Cozy Butter · How Insulin Resistance Progresses

Why it goes unfound for so long

A routine blood panel measures glucose. It does not measure insulin. For as long as the pancreas can hold blood sugar in the normal range by working harder, the panel reports a normal result, and the report is accurate. It is answering a different question from the one that matters at that point.

The scale of this is in the national figures. The CDC estimates that 115.2 million American adults, more than two in five, have prediabetes, and that eight in ten of them do not know. Prediabetes is already the stage a standard test can detect. Everyone in the two stages before it is uncounted.

The clearest measurement of how early this starts comes from the Whitehall II cohort, which followed 6,538 British civil servants and then looked backward from the point of diagnosis. People who went on to develop type 2 diabetes already differed from those who did not 13 years before the diagnosis, which was as far back as the study could see. Their insulin sensitivity fell more steeply over the last five years, and their beta cell function rose until about three years before diagnosis and then fell, which is the same rise and fall the bars above show.

If you have been told for years that your bloodwork is fine and then received a prediabetes or type 2 result that felt sudden, this is the reason. The result was not sudden. The test that could see it only starts working near the end.

Why two people at the same weight get different results

Researchers at Newcastle proposed that each person has an individual amount of fat they can store in fat tissue before it starts accumulating inside the liver and the pancreas instead. They called it a personal fat threshold. Once fat is inside those two organs, the liver becomes less responsive to insulin and the pancreas releases less of it, and each of those makes the other worse.

The threshold sits at a different place for different people, which is the proposed reason one person develops type 2 diabetes at a weight another person carries without it, including people whose BMI is in the range usually called normal. It also explains why weight loss reverses the process in the order it does. Liver fat falls within days to weeks, and pancreas fat falls more slowly, which matches how quickly fasting glucose improves compared with the insulin response.

Insulin resistance, set point and weight regain

You will hear that insulin resistance holds your weight in place and pulls it back up after you lose it. Most of that picture is well supported. One specific piece of it is not, and the piece that fails is narrower than people assume. Setting out which part is which is more useful than picking a side, because the supported part is the part that explains what happens to people.

Body weight is biologically defended, and the evidence is strong

This is not a theory resting on associations. Much of it comes from experiments in which researchers deliberately changed people's weight and measured what the body did about it.

01

Energy expenditure falls, and stays down

In a study at Rockefeller, people were held at a weight 10 percent or more below their usual one, and their energy expenditure fell by more than their smaller body accounts for. The measured gap was about 6 calories per kilogram of fat-free mass per day in people who had never been obese and about 8 in people who had. The authors concluded that these changes oppose the maintenance of a body weight different from the usual weight, and may account for the poor long term results of obesity treatment. Estimates of how large this effect is vary between studies. Its direction does not.

02

Replacing the missing hormone reverses it

The same group then gave weight-reduced people low-dose leptin, enough to restore it to what it had been before the weight loss. It reversed the drop in energy expenditure and returned muscle efficiency, sympathetic nervous system activity and thyroid hormones to baseline. This is the strongest piece of evidence on the page, because it is causal rather than observational.

03

The appetite hormones stay changed for at least a year

Fifty people who lost about 13.5 kg still had altered ghrelin, leptin, peptide YY, cholecystokinin and insulin a year later, along with increased hunger, even after regaining some weight. Those measurements are shown in full in the Food Noise Simulator.

04

The response is substantially inherited

Twelve pairs of identical twins were overfed 1,000 extra calories a day, six days a week, for a total of 84 days. How much weight each man gained, and where he put it, varied about three times as much between pairs as within them. What your body does with a surplus is not a matter of character.

The Endocrine Society's scientific statement on obesity gathers this literature and describes body weight as being actively defended, with the response to weight loss stronger than the response to weight gain. That asymmetry is what people are describing when they say the weight went on more easily than it came off.

What is still being worked out is the model, not whether it happens

Researchers do disagree about the shape of the system. A 2023 review by two of the people on opposite sides of that argument went through the set point model, the settling point model and several others together. Their conclusion was that no single model yet accounts for everything, and the one they found most workable proposes an upper and a lower threshold with a range in between, rather than one exact defended number.

That is a disagreement about the mechanism, not about whether the defence exists. Every model in that review has to account for the experiments above, because those results are not in dispute. If you use set point as shorthand for a weight your body works to protect, the science is behind you. What it does not support is a single fixed number you were assigned at birth and can never move.

The defended level can move upward

This is the part that matters for regain. The leading explanation is that the level the body defends can itself rise, and is then defended at the new, higher place. In rodents, inflammation and reactive changes appear in the hypothalamus within days of a high fat diet, before much weight has been gained. An MRI study of 34 people found evidence consistent with the same process in the mediobasal hypothalamus of people with obesity, and it tracked with BMI. The sample is small, imaging is indirect, and this is an active area rather than a settled one, but it fits what people describe, which is a body defending a weight it did not previously hold.

Where insulin actually fits

Two different things are called insulin resistance, and the popular version of this claim runs them together. Separating them is what makes the question answerable, because one half of it holds and the other does not.

