Supplements to Avoid (or Use with Caution)

Some supplements are not just useless — they actively raise mortality, cancer risk, or fracture risk at doses people routinely take. Here are the documented harms, with the trials that established them.

The pattern is consistent: several widely sold supplements were once enthusiastically recommended on the strength of observational data or mechanistic reasoning, and then large randomized controlled trials (RCTs) revealed a harm signal at the doses people actually use.

The clearest example is the antioxidant class as a whole. Pooling 78 randomized trials in almost 300,000 people, the antioxidant supplements studied did not prevent death — and when the analysis was restricted to the low-bias trials, beta-carotene, vitamin E, and higher-dose vitamin A were associated with slightly higher mortality.[1] The reason is now well understood: reactive oxygen species are not just damage, they are signalling molecules the body uses to trigger its own defenses, so flooding the system with isolated scavengers blunts the very adaptations (to exercise, to dietary stress) that the antioxidants were supposed to help. Antioxidants from a whole-food matrix — fruit, vegetables, olive oil, tea — do not carry this signal, because the doses are far lower and arrive packaged with the cofactors that regulate them.

Vitamin E (≥400 IU/day) — Strong harm signal

Avoid supplemental doses above ~30 IU. Food sources (nuts, seeds, olive oil) are sufficient.

The evidence for harm:

  • All-cause mortality ↑ at high doses — a 2005 meta-analysis found about 39 extra deaths per 10,000 people.[2]
  • Heart failure hospitalizations ↑ in the HOPE-TOO trial, in people with vascular disease or diabetes.[3]
  • Prostate cancer ↑ in men — SELECT reported about 17% more prostate cancer in men taking vitamin E (hazard ratio, HR, 1.17).[4]
  • Stroke — a trade-off, not a straight harm. Across nine trials in 118,765 people, the five that reported hemorrhagic (bleeding) stroke showed about 22% more of it — an estimate that only just cleared statistical significance — while the five reporting ischemic (clot) stroke showed about 10% less, and total stroke was unchanged in the seven trials that reported it. In the authors' own absolute terms, that is one extra hemorrhagic stroke for every 1,250 trial participants taking vitamin E, against one ischemic stroke prevented for every 476. They still advise against indiscriminate use, because bleeding strokes are the more damaging kind.[5]

The repeated finding across multiple large trials makes this one of the clearest "supplement causes harm" stories in medicine. Part of the mechanism: supplements deliver isolated alpha-tocopherol, but the diet supplies a balance of tocopherols, and high-dose alpha-tocopherol lowers the body's tissue levels of gamma-tocopherol — the form that does most of the anti-inflammatory work. So a megadose can leave you worse off than the food it was meant to concentrate.[6]

Beta-carotene in current or former smokers — Strong harm signal

Avoid entirely if you smoke or have ever smoked.

  • The ATBC trial showed an 18% increase in lung cancer in male smokers given beta-carotene.[7]
  • The CARET trial was stopped early due to a 28% increase in lung cancer.[8]
  • The US Preventive Services Task Force (USPSTF) issued a Grade D recommendation against routine vitamin E or beta-carotene for cardiovascular disease (CVD) or cancer prevention.[9]

Get carotenoids from food (orange/leafy vegetables) — no toxicity risk.

Preformed vitamin A above 3000 µg RAE/day — Caution

The concern is preformed vitamin A (retinol — from liver, supplements, and retinol-heavy multivitamins), not the beta-carotene in vegetables, which the body converts to vitamin A only as needed. Above roughly 3,000 µg RAE (retinol activity equivalents) per day:

  • ↑ hip fracture risk — in the Nurses' Health Study (72,337 postmenopausal women, 18 years of follow-up, 603 hip fractures), women taking in at least 2,000 µg/day of retinol had about 90% higher hip-fracture risk than women taking in under 500 µg/day (relative risk, RR, 1.89; the 95% confidence interval, 1.33 to 2.68, is the range in which the true value most plausibly lies, and because it stays above 1.0 chance alone is an unlikely explanation). The widely quoted 48% figure (RR 1.48) from the same study is a different contrast — total vitamin A above 3,000 µg/day measured in retinol equivalents, an older unit that credits dietary carotenoids more generously than RAE does — so it is not a preformed-vitamin-A number. Beta-carotene on its own was not significantly associated with fracture, and the retinol association was weaker in women taking postmenopausal estrogen.[10] Note that this fracture signal starts below the 3,000 µg RAE ceiling named above — that ceiling is set on liver toxicity and teratogenicity, not on bone.
  • Hepatic fibrosis (liver scarring) and teratogenicity (harm to a developing fetus) at chronic high intake.
  • Pregnancy: teratogenic at high doses; avoid retinol-heavy multivitamins.

