The Uncharted Universe Within

Cosmic Eye, by astrophysicist Danail Obreschkow (2018): a single continuous zoom out from a human eye to the cosmic web — and then back down through the cell, the molecule, the atom. The same journey, in both directions, ends somewhere we have barely mapped. Watch on YouTube ↗

Appetite suppression is the dominant, well-evidenced mechanism behind GLP-1 and dual GIP/GLP-1 drugs. It is probably most of the story — but a patient losing a stone in six weeks on a sub-licensed dose, with almost no dietary change, sits in the part it does not comfortably explain. The grown-up response to that gap is not to pretend, but to ask better questions. This series asks one of them.

Russell — not his real name — is not a patient I went looking for. He is an older gentleman from my community, from a large local family, a man I had come to know. And I should be clear about why he was in my consulting room at all: I am a cardiologist, and the patients whose stories I tell here come to me with hearts to worry about, not waistlines. Russell's problem was his blood pressure.

He farmed for most of his life, and even now his arms are strikingly muscular for his age. What had changed was his mobility: arthritis in both knees had all but stopped him walking, and he relied on transport for almost everything. He carried a great deal of weight, most of it around the middle.

When I took his dietary history I expected the usual hidden excess. I did not find it. Weetabix with milk and a single teaspoon of sugar; a mid-morning coffee with two; a traditional meat-and-two-veg lunch with a small potato; almost nothing in the evening beyond the occasional two squares of dark chocolate. No snacking, no biscuits, no alcohol, no smoothies. A diet essentially unchanged for years.

He is a man of obvious integrity, with no reason to mislead me. Which left me with a genuine puzzle before I had prescribed anything at all: how does someone become this large on this little?

And there was the matter that brought him to me in the first place. His blood pressure was not well controlled — running around 150/90 mmHg despite two antihypertensive medications. He had no family history of coronary heart disease; on the contrary, he came from large-built but long-lived stock, his parents reaching a grand old age. Whether his hypertension was simply "essential" or driven by his size, none of us could say for certain. But his waist-to-height ratio was 0.65 — well into the highest-risk band for central adiposity, the category that guidelines link directly to hypertension and cardiovascular disease.¹

We start everyone gently — a microdosing principle, beginning well below the licensed starting dose and building slowly, aiming for a steady pound or two a week at the lowest dose that delivers it. Russell began on a dose of tirzepatide (Mounjaro) well below the 2.5 mg licensed starting dose. He came back a week later down one to two kilograms; the same again the following week; and by six weeks he had lost close to a stone — with no side effects at all.

When I asked how much he had changed his diet, he said: not much. Still the sugar on his Weetabix, still the coffee, and he had not taken to the protein powder I suggested. Perhaps slightly less overall — on any honest reckoning, under a tenth.

The reflexive explanation is that patients like Russell must be under-reporting. We are right to be sceptical: people estimate their own intake poorly. But scepticism has limits here. He was not eating much to begin with; the drug is meant to work mainly by curbing appetite; and his appetite-driven intake had barely moved. The under-reporting story has to carry an enormous amount of weight to explain close to a stone in six weeks at a dose below the licensed floor.

The standard account of these drugs is about appetite. Semaglutide is a GLP-1 receptor agonist; tirzepatide is a dual agonist, acting at both the GLP-1 receptor and the GIP receptor.² They quieten the appetite centres, soften cravings, and slow gastric emptying so people feel full for longer.

The evidence is strong. In carefully controlled trials the dominant driver of weight loss is a genuine fall in how much people eat, and tirzepatide produced average reductions of around a fifth of body weight at the higher doses.³ For most patients, most of the time, appetite suppression is probably the lion's share of the explanation. But "lion's share" is not "the whole story" — and Russell sits awkwardly in the part that is left over.

Incretin signalling reaches well beyond the appetite centres. The GIP arm of a dual agonist is particularly interesting, because GIP acts directly on adipose tissue and on peripheral energy metabolism — which is precisely why the dual drug was expected to recruit metabolically active tissue that a pure GLP-1 drug does not, and why increased energy utilisation, not just reduced intake, has been part of the rationale from the start.²

Here is my current favourite — and I want to be clear it is a hypothesis, not a fact. It begins with the thyroid hormones and the enzymes, the deiodinases, that decide their fate in the tissues.

