Appearance
12.5 — The Pancreatic Islets and Glucose Control
You have about 4 grams of glucose in your bloodstream — roughly a teaspoon. Your brain alone consumes about 120 grams a day. So the entire circulating supply is turned over every twenty minutes, indefinitely, whether you eat or not, and the concentration barely moves.
Before 1922, a child diagnosed with type 1 diabetes lived on average a few months. The treatment was starvation, which delayed death by making there be less glucose to fail to handle. In January 1922 the first patient received an extract of pancreas, and children who had been dying in wards woke up. It remains one of the most dramatic single moments in medical history.
The islets
About one million islets, scattered through the pancreas, making up only 1 to 2 percent of its mass (Chapter 9.5).
They receive around 10 to 15 percent of the pancreatic blood flow — roughly ten times their share by mass — because a hormone must reach the circulation quickly.
Four cell types:
Beta cells (about 70 percent) — insulin. Alpha cells (about 20 percent) — glucagon. Delta cells (about 5 percent) — somatostatin, which inhibits both. PP cells — pancreatic polypeptide.
And the islet is arranged so the cells influence each other directly. Insulin suppresses glucagon release from neighbouring alpha cells, which means a single glucose rise both increases insulin and reduces glucagon — a coordinated switch rather than two independent adjustments.
The blood flows from the centre of the islet outward, so the beta cells' output bathes the alpha cells before leaving. The architecture enforces the hierarchy.
How a beta cell senses glucose
This is one of the most direct sensing mechanisms in the body — the sensor and the effector are the same cell (Chapter 4.7).
1. Glucose enters through a transporter that does not need insulin.
2. It is metabolised, producing ATP.
3. Rising ATP closes an ATP-sensitive potassium channel.
4. With potassium no longer leaking out, the membrane depolarises.
5. Voltage-gated calcium channels open, calcium enters.
6. Calcium triggers insulin vesicles to fuse and release.
So the cell measures glucose by metabolising it and reading the ATP. It is not detecting glucose with a receptor; it is detecting the consequence of using it, which means it reports usable fuel rather than merely the presence of a molecule.
And step 3 is a drug target that has been exploited for seventy years.
Sulfonylureas — gliclazide, glibenclamide — close that same potassium channel directly, without needing glucose. So they force insulin release regardless of the blood sugar.
Which is precisely why they cause hypoglycaemia and why the newer drug classes, which only act when glucose is high, are generally preferred.
And a mutation in that channel causes neonatal diabetes — and, remarkably, those infants can often be switched from insulin injections to sulfonylurea tablets, because the drug closes the faulty channel. A genetic diagnosis changing an infant from lifelong injections to a tablet is one of the better stories in precision medicine.
Insulin release is biphasic. A rapid first-phase burst from pre-formed vesicles within minutes, then a sustained second phase from newly made insulin.
And loss of the first phase is one of the earliest detectable abnormalities in type 2 diabetes, appearing years before blood sugar becomes abnormal.
What insulin does
Insulin is the hormone of the fed state. Everything it does is storage.
Muscle — inserts GLUT4 transporters into the membrane so glucose can enter (Chapter 1.4), and promotes glycogen and protein synthesis.
Fat — promotes glucose uptake and fat storage, and powerfully inhibits fat breakdown.
That last point is the key to diabetic ketoacidosis. The amount of insulin needed to suppress fat breakdown is much smaller than the amount needed to move glucose into cells. So in type 2 diabetes, where some insulin remains, fat breakdown stays suppressed and ketoacidosis is uncommon. In type 1, where there is none, fat breakdown runs unchecked (Chapter 1.6).
Liver — promotes glycogen storage, switches off glucose production, and promotes fat synthesis.
And a crucial anatomical point. Insulin from the pancreas drains into the portal vein and reaches the liver first, at concentrations two to three times higher than the rest of the body sees (Chapter 7.5).
Injected insulin does not. It enters the general circulation and reaches the liver at the same concentration as everywhere else. So injected insulin can never reproduce the natural distribution, and this is a fundamental limitation of insulin therapy that no delivery device solves.
