Evidence guide · South Africa

NMN drinks and "functional age": what a new 42-day study does and does not show

A sponsor-funded 42-day study of NMN drinks reported a 2.79-year fall in "functional age" and a clock drop of about 4.1 years, yet its exercise and cognitive comparisons were not significant. Where the numbers come from and how it compares with controlled NMN trials.

Review type: research summary. This page summarises published research and product labels, cited below.

A study of NMN-based drinks, published in the journal Nutrients on 23 October 2026 and added to PubMed on 14 November 2026, carries two numbers that suit an advertisement: a 2.79-year fall in "functional age" over 42 days, and a fall of about 4.1 years on an epigenetic clock. The same paper reports that its cardiopulmonary and cognitive comparisons between placebo and the most complex formula were not statistically significant, and its authors write that these age measures "do not establish biological age reversal".

Short answer: this is a small, open-label, sponsor-funded feasibility study. It is a reason to run a proper trial, not evidence that an NMN drink makes anyone younger, and its most quotable results are its least reliable.

What the study did

Forty-two Japanese adults aged 45 to 65, described as healthy on the basis of self-report, took part between 14 October and 26 December 2025. Two withdrew for personal reasons, leaving 40 in the analysis, 32 of them women. Each drank one 50 mL bottle a day after breakfast for 42 days: a placebo drink (10 people analysed), NMN alone (9), a core combination of NMN with D-beta-hydroxybutyrate (D-BHB, a ketone) and a ginseng-derived ingredient (10), or an expanded eight-ingredient formula (11) that added berberine, PQQ, coenzyme Q10, trans-resveratrol and spermidine. The final clinical study report lists 3,000 mg of beta-NMN per daily bottle in each NMN arm.

Abe Yoando Pharma Co., Ltd. funded the study and supplied the drinks, which the company developed and manufactured. The clinical research organisation SOUKEN Co., Ltd. handled recruitment, assessments, statistical analysis and the clinical report. The study was open-label, so participants were not blinded to what they drank, and the paper does not describe the whole four-arm study as a verified randomised controlled trial because allocation and concealment were not documented for all arms. The expanded-formula group was also older on average than the placebo group (60.8 v 55.8 years), which matters for age-sensitive measures.

What it found

  • Exercise testing: no statistically significant differences between placebo and the expanded formula in anaerobic threshold, workload, exercise oxygen uptake (VO2), carbon dioxide output, respiratory exchange ratio or ventilation. The direction was mixed: exercise VO2 changed by +0.0018 L/min with placebo and by -0.0585 L/min with the expanded formula.
  • Cognition: on a computerised battery (CogniTrax), several mean scores rose with the expanded formula and fell with placebo, for example the Neurocognitive Index (+3.09 v -1.90) and composite memory (+5.82 v -5.30). None of the placebo-versus-expanded-formula contrasts in these standardised scores was statistically significant (P = 0.801 and P = 0.495 for those two). The NMN-only group had the largest mean Neurocognitive Index rise (+13.11), and its contrast with placebo was not significant either (P = 0.062).
  • Blood markers: in a placebo-versus-expanded-formula subset, two of nine markers on an ageing panel differed nominally, CTACK (P = 0.015) and GDF-15 (P = 0.022), with no correction for multiple testing. GDF-15 rose with the formula, and the authors note it may reflect physiological stress in some contexts. Several other markers were below the assay's detection limit.
  • Safety: recorded adherence was 42 of 42 scheduled doses for all 40 analysed participants, and no serious adverse events were reported. The paper calls this recorded rather than objectively verified adherence. Participants were asked to report symptoms during dosing and for two weeks afterwards.

Where the "younger" numbers come from

The 2.79-year figure is not a measured biological age. Changes in 11 of the study's own outcomes (three exercise measures, four cognitive scores and four blood markers) were converted into age-equivalent units using fixed ageing slopes from the Baltimore Longitudinal Study of Aging framework, then combined. On that composite, estimated functional age fell by 2.79 years with the expanded formula and by 0.63 years with placebo. The paper calls the converted units "a mathematical transformation, not a validated estimate of a change in chronological or biological age and not a surrogate for clinical benefit". Because the composite is built from the study's own outcomes, including exercise and cognitive measures that did not differ significantly from placebo, it is not a second, separate confirmation of those results.

The 4.1-year epigenetic figure comes from a different group: eight employees aged 40 to 54 who took the candidate supplement for 42 days with no control group. The paper notes that prior NMN use was possible, that their methylation ages were already younger than their calendar ages before the intervention, that DNAm PhenoAge was one of several clocks examined and was not documented as chosen in advance, and that two post-intervention samples were excluded from a gene-expression heatmap for quality concerns. The authors write that the pilot "cannot be used to infer intervention efficacy or age reversal".

Reporting one favourable clock out of several is the pattern a Yale group cautioned against in Aging Cell in October 2026. When earlier single-clock intervention studies were extended to more clock models, no additional clocks reached statistical significance. The group proposed three signs that make a clock result more trustworthy: reliable principal-component-based clocks respond, mortality-trained or pace-of-ageing clocks respond, and multiple clocks move in the same direction.

