Diabetes

TL;DR. Diabetes is a failure of the system that keeps blood sugar in a narrow range. In type 1, the immune system destroys the cells that make insulin, so there is no key for the lock; it is an autoimmune disease, usually starts young, and is fatal within weeks without injected insulin. In type 2, which is about 19 in every 20 cases, insulin is present but the body responds to it poorly and the pancreas eventually cannot keep up; it develops over years, silently. The high sugar itself rarely causes symptoms at first. What it does is slowly damage blood vessels and nerves, and that damage is where blindness, kidney failure, amputations, heart attacks, and strokes come from. It is not contagious, it is not caused by eating a dessert, and type 2 is now sometimes reversible.

Key takeaways

• About 589 million adults live with diabetes, roughly 1 in 9 adults worldwide, and about 252 million of them do not know it. It caused around 3.4 million deaths in 2024, roughly one every nine seconds.
• Type 1 and type 2 share a symptom and share almost nothing else. Treating them as one disease is the most common misunderstanding about diabetes and it hurts patients in both directions.
• High blood sugar damages the body mainly through small blood vessels (eyes, kidneys, nerves) and large ones (heart, brain, legs). Most of the harm is vascular.
• The complications are largely preventable, and the evidence for that is unusually strong: two landmark trials (DCCT in type 1, UKPDS in type 2) showed that lowering average glucose lowers complications, with benefits still visible decades later.
• Type 2 diabetes can go into remission. In the DiRECT trial, a structured weight-loss programme put 46 percent of participants into remission at one year, and remission tracked almost entirely with how much weight was lost.
• It is not contagious. There is no way to catch diabetes from another person.

What it is

In short: Chronically high blood glucose, diagnosed on a threshold, with HbA1c giving a three-month running average that cannot be improved by behaving well the week before an appointment.

Every cell in your body runs on glucose. Blood carries it, and the concentration has to stay inside a narrow band: too little and the brain fails within minutes, too much and the sugar slowly damages proteins throughout the body. A healthy adult holds blood glucose roughly between 70 and 140 mg/dL (3.9 to 7.8 mmol/L) all day, through meals, fasting, and exercise.

Diabetes mellitus is the state in which that control has failed and blood glucose runs chronically high (hyperglycaemia). It is diagnosed by any of the following, confirmed on a second occasion unless symptoms are obvious:

HbA1c deserves a moment, because it appears in every later section. Glucose sticks irreversibly to haemoglobin inside red blood cells, and the more glucose there has been, the more of it sticks. Red cells live about three months, so the percentage of glycated haemoglobin is a running average of blood sugar over roughly 8 to 12 weeks. It cannot be faked by behaving well the week before an appointment, which is exactly why it became the standard measure.

The types

Don't be confused: type 1 is not "the childhood one" and type 2 is not "the one you gave yourself." Roughly half of type 1 diagnoses occur in adults. Type 2 now appears in teenagers. And while excess weight is the dominant risk factor for type 2, the genetic contribution is large (larger than for type 1 by twin studies), which is why plenty of heavy people never develop it and plenty of slim people do. Telling someone their diabetes is their fault is both cruel and factually shaky.

The history

In short: Described 3,500 years ago, uniformly fatal until 1922, when insulin was isolated in Toronto and the patent sold for one dollar.

Diabetes is one of the oldest recorded diseases and one of the most recently made survivable.

An Egyptian medical text, the Ebers papyrus (about 1550 BCE), describes a condition of excessive urination. Around the second century CE, the Greek physician Aretaeus of Cappadocia named it diabetes, meaning "siphon" or "to pass through," describing patients as melting flesh into urine. Physicians in India and China independently noted that such urine attracted ants and flies. In 1675 Thomas Willis added mellitus, Latin for honey-sweet, having tasted it. In 1776 Matthew Dobson demonstrated that the sweetness came from actual sugar.

