The short version: the category's foundational premise — that NAD+ declines with age — was tested in 2026 in over 300 people across seven cohorts and did not hold up in blood. Across roughly 31 human trials of NMN and NR, the reproducible finding is that a biomarker goes up. Clinical outcomes are mostly null. And in the two trials that measured both, blood NAD+ rose while the target tissue's NAD+ did not move at all.
Australia is now the first country to formally permit NMN in listed medicines, which makes this worth understanding properly rather than from a product page.
The three pathways, briefly
Your cells make NAD+ three ways:
- De novo (kynurenine) pathway — from tryptophan. Low flux, mainly liver.
- Preiss-Handler pathway — nicotinic acid (niacin) enters here, bypassing the rate-limiting enzyme NAMPT entirely.
- Salvage pathway — nicotinamide (NAM) enters here and is converted to NMN by NAMPT, which is the bottleneck. Nicotinamide riboside (NR) is phosphorylated to NMN by NRK1/2. NMN would enter at the NMN step — if it gets into the cell intact.
That last "if" is the entire marketing argument for NMN, and it is contested.
The claim: Grozio A, Mills KF, Yoshino J, et al., "Slc12a8 is a nicotinamide mononucleotide transporter," Nature Metabolism 1:47–57 (2019) — proposed a dedicated NMN transporter in mouse small intestine.
The pushback: Schmidt MS, Brenner C, "Absence of evidence that Slc12a8 encodes a nicotinamide mononucleotide transporter," Nature Metabolism 1:660–661 (2019) — argues the analytical methods and transport data "are not sound," notes prior evidence that NMN is dephosphorylated to NR before cellular entry (making a dedicated transporter unnecessary), and that Slc12a8 is annotated as a Na⁺/K⁺/Cl⁻ transporter.
The reply (Grozio et al., same journal, 662–665) defends the proteoliposome experiments but concedes that intracellular NMN doesn't accumulate in intact cells the way it does in proteoliposomes.
Status: unresolved. Note the conflict structure on both sides — Brenner is chief scientific adviser to ChromaDex, which sells NR; Imai's group has NMN commercial links.
And here's the part that matters most. Damgaard and Treebak's review (below) states that most orally administered NR is converted to NAM by CD157 in the small intestine, and partly to NA by gut bacteria, so "the primary circulating NAD+ precursors resulting from oral NR administration are NA and NAM." If the dephosphorylation evidence is right, the same logic applies to NMN.
You pay for the premium molecule; your tissues substantially receive cheap niacin derivatives.
Does NAD+ actually decline with age in humans?
This is the premise the entire category rests on. It is much weaker than the marketing implies, and 2026 made it weaker.
The strongest recent evidence says no
**Trętowicz MM, Scantlebery AML, Schomakers BV, et al., "Human whole-blood NAD+ levels do not vary with age or lifestyle interventions," Nature Metabolism 8(6) (2026).**
- n > 300 across seven independent cohorts — cross-sectional young vs old, a cardiac cohort, elite athletes, the Leiden Longevity Study (ages 63–87), an exercise intervention in frail older adults, and a twin-pair NR cohort as positive control
- Validated UHPLC–high-resolution mass spectrometry with stable-isotope internal standards
- In the cardiac cohort (late 20s to early 70s), age accounted for about 1% of the variation in NAD+
- In Leiden Longevity (63–87), the regression was flat: R² = 0.051, p = 0.991
- The assay detected NR supplementation as expected — so this is not a sensitivity failure
- Publicly funded. "The authors declare no competing interests."
What the pro-decline claim rests on
**Clement J, Wong M, Poljak A, Sachdev P, Braidy N, Rejuvenation Research (2019) — n=29, cross-sectional, plasma. Reported plasma NAD+ declining about 4% per year** (r² = 0.86, p = 4.4×10⁻¹³).
This is the study the industry's "NAD+ halves by middle age" claim traces back to. Twenty-nine people, cross-sectional, plasma rather than tissue, and an r² of 0.86 against a 4%/year slope is implausibly tight for a 29-person human cohort. It does not survive contact with the 300-person multi-cohort replication above.
Brain data exist (Zhu XH, et al., PNAS 2015, ³¹P-MRS at 7T, reporting age-dependent decline in brain NAD content) but the cohorts are small and the method was still being refined as recently as 2024.
