FitProven
Evidence Breakdown
Based on 8 studies
THE BIGGER QUESTION
Whether this one statement holds is settled above. What to actually do about it is a wider question, weighed across every claim that bears on it.
Evidence map
For & against, at a glance
Meta-Analysis
Pro
Zittermann A et al. · 2023European Journal of NutritionSystematic review and meta-analysis of 22 RCTs (12,952 participants) restricted to long-term daily doses of 3,200-4,000 IU — i.e. the top of the tolerable upper intake range rather than exotic megadoses. Compared with control, these doses raised hypercalcemia (RR 2.21, 95% CI 1.26-3.87, an absolute excess of about 4 cases per 1,000 people), falls (RR 1.25, 95% CI 1.01-1.55) and hospitalization (RR 1.16, 95% CI 1.01-1.33). There was no increase in kidney stones (RR 1.09, 95% CI 0.66-1.82), hypercalciuria (RR 1.40, 95% CI 0.91-2.17) or mortality (RR 1.07, 95% CI 0.75-1.52). The authors conclude 3,200-4,000 IU/day is "not completely safe" — real but small absolute harm.
0.85
Systematic review and meta-analysis of 22 RCTs (12,952 participants) restricted to long-term daily doses of 3,200-4,000 IU — i.e. the top of the tolerable upper intake range rather than exotic megadoses. Compared with control, these doses raised hypercalcemia (RR 2.21, 95% CI 1.26-3.87, an absolute excess of about 4 cases per 1,000 people), falls (RR 1.25, 95% CI 1.01-1.55) and hospitalization (RR 1.16, 95% CI 1.01-1.33). There was no increase in kidney stones (RR 1.09, 95% CI 0.66-1.82), hypercalciuria (RR 1.40, 95% CI 0.91-2.17) or mortality (RR 1.07, 95% CI 0.75-1.52). The authors conclude 3,200-4,000 IU/day is "not completely safe" — real but small absolute harm.
Design Meta-Analysis (1.0) × quality 0.85 = impact 0.85
View sourceMeta-Analysis
Pro
Malihi Z et al. · 2016The American Journal of Clinical NutritionSystematic review and meta-analysis of 48 long-term vitamin D supplementation trials (19,833 participants, supplementation for at least 24 weeks). Vitamin D significantly increased the risk of hypercalcemia (RR 1.54, 95% CI 1.09-2.18, P = 0.01, from 37 studies) and hypercalciuria (RR 1.64, 95% CI 1.06-2.53, P = 0.03, from 14 studies), but did not increase kidney stones (RR 0.66, 95% CI 0.41-1.09, P = 0.10, from 9 trials). Notably, the increased calcium-related risks were *not* dose related within the range studied, which cuts against a simple "more is more harmful" reading.
0.80
Systematic review and meta-analysis of 48 long-term vitamin D supplementation trials (19,833 participants, supplementation for at least 24 weeks). Vitamin D significantly increased the risk of hypercalcemia (RR 1.54, 95% CI 1.09-2.18, P = 0.01, from 37 studies) and hypercalciuria (RR 1.64, 95% CI 1.06-2.53, P = 0.03, from 14 studies), but did not increase kidney stones (RR 0.66, 95% CI 0.41-1.09, P = 0.10, from 9 trials). Notably, the increased calcium-related risks were *not* dose related within the range studied, which cuts against a simple "more is more harmful" reading.
Design Meta-Analysis (1.0) × quality 0.80 = impact 0.80
View sourceDouble-Blind RCT
Pro
Burt LA et al. · 2019JAMAThree-year double-blind RCT in 311 healthy, community-dwelling adults aged 55-70 without osteoporosis (287 completed), randomized to 400, 4,000 or 10,000 IU/day of vitamin D3 — the cleanest test of the claim in exactly the population it concerns. Higher doses produced *lower* radial volumetric bone mineral density: mean change -1.2% (400 IU), -2.4% (4,000 IU) and -3.5% (10,000 IU); tibial BMD fell significantly only at 10,000 IU. Bone strength (failure load) did not differ between groups, so the BMD loss did not translate into measurable weakening over three years. Hypercalcemia occurred in 0%, 4% and 9% of the three groups and hypercalciuria in 17%, 22% and 33%. High doses gave no bone benefit and a dose-dependent decrement.
