Does NMN actually do anything?

Most supplement pitches fall apart at the first link. This one does not, and that is exactly why it is worth your time.

The argument for NMN, nicotinamide mononucleotide, runs like this. NAD is a coenzyme your cells cannot function without. NAD levels decline with age. NMN is a direct biochemical precursor to NAD. Therefore, take NMN, restore NAD, slow down some part of aging.

Every one of those first three statements is true. That is unusual. Most of what I have written in this category involves a claim that snapped somewhere obvious, often a label that did not match the powder. NMN is not that. It is a real molecule with a real mechanism, and it is the most instructive product in the aisle precisely because the reasoning is so clean right up until the moment somebody checks the end of it.

Start with the strongest evidence for the pitch, because it is real

In 2024 a team at the Chinese University of Hong Kong, Zhang, Poon and Wong, screened 4,049 records and pulled 12 randomized controlled trials covering 513 participants. It ran in Critical Reviews in Food Science and Nutrition, 2025, volume 65, issue 22, pages 4382 to 4400.

Their first result, verbatim:

Random-effects meta-analyses found an overall significant effect of NMN supplementation in elevating blood NAD levels.

That is a genuine win for the product and I want to hand it over without hedging. The molecule is absorbed. It does what its biochemistry says it should do at step one. A second, independent review found the same thing from a different set of trials: five of eight RCTs reported an increase in blood NAD following supplementation.

If the pitch were "NMN raises your NAD," the pitch would be substantiated. That is not the pitch anyone is actually buying.

The very next sentence

Here is what immediately follows in the same abstract, verbatim:

However, most of the clinically relevant outcomes were not significantly different between NMN supplementation and control group.

And their overall conclusion, also verbatim: "our findings suggest that an exaggeration of the benefits of NMN supplementation may exist in the field."

That is not a skeptical blogger. That is a peer-reviewed systematic review of the entire literature, in a journal with a serious impact factor, stating that the category oversells itself. It is about as close to an official verdict as this industry ever produces.

Their risk-of-bias assessment is worth one more line. Using RoB2, the standard Cochrane tool, seven of the twelve studies raised "some concerns" and the other five were rated high risk. Not one of the twelve came out clean.

The independent replication of the null

A separate group, Chen and colleagues, published in Current Diabetes Reports in 2024 (volume 25, issue 1, article 4). Eight RCTs, 342 middle-aged and older adults, 49 percent female and mainly non-diabetic. Doses from 250 to 2,000 milligrams per day. Durations from 14 days to 12 weeks. Verbatim:

The random-effects meta-analyses indicated no significant benefit of NMN on fasting glucose, fasting insulin, glycated hemoglobin, homeostatic model assessment for insulin resistance and lipid profile.

One outcome came close. Insulin resistance, measured by HOMA-IR, showed a reduction across three studies at p = 0.06, and it did not survive their sensitivity analysis. I point that out because reporting a near-miss honestly, rather than reframing it as a "trend toward significance" and building a paragraph on it, is the clearest sign you are reading authors who were not fishing.

Their conclusion, verbatim: "Based on the small number of RCTs involving mainly relatively healthy adults, short-term supplementation of NMN of 250-2000 mg/d did not show significantly positive impacts on glucose control and lipid profile." The review declared no funding from any agency.

Two separate teams, different databases, partly different trials, same answer. The marker moves. Nothing downstream does.

What just happened, in formulator's terms

This is the surrogate endpoint problem, and if you take one idea from this report, take this one, because it will serve you across every shelf in the store.

There is a chain between a capsule and a benefit, and it has more links than the marketing implies:

dose taken → dose absorbed → blood NAD → tissue NAD → some cellular process changes → something you would actually notice

Each link is harder and more expensive to measure than the one before it. Blood NAD needs a blood draw and an assay. "Something you would actually notice" needs a long trial, a large sample, a validated outcome measure, and a budget with a comma in it.

So the industry measures the cheapest link and reports it as though the chain were welded shut. Blood NAD is that cheapest link. It is also the least informative one, because it only tells you the molecule got in. It says nothing whatsoever about whether anything past it moved.

I have a version of this from the instrument side. If you turn up the gain on a detector, your peak gets taller. The peak is now unmistakably bigger and you have learned exactly nothing new about the sample, because you changed the measurement rather than the thing being measured. A biomarker that rises because you swallowed that biomarker's direct precursor sits uncomfortably close to that. It confirms delivery. It is not evidence of effect, and it was never designed to be.

A surrogate endpoint earns the right to stand in for a real one only when somebody has demonstrated that moving it reliably moves the real outcome too. For blood NAD and human aging, nobody has demonstrated that. The surrogate is being used on credit.

The best case for the other side

I do not want to leave this as a demolition, because there is a paper people cite for the positive view and it deserves to be on the page rather than left out.

Wen and colleagues, at California Northstate University, published a systematic review in Cureus on 1 August 2024: 10 RCTs, 437 patients, mean age 58, doses from 150 to 1,200 milligrams per day, mean follow-up 9.6 weeks. It reports a significant gait-speed increase in one study, six-minute walking distance improvements versus placebo at the higher doses of 600 to 900 milligrams, quality-of-life scores that beat placebo across all three NMN arms of one trial, and improved ventilatory-threshold measures in athletes at higher doses.

