Skip to content

Insulin and insulin sensitivity during a fast

Also called: Selective insulin resistance

Written by Reviewed Sep 2026

Insulin falls fast and early, and the body's tissues respond to it less, but only when it comes to sugar. That detail is what separates a normal fasting change from the insulin resistance of metabolic disease.

What it is

Insulin is made by beta cells in the pancreas. They release it when blood sugar rises, and also in response to protein building blocks and certain gut hormones. Insulin is the "store it" signal. It moves sugar into muscle and fat by bringing the sugar doorway GLUT4 to the cell surface. It helps store sugar as glycogen and build fat. It stops the liver from pouring out sugar. It switches on mTORC1, the build switch. And, most important for a fast, it strongly stops stored fat from being broken down.

That last effect is extremely sensitive. Insulin holds back fat breakdown at levels far below what it takes to push sugar into muscle. So the fall in insulin is the event that lets everything else happen: fat release, ketone making, and the switch in which fuel the body burns.

Then comes the part that confuses people. Measure how well the body responds to insulin during a fast, using the standard test, and you will find it lower. On paper, the person has become insulin resistant. To see why that is the right answer for the body, and not a warning sign, you need one specific experiment in people.

How it works

Insulin falls early and falls a long way. Six healthy men were sampled seven times through a 72 hour fast. Their insulin fell about 50%, from 64.6 to 30.1 pmol/L. 70% of that whole drop had already happened in the first 24 hours. Blood sugar fell about 25% over the same time. The release of fat into the blood (glycerol and palmitic acid) roughly doubled. 60% of that rise came between 12 and 24 hours, and the single biggest jump came between 18 and 24 hours. The authors conclude that early fat release follows the fall in insulin, not changes in blood sugar.

The resistance that develops is selective, and that was measured directly. Six lean people had a series of clamp tests, where insulin is dripped in at a set rate while blood sugar is held steady. They were tested at four insulin rates, before and after a 48 hour fast. After fasting, their bodies used much less sugar at every insulin rate. But insulin's power to hold down fat in the blood, at a level of 30 microunits per mL, was almost exactly the same before and after the fast. Sugar handling became resistant. The brake on fat release did not.

That is exactly the setup a fasting body needs. Sugar has to be saved for the tissues that can use nothing else. Early on, that is mainly the brain. Red blood cells and the inner part of the kidney always need sugar. Making muscle and fat less responsive to insulin's sugar signal is how the body saves it. Meanwhile, the brake on fat release has to stay working, because unchecked fat release would drive ketone making out of control.

Rising fat in the blood is part of how this works. Burning fat raises certain by-products in the cell, and those slow down the enzymes that burn sugar. Fat by-products also get in the way of insulin's signal inside muscle. This is a competition between fuels, and it reverses when carbs come back.

What switches it on

Falling blood sugar takes away the signal that makes beta cells release insulin. Rising fat in the blood and rising pushback hormones (the ones that raise blood sugar) then cause the selective resistance in muscle and fat.

On the timing
The fall in insulin is the fastest change on this site, and it is well mapped in people. 70% of the whole 72 hour drop is done in the first 24 hours. At 36 hours, the first quick burst of insulin after sugar is slightly smaller. At the same time, insulin works better in the liver, and the body as a whole is more insulin resistant. By 48 hours, the body's sugar use during a clamp test has fallen from 39.8 to 24.1 micromol per kg per minute.

What has been measured

In people

Almost everything on this page was measured in people. Insulin is in the blood, and the clamp test was built for people. The key result, the one that clears up the confusion, is a 1983 study. It measured sugar use and fat release in the same people, at the same insulin levels.