Get vitamin A from food (orange/leafy vegetables = beta-carotene without toxicity risk; liver, dairy, eggs in moderation).

Mega-dose vitamin C (>2000 mg/day) — Caution

Vitamin C is genuinely benign at normal intakes — it's water-soluble and the excess is cleared in urine. The problem is only the gram-plus "immune megadose":

  • Kidney stones — plausible, not settled. The adult upper limit is 2,000 mg/day. High vitamin C intakes can raise the amount of oxalate and uric acid excreted in urine, and oxalate is what most stones are built from, which the NIH Office of Dietary Supplements says "could contribute to the formation of kidney stones, especially in individuals with renal disorders" — but the same factsheet calls the studies on oxalate excretion, spanning intakes from 30 mg to 10 g/day, conflicting, says it is "not clear whether vitamin C actually plays a role in the development of kidney stones", and places the best evidence narrowly, in people who already have hyperoxaluria (persistently high urinary oxalate).[11]
  • The one large human cohort on this question is in men. In a prospective cohort of Swedish men, those taking ascorbic-acid supplements — typically 1,000 mg tablets — were nearly twice as likely to have a first kidney stone as non-users, while multivitamin users, whose vitamin C dose is far lower, were not.[12] The same association has not been found in women.
  • Pro-oxidant chemistry — a test-tube result, not yet a human one. In a test tube, ascorbate reduces free iron and feeds the Fenton reaction that generates the hydroxyl radical, the most damaging oxidant the body makes. Whether that happens inside people is unsettled: the review most often cited for the effect says the in-vitro evidence is extensive but "there is no convincing and unambiguous evidence of such in vivo activity", probably because iron and copper are kept bound to carrier proteins rather than left loose.[13] The exception is genuine iron overload: in hereditary hemochromatosis, chronic high-dose vitamin C could worsen iron loading and damage tissue — while in healthy people excess iron absorption from vitamin C does not appear to be a concern.[14]

There's no benefit to chronic megadosing for a healthy adult. Food (citrus, peppers, kiwi, broccoli) covers requirements with no ceiling concern.

High-dose antioxidants taken around exercise — Caution

This one is specific to active people and easy to miss, because the supplements involved are otherwise considered harmless. The transient burst of reactive oxygen species your muscles produce during a hard session is the signal that drives the adaptation — mitochondrial biogenesis and improved insulin sensitivity. Blunt that signal with high-dose antioxidants and you blunt the adaptation. In a controlled trial, 1,000 mg vitamin C plus 400 IU vitamin E daily abolished the exercise-induced improvement in insulin sensitivity and prevented the upregulation of the body's own antioxidant enzymes.[15] That trial ran four weeks in 39 healthy young men, and the insulin-sensitivity result has not replicated: a later randomised trial put 21 men through 12 weeks of supervised endurance training on half the vitamin C dose alongside the same 400 IU of vitamin E, and insulin-stimulated glucose uptake rose about as much with the vitamins (17.2%) as without them (18.9%).[16] Blood levels of both vitamins rose in the supplemented group, so that null is not a compliance failure.

Practical: don't take high-dose vitamin C or E (or other concentrated antioxidants) in the hours around training, especially if the goal is fitness or metabolic adaptation. Eating antioxidant-rich food is fine — the issue is the concentrated peri-workout pill.

High-dose calcium supplements — Caution

Supplements above ~1000 mg/day in non-deficient adults carry a probable cardiovascular risk signal — probable rather than established, because the trial evidence points both ways. An updated systematic review of 4 randomised trials at low risk of bias and 27 observational studies concluded that calcium from food and supplements combined, up to the tolerable upper limit of 2,000–2,500 mg/day, is not associated with cardiovascular disease risk in generally healthy adults — a ceiling on total intake, not a supplement allowance — though cardiovascular events were a secondary endpoint in every trial it pooled, it was funded by the National Osteoporosis Foundation, and it drew an accompanying editorial and five published letters.[17] The evidence on the other side:

  • A 2010 meta-analysis found that calcium supplements taken without vitamin D were associated with about 31% more heart attacks: across five trials contributing individual participant data on 8,151 adults followed for a median 3.6 years, 143 people on calcium had a heart attack against 111 on placebo (hazard ratio 1.31, 1.02 to 1.67). Stroke and death were not significantly raised.[18] The restriction matters, because the trials pooled gave calcium on its own while most products on the shelf pair calcium with vitamin D.
  • In the Women's Health Initiative, 36,282 postmenopausal women aged 50–79 took either 500 mg calcium carbonate plus 200 IU vitamin D3 twice daily or a matching placebo for an average of seven years, and the supplemented group reported about 17% more urinary tract stones — 449 of 18,176 women against 381 of 18,106 on placebo, roughly 2.5% versus 2.1% over seven years.[19] The stones were self-reported as clinically diagnosed rather than independently adjudicated, and taking a calcium supplement with a meal is the practical mitigation; see Calcium.

Dietary calcium does not show this pattern.

Practical:

  • Prioritize dietary calcium (dairy, leafy greens, sardines, calcium-set tofu)
  • Supplement only the gap to reach 1000–1200 mg/day total
  • Avoid bolus doses >500 mg
  • Take calcium carbonate with meals (with food it absorbs as well as citrate); use calcium citrate (acid-independent absorption) if you take acid-reducers or dose between meals
  • Postmenopausal women with osteoporosis risk are an exception — work with a clinician

The mechanism behind the divide is the "bolus effect": a single large pill spikes serum calcium for hours, while food never does. See Calcium for the full dietary-versus-supplement story, the synergistic triad, and how to choose a form.

Iron in adult men and postmenopausal women without deficiency — Caution

  • Iron drives Fenton-reaction oxidative stress
  • Elevated ferritin (the blood marker of stored iron) associates with type-2 diabetes: in a meta-analysis of prospective cohorts, people in the highest ferritin category ran about 70% more risk of type 2 diabetes than those in the lowest, and about 63% more after adjusting for inflammation markers, which raise ferritin independently of iron stores.[20] Higher stored iron also associates with coronary artery disease risk.
  • Hereditary hemochromatosis — an inherited iron-overload disorder caused by a mutation in the iron-regulating HFE gene — is carried in double dose by about 1 in 150 people of European descent (Pilling's exact figure is 1 in 156): 2,890 of 451,243 UK Biobank volunteers aged 40–70 were homozygous for the main variant, HFE p.C282Y.[21] Carrying the genotype is not the same as having the disease — hemochromatosis had been diagnosed in 21.7% of homozygous men and 9.8% of homozygous women by the end of about seven years of follow-up — but it is not benign either: over that period 15.7% of homozygous men developed at least one associated condition against 5.0% of men without the variant, and homozygous men were about four times as likely to have liver disease.

Don't supplement iron without documented deficiency. Most adult-male and senior multivitamins are intentionally iron-free.

Premenopausal women with low ferritin are the exception; supplement to a ferritin above 30 µg/L (or 40–70 µg/L for hair regrowth and athletic performance per some clinicians). Take it on alternate days rather than daily: in iron-depleted women given 60 mg of iron, alternate-day dosing raised the share of each dose actually absorbed from 16.3% to 21.8% over the course of treatment, because a dose taken the day before leaves the hormone hepcidin elevated and blunts the next one.[22] That trial used ferrous sulfate, and it measured absorption over days rather than haemoglobin recovery — its authors asked for confirmation in iron-deficient anaemic patients. Ferrous bisglycinate is a reasonable choice if plain sulfate upsets your stomach, but it is not the form that was tested here.

Chronic high-dose zinc (>40 mg/day for years) — Caution

  • Chronic intake above 40 mg/day depletes copper
  • US tolerable upper limit (UL): 40 mg/day[23]; the European Food Safety Authority (EFSA) sets a stricter UL of 25 mg/day[24]
  • Critical for men: In the Health Professionals Follow-up Study — 46,974 men followed for 14 years — those taking more than 100 mg/day of supplemental zinc had more than double the rate of advanced prostate cancer compared with men taking none (relative risk 2.29, 1.06 to 4.95 — it only just clears no effect, and rests on 10 advanced cancers). Taking supplemental zinc for 10 years or more, at any dose, carried a separate signal of similar size. Total prostate cancer was not raised by either exposure, and the authors could not rule out confounding by supplemental calcium, or men already self-medicating prostate symptoms with zinc.[25]

Practical: 8–15 mg/day matches the recommended dietary allowance (RDA). For a cold you already have, zinc lozenges may shorten it by roughly two days, but Cochrane grades that benefit low-certainty and is more confident of the downside: zinc used to treat a cold probably increases non-serious side effects.[26] The trials that showed anything used 75 mg/day or more, for 1–2 weeks. Avoid daily high-dose zinc long-term.