The authoritative review agrees the premise is shaky. Nature Metabolism (2025), "NAD+ precursor supplementation in human ageing": age-related NAD+ decline "has been consistently observed only in a limited number of studies"; "the published body of data on NAD+ dynamics in human tissues remains sparse"; "the extrapolation of rodent-based data is not straightforward."
And even if it did decline, that wouldn't settle anything. As RMIT's Oliver Jones put it: even if NAD concentrations did decline with age, that would not mean the decrease in NAD caused the ageing.
What the human trials actually found
Two framing quotes before the detail.
**Gallagher C, Owoturo OE, "NAD⁺ supplementation for anti-aging and wellness: A PRISMA-guided systematic review," Ageing Research Reviews 116:103057 (April 2026):**
"Oral NR and NMN consistently demonstrated biochemical target engagement" yet "effects on functional, metabolic, vascular, and other healthspan-relevant outcomes were heterogeneous and often null or endpoint-specific." Conclusion: "NAD⁺ augmentation shows clear biological activity, but clinical effectiveness for anti-aging or wellness outcomes remains inconclusive."
**Damgaard MV, Treebak JT, "What is really known about the effects of nicotinamide riboside supplementation in humans," Science Advances 9(29), July 2023**, reviewing 25 trials:
"Collectively, oral nicotinamide riboside supplementation has displayed few clinically relevant effects, and there is an unfortunate tendency in the literature to exaggerate the importance and robustness of reported effects."
Nicotinamide riboside
| Study | n | Dose | Duration | Result |
|---|---|---|---|---|
| **Martens CR, et al. Nat Commun 2018** | 30 randomised, 24 completed | 1000 mg/day | 2 × 6-week crossover, healthy 55–79 | PBMC NAD+ +~60%. Blood pressure and arterial stiffness were NOT significant — systolic −3.9 mmHg, diastolic −2.0 mmHg, both NS after correction; pulse wave velocity NS. The famous "9 mmHg" is a post-hoc subgroup of 13 people with no statistical inference reported. |
| **Dollerup OL, et al. Am J Clin Nutr 2018** | 40 | 2000 mg/day | 12 weeks, obese sedentary men | Primary endpoint null. Insulin sensitivity by hyperinsulinaemic euglycaemic clamp "was not affected" (interaction p = 0.71). Also null: glucose production, energy expenditure, lipolysis, body composition, hepatic lipid. Follow-up analyses found skeletal muscle NAD+ showed "no clear change." COI: "CB serves as chief scientific adviser of ChromaDex, holds stock in ChromaDex, and has received research grants from ChromaDex." |
| **Elhassan YS, et al. Cell Reports 2019** | 12 | 1000 mg/day | 21 days, men 70–80 | Muscle NAD+ metabolome augmented, anti-inflammatory signature. Null: no change in muscle mitochondrial function, no improvement in glucose tolerance. |
| **Remie CME, et al. Am J Clin Nutr 2020** | 13 | 1000 mg/day | 6 weeks crossover | Insulin sensitivity unchanged; ectopic lipid unchanged. Small body-composition shifts at p = 0.02–0.05 in a 13-person crossover. |
| Conze D, Brenner C, Kruger CL. Sci Rep 2019 | 140 | 100–1000 mg | 8 weeks | Dose-dependent blood NAD+ rise; no meaningful clinical effects |
| Dellinger RW, et al. 2017 (NR + pterostilbene) | 120 | 250–500 mg | 8 weeks | NAD+ up; no meaningful clinical improvement; concerning LDL/cholesterol increases |
| Wang DD, et al. 2022 | 30 | up to 2000 mg | 12 weeks, heart failure | Blood NAD+ up; BP, glucose, potassium, renal and hepatic markers all unaffected |
| Brakedal B, et al. Cell Metab 2022 (NADPARK) | 30 | 1000 mg | 30 days, Parkinson's | Brain NAD up on MRS; clinical benefit required post-hoc subgroup analysis |
| **Berven H, et al. Nat Commun 2023 (NR-SAFE)** | 20 | 3000 mg/day | 4 weeks, Parkinson's | All completed, all adverse events mild, no significant difference vs placebo. Importantly: SAM, SAH, the SAM/SAH ratio and methionine were unchanged — see the methyl-donor section |