In the brain, this holds

Insulin is one of the signals that reports your fat stores

Insulin has been understood for decades as a signal that reports body fat to the brain. It is released in proportion to how much fat you carry, it reaches the brain, and it acts there alongside leptin in the feedback loop that regulates eating and body weight. The human brain is an insulin-sensitive organ, and insulin acts on a limited number of specific areas, including ones governing eating behaviour and whole-body metabolism. Impaired insulin action in the brain is found in people with obesity and can be studied in them directly. So insulin does sit inside the system that sets the defended level. How much it moves that level is still being established, and that part is honestly unfinished rather than settled.

In the body, this does not

Your fasting insulin number does not predict your regain

The insulin resistance a clinician could measure in your blood is a different quantity from the brain signal. In the RISC study, neither insulin sensitivity nor insulin secretion predicted who would gain weight. A review of the weight loss studies that tracked leptin, ghrelin and insulin found those changes, taken alone, were not sufficient to predict who regained, and for insulin specifically one study found an association while five found none.

So the honest version is that insulin is part of the system that defends body weight, through what it does in the brain, and that is a real answer rather than a hedge. What does not hold up is the narrower claim that the peripheral insulin resistance measured in your bloodwork is what predicts or explains an individual person's regain. Both of those can be true at once.

The clearest measured account of what happens when weight comes back is still the appetite hormones and the drop in energy expenditure, which are shown in the Food Noise Simulator. When people stop a GLP-1, about two thirds of the lost weight returns within a year and the metabolic improvements move back toward baseline with it, which is what a defended weight looks like when the thing holding it down is removed.

What actually moves the progression

The Diabetes Prevention Program is the trial worth knowing here. It enrolled 3,234 adults at high risk, who had both impaired glucose tolerance and a raised fasting glucose, and followed them for an average of about three years. In the placebo group, around 11 percent developed type 2 diabetes each year. That rate belongs to this high-risk group rather than to everyone with prediabetes.

58%

Lower rate with an intensive lifestyle program

A structured program aiming at 7 percent weight loss and 150 minutes of activity a week. Progression fell to about 5 percent a year.

31%

Lower rate with metformin

The medication arm of the same trial, at 850 mg twice daily, compared with placebo.

Neither number is a promise about one person, and the lifestyle result was produced by a supported program rather than by advice alone. What the trial establishes is that the progression is not fixed, and that the earlier stages are the ones that move most.

This is education, not medical advice

This page explains published research. It does not diagnose anything and it does not measure your own insulin or blood sugar. The bars are illustrative directions rather than units. Insulin resistance is not diagnosed from a single number, and fasting insulin testing is not a routine screening test. If you want to know where you are, that is a conversation with your clinician about which tests are appropriate for you.

Sources

Every source below is a peer-reviewed journal, a government agency or a professional society statement.

