Cart
0
Read the product page for almost any cognitive supplement and you will meet a particular kind of sentence. The ingredient "increases BDNF," or "crosses the blood-brain barrier," or "reduces oxidative stress in neurons," and the sentence stops there, as though the mechanism were the destination. It is not. A mechanism is a hypothesis about why a compound might help, and the history of medicine is a long catalogue of mechanisms that were biologically elegant, entirely real, and clinically useless or worse. Learning to notice the gap between "here is how it could work" and "here is proof that it does" is the most protective habit a supplement buyer can build, because that gap is exactly where marketing lives.
The point is not that mechanisms are worthless. They are how science generates candidates worth testing. The error is treating the candidate as if it had already passed the exam. What follows is a tour of how often, and how badly, plausible mechanisms fail when someone finally runs the trial.
The graveyard between the bench and the clinic
Drug development is, in effect, an industrial-scale test of whether mechanisms translate into benefits, and the failure rate is sobering. Analyzing more than 400,000 clinical trial records, researchers estimated that a compound entering Phase 1 testing has roughly a 13.8 percent chance of eventual approval [1]. The overwhelming majority of these candidates entered trials on the strength of a compelling mechanism and promising laboratory data, and most still failed. When investigators catalogued why drugs collapse in the expensive late stages, lack of efficacy accounted for approximately half of all Phase 2 and Phase 3 failures during 2013 to 2015 [2]. These were not compounds that looked implausible on paper; they were compounds that worked exactly as designed at the molecular level and then did not help patients.
The chasm is well enough recognized to have a name. Reviewers describe a "valley of death" between preclinical promise and clinical reality, driven by oversimplified disease models and findings that do not reproduce once the stakes rise [3]. Supplements are not exempt from this biology. They are simply exempt from the requirement to prove they crossed the valley before being sold.
Why the test tube and the mouse mislead
Two features of preclinical research make mechanisms look more decisive than they are. The first is that animals are not small humans. A systematic review comparing treatment effects in animal experiments against the corresponding human trials found agreement in only about half of the cases examined, and in some the divergence was stark, as when corticosteroids that protected animals after head injury turned out to increase deaths in patients [4]. A later scoping review of reported animal-to-human concordance rates found they scattered across the entire possible range, from zero to one hundred percent, which is another way of saying that animal evidence alone predicts human outcomes no better than a coin in many domains [5].
The second feature is concentration. A mechanism demonstrated in a dish is demonstrated at whatever dose the researcher pipettes in, and that dose is frequently unreachable in a living person. Curcumin is the textbook case: it shows genuine and varied activity in cell studies, yet its oral bioavailability is so poor, through weak absorption, rapid metabolism, and fast elimination, that plasma levels remain very low even after large doses [6]. Resveratrol tells the same story in numbers. Many of its celebrated in vitro effects appear at concentrations up to around 200 micromolar, while human plasma after supplementation typically sits three to four orders of magnitude lower, a review of its clinical trials noted, which leaves a great deal of the mechanistic literature describing events that cannot physically occur in the body at normal intake [7]. A molecule that never reaches its target at an active concentration has a mechanism only in principle.
When the biomarker improves and the patient does not
Even inside the human body, moving a biological marker in the "right" direction is no guarantee of benefit, and this is where the mechanism-equals-benefit error becomes dangerous rather than merely disappointing. The methodological warning was issued plainly decades ago: a review of surrogate endpoints concluded that biological markers repeatedly served as correlates of clinical outcomes yet failed to predict the effect of treatment on those outcomes, so that changing the marker told you little about changing the disease [8].
The clinical proof came from trials that should be required reading. In the Cardiac Arrhythmia Suppression Trial, drugs that successfully suppressed the abnormal heartbeats thought to cause sudden death, the perfect mechanistic target, actually increased mortality, with deaths from arrhythmia or cardiac arrest reaching 5.7 percent on the drugs against 2.2 percent on placebo [9]. Years later, torcetrapib raised HDL cholesterol, the "good" cholesterol, by roughly 72 percent and lowered LDL, and by every biomarker it was a triumph, yet it increased the risk of death and cardiovascular events so clearly that the trial of more than 15,000 patients was halted [10]. In both cases the mechanism performed flawlessly and the people got worse.
