On Vitamins and Supplements: When Less Is More

Most patients I see take one or more vitamin or nutritional supplements. They take them for various reasons, but primarily in the hope that they will preserve health and prevent chronic disease — heart disease, cancer, high blood pressure, diabetes. More often than not, these supplements make a significant dent in their finances.

Are they worth the money? Do they deliver on any of these promises? Is there reason for concern?

What I wrote the first time, and why I am revisiting it

I first wrote this article after the Archives of Internal Medicine published the Iowa Women’s Health Study analysis: 38,772 older women followed from 1986 to 2008, reporting their supplement use along the way.¹ The supplements examined included vitamin B6, folic acid, iron, magnesium, zinc, copper and calcium.

The results were striking. Not only was there no mortality benefit — for most of the supplements studied there was an increase in mortality. Multivitamin use was associated with roughly a 2.4% absolute increase in risk of death over the follow-up period, and iron looked worse the more of it women took. Calcium was the sole exception, associated with lower mortality.

At the time, that was one large observational study, and I said so. Observational studies show associations, not causes; people who take supplements differ from people who don’t in a hundred ways researchers cannot fully measure. The honest position in 2015 was that the promise had not been demonstrated and the evidence was pointing the wrong way.

Since then, the large randomised trials have reported. Randomised trials settle the question observational studies can only raise, because the coin toss removes the confounding. We now have a much clearer answer, and it is close to the one the Iowa data hinted at.

What eleven more years of evidence showed

In 2022 the US Preventive Services Task Force — the body that grades preventive interventions for American medicine — completed its review of 84 studies including 52 published since its previous look.^2,3 Its conclusions were plain:

The individual trials tell the same story in more detail. VITAL randomised 25,871 American adults to 2,000 IU of vitamin D and to one gram of marine omega-3 daily, and followed them for a median of 5.3 years. Neither reduced the primary cancer or cardiovascular endpoints.^4,5 The fracture arm found that vitamin D did not reduce fractures either — not total, not hip, not wrist — in adults who were not vitamin D deficient to begin with.⁶ Australia’s D-Health trial gave more than 21,000 older adults a monthly high dose for five years and found no reduction in all-cause mortality.⁷

And in 2024, the National Cancer Institute pooled three American cohorts — 390,124 generally healthy adults followed for up to 27 years — specifically designed to overcome the biases that plague this literature. Daily multivitamin use was associated with no mortality benefit whatsoever. If anything the hazard ratio sat fractionally above 1.0.⁸

I want to be fair to the other side, because there is one genuine exception worth knowing about. The COSMOS trial, which randomised more than 21,000 older adults, found that a daily multivitamin modestly improved memory and slowed cognitive ageing across three separate substudies.^9,10 The effect was small — on the order of slowing cognitive ageing by a couple of years on a memory test — and it has not yet been replicated by an independent group. It is the most promising finding in favour of multivitamins in a generation, and it is still a long way from a reason to expect a longer or healthier life from the bottle.

Where supplements did not merely fail — they harmed

This is the part that surprises people most, so it deserves specifics. These are not theoretical risks; they are the reasons trials were stopped early.

• Beta-carotene. Two large trials in smokers — ATBC in Finland and CARET in the United States — found more lung cancer and more deaths in the supplemented groups. CARET was halted 21 months early.^11,12 An antioxidant that is protective in carrots was harmful in a capsule.
• Vitamin E. The SELECT trial randomised 35,533 healthy men. Men taking 400 IU of vitamin E daily had a 17% higher rate of prostate cancer — 76 cases per 1,000 men versus 65 on placebo over seven years.¹³
• Fish oil. A 2024 analysis of 415,737 UK Biobank participants found that in people without existing heart disease, regular fish oil use was associated with a 13% higher risk of atrial fibrillation and a 5% higher risk of stroke. In people who already had cardiovascular disease, the same supplement was associated with benefit.¹⁴ The same pill, opposite directions, depending on who swallows it.

A nutrient that is beneficial in food, at food doses, in the company of everything else the food contained, is not the same thing as that nutrient isolated, concentrated, and taken alone for years.

You do not actually know what is in the bottle

Under the Dietary Supplement Health and Education Act of 1994, supplements in the United States are regulated more like food than like medicine. A manufacturer does not have to prove a product works, or even that it is safe, before selling it. The FDA’s power is largely to act after harm is reported.

What that means in practice is measurable. When researchers bought 25 melatonin gummy products and assayed them, 22 were inaccurately labelled. Actual melatonin content ranged from 74% to 347% of what the label claimed. One product contained no melatonin at all — but did contain 31 mg of CBD.¹⁵ These are products routinely given to children.

