Multimineral Supplements and Biological Aging

A Nature Medicine Paper, a Centrum Silver Bottle, and a Very Loaded Word

On March 9, 2026, Nature Medicine published a study with a finding that sounded almost too convenient: taking a daily multivitamin-multimineral supplement — specifically Centrum Silver — slowed the rate of biological aging in older adults.¹ Within hours, health media lit up. "Multivitamins reverse aging," declared dozens of headlines. Supplement stock tickers twitched. Wellness influencers who'd spent years dismissing multivitamins as "expensive urine" suddenly found religion.

I want to be careful here, because the study itself is good. It's a prespecified ancillary analysis of a large, double-blind, randomized controlled trial — the kind of evidence most supplement claims never get anywhere near. The investigators are credible. The journal is credible. The statistics are real. But the gap between "statistically significant epigenetic clock slowing" and "your multivitamin reverses aging" is a canyon, and the headlines drove right off the edge.

Let me walk you through what the study actually found, why it matters, why it might not, and what you should do with this information if you're standing in a pharmacy aisle right now staring at a bottle of Centrum Silver wondering whether it's a fountain of youth or a $12 placebo.

COSMOS: 21,000 People, Two Supplements, One Massive Factorial Design

The study comes from COSMOS — the COcoa Supplement and Multivitamin Outcomes Study — one of the largest supplement trials ever conducted in the United States. COSMOS enrolled 21,442 participants: women aged 65 and older, men aged 60 and older, all without recent cancer, heart attack, or stroke. It was a 2×2 factorial design, meaning participants were randomized to one of four groups: daily Centrum Silver, daily cocoa extract (rich in flavanols), both, or double placebo.²

The primary outcome of COSMOS was cardiovascular disease prevention. On that front, the results were largely negative — neither cocoa nor multivitamins significantly reduced heart attacks, strokes, or cardiovascular death.³ But COSMOS was designed to be a platform trial, with dozens of prespecified ancillary studies examining secondary outcomes. This epigenetic aging analysis was one of them.

For the epigenetic substudy, investigators collected blood samples from 958 participants at baseline and again at two years. From those samples, they measured DNA methylation patterns across the genome — the chemical tags that sit on top of your DNA and change predictably with age, disease, and environmental exposure. These patterns are the raw material for epigenetic clocks.

Five Clocks, Two Generations, and a Fundamental Question About What They Actually Measure

To understand what this study found, you need to understand what an epigenetic clock is — and what it isn't. The concept is elegant: as we age, specific sites on our DNA accumulate methyl groups in predictable patterns. By measuring hundreds of these sites simultaneously, researchers can estimate a person's "biological age" — how old their cells look, regardless of their birthday.⁴

The COSMOS team measured five different clocks, which fall into two generations. The first-generation clocks — Horvath and Hannum — were trained to predict chronological age from DNA methylation. They're accurate at telling you how old someone is, but they don't tell you much about how fast they're aging or how likely they are to get sick.⁵

The second-generation clocks are more clinically interesting. PCPhenoAge was trained to predict a composite of biomarkers associated with disease and mortality. PCGrimAge was trained to predict time to death — it incorporates smoking pack-years, plasma protein surrogates, and other mortality-associated markers.⁶ These clocks don't just estimate age; they estimate risk. And they're the ones that responded to the multivitamin.

Here's the critical nuance: epigenetic clocks are population-level research tools. They've been validated to predict disease and mortality in large cohorts. But they have not been validated for individual-level clinical decision-making. Telling a single patient "your GrimAge went down 0.1 years, so you're aging more slowly" is like telling them their cholesterol dropped 2 points — statistically detectable, practically ambiguous, and possibly meaningless for their personal trajectory.⁷

2.6 Months Per Year. That's the Number. Let's Talk About What It Means.

The headline finding: participants randomized to Centrum Silver showed a reduction in the yearly rate of increase in PCPhenoAge of 0.217 years — roughly 2.6 months — compared to placebo. For PCGrimAge, the reduction was 0.113 years per year, or about 1.4 months.¹ Over the full two-year study, that translates to approximately 4 months of cumulative epigenetic "age deceleration" on the most responsive clock.

First-generation clocks (Horvath and Hannum) showed no significant response. This is important — it suggests the multivitamin isn't making cells "younger" in a chronological sense, but rather modulating the disease- and mortality-associated methylation patterns that second-generation clocks are designed to capture.

