David Sinclair’s Longevity Protocol: What Holds Up?

The world’s best-known advocate of reversing ageing has changed his own regimen repeatedly. That is not necessarily a contradiction. It is an opportunity to ask a sharper question: which of his ideas have survived contact with experiments on humans?

In April 2026, Peter Diamandis published an account of David Sinclair’s latest longevity routine. It included familiar names, such as NMN and resveratrol, alongside newer additions, such as nattokinase, spermidine and glycine. Sinclair had also reduced his alcohol intake, and his approach to glucose-lowering drugs had evolved. The list attracted attention because its author is not merely an enthusiastic consumer. He is a Harvard genetics professor whose laboratory has helped shape modern arguments about biological ageing.

The tempting response is to copy the list. The more interesting response is to examine it. If a molecule increases NAD+ in blood, has it changed the ageing process? If an artery looks less obstructed after a year, was the treatment responsible? And if a scientist’s experiments restore vision in old mice, what follows for a healthy person buying supplements?

This review treats three things separately: what Sinclair has reported doing, what experiments actually measured, and what remains a hypothesis. His personal regimen is not clinical advice, and public interviews cannot independently verify what he takes every day.

Key Takeaways

• His laboratory science is more consequential than his supplement stack. Partial epigenetic reprogramming has produced notable results in mice; it has not established a general way to reverse human ageing.
• NMN affects NAD-related biology, but clinical benefits remain uncertain. A 2026 review covering 15 trials found no convincing broad metabolic improvement and could address only short-term safety.
• Resveratrol is not an empty compound or a proven longevity treatment. Some human outcomes improve in selected groups; no trial establishes longer human life.
• Some medicines in his regimen have real clinical uses. Statins and metformin are effective in particular patients. That does not make them universal anti-ageing medicines.
• The most transferable lessons are comparatively ordinary. Physical activity, good sleep, appropriate cardiovascular risk assessment, and sustainable eating habits have a stronger human evidence base than the experimental stack.

The scientist and the protocol are different stories

Sinclair’s work matters independently of anything he eats for breakfast. In a 2020 Nature study, researchers including Yuancheng Lu and Sinclair expressed three reprogramming factors, commonly called OSK, in mouse retinal neurons. The treatment restored some more youthful molecular patterns, promoted nerve regeneration and improved vision in experimental models of glaucoma and ageing. A 2023 Cell paper developed the broader argument that loss of epigenetic information can contribute causally to mammalian ageing.

Those are substantial experimental findings. They are also easy to overextend. Reversing a measurable change in mouse retinal tissue is not proof that an intervention extends human life, reverses the ageing of an entire human body, or can be reproduced by swallowing a precursor molecule. Restored function, a younger molecular signature, fewer diseases and longer survival are four different endpoints.

The distinction is especially important because the cellular reprogramming intervention is not the same as Sinclair’s daily supplement routine. One is an experimental manipulation of gene expression in particular tissues. The other is a collection of dietary choices, drugs and supplements, most of which have never been tested as a combined longevity protocol.

The history of longevity research is full of plausible mechanisms that proved harder to translate into clinical benefits than expected. This does not mean mechanism is irrelevant. It means that a mechanism opens a research programme; it does not complete one.

What Sinclair reported taking in 2026

A 21 April 2026 account by Peter Diamandis, following an interview recorded in March, lists approximately 1 g of NMN and 1 g of resveratrol per day and describes cycling between metformin and berberine. It also names nattokinase, spermidine, glycine, niacin, vitamin D3/K2 and low-dose aspirin. Sinclair describes a mainly plant-based diet, time-restricted eating, physical activity and much less alcohol than in earlier interviews.

This is a dated, attributed report, not an independently audited medication record. Dosages and frequency may change. Some choices relate to his stated personal cardiovascular risk rather than ageing in healthy adults. Diamandis is also a personal and commercial associate of Sinclair, not an independent clinical investigator. We therefore treat his account as evidence of what was publicly reported, not evidence that any component works.

Older versions of Sinclair’s routine mentioned CoQ10, melatonin, GABA, quercetin, fisetin and particular daily schedules. They should not be silently carried forward as a verified 2026 regimen. Equally, a routine described in an interview cannot establish the exact contents of every subsequent day’s pill organiser.

How the public record changed, 2019–2026

There is no single, immutable “Sinclair protocol”. The primary record is a series of interviews, a book, podcast episodes and later accounts by associates. The date and the speaker matter: an item described in 2022 is not automatically still taken in 2026. These milestones track public statements, not independently verified adherence.

