The Science Behind Multivitamins & General Health Supplements

Therapeutic dose: 1,000–2,000 IU/day (higher to correct deficiency)

See ranked Vitamin D products→

• Pooled 8 trials in adults 65+: 700–1,000 IU/day cut fall risk by ~19%. Lower doses showed no benefit. Strongest clinical case is the older, at-risk population.
• Individual-participant meta-analysis of 25 RCTs: daily/weekly vitamin D modestly reduced acute respiratory infections, largest effect in those starting below 25 nmol/L. Bolus dosing did not work.PubMed ↗
• VITAL, the largest primary-prevention trial: 2,000 IU/day for 5+ years did not reduce cancer or cardiovascular events in generally healthy, already-replete adults.
• D-Health: monthly vitamin D in 21,315 older Australians for ~6 years showed no reduction in all-cause, cancer, or cardiovascular mortality. Converged with VITAL.
• A single annual 500,000 IU mega-dose in older women INCREASED falls and fractures, concentrated in the months after dosing. More is not better.

Therapeutic dose: 200–400 mg/day elemental

See ranked Magnesium products→

• Pooled 34 RCTs: supplementation lowered systolic BP ~2.0 mmHg and diastolic ~1.8 mmHg, with larger effects in low-magnesium, hypertensive, or diabetic subgroups.
• Pooled migraine trials: 300–600 mg/day for 8–12 weeks consistently reduced attack frequency and intensity — a modest but clinically useful effect comparable to some first-line prophylactics.
• Meta-analyses converge on modest improvements in fasting glucose, HbA1c, and insulin resistance in adults with type 2 diabetes or pre-diabetes.
• Cochrane review of 11 RCTs found NO benefit over placebo for idiopathic muscle cramps in older adults — the popular 'magnesium fixes cramps' claim does not survive controlled trials.

Therapeutic dose: 250–500 mg/day combined EPA+DHA (general); up to 4 g/day for triglycerides

• Classic dose-response: each 1 g/day of combined EPA+DHA lowers fasting triglycerides ~5–10%, with larger absolute drops at higher baseline. The most reproducible omega-3 finding.PubMed ↗
• REDUCE-IT: 4 g/day pure EPA (icosapent ethyl) in 8,179 high-risk statin-treated adults cut major cardiovascular events 25%. A pharmaceutical dose and form, not a grocery-store capsule. Industry-funded.PubMed ↗
• VITAL: 1 g/day combined EPA+DHA did not reduce major cardiovascular events in generally healthy adults. Closes the door on routine fish oil for primary prevention in well-fed populations.PubMed ↗
• STRENGTH: combined EPA+DHA at the same 4 g/day dose as REDUCE-IT was stopped early for futility — no cardiovascular benefit, and atrial fibrillation rose ~70%. Form and ratio matter, not just total dose. Industry-funded.PubMed ↗
• Pooled 13 RCTs in major depression: a small-to-moderate benefit driven entirely by EPA-dominant formulas; DHA-pure blends showed no signal. Best as an add-on to antidepressants.PubMed ↗

Therapeutic dose: 30–80 mg/day elemental (deficiency); alternate-day dosing absorbs better

See ranked Iron products→

• 198 non-anemic women with low ferritin and fatigue: 80 mg/day ferrous sulfate cut fatigue 47.7% vs 28.8% on placebo over 12 weeks. The mechanism is repletion, not pharmacology.PubMed ↗
• Cochrane pooled 44 trials (43,274 women): daily iron in pregnancy cut maternal anemia at term ~70% and low birthweight ~20%. WHO-recommended in low-resource settings.PubMed ↗
• Pooled trials in school-age children: iron improved attention, concentration, and IQ — but the effect was concentrated in iron-deficient kids; replete children gained little.PubMed ↗
• In iron-replete adults, no strong RCT evidence that iron reduces cardiovascular events, mortality, or cancer — and high body iron stores may track higher cardiovascular risk, especially in men.

