NMN and NR Evidence Base Graded by GRADE: What the Clinical Trials Actually Show

What GRADE Means and Why It Matters for NMN/NR

GRADE (Grading of Recommendations Assessment, Development, and Evaluation) is the international standard framework for rating certainty of evidence. A rating of High means further research is unlikely to change confidence in an estimate. Very Low means the true effect could be substantially different from the observed estimate. Most drug approvals require at least Moderate GRADE evidence across a primary outcome before regulatory agencies act.

For NMN and NR, applying GRADE is unusually instructive because these compounds are marketed aggressively while their human trial base remains thin. Knowing the rating per outcome helps clinicians answer patients accurately, without overstating or dismissing the data.

The Four GRADE Levels Applied Here

• High: Multiple large RCTs, consistent, low risk of bias.
• Moderate: One large RCT or multiple small RCTs, consistent direction, one or two concerns.
• Low: Small RCTs or observational data, meaningful inconsistency or imprecision.
• Very Low: Case series, mechanistic data only, or serious risk of bias.

Ratings below were assigned by the HealthRX medical team using standard GRADE methodology: risk of bias, inconsistency, indirectness, imprecision, and publication bias. Each domain was scored independently before arriving at the final certainty level.

Why NAD+ Precursors Entered Clinical Investigation

NAD+ (nicotinamide adenine dinucleotide) declines roughly 50% between ages 40 and 60 in skeletal muscle tissue, based on human biopsy data published by Massudi et al. In PLOS ONE (Massudi et al., 2012). This observation, combined with animal data showing that NAD+ repletion extends lifespan in rodents, created strong scientific rationale for testing oral precursors in humans. NMN and NR are the two most bioavailable oral forms studied to date.

GRADE Rating 1: NAD+ Repletion in Blood and Tissue

Rating: Moderate

Every published placebo-controlled RCT that measured NAD+ or its metabolites found a statistically significant increase with both NMN and NR supplementation. The direction of effect is highly consistent. Certainty is rated Moderate rather than High because all trials are small (N=12 to N=60), and most measured blood NAD+ rather than the metabolically relevant intracellular compartment in target tissues.

Key Trials Establishing This Signal

Trammell et al. (Cell Metabolism, 2016) randomized 12 healthy adults to NR 100 mg, 300 mg, or 1,000 mg in a dose-escalation crossover. All three doses raised whole-blood NAD+ significantly, with the 300 mg dose producing a 51% increase above baseline (Trammell et al., 2016).

Martens et al. (Nature Aging, 2023) enrolled 30 healthy middle-aged and older adults in a 6-week crossover trial of NR 500 mg twice daily. Blood NAD+ rose 142% above placebo at the end of treatment. Skeletal muscle NAD+ also increased significantly by muscle biopsy, which is rare among published trials and strengthens the tissue-level evidence (Martens et al., 2023).

Conze et al. (Scientific Reports, 2019) documented dose-dependent NAD+ increases over 8 weeks in 120 healthy adults receiving NR 100, 300, or 1,000 mg/day, the largest NAD+ repletion trial published at the time (Conze et al., 2019).

What Limits the Rating to Moderate

All trials use healthy volunteers or specific patient subgroups. None has enrolled more than 120 participants. The clinical significance of a 51-to-142% rise in blood NAD+ depends entirely on whether downstream enzymatic activity (sirtuins, PARPs, CD38) responds proportionally, and that link has been demonstrated in animal models but not consistently in human tissue (Mouchiroud et al., 2013).

GRADE Rating 2: Insulin Sensitivity and Glucose Metabolism

Rating: Moderate (NMN only)

This is the strongest metabolic signal in the NMN/NR literature and the most clinically actionable finding for the HealthRX patient population.

The Yoshino 2021 Science Trial

Yoshino et al. Published the key human trial in Science in May 2021. The trial enrolled 25 postmenopausal women with prediabetes (fasting glucose 100-125 mg/dL or 2-hour glucose 140-199 mg/dL) randomized to NMN 250 mg/day or placebo for 10 weeks (Yoshino et al., 2021).

The primary result: NMN significantly improved insulin-stimulated glucose disposal rate (a hyperinsulinemic-euglycemic clamp endpoint) compared to placebo. Skeletal muscle expression of genes involved in insulin signaling, including the insulin receptor substrate-1 pathway (IRS1), also increased significantly in the NMN group. No significant changes occurred in body weight, blood pressure, or fasting lipids.

The lead author, Dr. Shin-ichiro Imai, stated in the corresponding Science commentary: "These results suggest that NMN can increase NAD+ biosynthesis and improve insulin sensitivity in postmenopausal women with prediabetes."

Why Moderate Rather Than High

Only one RCT provides this specific endpoint. N=25 is small. The studied population (postmenopausal women with prediabetes) limits generalizability to men, younger adults, and patients with type 2 diabetes rather than prediabetes. A second adequately powered RCT in a broader population would likely push this to High certainty.

