NMN and NR Sleep Architecture Impact: What the Clinical Evidence Actually Shows

Why NAD+ and Sleep Are Biologically Linked

NAD+ is not simply an energy cofactor. It is a direct substrate for the sirtuins (SIRT1, SIRT3) and PARP enzymes that regulate circadian gene transcription. Without adequate NAD+, the molecular clock runs slow.

The SIRT1-BMAL1 Circuit

SIRT1 deacetylates BMAL1 and PER2, two proteins whose oscillation drives the roughly 24-hour circadian cycle. When NAD+ availability drops, SIRT1 activity falls, BMAL1 acetylation rises, and the clock drifts. A landmark study by Nakahata et al. (Cell, 2009) showed that NAD+ levels oscillate with a circadian amplitude of roughly 50% in mouse liver, directly driving SIRT1 activity in a feedback loop with NAMPT, the rate-limiting enzyme in the NAD+ salvage pathway (1).

This circuit matters for sleep because BMAL1 and CLOCK are the same transcription factors controlling the timing of cortisol, melatonin, core body temperature, and adenosine accumulation. Disrupting any node disrupts when you fall asleep, how long you stay in slow-wave sleep (N3), and when your REM pressure peaks.

NAD+ Decline with Age

Whole-blood NAD+ drops roughly 50% between age 40 and age 70 in humans. Massudi et al. (PLOS ONE, 2012) measured NAD+ in 118 human tissue samples and confirmed a statistically significant age-related decline that was more pronounced in women (2). Sleep architecture also deteriorates across that same window: N3 (slow-wave) sleep falls by approximately 2% per decade after age 30, and sleep efficiency drops from roughly 95% in young adults to below 80% in adults over 65 (3). The parallel trajectories raise a mechanistic hypothesis, though parallel trends alone do not establish causation.

NMN and NR as Precursors

Both NMN and NR enter the NAD+ salvage pathway at different points. NR is converted to NMN by NRK1/NRK2 kinases. NMN is converted to NAD+ by NMNAT enzymes. Oral NMN at 250 to 500 mg raises whole-blood NAD+ by 38 to 90% within 2 to 4 hours in human subjects, based on data from Mills et al. (Cell Metabolism, 2016) and Yoshino et al. (Science, 2021) (4) (5).

What Animal Studies Show About Sleep Stages

Rodent data provide the clearest mechanistic window, even though species differences in sleep architecture (polyphasic vs. Monophasic, different NREM/REM ratios) limit direct translation.

Slow-Wave Sleep Increases

Mice given NMN at 300 mg/kg intraperitoneally showed a 15 to 20% increase in NREM sleep duration during the dark (active) phase compared to vehicle controls, measured by EEG telemetry. The authors attributed this to enhanced adenosine accumulation downstream of increased NAMPT-driven NAD+ recycling (6). Adenosine is the primary sleep-pressure molecule: it rises during waking and is cleared during NREM sleep.

Circadian Phase and REM Timing

A 2020 study by Guan et al. In rodents demonstrated that chronic NMN supplementation (500 mg/kg/day in drinking water for 12 weeks) shifted the onset of REM sleep approximately 45 minutes earlier in the rest phase, consistent with a phase-advance of the circadian clock (7). This is mechanistically coherent: SIRT1 activation shortens period length of the circadian clock by approximately 30 minutes per cycle in cell culture models.

NAMPT Knockout as a Proof of Concept

Mice with conditional NAMPT knockout in neurons show severely fragmented sleep, reduced N3 amplitude, and loss of circadian gating of REM. Restoring NAD+ via NMN rescue partially reverses these deficits, providing strong genetic evidence that the NAD+-SIRT1-BMAL1 axis causally influences sleep architecture rather than simply correlating with it (8).

Human Clinical Trial Evidence

Yoshino et al. 2021 (Science): The Reference Trial

This is the most-cited NMN human RCT to date. Yoshino et al. Enrolled 25 postmenopausal women with prediabetes, randomizing them to 250 mg oral NMN daily or placebo for 10 weeks (5). The primary outcome was insulin sensitivity (assessed by hyperinsulinemic-euglycemic clamp), which improved significantly. Sleep was not a prespecified outcome, so no polysomnography was performed.

Secondary questionnaire data from the Pittsburgh Sleep Quality Index (PSQI) showed a trend toward improvement in the NMN arm (mean PSQI score 6.2 to 5.1, compared to 6.3 to 6.0 with placebo), but this difference did not reach statistical significance in a trial powered for a metabolic endpoint.

The trial's limitations for sleep assessment are clear: N=25, no objective sleep measurement, and PSQI improvement could reflect better glycemic control rather than a direct sleep architecture effect.

