Story
Energy Balance vs Carbohydrate-Insulin · An Ongoing Debate
Last updated
In one pass You and a colleague eat about the same, yet they stay lean and you gain weight. Not this — Insulin is the sole driver of fat gain — In metabolic-ward trials where every meal was controlled (Hall 2015 and 2021), a low-fat diet lost more body fat than a low-carb one at equal calories; eating freely, the low-fat plant-based group ate 689 fewer kcal a day. Under controlled testing, the carbohydrate-insulin model fails.
Educational content, not medical advice — consult a clinician.
Story path
Chapter 1
Eating too much, or eating wrong?
You and a colleague eat about the same, yet they stay lean and you gain weight. What is to blame? Nutrition science has spent a decade arguing over the serious version of that question: do people get fat because they eat too much, or because they eat the wrong things? So far, tightly controlled feeding trials favor the first answer.
The two camps answer differently:
The energy balance model (EBM) is the mainstream view. If you take in more than you use over the long run, you gain weight. How the calories split among carbohydrate, fat and protein is secondary; the split works mainly through how full you feel and whether you can stick with it, which in turn shape how much you actually eat.The carbohydrate-insulin model (CIM) reverses the causal arrow. Refined carbohydrate pushes insulin up first, insulin locks energy away in fat, and so you get hungry again quickly and have to eat more. On this account counting calories is wasted effort; what you should do is cut carbohydrate.
This is not a private quarrel among academics. The lines you hear in short videos, carbs make you fat and insulin is the master switch of obesity, all come from the CIM story. The point is not to pick a side, but to see how far apart a mechanism that sounds airtight and an experiment that was actually run can be.
The two camps answer differently:
The energy balance model (EBM) is the mainstream view. If you take in more than you use over the long run, you gain weight. How the calories split among carbohydrate, fat and protein is secondary; the split works mainly through how full you feel and whether you can stick with it, which in turn shape how much you actually eat.The carbohydrate-insulin model (CIM) reverses the causal arrow. Refined carbohydrate pushes insulin up first, insulin locks energy away in fat, and so you get hungry again quickly and have to eat more. On this account counting calories is wasted effort; what you should do is cut carbohydrate.
This is not a private quarrel among academics. The lines you hear in short videos, carbs make you fat and insulin is the master switch of obesity, all come from the CIM story. The point is not to pick a side, but to see how far apart a mechanism that sounds airtight and an experiment that was actually run can be.
Background · What each camp is betting on
The energy balance model (EBM) is the mainstream view; its leading voices include Kevin Hall at the US National Institutes of Health (NIH) and Dale Schoeller. It does not deny that hormones are at work. It places them as intermediate variables: insulin, leptin and ghrelin all influence how much you eat and how much you move, but they cannot rewrite the ledger of what came in minus what was used.The carbohydrate-insulin model (CIM) was set out systematically by David Ludwig and Cara Ebbeling at Harvard (Ludwig 2018). It turns the whole causal chain around: it is not you ate too much, so you got fat, but what you ate made you eat too much. Carbohydrate with a high glycemic index (, how fast a food raises blood glucose) → a sharp rise in insulin after the meal → more fat storage → blood glucose falls into a trough and you get hungry → you eat more → obesity. The conclusion it draws is blunt: carb restriction and keto are the real route to fat loss, and counting calories alone is futile.
Why this debate is worth knowing for ordinary people
People selling low-carb products mostly bet on CIM. The marketing lines are few: carbs make you fat, insulin is the master switch of obesity, forget calories, just quit sugar, keto cures everything. They sound far more exciting than eat less, move more, and they make it easy to sell books, courses and low-carb foods.If CIM is right, decades of mainstream public-health advice (cut total calories, eat a balanced diet) were wrong, and dietary guidelines would need rewriting.If CIM is wrong, you were sold a simplified causal story, and you missed the levers that actually help: total energy, protein, food quality and sustainability.
