NMN Dosage: What the Clinical Trials Actually Used (100mg vs 250mg vs 600mg vs 900mg)

Clinical trials of NMN in humans have used oral doses ranging from 100 mg to 900 mg per day. The most-cited trials — Irie 2016, Yoshino 2021, Okuyama 2022, Yamaguchi 2022, Akasaki 2022, Kim 2022, Igarashi 2022, Pencina 2023, Katayoshi 2023 — show dose-dependent increases in blood NAD+ across this range, but the evidence for clinical outcomes beyond the NAD+ biomarker remains uneven, and no trial has established an “optimal” dose for any defined endpoint.

What doses have clinical trials actually used?

Human NMN research is a young field. The first first-in-human pharmacokinetic study was published in 2016. Since then, fewer than two dozen placebo-controlled or dose-ranging trials have reached peer review. Doses have clustered at four levels — roughly 100-250 mg, 250-300 mg, 500-600 mg, and 900 mg — reflecting the field's gradual escalation from tolerability to efficacy to safety characterization at higher doses.

The table below summarizes the doses, sample sizes, durations, and primary outcomes across the published trials most frequently cited in the literature and consumer discourse. It is not a complete registry of all NMN human trials — it focuses on the studies with peer-reviewed publication and clear dose documentation.

Dose-response summary across published trials

What does the dose-response curve look like?

The chart below plots the approximate peak blood NAD+ fold-increase reported across the dose range, drawn from the summaries in Irie 2020, Pencina 2023, and Yamaguchi 2022. Published point estimates vary by assay (LC-MS vs enzymatic), compartment (whole-blood vs PBMC vs plasma), and timing of measurement, so these values are approximate and should be read as indicating a trend rather than precise effect sizes.

What did each trial actually find?

Irie et al. 2016/2020 — first-in-human single dose (100-500 mg)

Irie et al. (2020, Endocrine Journal, PMID: 32357765) reported single-dose administration of 100, 250, and 500 mg NMN in ten healthy men. The trial did not measure blood NAD+ directly as its primary endpoint — it was designed to establish tolerability and pharmacokinetics. No serious adverse events were reported. NMN metabolites appeared in plasma in a dose-dependent pattern. The study established that a single oral NMN dose up to 500 mg is broadly tolerable in healthy men (evidence grade: emerging — single small study, short duration, narrow population).

Yoshino et al. 2021 — 250 mg, 10 weeks, mechanistic positive result

Yoshino et al. (2021, Science, PMID: 33888596) remains the most widely cited human NMN trial. Twenty-five postmenopausal women with prediabetes received 250 mg/day NMN or placebo for 10 weeks. Muscle insulin sensitivity (the primary endpoint, measured by hyperinsulinemic-euglycemic clamp) improved significantly in the NMN arm but not in placebo. Gene expression changes in skeletal muscle pointed toward increased muscle remodeling activity. Our Yoshino trial summary covers the effect sizes and limitations in more detail.

The caveats are essential: n=25, a single population (postmenopausal, prediabetic, BMI 25-42), a single dose, and a single center. The effect was statistically significant but modest in absolute terms, and the trial did not measure long-term metabolic outcomes or body composition. Yoshino is best read as proof that a specific mechanism can be engaged in humans with 250 mg/day NMN — not as proof that NMN improves diabetes risk.

Okuyama et al. 2022 — 250 mg evening dose, sleep and fatigue subgroups

Okuyama et al. (2022, Nutrients, PMID: 35817887) tested 250 mg NMN taken in the afternoon or evening in older adults for 12 weeks. Primary outcomes focused on drowsiness and fatigue patterns. Evening-dosed NMN was associated with modest reductions in daytime drowsiness in subgroups. The design did not include a morning-dose arm as an active comparator, and population was small, so “evening is better” interpretations are premature (evidence grade: emerging).

