6 Needs context
In the COSMOS trial, participants given a multivitamin slowed their brain aging by 2.1 years.
"So, this was the COSMOS trial, and at the end of the trial, the people that were given the multivitamin had slowed their brain aging by 2.1 years." (said at 0:00:32)
In the randomized COcoa Supplement and Multivitamin Outcomes Study (COSMOS) cognitive substudies (including COSMOS-Mind, COSMOS-Web, and COSMOS-Clinic), daily multivitamin-mineral supplementation significantly improved cognitive performance (global cognition and episodic memory) relative to placebo in older adults. By benchmarking effect sizes against cross-sectional age-related cognitive decline, trial investigators estimated that multivitamin use was equivalent to slowing cognitive aging by approximately 1.8 to 3.1 years (and by 2 years in the meta-analysis of the three substudies). The claim requires context because the trial evaluated neuropsychological test scores of cognitive function rather than direct biological or structural brain aging.
Epigenetic clocks have a correlation of approximately 0.5 with one another after regressing out chronological age, sex, and other covariates.
"Why? Because these clocks are correlated with each other. And just to throw out a number, correlation 0.5 after you regress out age, sex, and various variables, but there's still a fairly good agreement." (said at 0:34:22)
The speaker's figure of r ≈ 0.5 is a reasonable approximation for the moderate inter-clock agreement observed between certain DNA methylation clocks, but requires qualification. Epigenetic clocks derive measures of biological age acceleration as the residuals after regressing estimated DNA methylation age on chronological age (often alongside sex and leukocyte composition). In observational cohorts evaluating multiple clocks (such as the first-generation Horvath and Hannum clocks and second-generation PhenoAge and GrimAge clocks), pairwise residual correlations vary considerably—ranging from weak (r ≈ 0.1–0.3) between clocks trained on different target outcomes (e.g., chronological age vs. mortality/morbidity biomarkers) to moderate (r ≈ 0.4–0.6) between structurally similar clocks.
HIV-positive individuals exhibit 5 to 7 years of epigenetic age acceleration in blood, which is reversed by 4 to 5 years following antiretroviral therapy.
"HIV-positive people exhibit epigenetic age acceleration. It's actually a pronounced pro-aging effect, maybe 5 to 7 years in blood. And sure enough, if they stick to their antiretroviral therapy, that will reverse their epigenetic age... Several years, to give you a number: 4 or 5 years." (said at 0:40:03)
Studies evaluating DNA methylation clocks in blood confirm that untreated HIV infection leads to significant epigenetic age acceleration, which is partially reversed following antiretroviral therapy (ART). However, the exact magnitude depends heavily on the specific epigenetic clock used. In a substudy of the NEAT001/ANRS143 trial, untreated individuals with HIV showed a mean epigenetic age acceleration of 7.3 years using PhenoAge and 2.5 years using Horvath's clock compared to uninfected controls. After 2 years of ART, PhenoAge acceleration decreased by nearly 3.6 years (remaining 3.69 years higher than controls), while Horvath acceleration largely normalized. Longitudinal tracking over 17 years similarly demonstrates that epigenetic age accelerates during untreated infection (0.36–0.69 years per untreated year depending on the clock) and decelerates during suppressive ART (-0.26 to -0.49 years per treated year). The speaker's numerical ranges align closely with estimates from the PhenoAge clock, though other clocks reflect smaller baseline shifts.