Centers for Disease Control and Prevention. National Diabetes Statistics Report, January 2026. cdc.gov. Source of the estimated 115.2 million US adults aged 18 and over with prediabetes. The accompanying CDC page Prediabetes: Could It Be You? is the source of that being more than two in five adults, and of eight in ten of them not knowing they have it.
Tabák AG, Jokela M, Akbaraly TN, Brunner EJ, Kivimäki M, Witte DR. Trajectories of glycaemia, insulin sensitivity, and insulin secretion before diagnosis of type 2 diabetes: an analysis from the Whitehall II study. The Lancet. 2009;373(9682):2215–2221. PMID 19515410. A cohort of 6,538 people, with trajectories traced back 13 years from diagnosis. Source of the timeline, of insulin sensitivity falling more steeply over the last five years, of beta cell function rising until about three years before diagnosis and then falling, and of fasting glucose rising steeply only in the final three years. The measured basis for the rise and fall shown in the two bars.
DeFronzo RA. Banting Lecture. From the triumvirate to the ominous octet: a new paradigm for the treatment of type 2 diabetes mellitus. Diabetes. 2009;58(4):773–795. PMID 19336687. Source for the compensating phase, for insulin rising before glucose does, and for people in the upper third of the impaired glucose tolerance range, defined on the two hour glucose test, having lost more than 80 percent of beta cell function while being close to maximally insulin resistant.
UK Prospective Diabetes Study Group. UK Prospective Diabetes Study 16: overview of 6 years' therapy of type II diabetes, a progressive disease. Diabetes. 1995;44(11):1249–1258. PMID 7589820. Source for beta cell function measuring about half of normal at the point of a type 2 diagnosis.
Taylor R. Aetiology of type 2 diabetes: an experimental medicine odyssey. Diabetologia. 2025. PMC12176950. Source of the twin cycle hypothesis and the personal fat threshold, of liver fat falling within days to weeks against pancreas fat falling more slowly, and of the individual variation in threshold proposed to explain different outcomes at the same BMI.
Diabetes Prevention Program Research Group. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. New England Journal of Medicine. 2002;346(6):393–403. NIDDK programme page. Source of the 3,234 participants, the roughly 11 percent annual progression rate on placebo falling to about 5 percent, and the 58 percent and 31 percent reductions for the lifestyle and metformin arms.
Wilding JPH, Batterham RL, Davies M, et al. Weight regain and cardiometabolic effects after withdrawal of semaglutide: the STEP 1 trial extension. Diabetes, Obesity and Metabolism. 2022;24(8):1553–1564. doi:10.1111/dom.14725. Source for participants regaining about two thirds of their lost weight in the year after stopping, and for the cardiometabolic improvements reverting toward baseline along with the weight.
Leibel RL, Rosenbaum M, Hirsch J. Changes in energy expenditure resulting from altered body weight. New England Journal of Medicine. 1995;332(10):621–628. PMID 7632212. Source for the measured reduction in total energy expenditure at a weight 10 percent or more below the usual one, reported as 6 plus or minus 3 calories per kilogram of fat-free mass per day in people who had never been obese and 8 plus or minus 5 in people who had, and for the authors' own conclusion that these changes oppose maintaining a weight different from the usual one and may account for the poor long term results of obesity treatment. Widely quoted daily-calorie versions of this figure are conversions rather than the number the paper reports, so this page uses the paper's own units.
Rosenbaum M, Goldsmith R, Bloomfield D, et al. Low-dose leptin reverses skeletal muscle, autonomic, and neuroendocrine adaptations to maintenance of reduced weight. Journal of Clinical Investigation. 2005;115(12):3579–3586. doi:10.1172/JCI25977. Source for leptin replacement reversing the decline in total and non-resting energy expenditure and restoring muscle work efficiency, sympathetic nervous system activity and thyroid hormones to baseline. The experimental evidence that the defence of body weight is causal rather than an association.
Bouchard C, Tremblay A, Després JP, et al. The response to long-term overfeeding in identical twins. New England Journal of Medicine. 1990;322(21):1477–1482. PMID 2336074. Source for twelve pairs of identical twins overfed 1,000 calories a day, six days a week, for a total of 84 days within a 100 day period, and for about three times as much variance in the response between pairs as within them.
Thaler JP, Yi CX, Schur EA, et al. Obesity is associated with hypothalamic injury in rodents and humans. Journal of Clinical Investigation. 2012;122(1):153–162. doi:10.1172/JCI59660. Source for hypothalamic inflammation appearing in rodents within days of a high fat diet and before substantial weight gain, and for the MRI finding of gliosis in the mediobasal hypothalamus in people with obesity, correlated with BMI, in 34 subjects.
Heni M, Kullmann S, Preissl H, Fritsche A, Häring HU. Impaired insulin action in the human brain: causes and metabolic consequences. Nature Reviews Endocrinology. 2015;11(12):701–711. PMID 26460339. Source for the human brain being an insulin-sensitive organ, for insulin acting on the hypothalamus in the regulation of whole-body metabolism and food intake, and for impaired brain insulin action being measurable in people with obesity.
Schwartz MW, Seeley RJ, Zeltser LM, et al. Obesity pathogenesis: an Endocrine Society scientific statement. Endocrine Reviews. 2017;38(4):267–296. PMID 28898979. Source for the biological defence of fat mass, for the tendency of lost weight to be regained, and for that defence being asymmetric, meaning stronger against weight loss than against weight gain.
Rebelos E, Muscelli E, Natali A, et al. Body weight, not insulin sensitivity or secretion, may predict spontaneous weight changes in nondiabetic and prediabetic subjects: the RISC study. Diabetes. 2011;60(7):1938–1945. PMID 21617179. Source of the finding that neither insulin sensitivity nor insulin secretion predicted spontaneous weight gain.
Strohacker K, McCaffery JM, MacLean PS, Wing RR. Adaptations of leptin, ghrelin or insulin during weight loss as predictors of weight regain: a review of current literature. International Journal of Obesity. 2014;38(3):388–396. doi:10.1038/ijo.2013.118. Source of the conclusion that changes in these hormones taken alone are not sufficient to predict weight regain, and of the count for insulin specifically, one study finding an association and five finding none.
Speakman JR, Hall KD. Models of body weight and fatness regulation. Philosophical Transactions of the Royal Society B. 2023;378(1888):20220231. doi:10.1098/rstb.2022.0231. Source for no single model of body weight regulation explaining all of the available evidence, for the specific problems with a strict set point, and for the dual intervention point model proposing an upper and a lower threshold with a range between them.
Sumithran P, Prendergast LA, Delbridge E, et al. Long-term persistence of hormonal adaptations to weight loss. New England Journal of Medicine. 2011;365(17):1597–1604. doi:10.1056/NEJMoa1105816. Source for the appetite hormone changes after weight loss persisting a year later. These are shown in full in the Food Noise Simulator.
Lean MEJ, Leslie WS, Barnes AC, et al. Primary care-led weight management for remission of type 2 diabetes (DiRECT): an open-label, cluster-randomised trial. The Lancet. 2018;391(10120):541–551. PMID 29221645. Source of the 46 percent remission figure at 12 months and of more than 80 percent remission among those who lost over 15 kg.
American Diabetes Association Professional Practice Committee. 2. Diagnosis and classification of diabetes: Standards of Care in Diabetes 2026. Diabetes Care. 2026;49(Supplement 1):S27. diabetesjournals.org. Source of the fasting glucose and A1C thresholds used at each stage.