Brain — and this is worth being clear about. The brain does not need insulin for glucose uptake. Its transporters are insulin-independent.
Which is exactly why hypoglycaemia is so dangerous. The brain cannot store glucose, cannot make it, and cannot use fatty acids (Chapter 1.6). It is entirely dependent on a continuous supply from the blood, and when that supply falls the brain fails within minutes.
Glucagon and the counter-regulatory hormones
Glucagon is the hormone of the fasted state. Released by alpha cells when glucose falls, and it acts almost entirely on the liver: break down glycogen, manufacture glucose, and produce ketones.
And glucose is defended by four hormones while it is attacked by only one. Glucagon, adrenaline, cortisol and growth hormone all raise blood glucose; only insulin lowers it.
That asymmetry is not an accident. A high blood sugar damages you over years; a low blood sugar kills you in minutes. The system is deliberately biased toward the error that is survivable.
The fed and fasted states
After a meal: glucose rises, insulin rises, glucagon falls. Glucose enters muscle and fat, glycogen is made, the liver stops producing glucose.
Fasting (a few hours): insulin falls, glucagon rises. Liver glycogen is broken down, supplying glucose.
Prolonged fasting (over 24 hours): glycogen runs out — there is only about a day's worth (Chapter 1.2). The liver manufactures glucose from amino acids and glycerol, and begins producing ketones from fat.
Extended fasting (days): the brain adapts to using ketones for up to about two thirds of its needs, which is what makes prolonged fasting survivable. This adaptation is why muscle breakdown slows after the first few days — the body has found a way to fuel the brain without dismantling itself.
Diabetes
The most common endocrine disease in the world — over 500 million people, and rising.
Chapter 18.7 covers the clinical management; here is what each type actually is.
Type 1
Autoimmune destruction of beta cells. Around 5 to 10 percent of diabetes.
By diagnosis, 80 to 90 percent of beta cells are gone.
Usually presents in childhood or adolescence, though it can occur at any age — and adult-onset type 1 is regularly misdiagnosed as type 2, which matters because the treatment differs completely.
The presentation is characteristic and fast: excessive thirst, excessive urination, weight loss despite eating, and fatigue, developing over weeks. Often the first presentation is diabetic ketoacidosis, particularly in young children.
And the mechanism of the symptoms follows from Chapter 10.2. Glucose above the renal threshold spills into the urine, dragging water with it osmotically, causing polyuria and then dehydration and thirst.
There is a partial remission phase — the "honeymoon period" — in the months after starting insulin, when the remaining beta cells recover somewhat and insulin requirements fall. It is temporary, and patients should be warned, because otherwise it feels like a cure that is then taken away.
Treatment is insulin, for life. There is no alternative, and the developments have been in how it is delivered rather than in avoiding it: multiple daily injections, pumps, continuous glucose monitors, and increasingly closed-loop systems where the monitor and pump communicate and adjust automatically. Those systems have made a substantial difference to both control and quality of life.
Islet and whole-pancreas transplantation work and are limited by donor supply and by the need for lifelong immunosuppression. Stem cell-derived beta cells are in trials, and encapsulation approaches that protect them from immune attack are the main hope for a genuine cure.
Type 2
Insulin resistance plus progressive beta cell failure. Around 90 percent of diabetes.
Both parts are needed. Insulin resistance alone does not cause diabetes — the beta cells compensate by producing more, and a person can maintain normal glucose for years with insulin levels several times normal. Diabetes appears when the beta cells can no longer keep up.
So type 2 diabetes is a disease of beta cell failure occurring in the context of insulin resistance, and framing it as simply "too much sugar" misses the part that determines who develops it.
Insulin resistance is driven by excess visceral fat, physical inactivity, genetics — the genetic contribution to type 2 is stronger than to type 1, with concordance in identical twins of around 70 to 90 percent — and age.
Why fat causes it: fat tissue releases inflammatory signals and free fatty acids, and fat deposited inside the liver and muscle interferes directly with insulin signalling. Visceral fat matters much more than subcutaneous fat, which is why waist measurement predicts risk better than weight.