So what this story really teaches is not which camp won, but the distance between a mechanism story and clinical evidence: a mechanism that sounds entirely reasonable may not hold up once it is put inside a tightly controlled metabolic-ward experiment.
Chapter 2
What the carb-insulin model claims
This camp's claim is that what makes you fat is not eating too much but what you eat.
After a bowl of white rice, glucose floods into the blood and insulin is released in large amounts. In fat tissue, insulin does three things at once. It shuts the gate that lets fat be broken down, so stored fat cannot get out. It moves fat circulating in the blood into fat cells. And it hands surplus glucose to the liver to be turned into fat. According to CIM, energy is therefore rerouted into storage instead of being sent to muscle to be burned.
The trouble, it says, comes next. A strong surge of insulin tends to overshoot, blood glucose is driven low, the brain reads not enough energy and switches hunger on early, so you want to eat before the next meal is due. You eat, and the cycle turns again, a little larger.
Every step has a real biochemical basis and matches how many people feel, which is why the idea spread so fast. But it contains one sentence an experiment can test directly: at the same calories, low-carb should lose more fat than low-fat. The metabolic-ward trials that followed were built to test exactly that sentence.
After a bowl of white rice, glucose floods into the blood and insulin is released in large amounts. In fat tissue, insulin does three things at once. It shuts the gate that lets fat be broken down, so stored fat cannot get out. It moves fat circulating in the blood into fat cells. And it hands surplus glucose to the liver to be turned into fat. According to CIM, energy is therefore rerouted into storage instead of being sent to muscle to be burned.
The trouble, it says, comes next. A strong surge of insulin tends to overshoot, blood glucose is driven low, the brain reads not enough energy and switches hunger on early, so you want to eat before the next meal is due. You eat, and the cycle turns again, a little larger.
Every step has a real biochemical basis and matches how many people feel, which is why the idea spread so fast. But it contains one sentence an experiment can test directly: at the same calories, low-carb should lose more fat than low-fat. The metabolic-ward trials that followed were built to test exactly that sentence.
Mechanism · The four-step chain in full
Ludwig and Ebbeling 2018 (JAMA Internal Medicine) wrote the carbohydrate-insulin hypothesis up as a heavily cited paper. Its four steps connect like this.Step 1 · A high glycemic-index food pushes up blood glucose
After a bowl of white rice or a glass of juice, glucose floods into the blood within 30-60 minutes. Insulin is released in large amounts in response and pushes the glucose into cells.
Step 2 · Insulin is the fat-storage signal
Insulin suppresses lipolysis in fat tissue: it holds down the enzyme that breaks fat apart, hormone-sensitive lipase (HSL), so fat already stored cannot get out.Insulin activates lipoprotein lipase (), so in the blood are pulled into fat cells.It also stimulates the liver's de novo lipogenesis (DNL), making fat from scratch out of surplus glucose.According to CIM, the net effect is that energy is forced into fat storage rather than burned by muscle.
These three only carry weight when read together: one blocks the exit, one opens the entrance, and one keeps adding more. All three point at the warehouse, which is why CIM feels entitled to say that energy has been locked away.
Step 3 · Glucose trough → hunger → eating more
According to CIM, insulin overshoots and 1-3 hours later blood glucose falls into a trough (in severe cases, reactive hypoglycemia). The brain reads not enough energy and sends out hunger signals: ghrelin, the hunger hormone, rises, and the fullness feedback from leptin grows dull. You are hungry an hour early, grab another cookie, and the cycle amplifies.
Step 4 · The conclusion
It is not that eating too much made you fat; it is that what you ate made you eat too much.The fix: restrict carbohydrate or go keto → insulin stays low → fat can be released and you are not hungry → you naturally eat less.
Why this story spread: every step has a real biochemical basis. Insulin really does suppress lipolysis, and nobody disputes that part. The real argument is not over whether insulin does these things, but over whether their net result is large enough to decide whether you get fat.
