How to Make a Supplement That Lowers Death Look Deadly

The press just took one of the most consistently life-extending supplements ever studied — and taught you to be afraid of it

On July 7, 2026, The Washington Post ran a headline engineered to plant a single seed of doubt: “Should you take glucosamine for joint pain? A study suggests it might have risks.” The on-page deck softened it only slightly — “Many people use glucosamine, though evidence of its effectiveness is mixed at best.” The message a busy reader takes away in three seconds is unmistakable: the joint pill in your cabinet might be hurting you.

Here is what that reader will almost certainly never learn from the coverage. Glucosamine is one of the most consistently mortality-lowering compounds ever tracked in large human populations. In nearly half a million adults in the UK Biobank, regular users had a 15% lower risk of dying from any cause, with fewer deaths from cardiovascular disease, cancer, respiratory illness, and digestive disease alike. A separate U.S. national cohort found regular users about 39% less likely to die of any cause and 65% less likely to die of cardiovascular disease after adjustment. A dedicated UK Biobank cardiovascular analysis in The BMJ found a 15% lower risk of total cardiovascular events, including lower risk of coronary heart disease, stroke, and cardiovascular death.

That is the actual weight of the evidence on this “risky” supplement. So how does a compound repeatedly linked to living longer get rebranded, in a single news cycle, as something you should fear?

By doing exactly what the coverage did.

The study everyone is citing — and what it actually found

The trigger is a single paper in Nature Metabolism (June 9, 2026), titled “Hyperglycosylation is a metabolic driver of Alzheimer’s disease,” from a University of Florida team led by biochemist Ramon Sun. Strip away the headlines and here is precisely what it reported.

The researchers pulled the clinical records of roughly 25,000 patients who already had Alzheimer’s disease or a related dementia. Among those patients, the ones who had documented glucosamine use for at least a year after their diagnosis were about 25% more likely to die over the following decade. In a separate group with mild cognitive impairment, glucosamine use was tied to a 25% higher chance of progressing to full dementia — but, notably, no increase in mortality at all. They then gave glucosamine to mice bred to carry Alzheimer’s genes and saw worse social memory and more brain “glycans”; mice without those genes, given the same supplement, showed no such effect.

Read that again, because the entire story lives in what it does not say. This was a study of sick brains — people already diagnosed, and mice genetically engineered to get sick. It was not a study of you, or of the millions of healthy adults who take glucosamine for an aching knee. As the Post itself concedes, buried mid-article: “The research did not look at healthy participants, but previous studies of glucosamine use in the general population did not demonstrate worsening of any disease.” Even the senior author frames it as a caution for people who “have any signs of dementia” to talk to their doctor — not as a warning for the general public.

The finding may well be real and worth studying. Glycan biology in the diseased brain is a legitimate frontier. But that is a narrow, mechanism-flagged signal in an already-vulnerable population. It is not a safety verdict on a supplement — and it certainly is not the story a headline about whether you “should take glucosamine for joint pain” implies.

The confound the study’s own critics named out loud

Here is the part that should end the scare on its own, and it comes not from me but from the researchers quoted in the Post’s own article.

The human data are correlational — they cannot show causation. And by the admission reported in the piece, the researchers “only accounted for sex, age and ethnicity as potential confounding factors.” That is a strikingly thin set of adjustments for a claim about who lives and who dies.

Angela Zivkovic, a nutrition professor at UC Davis who reviewed the work, put the obvious problem plainly: “people who take glucosamine because of joint pain may move less and be more physically frail, which is associated with higher dementia risk.” Think about who reaches for glucosamine after a dementia diagnosis: the patients whose joints hurt the most, who are the least mobile, the most frail, the sickest. Frailty and immobility are themselves powerful predictors of faster decline and death. A study that adjusts only for age, sex, and ethnicity has no way to separate “glucosamine harmed them” from “the people who needed glucosamine were already worse off.” Zivkovic was blunt about the mechanistic claims too: the team “used a really cool tool but applied it in a way that kind of missed a lot of the underlying biology,” and glycomics remains “the last frontier” that “we have very little understanding of how it works.”