| Stocks B, et al. 2021 | 8 | 1000 mg | 7 days | No effect on metabolic rate, substrate use or exercise performance |
| Ahmadi A, et al. 2023 | 25 | — | 6 weeks, CKD | VO₂peak unchanged |
NMN
| Study | n | Dose | Duration | Result |
|---|---|---|---|---|
| **Yoshino M, Yoshino J, Kayser BD, et al. Science 2021** | 25 | 250 mg/day | 10 weeks, prediabetic postmenopausal women | Muscle insulin sensitivity +25%. Null: hepatic and adipose insulin sensitivity, body weight, body composition. And critically: PBMC NAD+ rose ~50% but muscle NAD+ was unchanged — the proposed mechanism did not move in the tissue where the effect appeared. (A formal peer-reviewed technical Comment on this paper was published in Science, with an author Response.) |
| **Igarashi M, et al. npj Aging 2022** | 42 | 250 mg/day | 12 weeks, Japanese men ≥65 | Blood NAD+ up. Gait speed p = 0.033, left grip p = 0.019 — both described as "nominally significant"; right grip not. Null: body composition, muscle mass, visceral fat, liver fat, HOMA-IR, HbA1c, glucose, arterial stiffness. Mitsubishi-funded. |
| **Katayoshi T, et al. Sci Rep 2023** | 36 | 250 mg/day | 12 weeks, healthy 40–59 | Primary vascular endpoint null — arterial stiffness between-group p = 0.097. Null: SIRT1 mRNA, urinary 8-OHdG, skin AGEs. A BMI subgroup was significant, post-hoc. "Supported by a research grant from the DHC Corporation" — which sells NMN. |
| **Kim M, Seol J, Sato T, et al. Nutrients 2022** | 108 | 250 mg/day | 12 weeks, Japanese ≥65 | Primary endpoints essentially null — no significant interaction for sleep quality. Significant only for drowsiness (p = 0.02) and sit-to-stand (p = 0.04). Null: grip strength, 5-m walk. Mitsubishi provided funds and supplements. |
| **Liao B, et al. J Int Soc Sports Nutr 2021** | 48 | 300/600/1200 mg | 6 weeks, recreational runners | VO₂max NULL — "no difference in VO₂max… in any of these groups." Also null: oxygen pulse, peak power, peak workload. The paper's title says "enhances aerobic capacity." Its own VO₂max result says otherwise. |
| **Pencina KM, et al. J Gerontol A 2023** | 32 | MIB-626 1000–2000 mg | 14 days | Blood NAD AUC +1.7× to +3.7×, p<0.001. No significant metabolic effects. Authors concede the trial was "neither large enough nor long enough." Metro International Biotech funded; authors employed by sponsor. |
| **Pencina KM, et al. J Clin Endocrinol Metab 2023** | 30 (21 active / 9 placebo) | MIB-626 2000 mg | 28 days | Significant: total cholesterol −26.89 mg/dL (p=0.004), LDL −18.73 (p=0.007), diastolic BP −7.01 mmHg (p=0.034), weight −1.9 kg (p=0.008). Null: muscle strength, aerobic capacity, insulin sensitivity, hepatic fat. Note 9 placebo subjects and ~8 endpoints without multiplicity correction. Sponsor-funded, authors held equity. |
The pattern
Across roughly 31 trials, the reproducible finding is the biomarker. Blood or PBMC NAD+ rises, dose-dependently, reliably. That is pharmacology, not benefit.
Clinical endpoints are heterogeneous and mostly null. Where positive, they are frequently (a) post-hoc subgroups — Martens' 9 mmHg, Katayoshi's BMI subgroup, Brakedal's Parkinson's benefit; (b) secondary among many endpoints without correction; (c) from trials of n ≤ 40; or (d) industry-funded. Often several at once.
And in the two trials that measured both, blood NAD+ rose while the target tissue's NAD+ did not (Yoshino 2021, muscle; Dollerup 2020, muscle). So the biomarker doesn't even reliably demonstrate target engagement where it matters.
The largest trials in the entire field are n=140, n=120, n=111 and n=108. None is an outcome trial. There is no NMN or NR trial with a hard clinical endpoint, and none longer than 26 weeks.
