The NAD+ Precursor Trials That Found Nothing, and What They Still Tell Us

The controlled human trials of NMN, NR and urolithin A that missed their main outcome, set out with dose, size, duration and limits. Blood NAD+ rises reliably. Cognition, strength, insulin sensitivity and blood pressure mostly did not move in trials this short. The largest trial, a 52-week study of nicotinamide riboside in early Parkinson's disease, favored placebo. This is a Mt. Angel Vitamins company white paper. It is not a journal article and has not been through journal peer review.

If you're not a researcher

  • The question: do NMN, NR or urolithin A deliver the benefits they are sold for, in controlled human trials? Most trials that tested a function outcome found no benefit.
  • The answer: blood NAD+ rises reliably after NMN or NR. Strength, cognition, insulin sensitivity and blood pressure mostly did not move in the trials checked here.
  • How strong: the NAD+ rise is graded Strong. The function outcomes are graded Insufficient, because the trials are small and short and a few positive signals conflict with the nulls.
  • The doses: the nulls used 250 to 2,000 mg a day of NMN or NR, or 1,000 mg a day of urolithin A, for 4 to 52 weeks.
  • The one thing to do: read a null as a limit on what was tested, and check the milligrams on any label against the trial doses.
  • Full evidence: the NMN monograph and the NR monograph.

In the controlled human trials we could find, NMN and NR raised blood NAD+ in each trial in the table that measured it. The functions those trials were built to test mostly did not improve. The largest trial in the set, a 52-week study of nicotinamide riboside (NR) in early Parkinson's disease, found the placebo group did better on the primary rating scale. This page lists the trials that missed their main outcome, with dose, size, duration and limits. It then covers the few trials that found a signal, so the whole picture sits on one page.

What we searched, and what we left out

We searched PubMed on 20 October 2026 with six queries. Three looked for randomized, placebo-controlled trials of NR, NMN and urolithin A, returning 41, 25 and 21 records. One paired those compounds with null-result wording such as "no significant" or "did not" (39 records). One looked for reviews (28 records, used for context only). One looked for urolithin A trials in athletes (11 records). The six searches returned 126 records before we removed duplicates by PMID. We read titles and abstracts for every candidate. We read full text only where an abstract left out the sample size or dose (Martens 2026 and Igarashi 2022). We checked the Whitfield registry entry on ClinicalTrials.gov (NCT04783207). Every other NCT number comes from its fetched PubMed record.

Inclusion: controlled human studies of NMN, NR or urolithin A with a prespecified primary or key secondary outcome and a reported result. Positive and null trials sit in the same table. Exclusions: animal and cell work (not counted here); trials whose only prespecified outcomes were safety or blood NAD+, judged by title and abstract (Fukamizu 2022, whose title describes a safety evaluation, and Pencina 2025 in hospitalized COVID-19 patients); trials where the precursor could not be separated from other ingredients (Dellinger 2022, NR with pterostilbene; Roy 2026, an NAD+ precursor with ketones; Yulug 2025, a combined metabolic activator); and systematic reviews, which we used for context and not for numbers.