0.77
Three-year double-blind RCT in 311 healthy, community-dwelling adults aged 55-70 without osteoporosis (287 completed), randomized to 400, 4,000 or 10,000 IU/day of vitamin D3 — the cleanest test of the claim in exactly the population it concerns. Higher doses produced *lower* radial volumetric bone mineral density: mean change -1.2% (400 IU), -2.4% (4,000 IU) and -3.5% (10,000 IU); tibial BMD fell significantly only at 10,000 IU. Bone strength (failure load) did not differ between groups, so the BMD loss did not translate into measurable weakening over three years. Hypercalcemia occurred in 0%, 4% and 9% of the three groups and hypercalciuria in 17%, 22% and 33%. High doses gave no bone benefit and a dose-dependent decrement.
Design Double-Blind RCT (0.85) × quality 0.90 = impact 0.77
View sourceDouble-Blind RCT
Con
Malihi Z et al. · 2019The American Journal of Clinical NutritionSafety analysis of the ViDA trial: 5,110 adults aged 50-84 in Auckland randomized double-blind to 100,000 IU vitamin D3 monthly (roughly 3,300 IU/day equivalent) or placebo, median follow-up 3.3 years — the largest randomized test of sustained high-dose vitamin D safety. Kidney stone events were no more common on vitamin D (76 vs 82 events; HR 0.90, 95% CI 0.66-1.23), nor were hospitalized urolithiasis events (7 vs 11; HR 0.62, 95% CI 0.24-1.26). Hypercalcemia was essentially absent (0 cases in the vitamin D arm vs 1 on placebo in the blood-tested subsample). A large, sustained high dose produced no detectable calcium-related harm.
0.72
Safety analysis of the ViDA trial: 5,110 adults aged 50-84 in Auckland randomized double-blind to 100,000 IU vitamin D3 monthly (roughly 3,300 IU/day equivalent) or placebo, median follow-up 3.3 years — the largest randomized test of sustained high-dose vitamin D safety. Kidney stone events were no more common on vitamin D (76 vs 82 events; HR 0.90, 95% CI 0.66-1.23), nor were hospitalized urolithiasis events (7 vs 11; HR 0.62, 95% CI 0.24-1.26). Hypercalcemia was essentially absent (0 cases in the vitamin D arm vs 1 on placebo in the blood-tested subsample). A large, sustained high dose produced no detectable calcium-related harm.
Design Double-Blind RCT (0.85) × quality 0.85 = impact 0.72
View sourceMeta-Analysis
Pro
Zittermann A et al. · 2023European Journal of NutritionSystematic review and meta-analysis of 22 RCTs (12,952 participants) restricted to long-term daily doses of 3,200-4,000 IU — i.e. the top of the tolerable upper intake range rather than exotic megadoses. Compared with control, these doses raised hypercalcemia (RR 2.21, 95% CI 1.26-3.87, an absolute excess of about 4 cases per 1,000 people), falls (RR 1.25, 95% CI 1.01-1.55) and hospitalization (RR 1.16, 95% CI 1.01-1.33). There was no increase in kidney stones (RR 1.09, 95% CI 0.66-1.82), hypercalciuria (RR 1.40, 95% CI 0.91-2.17) or mortality (RR 1.07, 95% CI 0.75-1.52). The authors conclude 3,200-4,000 IU/day is "not completely safe" — real but small absolute harm.
0.85
Systematic review and meta-analysis of 22 RCTs (12,952 participants) restricted to long-term daily doses of 3,200-4,000 IU — i.e. the top of the tolerable upper intake range rather than exotic megadoses. Compared with control, these doses raised hypercalcemia (RR 2.21, 95% CI 1.26-3.87, an absolute excess of about 4 cases per 1,000 people), falls (RR 1.25, 95% CI 1.01-1.55) and hospitalization (RR 1.16, 95% CI 1.01-1.33). There was no increase in kidney stones (RR 1.09, 95% CI 0.66-1.82), hypercalciuria (RR 1.40, 95% CI 0.91-2.17) or mortality (RR 1.07, 95% CI 0.75-1.52). The authors conclude 3,200-4,000 IU/day is "not completely safe" — real but small absolute harm.
