How Should We Evaluate Vitamins and Supplements? An Evidence-Based Review from NMN Anti-Aging Claims to Supplementation Risks

Vitamins and Supplements: The Existence of Research Does Not Establish Effectiveness

The conclusion in one sentence

For adults without a clearly established deficiency or specific indication, supplements should not be presumed necessary for extending life, reversing aging, or preventing chronic disease. Every claimed benefit must be tested individually in relation to the product, population, and clinical endpoint. PMID: 35727272；PMID: 38922615

NMN: More human evidence, but no proof of age reversal

When Professor Ching-Shun Lin published his original article, he questioned the promotion of NMN as a means of “anti-aging and age reversal” and distinguished animal experiments and preliminary human studies from genuine clinical benefits. His article and historical position should be understood in the context in which they were published, not rewritten as claims based on evidence that emerged later.

The literature has been updated: subsequent human trials and meta-analyses have indeed made the evidence base more complete. It is therefore no longer appropriate to state categorically that “not even the most preliminary human evidence exists.” Nevertheless, newer overall analyses still have not shown that NMN provides robust anti-aging or metabolic benefits in adults generally. The central judgment remains unchanged: claims of age reversal or life extension are not supported. PMID: 39531138；PMID: 40275690；PMID: 42514320

Some studies have observed improved muscle insulin sensitivity in a specific group of women with prediabetes, while others have reported weak signals in walking speed or measures of insulin resistance. These are preliminary findings that require confirmation; they cannot be extrapolated to mean that everyone who takes NMN will become younger. PMID: 33888596；PMID: 39185644

The overall evidence in healthy adults has not shown consistent, significant improvements in common clinical measures such as blood glucose, blood lipids, or blood pressure. Being generally tolerable in the short term does not mean that NMN works over the long term, much less that it extends life. PMID: 42514320；PMID: 39531138

The key to evaluating supplements: Do not substitute one endpoint for another

“Increasing a molecule in the body,” “changing a biological-aging clock,” and “lowering an oxidative-stress marker” are not synonyms for mortality, disease incidence, disability, or quality of life. If a study measures only a surrogate marker, its conclusion must remain limited to that marker; it cannot be elevated directly into a claim of treating disease, extending life, or reversing aging.

This is the most common gap in claims about both NMN and AKG. The original evidence that AKG might extend life came mainly from mice, while information from human trials remains insufficient to establish efficacy. Changes in the DNA methylation clocks of product users also cannot be equated with actual increases in lifespan. PMID: 32877690；PMID: 40819772；PMID: 34847066；PMID: 39077104

Similarly, NR can raise NAD+ levels in humans, and short-term trials indicate that it is generally tolerable, but those findings alone do not prove life extension. With the literature now updated, functional signals have appeared in certain patient populations, so it is no longer accurate to say that there are “absolutely no clinical benefits of any kind.” More precisely, the limited signals still do not support advertising that promises rejuvenation. PMID: 29992272；PMID: 31278280；PMID: 38871717

Vitamins are not a chronic-disease insurance policy that everyone needs

In generally healthy adults without a known nutritional deficiency, routine vitamin and mineral supplementation has not been shown to clearly reduce mortality or cardiovascular risk, and there is no evidence that multivitamins extend life. This does not mean that nutrients are unimportant. It means that “the human body needs nutrients” must not be transformed into “everyone needs to take additional supplements.” PMID: 35727272；PMID: 39891733；PMID: 41565231；PMID: 38922615

There are still legitimate reasons to supplement, including a confirmed deficiency, particular medical conditions, or a clear risk created by one's dietary pattern. The need for people following vegan or strict plant-based diets to obtain vitamin B12 deliberately is an important example that cannot be erased by the broad statement that “supplements are usually unhelpful.” PMID: 41382333

Vitamin D must likewise be assessed by population and intended use. A large trial did not show that supplementation in adults generally reduced invasive cancer or major cardiovascular events. Nor should asymptomatic, community-dwelling adults be indiscriminately screened and given supplements. PMID: 30415629；PMID: 29671532

Conversely, excessive or inappropriate long-term vitamin D supplementation can cause hypercalcemia, toxicity, and kidney injury. High doses have also not demonstrated skeletal benefits in healthy adults whose baseline levels are not low. “Possibly unnecessary” and “always harmless if taken” are two different questions. PMID: 30294301；PMID: 37402617；PMID: 31454046

The evidence for vitamin C is also not all-or-nothing. Regular supplementation has not markedly reduced the incidence of colds in the general population, and starting it only after a cold begins has not produced consistent therapeutic effects. Regular supplementation may, however, shorten the duration of illness. Describing vitamin C as either a cure-all or entirely useless therefore goes beyond the evidence. PMID: 23440782

Collagen, coenzyme Q10, and other popular ingredients

Reconsidered in light of newer literature, the assertion that oral collagen “is digested and therefore cannot possibly have any physiological effect” does not align with current clinical evidence. Studies have reported measurable skin and related outcomes, although substantial limitations remain and the findings cannot be expanded into a claim of comprehensive anti-aging effects. This is an updated assessment made possible by scientific progress; it does not rewrite later evidence as an updated personal position of the professor. PMID: 41809116；PMID: 41924746

The evidence on coenzyme Q10 has also evolved. Newer meta-analyses suggest a more favorable evidence base than earlier research for its use as an adjunctive treatment in heart failure. Signals of improvement have also emerged for statin-associated muscle pain, although trial size, duration, and consistency continue to limit certainty. Greater absorption or a better result for a single measure still does not automatically translate into greater clinical benefit. PMID: 39462324；PMID: 41158831；PMID: 41789786

Selenium illustrates why “essential element” does not mean “the more supplementation, the better.” A large randomized trial did not support selenium supplementation to prevent major cancers in generally healthy men, and high or unnecessary supplementation is not risk-free. PMID: 19066370；PMID: 40243093

Safety and research credibility are also questions of effectiveness

Supplements can cause liver injury, and the absence of adverse effects on a package label is not proof of safety. An assessment should examine the ingredients, dose, interactions, duration of use, and the individual's medical conditions rather than relying solely on the word “natural” or a celebrity endorsement. PMID: 31720699；PMID: 34307603

Finally, the existence of a paper does not mean that the evidence is reliable. Paper mills and potentially fraudulent biomedical studies are growing problems, and retracted studies may continue to be cited after retraction. Study design, conflicts of interest, retraction status, and independent reproducibility should therefore all be part of any review of a supplement's claimed effects. PMID: 41611528；PMID: 40991025；PMID: 38512253

Practical decision principles

When considering any supplement, ask these questions in order: Do laboratory testing or a clinical assessment show that I have a deficiency? Did the study use the same ingredient and formulation? Were its participants similar to me? Was the endpoint a biomarker or an actual health outcome? Do the benefits outweigh the adverse effects, cost, and interactions? If these questions do not have clear answers, the most robust conclusion is usually neither “it definitely works” nor “it definitely does not,” but that the evidence does not yet support the product's claim.

This approach also preserves science's essential capacity to update itself. It respects the evidence and warnings available when the original article was written, while allowing newer, higher-level research to revise today's assessment. What changes is the conclusion drawn from the literature, not the historical position already published by an individual.