The trials that missed their main outcome

Trial Population and n Dose and duration Primary or key outcome Result (direction, CI where given) Limits Sponsor type
Brakedal 2026 (NOPARK) Early Parkinson's disease; 410 randomized, 393 analysed NR 500 mg twice daily (1,000 mg/day), 52 weeks MDS-UPDRS parts I to III (primary) Worse on NR: adjusted difference 2.72 points favoring placebo (95% CI 0.47 to 4.98). Nonmotor score also worse on NR (4.22; 1.07 to 7.37) Longest and largest in the set; one country Academic (Norway); funding not checked
da Costa Pinaffi-Langley 2026 Adults aged 40 to 64 with obesity; 50 randomized (25 per arm) Urolithin A 1,000 mg/day, 4 weeks Brachial flow-mediated dilation (primary) No significant change: control 0.3% vs active 1.3% Four weeks; single site Academic; sponsor not checked
Wu 2025, long-COVID Adults with long-COVID; 58 randomized 2:1 NR 2,000 mg/day, 20 weeks (one arm had 10 weeks of placebo first) Cognition: ECog, RBANS, TMT-B (primary) No difference (p 0.47 to 0.74); fatigue, sleep and mood unchanged. NAD+ rose 2.6 to 3.1-fold Single center; dropout 32% to 51% in the NR arm Academic; an NR supplier provided support (named in abstract)
Whitfield 2025 Male distance runners; 42 randomized (22 urolithin A, 20 placebo) Urolithin A 1,000 mg/day, 4 weeks at altitude camp 3,000 m time trial (registry primary) No significant gain within either group (p 0.116 and 0.771). Perceived exertion lower with urolithin A (p 0.02) Four weeks; the registry summary says 36 and its enrollment field says 42 Industry-linked; registry sponsor is the urolithin A ingredient developer
Lin 2025 Adults aged 55 and older with systolic BP of 130 or higher; 54 randomized, 49 completed NR 1,000 mg/day with supervised walking, 6 weeks Daytime systolic BP (primary) NR plus exercise did not beat placebo plus exercise (mean change +5.2 vs -2.7 mmHg) Pilot; 15 to 18 completers per arm Academic
Wu 2025, subjective decline and MCI Adults over 55 with subjective cognitive decline or MCI; 46 randomized, 37 completed NR 1 g/day, 8 weeks, crossover RBANS cognition (primary) No between-group difference. Plasma pTau217 fell on NR (p = 0.02) Eight weeks; the pTau result was secondary Academic; sponsor not checked
Martens 2026 Older adults with amnestic MCI; 52 randomized, 42 completed (22 NR, 20 placebo) NR 1,000 mg/day, 12 weeks Cognitive function (primary) No group-by-time gain. Blood NAD+ rose about twofold Pilot; regional brain blood flow was exploratory and uncorrected Academic; capsules from the ingredient maker under a research agreement
Orr 2023 Older adults with MCI; 20 randomized NR escalated to 1,000 mg/day over 10 weeks MoCA cognition (primary) No change versus placebo. Blood NAD+ rose 2.6-fold Twenty people; 10 weeks; a default-mode blood flow fall did not survive multiple-test correction Academic; sponsor not checked
Katayoshi 2023 Healthy middle-aged adults; 36 randomized NMN 125 mg twice daily (250 mg/day), 12 weeks Arterial stiffness (pulse wave velocity) Tended lower with NMN; no significant difference Small; 12 weeks Industry-linked; all authors list the product maker's laboratories
Pencina 2023 Adults 45 and older with overweight or obesity; 30 randomized 2:1 Beta-NMN (MIB-626), 2 x 500 mg twice daily (2,000 mg/day), 28 days Listed outcomes: safety and NAD+ first, then insulin sensitivity, strength, aerobic capacity Insulin sensitivity, strength, aerobic capacity and stair power unchanged. Weight fell 1.9 kg more than placebo (95% CI -3.3 to -0.5); diastolic BP fell 7.0 mmHg more (-13.4 to -0.6) 28 days; mixed outcome set Industry-linked; ingredient developer named in abstract
Akasaka 2023 Men aged 65 and older with diabetes and low grip or walking speed; 14 analysed NMN 250 mg/day, 24 weeks Grip strength and walking speed (primary) No between-group difference. Grip change with NMN +1.25 kg (95% CI -2.31 to 4.81) vs -0.44 kg on placebo Fourteen analysed; wide intervals Academic; two authors list company affiliations
Igarashi 2022 Healthy men aged 65 and older in Japan; 42 randomized NMN 250 mg/day, 12 weeks Primary: blood NAD+ and skeletal muscle index NAD+ rose. Muscle index unchanged. Gait speed and left grip rose, nominal and uncorrected A supplier error at week 6 led to 22 participants' 12-week data being excluded; 20 analysed. Authors' power estimate was 0.99 for a 0.45 kg/m2 change Academic; NMN supplied by a company
Liu 2022 Adults aged 65 to 90; 66 randomized (33 per arm) Urolithin A 1,000 mg/day, 4 months Primary: 6-minute walk and hand muscle ATP production Primary not significant (walk +60.8 m vs +42.5 m on placebo). Endurance in two muscles improved at 2 months (secondary) Mixed result; four months Industry-linked; ingredient developer among the co-authors
Remie 2020 Healthy overweight or obese adults; 13 randomized, crossover NR 1,000 mg/day, 6 weeks per arm Insulin sensitivity and mitochondrial function (stated aims) No change in insulin sensitivity or mitochondrial function. Fat-free mass share rose 1.34 percentage points over placebo (P = 0.02) Thirteen people; six weeks Sponsor not checked
Dollerup 2018 Sedentary men aged 40 to 70 with BMI over 30; 40 randomized NR 1,000 mg twice daily (2,000 mg/day), 12 weeks Insulin sensitivity by clamp (primary) No improvement; no serious adverse events Men only; 12 weeks Investigator-initiated (NCT02303483)