For the next 150 years there was no treatment. Type 1 diabetes was a death sentence, usually within a year of diagnosis, and the only intervention that bought time was near-starvation diets that left children skeletal.

The mechanism arrived in 1889, when Oskar Minkowski and Joseph von Mering removed a dog's pancreas and observed that it developed diabetes. Something in the pancreas controlled blood sugar.

In the summer of 1921 in Toronto, Frederick Banting, Charles Best, J.J.R. Macleod, and the biochemist James Collip extracted and purified that something. On 11 January 1922 they injected an extract into Leonard Thompson, a 14-year-old dying of diabetic ketoacidosis. The first, impure batch produced only an abscess. Collip's refined preparation, given twelve days later, drove his blood sugar down and revived him. He lived another 13 years. The discoverers sold the patent to the University of Toronto for one dollar each, on the principle that insulin belonged to the world, an intention the modern insulin market has not honoured.

The dates that follow matter because they explain the treatments in this chapter:

What actually goes wrong

In short: Type 1 destroys the insulin factory; type 2 is cells ignoring the signal until the factory wears out compensating.

The normal system

After a meal, glucose enters the blood from the gut. Beta cells in the pancreas sense the rise and release insulin, a small protein hormone. Insulin binds receptors on muscle, fat, and liver cells and does four things:

1. Muscle and fat: it triggers transporter proteins called GLUT4 to move to the cell surface, opening doors for glucose to enter. Without insulin, those doors are mostly closed.
2. Liver: it tells the liver to stop producing glucose and to store it as glycogen.
3. Fat tissue: it stops the breakdown of fat.
4. Overall, it signals "fed state: store, do not release."

Between meals, insulin falls and glucagon (from pancreatic alpha cells) rises, telling the liver to release stored glucose so the brain keeps running. The pair behaves like a thermostat with a heater and a cooler.

Type 1: the factory is destroyed

In type 1 diabetes, the immune system's T cells attack and destroy beta cells, over months to years, until insulin production essentially stops. Autoantibodies against beta cell proteins (GAD65, IA-2, insulin itself, ZnT8) are detectable in blood years before symptoms, which is what makes early screening and delay strategies possible. Risk is strongly linked to particular HLA immune genes, and something environmental (enteroviral infection is the leading suspect) appears to trigger it, though no trigger is proven.

Symptoms appear abruptly, usually once about 80 to 90 percent of beta cells are gone. With no insulin, muscle and fat cannot take glucose in, so blood glucose climbs while the cells starve. The liver, reading the absence of insulin as fasting, pours out more glucose, which makes it worse. Fat is broken down for fuel, producing acidic ketones. Blood turns acidic: diabetic ketoacidosis (DKA), which without treatment is fatal.

Type 2: the locks stop responding, then the factory wears out

Type 2 is a two-part failure over years.

Insulin resistance comes first. Muscle, liver, and fat cells respond weakly to insulin. The dominant driver is fat stored where it does not belong: inside liver and muscle cells, and in the abdomen around the organs (visceral fat), rather than under the skin. Fat inside these cells interferes with the insulin signalling cascade, and visceral fat secretes inflammatory cytokines that interfere further (the chronic inflammation of Chapter 13). Physical inactivity worsens it; muscle is the largest glucose sink in the body, and unused muscle takes up less.

The pancreas compensates, for a long time. Beta cells increase output, so blood glucose stays normal while insulin levels run two or three times high. This phase can last a decade or more with no symptoms and no abnormal glucose test.

Then compensation fails. Beta cells lose function and mass, from a mix of exhaustion, fat accumulation in the pancreas itself, and the toxic effect of high glucose and fatty acids on the cells themselves. Glucose rises into the prediabetic range, then the diabetic range. By the time type 2 diabetes is diagnosed, a substantial share of beta cell function is already lost, which is why it is often described as progressive.