- supports: Epigenetic age acceleration changes 2 years after antiretroviral therapy initiation in adu… (The lancet. HIV 2021) · cited 107x in the literature
"Compared with the HIV-uninfected group, ART-naive participants with HIV showed higher epigenetic age acceleration (EAA) according to all EAA estimators (mean 2·5 years, 95% CI 1·89-3·22 for Horvath-EAA; 1·4 years, 0·74-1·99 for Hannum-EAA; 2·8 years, 1·97-3·68 for GrimAge-EAA; and 7·3 years, 6·40-8·13 for PhenoAge-EAA)... After 2 years of ART, epigenetic age acceleration was reduced, although PhenoAge and GrimAge remained significantly higher in participants with HIV compared with participants without HIV (mean difference 3·69 years, 95% CI 1·77-5·61; p=0·0002 and 2·2 years, 0·47-3·99; p=0·013, respectively)." (abstract, results, passage verified)
pubmedfull study (doi) - supports: Epigenetic ageing accelerates before antiretroviral therapy and decelerates after viral su… (The lancet. Healthy longevity 2023) · cited 52x in the literature
"Per year of untreated HIV infection (median observation 8·08 years, IQR 4·83-11·09), mean EAA was 0·47 years (95% CI 0·37 to 0·57) for Horvath's clock, 0·43 years (0·3 to 0·57) for Hannum's clock, 0·36 years (0·27 to 0·44) for SkinBlood clock, and 0·69 years (0·51 to 0·86) for PhenoAge. Per year of suppressive ART (median observation 9·8 years, IQR 7·2-11), mean EAA was -0·35 years (95% CI -0·44 to -0·27) for Horvath's clock, -0·39 years (-0·50 to -0·27) for Hannum's clock, -0·26 years (-0·33 to -0·18) for SkinBlood clock, and -0·49 years (-0·64 to -0·35) for PhenoAge." (abstract, results, passage verified)
pubmedfull study (doi)
DunedinPACE was developed to track changes in BMI and pace of aging, whereas GrimAge was trained on mortality.
"DunedinPACE again was uh um trained—that's the lingo of machine learning—but it was developed to track changes in BMI. So yes, it picked it up. By contrast, GrimAge was never trained to look at weight loss. It was trained on mortality." (said at 1:09:04)
The speaker is correct regarding the general training objectives, but with important nuance regarding DunedinPACE's design. GrimAge was trained on time-to-death (all-cause mortality) alongside DNA methylation surrogates of plasma proteins and smoking pack-years. DunedinPACE was not developed solely to track BMI or weight loss; rather, it was trained on the multi-system 'Pace of Aging,' a composite measure tracking longitudinal change across 19 physiological biomarkers of organ-system integrity over 20 years (ages 26 to 45 in the Dunedin Study), which included BMI and waist-to-hip ratio alongside cardiovascular, metabolic, pulmonary, kidney, and immune indicators.
- supports: DNA methylation GrimAge strongly predicts lifespan and healthspan. (Aging 2019) · cited 2724x in the literature
"The resulting predictor of lifespan, DNAm GrimAge (in units of years), is a composite biomarker based on the seven DNAm surrogates and a DNAm-based estimator of smoking pack-years. Adjusting DNAm GrimAge for chronological age generated novel measure of epigenetic age acceleration, AgeAccelGrim .Using large scale validation data from thousands of individuals, we demonstrate that DNAm GrimAge stands out among existing epigenetic clocks in terms of its predictive ability for time-to-death" (abstract, results, passage verified)
pubmedfull study (doi) - context: DunedinPACE, a DNA methylation biomarker of the pace of aging. (eLife 2022) · cited 1058x in the literature
"We used data from the Dunedin Study 1972-1973 birth cohort tracking within-individual decline in 19 indicators of organ-system integrity across four time points spanning two decades to model Pace of Aging. We distilled this two-decade Pace of Aging into a single-time-point DNA-methylation blood-test using elastic-net regression" (abstract, results, passage verified)
pubmedfull study (doi) - context: Effect of long-term caloric restriction on DNA methylation measures of biological aging in… (Nature aging 2023) · cited 189x in the literature
"We found that CALERIE intervention slowed the pace of aging, as measured by the DunedinPACE DNAm algorithm, but did not lead to significant changes in biological age estimates measured by various DNAm clocks including PhenoAge and GrimAge." (abstract, results, passage verified)
pubmedfull study (doi)
In the DO-HEALTH trial, the combined intervention of 1g omega-3, 2000 IU vitamin D, and home exercise delayed PhenoAge biological aging by 3.8 months over 3 years, while reducing metastatic cancer risk by 61% and pre-frailty by approximately 20%.