One discrepancy is worth preserving rather than smoothing away: Diamandis’s April 2026 account presents nattokinase as a daily 10,000-unit component, whereas in his July 2026 interview Sinclair says he is trying it. The distinction may reflect changed practice or different levels of detail. We cannot infer stable daily use from either snapshot alone.

The revealing pattern is not that he changed his mind. Revising a regimen when evidence or tolerability changes is reasonable. The problem arises when older anecdotes are recycled as a definitive current protocol, or when a change in a biomarker, an animal experiment or a commercial claim is presented as proof of extra years of human life.

The evidence, at a glance

This table deliberately separates an effect on a surrogate marker from a measured clinical benefit. A statistically significant change is neither automatically large nor automatically useful; the population, absolute magnitude, uncertainty and adverse effects all matter.

Resveratrol: how a fluorescent tag changed the story

Resveratrol became a celebrity molecule because a laboratory hypothesis seemed to connect red wine, cellular stress responses and longevity. It was proposed to activate SIRT1, an enzyme linked to important regulatory processes. The biochemical story was elegant enough to travel far beyond the laboratory.

Then researchers changed the measurement. In 2005, independent investigators found that apparent activation depended on an artificial fluorescent tag attached to the substrate. With an unlabelled substrate, the effect was absent in those assays. Further work in 2010 challenged whether resveratrol and related compounds directly activated SIRT1 with native substrates.

This did not demonstrate that resveratrol has no biological activity. Later work continued to investigate substrate-specific mechanisms and indirect signalling. But it dismantled the easy equation: resveratrol activates a longevity gene, therefore a pill should slow human ageing.

Human evidence now tells a more nuanced story. An April 2026 umbrella review assessed 45 systematic reviews and meta-analyses across 129 outcome associations. Of 35 statistically significant associations, only four received a high-certainty rating. They included a mean waist-circumference difference of −0.80 cm (95% CI −1.40 to −0.20) and lower systolic blood pressure in people with type 2 diabetes of −7.97 mmHg (95% CI −10.63 to −5.31).

Those are possible clinical effects in defined populations, not evidence of rejuvenation. The review aggregates overlapping research rather than representing 45 independent experiments, and its estimates must be read against trial heterogeneity and population differences. No included outcome demonstrates that resveratrol extends human lifespan.

Sinclair’s reported 1 g daily dose is consequently an experiment of personal preference and biological plausibility, not an established longevity prescription. Whether fat taken alongside a dose alters absorption does not answer whether the absorbed compound improves survival.

NMN: changing the marker is not changing the outcome

NMN is a precursor to NAD+, a coenzyme involved in energy metabolism and the activity of multiple enzymes. That makes it an attractive candidate for research into age-related biology. It also creates a tempting shortcut: if NAD-related measures rise, perhaps ageing must slow.

The human results have not justified that shortcut. A 2026 systematic review and meta-analysis included 15 randomised trials of oral NMN or related preparations, with studied doses generally between 250 and 2,000 mg per day and durations from 14 days to 24 weeks. Broad improvements were not evident for body weight, BMI, fasting glucose, HbA1c, lipid profiles or systolic blood pressure. A small diastolic-pressure signal warranted further investigation. Short-term adverse-event results were generally reassuring, but cannot establish safety over years.

An earlier synthesis of eight trials involving 342 adults likewise found no significant aggregate effect on glucose control or lipid profiles. Short, small trials can miss modest benefits; equally, they cannot be stretched into proof of a lifespan advantage.

Most importantly, the trials did not measure the outcome that gives the protocol its name: living longer. Nor did they demonstrate that a person whose NAD+-related biomarker moves the most will have a better clinical future.

Metformin and berberine: from diabetes to longevity

Metformin works for a disease we can define and measure: type 2 diabetes. Its extension into longevity research rests on observations in model organisms, molecular mechanisms, and epidemiological comparisons among people taking different diabetes treatments. The leap to healthy adults is much less secure.

A 2025 critical review examined follow-up studies, limitations in influential observational findings and trials in people without type 2 diabetes. Its conclusion was not that metformin has been disproved as a gerotherapeutic, but that confidence in its anti-ageing potential has weakened as better evidence has accumulated.

There is an important methodological trap here. People prescribed different glucose-lowering drugs may differ in disease severity, weight, kidney function, healthcare use and other factors. Even sophisticated statistical adjustment cannot automatically make them equivalent to randomly assigned healthy adults. Comparing their mortality can generate hypotheses, but cannot by itself prove that a healthy person will live longer on metformin.