The Australian regulatory change
On 10 December 2025, the TGA added NMN to the Therapeutic Goods (Permissible Ingredients) Determination, permitting NMN in listed (AUST L) medicines. Reported as Permissible Ingredients Determination (No. 4) 2025.
Conditions, as reported:
- Maximum recommended daily intake 500 mg
- Adults only; excludes pregnant and breastfeeding women
- Oral dosage forms only
- Recommended duration 12 weeks or less
- Two-year market exclusivity to the applicant (Longevity Life Sciences Pty Ltd), 10 December 2025 to 10 December 2027
Permitted claims are ordinary listed-medicine boilerplate — "maintain or support energy production," "maintain or support general health and well-being," "reduce free radicals formed in the body." These are permitted-claim templates, not findings about NMN.
The "world first" framing circulating in coverage is a company representative's quote, and the trade press that carried it hedged to "one of the few countries." Treat it as a claim, not a finding.
**Important context: the TGA previously treated NMN as not permissible** and had issued safety-alert and guidance pages about NAD, NAD+, NADH and NMN medicines sold in Australia. This is a reversal, not a longstanding position.
The United States
The FDA excluded NMN from the dietary supplement definition in October 2022 under the drug-preclusion clause. After a 2023 citizen petition and 2024 litigation, on 29 September 2025 the FDA issued two letters reversing its interpretation, reasoning via the "race to market" provision — NMN was marketed as a supplement before it was authorised for drug investigation, and the FDA clarified it "does not require proof that the ingredient was lawfully marketed." Confirmation letters to ingredient suppliers followed in December 2025.
But NMN remains a New Dietary Ingredient in the US: a notification is required before marketing, and self-affirmed GRAS does not substitute.
So Australia permits NMN inside a regulated medicines framework with a dose cap and a duration note; the US merely stopped excluding it. That's a real difference.
And note the asymmetry with NR
NR has been permitted in Australian listed medicines since November 2019 at 300 mg/day, registered as AUST L 328365 in January 2020, with EFSA novel-food approval in 2019 and EU authorisation in 2020. NMN has no EU approval as of late 2025.
The "Australia first" headline is about NMN specifically, not about NAD+ precursors.
Why the market looks the way it does
The NR-versus-NMN consumer framing is downstream of a patent war, not a scientific disagreement.
ChromaDex sued Elysium Health in December 2016 over NR supply, adding trade-secret claims. A September 2021 jury returned a mixed verdict: Elysium to pay ChromaDex roughly $2.25–2.5 million, and ChromaDex to pay Elysium $1.9 million — including $1.025 million in punitive damages for fraud. ChromaDex's trade-secret and fiduciary-duty claims were rejected. The same month, a Delaware federal judge invalidated two NR patents held by ChromaDex and Dartmouth.
NMN had no equivalent US gatekeeper, which is precisely why the FDA's drug-preclusion exclusion mattered so much commercially.
Safety
What's actually known is short-term tolerability. The longest NMN trial is 24 weeks (n=14); the longest NR trial 26 weeks (n=111). The overwhelming majority run 4–12 weeks at n under 50. Highest doses formally studied: NR 3000 mg/day for 4 weeks (n=20), NR 2000 mg/day for 12 weeks (n=40), NMN 1200 mg/day for 6 weeks (n=48).
There is no long-term human safety data for either compound. Anyone taking these for years is outside the evidence entirely.
Methyl donors and TMG — the concern was tested, and largely didn't materialise
The theory: nicotinamide is cleared by NNMT using S-adenosylmethionine, so high loads could drain the methyl pool. This is why TMG (betaine) is co-marketed with NMN.
NR-SAFE measured it directly at 3000 mg/day for 4 weeks. Methylated metabolites rose sharply, but SAM, SAH, the SAM/SAH ratio and methionine were unchanged — "NR even at a dose of 3000 mg for 4 weeks does not cause depletion of the methyl-group pool." Homocysteine rose mildly and transiently, then stabilised, mostly within normal range.
We could find no RCT of NMN or NR with versus without TMG. The co-marketing is vendor reasoning, not data.
