Fasting to Reverse Insulin Resistance & Diabetes

Picture a doorman at a busy venue. Insulin is the pass that tells him to open the door so glucose can come inside the cell and be burned for energy. Early on, the system is elegant: the pass appears, the door opens, the sugar goes in. But if passes are waved at that door every ninety minutes, all day, for twenty years — breakfast, coffee with milk, mid-morning pastry, lunch, afternoon biscuit, dinner, evening snack — the doorman stops taking them seriously. The pancreas responds the way anyone would when ignored: it shouts louder, releasing more insulin. The doors stay shut anyway. That is insulin resistance, and it sits underneath type 2 diabetes, metabolic syndrome, fatty liver and a large share of cardiovascular risk.

The crucial and often-missed detail is that high insulin is not only the response to resistance — it is also a cause of it. A substantial literature, including work in Cell Metabolism and Diabetologia, shows that chronically elevated insulin drives downregulation of its own receptors. The pancreas secretes more to overcome the blockage, and that hyperinsulinemia deepens the very desensitization it was trying to fix. It is a loop, and most glucose-lowering medications do not exit it, because they lower the number on the meter while leaving insulin exposure high or higher. Fasting interrupts the loop at a different point. When no food arrives, blood glucose drifts down, insulin secretion falls sharply, and receptors on muscle, liver and fat tissue get something they may not have had in years: a period of quiet. Studies using HOMA-IR, the standard clinical index of insulin resistance, consistently show marked improvement after even short fasting periods.

Underneath the receptor story sits a structural one. During the first twelve to twenty-four hours without food, the body draws down glycogen, the sugar stored in liver and muscle — the metabolic equivalent of running the house on battery after the grid fails. When those stores thin out, the body switches to fat, breaking triglycerides into fatty acids and converting them into ketone bodies, principally beta-hydroxybutyrate. That switch matters because of where the fat comes from. Ectopic fat — fat lodged inside the liver and pancreas, organs never designed to store it — is now understood as a primary driver of type 2 diabetes. A fatty liver keeps manufacturing glucose even when blood sugar is already high. A fat-infiltrated pancreas loses its timing, releasing insulin late and imprecisely. Roy Taylor's group at Newcastle University showed that removing even modest amounts of this organ fat was enough to restore near-normal insulin secretion in people who had carried a diagnosis for years. Fasting accelerates that mobilization because it forces the body to burn internal stores as its main fuel.

The clinical evidence is real but should be described honestly. A widely cited BMJ Case Reports series followed three medically supervised patients with type 2 diabetes through therapeutic fasting protocols; all three came off insulin. That is three people, uncontrolled — striking, not definitive. The Virta Health trials, using nutritional ketosis rather than fasting but working through the same insulin-lowering mechanism, are much larger: most insulin-using participants reduced or eliminated their doses within a year, and a substantial proportion met remission criteria at two years. That is strong signal. What we do not yet have is a large randomized trial of extended water fasting in diabetes with long follow-up. The mechanism is well-supported; the outcome data is promising and incomplete.

In working life, the problem rarely looks like overeating. It looks like grazing. A protein bar at the desk, oat milk lattes, tasting things while cooking, a glass of wine at nine. None of it is a meal, and all of it is an insulin signal. Someone can eat modestly, exercise, and still keep insulin elevated for sixteen hours a day. Frequency, not just quantity, is the exposure that matters.

Three misreadings do real damage. The first is treating the glucose number as the disease. Normal readings maintained by rising insulin doses describe compensation, not health, which is why fasting insulin and HOMA-IR are more revealing than glucose alone. The second is assuming fasting is simply eating less; the hormonal quiet of a genuine fast is not reproduced by spreading the same calories across six small meals. The third is the most consequential: fasting is not equally safe on all medications. Metformin is generally well tolerated, but sulfonylureas and injected insulin carry genuine hypoglycemia risk when food is withheld, and doses need active adjustment by a prescribing clinician. Anyone with type 1 diabetes, a history of eating disorders, or who is pregnant should not undertake extended fasting at all.

The counterintuitive closer: the people who find early fasting hardest are often the ones with the most insulin resistance. Their hunger and shakiness is not evidence that fasting is wrong for them — it partly reflects a metabolism that has forgotten how to access its own fat. The difficulty is a symptom of the condition, not a verdict on the intervention.