Where the evidence points

  • Blood or muscle NAD+ rise: Strong. Six controlled trials in the table measured it, and all six found a rise. Three report about two- to three-fold increases (Martens 2026, Orr 2023, Wu 2025).
  • Cognition: Insufficient. Four trials had null primary outcomes (Martens 2026, Orr 2023, both Wu 2025 studies). The 52-week NOPARK trial moved against NR on its disease scale.
  • Insulin sensitivity: Insufficient. Three trials were null (Dollerup 2018, Remie 2020, Pencina 2023). One positive study in women with prediabetes conflicts with them (Yoshino 2021).
  • Strength, endurance and walking: Insufficient. Akasaka 2023 was null, Igarashi 2022 showed nominal gains only, and both urolithin A primaries were null. Positive signals exist (see below).
  • Blood pressure and arterial stiffness: Insufficient. Lin 2025 was null on its primary. Katayoshi 2023 showed no significant change.
  • Short-term safety: Moderate. The trials up to 12 weeks reported no excess of serious adverse events (Igarashi 2022, Martens 2026, Akasaka 2023). NOPARK reported serious adverse events in 8.3% on NR and 14.1% on placebo.
  • Lifespan or healthspan in people: Insufficient. None of the trials measured it.

Why blood NAD+ can rise while function stays flat

Six trials in the table measured NAD+ metabolites and found a rise. Function did not follow in the four with functional or cognitive primaries. Whole blood is the easiest tissue to sample, and it is not the tissue where most of these functions happen. Martens's authors name that gap directly. They could not show the blood rise reached the brain, and they suggest a longer course and larger groups may be needed, because brain uptake can be slow and variable.

The other reasons are about design. Most trials ran four to twelve weeks. Several enrolled people who were healthy or only mildly impaired, which leaves little room to improve. Several used noisy tests, such as one 3,000 m time trial, a grip test, or a rating scale that moves slowly. A null result in that setting says the test did not catch a change. It does not show that the change is absent. For how the two precursors differ on the evidence, see NMN vs NR.

How big and how long these trials were

Fourteen of the fifteen trials in the table enrolled 66 people or fewer. The shortest ran four weeks. Most ran six to twelve weeks. Four ran 17 weeks or longer (Liu 2022, Wu 2025 long-COVID, Akasaka 2023 and NOPARK). Only NOPARK, with 393 people analysed over 52 weeks, was large enough to detect a difference of a few points on its scale.

Small trials can miss effects of several kilograms or several meters. Akasaka's 95% interval around the NMN group's grip change ran from -2.31 to +4.81 kg. That interval is too wide to rule out a gain of a few kilograms in 14 men. Igarashi's authors calculated 0.99 power to detect a 0.45 kg/m2 change in muscle index, which is a statement about one endpoint and one analysed group, not about the whole question.

What null results rule out, and what they leave open

A null result from a trial this size rules out a large, reliable benefit on the tested outcome, in the tested group, at the tested dose and duration. It does not rule out a smaller benefit, a benefit that needs years, a benefit at another dose, a benefit in another group, or a benefit on an outcome nobody measured. The long, large NOPARK trial found a disease-course result that favored placebo, which is a reason to read disease claims about this ingredient with care. "No significant difference" describes the test, not the ingredient.

For urolithin A, the dose shape matters as well. See the urolithin A dose analysis for how the response curve is read.

The trials that found a signal

Not every trial was flat. Yoshino et al. (Science, 2021, PMID 33888596) reported that NMN increased muscle insulin sensitivity in women with prediabetes. We read the title and metadata only. The design, 25 women over 10 weeks at 250 mg a day, comes from the introduction of the 2022 Igarashi paper, not from Yoshino's own text. In peripheral artery disease, a six-month NR trial with or without resveratrol (90 people) reported a 17.6-meter gain in six-minute walk distance over placebo, with a 90% interval starting at +1.8 meters. The authors call for a larger trial (McDermott et al. 2024). The urolithin A athletes' trial is covered in more detail in its own article. Pencina's 2023 trial saw body weight and diastolic pressure fall more than placebo. Remie's crossover found a larger fat-free mass share on NR. Each of these is a single trial, most of them small, and several carry sponsor links. They are reasons to keep following the field, not reasons to buy on them.