Myth · Those first-two-week kilos are mostly not fat
Low-carb and keto diets drop weight fast at the start, and that is the most direct experience that makes many people believe in CIM. But what is dropping is the wrong thing.The short-term weight numbers look supportive: low-carb and keto diets really do lose weight faster in the first 1-2 weeks.
But most of that fast loss is water and glycogen, not fat. When the liver stores glucose as glycogen, every 1 g of glycogen is stored together with 3-4 g of water: glycogen attracts water and sits in tissue in a soaked state. Once you cut carbohydrate, glycogen is used first and not replaced, and that water leaves with it, so you can be 2-3 kg lighter within a week. The scale looks great, but that part is not fat loss.
This explains two things. First, why almost everyone gets keto's magic first two weeks. Second, why weight comes back so puzzlingly fast once you eat carbohydrate again: glycogen refills, the water returns with it, and the fat has barely moved in between.
So if you want to know whether a way of eating actually took off fat, the scale changes of the first two weeks are nearly useless. Either look over a longer period, or measure body composition directly; metabolic-ward trials do the latter.
The sentence of CIM that can truly be falsified is this: at the same calories, low-carb should lose more fat than low-fat. It ignores the scale and the water and looks only at fat itself, so an experiment can answer it directly.
Chapter 3
What metabolic-ward trials found
A few short trials run on a metabolic ward at the US National Institutes of Health (NIH) did not support the key predictions of the carbohydrate-insulin model.
Kevin Hall did not compete over whose mechanism sounded better. He had people live on a metabolic ward: every bite of food was weighed, activity was controlled, energy expenditure was tracked around the clock by indirect calorimetry, and body composition was measured directly with dual-energy X-ray absorptiometry () and other methods. How much you think you ate, the biggest source of error, was taken out of the picture.
Calories fixed, only the macronutrient split changed (6 days on each diet): on the days with less carbohydrate, fat loss was smaller; on the days with less fat, it was larger. The direction was the opposite of CIM's prediction.Eat as much as you like (two weeks on each diet): the keto side did not show the fabled freedom from hunger; it ate more instead. The low-fat plant-based side ate a good deal less of its own accord.A finding along the way: with the nutrients broadly matched and only the degree of processing changed, the ultra-processed side ate several hundred kilocalories more a day.
These trials lasted only days to two weeks. They do not say that low-carb or keto diets fail; these ways of eating do produce weight loss in many people. They say that the reason they work is not the one CIM gives.
Kevin Hall did not compete over whose mechanism sounded better. He had people live on a metabolic ward: every bite of food was weighed, activity was controlled, energy expenditure was tracked around the clock by indirect calorimetry, and body composition was measured directly with dual-energy X-ray absorptiometry () and other methods. How much you think you ate, the biggest source of error, was taken out of the picture.
Calories fixed, only the macronutrient split changed (6 days on each diet): on the days with less carbohydrate, fat loss was smaller; on the days with less fat, it was larger. The direction was the opposite of CIM's prediction.Eat as much as you like (two weeks on each diet): the keto side did not show the fabled freedom from hunger; it ate more instead. The low-fat plant-based side ate a good deal less of its own accord.A finding along the way: with the nutrients broadly matched and only the degree of processing changed, the ultra-processed side ate several hundred kilocalories more a day.
These trials lasted only days to two weeks. They do not say that low-carb or keto diets fail; these ways of eating do produce weight loss in many people. They say that the reason they work is not the one CIM gives.
Numbers · The isocaloric head-to-head
Hall 2015 (Cell Metabolism) · the isocaloric head-to-headDesign: 19 adults with obesity lived on a metabolic ward. After a 5-day baseline diet, they spent 6 days on a carbohydrate-restricted diet and 6 days on a fat-restricted diet, in two separate inpatient stays in random order, with the same calories on both diets.Diet A: calories cut by 30%, all of it from carbohydrate, with fat unchanged (low-carb).Diet B: calories cut by 30%, all of it from fat, with carbohydrate unchanged (low-fat).CIM's prediction: low-carb should lose more fat.Actual result: calculated directly from fat balance (daily fat intake minus the measured net fat oxidation), body fat fell by 53 g a day on carbohydrate restriction and by 89 g a day on fat restriction, significantly more with fat restriction. Cutting carbohydrate did make the body burn more fat, but no less fat was coming in; cutting fat left fat burning unchanged, but far less fat was coming in.
