING · No significant reductions in invasive cancer or major cardiovascular events in the primary analyses — the boundary-drawing null for general-population supplementation.
LIMITATION · The population was largely replete (baseline ~28 ng/mL), and the null defines where supplementation doesn't help — not a refutation of deficiency correction. Subgroup analyses continue; the headline is the boundary.