The twin cycle model, developed by Roy Taylor's group, ties this together: excess fat in the liver drives excess fat export to the pancreas, and removing that fat, which weight loss does, can restart beta cell function. That model predicted, and the DiRECT trial confirmed, that type 2 diabetes can be reversed in many people.

What it does to the body

In short: High glucose damages small vessels in the eyes, kidneys, and nerves and large ones supplying heart, brain, and legs, so almost all the harm is vascular.

Persistent high glucose damages tissue through several parallel routes. Glucose sticks non-enzymatically to proteins, forming advanced glycation end products that stiffen collagen and blood vessel walls. Excess glucose entering the polyol pathway in cells that do not need insulin to take glucose up (nerve, retina, kidney, lens) consumes antioxidant capacity and produces osmotic stress. High glucose promotes inflammation and injures the lining of blood vessels.

The result is damage sorted into two families.

Microvascular (small vessel) damage is specific to diabetes:

Macrovascular (large vessel) damage is accelerated atherosclerosis (Chapter 21). Diabetes roughly doubles the risk of heart attack and stroke, and cardiovascular disease, not high sugar itself, is what most people with type 2 diabetes eventually die of.

The diabetic foot is where several of these meet, and it is worth understanding because it is the most preventable disaster in the disease. Neuropathy removes pain, so a blister or a stone in a shoe goes unnoticed. Poor circulation slows healing. High glucose impairs neutrophil function, so infection takes hold. An unnoticed ulcer becomes an infected one, then bone infection, then amputation. Diabetes remains the leading cause of non-traumatic lower limb amputation in most countries, and the majority of those amputations begin with an ulcer that a daily foot check would have caught.

Other effects: gastroparesis (damaged nerves to the stomach cause bloating and unpredictable digestion), erectile dysfunction (vascular and neural, often an early sign), frequent infections, and slow wound healing.

The two acute emergencies

Diabetic ketoacidosis (DKA), mostly type 1: high glucose, ketones, acidic blood, vomiting, deep rapid breathing, drowsiness, coma. Develops over hours to a day. Requires emergency intravenous fluids, insulin, and potassium.

Hyperosmolar hyperglycaemic state (HHS), mostly type 2 in older people: extremely high glucose (often above 600 mg/dL, 33 mmol/L) with severe dehydration and confusion, developing over days, with a higher death rate than DKA.

And the emergency created by treatment: hypoglycaemia, blood sugar too low, caused by insulin or sulfonylureas. Sweating, shaking, hunger, confusion, and at the extreme seizure and unconsciousness. It is treated with fast sugar, and it is the main reason tight control is not pushed to the limit in frail or elderly patients.

Is it deadly?

In short: About 3.4 million deaths a year, almost all through complications, and the trials show those complications are largely preventable.

Yes, though usually slowly and usually through its complications.

• About 3.4 million deaths in 2024 were attributed to diabetes, and that undercounts, since many people with diabetes die of heart disease or kidney failure and are recorded that way.
• Type 2 diabetes is associated with roughly 6 years of reduced life expectancy on average when diagnosed in middle age, and considerably more when diagnosed young.
• Type 1 diabetes, before insulin, was fatal within months to a couple of years. Today, with modern care, life expectancy is within several years of the general population and closing.
• Untreated DKA is rapidly fatal. With treatment, mortality in well-resourced settings is under 1 percent; in places without reliable insulin and emergency care, it remains a major killer of children.

The strongest counterpoint in the whole chapter: complications are not inevitable. The DCCT trial in type 1 showed that intensive glucose control cut retinopathy by 76 percent and nephropathy by around 50 percent, and its follow-up study found the treated group still had less heart disease decades later, a phenomenon named metabolic memory. UKPDS showed comparable benefits in type 2, and showed that controlling blood pressure mattered as much as controlling glucose.

Is it contagious?

No. Diabetes cannot be transmitted between people by any route: not by contact, saliva, blood, sex, sharing food, or living together. It is not caused by a microbe.