"high dosage vitamin D, um omega-3 plus exercise. And according to PhenoAge, that treatment arm did the best... I think it was 3.8 months the PhenoAge delayed the biological aging was delayed by 3.8 months, yeah, over three years of that... but also that was associated with outcomes that were important: 61% reduced chance of getting metastatic cancer, it was like a 20% reduction in pre-frailty" (said at 1:27:34)
The speaker's summary closely reflects published analyses from the DO-HEALTH randomized trial, with minor imprecision regarding cancer staging. In a post hoc analysis of 777 DO-HEALTH participants, the combination of 2,000 IU/day vitamin D3, 1 g/day omega-3, and a simple home exercise program showed additive benefits on the PhenoAge DNA methylation clock, slowing biological aging by up to 3.8 months over 3 years. Furthermore, in the trial's exploratory analysis of cancer risk, the triple combination was associated with a 61% reduction (adjusted HR 0.39, 95% CI 0.18–0.85) in the risk of 'any verified invasive cancer', rather than specifically 'metastatic cancer'.
- partial: Combined Vitamin D, Omega-3 Fatty Acids, and a Simple Home Exercise Program May Reduce Can… (Frontiers in aging 2022) · cited 31x in the literature
"The aim of this study was to test the individual and combined benefit of vitamin D, omega-3, and a simple home strength exercise program on the risk of any invasive cancer... For all three treatments combined, the adjusted HR was 0.39 (0.18-0.85; 4 vs. 12 cases)." (abstract, results)
pubmedfull study (doi) - supports: Individual and additive effects of vitamin D, omega-3 and exercise on DNA methylation cloc… (Nature aging 2025) · cited 126x in the literature
"Omega-3 alone slowed the DNAm clocks PhenoAge, GrimAge2 and DunedinPACE, and all three treatments had additive benefits on PhenoAge. Overall, from baseline to year 3, standardized effects ranged from 0.16 to 0.32 units (2.9-3.8 months)." (abstract, results, passage verified)
pubmedfull study (doi)
In the Berlin Aging Study II (BASE-II), vitamin D supplementation slowed epigenetic aging in participants who were deficient, but showed no effect in those who were already sufficient.
"where they took which was the thing that was nice about that was they had a deficient population and then a sufficient population and gave them vitamin D... you re reverse aging if you're if you're deficient and fill that sufficiency. But the people that were not deficient, actually there was no effect" (said at 1:31:47)
A longitudinal analysis of the Berlin Aging Study II (BASE-II) and its follow-up GendAge study (PMID 35562603) found that participants with baseline vitamin D deficiency who began vitamin D supplementation had significantly lower DNA methylation age acceleration (2.6 years lower on the 7-CpG clock and 1.3 years lower on Horvath's clock) compared to untreated deficient individuals, reaching epigenetic ages comparable to vitamin D-sufficient controls. However, this was an observational, quasi-interventional longitudinal study (where participants chose to take supplements over ~7.4 years of follow-up), not a randomized controlled trial that administered vitamin D to both deficient and sufficient cohorts.
- context: Vitamin D supplementation is associated with slower epigenetic aging. (GeroScience 2022) · cited 56x in the literature
"Vitamin D-deficient participants who chose to start vitamin D supplementation after baseline examination showed a 2.6-year lower 7-CpG DNAmAA (p = 0.011) and 1.3-year lower Horvath DNAmAA (p = 0.042) compared to untreated and vitamin D-deficient participants. DNAmAA did not statistically differ between participants with successfully treated vitamin D deficiency and healthy controls (p > 0.16). Therefore, we conclude that intake of vitamin D supplement is associated with lower DNAmAA in participants with vitamin D deficiency." (abstract, results, passage verified)
pubmedfull study (doi)
Unverified means no publication matching the claim was located; it does not prove the claim false. Spotted an error? See the corrections policy - disputes from the people quoted are prioritized.