Do Health Supplements Really Work? An Evidence-Based Appraisal of Arginine, Collagen, NMN, and More

The Evidence-Based Truth About Health Supplements: Do Not Confuse Presence, Absorption, and Clinical Efficacy

The conclusion in one sentence

Health supplements must be evaluated separately by product, formulation, population, and clinical endpoint. The presence of an ingredient, its absorption, or a change in a surrogate marker does not automatically prove that it treats disease, slows aging, or improves overall health. PMID: 42255715

The central message of Professor Lin’s original article

Professor Ching-Shun Lin has long reminded readers not to translate observational associations, cell experiments, animal studies, mechanistic accounts of ingredients, or marketing language directly into established efficacy in humans. This framework remains applicable to claims that arginine increases height, NMN slows aging, bitter melon peptides lower blood glucose, skin carotenoid readings measure health, and salmon nasal cartilage proteoglycans broadly treat disease.

Even if dietary studies in children find an association between arginine intake and growth, they do not prove that arginine supplementation will reliably make children taller. The relevant research itself did not establish arginine supplementation as a clinical strategy. PMID: 23046689

NMN appeared in human metabolic research long ago, so contemporary researchers should not be described as its discoverers. Existing human studies are also insufficient to prove that NMN can treat multiple diseases of aging or cancer, or substantially extend human life. PMID: 14832305 PMID: 36482258 PMID: 37619764

The status of bitter melon preparations as supplements, together with their advertising, is not enough to elevate them to an established treatment for diabetes. PMID: 22895968 PMID: 38274207

Three levels of evidence that are often confused

First, being able to measure an ingredient does not mean that the body is healthier. Reflectance or Raman spectroscopy can measure skin carotenoids noninvasively and use them as a biomarker of dietary exposure. A single reading, however, is not a composite score of whole-body antioxidant status or overall health. PMID: 42255715

Second, the fact that something is digested after ingestion does not mean it must be entirely ineffective. Most dietary proteins are indeed hydrolyzed, but certain collagen-derived dipeptides and tripeptides can enter the plasma intact. Absorption establishes exposure only; effects in the target tissue and clinical benefits still require separate proof. PMID: 39149544

Third, improvement in a surrogate marker does not mean that a disease outcome has improved. Newer meta-analytic findings for pine bark extract are positive for certain cardiometabolic surrogate markers, but study quality, industry funding, and hard clinical endpoints still warrant careful scrutiny. PMID: 39987124

Oral collagen in the context of scientific progress

Professor Lin’s earlier articles, based on protein digestion and the limited human research available at the time, rejected the claim that oral collagen necessarily reaches the skin directly and produces cosmetic benefits. That historical position should not be rewritten. The literature has since advanced, however, and later syntheses of randomized trials show positive signals that oral hydrolyzed collagen or collagen peptides may benefit skin hydration, elasticity, radiance, and wrinkles. It is therefore no longer appropriate to summarize the evidence as showing that oral collagen is categorically of no benefit to the skin. PMID: 41809116 PMID: 41924746

This update does not mean that every brand, formulation, and claim is effective. Nor does detecting a particular small peptide in the blood justify the direct assertion that it must reach the target tissue and carry out the mechanism described in marketing materials. PMID: 39149544

Topical collagen is a separate issue. Native collagen is a macromolecule, and the intact skin barrier limits its passive penetration. Without a specialized delivery technology, surface coating and moisturization should not be described as collagen becoming part of the skin. PMID: 41127992

Oral enzymes and peptides cannot be judged with absolute statements

Professor Lin’s original article questioned enzyme drinks and oral peptides on the basis of gastric acid, digestion, and the intestinal barrier. This remains important for preventing exaggerated promotion. With the advancement of science, the more precise formulation is that the gastrointestinal environment and dosage form do constrain oral enzymes, but this cannot be extended into a claim that no oral enzyme can ever have an effect. Lactase has specific established uses, whereas the evidence for bromelain remains limited. PMID: 41471052 PMID: 42245557

Literature reporting that orally administered bromelain can retain proteolytic activity in serum requires an update to the absolute proposition that it can “never enter the circulation in an active form.” This still does not mean that bromelain has been proven to treat multiple diseases. PMID: 37157782

Oral delivery of peptide drugs continues to face major barriers, and most peptides are still not administered orally. Nevertheless, approved oral peptide drugs now exist, so the historical statement that “there are no oral peptide drugs” should be understood as a description that the literature has superseded. PMID: 42556625

Signals, insufficient evidence, and claims that should not be extrapolated

Human trials of salmon nasal cartilage proteoglycans have reported improvement in some skin measures, so it is no longer accurate to say that oral consumption “will have no effect whatsoever.” These findings, however, cannot be extrapolated into claims that the product treats cancer, the common cold, or injuries. PMID: 40613544

A small randomized human trial of fucoidan has now appeared, making the statement that there are “no human clinical trials whatsoever” another point on which the literature has been updated. A single small trial remains insufficient to support broad health-benefit claims. PMID: 37405785

Studies of astaxanthin show signals in some recovery biomarkers, but primary measures of athletic performance do not necessarily improve. The evidence as a whole still supports the cautious conclusion that preliminary signals may exist and further trials are needed. PMID: 42197030

How to make a purchasing decision

Start by rewriting the product advertisement as a set of testable questions. Did the research actually use the same ingredient and formulation? Were the participants similar to you? Were the outcomes symptoms and disease events, or merely surrogate markers? Has an independent team replicated the effect? Do the researchers or manufacturers have conflicts of interest?

A systematic review of pine bark extract could not reach firm conclusions about efficacy and safety because of study size, differences in outcome measurement, and reporting quality. Manufacturer participation in producing the evidence also makes independent replication of positive findings especially important. PMID: 32990945 PMID: 38757130

Findings on BCAA for athletic performance and body composition in healthy adults are mixed, so it should not be presented as essential for everyone. By comparison, creatine has more consistent positive evidence for muscle outcomes in resistance training and older adults, although people with pre-existing kidney disease should still receive an individualized and cautious assessment. PMID: 41346907 PMID: 42141930 PMID: 41404326

The most defensible conclusion is neither that “all supplements work” nor that “all supplements are useless.” It is to accept only the narrow portion of a claim that the evidence actually covers, while clearly distinguishing what remains unproven, what has changed as the literature evolved, and what now has evidence of a specific benefit.