A two-year randomized trial found that a daily multivitamin slowed biological aging across all five epigenetic clocks tested in a group of 958 older adults. Participants who entered the study with a higher biological age relative to their chronological age saw the largest gains, estimated at roughly four months of slowed aging. The results, drawn from a prespecified analysis within the COcoa Supplement and Multivitamin Outcomes Study, raise a sharp question: can a cheap, over-the-counter supplement meaningfully change how fast people age at the molecular level?
What is verified so far
The trial behind these findings is COSMOS, a randomized, double-blind, placebo-controlled study that used a 2×2 factorial design to test both cocoa extract and a daily multivitamin-multimineral supplement against placebo. The specific multivitamin tested was Centrum Silver, given daily for two years. COSMOS enrolled older adults and was registered as NCT02422745 with the U.S. National Library of Medicine. The trial was funded by the NIH.
The epigenetic analysis published in Nature Medicine drew on 958 participants from the parent trial. Researchers measured five DNA methylation clocks: PCHannum, PCHorvath, PCPhenoAge, PCGrimAge, and DunedinPACE. Each clock captures a slightly different dimension of biological aging. The Horvath clock, for instance, estimates DNA methylation age from CpG methylation patterns across human tissues, while DunedinPACE is designed to measure the pace of aging rather than a single-point age estimate. All five clocks showed slower aging in the multivitamin group compared with placebo.
Howard Sesso, an epidemiologist at the Harvard T.H. Chan School of Public Health, provided statements on the findings. According to the Harvard Gazette, the multivitamin effect was estimated as equivalent to approximately four months of aging. The same institutional summary confirmed that participants who were biologically older at baseline showed greater benefit, which aligns with the hypothesis that people whose cells are aging faster have more room for correction through micronutrient supplementation.
The trial’s design and baseline characteristics were detailed in a methods report describing a pragmatic, mail-based operation for recruitment, follow-up, and adherence tracking. That structure allowed COSMOS to run at scale without requiring participants to visit clinical sites, a feature that makes the design relatively easy to replicate and potentially adaptable to other large prevention trials.
What remains uncertain
The most pressing gap is whether slower epigenetic clocks actually translate into fewer heart attacks, cancers, or years of frailty. The published COSMOS epigenetic analysis reports clock changes, not clinical events. No direct mapping of the observed clock slowing to hard health outcomes appears in the primary results. Epigenetic clocks are validated biomarkers of aging, but they remain surrogate endpoints. A person whose DunedinPACE score improves may or may not live longer or stay healthier as a result, and the current data cannot settle that question.
Raw individual-level methylation values and detailed adherence logs from the 958-participant subset have not been released alongside the Nature Medicine paper. Without those data, independent researchers cannot yet verify whether participants who took the supplement most consistently saw the strongest clock changes, or whether certain baseline nutrient deficiencies predicted larger effects. Baseline and follow-up serum micronutrient concentrations linked to each participant’s clock changes are also absent from the published record, which limits insight into how closely the epigenetic response tracks with changes in vitamin or mineral status.
The mechanism behind the effect is similarly unresolved. Centrum Silver contains dozens of vitamins and minerals. It is not clear which specific nutrients drove the epigenetic signal, or whether the benefit came from correcting deficiencies common in older adults rather than from any single ingredient. One plausible hypothesis is that the effect size was largest among participants whose baseline nutrient shortfalls aligned with the specific CpG sites weighted most heavily in the PCGrimAge and DunedinPACE algorithms, but the published data do not yet test that idea directly. Without mechanistic work, it remains possible that the observed changes are the composite result of many small micronutrient corrections rather than a targeted anti-aging action.
The cocoa extract arm of the 2×2 factorial design also warrants attention. The Nature Medicine paper tested both interventions, yet the public discussion has focused almost entirely on the multivitamin results. Whether cocoa extract showed any independent or interactive effect on epigenetic clocks is a question the full paper addresses, but the institutional summaries have not emphasized it. If cocoa had a neutral effect, that would simplify interpretation; if it had its own influence on methylation patterns, teasing apart the contributions of each intervention will require more granular analysis.
How to read the evidence
The strongest evidence here comes from the trial’s prespecified design. This was not a post-hoc fishing expedition through old blood samples. The epigenetic analysis was planned before the data were unblinded, which reduces the risk that researchers cherry-picked a favorable outcome from many possible analyses. The parent trial’s randomized, placebo-controlled structure also means that confounders like diet, exercise, and socioeconomic status were distributed equally between groups at baseline, strengthening the causal interpretation that the multivitamin itself produced the observed clock differences.
That said, a four-month effect on an epigenetic clock is a modest signal. It is statistically robust in the context of the COSMOS dataset, but small enough that individual participants would not notice any tangible difference in day-to-day function. Biological aging is a slow, cumulative process; even a modest shift could matter over decades, yet such projections are speculative until linked to real-world outcomes. For now, the result is best understood as a biomarker change consistent with slightly slower molecular aging, not as proof of extended lifespan or guaranteed disease prevention.
Another important nuance is generalizability. COSMOS enrolled older adults who were healthy enough to participate in a long-term mail-based trial and willing to take daily pills. These volunteers may differ from the broader population in health behaviors, comorbidities, and access to medical care. The findings therefore apply most directly to similar older adults and may not extrapolate to younger people, those with severe chronic disease, or individuals with very different dietary patterns. Future work would need to test whether comparable epigenetic effects appear in other age groups or in populations with higher burdens of nutrient deficiency.
For clinicians and patients, the current evidence does not mandate a change in practice, but it does nudge multivitamins into a more intriguing scientific light. Until clinical endpoints are reported, the primary reasons to consider a daily multivitamin remain straightforward: filling potential nutritional gaps and aligning with existing guidelines for certain older adults. The COSMOS epigenetic analysis adds a layer of mechanistic plausibility that such supplementation might also influence the biology of aging, while still falling short of demonstrating concrete health benefits.
For researchers, the trial highlights both the promise and the limitations of epigenetic clocks as tools in large-scale intervention studies. They offer a way to detect subtle biological shifts over relatively short time frames, which is invaluable when waiting decades for mortality data is impractical. At the same time, the COSMOS experience underscores the need for transparent sharing of underlying methylation and adherence data, careful attention to multiple comparisons, and parallel tracking of clinical outcomes whenever possible.
In the coming years, follow-up analyses from COSMOS and independent replication efforts will be crucial. If similar multivitamin effects on epigenetic aging appear in other randomized cohorts, and if those changes can be tied to reductions in disease or disability, the case for routine supplementation as a low-cost longevity strategy will strengthen. Until then, the current findings are best viewed as a rigorously obtained but still preliminary signal that a simple daily pill may slow certain molecular markers of aging, while the larger story of how that translates into longer, healthier lives remains to be written.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