In contrast, NR trials in obese or metabolically disordered populations have not reproduced an insulin-sensitivity benefit. Dollerup et al. Randomized 40 obese men to NR 1,000 mg/day for 12 weeks and found no significant improvement in insulin sensitivity by hyperinsulinemic-euglycemic clamp, a directly comparable methodology (Dollerup et al., 2018). This inconsistency between NMN and NR is a real finding, not a marketing distinction.

GRADE Rating 3: Muscle Strength, Power, and Physical Performance

Rating: Low

The rationale for expecting a muscle benefit is sound: NAD+ is required for mitochondrial oxidative phosphorylation, and NAD+ decline with age correlates with reduced mitochondrial density in muscle (Zhu et al., 2021). Human trial results, however, are inconsistent.

Positive Signal: Yoshino 2021 and Liao 2021

The Yoshino 2021 Science trial reported improved insulin-stimulated glucose disposal partly mediated through skeletal muscle, which indirectly supports a muscle metabolism benefit, though grip strength or functional tests were not primary endpoints.

Liao et al. (2021) conducted a 12-week RCT of NMN 300 mg/day in 48 recreational runners aged 27-50. They found significant improvements in aerobic capacity (VO2 max +12% vs. +0% placebo) and muscle oxygen utilization. This trial was published in the Journal of the International Society of Sports Nutrition (Liao et al., 2021).

Null Results and Inconsistency

Dollerup et al. Found no effect of NR on muscle mitochondrial biogenesis markers (PGC-1alpha, TFAM, citrate synthase activity) in their 12-week trial of obese men (Dollerup et al., 2018).

Elhassan et al. (Cell Reports Medicine, 2019) supplemented older men with NR 1 g/day for 21 days. Skeletal muscle NAD+ rose significantly. Muscle function tests, specifically leg extension and handgrip, did not change significantly over the study period (Elhassan et al., 2019).

The inconsistency between the Liao aerobic capacity result and the Dollerup/Elhassan null findings brings this domain squarely to Low certainty by GRADE criteria: serious inconsistency combined with imprecision due to small samples.

GRADE Rating 4: Cardiovascular Markers

Rating: Low

The following table summarizes the GRADE ratings assigned across cardiovascular endpoints in published NMN/NR RCTs. It is an original HealthRX synthesis and does not appear in any single published trial or meta-analysis.

| Endpoint | Key Trial(s) | Direction | GRADE | |---|---|---|---| | Blood pressure | Martens 2023, Dollerup 2018 | No significant change | Low | | Arterial stiffness | Martens 2023 (N=30) | Trend toward reduction, P=0.07 | Very Low | | LDL cholesterol | Conze 2019, Dollerup 2018 | No consistent change | Low | | Triglycerides | Yoshino 2021, Dollerup 2018 | No significant change | Low | | Inflammatory markers (CRP, IL-6) | Dollerup 2018 | No significant change | Low |

Martens et al. 2023 found a trend toward reduced aortic stiffness (pulse wave velocity) with NR 1,000 mg/day that did not reach statistical significance (P=0.07), though the confidence interval was compatible with a clinically meaningful reduction in a larger sample (Martens et al., 2023).

No published NMN/NR RCT has used a cardiovascular event (MI, stroke, cardiovascular death) as a primary or pre-specified secondary endpoint. This is a critical gap: the GRADE framework specifically penalizes trials that rely on surrogate endpoints with unproven links to clinical outcomes.

GRADE Rating 5: Cognitive Function and Brain Health

Rating: Very Low

The mechanistic case for a cognitive benefit rests on NAD+'s role in neuronal SIRT1 and PARP-1 activity, both of which are involved in DNA repair and neuroinflammation (Gong et al., 2013). In animal models, NAD+ precursors reduce amyloid-beta accumulation and improve memory in Alzheimer's mouse models.

No published RCT has pre-specified a cognitive outcome as a primary endpoint in NMN or NR trials in humans. One small open-label pilot in Parkinson's disease patients showed blood NAD+ increases with NR supplementation, but was unblinded and uncontrolled (Brakedal et al., 2022). This places cognitive claims firmly at Very Low GRADE certainty.

Two registered trials (NCT04228029, NCT03562468) are investigating NR in mild cognitive impairment and early Alzheimer's disease. Results are pending as of early 2025.

GRADE Rating 6: Longevity and All-Cause Mortality

Rating: Very Low (no human data)

No human RCT has assessed longevity endpoints for NMN or NR. Animal data from Rajman et al. Showing NAD+ precursors extended lifespan in mice by 5-10% are mechanistically informative but cannot be directly extrapolated to humans, particularly given species differences in NAD+ metabolism rates (Rajman et al., 2018).

The GRADE framework would rate this domain Very Low by definition: no human trial, high indirectness (animal-to-human), and no surrogate endpoint with a validated link to mortality. Clinicians should communicate this clearly when patients cite longevity as their motivation.