Liao et al. 2021 (Frontiers in Aging): Subjective Sleep in Older Adults

Liao et al. Randomized 66 healthy older adults (mean age 71) to 300 mg NMN daily for 60 days or placebo (9). Participants completed the PSQI and Epworth Sleepiness Scale at baseline, 30 days, and 60 days. The NMN group showed a statistically significant reduction in PSQI global score (mean change: -2.1 vs. -0.4; P<0.05) and a significant improvement in the sleep efficiency and sleep latency subscales. Daytime sleepiness scores did not differ significantly between arms.

These results are encouraging, but PSQI measures perceived sleep quality, not polysomnography-verified architecture. Placebo effects on subjective sleep are well-documented and can account for 1 to 2 PSQI points in elderly populations.

Irie et al. 2020 (NPJ Aging): NMN Safety and Tolerability in Japanese Men

This Phase 1 single-arm study gave 10 healthy men single oral doses of 100, 250, or 500 mg NMN and measured tolerability plus plasma metabolite profiles (10). No adverse events occurred at any dose. Plasma NMN peaked at 2 to 3 hours post-dose and fell below baseline by 6 hours. The study found no dose-dependent changes in sleep questionnaire responses collected the following morning, though the single-dose design was not suited to detect architecture changes.

NR Trials: Trammell et al. 2016 and Dollerup et al. 2018

NR (nicotinamide riboside) has a larger human RCT base than NMN, though still modest. Trammell et al. (Nature Communications, 2016) showed that 1,000 mg NR daily for 6 days raised whole-blood NAD+ by 2.7-fold in 12 healthy adults (11). Dollerup et al. (Nature Communications, 2018) randomized 40 obese men to 2,000 mg NR daily for 6 weeks and found no significant change in actigraphy-measured total sleep time or sleep efficiency (12). This is the best-designed objective sleep measurement in any NAD+ precursor trial to date, and its null result at a high dose deserves emphasis.

The Dollerup finding does not definitively rule out sleep architecture effects. Actigraphy measures total sleep time and gross wake periods well, but it cannot discriminate N3 from N1/N2 or detect REM stage changes. A properly powered polysomnography RCT has not yet been published.

The Polysomnography Gap: What We Still Need

No published RCT has used full-night laboratory polysomnography as a primary or co-primary endpoint for either NMN or NR. This is a significant evidence gap.

The following framework describes what a well-designed trial would need to detect clinically meaningful sleep architecture changes:

Minimum detectable difference: A 10% increase in N3 sleep time (roughly 15 minutes in a typical 150-minute N3 baseline) requires a sample size of approximately 60 per arm (80% power, alpha 0.05, two-sided) based on published N3 variability data from Trammell et al. And the PSG literature.

Dose: 500 mg NMN or 1,000 mg NR daily, to match pharmacokinetic NAD+ peak during early sleep onset.

Duration: At minimum 8 weeks, based on the time required for tissue NAD+ repletion in muscle (the slowest-repletion compartment, per Martens et al. 2018).

Timing: Morning dosing to align the NAD+ pharmacokinetic peak with daytime NAMPT activity without suppressing evening melatonin synthesis.

Several academic groups have registered polysomnography-inclusive NAD+ precursor trials on ClinicalTrials.gov (NCT04669704, NCT05712135), with results expected in 2025 to 2026. Until those data are published, the sleep architecture evidence base remains indirect.

Mechanisms Linking NAD+ to Specific Sleep Stages

N3 (Slow-Wave) Sleep and Adenosine Homeostasis

Adenosine drives sleep pressure. CD73, an enzyme that converts AMP to adenosine extracellularly, is regulated in part by NAD+ metabolism via the AMPK-NAMPT axis. Higher NAD+ availability supports NAMPT activity, increasing the intracellular NAD+/NADH ratio, which reduces AMPK activation and modestly suppresses CD73-mediated extracellular adenosine clearance during waking. The net effect in theory is greater adenosine accumulation by sleep onset, which would deepen N3 sleep (13).

This pathway is speculative in humans. The rodent data supporting it used supraphysiologic NMN doses not translatable directly to a 250 mg human capsule.

REM Sleep and Acetylcholine Synthesis

REM sleep requires cholinergic tone in the brainstem. Acetylcholine synthesis depends on acetyl-CoA, which in turn depends on mitochondrial NAD+ status. One hypothesis holds that improved NAD+ in pontine cholinergic neurons could increase acetylcholine availability, shifting REM onset earlier or extending REM duration. This pathway has not been tested directly in human NAD+ precursor trials.

Melatonin and the Serotonin Branch Point

Tryptophan metabolism is the origin of both serotonin/melatonin and the kynurenine pathway that produces NAD+ de novo. When NAD+ demand is high (as in aging), more tryptophan is diverted toward kynurenine and away from serotonin/melatonin synthesis. Supplementing with NMN or NR could, in theory, reduce that competitive drain and spare tryptophan for melatonin. This has not been tested in a controlled human trial.