Therapeutic dose: 2.4 µg/day (RDA); 500–1,000 µg/day to correct deficiency

See ranked Vitamin B12 products→

• 600 Nepalese infants at risk of poor B12 status: daily B12 improved neurodevelopmental scores vs placebo, most in those with the lowest baseline. The deficiency-correction story in miniature.
• 60 adults with megaloblastic anemia: high-dose ORAL B12 was non-inferior to intramuscular injections for normalizing hematology and B12 at 90 days. Injections rarely necessary.
• Meta-analysis of 29 studies (8,089 patients): metformin use roughly 2.5x the odds of B12 deficiency. Long-term metformin means monitor B12.
• 195 older adults with mild biochemical deficiency: B12 levels rose substantially but cognition did NOT improve over 24 weeks in those without anemia or overt neurologic disease.
• Long-term multivitamin (with B12) in 5,947 male physicians 65+, ~12 years: no effect on cognitive decline or dementia. Routine supplementation in the non-deficient is mostly placebo.

Therapeutic dose: 400 µg/day (preconception); 4 mg/day with prior NTD pregnancy

• MRC trial: in women with a prior neural-tube-defect pregnancy, 4 mg/day folic acid reduced recurrence 72%. Stopped early once results were clear — the basis for global fortification policy.PubMed ↗
• Czeizel-Dudás: 4,156 women with no prior NTD history; the folic-acid multivitamin group had zero NTDs vs six in controls. Established prevention of FIRST occurrences too.PubMed ↗
• Cochrane pooled 5 trials (7,391 women): folic acid cut neural tube defects ~70% (RR 0.31), across 360 µg–4 mg/day. The strongest evidence in the entire supplement-and-pregnancy literature.
• NORVIT: lowering homocysteine ~28% in 3,749 post-heart-attack patients did not reduce recurrent events, with a hint of harm in the combined-vitamin arm. Homocysteine-lowering ≠ fewer events.PubMed ↗
• 1,021 adults with prior colorectal adenomas: 1 mg/day folic acid did not reduce recurrence and trended toward MORE advanced/multiple adenomas. Don't mega-dose folic acid long-term.PubMed ↗

Therapeutic dose: 8–11 mg/day (RDA); cap chronic intake well below 40 mg/day

See ranked Zinc products→

• Pooled 7 RCTs (575 cold patients): high-dose zinc lozenges (≥75 mg/day) cut cold duration 33% — but only when started within 24 hours of symptom onset.PubMed ↗
• 50 older adults: 45 mg/day zinc gluconate for 12 months reduced infection incidence and inflammatory markers. Older adults often run subclinically deficient — repletion appears protective.PubMed ↗
• 46,974 men followed 14 years: zinc above 100 mg/day was linked to a 2.29x higher risk of advanced prostate cancer. More is not better — chronic high-dose zinc also causes copper deficiency.PubMed ↗
• 470 hospitalized COVID-19 patients: oral zinc did not reduce mortality or ICU admission vs placebo. The COVID-zinc hypothesis did not survive a properly powered trial.PubMed ↗

Therapeutic dose: 75–90 mg/day (RDA); plasma saturates around 200 mg/day

• Cochrane pooled 29 trials (11,306 people): daily vitamin C did NOT prevent colds in the general population. It shortened duration modestly — ~8% in adults, ~14% in children.
• Experimental scurvy: volunteers on a vitamin-C-free diet developed bleeding gums and perifollicular hemorrhages within ~4 weeks; 10 mg/day fully reversed it. The basis for the RDA.
• 8-year trial in 14,641 male physicians: 500 mg/day vitamin C produced no reduction in total cancer. The largest, cleanest cancer-prevention test came back null.
• LOVIT: 863 ICU sepsis patients on IV vitamin C did WORSE than placebo (44.5% vs 38.5% death or organ dysfunction at day 28). Reversed earlier enthusiasm.

