NAD+ (Nicotinamide Adenine Dinucleotide): An Evidence Reference

Verified against PubChem PUG-REST, CID 5892 (title: Nadide). Formula, molecular weight and CAS retrieved directly. The CAS 53-84-9 applies to the free acid / beta-NAD form; salt and hydrate forms carry different registry numbers and different masses. Fields marked “not verified” are ones we could not confirm against a primary chemical database — we leave them blank rather than guess.

Two distinct jobs, often conflated

NAD+ appears in the literature in two roles that are frequently merged into one, and merging them produces most of the confusion in secondary writing about the molecule.

In the first role it is a redox carrier. NAD+ accepts a hydride to become NADH and returns to the oxidised form, cycling continuously without net consumption. This turnover drives the electron-transfer steps of glycolysis, the tricarboxylic acid pathway, fatty acid oxidation and mitochondrial oxidative phosphorylation. The NAD+ to NADH ratio is used across the literature as a readout of a cell's redox and metabolic state (Covarrubias et al. 2021; Navas and Carnero 2021).

In the second role NAD+ is not recycled at all but cleaved and consumed. Three enzyme families do this: the sirtuins, which are NAD+-dependent protein deacylases; the poly(ADP-ribose) polymerases, which consume NAD+ during DNA-damage responses; and the glycohydrolases CD38 and CD157, which generate ADP-ribose and cyclic ADP-ribose.

The consequence of the second role is the whole reason biosynthesis matters. Because these enzymes destroy the molecule rather than pass it along, cellular pools must be continuously regenerated. The salvage route through NAMPT, the rate-limiting step, regenerates most NAD+ from nicotinamide, with the de novo route from tryptophan and the Preiss-Handler route from niacin contributing (Yoshino, Baur and Imai 2018).

What the largest human measurement of blood NAD+ found

The claim that human NAD+ falls by roughly half by middle age is repeated across the internet more often than almost any other biochemical figure. The largest human dataset located here does not support it.

Yang and colleagues measured whole-blood NAD+ by cycling assay and LC-mass spectrometry in 1,518 community-based adults from the Jidong cohort, aged over 18, free of cardiovascular disease and cancer, sampled between 2019 and 2020 (PMID 35388296). Mean age was 43.0 years and 52.6 percent were men. Mean whole-blood NAD+ was 33.0 plus or minus 5.5 micromolar.

The age association was small and sex-dependent. Across the full sample a declining trend appeared before age 50 and reached significance only in the 40-to-49 group, with an adjusted beta of -1.12 micromolar (95% CI -2.18 to -0.06). After 50 the trend disappeared. The interaction with sex was significant at p=0.003: in men the coefficient decreased with age and reached significance only in the 60-and-over group (beta -2.16, 95% CI -4.16 to -0.15), while in women whole-blood NAD+ did not differ significantly across any of the five age groups.

A decrement of roughly one to two micromolar against a mean of 33 micromolar, present in one sex and in one or two age bands, is a different quantity from a fifty percent lifespan collapse. Blood is also not tissue, and the authors are explicit that sex must be accounted for in future work.

Human interventional data studies precursors, not NAD+

An important structural feature of this evidence base: the randomised human trials located here administered NAD+ precursors, not NAD+ itself.

Remie and colleagues ran a randomised, double-blind, placebo-controlled crossover study of nicotinamide riboside at 1000 mg per day for six weeks in 13 healthy overweight or obese men and women, with hyperinsulinaemic-euglycaemic clamps, magnetic resonance spectroscopy and muscle biopsies (PMID 32320006, NCT02835664). Markers of NAD+ synthesis rose in skeletal muscle, confirming the compound reached its target. Fat-free mass increased from 61.32 plus or minus 2.58 percent to 62.65 plus or minus 2.49 percent (change 1.34 plus or minus 0.50, p=0.02), and muscle acetylcarnitine rose from 3025 plus or minus 316 to 4558 plus or minus 749 pmol/mg dry weight (p=0.04).