Crossover is the key here: each person ate both diets in turn and served as their own control. Basal metabolic rate, activity habits and gut bacteria, which differ hugely from person to person, all cancel out, so whatever difference remains can only come from the diet itself.
Conclusion: at equal calories, cutting fat rather than carbohydrate produced more body-fat loss as measured directly by fat balance. CIM's core assumption (that low-carb gives a fat-storage advantage) was not supported in this trial.
What is worth noticing is the direction of the gap, not only its size. CIM predicted more fat loss on the low-carb side; the measurement showed more on the low-fat side. A hypothesis whose prediction comes out backwards is in far worse shape than one whose predicted effect was not large enough: the second can still be explained away as a small effect, the first cannot.
The limits need stating too. This was a 6-day trial. The CIM side replies that the body had not yet adapted to low-carb eating, and the authors' own mathematical model predicted that over longer periods the body acts to shrink the body-fat difference between the two diets. So what it shows is that CIM's predicted direction came out backwards in the short term; it does not show that low-fat eating will always lose more fat in the long run.
Numbers · The eat-as-much-as-you-want trial
Hall 2021 (Nature Medicine) · the eat-freely versionDesign: 20 adults stayed at the NIH Clinical Center and ate each diet for 2 weeks, one period straight after the other, in random order.Diet A: a low-fat, mainly plant-based diet (75% carbohydrate, 10% fat).Diet B: a low-carb, mainly animal-based ketogenic diet (10% carbohydrate, 75% fat).The key point: ad libitum eating, as much as people wanted.CIM's prediction: without insulin spikes, the keto group should be less hungry and naturally eat less.Actual result: over the two weeks, the low-fat plant group ate 689 kcal a day less than the keto group; in the final week of each period alone the gap was 544 kcal, in the same direction. The keto group showed none of the fabled near-absence of hunger and ate more instead.
This trial tested CIM's most central behavioral claim. An isocaloric experiment can only answer where the fat went at the same calories; an ad libitum experiment answers how much you will eat on your own, and how much you will eat on your own is the link CIM uses to explain obesity. Let that link run freely, and the result still went against the prediction.
One detail often skipped: insulin levels in the keto group really were much lower. In other words, the mechanism half was right and the outcome half was not: insulin moved the way CIM said it would, but people did not eat less because of it. That is exactly where a mechanism story fools people most easily: every middle step can be right and the endpoint can still be wrong.
The limits need stating here too. Each diet lasted only two weeks, so adaptation to keto may not have been complete (the narrower gap in the final week hints at that), and ward meals are not real life. The trial answers who eats more when eating freely for two weeks, not who is leaner a year later.
Numbers · The finding picked up along the way
The Hall 2019 randomized trial of ultra-processed food (Cell Metabolism): the two diets were served broadly matched for calories, macronutrients, sugar, sodium and fiber, and people ate as much as they wanted. The result: people ate 508 kcal a day more in the ultra-processed two weeks than in the unprocessed two weeks, and were 0.9 kg heavier after 14 days.Notice that this trial did not change the macronutrient split: the shares of carbohydrate, fat and protein were the same on both sides. Energy density was matched only when drinks were counted in; leaving drinks out, the ultra-processed food itself was clearly more energy-dense. What changed was a set of properties that come with the degree of processing: the form and texture of the food, its energy density, and how easy it is to eat one bite after another. People also ate faster during the ultra-processed weeks. The authors think energy density and eating speed may both play a part, but the trial cannot tell which one matters most.
