A randomized clinical trial involving 958 older adults suggests that taking a daily multivitamin for two years may produce a small, statistically significant slowdown in two measures of biological aging. A Nature news analysis described the effect as roughly four months of slower biological aging compared with placebo over the study period, a result that has energized researchers studying how supplements interact with the molecular machinery of aging. Yet the finding sits in tension with larger, longer studies showing that multivitamins do not lower the risk of death, raising a pointed question about what “slower aging” on a lab test actually means for a person’s health.
What the COSMOS Epigenetic Study Found
The new results come from an ancillary study nested inside the larger COcoa Supplement and Multivitamin Outcomes Study, a randomized trial that enrolled more than 20,000 older U.S. adults in a two-by-two factorial design testing both cocoa extract and a daily multivitamin-multimineral supplement. The parent trial’s design and baseline characteristics have been detailed in a report on older participants, highlighting its focus on generally healthy men and women without major recent cardiovascular events.
For the epigenetic analysis, researchers drew blood samples from a subset of 958 participants with a mean age around 70 and measured changes in five DNA-methylation-based aging clocks over two years. The methods and main findings of this substudy are described in a Nature Medicine report that emphasized its randomized, placebo-controlled structure and blinded laboratory assessment.
DNA methylation is a chemical modification that accumulates on genes over a lifetime, and scientists have developed algorithms, often called epigenetic clocks, that translate methylation patterns into an estimate of biological age. In COSMOS, participants assigned to the multivitamin group showed statistically significant slowing on two of the five clocks tested, specifically newer “second-generation” measures designed to predict health outcomes rather than simply track chronological age. A news analysis in Nature framed the effect as about four months of slower biological aging compared with placebo over the two-year period.
That four-month figure is modest on an individual level, but the pattern within the data may matter more than the average. The researchers reported that the slowing appeared larger among participants whose baseline measurements suggested accelerated aging, hinting that the supplement’s impact could be greater in people whose biology is already trending in a worse direction. If future trials replicate that pattern, the conversation could shift away from blanket recommendations toward targeted use in higher-risk subgroups.
Epigenetic Clocks as Stand-Ins for Health
The central challenge with this study is not the measurements themselves but what they can prove. Epigenetic clocks are statistical constructs trained to predict age or health outcomes based on methylation patterns; they are not direct measurements of disease, disability, or quality of life. A commissioned expert review in Nature Medicine examined whether these tools can function as reliable surrogate endpoints in clinical trials and laid out a cautious framework for using methylation-based biomarkers to evaluate interventions.
Surrogate endpoints are lab values that stand in for clinical outcomes, the way LDL cholesterol stands in for heart attack risk or blood pressure for stroke risk. If epigenetic clocks eventually earn that status, researchers could test potential anti-aging therapies far more quickly and cheaply than by waiting decades for mortality or major disease data to accumulate. But the bar for acceptance is high: a valid surrogate must consistently track with meaningful health outcomes across diverse populations and interventions.
The commentary emphasized how difficult that standard will be to meet. Slowing a methylation-based score by a few months does not automatically mean a person will live longer, avoid dementia, or remain physically independent. Clocks such as DunedinPACE and related measures have shown promising correlations with functional decline, chronic disease, and mortality in prior cohorts, yet the expert panel underscored that substantial calibration is still needed before clinicians can interpret small changes as actionable signals for patients.
When Biomarkers and Hard Outcomes Diverge
The tension becomes sharper when the epigenetic findings are placed next to the parent trial’s clinical results. The full COSMOS trial, which followed participants for a median of roughly 3.6 years, tested multivitamins and cocoa extract against cancer and cardiovascular endpoints. In a detailed outcomes paper, investigators reported that daily multivitamin use did not clearly reduce major cardiovascular events or total invasive cancer, though there were exploratory signals in some subgroups described in the trial report.
