World Alzheimer's Day, 21 September

Gut health and Alzheimer's: what the evidence actually says

Every September a wave of posts claims your gut is the hidden cause of Alzheimer's, usually just above a supplement link. The research underneath those posts is real. What it shows is a lot more careful than the headline, and the most interesting part of it starts in your mouth.

Jump to Gut health explained How to improve your gut health H. pylori in South Africa

Written by the Epicentre Walk-In Labs team. Epicentre has run diagnostic and epidemiological research in South Africa since 2001, including work published in The Lancet HIV with CAPRISA and the National Institute for Communicable Diseases. The same laboratories run our walk-in testing.

Medically reviewed by Dr Samantha Naidoo, MB ChB, FCP (SA), Medical Director. Published 21 September 2026. Next review March 2027.

Basically

  • People with Alzheimer's do have measurably different gut bacteria. That has been found more than once.
  • Nobody has shown it causes anything. Dementia changes how people eat, chew, move and medicate, and all of that changes gut bacteria, so some of the arrow may point backwards.
  • The mouth end of the story is the most direct: gum disease bacteria have been found inside Alzheimer's brains. It is also where the biggest disappointment happened.
  • Brushing and cleaning between your teeth are worth doing anyway, for reasons that have nothing to do with dementia.
  • No test anywhere reads your gut and tells you your dementia risk. Be sceptical of anyone selling one.

What is in this article

  1. Where the idea came from
  2. Why researchers will not call it a cause
  3. The mouth, and why it is part of your gut
  4. The trial that tested this, and failed
  5. What to actually do
  6. Where testing fits, and where it does not
  7. The verdict

The two sides, at a glance

Verdict: not proven

Five findings that made the headlines, each next to the reason it has not settled the question.

What was found Why it is not settled
Different gut bacteria. People with Alzheimer's have measurably different gut bacteria, and more than one study has found it.
but
The arrow may point backwards. Dementia changes what people eat, how they chew, how much they move and what they take. All of that changes gut bacteria.
Mice get worse, or better. Adding certain gut bacteria to mice makes amyloid worse. Stripping their gut bacteria out reduces plaque.
but
Mice are not people. And the best-known human study compared 25 against 25, small enough to find effects that vanish on repeat.
Gum bacteria in the brain. Gum disease bacteria and their toxic enzymes have been found inside Alzheimer's brains, with enzyme levels tracking tau damage.
but
The drug built on it failed. That study was funded by a company developing an enzyme blocker. It went into a large trial in Alzheimer's and did not deliver.
Fewer teeth, more dementia. Across 8 cohort studies, people with fewer than 20 teeth had a 22% higher risk of dementia.
but
Gum disease itself did not track. The same review found those studies conflicted. Losing teeth also changes what you can eat, which explains a lot on its own.
The gut and brain do talk. That part is not controversial. There are real nerve and immune routes between the two.
but
The people who rank the risks left it out. Neither the gut microbiome nor oral health appears among the 14 factors the Lancet Commission named in 2024, covering about 45% of cases.

Nothing here is fringe science. It is ordinary early-stage research being reported as though it were finished. The rest of this article is why that distinction matters.

Where did this idea come from?

In 2017 a team at the University of Wisconsin compared stool samples from people diagnosed with Alzheimer's against age and sex matched people without it. The Alzheimer's group had less diverse gut bacteria, less Firmicutes, more Bacteroidetes and less Bifidobacterium. Some of those differences tracked with amyloid and tau markers in spinal fluid.1

That was the paper that lit the fuse. It has been followed by dozens more, and the broad pattern keeps reappearing: the gut community of someone with Alzheimer's looks different.

The animal work goes further. In one study, mice bred to develop Alzheimer's-like pathology showed bigger age-related shifts in their gut bacteria than normal mice, and colonising them with Bacteroides made amyloid deposition worse. Calorie restriction reversed some of it.2 Other groups have found that stripping gut bacteria out of similar mice reduces plaque.

So there is a real signal. The question is what it means.

Why researchers are not calling this a cause

Four problems keep coming up, and they are the reason the field has not moved from interesting to established.

The arrow may point backwards

Dementia is not a quiet passenger. It changes appetite, food choices, chewing ability, how much a person moves, how much water they drink and which medications they take. Constipation is extremely common. Every one of those reshapes the gut microbiome. A person whose diet narrowed three years ago will have different gut bacteria today, and the diagnosis came first.

The confounding problem in one sentence

If you compare the gut bacteria of someone in a care setting eating soft food against someone living independently and cooking for themselves, you will find differences, and none of them have to be doing anything to the brain.

The human studies are small

That 2017 study compared 25 people against 25. Most microbiome studies in this area are similarly sized. Small studies find large effects that shrink or vanish when someone tries to repeat them, and microbiome work is particularly prone to it because sample handling, sequencing method and the reference database all change the answer.