Yamaguchi et al. 2022 — up to 900 mg, safety and dose-dependent NAD+ rise

Yamaguchi et al. (2022, Endocrine Journal, PMID: 35922840) is the key reference for tolerability of higher doses. Healthy older adults (mean age ~65) received NMN titrated up to 900 mg/day over 12 weeks. Whole-blood NAD+ rose in a dose-dependent fashion. No serious adverse events were reported. Mild, transient events — gastrointestinal symptoms, headache — occurred at rates comparable to placebo. This trial is routinely cited to support the proposition that doses in the 600-900 mg range are tolerable over a medium-duration window.

Akasaki et al. 2022 — 250 mg, largest NMN trial to date

Akasaki et al. (2022, Journal of Medical Investigation, PMID: 36543531) enrolled 108 older adults to receive 250 mg/day NMN or placebo for 12 weeks. Physical-function endpoints — walking speed, grip strength, sit-to-stand — showed improvements in subgroup analyses. The trial is notable for its sample size relative to most NMN studies, but the effect sizes on functional endpoints were modest and specific improvements were not uniform across measures.

Igarashi et al. 2022 — 250 mg, aging markers

Igarashi et al. (2022, NPJ Aging, PMID: 36482258) reported a 12-week, 250 mg/day trial in older adults evaluating aging-related markers. Outcomes included biological-age proxies and reported eye-related fatigue. Reported improvements were modest and the primary biomarker shifts were consistent with blood NAD+ elevation. Like other 250 mg trials, the outcome effect sizes are small and the population is narrow.

Kim et al. 2022 — 250 and 500 mg, walking distance and fatigue

Kim et al. (2022, Frontiers in Aging, PMID: 36745179) is the closest published trial to a formal dose comparison at the low end of the dose range. Eighty older adults were randomized to placebo, 250 mg, or 500 mg NMN for 60 days. Six-minute walking distance improved in both NMN arms relative to placebo, with a larger effect at 500 mg. Blood NAD+ rose in a dose-dependent fashion. Short duration (60 days) and older-adult-specific population limit direct generalization; safety laboratory panels showed no clinically meaningful changes at either dose.

Pencina et al. 2023 — 300, 600, 900 mg direct dose comparison

Pencina et al. (2023, Journal of Nutrition, PMID: 36796762) remains the cleanest dose-ranging study in the NMN literature at this writing. Eighty healthy middle-aged adults were randomized to placebo, 300 mg, 600 mg, or 900 mg/day NMN for 14 days. Blood NAD+ rose significantly at all three doses relative to placebo. The response was dose-dependent but non-linear: the absolute NAD+ elevation at 600 mg was meaningfully higher than at 300 mg, while 900 mg produced only a modest further increment above 600 mg.

Pencina is a short-duration pharmacokinetic trial, not a clinical-outcome trial. It cannot answer whether 600 mg produces better functional endpoints than 300 mg — only that it produces more blood NAD+. Whether additional NAD+ at the biomarker level translates to additional biological effect remains an open question, and is the central scientific limitation of the current dose-response picture.

Katayoshi et al. 2023 — 125 and 250 mg, arterial stiffness

Katayoshi et al. (2023) reported a 12-week trial of 125 and 250 mg/day NMN in older adults evaluating arterial stiffness and cardiovascular markers. A modest improvement in arterial stiffness was observed in the higher-dose arm. Our Katayoshi trial summary covers the details. Cardiovascular endpoints in NMN trials remain exploratory, and single-trial effects have not been replicated.

What does the safety data across these doses actually show?

Safety across the published NMN literature has been consistently favorable at studied doses and durations (evidence grade: moderate for short- to medium-duration safety; limited for long-term). No trial has reported serious adverse events attributed to NMN. The most commonly reported mild events are gastrointestinal — nausea, transient abdominal discomfort — occurring at rates comparable to placebo in controlled trials.

Kim et al. (2022) specifically examined safety parameters across 250 and 500 mg doses and found no clinically meaningful changes in hepatic, renal, or hematologic safety labs. Yamaguchi et al. (2022) examined up to 900 mg/day for 12 weeks and reported similarly clean safety labs. Pencina 2023 reinforced short-term safety at 900 mg over 14 days. Long-term safety (multi-year daily use) has not been characterized in peer-reviewed human data.

How do NMN doses compare to NR dosing in trials?