Selenium ≥200 µg/day in already-replete people — Caution

  • SELECT trial: in 35,533 men, 200 µg/day of selenium did not protect the prostate — prostate cancer was marginally, and not significantly, more common than on placebo (hazard ratio 1.04, 99% confidence interval 0.87–1.24).[27] A later analysis of the same trial suggests it may have harmed men who were already selenium-replete.[28]
  • People in low-selenium-soil regions (much of inland Europe) may benefit from modest supplementation (50–100 µg/day selenomethionine) — though selenium is not free of risk: in a randomised, placebo-controlled trial among 1,202 adults with a history of skin cancer in a low-selenium region of the eastern United States, 200 µg/day, double the top of that range, produced more type 2 diabetes than placebo over an average of 7.7 years, 12.6 versus 8.4 cases per 1,000 person-years, and the excess concentrated in the third of participants who started with the highest blood selenium[29]. Stay below the European Food Safety Authority upper limit of 255 µg/day, a ceiling the agency derived from hair loss as the earliest observable effect of selenium excess rather than from any cancer signal.[30]

Chronic high-dose B6 — Caution

  • Chronic B6 (pyridoxine) excess causes a sensory peripheral neuropathy — tingling and numbness in the hands and feet. Symptoms usually stop if the supplement is discontinued as soon as they appear[31], but recovery is not guaranteed: in Australia's 2025 review of vitamin B6, 103 people reported severe and sometimes permanent effects, including nerve damage and muscle weakness.[32]
  • The regulators disagree sharply on where the ceiling sits, and the conservative number is the one to use: EFSA sets the tolerable upper limit at 12 mg/day, counting food and supplements together rather than the supplement alone[33], and Australia, whose own upper level is 50 mg/day, has required a neuropathy warning on products above 10 mg/day since 2022 and in November 2025 decided that, from 1 June 2027, products above 50 mg per daily dose will need a pharmacist's advice to buy (above 200 mg already needs a prescription), its regulator concluding that neuropathy can occur at 50 mg/day or even lower.[34] The US UL is 100 mg/day — an outlier, and not a target.[35]
  • Many US "stress vitamin" or "energy" formulations contain 50–100+ mg — several times the European ceiling. Check the label. See B vitamins.

Practical: Avoid B6 megadoses unless treating a specific deficiency under medical supervision.

Over-the-counter sleep aids (diphenhydramine, doxylamine) — Caution

  • Anticholinergic burden (these drugs block acetylcholine, a key nerve-signaling chemical)
  • Daytime sedation, cognitive effects
  • Dementia. In a prospective cohort of 3,434 adults aged 65 and over, followed about seven years, the heaviest cumulative anticholinergic exposure — roughly a standard daily dose sustained for more than three years — carried about 54% higher risk of dementia than no use (hazard ratio 1.54, 1.21 to 1.96), with a clear dose-response across exposure levels; lighter use showed no significant increase.[36]
  • This is observational, and the causal reading has softened: a study roughly seventeen times larger, in the same journal, broke the signal down by drug class,[37] and a pooled analysis across 26 studies and more than 600,000 people puts the class-level association well below 54%.[38] Early dementia may itself drive both the sleep trouble and the drug use. The direction has held up; the size of it, for antihistamines specifically, is less certain than a single 54% figure suggests.

Avoid chronic use at any adult age. The 2023 update of the American Geriatrics Society Beers Criteria — the standard list of medications usually best avoided in older people — lists first-generation antihistamines, diphenhydramine and doxylamine among them, as drugs to avoid from age 65: a strong recommendation on moderate-quality evidence.[39] Below 65 there is no age threshold in the evidence; the reason to skip these earlier is that they work poorly, not that a cut-off exists.