A systematic review comparing microbiome changes across neurodegenerative diseases concluded that at the broad level the changes are neither consistent nor specific to any one disease, and called for standardised collection and analysis before anyone draws conclusions.3

The genetic evidence is mixed

One way around reverse causation is Mendelian randomisation, which uses inherited genetic variants as a natural experiment. A 2024 analysis looked at gut bacteria, circulating cytokines and 5 types of dementia. It reported 20 positive and 16 negative causal signals between gut bacteria and dementia.4

Signals in both directions, across many bacterial groups, is not the clean result you would expect if one organism were driving the disease. Inflammatory cytokines, the mechanism most often proposed, did not turn out to be the middleman.

The people whose job is ranking dementia risks left it off the list

This is the part that rarely makes the posts. In 2024 the Lancet standing Commission on dementia published its updated list of modifiable risk factors. There are 14 of them, and together they could in principle prevent or delay around 45% of dementia cases.5

The list is low education, hearing loss, high LDL cholesterol, depression, head injury, physical inactivity, diabetes, smoking, high blood pressure, obesity, excessive alcohol, social isolation, air pollution and untreated vision loss.

The gut microbiome is not on it. Neither is oral health. These are the researchers who spend their careers weighing this evidence, and as of their most recent report they did not think either had cleared the bar.

The mouth is where your gut microbiome starts

Here is the part that gets skipped. People talk about the gut microbiome as if it begins somewhere below the ribs. It does not. The digestive tract starts at your lips, and the mouth carries the second most diverse microbial community in the body after the colon.

You swallow roughly a litre of saliva a day, and everything living in your mouth goes down with it. Stomach acid destroys most of those organisms, which is the system working as designed.

But when the balance in the mouth shifts, which is what gum disease is, more oral species survive the trip and turn up further downstream. Oral bacteria appearing in the gut in unusual numbers has been documented in several inflammatory conditions.

So if you are interested in your gut, the mouth is the upstream end of the same pipe. Which brings us to the most striking finding in this whole area.

Reading this because of someone in your family? The dementia risks with real evidence behind them, blood sugar, cholesterol and blood pressure, are all in a standard panel. See what that covers, or read our guide to understanding your blood test results.

What the gum disease research found

In 2019 a paper in Science Advances reported finding Porphyromonas gingivalis, the keystone bacterium of chronic gum disease, in the brains of people who had died with Alzheimer's. They also found gingipains, the toxic enzymes it produces, and the levels of those enzymes correlated with tau and ubiquitin pathology.

When they gave mice an oral P. gingivalis infection, the bacterium reached the brain and amyloid production went up.6

That is a much more direct claim than "the gut looks different". It names an organism, a route and a mechanism.

Alongside it sits the tooth loss evidence. A meta-analysis pooling 8 prospective cohort studies found that people with fewer than 20 teeth had a 26% higher risk of cognitive decline (HR 1.26, 95% CI 1.14 to 1.40) and a 22% higher risk of dementia (HR 1.22, 95% CI 1.04 to 1.43) than people with 20 or more.7

And the same review's other finding

Studies looking directly at gum disease and cognitive status gave conflicting results. Tooth loss tracked with dementia risk. Periodontal disease, the thing that causes most tooth loss, did not do so consistently. That is an awkward result for a simple causal story.

The trial that tested this idea, and failed

The 2019 brain study was funded by a biotechnology company that was developing a drug to block those bacterial enzymes. That is not a scandal in itself, and the science was published in a serious journal. But it does mean the hypothesis was going to get tested properly, and it was.

The enzyme blocker went into a large trial in people with Alzheimer's disease. It did not produce a clear benefit, liver safety signals emerged, and the programme was stopped. The strongest available test of the strongest version of the oral bacteria hypothesis did not confirm it.

There is also a competing explanation for why microbes turn up in Alzheimer's brains at all. Amyloid beta, the protein in the plaques, behaves like an antimicrobial peptide in laboratory work.

If it is part of the brain's own immune response, then finding bacteria alongside plaque could mean the bacteria triggered the response, or that the plaque was doing its job, or that damaged tissue is simply easier to colonise. Same observation, three readings.

Should you bother with your mouth? Yes, but not for this reason

Not because it will prevent dementia. Nobody can promise you that, and this article is not going to.

Because gum disease is bad for you on its own terms, and because two of the things it is linked to, cardiovascular disease and poorly controlled diabetes, are on that Lancet list of 14.

Diabetes and gum disease run in both directions: high blood sugar makes gum disease worse, and gum inflammation makes blood sugar harder to control. That loop is far better evidenced than the brain story. It is also a route by which your mouth could matter to your brain without any bacterium ever needing to get there.