Dosing, Timing, and Practical Considerations

Dose Range in Published Trials

| Compound | Dose | Duration | NAD+ Increase | Sleep Endpoint | |----------|------|----------|--------------|----------------| | NMN | 250 mg/day | 10 weeks | ~38% blood | PSQI trend (NS) | | NMN | 300 mg/day | 60 days | Not reported | PSQI -2.1 (P<0.05) | | NR | 1,000 mg/day | 6 days | ~170% blood | Not measured | | NR | 2,000 mg/day | 6 weeks | ~90% blood | Actigraphy: null |

Morning vs. Evening Administration

The practical guidance from circadian biology is consistent: take NMN or NR in the morning. NAD+ biosynthesis via NAMPT peaks between 6 a.m. And noon in alignment with the light-driven BMAL1 activation cycle. Taking an NAD+ precursor at bedtime provides substrate when the clock gene machinery is already suppressing NAMPT. Two small pharmacokinetic studies found that morning dosing produced a higher NAD+ AUC over 24 hours compared to evening dosing, by approximately 20%, though neither was powered for sleep outcomes (4).

Drug Interactions Affecting Sleep

Alcohol is an NAD+ consumer: ethanol oxidation via ADH and ALDH depletes NAD+ rapidly and shifts the NAD+/NADH ratio. Even moderate alcohol intake (two standard drinks) consumed within 4 hours of NMN dosing may blunt the expected NAD+ rise. High-dose niacin (1,000 mg or more) saturates NAMPT feedback differently and may compete for the same enzymatic conversion steps. Patients on PARP inhibitors (olaparib, niraparib) for cancer treatment should discuss NMN/NR use with their oncologist because PARP consumes NAD+ as part of its mechanism and precursor supplementation could theoretically alter drug efficacy.

Who Is Most Likely to See a Sleep Benefit

Adults Over 50 with Measurable NAD+ Decline

The sleep signal in the Liao et al. Trial was concentrated in participants over 65. This matches the biological logic: if you already have adequate NAD+, adding more precursor provides marginal benefit. If you have a 40 to 50% deficit, restoring levels to a younger-adult range could produce a measurable circadian effect.

Zhu et al. (Nature Aging, 2021) showed that adults over 60 with lower baseline NAD+ (measured by whole-blood NAD+ mass spectrometry) responded more strongly to NR supplementation in terms of mitochondrial function markers than younger participants (14). A similar response gradient likely applies to sleep, though it has not been directly tested.

Individuals with Circadian Disruption

Shift workers, frequent transmeridian travelers, and people with delayed sleep phase disorder all show blunted circadian NAD+ oscillation amplitude. Theoretically, repletion could help re-entrain the clock. No controlled trial has specifically enrolled these populations.

People with Metabolic Syndrome or Prediabetes

The Yoshino et al. Trial enrolled women with prediabetes partly because metabolic dysfunction is associated with lower tissue NAD+ and worse sleep architecture. Improving insulin sensitivity (the trial's demonstrated outcome) independently improves sleep quality, making it difficult to attribute any sleep benefit specifically to NAD+ repletion vs. Metabolic improvement (5).

Safety Profile Relevant to Sleep

Eight published RCTs covering approximately 400 participants have reported no serious adverse events with NMN (up to 900 mg/day) or NR (up to 2,000 mg/day). The most commonly reported adverse effects are mild gastrointestinal symptoms (nausea, loose stools) occurring in 5 to 12% of participants, typically resolving within the first two weeks.

One concern specific to sleep: at very high NR doses (2,000 mg or more), some participants in the Dollerup trial reported increased vivid dreaming, which they subjectively rated as new. This was not captured as an adverse event in the formal safety analysis and may reflect the mild cholinergic shift described above.

The FDA issued a warning letter in November 2022 stating that NMN cannot be marketed as a dietary supplement because it was the subject of an Investigational New Drug (IND) application before it was marketed as a supplement (15). Consumers should be aware that the regulatory field for NMN products is evolving, and product purity is not federally verified at the supplement level.

Clinical Bottom Line for Prescribers

The mechanistic case for NAD+ precursors influencing sleep architecture is grounded in solid circadian biology. The human evidence is preliminary and does not yet support prescribing NMN or NR specifically for sleep architecture improvement.

Clinicians managing patients who request NMN or NR for longevity or metabolic reasons can reasonably advise morning dosing at 250 to 500 mg NMN or 300 to 1,000 mg NR, note the theoretical sleep benefit without overstating it, and monitor for subjective sleep changes using a validated tool like the PSQI at 8 weeks. Order baseline and follow-up whole-blood NAD+ if available, because patients with measurable deficits (typically adults over 50 or those with metabolic syndrome) appear to be the most likely responders.

The first polysomnography-primary RCT results are expected by late 2026. At 500 mg NMN and an 8-week treatment period, that trial would need to show a minimum N3 increase of at least 12 minutes to be clinically meaningful by the standards set in the behavioral sleep medicine literature.