The confusion is worth addressing directly because it causes real harm. Families cluster for diabetes because they share genes and because they share meals, kitchens, and neighbourhoods, not because it spreads. Children with type 1 diabetes have been excluded from schools and camps by adults who believed otherwise, and adults with type 2 have been treated as contaminated. Neither has any basis.

The one genuine link between infection and diabetes runs the other way: high blood glucose impairs immune function, so people with poorly controlled diabetes catch infections more easily and recover from them more slowly, including tuberculosis, severe COVID-19, and skin infections.

Who gets it

In short: 589 million adults, over 250 million of them undiagnosed, with disease arriving at lower body weights in South Asian, East Asian, African, and Pacific populations.

Scale. 589 million adults aged 20 to 79, about 1 in 9. Projected to reach 853 million by 2050. About 252 million are undiagnosed, and in some regions more than half of all cases are undiagnosed. Global spending exceeded 1 trillion US dollars in 2024, around 12 percent of all health expenditure.

Geography. The largest absolute numbers are in China and India. The highest prevalence rates are in Pacific Island nations and in the Gulf states, where several countries exceed 20 percent of adults. Sub-Saharan Africa has the highest proportion of undiagnosed cases.

Ethnicity and body type. People of South Asian, East Asian, African, Hispanic, and Pacific Islander descent develop type 2 diabetes at lower body weights than people of European descent, at younger ages, and with less visible obesity. South Asian populations in particular carry more visceral and liver fat at a given BMI. This is why several countries use lower BMI screening thresholds for these groups. It is a real biological difference in fat distribution and beta cell reserve, not a difference in discipline.

Genes. Type 2 is highly polygenic, with over 500 identified risk loci (TCF7L2 being the strongest common one), and identical twin concordance is high. Type 1 is strongly linked to HLA genes but has lower twin concordance, meaning environment matters more for its onset than most people assume. Having a parent with type 2 diabetes roughly doubles to triples risk.

Sex and age. Type 2 rises steeply with age. Men develop it at slightly lower BMI than women. Gestational diabetes affects a substantial minority of pregnancies and identifies women at high subsequent risk of type 2.

Income. The gradient flipped over a century. Diabetes was once a disease of the wealthy; today, within high-income countries, it is concentrated among poorer people, tracking food environment, working conditions, stress, and access to care. Between countries, more than three-quarters of people with diabetes now live in low- and middle-income countries.

Treatment, and how each treatment works

In short: Type 1 requires insulin replacement; type 2 has a ladder of drug classes, two of which now prevent heart and kidney disease beyond anything they do to blood sugar.

Type 1: replace what is missing

There is no alternative to insulin, and the goal is to imitate a working pancreas.

Basal-bolus insulin uses two kinds: a long-acting insulin (glargine, detemir, degludec) once or twice daily to cover the liver's background glucose output, and a rapid-acting insulin (lispro, aspart, glulisine) with each meal, dosed to the carbohydrate eaten. The analogues are human insulin with small amino acid changes that alter how quickly the molecules clump and dissolve, and therefore how quickly they act.

Insulin pumps deliver rapid-acting insulin continuously through a small cannula, with programmable rates and mealtime boluses.

Continuous glucose monitors (CGM) measure glucose in the fluid under the skin every few minutes, showing not just a value but a direction, with alarms.

Closed-loop systems (the "artificial pancreas") connect the two: an algorithm reads the CGM and adjusts pump delivery automatically every few minutes. These are now standard care in many health systems and represent the largest practical improvement in type 1 management since the insulin pen.

Teplizumab is an antibody against the CD3 protein on T cells. Given to people in the early autoimmune stage before symptoms, it blunts the attack on beta cells and delays clinical type 1 diabetes by roughly two years on average. It does not prevent it.