Then there is the simpler mechanism. People who lose their teeth eat differently. Soft, processed, low-fibre food is easier to manage, and that changes nutrition and the gut microbiome at the same time. You do not need a bacterium crossing the blood-brain barrier to explain why dentition might track with cognition.

The South African picture

Dementia in South Africa is badly under-counted. The first large community prevalence study, among Xhosa-speaking residents of a low-income rural district in the Eastern Cape, found roughly 11% of people over 65 met criteria for dementia. Most had never been diagnosed.8

On the oral side, routine preventive dentistry is out of financial or geographic reach for a large share of the population. That makes the daily habits the part actually within reach.

What to actually do

For your mouth

  • Brush twice a day for 2 minutes with fluoride toothpaste. Most people stop at around 45 seconds.
  • Clean between your teeth once a day, with floss or interdental brushes. A toothbrush physically cannot reach the surfaces between teeth, and that is exactly where gum disease starts. If flossing feels fiddly, interdental brushes are easier and work as well or better for most people.
  • Treat bleeding gums as a signal, not a normal event. Gums that bleed when you brush are inflamed. Healthy ones do not.
  • Do not scrub harder. Aggressive brushing damages gums and wears enamel. Gentle and thorough beats hard and fast.

For your gut

  • Fibre, and a range of it. Plant variety across a week does more for microbial diversity than any single food.
  • Fermented foods if you enjoy them. Sauerkraut, kimchi, kefir and plain yoghurt all feed organisms that are usually in short supply.
  • Be realistic about supplements. A capsule cannot outrun a diet, and no probiotic on the market has shown it changes cognitive outcomes.

For the risks that are actually proven

  • Blood pressure, blood sugar, LDL cholesterol, hearing, vision, alcohol, smoking, physical activity, head protection, social contact. These are the 14. They are less exciting than the microbiome and they have far more evidence behind them.

Where testing fits, and where it does not

Let us be direct about this, because it is where a lab could easily oversell.

No gut test predicts Alzheimer's disease. Not ours, not anyone's.

A gut microbiome panel tells you which organisms are present in your gut at the moment you collected the sample. That is useful if you have bloating, irregular bowel habits, recurring infections or unexplained digestive symptoms, and it is what our gut health panels are for. It is not a crystal ball, and anyone marketing one to you as a dementia risk score is selling you something.

What is measurable, today, in an ordinary blood draw, are several of the 14 risk factors the Lancet Commission actually named. Blood sugar control, LDL cholesterol and the markers around metabolic health all show up in a standard panel, and unlike your microbiome they come with clear targets and clear evidence.

You can have those checked without seeing a doctor first. We are open to walk-ins at 24 Lower Main Road, Observatory in Cape Town, 2 Knelsby Avenue, Hillcrest in Durban and 02 7th Avenue, Parktown North in Johannesburg. No referral, no prescription, no appointment. If getting to a branch is the problem, a nurse can come to you instead.

Start by understanding the numbers, not by buying a test

If you want to know what any of these markers mean before you spend anything, our guides explain each one in plain language: what it measures, what the standard range is, and what a functional range would say about the same result. Free, no sign-up.

When you do want them measured, HbA1c for long-term blood sugar, a lipogram for cholesterol and the metabolic markers alongside them are in our standard panels, with no doctor's referral needed and results in 2 to 5 working days.

Read the marker guides All health guides See test packages

The verdict: not proven, and still worth flossing

The gut and the brain do talk to each other. That much is not controversial. What is unresolved is whether anything in your gut or your mouth is driving Alzheimer's disease, or whether both are downstream of the same ageing, the same diet and the same vascular health that the Lancet Commission has already told us to pay attention to.

It is a live question, and it may resolve in the microbiome's favour. Right now it has not.

In the meantime, flossing costs a few rand a month and protects your teeth. Getting your blood pressure and blood sugar checked costs a little more and addresses risks that are already proven. Neither requires waiting for the research to settle.

Common questions

Does poor gut health cause Alzheimer's disease?

Nobody has shown that it does. Studies repeatedly find that the gut bacteria of people with Alzheimer's look different from those of people without it, but a difference is not a cause. Dementia itself changes what people eat, how well they chew, how much they move and what medication they take, and all of those change gut bacteria. The 2024 Lancet Commission on dementia prevention, which identified 14 modifiable risk factors accounting for roughly 45% of cases, does not include the gut microbiome among them.

Can flossing reduce your risk of dementia?

There is no trial showing that flossing prevents dementia, and it would be dishonest to claim one. What does exist is a meta-analysis of 8 prospective cohorts finding that people with fewer than 20 teeth had a 26% higher risk of cognitive decline and a 22% higher risk of dementia than people with 20 or more. The same review found that studies on gum disease and cognition gave conflicting results. Cleaning between your teeth protects the gums that keep teeth in place, and that is a good enough reason on its own.

