The supplement I trusted is now linked to Alzheimer's — and I don't know what to tell my patients

I’ve recommended glucosamine to patients for joint pain for years. Not every patient, but enough — the 62-year-old with creaky knees who doesn’t want to go near a cortisone injection, the 70-year-old who swears by it and just wants me to sign off. It was one of the easy recommendations. Over-the-counter, cheap, generally well-tolerated. The evidence for joint pain was modest but real. And when patients asked about long-term safety, I said what most GPs say: there’s no signal of harm.

A study out this week has me rethinking that conversation entirely.

Published June 9 in Nature Metabolism, researchers from the University of Florida tracked roughly 65,000 patients through a dozen years of health records and found something unsettling. Among people who already had mild cognitive impairment — the stage where memory slips start to become noticeable but daily function holds — those taking glucosamine were 25 per cent more likely to progress to full Alzheimer’s disease. For patients already diagnosed with dementia, glucosamine use was associated with 25 per cent higher mortality.

These aren’t small numbers on a small sample. The UF team pulled records from 24,000 Alzheimer’s and dementia patients and another 41,000 with mild cognitive impairment. Roughly 8 per cent of both groups were taking glucosamine — which, if those figures scale, means thousands of people taking a supplement that might be accelerating their decline.

Let me sit with that for a moment. Glucosamine isn’t niche. Roughly 40 million Americans take it every year. The Australian numbers are harder to pin down — supplements don’t require prescriptions, so nobody is counting — but walk through any chemist in Sydney and you’ll find it in the joint-health aisle, often on special, two bottles for $35. It’s one of those supplements that crossed over from “maybe it helps” to “everyone’s mum takes it” without anyone really noticing.

A lot of these people actively take an over-the-counter supplement that could be making their disease progression worse.
— Ramon Sun, lead author, University of Florida

That’s Ramon Sun, the senior author and director of UF’s Center for Advanced Spatial Biomolecule Research. His team didn’t just find an association and stop there — which is what makes this paper different from most supplement-safety scares. They traced a mechanism.
The mechanism, it turns out, sits in a cellular process most of us last heard about in first-year biochemistry. Glucosamine crosses the blood-brain barrier. Once inside the brain, it fuels something called hyperglycosylation — an overactive sugar-tagging system that attaches too many sugar structures to proteins in the Alzheimer’s brain. The N-glycan pathway is normally routine cellular housekeeping. In the Alzheimer’s brain, it runs hot. Matt Gentry, who chairs UF’s biochemistry department and co-authored the study, put it plainly.

What we found in Alzheimer’s is that this sugar-tagging system appears to be overactive. The Alzheimer’s brain is adding too many of these sugar structures, and this seems to contribute to the disease rather than protect against it.
— Matt Gentry, UF Department of Biochemistry and Molecular Biology

Then they went further, which is rare in supplement research. They tested whether blocking this pathway could reverse the damage. In mice, it did. Cognitive deficits improved when the hyperglycosylation pathway was shut down. That’s not a cure — mice aren’t humans, and the jump from murine models to clinical therapy is measured in decades, not months. But it moves the finding from “concerning association” to “identifiable mechanism with a plausible intervention target.”

Yasuhiko Kizuka, a glycobiologist at Gifu University in Japan who was not involved in the study, wrote in an accompanying commentary for Nature Metabolism that the work “provides new mechanistic insight” into how dietary glucosamine uptake might worsen Alzheimer’s pathology. That’s the kind of cautious endorsement that gets attention from researchers who spend their careers studying sugar biology — and it signals that the hyperglycosylation finding isn’t a one-off.

This is where it gets hard, because this study doesn’t land in a vacuum. It lands on top of earlier evidence that pointed the other way.

A 2023 UK Biobank study involving nearly half a million people found that regular glucosamine use was associated with an 18 per cent lower risk of developing dementia. A 2024 Mendelian randomisation study — which uses genetic variants to mimic a randomised trial — similarly suggested a protective effect against vascular dementia. For years the glucosamine story read like a quiet win: cheap, accessible, possibly neuroprotective.

The UF paper doesn’t contradict the earlier work so much as it draws a line through it. Sun told MedPage Today that the effect appears to depend entirely on whether neurodegeneration has already started.