So the levers that really decide how much you eat may rank ahead of the macronutrient split: what the food is like, how long it takes to chew, how dense it is, and how easy it is to keep taking bites. That also explains how low-carb works and low-fat plant eating works can both be true: in real life both push out a large share of ultra-processed food. For the full story of this trial, see Ultra-processed Foods.
This chapter in sum: in the metabolic-ward trials published so far, CIM's key predictions were not supported. The CIM camp has feeding trials of its own whose conclusions and analysis are still disputed, and all these ward trials were short. Low-carb eating can still be a good choice, but the reason needs changing: the chapter further on about what really drives weight loss gives a different one.
Chapter 4
What actually drives weight loss
After a decade of debate, the mainstream view has settled into a few plain sentences, and consensus statements from bodies such as the American Diabetes Association also hold that no single macronutrient split is best for everyone.
The energy deficit is the deciding factor. Over months and longer, fat is lost because less comes in than goes out. Insulin and other hormones are intermediate variables: they affect how much you eat and how much you move, but they cannot change the ledger itself.
The macronutrient split is a secondary variable. At the same energy deficit, different splits produce only small differences in fat loss in . Where the split really matters is elsewhere: it decides whether you feel full, whether you enjoy the food, and whether you can stick with it, and those three work through how much you eat in total.
So low-carb works for some people and low-fat plant eating works for others, in both cases because it is easier to stick with, not because it bypasses some switch.
Nutrition has no macronutrient split that is best for everyone. An energy deficit, good food quality, enough protein, and a plan you can keep: those four are what to hold on to.
The energy deficit is the deciding factor. Over months and longer, fat is lost because less comes in than goes out. Insulin and other hormones are intermediate variables: they affect how much you eat and how much you move, but they cannot change the ledger itself.
The macronutrient split is a secondary variable. At the same energy deficit, different splits produce only small differences in fat loss in . Where the split really matters is elsewhere: it decides whether you feel full, whether you enjoy the food, and whether you can stick with it, and those three work through how much you eat in total.
So low-carb works for some people and low-fat plant eating works for others, in both cases because it is easier to stick with, not because it bypasses some switch.
Nutrition has no macronutrient split that is best for everyone. An energy deficit, good food quality, enough protein, and a plan you can keep: those four are what to hold on to.
Numbers · A full-year head-to-head, nearly a draw
DIETFITS (Gardner's team, JAMA 2018) · Stanford's 12-month head-to-headDesign: 609 overweight adults were randomized to a healthy low-fat diet or a healthy low-carb diet. Both groups were encouraged to eat whole foods, less sugar and fewer refined grains, with no calorie target.Result: weight loss at 12 months was almost the same: −5.3 kg on low-fat and −6.0 kg on low-carb, a difference that was not statistically significant.The pre-specified subgroups found no difference either: sorted by genotype or by baseline insulin secretion, there was no significant interaction showing that some people suit low-carb and others low-fat. The precision-nutrition prediction that follows from CIM did not hold up here either.
That last point deserves a second look. The CIM camp's later fallback was precision nutrition: even if the averages match, surely people who secrete a lot of insulin do better on low-carb? DIETFITS wrote that hypothesis into its protocol in advance and looked for a difference by genotype and by insulin secretion, and found none. Stating a hypothesis first and then looking is not the same as fishing a result out of the data afterwards; failing to find it the first way carries far more weight.
The key insight: diet quality matters more than diet type. Both groups were asked to eat whole foods, less sugar and fewer refined grains, and both lost weight; what the two diets shared clearly contributed more than the carbs-or-fat difference between them.
Mechanism · Why the macronutrient split comes second
Energy conservation is not a slogan. It is the floor of this whole debate.I eat almost nothing and still do not lose weight is, most of the time, not a broken metabolism but intake counted too low and activity counted too high. People tend to underestimate how much they eat and overestimate how much they move; stack the two errors and they can wipe out a deficit that really exists. That is also why metabolic wards, expensive as they are, still have to be used: what they buy is switching off both errors at once. The Hall 2017 review went through these ward trials and concluded that several key predictions of the carbohydrate-insulin model failed under tight control, and that the model is too simple.
