Intervention
Vitamin D3 supplementation
Research Observatory
Evidence Update
Current evidence
Vitamin D3 at 2000 IU/day was compared with placebo in factorial VITAL and DO-HEALTH trials of generally healthy older adults, assessing fractures and other clinical outcomes. Vitamin D3 did not reduce incident fractures in VITAL's broadly recruited, non-deficiency-selected population. The selected VITAL and DO-HEALTH primary results do not support a broad disease-prevention claim for routine vitamin D supplementation in generally healthy older adults.
In VITAL Rhythm, vitamin D3 at 2,000 IU daily did not significantly prevent atrial fibrillation, an irregular heart rhythm, compared with placebo in adults without that condition at entry. The finding does not prove equivalence or address treatment of established vitamin D deficiency. This comparison shares participants with the trial’s omega-3 comparison rather than providing an independent replication. Population: 25119 adults without baseline cardiovascular disease, cancer or atrial fibrillation; mean 66.7, 50.8% women. Species: Homo sapiens Intervention and comparator: Vitamin D3 2000 IU daily Comparator: Matching vitamin-D placebo within the VITAL factorial design Endpoint: Incident atrial fibrillation; vitamin-D factorial contrast Reported result: 469 versus 431 events; HR 1.09 (0.96–1.25), P=0.19. Follow-up: Median 5.3 years
In D2d, high-dose vitamin D did not significantly reduce new diabetes diagnoses overall compared with placebo in adults with prediabetes who were not selected for vitamin D deficiency. The uncertainty includes both benefit and no effect. This does not establish equivalence, confirm a favourable subgroup result, or argue against treating an established deficiency. Population: 2423 adults with prediabetes, not selected for vitamin-D deficiency; 1211/1212; age/sex distribution not extracted. Species: Homo sapiens Intervention and comparator: Vitamin D3 4000 IU daily Comparator: Placebo Endpoint: Primary new-onset diabetes Reported result: 293 versus 323 cases; HR 0.88 (0.75–1.04), P=0.12. Follow-up: Median 2.5 years
In D-Health, monthly vitamin D was compared with placebo in Australian adults aged 60–84. The cardiovascular analysis favoured vitamin D numerically, but its uncertainty included no effect. Differences according to cardiovascular drug use were not statistically established. This does not make the overall trial positive or resolve the separate, still-unread correction to the mortality report. Population: 21315 randomized Australians aged 60–84; CV analysis 21302, mean 69, 54% men; mortality original analysis 21310. Species: Homo sapiens Intervention and comparator: Vitamin D3 60000 IU monthly Comparator: Placebo Endpoint: Major CV events secondary to the overall trial; main endpoint of this analysis Reported result: 637 versus 699; HR 0.91 (0.81–1.01); five-year risk difference −5.8/1000 (−12.2 to 0.5). Follow-up: Up to 5 years intervention; original mortality median 5.7 years
What changed
In the reviewed generally healthy populations, routine vitamin D did not reduce fractures or the selected primary cancer/cardiovascular outcomes.VITAL Rhythm did not demonstrate atrial-fibrillation prevention with vitamin D3 2000 IU daily.D2d did not show statistically significant diabetes prevention with vitamin D in its overall prediabetic population.The later D-Health cardiovascular analysis had a favorable point estimate whose confidence interval included no effect.
What did not change
These trials do not answer treatment of established deficiency, osteomalacia or every high-risk skeletal indication.Same event counts across separate factorial contrasts do not mean the same treatment; not deficiency-treatment evidence.CI includes clinically relevant benefit and null; not evidence against treating established deficiency or confirmation of a favorable subgroup.Drug-use interaction P=0.12 was not significant; component MI result does not make the overall trial positive. Does not resolve mortality correction.
Why it matters
Makes a clinically meaningful null visible without overgeneralizing it.
Main limitation
Not selected for vitamin D deficiency, low bone mass or osteoporosis; does not answer deficiency treatment or osteomalacia replacement. Not a claim about every dose, deficient patient, secondary endpoint or medical indication. Preserve multiplicity threshold.Same event counts across separate factorial contrasts do not mean the same treatment; not deficiency-treatment evidence. Design: Randomized 2x2 factorial ancillary trial Safety: AF is itself a clinically relevant adverse outcome; these doses do not settle high-dose formulation safety. Interpretation: Factorial contrasts can coincidentally have identical event counts.CI includes clinically relevant benefit and null; not evidence against treating established deficiency or confirmation of a favorable subgroup. Design: Randomized placebo-controlled trial Safety: No significant adverse-event difference reported; not proof of equivalence or safety in unstudied populations. Interpretation: P=0.12 is neither benefit confirmation nor equivalence.Drug-use interaction P=0.12 was not significant; component MI result does not make the overall trial positive. Does not resolve mortality correction. Design: Randomized placebo-controlled trial with secondary outcome report Safety: Original mortality sensitivity analysis cannot establish cancer harm; unresolved correction prevents updated primary numerical recommendation. Interpretation: Mortality correction unread; cardiovascular CI includes null and interaction nonsignificant.
Next proof
Indication-specific randomized evidence in deficient or high-risk populations, with clinically meaningful outcomes and prespecified multiplicity control.Deficiency-specific, prespecified clinical evidence if that indication is proposed.Prespecified deficiency-stratified clinical estimates or independently qualified synthesis.Prespecified replication or prospectively qualified clinical synthesis.
Summary
Vitamin D3 at 2000 IU/day was compared with placebo in factorial VITAL and DO-HEALTH trials of generally healthy older adults, assessing fractures and other clinical outcomes. Vitamin D3 did not reduce incident fractures in VITAL's broadly recruited, non-deficiency-selected population. The selected VITAL and DO-HEALTH primary results do not support a broad disease-prevention claim for routine vitamin D supplementation in generally healthy older adults.
Intelligence Take
Makes a clinically meaningful null visible without overgeneralizing it.
Limitations
Not selected for vitamin D deficiency, low bone mass or osteoporosis; does not answer deficiency treatment or osteomalacia replacement. Not a claim about every dose, deficient patient, secondary endpoint or medical indication. Preserve multiplicity threshold.