Evidence Against

The assumption: an antioxidant from carrots, taken for lung and general health, obviously harmless.

What the trials found: two large randomized, placebo-controlled trials tested isolated high-dose beta-carotene in the people most likely to take it for their lungs, and it made things worse. ATBC randomized 29,133 male smokers to 20mg per day and recorded higher lung cancer incidence. CARET gave 18,314 smokers and asbestos-exposed adults 30mg per day plus retinol, and was stopped nearly two years early in 1996 after the active group showed a 28 percent increase in lung cancer and a 17 percent increase in deaths.

The textbook case of a "healthy" isolated antioxidant doing net harm at supplemental doses. Carotenoids from whole vegetables are not the same as a high-dose capsule.

The assumption: a heart-protective, anti-aging antioxidant worth taking daily.

What the trial found: SELECT randomized 35,533 men to vitamin E 400 IU per day. It was designed to show cancer prevention. Instead, vitamin E was associated with a statistically significant Unlikely to be explained by chance alone at the threshold the study set. It says nothing about whether the difference is large enough to matter. increase in prostate cancer, roughly 76 cases per 1,000 men versus 65 on placebo An inactive substitute made to look like the real treatment, so participants cannot tell which one they received.. An earlier meta-analysis A study that pools the numbers from several trials into one combined estimate. had already linked high-dose vitamin E to higher all-cause mortality.

The dose and isolated form matter. This is not a statement about the vitamin E you get from food.

The assumption: menopausal hormone therapy is broadly protective for the heart and bones.

What the trial found: the Women's Health Initiative randomized 16,608 healthy postmenopausal women and was stopped early. Estrogen plus progestin raised invasive breast cancer (HR 1.26), stroke (1.41), pulmonary embolism (2.13), and coronary events, with the global balance of risks outweighing the benefits seen in fracture and colorectal cancer.

Important nuance: timing and formulation matter, and for symptom relief in younger, recently menopausal women the risk-benefit picture is different. What failed was the blanket claim that combined HRT protects the heart.

The assumption: niacin raises HDL cholesterol, so adding it to a statin should cut heart risk further.

What the trial found: HPS2-THRIVE added extended-release niacin to statin therapy in 25,673 patients and found no reduction in vascular events, plus a significant excess of serious adverse effects: new-onset diabetes, infection, bleeding, and gastrointestinal and skin problems.

A clean example of a plausible mechanism, raising HDL, both failing to help and causing harm in a large trial.

The assumption: folate protects against colorectal cancer, so more is better.

What the trial found: a randomized trial gave 1,021 people with a history of adenomas 1mg per day of folic acid. It did not reduce adenomas, and significantly increased the risk of advanced and multiple adenomas. A separate trial later linked folic acid plus B12 to higher colorectal cancer.

Timing appears to matter: folate may protect before neoplasia exists but promote growth once precancerous cells are present. This concerns high-dose supplementation, not dietary folate.

The assumption: an antioxidant trace mineral that supports metabolism and may prevent cancer.

What the trial found: a randomized trial of 1,202 adults given 200 mcg per day found selenium increased the incidence of type 2 diabetes (HR 1.55), with the highest risk in those who already had high baseline selenium.

For a trace mineral, more is not better, and the risk concentrated in people who were already replete.

What it is: a large group of common drugs that block acetylcholine. Older antihistamines such as diphenhydramine (Benadryl), some sleep aids, overactive-bladder drugs, and tricyclic antidepressants all carry anticholinergic activity, and the effect stacks when several are taken together.

The signal: a 2026 Karolinska Institutet cohort in BMC Medicine followed 508,273 adults aged 45 and older for a median of 14 years and recorded 118,266 cardiovascular events. Higher cumulative anticholinergic burden was associated with more heart failure, arrhythmias, heart attacks, and strokes, in a dose-response pattern that held after adjustment. This sits on top of a well-established prior link to faster cognitive decline and dementia.

Why a watch, not a grade: it is observational. No one can ethically randomize people to years of a suspected harm. The association is large, replicated The same finding reproduced by a separate research group. A result seen once can be chance or bias; a replicated one is harder to explain away., dose-responsive, and mechanistically plausible, which earns a flag rather than silence.

What to do with it: review your total anticholinergic load with your physician. Stopping a prescription abruptly carries its own risks. Lower-burden alternatives are a question for the prescriber.

What it is: the long-standing "moderate drinking is good for your heart" idea, which leaned on observational studies with known abstainer-bias problems.

The signal: the 2018 Global Burden of Disease analysis in The Lancet concluded that the level of consumption minimizing overall health risk is zero, with risk rising for most outcomes including all cancers. Only ischemic heart disease showed a small protective dip, which was offset by cancer risk. Mendelian-randomization studies have not found a cardioprotective effect.

Honest caveat: the "no safe level" framing has drawn methodological criticism, and the absolute risk The actual chance of an event in a group. A fall from 2 percent to 1 percent is a 1-point absolute reduction, even though it is a 50 percent relative one. at light intake is small. This is population-level burden data, not an individual prescription. The defensible statement is that the old "a daily drink is heart-healthy" claim has not held up.

The evidence: ASPREE randomized 19,114 healthy adults aged 70 and older with no cardiovascular disease to daily low-dose aspirin or placebo. Aspirin produced no cardiovascular benefit, more major bleeding, and higher all-cause mortality, driven mainly by cancer death.

The context: this is about primary prevention in healthy older people. Aspirin remains valuable in secondary prevention, after a heart attack or stroke. So the message is "if you take a daily aspirin just in case, ask whether you still should," not "stop your aspirin."

The evidence: PPIs are necessary and appropriate for many conditions, but they are heavily overprescribed and often continued indefinitely without review. Well-established long-term concerns include vitamin B12 and magnesium depletion A fall in a nutrient's level caused by a medication, for example through reduced absorption or increased loss., fracture risk, kidney issues, and enteric infection. Dementia and mortality signals appear in some cohorts (a UK Biobank analysis found higher dementia risk) but not others (an analysis within ASPREE found none), so those remain unsettled.

The context: a classic deprescribing target. If you have been on a PPI for years without a current indication, that is a conversation worth having, including a tapering plan, since stopping abruptly causes rebound acid.

The evidence: for older adults, these sit on the Beers list of potentially inappropriate medications. The most robust harm is falls and fractures. Simple cohorts have linked them to higher mortality (about four excess deaths per 100 people over roughly eight years) and to dementia, but better-controlled studies weaken the mortality and dementia signals, pointing toward confounding by indication.

The context: a deprescribing target, especially when used nightly for sleep. Stopping abruptly carries its own risks. Benzodiazepine withdrawal can be dangerous, so any reduction should be a supervised taper.