Therapeutic dose: 1,000–1,200 mg/day total (food first); pair with vitamin D

• Pooled 29 trials (63,897 adults 50+): calcium with or without vitamin D reduced fractures 12%, with a much larger effect at ≥1,200 mg/day calcium plus ≥800 IU/day vitamin D. Adherence drove the size.PubMed ↗
• WHI: 1,000 mg calcium + 400 IU vitamin D in 36,282 postmenopausal women raised hip BMD 1.06% but only modestly, non-significantly cut hip fracture in intention-to-treat. Kidney stones rose 17%.PubMed ↗
• Pooled 15 trials: calcium supplements WITHOUT coadministered vitamin D raised heart-attack risk ~31%. Triggered the modern debate over calcium-only supplementation.PubMed ↗
• Pooled 13 pregnancy trials (15,730 women): ≥1 g/day calcium roughly halved pre-eclampsia risk, largest in low-baseline-intake women. WHO recommends it for low-intake populations.PubMed ↗

Therapeutic dose: 150 mcg/day (RDA); 220 mcg/day in pregnancy

• Universal salt iodization eliminated endemic goiter and cretinism and raised population cognition in formerly deficient regions — one of the great 20th-century public-health wins.
• RCT in mildly iodine-deficient New Zealand children: 150 mcg/day for 28 weeks improved a cognitive composite by 0.19 SD vs placebo.PubMed ↗
• In iodine-replete adults, supranormal intake (kelp, mega-dosing, amiodarone) can trigger iodine-induced hyperthyroidism and worsen Hashimoto's — the dose-response is U-shaped.
• RCT in mildly deficient pregnant women found no clear improvement in maternal thyroid function or child neurodevelopment — benefit is concentrated in severe deficiency, not marginal.PubMed ↗

Therapeutic dose: 55 mcg/day (RDA); UL 400 mcg/day

• Population-scale selenium repletion essentially eliminated Keshan disease (a fatal cardiomyopathy) in low-soil regions of China — the strongest case for selenium, and it is deficiency correction.
• SELECT (35,533 healthy men): selenium produced no reduction in prostate cancer and was halted early for futility; vitamin E increased prostate cancer by 17%.PubMed ↗
• Pooled Hashimoto's trials: 200 mcg/day modestly lowered anti-TPO antibody titers over 3-6 months, but effects on actual thyroid function (TSH, free T4) were small and inconsistent.PubMed ↗
• Secondary analysis of the NPC trial found increased incident type 2 diabetes with 200 mcg/day, concentrated in those starting from high baseline selenium.PubMed ↗

Therapeutic dose: Strain-specific (e.g. ≥5 billion CFU/day L. rhamnosus GG or S. boulardii for AAD)

• Cochrane review of 39 RCTs (8,672 patients): probiotics during antibiotics cut C. difficile-associated diarrhea risk by ~60%, with benefit concentrated in higher-baseline-risk patients.PubMed ↗
• Meta-analysis of 53 RCTs (5,545 IBS patients): probiotics beat placebo on global symptoms, but no single strain emerged as clearly superior.PubMed ↗
• Pediatric Cochrane review (33 trials): high-dose L. rhamnosus GG or S. boulardii roughly halved antibiotic-associated diarrhea, NNT ~9; low-dose multi-strain blends did not reliably help.PubMed ↗
• The strain number is the whole story — 'general gut/immune health' and generic '30 billion CFU multi-strain' claims have no clean RCT support in healthy adults.

Therapeutic dose: 1.3–1.7 mg/day (RDA); UL 100 mg/day

• Pooled review of 9 PMS trials (940 women): up to 100 mg/day roughly doubled the rate of overall symptom improvement vs placebo, though included trials were small and mixed in quality.PubMed ↗
• RCT in pregnancy nausea: 25 mg every 8 hours significantly reduced severe nausea — basis for ACOG's first-line B6 (+doxylamine) recommendation.PubMed ↗
• Case series established that sustained high-dose pyridoxine causes sensory neuropathy — the irony being it can cause the very nerve damage it is sometimes taken to treat.PubMed ↗
• HOPE-2 (5,522 high-risk adults): B6/folate/B12 lowered homocysteine but did not reduce the cardiovascular composite — the homocysteine hypothesis failed at hard endpoints.