NR human trials have typically used higher milligram doses than NMN trials. Martens et al. (2018, Nature Communications, PMID: 29599478) used 1,000 mg/day for six weeks. Dollerup et al. (2018, AJCN, PMID: 29992272) used 2,000 mg/day for 12 weeks. Conze et al. (2019, Sci Rep, PMID: 31316207) tested NR at up to 2,000 mg/day across 8 weeks. Elhassan et al. (2019, Cell Reports, PMID: 31412242) administered 1,000 mg/day for 21 days and is the only trial to document tissue-level NAD+ elevation via muscle biopsy.

A direct milligram-for-milligram comparison between NR and NMN is not scientifically meaningful. The two compounds have different molecular weights (NR 255 g/mol vs NMN 334 g/mol), enter the salvage pathway at different points, and have different pharmacokinetic profiles. More importantly, no head-to-head randomized trial has compared them at equivalent molar doses in the same population. Our detailed NR vs NMN comparison covers why indirect cross-trial comparisons are weak evidence for choosing between precursors.

What the dose-response data does not tell you

Three limitations of the current literature matter for anyone trying to interpret what a “good” dose is from published trials.

Blood NAD+ is a biomarker, not an outcome. Every dose from 100 mg upward has been shown to raise blood NAD+ relative to placebo in at least one trial. Whether the additional NAD+ at higher doses translates to additional functional benefit is only partially answered. Kim 2022 hints at a dose-response in walking distance, but the trial was small and 60 days is short. Pencina 2023 established a dose-response in NAD+ but did not measure functional endpoints.

Tissue NAD+ has not been measured in humans for NMN. Elhassan 2019 is the only precursor trial to directly measure tissue-level NAD+ elevation (in skeletal muscle, with NR). Tissue-level NAD+ decline with age is the endpoint that ultimately matters for any longevity hypothesis. Assuming blood NAD+ elevation implies tissue NAD+ elevation is an extrapolation, not an observation, in the NMN literature.

Populations are narrow. Most published NMN trials are in older adults (often 55-80), in Japanese or Korean populations, with specific health profiles. The applicability of dose-response data from a 65-year-old postmenopausal Japanese woman to a 35-year-old in any other population is unknown. No trial has systematically varied population characteristics to evaluate dose-response in subgroups.

How do trial doses compare to commercial product labels?

Commercial NMN products commonly label 250 mg, 500 mg, or 1,000 mg per serving — driven in part by consumer expectations anchored to the Yoshino 250 mg finding and the Yamaguchi 900 mg ceiling. The relationship between label dose and trial evidence is imperfect. A 500 mg label has weaker evidence for clinical endpoints than a 250 mg label (fewer trials at 500 mg with functional endpoints), even though 500 mg produces more blood NAD+ elevation than 250 mg in direct comparisons. The precursor comparison matrix lays out the regulatory status and evidence bases side-by-side without brand or label framing.

Bottom line on NMN trial doses

Published clinical trials of NMN in humans have used doses from 100 mg to 900 mg/day for durations of 14 days to 12 weeks in populations ranging from 10 to 108 participants. Every dose in this range raises blood NAD+ relative to placebo in at least one trial. The dose-response curve at the blood NAD+ biomarker level is approximately monotonic but non-linear — larger incremental gains below 600 mg than above. Clinical-endpoint evidence is uneven: the strongest single result (muscle insulin sensitivity in postmenopausal prediabetic women) comes from 250 mg. Safety data at 900 mg is short- to medium-duration but consistent.

What the literature does not support: a specific recommended dose for any clinical endpoint, a direct comparison with NR at equivalent molar doses, tissue-level NAD+ elevation in humans for NMN, or long-term (multi-year) safety at any dose. Interpreting trial doses as consumer recommendations requires inferential steps that the underlying evidence does not yet support.

For context on why NAD+ restoration strategies matter and where the broader precursor landscape stands, our primer on NAD+ decline with age covers the three mechanisms (CD38, NAMPT, PARP) that drive age-related NAD+ loss. For regulatory and bioavailability differences between NMN and NR, the head-to-head precursor comparison is the companion reference.