"Cleanses," detox protocols, and most herbal "liver support" — No benefit; some are liver-toxic

  • No evidence of benefit for normal liver/kidney function
  • Some herbal preparations can damage the liver (hepatotoxicity). The US National Institutes of Health LiverTox database rates kava a well-known cause of clinically apparent liver injury — rare and idiosyncratic, but occasionally fatal, with 50 to 100 published cases and onset typically two to twenty-four weeks after starting; it is banned or restricted in several countries.[40] Oral comfrey is rated a probable cause: its pyrrolizidine alkaloids are converted in the liver into metabolites that damage the small hepatic blood vessels (sinusoidal obstruction syndrome), which is why oral forms are banned or restricted in most countries.[41] Of more than 7,000 Chinese herbal medicines only a few have been linked to clinically apparent liver injury — most often Jin Bu Huan, Shou Wu Pian, Ma Huang (ephedra), Ba Jiao Lian and Chi R Yun — but these products are frequently poorly labelled and can be contaminated with metals, bacteria or conventional drugs.[42]
  • The liver and kidneys handle their own "detox"

If you have liver concerns, get liver function tests (LFTs); don't self-treat with supplements.

Two of the best-documented supplement liver injuries are covered in detail elsewhere on this site: concentrated green-tea-extract catechin pills, where brewed tea itself carries no such signal (see Tea), and high-absorption curcumin formulations using piperine or lipid carriers, where culinary turmeric does not (see Curcumin).

"Adrenal support" supplements — No evidence of benefit

  • "Adrenal fatigue" is not a recognized medical diagnosis (per the Endocrine Society). A systematic review that screened 3,470 papers and analysed the 58 that qualified found no substantiation for it, and noted that no endocrinology society recognises the condition.[43]
  • Adrenal glandulars from animal sources, mega-dose vitamin C/B5 protocols, etc. — no evidence of benefit
  • Real adrenal insufficiency requires medical evaluation and proper hormone replacement

High-dose niacin (≥1 g/day) — Caution (no benefit; harm signal)

  • Older lipid trials showed flushing-dose niacin lowered LDL ("bad") cholesterol and raised HDL ("good") cholesterol, but
  • The AIM-HIGH[44] and HPS2-THRIVE[45] RCTs showed no benefit on cardiovascular events when added to statins, and HPS2-THRIVE put numbers on the harm. Across its 25,673 participants over a median 3.9 years, serious infections were 1.4 percentage points more common on niacin than on placebo, new diabetes diagnoses 1.3 points, serious gastrointestinal events 1.0 point and serious skin events 0.3 points; serious musculoskeletal events — the category the "muscle damage" signal falls into — and serious bleeding were each up 0.7 points.[46] A larger excess, 3.7 percentage points of serious loss of blood-sugar control, belongs to the 8,299 participants who already had diabetes when they enrolled rather than to the trial as a whole.[47] In that group serious disturbance of diabetes control ran 56% higher than on placebo, an excess of 12 per 1,000 person-years.[48] The niacin was given combined with laropiprant, an anti-flushing drug not sold as a supplement, so not every excess is provably niacin's alone.
  • Routine high-dose niacin for cardiovascular protection is not recommended. The 2025 American Association of Clinical Endocrinology guideline on managing dyslipidemia (abnormal blood lipids) strongly recommends against niacin for adults with high triglycerides who have, or are at increased risk of, atherosclerotic cardiovascular disease — its strongest form of words, where the guideline's other drug recommendations only "suggest".[49] Niacin remains a prescription lipid drug for specific indications and the treatment for pellagra; what is not recommended is routine use for heart protection.

NR (nicotinamide riboside) and NMN (nicotinamide mononucleotide) are different molecules; see Geroprotectors.

Red yeast rice — Caution (an unlabelled statin dose, and a contamination record)

Red yeast rice is rice fermented with the mould Monascus, sold over the counter for lowering cholesterol. It is not a gentle botanical alternative to a statin. Its active constituent, monacolin K, is in its lactone form the same molecule as lovastatin, a prescription statin, and the amounts delivered by supplements can reach the therapeutic dose range of that drug.[50] Since June 2022 EU law has capped how much a product sold in Europe may deliver: a daily portion must provide less than 3 mg of monacolins, the label must state the content per portion, and it must warn against use in pregnancy, under-18s, over-70s, and alongside cholesterol-lowering medication.[51] Products bought outside the EU carry no such limit.