The antioxidant hypothesis: a mechanism that harmed
No story illustrates the trap better than antioxidants. The reasoning was impeccable: oxidative damage contributes to cancer and heart disease, antioxidants neutralize that damage, therefore antioxidant supplements should prevent those diseases. Then the trials arrived. A large study in male smokers found that beta-carotene supplementation increased lung cancer incidence by 18 percent rather than reducing it [11], and a second major trial of beta-carotene with vitamin A was stopped early after the treatment group showed a 28 percent higher rate of lung cancer and higher overall mortality [12]. These were not fringe results. A Cochrane review pooling 78 trials and nearly 300,000 participants found that in the studies at lowest risk of bias, antioxidant supplements significantly increased mortality [13], and a separate meta-analysis of more than 135,000 people linked high-dose vitamin E specifically to increased all-cause death [14]. When vitamin E was put to a long-term randomized test for cardiovascular protection, it not only failed to help but was associated with a higher risk of heart failure [15]. A more perfect mechanism has rarely produced a more consistent record of null and harmful outcomes.
Moving the marker, missing the point
The pattern repeats wherever a supplement is sold on its effect on a number. Homocysteine, an amino acid whose elevated levels correlate with both cardiovascular disease and cognitive decline, seemed an obvious target, and B vitamins reliably lower it. They do so impressively: in a meta-analysis of eleven trials with cognitive data on 22,000 people, B vitamins reduced homocysteine by about 28 percent. The cognitive benefit was zero, with no significant effect on any domain of thinking [16]. The cardiovascular story matched it, as a Cochrane review of homocysteine-lowering trials in more than 70,000 participants found no meaningful reduction in heart attacks or death despite the marker falling as intended [17]. The biomarker was a spectator, not a lever.
Cognitive supplements sit squarely in this tradition. Ginkgo biloba has a genuinely attractive mechanism, improving blood flow and offering antioxidant activity in the laboratory, which is precisely why it was so widely sold for memory. When the mechanism was finally tested at scale, the Ginkgo Evaluation of Memory trial followed more than 3,000 older adults for roughly six years and found that the extract did not reduce the incidence of dementia or Alzheimer's disease compared with placebo [18]. The plausibility survived; the benefit never materialized.
How to read a mechanism claim without being fooled
None of this means you should dismiss mechanistic reasoning. It means you should file it correctly. When a label tells you an ingredient modulates a receptor, raises a growth factor, or scavenges a free radical, the honest translation is that someone has a reason to run a clinical trial, not that the trial has been run and won. The questions that follow are the useful ones. Has the compound been tested in humans, at a dose that actually reaches the target, measuring an outcome you care about rather than a blood marker standing in for it? Did the benefit hold up in a randomized, placebo-controlled design, and ideally in more than one? The evidence hierarchy that clinicians use exists for exactly this reason, placing controlled trials and their systematic reviews above mechanistic and animal data, because the lower rungs so often fail to predict the higher ones.
This is also why the order of evidence matters more than its sheer volume. A hundred cell-culture studies and a dozen mouse experiments can all point the same direction and still be overturned by a single adequately powered human trial, because the lower tiers share the same blind spots. They cannot see whether a compound actually reaches its target in a living person, at a dose the body tolerates, over a realistic span of time, without some countervailing effect that only a whole organism will reveal. Weight of evidence is not a headcount of supportive studies; it is a judgment about which studies were built to answer the question that matters.
A mechanism is the beginning of an argument, and a good one is genuinely exciting. What it is not, and can never be on its own, is proof that a capsule will do something for you. The compounds worth your money are the ones that closed the gap between could and does, and the single most clarifying thing you can ask of any supplement is not how it works, but whether anyone has actually shown that it does.
