Multivitamins

Introduction

Micronutrient deficiencies are widespread. Up to 70% of Americans don’t get enough vitamin D, nearly 50% fall short of their magnesium needs, and 35% have an insufficient intake of calcium.

These deficiencies and insufficiencies ravage our health. Inadequate vitamin D increases mortality risk. A magnesium deficiency can accelerate biological aging. Folate deficiency by itself can produce a level of DNA damage that is equivalent to or in excess of an acute exposure to radiation (although this occurs at levels uncommonly seen in human populations).

Addressing these deficiencies seems to hold a remarkable benefit. For example, correcting vitamin D and magnesium deficiencies can reverse epigenetic aging and protect against DNA damage.

Diet should be the primary source of most of our nutrients, but it can be hard to obtain optimal nutrition through whole-food sources.

In light of this, the benefit of a multivitamin seems clear: the downside is low when you consider the overall benefit of potentially shoring up highly individual minor nutritional deficiencies. And the notion that multivitamins have some benefit is shared by many. Around one third of adults report that they take a multivitamin.

A health investment or expensive urine?

A multivitamin is best understood as nutritional insurance—not a substitute for a varied diet and not a proven longevity drug. The most informative modern test of that idea is COSMOS (the COcoa Supplement and Multivitamin Outcomes Study), a randomized, double-blind, placebo-controlled factorial trial of 21,442 U.S. adults: women age 65 or older and men age 60 or older. Participants received a standard daily multivitamin-mineral supplement (Centrum Silver) or placebo, with a separate cocoa-extract assignment.^[1]

What COSMOS found

The strongest positive signal was cognitive. Across three COSMOS cognitive substudies using different testing formats, a pooled analysis found small average advantages for global cognition and episodic memory over two to three years—about 0.07 and 0.06 standard deviations, respectively.^[2][3][4][5] Investigators translated the global-cognition difference into roughly two years of age-related cognitive change. That is a model-based comparison of test scores, not evidence that participants’ brains literally became two years younger, and the related substudies should not be treated as three independent replications.

COSMOS also showed that the supplement changed nutritional biomarkers. Folate, vitamin B12, vitamin D, carotenoids, and alpha-tocopherol increased, demonstrating that the intervention had measurable biological effects.^[6] In a separate analysis, two of five DNA-methylation clocks shifted modestly—roughly 0.2 to 0.4 years over two years—while three clocks did not change significantly.^[7]

A newer metabolomics analysis of 399 participants found nominal increases in DHA and total omega-3 levels and a smaller rise in creatinine. It also reported directional changes in several metabolomic aging clocks and disease-risk scores.^[8] But the analysis tested roughly 200 outcomes, and none of the reported metabolomic-clock or disease-risk-score findings remained statistically significant after correction for multiple comparisons. These results are hypothesis-generating, not proof that a multivitamin slows aging or prevents disease.

Multivitamins and mortality

When it comes to major diseases and mortality, the data aren’t as clear or convincing.

• Multivitamins don’t seem to reduce the risk of cardiovascular disease, cancer, or all-cause mortality.
• Multivitamins appear to have no effect on total cancer incidence or several types of cancers including breast, colorectal, prostate, or skin (melanoma), although the risk of lung cancer was 38% lower for participants taking a multivitamin.

These conclusions seem to hinge on a lack of evidence that prevents a balance of the benefits and harms from being determined.

Maybe multivitamins don’t need an overwhelming standard of evidence for a large benefit because the overall evidence that each of us may have individual minor nutritional deficiencies is strong. By taking a multivitamin, we hedge our bets that we can reduce the impact of suboptimal nutrition.

But what if there was even a slim chance multivitamins were actually harmful?

One study published in 2024 asserted this, finding that:

• People who reported taking a multivitamin every day had a 4% increase in all-cause mortality risk and a 6% increase in heart disease mortality risk compared to nonusers.
• Taking a multivitamin less than daily was associated with a 9% increase in all-cause mortality risk compared to nonusers at a 12 year follow-up.
• Taking a multivitamin was not associated with a greater or lower risk for cancer or cerebrovascular disease, nor was it associated with mortality from all-causes or specific causes after 15 years or more of follow-up.

