What is actually being claimed
Three quite different claims travel together under the NAD banner, and separating them is most of the work of grading them.
The first is biochemical: that nicotinamide adenine dinucleotide, NAD, is central to how cells make and spend energy, and that concentrations of it fall in ageing tissue. The second is pharmacokinetic: that swallowing a precursor such as nicotinamide mononucleotide (NMN) or nicotinamide riboside (NR) raises NAD-related metabolites that can be measured afterwards in blood. The third is clinical: that doing so makes a person age more slowly, live longer, or stay functional for longer.
The first claim is well supported. The second is supported in humans over short periods. The third is the one sold on the label, and it is the one with almost nothing behind it. Our grade attaches to the third claim, because that is the claim a reader is being asked to spend money on.
This distinction is not pedantry. It is the same structure as almost every failed intervention in the history of preventative medicine: a real mechanism, a real biomarker movement, and no demonstrated benefit at the level of the patient. Our explainer on surrogate endpoints sets out why that pattern recurs.
The mechanism, and how well it is established
NAD works as a carrier of electrons in metabolism, cycling between oxidised and reduced forms, and it is also consumed as a substrate by two families of enzymes that matter to ageing biology: the sirtuins, which remove chemical marks from proteins including histones, and the PARPs, which are recruited when DNA is damaged. Because those enzymes eat NAD rather than merely borrowing it, a cell under sustained DNA damage or inflammatory load has a genuine reason to run short.[3]
Falling tissue NAD with age has been reported repeatedly in laboratory organisms and in animal tissue. In humans the picture is less clean, because the measurement is harder: concentrations vary between tissues, blood is a poor proxy for muscle or brain, and sample handling affects the result. The strong version of the claim, that every human tissue shows a steady age-related decline, is more confident than the human measurements presently support.
The precursor logic then runs: if the pool is depleted, refill it. NR and NMN both feed into the salvage pathway that recycles nicotinamide back into NAD, and both are absorbed orally. There is a longstanding technical argument about whether NMN enters cells intact through a dedicated transporter or is first broken down to NR at the gut wall. It matters for how the compounds are designed and marketed. It does not much affect the grade, because the downstream question is unanswered either way.
Worth stating plainly: mechanistic plausibility is the cheapest thing in this field. Nearly every intervention reviewed in this journal has a plausible mechanism. Plausibility is the entry ticket, not the evidence.
What the human trials have shown
Human trials of NR and NMN exist in reasonable number now, and they share a shape. They are small. They usually run for a number of weeks, occasionally a few months. They recruit healthy adults or a specific patient group, and their primary outcome is almost always the blood concentration of NAD or its metabolites, sometimes with tolerability and safety bloods alongside.
On those terms, they generally succeed. Oral dosing raises measurable NAD metabolites, and short-term tolerability has been reported as acceptable across trials. The safety signal so far is unremarkable, which is not nothing, though short trials in healthy volunteers are the setting least likely to reveal a slow problem.
Where the literature thins dramatically is at the next step. Trials that go on to measure something functional, such as walking speed, muscle strength, insulin sensitivity, aerobic capacity or cognitive performance, report results that are inconsistent between studies and between the outcomes within a single study. Some report a change in one measure and not others. Some report nothing. Independent replication of any specific functional finding is scarce.
Funding is a live issue. A substantial share of this literature is sponsored by companies that manufacture or sell the compound, or is conducted with material supplied by them. That does not make a result wrong, and disclosed industry funding is far preferable to undisclosed. It does mean the incentive to publish a null result is weak, and the field should be read with that in mind. Our guide to reading a clinical trial covers how to find the funding statement and what to do with it.
What does not exist, at the time of writing, is a large, long, independently funded randomised trial in older adults with a pre-registered clinical or functional primary endpoint. Until one reports, the accurate summary is that we know these compounds raise a blood measurement and do not know whether that helps anyone.
The limitations that hold the grade down
| Limitation | Why it matters for the grade |
|---|---|
| Surrogate primary endpoints | Blood NAD is a marker of exposure, not of benefit. Raising it confirms the pill was absorbed, nothing more. |
| Short duration | Ageing is a decades-long process. Weeks of exposure cannot test a claim about decades. |
| Small samples | Small trials are noisy in both directions. They generate hypotheses; they do not settle them. |
| Sponsor involvement | Publication and outcome-reporting incentives skew towards positive findings. |
| Healthy volunteers | People with the most room to improve are usually the least represented. |
| Product heterogeneity | Purity, dose and formulation differ across trials and across what is sold, so results do not transfer between products. |
There is also a regulatory wrinkle that readers in the United Kingdom should know about. NMN and NR do not occupy the same position in food law here, and the status of a given ingredient can change. A supplement being on sale is not a statement that it has been assessed and found effective, and in some cases it is not even a statement that it has been authorised. We set this out in full in how supplements are regulated in the UK.
What would change the grade
Grade C would follow from consistent, independently replicated effects on a functional outcome in humans, even in trials of modest length: for example the same measure of physical function moving in the same direction in two or more separately funded randomised trials.
Grade B would require at least one adequately powered randomised trial, running for years rather than weeks, in an older population, with a pre-registered clinical or functional primary endpoint, showing a benefit that survives the pre-specified analysis.
Grade A would require that result to be replicated by an independent group in a different population, with the effect direction and rough magnitude holding.
None of that is unreachable. It is simply expensive and slow, and the commercial return on selling the compound today does not depend on any of it happening. That asymmetry, more than any scientific difficulty, is why this literature has stalled at the stage it has.