Where our serving sits

Our NMN + NR capsules list a two-capsule serving with 500 mg NMN and 200 mg NR, plus 250 mg resveratrol, 50 mg TMG, 50 mg Bioperine and 10 mg zinc. The live product page lists 30 servings per 60-capsule bottle at $59.99, which works out to $2.00 a day. The NMN trials on this page used 250 to 2,000 mg a day, so the 500 mg NMN falls inside that range. The NR trials used 1,000 to 2,000 mg a day, so 200 mg of NR sits below that range. None of the trials on this page tested this five-ingredient combination. The full dose comparison is in the NMN and NR doses guide. The label checklist shows what to verify on any bottle. Read the label and today's price on the product page.

What is established, what is proposed, and what would settle it

  • Established: NMN and NR raise blood NAD+ metabolites in adults at 250 to 2,000 mg a day over four to twelve weeks. Those short trials reported the doses as well tolerated. Most controlled trials of a function outcome found no benefit on their primary measure.
  • Proposed: NAD+ precursors may help specific outcomes in specific people, such as walking distance in peripheral artery disease, insulin sensitivity in women with prediabetes, or cerebral blood flow in people with cerebrovascular risk. We found no independent replication of these signals in the searches.
  • What would settle it: larger and longer trials with a functional primary outcome set in advance and a defined subgroup; measurements in the target tissue, not only whole blood; head-to-head trials of NMN and NR on the same functional endpoint; and independent replication of the walking and prediabetes signals.

Glossary

  • NAD+: nicotinamide adenine dinucleotide, a coenzyme that cells use for energy and for repair and signaling enzymes.
  • NMN: nicotinamide mononucleotide, an NAD+ precursor.
  • NR: nicotinamide riboside, an NAD+ precursor related to vitamin B3.
  • Urolithin A: a compound that gut bacteria make from ellagitannins, studied for mitophagy, the clearing of worn mitochondria.
  • Randomized controlled trial: participants are assigned by chance to a test arm or a control arm.
  • Placebo: an inactive capsule that looks the same as the product.
  • Crossover: each participant takes the product and the placebo in turn.
  • Primary outcome: the one measure a trial is designed to test. Key secondary outcomes are named in advance and come next.
  • Confidence interval (CI): the range of differences the data are consistent with. A wide interval means a small trial.
  • P value: the chance of a difference this large if there were no real effect. It does not measure size.
  • Insulin sensitivity: how well muscle and liver respond to insulin. The hyperinsulinemic clamp is the reference test.
  • Mild cognitive impairment (MCI): memory or thinking problems beyond normal aging but short of dementia.