  • The largest study: 33 healthy people through a 72 hour fast, used to set normal ranges. Insulin, C-peptide and proinsulin all fell, along with blood sugar. Glucagon and fat in the blood rose steadily. People with a higher BMI had higher insulin and C-peptide all through the fast (American Journal of Physiology: Endocrinology and Metabolism, 2001).
  • 36 hours against 12 hours, in 13 healthy young men. At 36 hours, the first quick burst of insulin was slightly smaller. At the same time, insulin worked better in the liver, and the whole body was more insulin resistant. The authors' own take is worth keeping: the smaller insulin release at 36 hours may be a healthy response to the liver working better, not a fault (American Journal of Physiology: Endocrinology and Metabolism, 2021).
  • The key selectivity study, in 6 lean people. They had clamp tests at four insulin rates, from low to very high, before and after a 48 hour fast. Fasting sharply cut sugar use at every insulin rate. But the drop in fat in the blood at an insulin level of 30 microunits per mL was almost exactly the same before and after fasting. Blood phosphate fell during the insulin drip in everyone, and the bigger the fall, the more sugar was used. Blood potassium also fell in everyone, but that had nothing to do with fasting or sugar use. The authors conclude that starving causes selective insulin resistance (Metabolism, 1983).
  • The hour-by-hour record, in 6 healthy men, sampled at 12, 18, 24, 30, 42, 54 and 72 hours with tracer tests. Blood sugar went from 5.58 to 4.14 mmol/L. Insulin went from 64.6 to 30.1 pmol/L, with 70% of the drop in the first 24 hours. Fat release into the blood roughly doubled, with 60% of the rise between 12 and 24 hours. Early fat release followed the fall in insulin, not any change in blood sugar (American Journal of Physiology, 1993).
  • 48 hours, tested with a clamp, in 6 normal-weight men without diabetes. The body's sugar use fell from 39.8 to 24.1 micromol per kg per minute. The rate of sugar burned for energy crashed, from 21.8 to 3.9. The rate of sugar stored did not drop, going from 18.0 to 20.2. Sugar uptake in the forearm fell from 59.4 to 15.4. That pattern, burning crashes while storing holds, is a change in fuel choice. It is not the kind of sugar problem seen in diabetes (Metabolism, 1990).
  • Whether repeated fasting improves insulin response over time, with weight loss taken out of the picture. Men with prediabetes ate either within a 6 hour window with dinner before 3pm, or across a normal 12 hour day, for 5 weeks. Then they switched. All their food was provided, with enough to keep their weight steady. The early eating window improved insulin response, beta cell response, blood pressure, oxidative stress (a kind of wear on cells), and appetite, all without weight loss (Cell Metabolism, 2018).

In animals and in cells in a dish

The fine detail of how insulin's signal travels inside a cell, and how fat causes resistance, comes mostly from rodent muscle and liver, and from fat and muscle cells grown in a dish. That work explains how it happens. The clamp studies in people show what actually happens in a person during a fast.

  • In rodent muscle and liver, and in fat and muscle cells in a dish: the chain of signals from the insulin receptor to the sugar doorway GLUT4, and the fat by-products that are linked to insulin resistance.
  • A review that combines rodent and human evidence on how the liver's sugar output is controlled. It covers how fat makes the liver resistant to insulin, and how new sugar making is controlled both directly in the liver and from outside it (Nature Reviews Endocrinology, 2017).

Why it matters

Two opposite mistakes get passed around about this, and the selectivity finding knocks down both. One is alarm: fasting causes insulin resistance, so fasting is bad for your metabolism. The other is brushing it off: the resistance seen in a fast proves that insulin resistance in general is harmless. Neither holds up against the measurement.

The fasting kind and the disease kind differ on three points you can check. How long it lasts: the fasting kind clears within hours to a day of eating carbs again, while the disease kind lasts months or years. Insulin level: the fasting kind happens with low insulin in the blood, while the disease kind happens with high insulin, because the pancreas is working overtime to make up for it. What it spares: the fasting kind leaves insulin's brake on fat release working, while the disease kind breaks it. That is why fat in the blood stays high after meals in metabolic disease, the opposite of the fasting picture.

There is a practical point worth knowing. A glucose tolerance test done on someone who just finished a fast of several days, or who has been eating very low carb, will look abnormal for reasons that have nothing to do with disease. The standard fix is several days of normal carb intake before the test.

A claim you will see repeated

The claim that fasting lowers insulin and so improves insulin response mixes up two things measured at different times. During the fast, sugar handling is measurably worse, not better. Whether repeated fasting improves insulin response afterwards is a separate question. The best-controlled answer in people is a study where all food was provided, in men with prediabetes, and their insulin response did improve without weight loss. Note that study's limits. It was men only, with prediabetes, small, and 5 weeks long. It used an early eating window that ended mid-afternoon. Body clock timing was part of the idea being tested, so the result does not carry over automatically to the late eating window that most people actually use.

Sources

  1. Study in people1983
    Insulin action during acute starvation: evidence for selective insulin resistance in normal man

    Metabolism

    6 lean people were tested before and after a 48 hour fast, with insulin given at four levels while blood sugar was held steady. After the fast, their bodies used much less sugar at every insulin level. But insulin held back fat release just as well as before. So the fast made them resist insulin for sugar, but not for fat. Blood phosphate fell during insulin in everyone.

    6 lean subjects, sequential euglycaemic glucose clamps at four insulin infusion rates from 6 to 442 mU per square metre per minute, before and after a 48 hour fast. Fasting markedly reduced glucose utilisation at all infusion rates while the decline in free fatty acids at 30 microunits per mL insulin was virtually identical before and after. Serum phosphate fell during insulin infusion in all subjects (P < 0.001) and correlated with glucose disposal rate at r = 0.76, P < 0.005. Plasma potassium fell but did not relate to fasting or to glucose disposal.

  2. Study in people1993
    Progressive alterations in lipid and glucose metabolism during short-term fasting in young adult men

    American Journal of Physiology

    6 healthy young men were sampled from 12 to 72 hours into a fast. Blood sugar fell from 5.58 to 4.14 mmol/L. Insulin fell about 50%, and 70% of that drop came in the first 24 hours. Fat release rose, with 60% of the rise between 12 and 24 hours. Early fat release followed insulin, not blood sugar.