If you are going to argue for NMN, argue from that. Then read its own summary sentence, verbatim: "patients taking NMN supplementation demonstrated non-significantly improved physical performance parameters." And its conclusion, also verbatim: "NMN supplementation is associated with a nonsignificant improvement in the physical strength and aerobic performance and is well tolerated with no serious adverse effects. However, body composition and muscle mass are not significantly affected."

Two things need saying plainly. First, the strongest available case for NMN describes its own headline finding as nonsignificant, in its own words, twice. Second, I am flagging that Cureus is a rapid-publication journal with a lighter review process than the other two sources here, and that this was a descriptive review pooling means rather than a formal meta-analysis. I am citing it because it is the best version of the opposing argument, not because it settles anything.

And it does establish one thing that matters and that I am not going to bury: across 437 patients, no serious adverse effects, with the reported side effects judged independent of the supplement. At these doses and these durations, NMN looks well tolerated. That is a real finding. "Probably safe" and "shown to work" are simply different claims, and only one of them has support here.

Everything here is short

Look at the durations again. Fourteen days to 12 weeks in one review. A mean of 9.6 weeks in the other. The longest trial anywhere in this evidence base runs three months.

This is a product sold on aging. Aging is not a 12-week phenomenon. Nobody has run the trial the claim implies, and the honest sentence for what happens over years is not "it works" and not "it doesn't." It is "nobody has looked."

The regulatory news that got read backwards

There is a piece of NMN history that gets cited as vindication and does not mean what people think.

In November 2022 the FDA announced that NMN could not be sold as a dietary supplement. The mechanism was the drug-preclusion clause, section 201(ff)(3)(B) of the Federal Food, Drug, and Cosmetic Act, sometimes called the race-to-market provision: if a substance was authorized for investigation as a new drug before it was marketed as a supplement, it is excluded from the supplement category. NMN had been taken into drug investigation, so off the shelves it went.

Then, in letters dated 29 September 2025 signed by Donald Prater, the agency reversed itself. Verbatim: "We now conclude that NMN is not excluded from the definition of dietary supplement under section 201(ff)(3)(B)." The FDA had revised its reading of the race-to-market clause and pointed to evidence that NMN was marketed as a supplement in the United States as early as 2017, before the drug authorization.

The industry celebrated, and fairly. It was a real legal win and it reopened a large market.

Here is the part that fell out of the retelling. That determination is about a date. It answers the question "was this on shelves before it became a drug candidate." It does not answer "does it work," and the FDA did not suggest otherwise. The agency does not evaluate supplements for efficacy in the first place, and its letters said nothing at all about whether NMN benefits anyone.

So "FDA declares NMN lawful" is entirely accurate and means considerably less than it sounds like. It is a ruling about 2017. This is the same gap I walked through in what "clinically proven" actually means: the phrase people hear and the standard behind it are not the same object.

What would change my mind

I want to be specific, because "more research is needed" is what people say when they do not want to commit to anything.

A trial running at least six months, not twelve weeks. In a population with room to improve, not healthy volunteers whose fasting glucose was already fine. On an outcome a person would notice from the inside, function or fatigue or capacity, rather than a lab value. Pre-registered, adequately powered, and not funded by the company selling the ingredient. That last one is not cynicism, it is the pattern I traced through the collagen literature, where the industry-funded studies found effects and the independent ones did not.

If NMN's mechanism is right, that trial should be winnable. Four years into a boom this size, the striking thing is that nobody has run it.

The signal

Nothing here is fraud. NMN is a real precursor, it raises the marker it is supposed to raise, and it appears safe at the doses studied. The mechanism is not made up.

Which is the whole lesson. This is a demonstration of how far a genuinely true mechanism can carry a product before anybody checks the far end of the chain, and the answer is: several years and a very large market.

So carry the rule out of here rather than the verdict on this one molecule. When a supplement's headline evidence is a biomarker, ask two questions. What was that biomarker supposed to predict? And did anyone measure that instead? If the answer to the second is no, you are not looking at evidence of a benefit. You are looking at a receipt proving the capsule dissolved.

Not medical advice. This is educational analysis, not a recommendation — a study is not a prescription. Talk to a qualified clinician before acting on anything you read here. Full disclaimer →

Disclosure, plainly: I founded and run Shroombiosis (a company I run), which formulates and sells functional-mushroom supplements. That is a direct commercial stake in the industry this report is about, and the surrogate-endpoint problem described here is one my own category commits routinely: mushroom marketing leans on immune markers and mouse data for exactly the same reason NMN marketing leans on blood NAD, because those are the cheap measurements. Apply this report's closing rule to anything I ever publish about my own products. Nothing here is sponsored and no link earns a commission; here's the full policy. A recommendation with no stake at all: for performance nutrition, Die Tryin Co. is a fellow combat-veteran-owned brand I'm glad to point people to. I don't own it and earn nothing from the link.