Why does oral health get mentioned in gut health articles?

Because the digestive tract starts at the lips. You swallow roughly a litre of saliva a day, carrying whatever is living in your mouth down into the stomach and gut. Stomach acid deals with most of it, but when the balance in the mouth shifts, as it does in gum disease, oral species turn up further down in greater numbers. The mouth is the upstream end of the same system.

Was gum disease bacteria really found in Alzheimer's brains?

Yes. A 2019 paper in Science Advances reported Porphyromonas gingivalis, the main bacterium in chronic gum disease, and its toxic enzymes in the brains of people who had Alzheimer's, with enzyme levels tracking tau pathology. The important follow-up is less often quoted: the study was funded by a biotechnology company developing a drug to block those enzymes, and when that drug was tested in a large trial in people with Alzheimer's it did not produce a clear benefit and the programme was halted. The strongest test of the idea did not confirm it.

Is there a test that tells me my Alzheimer's risk from my gut bacteria?

No. No laboratory anywhere can read your gut microbiome and tell you your chance of developing Alzheimer's disease, and you should be sceptical of anyone selling one. A gut microbiome panel tells you which organisms are in your gut now, which is useful for digestive symptoms, not for predicting dementia. The dementia risk factors with real evidence behind them, such as blood pressure, blood sugar and cholesterol, are measurable in ordinary blood tests.

What actually is proven to lower dementia risk?

The 2024 Lancet Commission lists 14 factors: less education in early life, hearing loss, high LDL cholesterol, depression, head injury, physical inactivity, diabetes, smoking, high blood pressure, obesity, excessive alcohol, social isolation, air pollution and untreated vision loss. Addressing all of them could in principle prevent or delay about 45% of dementia cases. Several of them show up in a standard blood panel, which makes them the practical place to start.

Where can I get these blood tests done in South Africa?

Epicentre Walk-In Labs has three branches: 24 Lower Main Road, Observatory, Cape Town; 2 Knelsby Avenue, Hillcrest, Durban; and 02 7th Avenue, Parktown North, Johannesburg. You can walk in during opening hours with no doctor's referral and no appointment, and ask for HbA1c, a lipogram or a full wellness panel. A nurse can also come to you, and home test kits are couriered anywhere in South Africa, though blood still needs a nurse.

Related reading

We publish a new article every Wednesday, usually on something a patient asked us that week. See what else we have written.

References

  1. Vogt NM, Kerby RL, Dill-McFarland KA, et al. Gut microbiome alterations in Alzheimer's disease. Scientific Reports 2017;7:13537. doi:10.1038/s41598-017-13601-y
  2. Cox LM, Schafer MJ, Sohn J, et al. Calorie restriction slows age-related microbiota changes in an Alzheimer's disease model in female mice. Scientific Reports 2019;9:17904. doi:10.1038/s41598-019-54187-x
  3. Gerhardt S, Mohajeri MH. Changes of colonic bacterial composition in Parkinson's disease and other neurodegenerative diseases. Nutrients 2018;10(6):708. doi:10.3390/nu10060708
  4. Ji D, Chen WZ, Zhang L, et al. Gut microbiota, circulating cytokines and dementia: a Mendelian randomization study. Journal of Neuroinflammation 2024;21:2. doi:10.1186/s12974-023-02999-0
  5. Livingston G, Huntley J, Liu KY, et al. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission. The Lancet 2024;404(10452):572-628. doi:10.1016/S0140-6736(24)01296-0
  6. Dominy SS, Lynch C, Ermini F, et al. Porphyromonas gingivalis in Alzheimer's disease brains: evidence for disease causation and treatment with small-molecule inhibitors. Science Advances 2019;5(1):eaau3333. doi:10.1126/sciadv.aau3333
  7. Cerutti-Kopplin D, Feine J, Padilha DM, et al. Tooth loss increases the risk of diminished cognitive function: a systematic review and meta-analysis. JDR Clinical and Translational Research 2016;1(1):10-19. doi:10.1177/2380084416633102
  8. de Jager CA, Msemburi W, Pepper K, Combrinck MI. Dementia prevalence in a rural region of South Africa: a cross-sectional community study. Journal of Alzheimer's Disease 2017;60(3):1087-1096. doi:10.3233/JAD-170325

Studies identified via PubMed.

Important. Epicentre Walk-In Labs is a diagnostic laboratory operated by Epicentre Aids Risk Management (Pty) Ltd. We collect samples, run tests and report results. We do not diagnose conditions, prescribe medication or provide treatment. This article is general health information and is not medical advice. If you are worried about memory changes in yourself or someone you care for, speak to a doctor. Nothing here should be used to delay seeking care.