And a genuinely important finding: type 2 diabetes can be reversed. Substantial weight loss — around 15 kg — puts a large proportion of people into remission, with normal glucose off all medication. The DiRECT trial achieved remission in 46 percent at one year and 36 percent at two years using a structured low-calorie diet programme in primary care.
Remission is most likely with a shorter duration of diabetes and greater weight loss, which is a strong argument for acting early rather than escalating medication over years.
Bariatric surgery produces even higher remission rates, and strikingly, glucose control improves within days — before any meaningful weight loss — which points to gut hormone changes rather than weight alone as the mechanism.
This is the most hopeful fact in the whole of diabetes, and it deserves to be much better known than it is.
The other types
Gestational diabetes — first appearing in pregnancy, affecting 5 to 10 percent of pregnancies. Caused by the placenta producing hormones that oppose insulin, on top of any existing resistance. It usually resolves after delivery — and it identifies a woman at substantially raised lifetime risk of type 2, which makes it a valuable early warning that is often not followed up.
MODY — maturity-onset diabetes of the young. A group of single-gene conditions, inherited dominantly, presenting in young adults. Frequently misdiagnosed as type 1 or type 2. And the diagnosis matters enormously: some forms need no treatment at all, and others respond beautifully to low-dose sulfonylureas rather than insulin. A family history of diabetes across three generations, diagnosed young, without obesity or antibodies, should prompt genetic testing.
Secondary diabetes — from pancreatic disease, Cushing's syndrome, acromegaly, or steroids.
Hypoglycaemia
Blood glucose below about 3.9 mmol/L, and it is the acute danger of diabetes treatment rather than of diabetes itself.
Symptoms come in two phases, and the order is the safety system.
Autonomic first (around 3.6 mmol/L) — sweating, trembling, palpitations, hunger, anxiety. These are adrenaline symptoms, and their purpose is to make you eat.
Neuroglycopenic second (below about 2.8) — confusion, difficulty concentrating, slurred speech, odd behaviour, drowsiness, then seizures and coma. These are the brain running out of fuel.
And the crucial safety point is the order: warning symptoms come before the brain fails.
Hypoglycaemia unawareness is the loss of that warning. After repeated episodes, the autonomic response is blunted, so the first symptom becomes confusion — by which point the person may be unable to treat themselves.
It affects around 25 percent of people with long-standing type 1 diabetes, and it is a major reason for severe episodes, hospital admissions, and driving restrictions.
It is partly reversible. Scrupulously avoiding all hypoglycaemia for several weeks restores the warning symptoms in many people, which is a treatment worth knowing about.
Treatment of a conscious person: 15 to 20 grams of fast-acting carbohydrate — glucose tablets, a small glass of fruit juice, or sugary drink — then recheck after 15 minutes, then follow with longer-acting carbohydrate.
Chocolate is a poor choice, because the fat slows absorption considerably.
If unconscious: nothing by mouth. Intramuscular glucagon, which family members can be trained to give, or intravenous glucose. Chapter 23.11.
Monitoring
HbA1c — glycated haemoglobin. Glucose attaches non-enzymatically to haemoglobin in proportion to its concentration, and the red cell lives 120 days (Chapter 7.1). So HbA1c reflects the average glucose over the preceding 2 to 3 months.
It is the standard measure of control, and it has known limitations: it is unreliable in anaemia, in haemoglobin variants such as sickle cell trait, in kidney failure, and in pregnancy. And it is an average, so it hides a pattern of high and low swings.
Continuous glucose monitoring has largely solved that last problem. A small sensor worn on the skin measures glucose in the tissue fluid every few minutes.
And it introduced a genuinely better measure: "time in range" — the proportion of the day spent between 3.9 and 10 mmol/L. A target of 70 percent or more is now standard, and it captures something HbA1c cannot: two people with identical HbA1c can have completely different experiences, one stable and one swinging between dangerous highs and lows.
What the next page fixes
Two glands remain, and one of them is largely gone by adulthood. Chapter 12.6 covers the gonads as endocrine organs — the hormones that drive puberty, fertility and much else — before Part 15 returns to reproduction in full.