There was a subgroup finding worth noting: participants who started the trial with accelerated baseline biological aging — their epigenetic age was already running ahead of their chronological age — showed larger benefits, with a PCGrimAge reduction of approximately 0.233 years per year (about 2.8 months).¹ This makes intuitive sense: if your cells are already showing signs of accelerated aging, correcting micronutrient insufficiencies might have more room to help.

Meanwhile, cocoa extract showed no significant effect on any epigenetic clock. This is a useful negative finding. Cocoa flavanols have been marketed for their anti-inflammatory and vascular benefits, yet they produced zero measurable epigenetic aging signal in the same population.¹ The mechanisms driving epigenetic clock changes are clearly not as simple as "anti-inflammatory = anti-aging."

An accompanying editorial by Daniel Belsky and Calen Ryan — both respected aging researchers — called the findings "a major advance for the supplement field" while emphasizing the uncertainties. They noted that epigenetic clocks may be measuring processes related to but distinct from the biological mechanisms that actually determine healthspan.⁸

This Isn't the First Time COSMOS Made Multivitamins Look Good

This epigenetic aging finding doesn't exist in isolation. COSMOS has produced several ancillary analyses that have painted multivitamin supplementation in a surprisingly favorable light — and it's worth putting them all on the table.

The cognitive findings from COSMOS are notable but face similar interpretive challenges: the effect was measured via telephone-administered tests, not gold-standard neuropsychological evaluation, and no biomarker confirmation was obtained.⁹ Still, when you combine modest cognitive benefits, modest epigenetic clock slowing, and a clean safety profile, the pattern starts to feel like it means something. The question is whether "means something" translates to "helps you live longer and healthier." We don't know yet.

On the other side of the ledger, the U.S. Preventive Services Task Force conducted a comprehensive evidence review in 2022 — examining multivitamin supplementation for the primary prevention of cardiovascular disease and cancer — and found insufficient evidence to recommend routine supplementation for most adults.¹⁰ The Task Force noted a possible modest reduction in cancer incidence but emphasized that the evidence was inconsistent and the effect small.

The Gap Between a Moving Biomarker and a Longer, Healthier Life

I've been reviewing evidence in this newsletter for 30 issues now, and I've developed a reflex when I see impressive biomarker data: I immediately ask, "Does this translate to an outcome a patient would care about?" In this case, the honest answer is: we don't know.

Epigenetic clocks have been validated at the population level to predict mortality, cardiovascular disease, cancer incidence, and cognitive decline.^6,7 But validation is not the same as clinical utility. No physician in any practice uses epigenetic clock measurements to guide treatment decisions. No FDA-approved diagnostic exists. And critically, no intervention study has yet shown that slowing an epigenetic clock produces a measurable improvement in healthspan or lifespan.⁸

The study measured only DNA methylation changes. It did not report changes in disease incidence, mortality, disability, frailty, cognitive function, physical performance, or any outcome that would matter to a patient asking, "Will this help me stay healthy?" That's not a flaw in the study — measuring those outcomes would require far more participants and far more time. But it's a critical interpretive boundary that most media coverage ignored entirely.

There's also the question of who benefits. COSMOS enrolled generally healthy older adults without recent major disease. These are not the malnourished, the chronically ill, or the micronutrient-depleted populations where supplementation logic is strongest. If your diet already provides adequate vitamins and minerals — which most American adults consuming fortified foods achieve for many nutrients — the marginal benefit of a daily multivitamin may be negligible.¹¹

And then there's the funding. Pfizer provided the Centrum Silver tablets and matching placebos. They did not design the trial, conduct it, collect data, analyze results, or write the manuscript — all of which is clearly stated in the publication. But Pfizer now has a Nature Medicine paper to cite in marketing materials for its second-best-selling consumer health product (U.S. revenue: $164.5 million in 2019).¹² The investigators' independence appears genuine, but the commercial implications are enormous.

Finally, a methodological point that deserves more attention than it received: the cocoa extract arm showed no effect on any epigenetic clock. Cocoa flavanols are promoted for anti-inflammatory, antioxidant, and vascular benefits — mechanisms that plausibly should interact with biological aging. The null result suggests our understanding of what drives epigenetic clock changes is incomplete, and that not all "anti-aging" mechanisms are created equal.¹

For safety, Centrum Silver is iron-free (important for older adults), contains modest calcium (~200 mg), and has a clean adverse event profile across COSMOS. A 2023 meta-analysis of 18 RCTs confirmed no increased mortality risk with daily multivitamin use.¹³ The safety case is solid. The efficacy case for aging is what's still being built.

Real Science, Real Signal, Uncertain Significance