This is the oldest trap in observational research — mistaking the marker of illness for its cause. It is the same error that once “proved” that people who use canes fall more often. The cane did not cause the falls; frailty caused both.

What honest coverage would have said

A fair headline had all the ingredients sitting right there. It could have read: “A supplement linked to living longer may behave differently in brains already fighting Alzheimer’s.” That is true, interesting, and appropriately narrow. Instead, the framing was inverted — the reassuring, population-scale benefit was demoted to a parenthetical, and a preliminary finding in sick brains was promoted to the headline as a reason for the general public to be wary.

Let me be scrupulously fair to the other side, because a case worth making is one that can survive its own objections.

The general-population mortality findings are themselves observational, and they have real critics. A 2022 analysis argued that the striking mortality reductions seen in these “prevalent user” cohorts can be inflated by a statistical artifact called collider bias — the same class of confounding problem, pointed in the opposite direction. Healthy, health-conscious people are more likely both to take glucosamine and to live longer, so some of the apparent benefit may reflect who chooses to take it rather than the compound itself. A Mendelian randomization analysis using genetic proxies even found glucosamine associated with higher risk of certain outcomes, cautioning against treating it as a blanket preventive. And glucosamine’s benefit for the very thing people buy it for — joint pain — is genuinely modest; large randomized trials often fail to beat placebo, with clearer signal mainly in moderate-to-severe knee osteoarthritis.

All of that is true, and all of it cuts in the same direction as my point rather than against it: observational data cannot establish causation — in either direction. If the mortality benefit deserves the asterisk of “correlation, not proof,” then the mortality harm in a 25,000-patient chart review that adjusted for only three variables deserves a far bigger one. You cannot cite the weakness of observational data to dismiss decades of favorable signals and then, in the same breath, treat a single, thinly-adjusted observational finding as a reason for the public to worry. That is not skepticism. That is a thumb on the scale.

The pattern is the point

This is not really about glucosamine. It is about a reflex.

When a cheap, off-patent, over-the-counter compound accumulates a decade of evidence linking it to lower mortality, that news travels quietly, in the back pages, hedged into oblivion. When a single preliminary study surfaces a possible harm — in the sickest patients, adjusted for almost nothing — it becomes a headline aimed squarely at the healthy majority the study never examined. The benefit is always “just an association.” The risk is always a reason to reconsider.

Ask yourself who is served by teaching millions of people to distrust a supplement that, on the best population-scale evidence available, tracks with fewer deaths — and whose worst-case risk profile appears confined to people already living with a devastating disease. The prudent takeaway from the actual data is close to the opposite of the headline: if you have Alzheimer’s or clear signs of cognitive decline, talk to your doctor about glucosamine — a reasonable, narrow caution. And if you don’t, the science that has followed hundreds of thousands of people for years still points, on balance, toward benefit.

The curious thing was never whether glucosamine “might have risks.” Everything carries risk, including the NSAIDs the same article admits are “dangerous for older people.” The curious thing is how eagerly a supplement associated with living longer got dressed up as a threat — and how few readers will ever be given the numbers to see it.

NOTE: I’ve been indexing research on Glucosamine’s therapeutic potential for two decades. Take a look at some of the studies below.

---

Sources referenced throughout: Washington Post, July 7 2026; Nature Metabolism study, s42255-026-01538-4; UK Biobank all-cause & cause-specific mortality, Annals of the Rheumatic Diseases 2020 (PubMed); NHANES all-cause & CVD mortality, JABFM; UK Biobank cardiovascular disease, The BMJ; collider-bias critique, PubMed 36029480; “note of caution,” ARD 2022; Mendelian randomization, Frontiers in Genetics 2024.