NAD+, NMN, and NR: What the Human Clinical Trials Actually Show

NAD+ is the molecule the longevity field has spent the last decade arguing about. The biological case is real — NAD+ is an obligate substrate for sirtuins, PARPs, and CD38, and tissue NAD+ pools fall in cross-sectional human data with age — but the gap between the rodent results that built the popular narrative and the human clinical record is wider than most coverage admits. This article is a structured reading list of the human evidence: which precursor was given, at what dose, in which population, and what the trial actually measured.

Nothing here is a recommendation. NAD+ and its precursors are sold by PepMax for laboratory research use only. Treat the trials below as the published scientific record any researcher evaluating these molecules should be working from — not as off-label clinical guidance.

Why this molecule, why now

Nicotinamide adenine dinucleotide (NAD+) was identified by Harden and Young in 1906 as a dialyzable cofactor required for yeast fermentation, and characterized chemically by Warburg, von Euler-Chelpin, and others in the 1930s. For most of the twentieth century it was understood almost exclusively as a redox carrier — a molecule that cycles between oxidized (NAD+) and reduced (NADH) forms across glycolysis, the tricarboxylic acid cycle, fatty-acid β-oxidation, and the electron transport chain.

That framing is correct but incomplete. The modern resurgence of NAD+ as a research target follows three findings, none of them older than thirty years:

• NAD+ is consumed, not just cycled.Sirtuin deacetylases (SIRT1–7), poly(ADP-ribose) polymerases (PARPs), and CD38/SARM1 ecto-enzymes all hydrolyze NAD+ as a substrate, releasing nicotinamide. Cellular NAD+ pools must therefore be continuously replenished from precursors, and the rate of consumption depends on cellular state (DNA-damage load, inflammatory signaling, metabolic stress).
• NAD+ availability rate-limits sirtuin activity. Imai, Guarente, and colleagues established in the 2000s that the K_mof sirtuins for NAD+ sits near the physiological NAD+ concentration in many cellular compartments — meaning modest changes in NAD+ pool size translate directly into changes in deacetylation flux on histone and non-histone targets^[15].
• Tissue NAD+ pools fall with age. Multiple cross-sectional human and rodent datasets show declines in NAD+ in skin, skeletal muscle, brain, and whole blood with chronological age. Whether this decline is causal for any specific aspect of the aging phenotype, and whether restoring it via precursor supplementation produces a functional benefit at the human-population level, is the open question this article is about^[16].

These findings created the modern field. They did not create the clinical evidence that the public framing of the field implies.

The biology in one figure

Cellular NAD+ is biosynthesized through three pathways. The salvage pathway, which recycles nicotinamide released by NAD+-consuming enzymes back into NAD+ via the intermediates NMN and NaMN, accounts for the majority of NAD+ turnover in most tissues. The Preiss-Handler pathway converts dietary nicotinic acid (niacin) into NaMN. The de novo pathway synthesizes NAD+ from tryptophan via the kynurenine route. Each oral precursor used in clinical trials feeds these pathways at a different point.

The practical consequence: tissue-specific expression of NRK1/NRK2, NAMPT, and NAPRT determines which precursor each tissue can use most efficiently, and oral pharmacokinetics differ substantially between NR, NMN, NAM, and NA. A trial that supplements one precursor and assays NAD+ in whole blood is not interchangeable with a trial that supplements a different precursor and assays NAD+ in muscle biopsy.

A short field timeline

The precursor class at a glance

Five oral or parenteral entry points into the NAD+ pool have been studied in humans. The table below summarizes route, the trial that established blood-NAD+ pharmacokinetics for each, and the highest level of clinical evidence currently published.

Reading the precursor strategies

Five entry points exist not because the field has converged on the “right” precursor but because each tests a different pharmacological hypothesis.