At the same energy deficit, the difference in fat loss between macronutrient splits is small in , under 1 kg in a year-long trial such as DIETFITS. The difference is small because the split works by an indirect route: it first affects fullness and how well you stick with a plan, and those two then affect total intake. A path with two bends in it cannot produce a large effect.
Low-carb works for some people because it is easier to stick with, not because of metabolic magic
Some people find high-fat, high-protein meals keep hunger away better, stick more easily to low-carb eating, and so naturally eat less. The effect is real; the mechanism is ordinary: total energy fell, not because anyone bypassed insulin, the master switch. But who those people will be is hard to predict in advance: DIETFITS grouped people by baseline insulin secretion and still found no one who suited low-carb better. So this is something to test on yourself, not a rule you can apply ahead of time.
Low-fat plant eating works for other people
People who like carbohydrate and already eat plenty of vegetables and fruit often find a low-fat, high-fiber pattern very filling, and their heart and blood vessels may benefit at the same time.
Protein has to be enough (both camps agree)
A common recommendation during weight loss is 1.6-2.2 g of protein per kilogram of body weight (about 0.8-1.0 g per pound), to protect lean mass and help with fullness.The often-cited Longland 2016 was a proof-of-principle trial: 40 young men, at an energy deficit of about 40%, doing strength training plus high-intensity intervals 6 days a week, for 4 weeks. The group eating 2.4 g/kg of protein a day gained a little more lean mass and lost a little more fat than the group eating 1.2 g/kg. It shows that high protein helps under these extreme conditions; it cannot be applied directly to ordinary people losing weight.This point holds whether you side with EBM or CIM, because where it acts is not where the energy went but whether what comes off is fat or muscle.
Chapter 5
Pick a way of eating you can keep
Once the debate has run its course, it comes down to three sentences for you.
There is no macronutrient split that is best for everyone. In DIETFITS, the 12-month head-to-head, the average difference between the two groups was < 1 kg. Picking the wrong diet camp is not why you are not losing weight.
Pick the one you can keep up for the long term. Short fasts, seven-day keto and seven-day intermittent-fasting plans can all make you 2 kg lighter by next week, much of it water. A good diet is one that keeps you 5 kg lighter a year from now. There is one test: will I still be eating this way 3 years from now? If the answer is no, switch.
Be wary of anyone who turns a complicated thing into a single switch. Carbs make you fat, insulin is the master switch, calories are a scam, keto cures everything: the more decisive these four lines sound, the more reason to pause.
However elegant a mechanism story is, it still has to pass . Anyone selling you a shortcut that gets around the physics of your body is not selling science.
There is no macronutrient split that is best for everyone. In DIETFITS, the 12-month head-to-head, the average difference between the two groups was < 1 kg. Picking the wrong diet camp is not why you are not losing weight.
Pick the one you can keep up for the long term. Short fasts, seven-day keto and seven-day intermittent-fasting plans can all make you 2 kg lighter by next week, much of it water. A good diet is one that keeps you 5 kg lighter a year from now. There is one test: will I still be eating this way 3 years from now? If the answer is no, switch.
Be wary of anyone who turns a complicated thing into a single switch. Carbs make you fat, insulin is the master switch, calories are a scam, keto cures everything: the more decisive these four lines sound, the more reason to pause.
However elegant a mechanism story is, it still has to pass . Anyone selling you a shortcut that gets around the physics of your body is not selling science.