Type 2: a ladder of mechanisms

The GLP-1 class gets a chapter of its own. Semaglutide (sold as Ozempic and Wegovy) and tirzepatide (Mounjaro and Zepbound) are the most prescribed and least understood drugs in modern medicine, and the row above compresses a great deal. What GLP-1 is, why it is called that, how these drugs were engineered from a lizard peptide, their full safety profile, and the crucial fact that they cannot replace insulin are covered in Chapter 66.

The order matters. Historically the sequence was metformin, then sulfonylurea, then insulin, chosen by glucose lowering alone. It has changed, because outcome trials showed that SGLT2 inhibitors and GLP-1 agonists prevent heart attacks, heart failure, kidney failure, and death to a degree not explained by their glucose lowering. In someone with diabetes plus heart or kidney disease, those two classes are now recommended regardless of how good the HbA1c already is, which is a genuine shift from treating a number to treating an outcome.

Metabolic surgery

Gastric bypass and sleeve gastrectomy produce diabetes remission in a large share of patients, often within days, before meaningful weight loss has occurred, through changes in gut hormone signalling as well as through later weight loss. Randomised trials consistently show better glucose outcomes than medical therapy in people with obesity and type 2 diabetes.

Treating the rest of the risk

Because most people with type 2 diabetes die of vascular disease, complete treatment is not just glucose:

• Blood pressure control, usually with an ACE inhibitor or ARB, which also protects the kidneys.
• A statin, for nearly everyone with diabetes over 40, because diabetes raises cardiovascular risk into the range where statins clearly pay.
• Annual retinal photography, which catches treatable retinopathy before vision changes.
• Annual kidney testing (urine albumin and blood creatinine).
• Annual foot examination, plus daily self-inspection.

What treatment costs

In short: Each class has a characteristic price, from metformin's diarrhoea to insulin's hypoglycaemia, and most are managed by dose, timing, or switching.

What the person can do

In short: Weight loss can put type 2 diabetes into remission, and exercise moves glucose into muscle without needing insulin at all.

This section is where diabetes differs most from other chapters, because the daily management belongs to the patient, and because several of these interventions are as powerful as the drugs.

Weight loss, for type 2, is the closest thing to a cure. In the DiRECT trial, a structured programme of total diet replacement followed by food reintroduction and maintenance produced remission (HbA1c below 6.5 percent off all diabetes medication) in 46 percent at one year and 36 percent at two years. The relationship with weight was almost linear: about 86 percent of those who lost 15 kg or more were in remission, versus about 7 percent of those who lost less than 5 kg. Remission is most achievable within the first few years after diagnosis, while beta cells can still recover.

Exercise works through a mechanism worth knowing. Contracting muscle moves GLUT4 transporters to the cell surface without needing insulin. That is why a walk after a meal lowers blood glucose even in someone with severe insulin resistance, and why exercise both lowers glucose acutely and improves sensitivity for a day or two afterwards. Resistance training adds muscle, which enlarges the tank glucose can go into. Guideline targets (about 150 minutes a week of moderate activity plus two resistance sessions) are useful, but the largest single gain is from doing something rather than nothing.

Diet. No single named diet has proven superior for everyone. What consistently helps: fewer refined carbohydrates and sugary drinks (they produce the sharpest glucose spikes), more fibre, and whatever pattern the person can sustain to produce an energy deficit if weight loss is the goal. Low-carbohydrate approaches reliably lower post-meal glucose and HbA1c in the short to medium term; Mediterranean-style patterns have the strongest cardiovascular outcome evidence. Anyone on insulin or sulfonylureas who changes their carbohydrate intake substantially needs their doses reviewed, because the same dose becomes an overdose.

Sleep and stress. Short sleep and untreated sleep apnoea worsen insulin resistance measurably. Sleep apnoea is common in type 2 diabetes and often undiagnosed.

Stop smoking. Smoking multiplies the vascular damage of diabetes; the two risks compound rather than add.

Daily foot check. Look at both feet, including between the toes and the soles, every day, especially if sensation is reduced. Shoes that fit. Never walk barefoot. This single habit prevents a large share of amputations.