The cancer question — real preclinical literature, no human data
Tumour cells are NAD+-dependent and upregulate the salvage pathway; NAMPT is overexpressed across multiple cancers and "is capable of increasing the tumorigenic properties of cells."
**The cleanest way to see the tension: NAMPT inhibitors are an active oncology drug class.** Pharmaceutical companies are spending money trying to lower tumour NAD+ while supplement companies sell precursors to raise it.
The most-cited direct finding (Maric T, et al., Biosensors and Bioelectronics 220:114826, 2023): the aggressive triple-negative breast cancer line MDA-MB-231 had the highest NR uptake of those screened, and in a mouse model "NR supplementation results in a significant increase in cancer prevalence and metastases of TNBC to the brain."
Caveats, stated fairly: mouse xenograft; the paper's primary purpose was validating an imaging probe, not carcinogenicity; the dose isn't specified; and it concerns NR, not NMN. No human trial has been powered or long enough to detect a cancer signal.
The counterpoint matters too: nicotinamide reduced skin cancer in a phase 3 trial, and NAD+ also supports DNA repair. These compounds are not uniformly pro-tumour. The concern is specifically about existing malignancy, and it's a reason to raise it with an oncologist rather than a reason for general alarm.
What happened when niacin was properly tested
These are the only large, long, hard-endpoint trials in the entire B3 family, and they are the most relevant cautionary data in this article.
**AIM-HIGH (Boden WE, et al., NEJM 2011). n=3,414, established cardiovascular disease, all on simvastatin, extended-release niacin 1,500–2,000 mg/day. Stopped early for futility. Primary composite 16.4% vs 16.2%; HR 1.02 (95% CI 0.87–1.21), p = 0.79.** Meanwhile HDL rose 25% and triglycerides fell 28.6%.
The lipid biomarkers improved and the outcomes did not. That is the cleanest available cautionary tale for "blood NAD+ went up," and it comes from the same vitamin family. There was also an ischaemic stroke signal: 1.6% vs 0.9%.
**HPS2-THRIVE (NEJM 2014). n=25,673, all on statin, ER niacin 2 g plus laropiprant, median 3.9 years. Major vascular events 13.2% vs 13.7%; RR 0.96 (0.90–1.03), p = 0.29 — no benefit. Significant excess serious adverse events, all p<0.001: new-onset diabetes 5.7% vs 4.3%**, worsening glucose control 11.1% vs 7.5%, infection 8.0% vs 6.6%, bleeding 2.5% vs 1.9%.
Fair caveat: HPS2-THRIVE used niacin with laropiprant, so some harm attribution is contested. AIM-HIGH used niacin alone and showed futility plus a stroke signal.
Together: around 29,000 patients, four years each, zero benefit, documented harm. Compare that to a category whose entire human evidence base is roughly 1,300 people over 4 to 26 weeks.
Australian note: nicotinic acid 250 mg tablets are Schedule 3 (Pharmacist Only) in Australia — the regulator already treats pharmacologic-dose niacin as a drug with risks.
The cheap alternative nobody sells you
No head-to-head trial exists. We could find no human RCT comparing NMN or NR against nicotinamide or nicotinic acid at equimolar doses on any endpoint. This is the single largest evidence gap in the category. Every "NMN is superior" claim is an inference from rodent and cell work, not a human comparison.
What the cheap forms have achieved in humans is, awkwardly, more rather than less.
Nicotinic acid rescued tissue NAD+ where NR did not. Pirinen E, Auranen M, Khan NA, et al., "Niacin Cures Systemic NAD+ Deficiency and Improves Muscle Performance in Adult-Onset Mitochondrial Myopathy," Cell Metabolism 31:1078–1090 (2020). n=5 patients plus 10 matched controls, niacin escalated to 750–1,000 mg/day, patients treated 10 months.
- Patients had about 2-fold lower baseline blood NAD+ than controls
- Niacin raised blood NAD+ up to 8-fold, and muscle NAD+ reached healthy-control levels
- Strength gains across muscle groups; hepatic fat down ~50%, visceral fat down ~25%
- Adverse effects: flushing and tingling in everyone above 500 mg/day, improving over time
Contrast directly with Dollerup: NR at 2000 mg/day for 12 weeks produced "no clear change" in skeletal muscle NAD+. Cheap niacin normalised muscle NAD+; expensive NR did not move it. Caveats: n=5, open-label, a genuine deficiency state rather than healthy ageing, and 10 months versus 12 weeks. But it is the strongest tissue-NAD+ result in the human literature and it belongs to the cents-per-gram compound.