• NR bypasses the rate-limiting NAMPT step of the salvage pathway. NRKs convert NR to NMN intracellularly, and the molecule has dose-dependent oral bioavailability characterized in humans by Trammell et al. and confirmed in subsequent trials. The strategic argument for NR is that it is the precursor with the deepest published human pharmacokinetic and chronic-safety record.
• NMNsits one enzymatic step from NAD+. The pharmacokinetic argument for NMN over NR is that it is closer to the product; the pharmacokinetic argument against NMN is that NMN is generally believed to be dephosphorylated to NR at the cell surface (by CD73 and related ecto-enzymes) before entering the cell — meaning orally administered NMN may, in the most strictly characterized tissues, deliver to the cell as NR. The Slc12a8 NMN transporter described in mouse small intestine adds tissue-specific complexity but does not resolve the broader question.
• NAMis the cheapest, longest-studied, and most regulatorily familiar precursor. It is the substrate for NAMPT, the rate-limiting salvage enzyme — which is the case both for and against it: NAM-driven NAD+ flux is gated by NAMPT expression, and high NAM concentrations are reported to inhibit sirtuins via product-inhibition feedback in cell-based assays. Whether that in-vitro feedback is quantitatively meaningful at oral-supplement doses in vivo is contested.
• NAenters the Preiss-Handler pathway via NAPRT, expressed at variable levels across tissues. Tissues low in NAPRT (notably the brain) cannot efficiently use NA to build NAD+. NA also activates GPR109A, producing the cutaneous flushing that limits its oral tolerability and that is mechanistically distinct from NAD+ biology. Most of NA’s clinical record is in lipid metabolism, not longevity.
• Intravenous NAD+bypasses the precursor pathways entirely — in principle. In practice, the Grant et al. pharmacokinetic study suggests that infused NAD+ is substantially metabolized in plasma to nicotinamide and adenosine-derived metabolites before tissue uptake, so the molecule that arrives at the tissue may be predominantly NAM. This is one of the central unresolved questions about IV NAD+: what fraction of the infused dose is delivered intact to any tissue.

Does NAD+ actually decline with age?

The headline claim of the modern NAD+ field — that NAD+ falls with age, and that restoring it can reverse age-associated decline — rests on a smaller and more heterogeneous human dataset than the popular framing implies.

Massudi et al. (2012) is the most-cited human cross-sectional dataset. The authors measured NAD+ in skin biopsies from 26 individuals across an age range of 20–87 years and reported a decline with age, alongside changes in oxidative-stress biomarkers^[13]. Subsequent cross-sectional and tissue-specific datasets have broadly confirmed the directionality of the effect in skin, skeletal muscle, brain, and whole blood, with varying magnitude.

Three caveats are load-bearing:

• The data are mostly cross-sectional.Within-individual longitudinal data — NAD+ measured in the same person across decades — is sparse. A cross-sectional decline can reflect cohort effects (different birth-year diets, different baseline metabolism) rather than within-individual aging.
• The magnitude is tissue-dependent.The frequently cited “NAD+ falls 50% by midlife” framing is not a stable result across tissues or assays. Skin and skeletal-muscle declines are the most consistently reported; whole-blood NAD+ declines are smaller in absolute magnitude and noisier.
• NAD+ flux is not the same as NAD+ pool size. A static measurement of NAD+ concentration does not capture the rate at which NAD+ is being consumed and replenished. Two tissues with the same NAD+ concentration but different turnover rates have very different NAD+-dependent biology, and few human studies measure flux.

Where the strongest evidence sits

Read against each other, the published trials cluster around a defined set of endpoints. The map below summarizes which precursor currently anchors the strongest published human evidence for each endpoint domain.

Nicotinamide riboside (NR)

NR is the precursor with the deepest published human pharmacokinetic and safety record. It is a phosphorylated nucleoside that bypasses NAMPT — widely cited as the rate-limiting step of the salvage pathway — and is converted to NMN by nicotinamide riboside kinases (NRK1 expressed broadly; NRK2 enriched in muscle, brain, and heart). The most-cited commercial form is NIAGEN (nicotinamide riboside chloride), which has been the test article in the majority of the published human trials.

The Trammell trial established that NR is orally bioavailable in humans and produces a dose-dependent rise in whole-blood NAD+ on a timescale and magnitude consistent with the NRK-driven mechanism characterized in mice^[1]. It is the foundational pharmacokinetic study for the entire modern precursor field.

The Martens trial extended the pharmacokinetic story into a 6-week chronic-dosing framework and added the first safety dataset of meaningful duration in older adults. The blood-pressure and aortic-stiffness signals were exploratory and underpowered, but they were the first human signal that chronic NR might reach a vascular endpoint. They motivated a series of larger CV-focused programs that have since reported^[2].

Subsequent NR trials have largely consolidated the safety profile. Conze et al. (2019) randomized 140 healthy overweight adults to NIAGEN 100, 300, or 600 mg/day vs. placebo for 8 weeks and reported sustained whole-blood NAD+ elevation across active doses with an adverse-event profile indistinguishable from placebo^[3].