Design Meta-Analysis (1.0) × quality 0.85 = impact 0.85
View sourceMeta-Analysis
Pro
Malihi Z et al. · 2016The American Journal of Clinical NutritionSystematic review and meta-analysis of 48 long-term vitamin D supplementation trials (19,833 participants, supplementation for at least 24 weeks). Vitamin D significantly increased the risk of hypercalcemia (RR 1.54, 95% CI 1.09-2.18, P = 0.01, from 37 studies) and hypercalciuria (RR 1.64, 95% CI 1.06-2.53, P = 0.03, from 14 studies), but did not increase kidney stones (RR 0.66, 95% CI 0.41-1.09, P = 0.10, from 9 trials). Notably, the increased calcium-related risks were *not* dose related within the range studied, which cuts against a simple "more is more harmful" reading.
0.80
Systematic review and meta-analysis of 48 long-term vitamin D supplementation trials (19,833 participants, supplementation for at least 24 weeks). Vitamin D significantly increased the risk of hypercalcemia (RR 1.54, 95% CI 1.09-2.18, P = 0.01, from 37 studies) and hypercalciuria (RR 1.64, 95% CI 1.06-2.53, P = 0.03, from 14 studies), but did not increase kidney stones (RR 0.66, 95% CI 0.41-1.09, P = 0.10, from 9 trials). Notably, the increased calcium-related risks were *not* dose related within the range studied, which cuts against a simple "more is more harmful" reading.
Design Meta-Analysis (1.0) × quality 0.80 = impact 0.80
View sourceDouble-Blind RCT
Pro
Burt LA et al. · 2019JAMAThree-year double-blind RCT in 311 healthy, community-dwelling adults aged 55-70 without osteoporosis (287 completed), randomized to 400, 4,000 or 10,000 IU/day of vitamin D3 — the cleanest test of the claim in exactly the population it concerns. Higher doses produced *lower* radial volumetric bone mineral density: mean change -1.2% (400 IU), -2.4% (4,000 IU) and -3.5% (10,000 IU); tibial BMD fell significantly only at 10,000 IU. Bone strength (failure load) did not differ between groups, so the BMD loss did not translate into measurable weakening over three years. Hypercalcemia occurred in 0%, 4% and 9% of the three groups and hypercalciuria in 17%, 22% and 33%. High doses gave no bone benefit and a dose-dependent decrement.
0.77
Three-year double-blind RCT in 311 healthy, community-dwelling adults aged 55-70 without osteoporosis (287 completed), randomized to 400, 4,000 or 10,000 IU/day of vitamin D3 — the cleanest test of the claim in exactly the population it concerns. Higher doses produced *lower* radial volumetric bone mineral density: mean change -1.2% (400 IU), -2.4% (4,000 IU) and -3.5% (10,000 IU); tibial BMD fell significantly only at 10,000 IU. Bone strength (failure load) did not differ between groups, so the BMD loss did not translate into measurable weakening over three years. Hypercalcemia occurred in 0%, 4% and 9% of the three groups and hypercalciuria in 17%, 22% and 33%. High doses gave no bone benefit and a dose-dependent decrement.
Design Double-Blind RCT (0.85) × quality 0.90 = impact 0.77
View sourceDouble-Blind RCT
Con
Malihi Z et al. · 2019The American Journal of Clinical NutritionSafety analysis of the ViDA trial: 5,110 adults aged 50-84 in Auckland randomized double-blind to 100,000 IU vitamin D3 monthly (roughly 3,300 IU/day equivalent) or placebo, median follow-up 3.3 years — the largest randomized test of sustained high-dose vitamin D safety. Kidney stone events were no more common on vitamin D (76 vs 82 events; HR 0.90, 95% CI 0.66-1.23), nor were hospitalized urolithiasis events (7 vs 11; HR 0.62, 95% CI 0.24-1.26). Hypercalcemia was essentially absent (0 cases in the vitamin D arm vs 1 on placebo in the blood-tested subsample). A large, sustained high dose produced no detectable calcium-related harm.
0.72
Safety analysis of the ViDA trial: 5,110 adults aged 50-84 in Auckland randomized double-blind to 100,000 IU vitamin D3 monthly (roughly 3,300 IU/day equivalent) or placebo, median follow-up 3.3 years — the largest randomized test of sustained high-dose vitamin D safety. Kidney stone events were no more common on vitamin D (76 vs 82 events; HR 0.90, 95% CI 0.66-1.23), nor were hospitalized urolithiasis events (7 vs 11; HR 0.62, 95% CI 0.24-1.26). Hypercalcemia was essentially absent (0 cases in the vitamin D arm vs 1 on placebo in the blood-tested subsample). A large, sustained high dose produced no detectable calcium-related harm.