Therapeutic dose: 90–120 mcg/day (AI); typical K2 supplements 90–180 mcg/day

• Vitamin K1 is the established cofactor for clotting factors II, VII, IX, X; routine newborn K1 prevents hemorrhagic disease — the one universally accepted role. Never start K with warfarin without a prescriber.
• Headline MK-7 bone trial (180 mcg/day, 3 years) showed modest BMD preservation in postmenopausal women — but it was funded by the dominant MK-7 supplier.PubMed ↗
• AVADEC (independent): 2 years of 720 mcg/day MK-7 + vitamin D did not slow aortic valve calcification vs placebo.PubMed ↗
• Pooled K2 trials in postmenopausal women showed a modest BMD effect but no significant fracture reduction — the only clinically meaningful endpoint.PubMed ↗

Therapeutic dose: 700–900 mcg RAE/day (RDA); UL 3,000 mcg (10,000 IU) preformed

• CARET (18,314 smokers/asbestos-exposed) was halted early: high-dose beta-carotene + retinol raised lung cancer ~28% and total mortality ~17%. Smokers should never take supplemental beta-carotene.
• ATBC (29,133 Finnish male smokers): supplemental beta-carotene increased lung cancer incidence ~18% — the harm signal that CARET replicated.
• Bjelakovic meta-analysis of 68 antioxidant trials linked supplemental vitamin A to a 16% rise in all-cause mortality; beta-carotene ~7% — the antioxidant cancer-prevention hypothesis collapsed in pooled data.
• Profound benefit only in deficiency: high-dose vitamin A cut child mortality ~12% in low-income settings (Cochrane, 1.2M children). No general-health benefit in well-fed adults; teratogenic at high dose in pregnancy.

Therapeutic dose: 15 mg/day (RDA); deficiency rare outside fat malabsorption

• SELECT (35,533 men): 400 IU/day alpha-tocopherol increased prostate cancer incidence by 17% — a hard contraindication for routine high-dose use in men.
• Miller meta-analysis (19 trials, 135,967 people): doses ≥400 IU/day associated with a ~4% increase in all-cause mortality — the result that drove guidelines against high-dose supplementation.
• HOPE / HOPE-TOO (9,541 adults, 7-year extension): 400 IU/day natural vitamin E showed no cardiovascular benefit and a 19% increase in heart-failure hospitalization.
• Women's Health Study (39,876 women, ~10 years): alternate-day vitamin E produced no reduction in cardiovascular events or cancer — adding to a consistent null prevention picture.

Therapeutic dose: 1.1–1.2 mg/day (RDA)

• Repletion is dramatic and life-saving in genuine deficiency (beriberi, Wernicke's, alcohol-related depletion) — but that is medical-grade deficiency correction, not nutritional supplementation.
• No controlled trial shows oral thiamine improves energy or fatigue in healthy adults with adequate intake — the 'B-vitamin energy boost' is mechanism, not evidence.
• ACTS trial (205 septic-shock patients): IV vitamin C + hydrocortisone + thiamine showed no improvement in organ failure or mortality — the 'metabolic resuscitator' hypothesis did not hold up.
• Benfotiamine showed only borderline, short-term symptom benefit in diabetic neuropathy (BENDIP, per-protocol p=0.033, ITT p=0.055) — the sole non-deficiency use case, and it is modest.

Therapeutic dose: 1.1–1.3 mg/day (RDA); 400 mg/day for migraine prophylaxis

• Schoenen RCT (55 adults): 400 mg/day for 3 months reduced migraine frequency with an NNT of 2.3 — striking, but a single small trial with limited, inconsistent replication.PubMed ↗
• No controlled trial shows riboflavin improves energy or fatigue in non-deficient adults — the B-complex 'energy boost' framing is mechanism repurposed as marketing.
• Repletion reliably resolves the cracked-lip, sore-tongue, inflamed-eye signs of frank deficiency — the only context with strong, settled evidence; deficiency is uncommon in adults eating dairy or fortified foods.

Therapeutic dose: 14–16 mg/day (RDA); supplemental UL 35 mg/day

• HPS2-THRIVE (25,673 statin-treated patients): adding niacin produced no reduction in vascular events and increased new diabetes diagnoses by ~33%, plus more bleeding, infection and gout.
• AIM-HIGH (3,414 statin-treated patients): extended-release niacin raised HDL and lowered triglycerides but produced zero reduction in cardiovascular events; stopped early for futility.
• Niacin reverses pellagra (dermatitis, diarrhea, dementia) as reliably as vitamin C reverses scurvy — but deficiency is rare with grain fortification, and replete adults gain nothing extra.