Where NR has been moved into clinical-endpoint trials, the results are mixed. Dollerup et al. (2018) randomized 40 obese men to NR 1000 mg/day for 12 weeks and reported no improvement in insulin sensitivity, mitochondrial function, hepatic lipid content, or body composition versus placebo^[4]. The Brakedal NADPARK trial in Parkinson disease — the highest-profile clinical-endpoint trial of NR to date — reported increased cerebral NAD+ on 31P-MRS, signal on selected clinical and metabolic biomarkers in NR-responsive patients, and a generally favorable safety profile^[5].

Nicotinamide mononucleotide (NMN)

NMN is the immediate precursor of NAD+ in the salvage pathway. The pharmacological case for NMN is that it is one enzymatic step from NAD+ (via NMNAT). The pharmacological case against it — or, more carefully, the open question — is that NMN at the cell surface is widely reported to be dephosphorylated to NR by CD73 and related ecto-enzymes before entering most tissues, so the molecule that crosses the membrane in those tissues may be NR rather than NMN. The Slc12a8 NMN-transporter described in mouse small intestine complicates this picture in a tissue-specific way, but does not resolve it for most tissues.

Yoshino 2021 is the strongest single positive clinical-endpoint result in the entire precursor literature, and the first to publish in a top-tier general-science journal with a clamp-based primary endpoint^[6]. The trial is small, single-site, and limited to a specific population, but its design was rigorous and its primary endpoint was a defined physiological measurement rather than a surrogate marker.

Subsequent NMN trials have added safety, pharmacokinetic, and small-effect functional data. Igarashi et al. (2022) reported a 12-week trial of NMN 250 mg/day in 42 healthy Japanese men aged 65 and older, showing dose-dependent elevation of whole-blood NAD+ and small but measurable changes in muscle-function endpoints^[7]. Liao et al. (2021) reported a 6-week dose-finding trial of NMN (300, 600, 1200 mg/day) in 48 amateur runners, with dose-dependent improvements in aerobic capacity (V̇O₂thresholds) without effects on peak V̇O₂^[8].

The MIB-626 program represents the most pharmaceutically rigorous NMN development to date. MIB-626 is a microcrystalline polymorph of β-NMN developed by MetroBiotech. Pencina et al. (2023) reported a Phase 1 trial in middle-aged and older adults showing dose-dependent increases in circulating NAD+ and downstream metabolites at 1000 mg/day with an adverse-event profile comparable to placebo^[9]. The MIB-626 program is one of the few NMN programs being conducted under an FDA IND, with the ambition of moving NMN from supplement to investigational drug.

Intravenous NAD+

Intravenous NAD+ is administered in private clinics in the United States and elsewhere for indications ranging from substance-use disorder to general “anti-aging” protocols. The published clinical record is dramatically thinner than the commercial availability suggests.

The Grant pilot is the only published human pharmacokinetic study of IV NAD+ infusion^[10]. Its central finding — that infused NAD+ does not produce a clean rise in plasma NAD+ and is instead substantially metabolized in circulation — is the foundational data point any honest evaluation of IV NAD+ has to start from. It does not rule out tissue-level effects (the metabolites NAM and the adenosine-derived fragments may themselves contribute to NAD+ pools after cellular uptake), but it does mean that “IV NAD+ raises NAD+” is not a claim supported by the available human pharmacokinetic data the way it is for oral NR or NMN.

Nicotinamide (NAM)

Nicotinamide is the oldest, cheapest, and most regulatorily familiar entry point into the NAD+ pathway. It is the salvage substrate for NAMPT, the rate-limiting enzyme of the salvage pathway. NAM has been used clinically since the resolution of the pellagra epidemic in the 1930s, and is the precursor with the largest controlled-trial outcomes record — one positive Phase 3 trial in skin-cancer chemoprevention, and one large negative Phase 3 trial in type 1 diabetes prevention.

ONTRAC is the strongest single Phase 3 result in the NAD+-precursor literature, full stop^[11]. The mechanism is generally interpreted as PARP-mediated DNA-repair support and ATP restoration after UV damage rather than a sirtuin-driven anti-aging effect. ONTRAC is an excellent example of how an intervention on the NAD+-precursor pathway can produce a meaningful, reproducible clinical benefit in a well-defined indication — and a reminder that “NAD+ pathway” and “longevity” are not the same thing.