Design Double-Blind RCT (0.85) × quality 0.85 = impact 0.72
View sourceShowing the 4 strongest of 8 studies. Tap any node to expand its detail.
Evidence
PRO (5)
PRO Double-Blind RCTn=287AGAINST FUNDER0.90 Burt LA, Billington EO et al. (2019)
Dose-dependent LOSS of radial volumetric BMD over 3 years: -1.2% (400 IU), -2.4% (4,000 IU), -3.5% (10,000 IU); no between-group difference in bone strength (failure load); hypercalcemia in 0% / 4% / 9%
Three-year double-blind RCT in 311 healthy, community-dwelling adults aged 55-70 without osteoporosis (287 completed), randomized to 400, 4,000 or 10,000 IU/day of vitamin D3 — the cleanest test of the claim in exactly the population it concerns. Higher doses produced lower radial volumetric bone mineral density: mean change -1.2% (400 IU), -2.4% (4,000 IU) and -3.5% (10,000 IU); tibial BMD fell significantly only at 10,000 IU. Bone strength (failure load) did not differ between groups, so the BMD loss did not translate into measurable weakening over three years. Hypercalcemia occurred in 0%, 4% and 9% of the three groups and hypercalciuria in 17%, 22% and 33%. High doses gave no bone benefit and a dose-dependent decrement.
Weighted 0.90 — 287 of 311 randomised completed the full 3 years (303 in the primary analysis) — a large, long, double-blind, dose-ranging RCT in exactly the population the claim concerns, with high retention. The funder is a private foundation that promotes high-dose vitamin D and several authors report Amgen/Eli Lilly ties, but the result runs directly against the funder's position, which strengthens rather than weakens it.
Funding: Pure North S'Energy Foundation (investigator-initiated grant; funder had no role in design, conduct, analysis or publication) — this result runs against the funder's own interest, which strengthens it.
JAMA
PRO Double-Blind RCTn=22560.90 Sanders KM, Stuart AL et al. (2010)
Annual 500,000 IU vitamin D increased falls (IRR 1.15, 95% CI 1.02-1.30, P = .03) and fractures (IRR 1.26, 95% CI 1.00-1.59, P = .047), with risk concentrated in the 3 months after each dose (fall RR 1.31)
Double-blind, placebo-controlled trial in 2,256 community-dwelling women aged 70+ at high fracture risk, given a single annual oral dose of 500,000 IU cholecalciferol for 3-5 years. The vitamin D group fell more (2,892 vs 2,512 falls; 83.4 vs 72.7 per 100 person-years; IRR 1.15, 95% CI 1.02-1.30, P = .03) and fractured more (171 vs 135; IRR 1.26, 95% CI 1.00-1.59, P = .047), with risk concentrated in the three months immediately after each dose (fall RR 1.31). Caveat on population: these were older, high-fracture-risk women rather than healthy adults, and the harm is tied to the annual bolus pattern rather than to a large daily intake.
Weighted 0.90 — n=2256 randomised, intention-to-treat, 6925 person-years over 3-5 years. Double-blind, placebo-controlled, independently funded (NHMRC), no financial disclosures. Large, long, and well powered for a harm signal that reached significance. Held just below the top band for self-reported falls (recall bias) and because the population is older high-fracture-risk women, so it is indirect for healthy adults; the harm is also tied to the annual bolus pattern, not to large daily intake.
Funding: National Health and Medical Research Council (grant 251682); Australian Government Department of Health and Ageing
JAMA
PRO Meta-Analysisn=129520.85 Zittermann A, Trummer C et al. (2023)
At 3,200-4,000 IU/day: hypercalcemia RR 2.21 (95% CI 1.26-3.87, about 4 excess cases per 1,000), falls RR 1.25 (1.01-1.55), hospitalization RR 1.16 (1.01-1.33); no increase in kidney stones, hypercalciuria or mortality
Systematic review and meta-analysis of 22 RCTs (12,952 participants) restricted to long-term daily doses of 3,200-4,000 IU — i.e. the top of the tolerable upper intake range rather than exotic megadoses. Compared with control, these doses raised hypercalcemia (RR 2.21, 95% CI 1.26-3.87, an absolute excess of about 4 cases per 1,000 people), falls (RR 1.25, 95% CI 1.01-1.55) and hospitalization (RR 1.16, 95% CI 1.01-1.33). There was no increase in kidney stones (RR 1.09, 95% CI 0.66-1.82), hypercalciuria (RR 1.40, 95% CI 0.91-2.17) or mortality (RR 1.07, 95% CI 0.75-1.52). The authors conclude 3,200-4,000 IU/day is "not completely safe" — real but small absolute harm.