Most active thyroid hormone, T3, is not secreted by the gland but made locally, by converting the prohormone T4. The deiodinases can take T4 down the activating path to T3, which raises tissue metabolism, or the inactivating path to reverse T3, which does not.⁴ Which path dominates is not fixed: in illness, starvation and inflammation the balance tips towards reverse T3 and away from active T3 — the well-described "low T3" or non-thyroidal-illness pattern, in which tissue metabolism falls.⁵

Now bring in visceral adipose tissue — the fat packed deep in the abdomen, the kind Russell carries. It is not inert storage but an active inflammatory organ, releasing cytokines that reach the whole body. So the chain I find myself sketching runs: accumulated visceral fat drives chronic low-grade inflammation; that inflammation tips the deiodinase balance toward reverse T3; tissue T3 falls; metabolism slows; weight is gained and held. One reading of recovery is purely indirect — shrink the fat, quieten the inflammation, release the brake.

But I would go further. The speed of Russell's response, at so small a dose, is hard to pin on fat loss alone — the metabolic shift seems to run ahead of the change in his body. That points to a second, direct action: the drug working on the deiodination step itself, inside the fat. And this is no longer pure conjecture. The GIP receptor is expressed in adipose tissue, and in mice, switching it on directly in the fat cell drives substantial weight loss and raises energy expenditure independently of appetite.⁶ More pointed still: a GLP-1 receptor agonist, liraglutide, has been shown to activate type 2 deiodinase in mouse brown and white fat, increasing local conversion of T4 to active T3 and adding to thermogenesis⁷ — almost exactly the step I am proposing, demonstrated at least in animals.

Why has this not simply been settled in patients? Because the blood does not tell you what the tissue is doing. Local deiodinase activity can lift tissue T3 signalling with no change at all in circulating hormone,⁵ and serum T3 and reverse T3 are difficult and expensive to assay reliably in any case. The very thing we most want to know — what the deiodinases are doing inside the fat — is largely invisible to a blood test. Which is why, for now, this has to be pursued in animal models rather than the clinic.

Is it proven in humans, for tirzepatide, in a man like Russell? No. Liraglutide is not tirzepatide; a mouse is not a man; and at least one direct adipose mechanism in the animal work runs through calcium cycling rather than thyroid hormone,⁶ so the deiodinase route is one candidate among several. But it is no longer a chain of disconnected pieces — the crucial link has been shown to exist. It is grounded, specific, and testable. That is the point.

If a treatment that works for reasons we cannot fully pin down makes you uneasy, it shouldn't — it is the normal condition of medicine, not the exception.

The SGLT2 inhibitors were built as diabetes drugs and turned out to be among the most powerful heart-failure treatments in decades, cutting deaths and admissions in patients with and without diabetes.⁸ Why? We have plausible answers — offloading sodium and water, easing cardiac workload, improving the heart's use of ketone fuels — but none alone accounts for the size of the benefit, and honest reviews still call the mechanism unresolved.⁹ Aspirin eased pain and prevented heart attacks long before anyone had heard of cyclo-oxygenase. Statins prevented events before we appreciated their reach beyond cholesterol. Anaesthetists switched off consciousness for over a century without a complete account of how. The treatment comes first; the explanation follows — sometimes by decades.

Watch the Cosmic Eye zoom once more and the point lands. The journey that takes you out past galaxies takes you, in the other direction, down into the cell, the molecule, the atom — and somewhere on that inward path, surprisingly close to home, the map runs out. The uncharted universe is not only out there. It is inside Russell, and inside all of us.

I worry that medicine sometimes wears more certainty than it has earned. Doctors are trained to give answers; patients understandably want them; the result can be a temptation to oversimplify. But in my experience patients are far more comfortable with uncertainty than we assume. Told that a treatment clearly works, that the evidence for its benefit is strong, but that the mechanism is still incompletely understood, people rarely lose confidence. If anything they lean in.

Russell's story sits at the heart of what The VAT Trap is about: visceral fat as an active organ whose reach extends far beyond simple calorie storage, and a web of links — to inflammation, metabolism, thyroid signalling and cardiovascular risk — still being drawn. That we cannot yet fully explain him is not a failure. It is an invitation.

Over the coming weeks, The Deiodinase Files follows that invitation in five parts — from why belly fat after fifty refuses to shift, to the thyroid clue almost nobody is testing, to why these drugs work faster than they should, and what a microdose can teach us about the mechanism. It opens next week.