Myth · The four loudest lines, one by one
Keto cures everything: no disease is cured by a macronutrient split. A ketogenic diet has genuine medical uses, such as epilepsy in children that drugs cannot control (drug-resistant epilepsy) and a few inborn metabolic disorders (for example glucose transporter 1 deficiency, GLUT1 deficiency), always under medical supervision; but those are specific indications, not a cure-all.Carbs make you fat: rice, oats, sweet potatoes, whole wheat and fruit are everyday staples in long-lived regions. What makes people fat is a long-running energy surplus, processed food and sitting, not the carbohydrate molecule. The same molecule inside different foods can have completely different effects: white sugar and sweet potato both contain carbohydrate, but the first needs almost no chewing and barely fills you up, while the second is bulky, has fiber and forces you to eat more slowly.
Insulin is the master switch of obesity: insulin is a hormone that influences where energy goes, not a fat-storing villain. There is a handy counterexample: protein also stimulates insulin, and some proteins (whey, for example) stimulate it no less than some carbohydrates do. If insulin really were that master switch, eating chicken breast would make people fat, and it plainly does not.
Calories are a scam: energy measurement has been validated again and again in randomized trials on metabolic wards. Everyday counting does carry error: package labels can be off by ± 20%, and your own estimates are often further off. But hard to measure precisely and does not exist are two different things; scales have error too, and nobody concludes that body weight is a scam.
The four lines share one shape: each squeezes a chain of many factors into a single switch, and a switch is the easiest thing to sell, because it promises that one flick and you are done. A real body has no such switch.
In practice · How much protein, and why protein
A common recommendation is 1.6-2.2 g of protein per kilogram of body weight during weight loss and 1.2-1.6 g/kg at maintenance. This holds whichever camp you side with.During weight loss, protein does three things:
It protects muscle. In an energy deficit the body breaks down both fat and muscle; when protein intake is high enough, muscle makes up a clearly smaller share of what is lost. Whether what comes off is fat or muscle decides what you look like once you are leaner, and affects whether your basal metabolism falls along with it.It keeps hunger away better. At the same calories, protein is clearly more filling than carbohydrate or fat, so you naturally eat a little less without forcing yourself, which loops back to the main thread of total energy.Digesting it costs energy. The thermic effect of food (TEF) is the energy the body spends digesting and absorbing a meal: roughly 20-30% for protein, 5-10% for carbohydrate and 0-3% for fat. So between a fifth and nearly a third of the energy in protein is taxed on the way through.
References · 7
- Hall, K. D., Bemis, T., Brychta, R., Chen, K. Y., Courville, A., Crayner, E. J., Goodwin, S., Guo, J., Howard, L., Knuth, N. D., Miller, B. V., Prado, C. M., Siervo, M., Skarulis, M. C., Walter, M., Walter, P. J., & Yannai, L. (2015). Calorie for calorie, dietary fat restriction results in more body fat loss than carbohydrate restriction in people with obesity. Cell Metabolism, 22(3), 427–436. 19 adults with obesity in a metabolic ward, each receiving both isocaloric diets for 6 days in random order after a 5-day baseline: cutting carbohydrate raised fat oxidation and lost 53 ± 6 g/day of body fat; cutting fat left fat oxidation unchanged but lost 89 ± 6 g/day (P = 0.002). Model simulations predicted the body minimises such differences over longer periods (abstract, PMID 26278052). 10.1016/j.cmet.2015.07.021