Weighted 0.85 — 22 RCTs and 12,952 participants, deliberately restricted to a single clinically relevant dose band rather than pooling wildly heterogeneous doses, which is what makes it usable for the question people actually ask. No commercial sponsor and no conflicts declared. Main limitation is that harms like falls and hypercalcemia were adverse-event or secondary outcomes in the source trials, not endpoints those trials were powered to detect.
Funding: none declared
European Journal of Nutrition
PRO Meta-Analysisn=198330.80 Malihi Z, Wu Z et al. (2016)
hypercalcemia RR 1.54 (95% CI 1.09-2.18) and hypercalciuria RR 1.64 (95% CI 1.06-2.53); no increase in kidney stones (RR 0.66, 95% CI 0.41-1.09)
Systematic review and meta-analysis of 48 long-term vitamin D supplementation trials (19,833 participants, supplementation for at least 24 weeks). Vitamin D significantly increased the risk of hypercalcemia (RR 1.54, 95% CI 1.09-2.18, P = 0.01, from 37 studies) and hypercalciuria (RR 1.64, 95% CI 1.06-2.53, P = 0.03, from 14 studies), but did not increase kidney stones (RR 0.66, 95% CI 0.41-1.09, P = 0.10, from 9 trials). Notably, the increased calcium-related risks were not dose related within the range studied, which cuts against a simple "more is more harmful" reading.
Weighted 0.80 — k=48 trials, 19,833 participants, all supplementing for at least 24 weeks — large and directly on topic. The weakness is sparse outcome reporting: hypercalcemia was reported in only 37 of the 48 trials and kidney stones in just 9, so the stone estimate is underpowered. The calcium-related risks were not dose-related, which the authors flag as inconsistent with a simple causal dose-response. Funders are academic, taken from the indexed grant record — the article itself is closed access, so the verbatim funding footnote could not be read.
Funding: China Scholarship Council and the University of Auckland
The American Journal of Clinical Nutrition
PRO Double-Blind RCTn=2000.65 Bischoff-Ferrari HA, Dawson-Hughes B et al. (2016)
Higher doses raised 25(OH)D above 30 ng/mL far more often but gave no benefit on lower-extremity function (P = .26) and MORE falls: 66.9% fell on 60,000 IU/month and 66.1% on calcifediol vs 47.9% on 24,000 IU/month (P = .048)
One-year double-blind RCT in 200 community-dwelling adults aged 70+ with a prior fall (mean age 78; 58% vitamin D deficient at baseline), comparing 24,000 IU/month, 60,000 IU/month, and 24,000 IU plus 300 µg calcifediol monthly. The higher-dose arms were far better at pushing 25(OH)D above 30 ng/mL but showed no benefit on lower-extremity function (P = .26) and more falls: 66.9% of the 60,000 IU group and 66.1% of the calcifediol group fell, versus 47.9% on 24,000 IU (P = .048). Population caveat: fall-prone seniors, not healthy adults — but it is a rare randomized signal that more is worse, not merely useless.
Weighted 0.65 — n=200, one year, double-blind and randomised — a rare randomised signal that more vitamin D is actively worse, not merely useless. Held to 0.65 by: modest n for a falls endpoint, a borderline P value (.048) on what was not the primary outcome, and serious indirectness — fall-prone adults aged 70+ with a prior fall, not the healthy adults the claim is about. Partly industry-supplemented funding (DSM, MSD), though the finding runs against the funders' interest.