Limitations the paper reports

  • Dose uncertainty: the caption to the formulation table says the listed quantities and the ginsenoside standardisation "require reconciliation against the actual 2025 administered-batch manufacturing records before publication". The methods add that an April 2026 patent-related formulation document gives different amounts of D-BHB, the ginseng preparation and resveratrol.
  • No primary endpoint: no single primary efficacy endpoint or plan for multiple comparisons was clearly documented, and the parent model for the reported four-arm Tukey comparisons was not documented in the materials available to the authors.
  • Eight ingredients at once: with NMN, D-BHB, ginseng, berberine, PQQ, CoQ10, resveratrol and spermidine in one drink, the design cannot show which ingredient had any effect, or whether they act together.
  • Practice and expectancy: the authors note that repeating a cognitive battery after six weeks invites practice effects, and that open-label designs can influence performance.
  • Competition framing: the blood-marker and functional-age analyses come from what the paper calls a competition-focused subset linked to an XPRIZE semifinal application, analysed separately from the four-arm comparisons.
  • Disclosures: the four listed authors give the Abe Yoan Aging Research Institute in Tokyo as their affiliation, one with a second affiliation at Nippon Medical School Hospital. The conflicts statement says the remaining authors declare no commercial or financial conflict, names an employee of SOUKEN who does not appear in the author list, and mentions patent application No. 2025-079089 concerning NMN plus BHB, for muscle, cognitive and anti-inflammatory effects.

How it compares with placebo-controlled NMN trials

The double-blind NMN trials cited here used lower daily doses. In a 60-day trial in 80 healthy middle-aged adults, 300, 600 and 900 mg a day raised blood NAD, six-minute walking distance increased more than with placebo, and a blood-test estimate of biological age (Aging.Ai 3.0) rose in the placebo group and stayed unchanged in the NMN groups; the lead author is an employee of Abinopharm and two other authors are employees of ABA Chemicals. In a 24-week trial in adults aged 65 and over, 1,250 mg a day raised NAD+ and related metabolites but did not improve the primary endpoint of auditory function or physical performance measures. In a four-week safety trial in 31 healthy adults funded by Mitsubishi Corporation Life Sciences, 1,250 mg once a day was well tolerated with no severe adverse events.

The 3,000 mg a day listed for the new study is well above the highest daily doses in those trials, and its safety information covers 42 days in 30 analysed participants across the three NMN arms. A 10-week placebo-controlled trial in 25 postmenopausal women with prediabetes who were overweight or obese found that 250 mg of NMN a day increased insulin-stimulated glucose disposal and muscle insulin signalling, while muscle NAD+ and physical function did not change.

Our view

  • The paper is candid about its limits, and the fair reading is the authors' own: it shows feasibility, not efficacy, ingredient effects, synergy or biological age reversal.
  • Expect "2.79 years younger" and "4.1 years younger" to appear in marketing. Neither is a measured change in biological age: one is a conversion of the study's own outcomes, the other an uncontrolled eight-person pilot.
  • If a multi-ingredient NMN drink is sold on this study, ask for the exact amount of each ingredient in the product you are buying, since the paper itself says the administered quantities still needed reconciling with batch records.
  • The useful next step is the one the authors describe: a registered, randomised, double-blind, placebo-controlled trial with a prespecified primary endpoint, verified formulation and longer safety follow-up.
  • Anyone pregnant, breastfeeding or taking regular medicines should speak to a doctor or pharmacist before starting any supplement.

South African rules

In South Africa, health supplements are regulated by SAHPRA as complementary medicines under the Medicines and Related Substances Act 101 of 1965, and SAHPRA publishes guidance on the safety and efficacy of health supplements. SAHPRA's complementary-medicines page limits health supplements to low-risk indications, such as general health maintenance, without reference to a specific disease. A label claim that an NMN product reverses biological age is not supported by the evidence above.

FAQ

Does NMN make you biologically younger?

No trial has shown that. In this study the age figures come from a mathematical conversion and an uncontrolled eight-person pilot, and the authors write that they do not establish biological age reversal.

Is 3,000 mg of NMN a day safe?

Controlled data at that dose are thin. The placebo-controlled trials above go up to 1,250 mg a day. The longest of those, in adults aged 65 and over, lasted 24 weeks. This study reported no serious adverse events over 42 days, which is too short and too small to settle long-term safety.

Did NMN alone improve memory in this study?

The paper does not report a significant memory benefit for NMN alone. The NMN-only group had the largest mean rise in the overall Neurocognitive Index, but the difference from placebo was not statistically significant (P = 0.062), in nine people who knew what they were drinking.

Sources

These are the documents this page relies on.

  1. Unlocking the Healthy Aging Potential of Novel NMN-Based Nutritional Formulations: Cardiopulmonary, Cognitive, and Aging-Related Outcomes in a 42-Day Open-Label Study. Nutrients. 2026;18(21):3490 (published 23 October 2026)Nutrients (MDPI)
  2. PubMed record 43030118: NMN-based nutritional formulations, 42-day open-label study (abstract, dates and affiliations)PubMed
  3. When to Trust Epigenetic Clocks: Avoiding False Positives in Aging Interventions. Aging Cell. 2026;25(10):e70756Aging Cell (PubMed 42850743)
  4. The efficacy and safety of beta-nicotinamide mononucleotide (NMN) supplementation in healthy middle-aged adults: a randomized, multicenter, double-blind, placebo-controlled, parallel-group, dose-dependent clinical trial. GeroScience. 2023;45(1):29-43GeroScience (PubMed 36482258)
  5. Sustained NAD+ augmentation by high-dose nicotinamide mononucleotide supplementation in older adults: a 24-week randomized, double-blind, placebo-controlled trial. npj Aging (published 25 September 2026)npj Aging
  6. Safety evaluation of beta-nicotinamide mononucleotide oral administration in healthy adult men and women. Sci Rep. 2022;12:14442Scientific Reports (PubMed 36002548)
  7. Nicotinamide mononucleotide increases muscle insulin sensitivity in prediabetic women. Science. 2021;372(6547):1224-1229Science (PubMed 33888596)
  8. Complementary medicinesSAHPRA
  9. Complementary medicines: health supplements safety and efficacySAHPRA