    6 healthy men sampled at 12, 18, 24, 30, 42, 54 and 72 hours with stable isotope tracers and indirect calorimetry. Glucose 5.58 plus or minus 0.08 to 4.14 plus or minus 0.10 mmol/L. Insulin 64.6 plus or minus 12.9 to 30.1 plus or minus 7.9 pmol/L, about a 50% fall, with 70% of the decline inside the first 24 hours. Glycerol rate of appearance 2.08 plus or minus 0.22 to 4.36 plus or minus 0.36 and palmitic acid 1.63 plus or minus 0.20 to 3.26 plus or minus 0.40 micromol per kg per minute, with 60% of the rise between 12 and 24 hours. Early lipid mobilisation tracked insulin, not glucose.

  3. Study in people2001
    Reference intervals for glucose, beta-cell polypeptides, and counterregulatory factors during prolonged fasting

    American Journal of Physiology: Endocrinology and Metabolism

    33 healthy adults fasted for 72 hours. Insulin and related markers fell along with blood sugar. Glucagon and free fats in the blood rose steadily, and stress hormones rose following the daily body clock. People with a higher BMI had higher insulin during the fast.

    33 healthy subjects across a 72 hour fast. Insulin, C-peptide and proinsulin all fell (P < 0.001) alongside glucose, while glucagon and free fatty acids rose linearly and the counter-regulatory hormones rose with a circadian rhythm superimposed. Higher BMI produced higher insulin and C-peptide during the fast.

  4. Study in people1990
    The effect of starvation on insulin-induced glucose disposal and thermogenesis in humans

    Metabolism

    6 healthy, normal-weight men without diabetes got a fixed dose of insulin while blood sugar was held steady, before and after 48 hours without food. After the fast, the body cleared much less sugar, 24.1 against 39.8. Burning sugar for fuel fell from 21.8 to 3.9, while storing it did not change. Sugar uptake in the forearm fell from 59.4 to 15.4. The heat boost from insulin seen when fed was gone after the fast.

    6 non-diabetic normal-weight men clamped at 100 mU per minute per square metre insulin and 3.5 mmol/L glucose after 48 hours of starvation. Whole-body glucose disposal 39.8 plus or minus 4.6 to 24.1 plus or minus 2.1 micromol per kg per minute (p < 0.01). Glucose oxidation 21.8 plus or minus 1.3 to 3.9 plus or minus 1.4 (p < 0.001), non-oxidative disposal unchanged. Forearm glucose uptake 59.4 to 15.4 micromol per minute per litre of forearm (p < 0.01). Net thermogenesis during the clamp was significant when fed and not significant after starvation.

  5. Study in people2021
    Impact of prolonged fasting on insulin secretion, insulin action, and hepatic versus whole body insulin secretion disposition indices in healthy young males

    American Journal of Physiology: Endocrinology and Metabolism

    13 healthy young men fasted 36 hours and, on another occasion, 12 hours. After 36 hours, the first burst of insulin was slightly smaller. The liver responded to insulin better, but the body as a whole responded worse. The authors think the smaller insulin burst may be a healthy response to the better liver response, not a defect.

    13 healthy young males, 36 hour versus 12 hour fasting. A minor reduction in first-phase insulin secretion alongside improved hepatic insulin action and increased whole-body insulin resistance. The authors frame the reduced secretion as potentially a healthy response to improved hepatic insulin action rather than a defect.

  6. Study in people2018
    Early Time-Restricted Feeding Improves Insulin Sensitivity, Blood Pressure, and Oxidative Stress Even without Weight Loss in Men with Prediabetes

    Cell Metabolism

    Men with prediabetes ate all their food in a 6 hour window ending before 3pm for 5 weeks, and on a 12 hour schedule for another 5 weeks. They were fed enough to keep their weight the same. The early window improved insulin response, insulin-making cells, blood pressure, oxidative stress and appetite, without weight loss.

    Supervised controlled feeding randomised crossover in men with prediabetes. A 6 hour eating window with dinner before 3pm versus a 12 hour control schedule for 5 weeks each, with participants fed enough to maintain weight. Insulin sensitivity, beta-cell responsiveness, blood pressure, oxidative stress and appetite all improved without weight loss.

  7. Review of studies2017
    Regulation of hepatic glucose metabolism in health and disease

    Nature Reviews Endocrinology

    A review of rat, mouse and human research on how the liver controls its sugar output. It covers how fat build-up makes the liver resist insulin, and control from the liver itself versus from other organs.

    Rodent and human evidence on hepatic glucose flux control, including lipid induced hepatic insulin resistance and the distinction between direct hepatic and indirect extrahepatic control of gluconeogenesis.