The Metabolic Brain

I’ve often said to patients, “You’re not diabetic.”

Most people are happy to hear that at a check-up, but it may be a little misleading. It is perfectly true but it just says that you are on the right side of a line drawn on a scale, and says nothing about where on that scale you are sitting, or which direction you have been travelling for the last fifteen years.

Diabetes is diagnosed when a test exceeds a threshold, and the threshold is set at roughly the point where the risk of damage to eyes, kidneys and nerves climbs steeply enough to warrant treating. What it is not is the point where harm begins. Blood sugar runs on a continuum, and so do its consequences.

So when I look through someone’s old results, the useful thing is almost never the most recent number. It is the shape of the line. A blood sugar measure that has climbed slowly and steadily for fifteen years, entirely within the normal range, never once flagged, is among the most informative things in a file and among the most consistently ignored.

Every one of those results was fine. The line joining them was not.

And for the purposes of this series, that line is a picture of something happening to the brain twenty years before anyone would think to worry about it.

Why the brain is exposed

The brain is an extraordinarily hungry organ. It runs almost entirely on glucose, takes about a fifth of your daily energy, and stores essentially no fuel of its own. Everything it does depends on a steady supply of energy arriving and being taken up efficiently. When the machinery for handling glucose starts to fail, the brain is among the first organs to feel it.

That machinery is insulin, and when it fails it fails in a particular way. Insulin is the hormone that tells cells to take glucose out of the blood and use it. Cells can become less responsive to that instruction, and when they do, the body compensates by producing more insulin to get the same job done. The result is a state where both glucose and insulin sit higher than they should, and neither works as well as it should. It is called insulin resistance, it develops silently across years to decades, and it is well underway long before anyone would use the word diabetes.

And it does not stop at the neck. The brain becomes insulin resistant too.

Insulin does more in the brain than you would think

In the brain, insulin does considerably more than move glucose into cells.

The brain is studded with insulin receptors, concentrated in the regions that matter most for memory, including the hippocampus. Insulin there supports synaptic plasticity — the strengthening and remodelling of connections that underlies learning. It influences neurotransmitter signalling. It plays a part in clearing amyloid, the protein that accumulates in Alzheimer’s disease.

So when insulin signalling in the brain degrades, what degrades with it is not just fuel supply. It is the machinery of memory itself.

Inside the Alzheimer’s brain

Post-mortem studies of Alzheimer’s brains consistently show reduced insulin signalling, fewer functioning insulin receptors, and impaired glucose metabolism in the affected regions. And the evidence linking that impairment to the core features of the disease — amyloid accumulation, tau tangles, inflammation, oxidative stress — is substantial and growing. Insulin resistance appears to worsen almost every process involved. Whatever else is going on in these brains, failing energy metabolism is not a bystander.

As discussed in the previous article, diabetes damages blood vessels and impairs blood flow to the brain. But impaired insulin signalling inside the brain is a separate insult arriving by a different route. Metabolic disease harms the brain twice over.

The number is normal, the trend is not

At population level the association is not in doubt. Type 2 diabetes substantially raises the risk of dementia, across large cohorts followed for decades, and the effect holds for both Alzheimer’s and vascular dementia. Diabetes sits on the Lancet Commission’s list of modifiable risk factors for good reason.

The more useful finding, for a series about prevention, is that the risk does not begin at the point of diagnosis. Insulin resistance and elevated glucose are associated with poorer cognitive performance — particularly processing speed and memory — well before anyone crosses the diabetic threshold. It is a continuum, not a switch.

Which brings us back to that drifting line. The years when blood sugar is climbing through the upper end of normal are not a period of safety before the real problem starts. They are the problem starting. “Not diabetic yet” describes precisely the window this series is about and where prevention is possible.

What about the drugs

A very large number of people are now taking medications that improve metabolic health — GLP-1 drugs like semaglutide, prescribed for diabetes and for weight, and SGLT2 inhibitors, prescribed for diabetes, heart failure and kidney disease. Given everything above, an obvious question follows: if these drugs repair the metabolic problem, do they also protect the brain?

The answer splits along the exact line this series is built on.

Start with the disappointing half. Semaglutide was tested in two large phase 3 trials — EVOKE and EVOKE+ — in people who already had early Alzheimer’s disease. The results were negative. Despite some improvement in disease biomarkers, semaglutide did not slow cognitive decline compared with placebo.

That pattern should feel familiar by now. It is the anti-amyloid story from the first article, repeated. Once disease is established, the damage is substantially done, and a drug acting on an upstream mechanism struggles to change the trajectory.

Now the encouraging half. A much larger body of observational evidence — hundreds of thousands of people with type 2 diabetes — suggests those taking GLP-1 drugs develop dementia at lower rates than those on some other treatments, with the clearest signal in long-term continuous use. The SGLT2 inhibitors show a similar pattern in the same kind of data, and there are plausible mechanisms for both: better glucose control, less inflammation, improved vascular function, and in the case of GLP-1s, receptors in the brain itself. Observational data cannot establish causation, and the findings are not perfectly consistent. But the direction repeats across many datasets.

So the drugs look more promising for prevention, taken during the long silent window, than for treatment once the disease has arrived — which is exactly where EVOKE tested them, and plausibly why it failed.

None of this is a reason to take either drug for your brain alone. Whether they prevent dementia in people without metabolic disease is unknown. But a great many people have a genuine metabolic, cardiac or renal reason to be on one of them already, and for those people a possible brain benefit is a fair part of the conversation. I have written about GLP-1s and healthspan more broadly elsewhere on the site, and the same caution applies: promising, plausible, not yet proven.

What actually moves the needle

Insulin resistance is among the most modifiable conditions in medicine, and almost all of that modification happens without a prescription.

The two most powerful levers get their own articles in this series. Physical activity improves insulin sensitivity more reliably than almost anything else available, and it is next. Underneath it sits what and how much you eat, and in particular visceral fat — the metabolically active fat around the abdominal organs, which is a principal driver of systemic insulin resistance and one of the most worthwhile things a person can lose.

None of this requires a diagnosis to begin. It requires knowing the numbers and acting while they are still only drifting.

Practical steps for this week

1. Ask for the trend, not the result. Request an HbA1c — a better measure than a fasting glucose, which is heavily influenced by what time you had the blood taken — and ask your GP to look at your old results alongside it. One number inside the normal range tells you very little. Ten years of numbers climbing through it tells you a great deal.
2. Measure your waist, not just your weight. Waist circumference is a better proxy for visceral fat than the scales. Measure at the midpoint between the lowest rib and the top of the hip bone; a rising waist is an early metabolic warning worth acting on.
3. Cut the liquid sugar first. You do not need an elaborate diet. For most people the highest-yield single change is removing sugar-sweetened drinks and reducing refined carbohydrate, which drive the glucose and insulin spikes that entrench insulin resistance.
4. If you have a metabolic reason to consider a GLP-1 or similar medication, have that conversation properly. Not for your brain alone, but as one of several benefits worth weighing with your doctor.