And nicotinamide has a positive phase 3 trial with a clinical endpoint. NMN and NR have none. ONTRAC (Chen AC, Martin AJ, Choy B, et al., NEJM 2015;373:1618–1626): n=386, nicotinamide 500 mg twice daily for 12 months in people with prior skin cancers. New non-melanoma skin cancers reduced 23% (95% CI 4–38), p = 0.02. Benefit did not persist after stopping. No clinically significant difference in adverse events. Funded by the Australian NHMRC — no commercial sponsor.
The verdict: the premium is not justified by comparative human evidence, because comparative human evidence does not exist. What does exist points the other way — the cheap forms have the only tissue-NAD+ rescue and the only positive phase 3 clinical endpoint in the family. The premium tracks patent position and regulatory exclusivity, not demonstrated superiority.
What to do with all this
If you want to try NMN or NR, the honest framing is that you are buying a reliable biomarker change with an unproven clinical benefit, at a price set by patents. That's a legitimate thing to spend money on if you understand it as such. It is not the same as buying something that works.
If the mechanism you're interested in is raising NAD+, the cheap compounds have better human data — and in the one direct tissue comparison available, niacin did what NR didn't.
If you have, or have had, a malignancy, discuss it with your oncologist before taking any NAD+ precursor. The preclinical literature is coherent enough to warrant the conversation and too thin to answer it.
And if you're in Australia: NMN is now legal in listed medicines at up to 500 mg/day for adults, with a 12-week duration note. A listed medicine means the TGA has assessed safety and quality — not that it has assessed whether the product works.
Frequently asked
Is NMN or NR better? Nobody has compared them head to head in humans, so there is no evidence-based answer. NR has a longer regulatory history in Australia (permitted since 2019 at 300 mg/day) and more published trials. NMN has been permitted here only since December 2025, at up to 500 mg/day. The consumer framing of one versus the other originates in a patent dispute, not a scientific one.
Does NMN actually work? It reliably raises blood NAD+ — that part is well established. Whether that produces any health benefit is not. Across roughly 31 human trials of NMN and NR, clinical outcomes were mostly null, the positive findings were frequently post-hoc subgroups or came from trials of fewer than 40 people, and most were industry-funded. There is no trial with a hard clinical endpoint and none longer than 26 weeks.
Does NAD+ really decline with age? Less clearly than the marketing says. The largest and best-controlled study — over 300 people across seven cohorts, published in Nature Metabolism in 2026 — found no age effect on whole-blood NAD+ at all, with age explaining about 1% of the variation. The widely cited "4% per year" figure comes from a 29-person cross-sectional study.
Is NMN legal in Australia? Yes, since 10 December 2025, when the TGA added it to the Permissible Ingredients Determination for listed medicines — adults only, oral forms only, maximum 500 mg per day, with a recommended duration of 12 weeks or less. One supplier holds two-year market exclusivity until December 2027.
Should I take TMG with NMN? There's no trial supporting it. The theory is that high nicotinamide loads drain methyl donors, but the one study that measured this directly — at 3,000 mg/day of NR for four weeks — found S-adenosylmethionine, SAH, the SAM/SAH ratio and methionine all unchanged. The co-marketing is reasoning, not data.
Why not just take niacin or nicotinamide? It's a fair question and the evidence doesn't answer it against you. Nicotinamide has a positive phase 3 trial with a clinical endpoint (skin cancer prevention, n=386, NHMRC-funded); NMN and NR have none. Nicotinic acid normalised muscle NAD+ in a small trial where high-dose NR did not. The counterweight is that high-dose niacin failed two large cardiovascular trials and caused harm in one of them — which is itself useful information about what happens when this vitamin family is finally tested properly.
Last reviewed 1 October 2026. Trial figures are from the published papers; funding and conflict-of-interest statements are reported where the papers disclose them. Australian regulatory details are from trade and specialist reporting of the December 2025 Determination — verify the current instrument on the Federal Register before relying on the dose cap. This page is general information, not medical advice.