Sinclair has described alternating metformin with berberine. Berberine is biologically active and may influence glucose or lipid measures in some studies, but calling it a natural substitute for metformin obscures differences in formulation, evidence, interactions and quality control. There is no credible basis to assume the two are interchangeable anti-ageing treatments.

Questions about exercise adaptation also need precision. Some studies raise concern that metformin may blunt particular training responses in some populations. That does not prove that everyone who takes metformin loses muscle or should interrupt prescribed treatment around exercise. Individual treatment decisions belong with a clinician.

What about the other supplements?

Nattokinase: the problem with 1,062 patients

The 2026 list includes nattokinase, an enzyme associated with fermented soybeans. A 2022 clinical report involving 1,062 participants described improvements in blood lipids and carotid ultrasound measures, particularly with a dose around 10,800 fibrinolytic units (FU) per day. The result sounds impressive because the sample is large and the imaging endpoints seem objective.

But a large study is not necessarily a strong causal experiment. This report did not establish treatment efficacy through a robust randomised, blinded, placebo-controlled comparison. Changes in ultrasound measures may reflect selection, concurrent behaviour, treatment changes, regression to the mean or measurement variability. The study also did not demonstrate fewer heart attacks, strokes or deaths.

There are additional practical concerns when a substance may affect clotting, especially if a person also takes aspirin or anticoagulants. No reader should infer a safe regimen by combining items from someone else’s list.

Spermidine, glycine, quercetin and fisetin

Spermidine is interesting because of its relationship to cellular maintenance and autophagy. Yet a 12-month randomised trial involving 100 older adults found no significant improvement in its primary memory outcome relative to placebo. That is not a universal test of longevity, but it is a useful antidote to the claim that a plausible mechanism guarantees a clinical effect.

Glycine has established physiological roles, and fisetin and quercetin remain interesting research compounds. None has demonstrated that routine supplementation extends human life. Animal experiments, changes in senescence-related markers and theoretical methylation benefits remain separate from clinically demonstrated human rejuvenation.

Statins, aspirin, niacin and vitamin supplements

Statins are an instructive contrast with speculative longevity products. Their cardiovascular benefits are supported by large randomised trial programmes in appropriately selected patients. Someone with high LDL cholesterol, elevated cardiovascular risk or a relevant medical history may have a clear reason to take one. The reason is risk reduction, not that every healthy person should copy Sinclair’s prescription.

Low-dose aspirin has a more complicated balance: bleeding can outweigh benefits in primary prevention for some people, whereas secondary prevention after particular cardiovascular events is different. Similarly, niacin can change lipid measures, including lipoprotein(a), but large outcome trials did not establish added cardiovascular benefit when combined with modern lipid-lowering treatment in broad populations. AIM-HIGH is a prominent example of why a changed number is not enough.

Vitamin D and K fulfil essential nutritional functions, but that does not justify high-dose supplementation without context. Vitamin D intake around 5,000 IU per day exceeds the usual adult upper limit of 4,000 IU unless specifically supervised. CoQ10, melatonin, GABA and other products in historical accounts of Sinclair’s routine should not be described as verified current daily habits without a dated source.

What is left when the experimental stack is removed?

Perhaps the most surprising finding is how ordinary the better-supported advice becomes. Sinclair has described walking, regular exercise, a predominantly plant-based diet, attention to cardiovascular risk and sleep. These are not as novel as epigenetic reprogramming. They are also much closer to interventions supported by extensive human evidence.

The World Health Organization recommends adults accumulate about 150–300 minutes of moderate aerobic activity weekly, or 75–150 minutes vigorous activity, with muscle-strengthening work on at least two days. These recommendations draw on substantial human evidence for cardiovascular and functional health. Observational mortality estimates still have their own confounding limits, but the evidence base is far broader than for a multi-supplement longevity stack.

His reported move away from habitual alcohol use is another notable revision. Earlier portraits described red wine with dinner. In the 2026 account, occasional celebratory drinking replaces a regular habit. The evidence on alcohol and long-term health is complex, but the old idea that daily red wine is a straightforward longevity intervention should not be repeated as established fact.

Time-restricted eating may be convenient and can improve some outcomes for some people, particularly where it changes overall energy intake. Yet a narrow eating window is not itself proven to lengthen human life. A proposed 72-hour fasting programme has additional safety and individual-context questions. One should not confuse an intriguing autophagy mechanism with a validated schedule for everyone.