The null findings in the same trial are equally informative and are usually dropped from summaries: no effect was observed on insulin sensitivity, mitochondrial function, hepatic or intramyocellular lipid accumulation, cardiac energy status, cardiac ejection fraction, ambulatory blood pressure, plasma inflammation markers or energy metabolism.

A separate Phase 2a study, NCT05038488, tested the beta-NMN compound MIB-626 at 1.0 g orally twice daily for 14 days against placebo in 42 hospitalised adults with COVID-19 and stage 1 acute kidney injury, sponsored by Metro International Biotech at Brigham and Women's Hospital. It completed in August 2023. No data have been posted to the registry.

Solution stability is a buffer problem, not only a pH problem

NAD+ is hygroscopic as a solid and unstable in solution, and the alkaline lability is genuinely documented: in basic solution NAD+ reversibly forms a 370 nm-absorbing transient that irreversibly decomposes to a 340 nm-absorbing species, and supplier documentation records rapid degradation on heating and marked lability in alkaline solution, particularly with phosphate, maleate or carbonate present.

What the simple pH framing misses is that buffer identity can matter as much as pH. Wolfe and colleagues held nicotinamide cofactors in common aqueous buffers for 43 days at 19 degrees Celsius and pH 8.5 (PMID 39598842). In Tris, NAD+ showed only a 4 percent decrease in absorbance at 260 nm across the entire period. In HEPES at the same pH and temperature, the cofactor was almost entirely lost, with visible spectral red-shifting. Sodium phosphate produced intermediate degradation.

That is a large divergence between three buffers at one pH, and it means a stability figure quoted without naming the buffer is not interpretable. The authors also flag a methodological limit: direct quantification of NAD+ concentration was confounded by overlapping absorbance from degradation products, so their NAD+ analysis is qualitative rather than quantitative.

The specific storage figures that circulate in supplier documentation are addressed separately in the unsourced table below.

Where NAD+ appears in laboratory research

The contexts in which NAD+ appears in published work are broad, and listing them accurately is more useful than compressing them into a claim.

It functions as the reference coenzyme in enzymology, in NAD+/NADH-coupled dehydrogenase reactions and spectrophotometric kinetics, where its absorbance behaviour is the measurement itself rather than the subject of it.

It appears in ageing research as a cofactor whose tissue concentrations were reported to decline with age in model organisms, reviewed by Covarrubias and colleagues in Nature Reviews Molecular Cell Biology in 2021. That review is a synthesis of preclinical work, not a primary measurement.

It appears as substrate and regulator in sirtuin studies, as substrate in PARP and DNA-repair studies, and in CD38/CD157 and cyclic ADP-ribose signalling within immunometabolism.

In cancer metabolism it appears from the opposite direction. Yaku and colleagues reviewed how NAMPT amplification supports the glycolytic phenotype in several cancers and how NAMPT-specific inhibitors deplete NAD+ and suppress proliferation (PMID 30631755). Navas and Carnero covered the same territory alongside stemness and immune regulation. Here the research interest is in lowering NAD+ availability, not raising it, which is a useful corrective to the assumption that more is uniformly the direction of interest.

What is not known

The redox biochemistry of NAD+ is settled; almost everything downstream of it is not. No randomised human trial located here administered NAD+ itself; the human interventional evidence concerns precursors, principally nicotinamide riboside and beta-NMN, which enter the pool by different routes and with different pharmacokinetics. The largest human precursor trial found here enrolled 13 people for six weeks, which is small and short, and its clamp-based insulin-sensitivity endpoint was null. The MIB-626 Phase 2a completed in 2023 with 42 participants and has posted no data, so its findings are unavailable. Human tissue NAD+ across the lifespan has not been mapped in the way blood has; the 1,518-person dataset measured whole blood, which may not track muscle, liver, brain or adipose concentrations. Sex differences are documented but unexplained, and the one large human study reporting them found the age association in men only. Nothing located addresses long-term administration in any population, effects in people with existing disease other than the two small trials noted, interactions with medication, or what a sustained change in the NAD+ pool does to the consuming enzymes over time. The cancer-metabolism literature reviewed here explores lowering NAD+ availability, and how that intersects with efforts to raise it is not resolved in any source retrieved.

Questions