Breaking down the study

Published in JAMA and titled “Multivitamin Use and Mortality Risk in 3 Prospective US Cohorts”, this new study attempted to more rigorously assess the link between multivitamin use and mortality by addressing some of the biases that have been present in previous observational studies, namely, the healthy user bias and the sick user effect.

To control for the sick user effect (the possibility that people who get sick will then start using a multivitamin, making them less healthy than nonusers), any participant who had cardiovascular disease, cancer, diabetes, or kidney disease at baseline was excluded.

In an attempt to control for the healthy user bias (the possibility that people who take multivitamins may have better lifestyle habits than nonusers), the statistical analyses were adjusted for confounding factors like age, sex, education, alcohol and tobacco use, BMI, physical activity, diet quality, family history of cancer, and the use of other supplements.

A total of 390,124 participants were pooled from 3 large prospective cohort studies. The participants were asked at least once whether or not they used a multivitamin and were then separated into one of three groups based on their answer:

• Daily multivitamin users (n=163,307)
• Nondaily multivitamin users (n=47,151)
• Nonusers (n=179,666)

The participants were followed for a median of 23 years, during which deaths from all-causes, heart disease, cancer, and cerebrovascular disease were recorded at two time points: the first 12 years (follow-up number one) and 15 years and beyond (follow-up number two).

Is daily supplementation risky?

During follow-up period number one—people who used multivitamins daily had a 4% greater risk for all-cause mortality than people who never used multivitamins, while people who used multivitamins less than daily had a 9% increase in all-cause mortality. No differences in all-cause mortality were observed during the second follow-up period—which was 15 years or more after the baseline assessment.

Other than heart disease mortality—the rate of which was 6% higher in daily users compared to nonusers during the first 12 years of follow up—there were no significant differences between multivitamin users and nonusers in the risk of dying from heart disease, cancer, or cerebrovascular disease. The same was true for nondaily users vs. nonusers.

In other words, taking a multivitamin wasn’t associated with protection against dying from these diseases, but it wasn’t associated with any harm either.

The researchers also conducted an analysis that included a second assessment in which participants indicated their multivitamin use (on average, 6 years after their baseline assessment). This is important, because it’s a better indicator of multivitamin use over time than a single assessment at baseline—it helps to control for multivitamin use habits that may have changed during the study.

The result? A similar finding to the first analysis: A 4% greater all-cause mortality risk in daily multivitamin users in the first 12 years of follow up and no difference in all-cause or cause-specific mortality at the 15-year and beyond follow up.

One more reason to question these results

The hazard ratios (HRs) for these outcomes were all very close to 1. An HR compares the likelihood of an event (such as heart disease or death) happening in one group compared to another. If the HR is 1, it means the risk is the same in both groups. If the HR is above 1, the risk is higher in the first group; if it's below 1, the risk is lower.

An HR close to 1 means that the risk of the events happening was nearly the same for both multivitamin users and non-users. For example, the hazard ratio of 1.04 for all-cause mortality in daily multivitamin users indicates a very small increase in risk. However, this increase is quite weak, and the confidence interval (1.02–1.07) shows the result is very close to no difference between the groups.

The increased risk is likely due to some as-yet unidentified factors that were different between multivitamin users and nonusers. Or perhaps it was a mere statistical artifact.

Conclusion

A multivitamin can be a reasonable nutritional backstop, but COSMOS does not show that every older adult needs one—or that a multivitamin prevents dementia, cancer, cardiovascular disease, or death. It also cannot replace protein, essential amino acids, essential fatty acids, fiber, or the many bioactive compounds supplied by a varied diet.

The formula and the person both matter. Vitamin K can interact with warfarin, and iron, preformed vitamin A, vitamin B6, iodine, selenium, zinc, and other nutrients can become excessive when products are stacked. Pregnancy, smoking history, kidney or liver disease, diet, medications, and laboratory findings can all change what is appropriate. Iron is best used for a demonstrated need or clinician-directed indication rather than taken automatically.

If you choose a multivitamin, look for an age-appropriate formula near recommended daily amounts and independent finished-product verification. Treat it as one layer of nutritional insurance. Exercise, sleep, diet quality, and appropriate medical care remain far more consequential for long-term health.