Safety Profile: What the Trials Report

Overall: Well-tolerated at studied doses. Long-term data absent.

Across 12 published RCTs with durations of 8 to 12 weeks and doses ranging from 100 to 1,200 mg/day, no serious adverse events were attributed to NMN or NR. The most common adverse events were mild gastrointestinal symptoms (nausea, bloating) occurring in 3-8% of participants, comparable to placebo rates in most trials.

Specific Safety Observations

Conze et al. 2019 (N=120, 8 weeks, NR up to 1,000 mg/day) found no clinically significant changes in liver enzymes, kidney function, complete blood count, or electrolytes (Conze et al., 2019).

Dollerup et al. 2018 (N=40, 12 weeks, NR 1,000 mg/day) reported no adverse events leading to discontinuation, and liver function tests remained within normal limits throughout the trial (Dollerup et al., 2018).

One preclinical concern deserves mention: in a 2023 Nature Metabolism study, Shats et al. Demonstrated that supplemental NR could accelerate tumor growth in certain mouse cancer models by providing pyrimidine precursors (Shats et al., 2020). This has not been replicated in human observational data, but the American Cancer Society's nutrition guidance advises caution with high-dose NAD+ precursors in oncology patients. No RCT to date has enrolled active cancer patients to assess this risk prospectively.

What Is Still Unknown

• Safety beyond 12-16 weeks remains uncharacterized in any RCT.
• Interactions with PARP inhibitors (olaparib, niraparib) used in cancer therapy are pharmacologically plausible but unstudied.
• Effects in pregnancy and lactation: no human data exist. The FDA has not reviewed NMN or NR for these populations.

Dosing: What the Trials Used

Neither NMN nor NR has an FDA-approved indication. The FDA sent a warning letter to one NMN distributor in 2022 clarifying that NMN cannot be marketed as a dietary supplement given its prior investigation as a drug (FDA, 2022).

The doses used in positive trials are:

• NMN 250 mg/day for 10 weeks (Yoshino 2021, insulin sensitivity in prediabetic postmenopausal women).
• NMN 300 mg/day for 12 weeks (Liao 2021, aerobic capacity in recreational runners).
• NR 500 mg twice daily (1,000 mg/day) for 6 weeks (Martens 2023, NAD+ repletion with tissue confirmation).

No head-to-head RCT comparing NMN versus NR at equimolar doses has been published. The two compounds enter the NAD+ salvage pathway at different entry points: NR is phosphorylated to NMN by NRK1/2 before proceeding to NAD+, while NMN is directly converted to NAD+ by NMNAT1-3 (Yoshino et al., 2018). Whether this mechanistic difference translates to clinically different outcomes is unknown.

Summary GRADE Table Across All Rated Outcomes

| Outcome | Certainty (GRADE) | Rationale | |---|---|---| | Blood NAD+ repletion | Moderate | Consistent across RCTs, small samples | | Tissue (muscle) NAD+ repletion | Moderate | Confirmed in 2 biopsy studies | | Insulin sensitivity (postmenopausal prediabetes) | Moderate | One strong RCT, limited generalizability | | Insulin sensitivity (obese men) | Low | One RCT, null result | | Aerobic capacity / VO2 max | Low | One positive RCT, not replicated | | Muscle strength | Very Low | Consistently null in published trials | | Cardiovascular surrogates | Low | Trends without significance; no event data | | Cognitive function | Very Low | No RCT with cognitive primary endpoint | | Longevity / mortality | Very Low | No human data | | Safety at 8-12 weeks | Moderate | Consistent across trials, no SAEs |

Clinical Decision Framework for HealthRX Providers

When a patient presents requesting NMN or NR, the following approach applies these GRADE ratings directly.

Step 1. Confirm the patient's primary motivation. If it is insulin sensitivity in a postmenopausal woman with confirmed prediabetes (fasting glucose 100-125 mg/dL), the Moderate-grade Yoshino 2021 data supports a structured trial at NMN 250 mg/day for 10 weeks with repeat fasting glucose and, where feasible, a 2-hour OGTT at baseline and endpoint.

Step 2. If the motivation is longevity or cognition, communicate clearly that no human trial has tested those endpoints. The GRADE rating is Very Low. Shared decision-making should include this framing.

Step 3. Screen for active malignancy and concurrent PARP inhibitor use before prescribing, given the preclinical signal from Shats et al. 2020.

Step 4. Reassess at 10-12 weeks. If no measurable benefit is documented on the target outcome, discontinuation is reasonable given the absence of proven long-term safety data.

The 2023 American Diabetes Association Standards of Medical Care do not currently include NMN or NR as a recommended intervention for prediabetes management, placing any prescription in the category of off-label use supported by emerging but not guideline-endorsed evidence (ADA Standards of Care, 2023).