Screening attendance. Eyes annually, kidneys annually, feet annually. All three detect damage during the window when it is still treatable and before it can be felt.

Vaccination. Diabetes raises the risk of severe outcomes from influenza, pneumococcal disease, and COVID-19, and these vaccines are recommended accordingly.

For anyone on insulin or sulfonylureas: carry fast-acting sugar, know the early symptoms of hypoglycaemia, and make sure someone at home and at work knows what to do.

Living with it

In short: The treatment is delegated to the patient, dozens of decisions a day, and the resulting burnout has a name and predicts worse control.

Diabetes is unusual in how much of the treatment is delegated to the patient. Someone with type 1 makes dozens of decisions a day about food, dose, activity, and correction, each with immediate consequences. The term diabetes distress describes the resulting burnout, which is common, distinct from clinical depression, and predicts worse glucose control. Depression itself is roughly twice as common in people with diabetes as without, in both directions of causation.

The financial burden is severe where insulin is not affordable. Insulin rationing has been documented in high-income countries with weak price controls, and it kills people; the resulting DKA is exactly the death the 1922 discovery was meant to end.

Two social points worth stating plainly. Type 1 and type 2 being conflated means people with type 1, an autoimmune disease no behaviour caused, absorb blame aimed at type 2. And blame aimed at type 2 is itself poorly founded and counterproductive, since shame is not a documented driver of sustained behaviour change while access, cost, and support are.

What's next

• Stem-cell derived islet transplantation. Lab-grown insulin-producing cells transplanted into people with type 1 have restored insulin production and eliminated insulin injections in early trials. The remaining obstacle is immune rejection, which currently requires immunosuppression; encapsulation and gene-edited hypoimmune cells are the two routes being pursued.
• Once-weekly insulin, reducing basal injections from 365 to 52 a year.
• Oral GLP-1 drugs, moving the most effective class from injection to tablet.
• Better closed-loop systems, including fully automatic ones that need no meal announcement, plus dual-hormone pumps delivering glucagon as well as insulin.
• Earlier interception of type 1, combining autoantibody screening of children with immune therapies like teplizumab.
• AI retinal screening, already deployed, where a camera and an algorithm perform the annual eye screen without a specialist, which matters most where specialists are scarce.

Sources and notes

Prevalence, undiagnosed proportion, mortality, and expenditure figures are from the International Diabetes Federation, IDF Diabetes Atlas, 11th edition (2025), reporting 2024 estimates: 589 million adults aged 20 to 79 (1 in 9), 252 million undiagnosed, 3.4 million deaths, over 1 trillion USD, and a projection of 853 million by 2050. Diagnostic thresholds are the American Diabetes Association Standards of Care, consistent with WHO criteria. DCCT: NEJM 1993, with EDIC follow-up. UKPDS: The Lancet 1998. DiRECT remission figures: Lean et al., The Lancet 2018 and Lancet Diabetes & Endocrinology 2019 (46 percent at 12 months, 36 percent at 24 months; remission by weight-loss band as cited). Twin cycle hypothesis: Taylor et al. Cardiovascular and renal outcome data for SGLT2 inhibitors and GLP-1 agonists come from EMPA-REG OUTCOME, LEADER, SUSTAIN-6, DAPA-HF, CREDENCE, and related trials. Teplizumab delay: Herold et al., NEJM 2019. Insulin discovery history: Bliss, The Discovery of Insulin. Life expectancy reduction estimates vary by cohort and age at diagnosis and are given as approximations.

Open questions. How durable type 2 remission is beyond five years, whether very early intensive treatment changes the disease trajectory permanently, and whether the muscle loss accompanying rapid GLP-1-induced weight loss has long-term consequences are all unresolved. The environmental trigger for type 1 remains unidentified.

Next: the condition that sits upstream of most type 2 diabetes, and is the most argued-about diagnosis in medicine. 👉