Therapeutic dose: 5 mg/day (AI)

• Named from the Greek for 'everywhere' — B5 is in nearly every food, frank deficiency is essentially never diagnosed, and no trial shows supplementation helps non-deficient adults.
• One small industry-funded RCT (48 adults) of a B5-based multi-ingredient formula reduced acne lesions at 12 weeks — but the effect cannot be attributed to B5 alone and was never replicated.
• 'Adrenal fatigue' is not a recognized diagnosis; the B5-for-adrenals/energy claim extrapolates from rat studies and mechanism, with no supporting human outcome trial.

Therapeutic dose: 30 mcg/day (AI)

• Systematic review of every published case of biotin for hair: improvement occurred ONLY in people with documented deficiency or a genetic biotin disorder. No RCT shows biotin grows hair in healthy, non-deficient adults.PubMed ↗
• High-dose biotin (≥5 mg/day) interferes with streptavidin-biotin immunoassays — falsely altering TSH, troponin, hCG and hormone results; the FDA has reported a death from a missed heart attack via false-low troponin.
• SPI2 phase 3 (642 progressive MS patients): pharmaceutical-grade 300 mg/day biotin failed its disability-improvement endpoint (12% vs 9%, not significant).
• The strongest non-deficiency signal is for brittle nails — two open-label 1990s studies at 2.5 mg/day, neither placebo-controlled; methodologically thin and far weaker than the marketing implies.

Therapeutic dose: 900 mcg/day (RDA); UL 10 mg/day

• The real-world concern is induced deficiency: chronic zinc >40 mg/day blocks copper absorption and causes anemia, neutropenia and a myelopathy whose neurological damage may not fully reverse.PubMed ↗
• Copper is a real cofactor (ceruloplasmin, cytochrome c oxidase, SOD, lysyl oxidase), but no controlled trial shows supplemental copper benefits skin, vascular or cognitive outcomes in non-deficient adults.
• A 2-year RCT found calcium + trace minerals (copper, zinc, manganese) arrested postmenopausal bone loss better than calcium alone — but the combination cannot isolate copper's contribution.PubMed ↗

Therapeutic dose: 1.8–2.3 mg/day (AI); UL 11 mg/day

• Frank manganese deficiency is one of the rarest nutritional disorders in humans — virtually any diet with grains, legumes, nuts, greens or tea exceeds the requirement.
• Chronic excess is the real concern: manganism (a parkinson-like syndrome with basal-ganglia deposition) occurs in welders, miners and high-manganese well-water exposure.PubMed ↗
• Children drinking high-manganese water show measurable IQ deficits with a consistent dose-response — reinforcing that more is not better.PubMed ↗

Therapeutic dose: 25–35 mcg/day (AI)

• Meta-analysis of 14 RCTs (875 patients): chromium picolinate — the most-marketed form — showed no significant effect on HbA1c in type 2 diabetes.PubMed ↗
• Western RCT in insulin-treated type 2 diabetes found no effect of chromium on glycemic control — directly contradicting the often-cited 1997 Chinese trial in a more deficient population.PubMed ↗
• Meta-analysis of 11 RCTs found no effect of chromium on blood pressure; no controlled trial supports the weight-loss marketing either.

Therapeutic dose: 45 mcg/day (RDA); UL 2,000 mcg/day

• Molybdenum is the metal center of four human enzymes (notably sulfite oxidase), but the requirement is tiny and reliably met by legumes, grains, nuts and organ meat.
• The only documented acquired deficiency was a single long-term parenteral-nutrition case before TPN routinely contained trace minerals — modern formulations have eliminated it.
• 'Detox' and sulfite-sensitivity marketing rests on enzyme roles, not data — no human trial shows supra-dietary molybdenum improves any health endpoint in healthy adults.

Therapeutic dose: 700 mg/day (RDA) — routinely exceeded by 50–100% from diet

• Population phosphate intake exceeds the RDA by 50–100%, driven by additives in processed meats, colas and baked goods — dietary insufficiency is essentially nonexistent in healthy adults.PubMed ↗
• Framingham cohort (3,368 adults): higher serum phosphorus, even within the normal range, was associated with increased incident cardiovascular events over long follow-up.PubMed ↗
• Phosphate replacement is essential only in clinical settings (refeeding syndrome, severe alcohol use disorder) — irrelevant to retail supplementation in healthy adults.