References
[1] Wong, C. H., Siah, K. W., & Lo, A. W. (2019). Estimation of clinical trial success rates and related parameters. Biostatistics, 20(2), 273–286. https://doi.org/10.1093/biostatistics/kxx069
[2] Harrison, R. K. (2016). Phase II and phase III failures: 2013–2015. Nature Reviews Drug Discovery, 15(12), 817–818. https://doi.org/10.1038/nrd.2016.184
[3] Seyhan, A. A. (2019). Lost in translation: the valley of death across preclinical and clinical divide – identification of problems and overcoming obstacles. Translational Medicine Communications, 4, 18. https://doi.org/10.1186/s41231-019-0050-7
[4] Perel, P., Roberts, I., Sena, E., Wheble, P., Briscoe, C., Sandercock, P., Macleod, M., Mignini, L. E., Jayaram, P., & Khan, K. S. (2007). Comparison of treatment effects between animal experiments and clinical trials: systematic review. BMJ, 334(7586), 197. https://doi.org/10.1136/bmj.39048.407928.BE
[5] Leenaars, C. H. C., Kouwenaar, C., Stafleu, F. R., Bleich, A., Ritskes-Hoitinga, M., De Vries, R. B. M., & Meijboom, F. L. B. (2019). Animal to human translation: a systematic scoping review of reported concordance rates. Journal of Translational Medicine, 17(1), 223. https://doi.org/10.1186/s12967-019-1976-2
[6] Anand, P., Kunnumakkara, A. B., Newman, R. A., & Aggarwal, B. B. (2007). Bioavailability of curcumin: problems and promises. Molecular Pharmaceutics, 4(6), 807–818. https://doi.org/10.1021/mp700113r
[7] Tomé-Carneiro, J., Larrosa, M., González-Sarrías, A., Tomás-Barberán, F. A., García-Conesa, M. T., & Espín, J. C. (2013). Resveratrol and clinical trials: the crossroad from in vitro studies to human evidence. Current Pharmaceutical Design, 19(34), 6064–6093. https://doi.org/10.2174/13816128113199990407
[8] Fleming, T. R., & DeMets, D. L. (1996). Surrogate end points in clinical trials: are we being misled? Annals of Internal Medicine, 125(7), 605–613. https://doi.org/10.7326/0003-4819-125-7-199610010-00011
[9] Echt, D. S., Liebson, P. R., Mitchell, L. B., Peters, R. W., Obias-Manno, D., Barker, A. H., et al. (1991). Mortality and morbidity in patients receiving encainide, flecainide, or placebo: the Cardiac Arrhythmia Suppression Trial. New England Journal of Medicine, 324(12), 781–788. https://doi.org/10.1056/NEJM199103213241201
[10] Barter, P. J., Caulfield, M., Eriksson, M., Grundy, S. M., Kastelein, J. J. P., Komajda, M., et al. (2007). Effects of torcetrapib in patients at high risk for coronary events. New England Journal of Medicine, 357(21), 2109–2122. https://doi.org/10.1056/NEJMoa0706628
[11] The Alpha-Tocopherol, Beta Carotene Cancer Prevention Study Group. (1994). The effect of vitamin E and beta carotene on the incidence of lung cancer and other cancers in male smokers. New England Journal of Medicine, 330(15), 1029–1035. https://doi.org/10.1056/NEJM199404143301501
[12] Omenn, G. S., Goodman, G. E., Thornquist, M. D., Balmes, J., Cullen, M. R., Glass, A., et al. (1996). Effects of a combination of beta carotene and vitamin A on lung cancer and cardiovascular disease. New England Journal of Medicine, 334(18), 1150–1155. https://doi.org/10.1056/NEJM199605023341802
[13] Bjelakovic, G., Nikolova, D., Gluud, L. L., Simonetti, R. G., & Gluud, C. (2012). Antioxidant supplements for prevention of mortality in healthy participants and patients with various diseases. Cochrane Database of Systematic Reviews, (3), CD007176. https://doi.org/10.1002/14651858.CD007176.pub2
[14] Miller, E. R., Pastor-Barriuso, R., Dalal, D., Riemersma, R. A., Appel, L. J., & Guallar, E. (2005). Meta-analysis: high-dosage vitamin E supplementation may increase all-cause mortality. Annals of Internal Medicine, 142(1), 37–46. https://doi.org/10.7326/0003-4819-142-1-200501040-00110
[15] Lonn, E., Bosch, J., Yusuf, S., Sheridan, P., Pogue, J., Arnold, J. M. O., et al. (2005). Effects of long-term vitamin E supplementation on cardiovascular events and cancer: a randomized controlled trial. JAMA, 293(11), 1338–1347. https://doi.org/10.1001/jama.293.11.1338
[16] Clarke, R., Bennett, D., Parish, S., Lewington, S., Skeaff, M., Eussen, S. J. P. M., et al. (2014). Effects of homocysteine lowering with B vitamins on cognitive aging: meta-analysis of 11 trials with cognitive data on 22,000 individuals. American Journal of Clinical Nutrition, 100(2), 657–666. https://doi.org/10.3945/ajcn.113.076349
[17] Martí-Carvajal, A. J., Solà, I., Lathyris, D., & Dayer, M. (2017). Homocysteine-lowering interventions for preventing cardiovascular events. Cochrane Database of Systematic Reviews, (8), CD006612. https://doi.org/10.1002/14651858.CD006612.pub5
[18] DeKosky, S. T., Williamson, J. D., Fitzpatrick, A. L., Kronmal, R. A., Ives, D. G., Saxton, J. A., et al. (2008). Ginkgo biloba for prevention of dementia: a randomized controlled trial. JAMA, 300(19), 2253–2262. https://doi.org/10.1001/jama.2008.683