The harm appears specific to a brain already in neurodegeneration. The supplement looks safe or possibly protective before disease onset, and harmful after.
— Ramon Sun

A supplement that helps protect a healthy brain might accelerate damage in an already-vulnerable one. Most people taking glucosamine don’t know which camp their brain falls into. MCI is, by definition, subtle — the kind of forgetfulness a patient dismisses as normal ageing, the kind a GP might not screen for unless prompted.

I keep coming back to the patient conversation. Say a 68-year-old woman sits in my office. She’s been taking glucosamine for ten years — her knees are better for it, or she believes they are. Her mother had Alzheimer’s. She’s worried, and she’s read the headline. What do I say?

There is no guideline for this yet. The TGA doesn’t test supplements for long-term neurological safety before they hit shelves — glucosamine is regulated as a complementary medicine, not a pharmaceutical. It doesn’t need to prove safety in the way a new drug does. The FDA framework is similarly light. Supplements that cross the blood-brain barrier occupy a regulatory blind spot: they’re treated as food even when they reach the one organ food isn’t supposed to directly alter. In Australia, complementary medicines sit on a list — the Australian Register of Therapeutic Goods — that requires evidence of manufacturing quality but not of long-term neurological safety. The bar for getting on a pharmacy shelf is closer to Weet-Bix than warfarin.

Sun, writing in The Conversation, argues that a standard placebo-controlled trial to settle the question is now off the table.

While glucosamine appears safe and potentially protective for a healthy brain, it may be harmful for a brain that is already experiencing cognitive decline.

You can’t randomise people with MCI to a supplement you now suspect harms them. That’s the ethical Catch-22 of post-market supplement safety: the moment the signal is clear enough to warrant a trial, the trial becomes unethical. The next-best design, Sun suggests, is tracking what happens when patients who are already taking glucosamine stop.

In the meantime, the burden falls on individual clinicians — and on patients themselves — to make calls with imperfect information.

So here’s where I’ve landed.

If you’re taking glucosamine and you have no cognitive concerns, no family history of dementia — the evidence doesn’t yet justify stopping. The UK Biobank data suggests glucosamine might be neutral or protective in healthy brains. The UF study’s harm signal was confined to brains where neurodegeneration had already begun. I’d keep an eye on the literature, but I wouldn’t clear out the medicine cabinet.

Now say there’s a strong family history. A parent with Alzheimer’s. An APOE4 genotype you already know about. That changes things. Sun told MedPage Today that “our data argue that conversation should happen.” I agree. If dementia runs in your family and you’re taking glucosamine daily, it’s worth sitting down with your GP and asking whether the joint-pain benefit is worth the uncertain cognitive risk. The answer isn’t obvious. But the question is.

And then there’s the group where the signal is clearest and the stakes are highest: patients with diagnosed MCI or early Alzheimer’s who are currently taking glucosamine. The 25 per cent progression risk is not trivial. The mechanism is biologically plausible. If you’re in this group, I’d be recommending you stop, and I’d be documenting that conversation carefully. Sun’s team is now following a cohort of patients who discontinue glucosamine to see whether their decline slows. Until those results arrive, caution is the only defensible position.

There’s something larger here too. It’s not just about glucosamine. We’re living through a moment where supplements have become ambient — part of the background noise of wellness, recommended by friends, Instagram, and chemists’ end-of-aisle displays. Collagen, turmeric, magnesium, ashwagandha, activated B12. Some of these have solid evidence. Many don’t. And the assumption that “natural” or “over-the-counter” equals “safe for everyone, forever” is, this week, harder to sustain.

The New Scientist asked last month whether turmeric and curcumin actually do anything. The answer was muddier than most supplement users would like. Earlier this year, dermatologists pushed back on collagen supplements, arguing the evidence for skin benefits is thin. The glucosamine finding fits a pattern: the supplement industry has outrun the safety data, and the data, when it finally arrives, sometimes points in uncomfortable directions.

I don’t think any of this is a reason to panic. But it’s a reason to pay attention — and to ask your GP questions the supplement bottle won’t answer. We’re in a strange moment where a person might spend more time researching a $400 phone than a supplement they put in their body every morning for a decade.

When my next patient asks me about glucosamine, I won’t have a neat script. I’ll talk about the study, about what we know and what we don’t, about the difference between a healthy brain and one already under strain. I’ll probably say something like: the joint-pain relief might be real, but the picture has changed, and we should weigh that together.

And I’ll be thinking about all the patients I recommended it to before this week. That’s the part of general practice nobody warns you about in medical school — the years-long lag between a well-intentioned recommendation and the evidence that might undo it.