ENDIT is the largest negative NAD+-precursor outcomes trial ever conducted^[12]. It was designed on the basis of mouse and observational data suggesting nicotinamide’s PARP-modulating effects might preserve β-cell function in pre-symptomatic type 1 diabetes. It did not. ENDIT is a useful corrective in any reading of the NAD+-precursor literature: the precursor pathway is broad and biological, but a positive mouse phenotype does not translate to a positive human outcome trial in the absence of indication-specific design.

Nicotinic acid (niacin)

Nicotinic acid (NA, niacin) enters NAD+ biosynthesis through the Preiss-Handler pathway via NAPRT. Most of NA’s clinical record is in lipid metabolism rather than longevity: NA at gram-scale doses lowers LDL cholesterol and triglycerides and raises HDL cholesterol via mechanisms partly independent of NAD+ biology, including activation of the GPR109A receptor. The two largest cardiovascular outcomes trials of NA on top of statin therapy — AIM-HIGH (2011) and HPS2-THRIVE (2014) — failed to demonstrate cardiovascular event benefit and reported increased adverse events^[17].

For the NAD+-precursor reader, NA is mainly a reminder that broad activation of a pathway does not equal benefit on every plausible endpoint. NA reaches NAD+ via a route that requires NAPRT — meaning tissues with low NAPRT expression (most notably the brain) cannot efficiently use NA to build NAD+. This is one of the reasons NR and NMN are typically preferred over NA for any NAD+-as-aging-substrate hypothesis.

Patterns across the precursors

• Oral precursors reliably raise blood NAD+. NR and NMN both produce dose-dependent, reproducible elevation in whole-blood NAD+ across multiple independent trials. The pharmacokinetic story is settled. Whether and where this elevation translates into a clinical-endpoint effect is a different and harder question.
• Tolerability is qualitatively similar across the oral precursors. NR, NMN, and modest doses of NAM all show adverse-event profiles indistinguishable from placebo in the published trials. NA is the exception: cutaneous flushing at gram-scale doses is dose-related and substantial, and is mediated by GPR109A rather than NAD+ biology.
• The strongest clinical-endpoint result is in skin cancer.ONTRAC (Chen 2015) is the highest-quality positive trial in the entire precursor literature. It is a useful counter to the framing that NAD+ precursors are an aging intervention: NAM’s clinically established benefit is in a defined, mechanistically explicable oncology-prevention indication, not in slowing chronological aging.
• The lifespan claim is not yet a human claim. No randomized human trial of any NAD+ precursor has reported a mortality or all-cause-aging endpoint. The rodent lifespan-extension data is real but has not been translated to human-population outcomes; the trials that would test that translation have not yet been run at the scale and duration required.

Open questions

The field is moving rapidly but several questions remain consequential for how this literature is interpreted.

• Cross-precursor comparative effectiveness. The single most useful missing trial is a randomized head-to-head comparison of NR vs. NMN at equimolar doses, with parallel pharmacokinetic and tissue-NAD+ readouts. Most cross-precursor claims in the broader literature are inferences from independent trials with different designs and different assays.
• Tissue-specific delivery. Most published trials measure whole-blood NAD+. The biologically interesting compartments (muscle, brain, liver, heart) have tissue-specific NRK, NAMPT, and NAPRT expression that determines which precursor each tissue can use. Trials with tissue NAD+ readouts are operationally hard but produce much more interpretable data.
• The NMN-to-NR conversion question. Whether orally administered NMN actually reaches cells as NMN, or is dephosphorylated to NR at the cell surface and enters as NR, is unresolved for most tissues. The Slc12a8 transporter is part of the answer in mouse small intestine; the broader resolution awaits more targeted human isotope-tracer studies.
• IV NAD+ pharmacokinetics and indication.The Grant pilot is the only published human study, and it is a pharmacokinetic pilot rather than a clinical-endpoint trial. The gap between IV NAD+’s commercial availability and its controlled-trial evidence base is the largest evidence gap in the entire NAD+-precursor space.
• The aging-endpoint translation.Rodent lifespan-extension claims for NAD+ precursors are real, but they have not been translated to a human-outcomes trial. The trials that would do so — multi-year, large-N, with hard endpoints — are not the trials currently being run in the supplement-driven part of the field.

Further reading

Each citation below links to the source paper. For methods context on what an HPLC purity number actually means for any compound in this class, see what ≥99% purity actually means and how we verify peptide purity. For the related mitochondrial-derived peptide reading, see our MOTS-c compound profile.