An even larger dataset points in the same direction. A long-term observational analysis of 390,124 U.S. adults followed for more than two decades found that regular multivitamin use was not associated with lower risk of death from any cause, including cardiovascular disease and cancer. The National Institutes of Health summarized this work by noting that daily multivitamins in otherwise healthy adults did not appear to reduce overall mortality, based on the large cohort examined.
These findings create a conceptual puzzle. If multivitamins genuinely slow biological aging in a way that matters, one might expect that effect to emerge eventually in survival statistics or major disease rates, especially across hundreds of thousands of people and decades of follow-up. The current evidence does not show that.
Several explanations are possible. The epigenetic signal could be real but too small to influence mortality in a generally healthy population over typical trial timeframes. Alternatively, methylation changes may be capturing aspects of cellular regulation that do not translate into fewer clinical events. A third possibility is that any benefit is concentrated in people with low baseline nutrient status or accelerated aging; in that case, large observational datasets that pool together supplement users with widely varying health profiles could easily dilute effects confined to a minority.
Cognitive Gains Add a Wrinkle
The COSMOS program did produce one functional outcome that aligns more closely with the epigenetic results. A companion trial known as COSMOS-Mind tested the same multivitamin supplement against placebo on cognitive endpoints and found a statistically significant benefit on global cognition for participants assigned to the multivitamin group. The investigators reported that cocoa extract had different or more limited effects across specific cognitive domains, according to the published neurocognitive analysis.
This cognitive signal matters because brain aging and systemic biological aging share overlapping mechanisms, including vascular health, oxidative stress, and inflammatory pathways. If a daily multivitamin both slows methylation-based aging and modestly improves cognitive performance in older adults, the two findings may be reflecting a common underlying process, such as correction of subtle micronutrient deficits that affect neuronal resilience.
Still, the size of the cognitive benefit was not dramatic, and the trial population skewed toward relatively healthy, well-educated older adults who were willing to enroll in a long-term study. That limits how broadly the results can be generalized. The convergence between cognition and epigenetic aging strengthens the case that something biologically meaningful is happening, but it does not yet clarify who stands to benefit most or how durable the effect might be over longer periods.
Funding, Limitations, and What Comes Next
Any discussion of supplement research requires attention to funding and potential conflicts of interest. The COSMOS program received support from the National Institutes of Health as well as industry partners that supplied study pills and packaging, and several investigators have disclosed relationships with supplement or pharmaceutical companies in trial publications. While the randomized, placebo-controlled design helps mitigate bias, readers should recognize that commercial stakeholders have a clear interest in positive findings around multivitamin use.
The epigenetic substudy itself also has important limitations. It examined a relatively small subset of participants, and the two-year follow-up captures only a brief window of an aging process that unfolds over decades. Only two of the five methylation clocks tested showed significant slowing, raising questions about how robust or generalizable the effect is across different measures of biological age. The participants were predominantly older, health-conscious adults, so the results may not apply to younger people, those with significant chronic illness, or populations with widespread nutritional deficiencies.
For now, these data do not justify taking multivitamins as an “anti-aging” therapy. Instead, they highlight how sophisticated biomarkers can reveal subtle biological shifts that traditional clinical endpoints may miss, at least over short time horizons. The logical next steps include longer follow-up of the COSMOS cohort, replication of the epigenetic findings in independent trials, and studies that specifically enroll individuals with evidence of accelerated aging or inadequate nutrient intake at baseline.
For clinicians and patients, the message is nuanced. Multivitamins remain reasonable for people with documented deficiencies, restrictive diets, or conditions that impair nutrient absorption, but routine use in otherwise healthy adults has not been shown to extend life or prevent major diseases. The new epigenetic and cognitive findings are intriguing, suggesting that small biological benefits may exist in certain groups, yet they are not strong enough on their own to outweigh established public-health advice that prioritizes balanced diets, physical activity, and control of known risk factors over blanket supplementation. As the science of biological aging advances, multivitamins may find a more clearly defined role, but that will depend on future trials that connect molecular changes to tangible gains in healthspan and longevity.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