References

  1. Brakedal et al. Nicotinamide riboside for patients with early-stage Parkinson disease: the NOPARK randomized clinical trial. JAMA. 2026. doi:10.1001/jama.2026.18828. PubMed
  2. da Costa Pinaffi-Langley et al. Urolithin A and endothelial and cerebrovascular function in middle-aged adults with obesity. JAMA Netw Open. 2026. doi:10.1001/jamanetworkopen.2026.38047. PubMed
  3. Wu et al. Effects of nicotinamide riboside on NAD+ levels, cognition, and symptom recovery in long-COVID. EClinicalMedicine. 2025. doi:10.1016/j.eclinm.2025.103633. PubMed
  4. Whitfield et al. Evaluating the impact of urolithin A supplementation on running performance, recovery, and mitochondrial biomarkers in highly trained male distance runners. Sports Med. 2025. doi:10.1007/s40279-025-02292-5. PubMed
  5. Lin et al. Nicotinamide riboside combined with exercise to treat hypertension in middle-aged and older adults. GeroScience. 2025. doi:10.1007/s11357-025-01815-2. PubMed
  6. Wu et al. Cognitive and Alzheimer's disease biomarker effects of oral nicotinamide riboside in older adults with subjective cognitive decline and mild cognitive impairment. Alzheimers Dement (N Y). 2025. doi:10.1002/trc2.70023. PubMed
  7. Martens et al. A phase-II randomized controlled pilot study of nicotinamide riboside supplementation in older adults with amnestic mild cognitive impairment. Alzheimers Dement. 2026. doi:10.1002/alz.71605. PubMed
  8. Orr et al. A randomized placebo-controlled trial of nicotinamide riboside in older adults with mild cognitive impairment. GeroScience. 2023. doi:10.1007/s11357-023-00999-9. PubMed
  9. Katayoshi et al. Nicotinamide adenine dinucleotide metabolism and arterial stiffness after long-term nicotinamide mononucleotide supplementation. Sci Rep. 2023. doi:10.1038/s41598-023-29787-3. PubMed
  10. Pencina et al. Nicotinamide adenine dinucleotide augmentation in overweight or obese middle-aged and older adults. J Clin Endocrinol Metab. 2023. doi:10.1210/clinem/dgad027. PubMed
  11. Akasaka et al. Effects of nicotinamide mononucleotide on older patients with diabetes and impaired physical performance. Geriatr Gerontol Int. 2023. doi:10.1111/ggi.14513. PubMed
  12. Igarashi et al. Chronic nicotinamide mononucleotide supplementation elevates blood nicotinamide adenine dinucleotide levels and alters muscle function in healthy older men. npj Aging. 2022. doi:10.1038/s41514-022-00084-z. PubMed
  13. Liu et al. Effect of urolithin A supplementation on muscle endurance and mitochondrial health in older adults. JAMA Netw Open. 2022. doi:10.1001/jamanetworkopen.2021.44279. PubMed
  14. Remie et al. Nicotinamide riboside supplementation alters body composition and skeletal muscle acetylcarnitine concentrations in healthy obese humans. Am J Clin Nutr. 2020. doi:10.1093/ajcn/nqaa072. PubMed
  15. Dollerup et al. A randomized placebo-controlled clinical trial of nicotinamide riboside in obese men. Am J Clin Nutr. 2018. doi:10.1093/ajcn/nqy132. PubMed
  16. McDermott et al. Nicotinamide riboside for peripheral artery disease: the NICE randomized clinical trial. Nat Commun. 2024. doi:10.1038/s41467-024-49092-5. PubMed
  17. Yoshino et al. Nicotinamide mononucleotide increases muscle insulin sensitivity in prediabetic women. Science. 2021. PubMed

Next reads

Related reading

The Bottom Line

The controlled trials of NMN, NR and urolithin A that missed their main outcome are not evidence that these compounds do nothing. They show that blood NAD+ rises reliably, while the functions most trials tested have mostly not moved in trials this short and small. Read a null as a limit on what was tested. Check the milligrams on any label against the trial doses, and do not expect a blood number to become a felt change.

FAQ

Do NMN and NR raise NAD+ in people?

Yes, in the trials checked here. Blood NAD+ metabolites rose after NMN or NR in every trial in the table that measured them, often two- to three-fold, at 250 to 2,000 mg a day for four to twelve weeks. That is a blood measure. It does not prove that NAD+ rose in the tissue that drives a given function.

Which trials found no benefit?

The clearest nulls are Dollerup 2018 (NR, insulin sensitivity), Akasaka 2023 (NMN, grip and walking speed), Martens 2026 (NR, cognition in 42 completers), Orr 2023 (NR, cognition) and NOPARK 2026 (NR, early Parkinson's disease, which favored placebo). The full table lists the others with doses, sizes and limits.

Why do NAD+ levels rise while function stays flat?

Blood is not the tissue where most of these functions happen. The trials were short and small, and several enrolled people with little room to improve. Some used noisy outcomes such as a single time trial. Most did not measure NAD+ in brain or muscle, so a rise in blood does not show a change in the tissue that matters.

How large and how long were these trials?

Fourteen of the fifteen trials in the table enrolled 66 people or fewer. Most ran six to twelve weeks. The longest, NOPARK, ran 52 weeks with 393 people analysed. Small trials can miss effects of several kilograms or several meters, so a null from 14 to 40 people is a weak test of a modest benefit.

What do null results rule out?

They rule out a large, reliable benefit on the tested outcome, in the tested group, at the tested dose and duration. They do not rule out a smaller benefit, a benefit that takes longer than the trial ran, a different dose or group, or an outcome nobody measured. A result that is not statistically significant describes the test, not the ingredient.

How should a buyer weigh a supplement whose trials were mostly null?

Check the milligrams of each named ingredient per serving on the label, not a proprietary blend total. Compare them with the doses the trials used, check the price per day, and look for lot-level identity testing and a certificate of analysis. Treat a blood NAD+ change as a laboratory result, and do not expect it to be felt.

More on NAD+ and Cellular Energy: The Complete Guide