  • No safe intake has been established. European regulators could not identify any daily intake of monacolins from red yeast rice that does not raise safety concerns. Severe reactions — muscle breakdown (rhabdomyolysis) and liver injury — have been reported at intakes as low as 3 mg/day of monacolin K, and a 2025 re-examination of newer post-marketing data reached the same conclusion.[52]
  • A fermented supplement carries a manufacturing risk you cannot check at home. After a 2024 recall of Beni-koji (red yeast rice) tablets in Japan, a two-phase nationwide survey of nephrologists collected 192 affected patients. Of those, 94% had reduced kidney function — an estimated glomerular filtration rate (eGFR) below 60 mL/min/1.73 m² — with the laboratory picture of Fanconi syndrome, in which the kidney tubules leak potassium, phosphate, uric acid and glucose. Biopsies showed tubulointerstitial nephritis in half the cases and tubular necrosis in about a third.[53]
  • The damage largely did not reverse. After the product was stopped and patients were treated, the Fanconi-syndrome markers improved, but 87% still had an eGFR below 60 mL/min/1.73 m² at last follow-up. The authors are explicit that further research is needed to establish causality.[54]
  • The culprit was a factory contaminant, not the mould. Citrinin, the nephrotoxin classically feared in red yeast rice, was not produced by the implicated strains; investigators traced the injury to puberulic acid made by Penicillium adametzioides growing in the production plant.[55]

Practical: If you need to lower LDL ("bad") cholesterol, take a prescribed statin at a known dose with monitoring rather than a fermented supplement whose statin content is neither labelled nor regulated. See Lipid management.

"Memory" formulas, "immune boosters," "fat burners" — No benefit; contamination risk

These categories are dominated by unproven mixtures, often with banned or unregulated ingredients. Quality control is poor, and the problem is documented rather than hypothetical: between 2007 and 2021 the US Food and Drug Administration (FDA) identified 1,068 unique supplement products adulterated with active pharmaceutical ingredients — undeclared prescription drugs. Sexual-enhancement and weight-loss products dominate the list, a single sexual-enhancement product has been found carrying as many as five such drugs, and the commonest weight-loss adulterant is sibutramine, a drug withdrawn from the market for cardiovascular harm.[56] That is a count of the products regulators caught, not a survey of the shelf, so it cannot tell you what share of supplements are spiked — and the FDA's categories do not include "memory" or "immune" formulas.

If you want immune support: vitamin D, zinc lozenges (acutely), and adequate sleep beat any "immune-boosting" supplement.

A summary table

SupplementDose threshold of concernEvidence type
Vitamin E≥400 IU/dayMultiple large RCTs
Beta-caroteneAny in smokersRCTs (ATBC, CARET)
Vitamin A (preformed)>3000 µg RAE/dayCohort + RCT
Vitamin C>2000 mg/dayMechanistic + UL (kidney stones)
Antioxidants around exerciseHigh-dose C/E peri-workoutControlled trial (blunted adaptation; not replicated)
Calcium supplements>1000 mg/day in repleteMeta-analysis (Bolland)
IronAny in repleted men/postmenopausal womenMechanistic + cohort
Zinc>40 mg/day chronicRCT (cohort for prostate cancer)
Selenium≥200 µg/day in repleteRCT (SELECT)
Vitamin B6>12 mg/day chronic (EFSA upper limit)Case-control + cohort
Diphenhydramine (sleep)Any chronicCohort (dementia association)
High-dose niacin≥1 g/dayRCT (AIM-HIGH, HPS2-THRIVE)
Red yeast riceAny (monacolin content unlabelled; no safe intake identified)Regulator safety opinion + national case series

Further reading

  • Bjelakovic G et al. Antioxidant supplements for prevention of mortality in healthy participants and patients with various diseases. Cochrane Database Syst Rev 2012.[57]
  • Ristow M et al. Antioxidants prevent health-promoting effects of physical exercise in humans. PNAS 2009.[58]
  • Miller ER 3rd et al. Meta-analysis: high-dosage vitamin E supplementation may increase all-cause mortality. Ann Intern Med 2005.[59]
  • Klein EA et al. Vitamin E and prostate cancer (SELECT). JAMA 2011.[60]
  • Bolland MJ et al. Effect of calcium supplements on risk of myocardial infarction and cardiovascular events: meta-analysis. BMJ 2010.[61]
  • Leitzmann MF et al. Zinc supplement use and risk of prostate cancer. J Natl Cancer Inst 2003.[62]
  • Gray SL et al. Anticholinergics and incident dementia. JAMA Intern Med 2015.[63]
  • ATBC: Beta-carotene and vitamin E in male smokers. NEJM 1994.[64]
  • CARET: Beta-carotene and retinol in smokers and asbestos workers. NEJM 1996.[65]
  • HPS2-THRIVE Collaborative Group. Effects of extended-release niacin with laropiprant in high-risk patients. NEJM 2014.[66]

— § —