NMN: what the human trials actually found

NMN is among the best-selling supplements in the longevity category, and it is also one of the few with enough randomised human trials behind it to be pooled and analysed. That combination is unusual and makes it a useful test case: the evidence has reached the top of the ladder, and it is worth seeing what it says when it gets there.

What the trials were

A 2026 systematic review and meta-analysis published in Nutrients searched six databases and included 15 parallel randomised trials comparing oral NMN or related preparations against placebo or an equivalent control. Ten of them contributed to the safety analyses. Doses ranged from 250 to 2,000 mg a day, and durations from 14 days to 24 weeks. The authors used random-effects models and graded the certainty of the evidence with GRADE, and registered the review in advance.

That is a well-conducted review of a real body of trials. It is worth stating plainly, because the honest criticism of NMN research is not that it does not exist.

What it found on safety

NMN did not increase overall adverse events, serious adverse events, withdrawals due to adverse events, or system-specific adverse events. It did not significantly raise ALT or AST, the two liver enzymes usually watched first when a new compound is taken daily.

The authors’ conclusion on this point is that short-term oral NMN showed favourable tolerability, with no clear increase in adverse events or hepatic biochemical abnormalities.

This is a genuine finding and the strongest thing that can currently be said for the supplement. It also has an expiry date attached: the longest trial pooled ran 24 weeks, so nothing here speaks to taking NMN for years.

What it found on benefit

No significant effects were observed on body weight, BMI, fasting glucose, HbA1c, lipid profiles or systolic blood pressure.

That sentence covers most of what NMN is marketed for. Metabolic health, weight, blood sugar control and cholesterol are the outcomes that appear on product pages, and pooled across 15 randomised trials they did not move.

Two measures were not flat. Diastolic blood pressure decreased slightly, and HOMA-IR, a measure of insulin resistance, showed a non-significant downward trend. The authors describe these as preliminary vascular-metabolic signals, particularly in older adults or people with early metabolic risk, and call for larger and longer trials to confirm efficacy and long-term safety.

A second 2026 meta-analysis in the same journal examined blood pressure specifically, pooling 10 randomised trials with 11 intervention arms and 349 participants in adults with elevated blood pressure. It found a statistically significant but modest reduction in resting diastolic pressure of 2.15 mmHg, and no significant reduction in systolic pressure.

How to read a result like this

A meta-analysis of randomised trials is the strongest form of evidence in this field, and the instinct on reading a null result from one is to assume the question is settled. That instinct is half right.

What is settled is narrow: over periods up to six months, at these doses, in these populations, NMN did not change these measures. What is not settled is what happens over years, which is the timescale the supplement is actually sold on. No trial has run that long, and a compound aimed at aging would not necessarily reveal itself in half a year.

The other thing worth noticing is that every measure here is a surrogate endpoint. Even had fasting glucose improved, that would be a marker moving, not a demonstration of a longer or healthier life. The distance between the two is the same distance that separates almost all longevity claims from what has been shown.

The Interventions Testing Program offers a useful corrective in both directions. It has tested dozens of compounds in mice under conditions far more controlled than any human trial, and most did nothing; it has also shown, in a 2024 reanalysis, that a compound can look inert under one statistical test and active under another. Confidence in either direction is expensive.

Where the evidence stands

Claim	What the pooled trials show
Raises NAD+	Supported; not in dispute
Safe over weeks to months	Supported: no increase in adverse events or liver enzymes to 24 weeks
Improves blood sugar control	Not supported: no significant effect on fasting glucose or HbA1c
Improves cholesterol	Not supported: no significant effect on lipid profiles
Aids weight loss	Not supported: no significant effect on weight or BMI
Lowers blood pressure	Partly: diastolic down about 2 mmHg; systolic not significant
Extends life or healthspan	Never measured in any trial

The summary that fits the data: NMN does what its chemistry says it does, is well tolerated for as long as anyone has studied it, and has not been shown to deliver the outcomes it is bought for. Whether longer trials would change that is an open question, and an open question is not a reason to assume either answer.