There is a further practical point. A protocol changes with age, inherited risks, disease, tolerance and new evidence. The sensible part is not memorising its exact ingredients; it is recognising when a claim requires better testing.

Commercial interests and scientific claims

Sinclair is a laboratory scientist, bestselling author and biotechnology entrepreneur. These roles can reinforce one another by bringing investment into neglected research. They can also create incentives to communicate possibilities before they have become reliable treatments. Financial or reputational interests do not make a finding false; they increase the importance of independent replication and careful wording.

The distinction became public in 2024 when a claim associated with an Animal Bioscience dog supplement was described as showing that ageing had been reversed. STAT reported criticism from other scientists and the absence, at that time, of published research establishing that the product reversed or slowed ageing in dogs. Sinclair disputed how the promotional language was presented, and the controversy later contributed to his departure from the presidency of a longevity researchers’ organisation.

None of this invalidates his 2020 mouse-vision paper. Nor does a successful mouse experiment validate a marketed supplement. The claims must be separated and assessed one by one. That is the standard we should apply to advocates, sceptics and ourselves alike.

So, should anyone copy his routine?

The fairest verdict is neither that Sinclair has found a way to live longer nor that his research is worthless. His group has contributed important experiments to the biology of ageing. His public regimen is an evolving collection of personal experiments, established treatments for particular risks and interventions whose human benefits remain unresolved.

A person seeking practical benefit should start from measurable needs: current health, medical history, blood pressure, clinically appropriate lipids and glucose measures, fitness, sleep, diet and exposure to known risks. New drugs and supplements then face a higher bar: do they improve outcomes that matter, compared with the alternatives, at an acceptable cost and risk?

There is no reason to abandon ambitious hypotheses. But the gap between an attractive biological mechanism and a human outcome is precisely where careful scientific journalism belongs.

Sources and method

Featured photograph: David Sinclair at the 2025 UN AI for Good Summit, Geneva. Photo by Xuthoria, Wikimedia Commons, licensed CC BY-SA 4.0. Cropped and resized by Mike’s Balance; the adapted image is also distributed under CC BY-SA 4.0.

Evidence cutoff: 10 October 2026. This is a critical synthesis of public statements and published studies, not an independent experiment or a verified record of Sinclair’s medication use. Statements about his personal regimen are attributed to dated interviews or associated accounts. Claims about efficacy are assessed separately, giving greater weight to randomised human trials and reviews than to mechanistic speculation. Absence of demonstrated benefit is not proof of zero effect; short trials provide limited evidence about decades-long outcomes.

1. Diamandis PH. Updated Longevity Protocols for 2026. 21 April 2026. First-person associate’s account, not an independent audit.
2. Sinclair DA, interviewed by Diamandis PH. Moonshots: Updated Protocols. Recorded 9 March; published 27 April 2026.
3. Lu Y et al. Reprogramming to recover youthful epigenetic information and restore vision. Nature. 2020.
4. Yang JH et al. Loss of epigenetic information as a cause of mammalian aging. Cell. 2023.
5. Kaeberlein M et al. Substrate-specific activation of sirtuins by resveratrol. J Biol Chem. 2005.
6. Pacholec M et al. SRT1720 and resveratrol are not direct activators of SIRT1. J Biol Chem. 2010.
7. Sun JN et al. Effects of resveratrol supplementation on multiple health outcomes: an umbrella review. Nutrition Journal. 2026.
8. Safety and metabolism-related outcomes of oral NMN: systematic review and meta-analysis. 2026.
9. Effects of NMN on glucose and lipid metabolism in adults: systematic review of randomised trials. 2024.
10. Keys MT et al. Emerging uncertainty on the anti-aging potential of metformin. Ageing Research Reviews. 2025.
11. Chen H et al. Nattokinase and atherosclerosis: clinical study of 1,062 participants. 2022. Interpret cautiously because of its non-randomised design.
12. Schwarz C et al. Spermidine supplementation and cognition: a randomised clinical trial. JAMA Network Open. 2022.
13. AIM-HIGH trial. J Am Coll Cardiol. 2013.
14. WHO Guidelines on Physical Activity and Sedentary Behaviour. 2020.
15. Molteni M. Harvard longevity scientist sparks furor with claim about reversing aging in dogs. STAT. 2024. Reporting on the public controversy, not primary clinical evidence.

Related reading: Resveratrol and longevity; Human trials of NMN; What epigenetic age tests can and cannot tell you; Why facial age estimates do not measure rejuvenation.