- Hall, K. D., Guo, J., Courville, A. B., Boring, J., Brychta, R., Chen, K. Y., Darcey, V., Forde, C. G., Gharib, A. M., Gallagher, I., Howard, R., Joseph, P. V., Milley, L., Ouwerkerk, R., Raisinger, K., Rozga, I., Schick, A., Stagliano, M., Torres, S., … Chung, S. T. (2021). Effect of a plant-based, low-fat diet versus an animal-based, ketogenic diet on ad libitum energy intake. Nature Medicine, 27(2), 344–353. 20 adults randomised to ad libitum minimally processed plant-based low-fat (10.3% fat, 75.2% carbohydrate) or animal-based ketogenic (75.8% fat, 10.0% carbohydrate) diets for 2 weeks each, inpatient crossover: the low-fat diet led to 689 ± 73 kcal/day less intake over 2 weeks and 544 ± 68 less in the final week, contrary to the carbohydrate-insulin model's prediction; one participant withdrew with hypoglycaemia on the low-carbohydrate diet (abstract, PMID 33479499). 10.1038/s41591-020-01209-1
- Gardner, C. D., Trepanowski, J. F., Del Gobbo, L. C., Hauser, M. E., Rigdon, J., Ioannidis, J. P. A., et al. (2018). Effect of low-fat vs low-carbohydrate diet on 12-month weight loss in overweight adults and the association with genotype pattern or insulin secretion: the DIETFITS randomized clinical trial. JAMA, 319(7), 667-679. 609 adults, 12 months; no significant weight-loss difference between healthy low-fat and healthy low-carbohydrate diets, and neither genotype pattern nor baseline insulin secretion predicted which diet worked better. 10.1001/jama.2018.0245
- Ludwig, D. S., & Ebbeling, C. B. (2018). The carbohydrate-insulin model of obesity: beyond 'calories in, calories out'. JAMA Internal Medicine, 178(8), 1098-1103. Narrative review setting out the model: processed, high-glycemic-load carbohydrates produce hormonal changes that promote calorie deposition in fat tissue, increase hunger and lower energy expenditure. The authors cite mechanistic, genetic and animal evidence and behavioural-trial meta-analyses (greater weight loss with reduced-glycemic-load than low-fat diets, with poor long-term compliance), and state that feeding studies have lacked the rigor and duration to test the model; pending definitive studies they offer low-glycemic-load principles as a practical alternative (abstract, PMID 29971406). 10.1001/jamainternmed.2018.2933
- Hall, K. D., Ayuketah, A., Brychta, R., Cai, H., Cassimatis, T., Chen, K. Y., et al. (2019). Ultra-processed diets cause excess calorie intake and weight gain: an inpatient randomized controlled trial of ad libitum food intake. Cell Metabolism, 30(1), 67-77.e3. 20 inpatients, 2 weeks per diet, crossover. Meal eating rate was greater on the ultra-processed diet by 17 +/- 1 kcal/min (7.4 +/- 0.9 g/min), p < 0.0001 - that is the between-diet difference; ratings of pleasantness and familiarity did not differ (full text, PMC7946062). Diets were matched for presented calories, energy density including beverages (1.024 vs 1.028 kcal/g), macronutrients, sugar, sodium and fiber (21.3 vs 20.7 g/1000 kcal, partly via fiber supplements added to ultra-processed meals); non-beverage energy density was 1.957 vs 1.057 kcal/g (~85% higher), which the authors say likely contributed. Intake was 508 +/- 106 kcal/day greater on the ultra-processed diet (full text, Table 1 and Results). 10.1016/j.cmet.2019.05.008
- Longland, T. M., Oikawa, S. Y., Mitchell, C. J., Devries, M. C., & Phillips, S. M. (2016). Higher compared with lower dietary protein during an energy deficit combined with intense exercise promotes greater lean mass gain and fat mass loss: A randomized trial. American Journal of Clinical Nutrition, 103(3), 738–746. 40 young men (20 per group), 4 weeks at a ~40% energy deficit with resistance training plus high-intensity intervals 6 days a week; 2.4 vs 1.2 g protein/kg/day. Lean body mass +1.2 ± 1.0 kg vs +0.1 ± 1.0 kg; fat mass -4.8 vs -3.5 kg; exercise performance improved similarly in both groups. The authors call it a proof-of-principle trial (abstract, PMID 26817506). 10.3945/ajcn.115.119339
- Hall, K. D. (2017). A review of the carbohydrate-insulin model of obesity. European Journal of Clinical Nutrition, 71(3), 323-326. Several logical consequences of the carbohydrate-insulin model were tested in a pair of carefully controlled inpatient feeding studies and failed; the author concludes that important aspects of the model have been experimentally falsified and the model is too simplistic. 10.1038/ejcn.2016.260