In rodents, NAD⁺ precursors behave enough like training that researchers started calling them exercise mimetics. In humans, the trials that measured exercise and muscle outcomes have mostly come back null — including the two that took muscle biopsies and looked directly at the mitochondria the whole theory rests on.
There is one real positive, and it is narrower and stranger than the headlines it produced. Working through both is the fastest way to understand why this category’s outcome evidence keeps disappointing people who read the animal work first.
The trial that found something
Forty-eight recreationally trained runners from a Guangzhou running team, training five or six times a week, were randomized into four arms for six weeks: placebo, or NMN at 300, 600 or 1,200 mg a day. Cardiopulmonary exercise testing was done at baseline and at six weeks.
Oxygen uptake, percentage of maximum oxygen uptake, power at the first ventilatory threshold and power at the second ventilatory threshold all increased more in the medium and high dose groups than in the control group. And then the sentence the coverage dropped: there was no difference in VO₂max, oxygen pulse, oxygen uptake related to work rate, or peak power in any group [1].
That is a submaximal result, not a ceiling result. Something shifted in how these runners used oxygen at threshold intensities; nothing shifted in their maximum. The authors attribute it to enhanced oxygen utilization in skeletal muscle, which is a reasonable reading of their own data and a long way from “NMN improves fitness.” It is also, so far, one trial in trained athletes at doses at the upper end of the published NMN range.
What happened when someone biopsied the muscle
The mechanism on every marketing page is mitochondrial. It has been tested twice, directly, and it did not hold.
In a randomized, placebo-controlled trial, 40 middle-aged obese and insulin-resistant men took 1,000 mg of NR or placebo twice daily for twelve weeks, with skeletal muscle biopsies before and after. Mitochondrial respiratory capacity was measured by high-resolution respirometry on single muscle fibers, and mitochondrial fractional area and network morphology by laser scanning confocal microscopy in a subset. Neither respiratory capacity nor the abundance of mitochondrial-associated proteins was affected, and there were no changes in mitochondrial fractional area or network morphology. Protein levels of NAMPT, an essential NAD⁺ biosynthetic enzyme in muscle, fell by 14%, while steady-state muscle NAD⁺ was unchanged [2].
A crossover trial reached a similar place with a more interesting wrinkle. Thirteen healthy overweight or obese men and women took 1,000 mg of NR daily for six weeks against placebo, with clamps, magnetic resonance spectroscopy, biopsies and ex vivo mitochondrial measurement. Markers of NAD⁺ synthesis rose in muscle. Fat-free mass increased slightly (62.65% ± 2.49% versus 61.32% ± 2.58%, p = 0.02) and sleeping metabolic rate rose. Muscle acetylcarnitine rose, and — the one exercise-specific finding — the capacity to form acetylcarnitine during exercise was higher on NR than placebo (2.99 ± 0.30 versus 2.40 ± 0.33 mmol/kg wet weight, p = 0.01). No effect on insulin sensitivity, mitochondrial function, liver or muscle fat, cardiac energy status, blood pressure, inflammation or energy metabolism [3].
Acetylcarnitine buffering is a real piece of exercise physiology and it is plausibly connected to threshold performance. It is also a biochemical intermediate measured in thirteen people, and the same trial found the mitochondrial function it was supposed to reflect unchanged.
Strength, recovery and older muscle
The recovery claim has had a proper test. Thirty-two people aged 55 to 80 were randomized to 1,000 mg of NR with 200 mg of pterostilbene or matched placebo. Two weeks in, a substantial muscle injury was induced by electrically stimulated eccentric work — confirmed by myoglobin and creatine kinase release, soreness, edema and a fall in strength — and biopsies were taken before, two hours after, and at 2, 8 and 30 days. The injury produced a large demand for muscle stem cell recruitment. Supplementation did not affect stem cell content, proliferation or cell size, nor muscle fiber area, central nuclei or embryonic myosin heavy chain, and the authors concluded it does not improve recruitment or any other measure of recovery from injury in older adults [4].
At the higher NMN doses, the physical-performance picture is the same. Thirty overweight or obese adults aged 45 and over, randomized 2:1 to two tablets each containing 500 mg of microcrystalline β-NMN, twice daily, or to placebo for 28 days, saw substantial rises in circulating NAD and meaningful falls in body weight, diastolic blood pressure and cholesterol. Changes in muscle strength, muscle fatigability, aerobic capacity and stair-climbing power did not differ significantly between groups [5].
Pooled, it holds. A 2025 systematic review and meta-analysis of randomized trials in adults with a mean age above 60 found NMN had no significant effect on skeletal muscle index (mean difference −0.42, 95% CI −0.99 to 0.14, p = 0.14), handgrip strength, gait speed (mean difference −0.01, 95% CI −0.08 to 0.06, p = 0.79) or the five-time chair stand test, with the narrative synthesis adding no improvement in knee extension strength, short physical performance battery or thigh muscle mass. The authors concluded that current evidence does not support either precursor for preserving muscle mass and function in that age group [6].
The question that matters most: does it add to training?
Nobody buys this instead of exercising. They buy it on top. Two trials have tested precisely that, and both are worth knowing.
Fifty-four sedentary adults aged 55 and over with daytime systolic blood pressure of 130 mmHg or more were randomized to six weeks of NR 1,000 mg a day plus three supervised 30-minute walks a week, the same walking program with placebo, or NR alone. Adherence was 93% to the exercise and 90% to the supplement. NR plus exercise did not reduce systolic blood pressure more than placebo plus exercise — a mean change of 5.19 ± 13.2 mmHg against −2.71 ± 10.5 mmHg — with a trend toward greater reduction in pulse wave velocity and, in a post hoc analysis restricted to participants not on antihypertensive medication, a trend toward greater nighttime blood pressure reduction [7].
Separately, a pilot randomized trial put 20 childhood cancer survivors with prediabetes on a structured telehealth exercise program with or without NR for six weeks. Adherence was high and there were no severe adverse events attributable to the interventions; secondary endpoints did not differ significantly between the arms, while the myostatin decrease that did appear tracked the number of exercise sessions completed rather than the supplement [8].
Both are small and both were framed by their authors as pilots. Neither found the additive effect the rodent literature predicts, and in the second one the variable that moved was how much people trained.
The honest read for an athlete
If you are already training and considering an NAD⁺ precursor, the published case for it is one six-week trial in runners showing a submaximal threshold shift with no change in VO₂max, plus a biochemical buffering finding in thirteen people. Against that sit two biopsy studies finding no change in mitochondrial function, a null recovery trial, a null physical-performance trial, a null meta-analysis and two trials where adding it to supervised exercise added nothing measurable.
That is not a reason to expect harm — the tolerability record for oral precursors is reasonable as far as it goes. It is a reason to buy at a price that matches the evidence, which means a capsule rather than a vial, and to be unmoved by a seller quoting the runners trial without its second paragraph. If anyone offers you an infusion for recovery, note that no trial of that format has measured an exercise outcome at all.