The consequences reach the hospital. An analysis of national surveillance data estimated about 23,000 emergency department visits a year in the United States attributable to supplement adverse events, most commonly cardiac symptoms in young adults taking weight-loss and energy products.¹⁶ And a 2024 analysis estimated that roughly 15 million American adults take at least one of six botanicals with documented reports of liver injury — turmeric, green tea extract, ashwagandha, garcinia cambogia, red yeast rice and black cohosh.¹⁷

I should be even-handed here: on a per-user basis those liver injuries are rare, and industry critics of that paper made the point fairly. But “rare” is a difficult trade when the expected benefit is zero. A small risk in exchange for a real benefit is medicine. A small risk in exchange for nothing is just a small risk.

The harm I actually see in clinic

Everything above concerns direct harm, and direct harm is the smaller problem. I do not think most supplements are killing anybody. What I think happens is subtler and far more common.

By investing their hopes for prevention — for more years, and better ones — in a daily handful of capsules, many people quietly stop placing the necessary emphasis on what actually works: how they eat, how they move, how they sleep. The supplement becomes a substitute for the thing it was meant to supplement.

Occasionally I meet patients who would rather spend their money on supplements than on good food. I find that genuinely troubling, particularly when I cannot restore their confidence that a decent diet is sufficient. It is the predictable result of aggressive marketing meeting an absence of meaningful regulation.

Let us be honest about why. It is far easier to swallow a few capsules each morning than to plan, shop, cook and sit down to a good meal. You do not break a sweat opening a bottle of nutraceuticals, the way you do walking for forty-five minutes.

When a supplement is exactly the right answer

None of this makes me anti-supplement. It makes me anti-guessing. There is a real and well-defined list of situations in which supplementation is correct, evidence-based medicine, and I prescribe from it regularly:

• Folic acid in pregnancy. 400 to 800 µg daily for anyone who could become pregnant, started before conception. This prevents neural tube defects and is one of the best-supported preventive measures in all of medicine.¹⁸
• Vitamin B12 for vegans and strict vegetarians, for people on long-term metformin or acid-suppressing medication, and for many adults over 60 whose absorption declines.
• Vitamin D for people who are genuinely deficient, and at the recommended daily allowance for children, pregnant women, adults over 75, and adults with high-risk prediabetes. Note what the 2024 Endocrine Society guideline also says: healthy adults under 75 are unlikely to benefit from more than the RDA, and do not need their vitamin D level tested at all.¹⁹
• Iron for documented iron deficiency — and only then. Iron is the supplement that looked worst in the Iowa data, and taking it without a deficiency is not a neutral act.
• The AREDS2 formulation for people with established intermediate or advanced age-related macular degeneration, where it slows progression.²⁰
• Calcium and vitamin D where dietary calcium genuinely falls short, particularly after bariatric surgery or with osteoporosis under treatment.
• Anything your own physician has prescribed for a diagnosed condition.

Notice the pattern. Every item on that list treats a specific, identified shortfall or condition. Not one of them is a healthy person taking something in the general hope of preventing something.

The question to ask before you buy anything

Here is what strikes me about the way supplements are actually bought. Almost nobody establishes that they are short of a nutrient before buying it. The sequence runs backwards: an advertisement, or a friend, or an article, suggests magnesium — and the bottle arrives before anyone has asked whether this particular person, eating this particular diet, is short of magnesium at all.

That question is answerable, and it does not require a blood test. Most of what you need to know is in what you already ate. Roughly half of Americans take in less magnesium than the estimated average requirement; nearly all fall short of the potassium recommendation from food alone.²¹ You may well be one of them — or you may not be, and the honest way to find out is to look.

This is precisely why I built VitalFill. Not to sell anyone a supplement, but to answer the question that should come first: am I actually short of anything? You log what you eat; it tracks calories and more than forty nutrients against targets set for your own age, sex, and whether you are pregnant or breastfeeding — and it shows you which ones are filled and which fall short.

If you have been short of vitamin B12 for a year, that is worth knowing and worth acting on. If your diet has been covering everything comfortably, that is worth knowing too, and it will save you a good deal of money.

Find out what you are actually short of

Log food by voice, camera, or barcode. VitalFill tracks 40+ nutrients against targets personalised to you — so you supplement what you are missing, and stop paying for what you are not.

 Coming to theApp Store ▶ Coming toGoogle Play

VitalFill is still in development. If you would like to know when it is ready, leave your name below and I will tell you — nothing else. I will never sell or share your email, and I use it only to write to you occasionally about this and similar tools I develop here.

In the meantime, you can try the nutrient tracker on this site, which runs the same calculations in your browser for a single day.

Conclusion

If little labour, little are our gains:
Man’s fortunes are according to his pains.
— Robert Herrick, Hesperides 752

Absorbing a bit too much of some nutrients, like the B vitamins, just results in “expensive urine”, because the excess is excreted.
— Kathi Kemper, MD, paediatrician

Eleven years and several hundred thousand randomised participants later, I would put it the same way I did then, with one addition.

If you want to reap the benefits of a healthy, fulfilling life, you have to be willing to do what it takes. There are no shortcuts. Not even supplements.

But if you are going to take something, take it for a reason you can name — a deficiency you have actually demonstrated, or a condition your doctor is actually treating. That is the difference between medicine and hope in a bottle.