Funding: Swiss National Science Foundation and the VELUX Foundations, with supplementary investigator-initiated funds from Merck Sharp & Dohme AG, WILD and DSM Nutritional Products
JAMA Internal Medicine
AGAINST (2)
AGAINST Double-Blind RCTn=51100.85 Malihi Z, Lawes CMM et al. (2019)
kidney stones 76 vs 82 events (HR 0.90, 95% CI 0.66-1.23); hospitalised urolithiasis 7 vs 11 (HR 0.62, 95% CI 0.24-1.26); hypercalcemia essentially absent (0 vs 1 case)
Safety analysis of the ViDA trial: 5,110 adults aged 50-84 in Auckland randomized double-blind to 100,000 IU vitamin D3 monthly (roughly 3,300 IU/day equivalent) or placebo, median follow-up 3.3 years — the largest randomized test of sustained high-dose vitamin D safety. Kidney stone events were no more common on vitamin D (76 vs 82 events; HR 0.90, 95% CI 0.66-1.23), nor were hospitalized urolithiasis events (7 vs 11; HR 0.62, 95% CI 0.24-1.26). Hypercalcemia was essentially absent (0 cases in the vitamin D arm vs 1 on placebo in the blood-tested subsample). A large, sustained high dose produced no detectable calcium-related harm.
Weighted 0.85 — 5,110 adults randomised double-blind with a median 3.3 years of follow-up — by far the largest randomised test of sustained high-dose vitamin D safety, and publicly funded. Held back from a higher weight by outcome ascertainment rather than size: kidney-stone events were largely self-reported by questionnaire, and serum calcium was measured in only an ~8% blood-tested subsample, so the hypercalcemia finding rests on a small fraction of the cohort and is underpowered for rare events.
Funding: Health Research Council of New Zealand
The American Journal of Clinical Nutrition
AGAINST Double-Blind RCTn=373CONFLICTED0.75 Billington EO, Burt LA et al. (2020)
Mild hypercalcemia (2.56-2.64 mmol/L) in 15 participants — 0% at 400 IU, 3% at 4,000 IU, 9% at 10,000 IU/day (P = .002), all transient and resolving on repeat testing; hypercalciuria in 87 participants (17%, 22%, 31% by dose, P = .01); clinical adverse events distributed similarly across arms
Prespecified safety analysis of the same three-year Calgary RCT (373 healthy adults aged 55-70 randomized to 400, 4,000 or 10,000 IU/day), so it is not an independent cohort from Burt 2019 but it reports the clinical-safety side. Mild hypercalcemia (2.56-2.64 mmol/L) occurred in 15 participants — 0% at 400 IU, 3% at 4,000 IU, 9% at 10,000 IU (P = .002) — and every case resolved on repeat testing. Hypercalciuria affected 87 participants (17%, 22%, 31% by dose, P = .01) but was common even in the low-dose arm. Clinical adverse events were distributed similarly across groups, and the authors concluded that doses up to 10,000 IU/day were safe and well tolerated, with hypercalcemia rare, mild and transient.
Weighted 0.75 — 373 randomised (the safety analysis population) over three years, double-blind and dose-ranging — a strong design. Two deductions: this is a prespecified safety analysis of the same Calgary cohort as Burt 2019, so it is not independent evidence, and the funder (Pure North S'Energy Foundation) runs vitamin D supplementation programmes and therefore has a stake in a favourable safety finding.
Funding: Pure North S'Energy Foundation — the funder profits from this result.
The Journal of Clinical Endocrinology and Metabolism
NEUTRAL (1)
NEUTRAL Narrative Review0.40 Galior K, Grebe S et al. (2018)
Toxicity at serum 25(OH)D of 150-1220 ng/mL and calcium 11.1-23.1 mg/dL, caused by manufacturing errors and gross overdosing
Review of 13 published case reports of vitamin D intoxication, which sets the dose scale at which unambiguous harm actually appears. Intoxicated patients presented with serum 25(OH)D of 150-1220 ng/mL (roughly 375-3050 nmol/L, versus a target range of about 30-50 ng/mL) and serum calcium of 11.1-23.1 mg/dL, with vomiting, dehydration, pain and appetite loss. The causes were supplement manufacturing errors and gross overdosing by patients or prescribers, not ordinary supplement use. Genuine toxicity is real but requires exposures orders of magnitude above the 4,000 IU/day upper limit.
Weighted 0.40 — A narrative review of 13 published case reports; no total patient count is stated, and case reports cannot establish incidence or a dose threshold. The article carries no funding statement at all. It is useful only for showing the magnitude of exposure at which harm appears - orders of magnitude above the 4,000 IU/day upper limit.
Nutrients