Tuesday, June 05, 2012

Insulin and the Rewards of overfeeding

I've been tempted away from the electron transport chain, origins of life and the suspected paleo prompt nuclear criticality on Mars by a neat paper from Liz.

This is what they did in the study: Over-fed 9 slim, young, fit, healthy blokes for two weeks while limiting their exercise. They establish an energy surplus of about 2000kcal/d. Here we go:

Needless to say, they gained weight. Here we go again, Table 3:

A total of about 1.5kg in two weeks.

Now, all you have to do is to go and ask any cutting edge, state of the art obesity researcher and you can be told that hyperinsulinaemia is a consequence of obesity, not a cause, and that carbohydrates are the worlds greatest slimming aid because insulin is a satiety hormone and, oh, did I fall asleep there?????? Sorry.

Back to the paper. The overfeeding was with utter crap

and produced a rise in fasting insulin from 35pmol/l to 74pmol/l in 3 days.

Personally, I found this quite amusing.

Aside: What would have happened if they had overfed with lard? That's another post, it has been done, rather badly, in Schwartz's lab using dogs. They gained weight. End aside.

By 3 days the fat mass had increase by an average of 100g. My, that is potent fat! Here's the table, day nought vs day three is the place to look:

We could leave it at that and just go away scratching our heads about what goes on in the minds of obesity researchers, but it does get quite interesting. Obviously, as obesity progressed the level of insulin should increase if insulin resistance is caused by fat mass. It doesn't. Fasting insulin falls progressively after the initial spike, and in a fairly linear manner, through days 3, 7 and 14, while weight (and especially visceral fat) actually increases over this time period.

So the idea that fasting insulin rises as a consequence of rising fat mass is, well, you know what it is.

Ah, but if insulin stores fat, why should the level of insulin fall progressively during a sustained hypercaloric eating episode? Surely you must need insulin to store those extra calories? In fact, as insulin levels fall, so does the rate of fat storage. The chaps gained, from Table 3, 1kg of fat mass in the first week and only 0.5kg of fat in the second week... Oh, I guess this must be because the subjects either (a) sneaked off to the gym in the second week or (b) flushed their Snicker Bars down the loo in the second week, without passing them through their gastro intestinal tract first (good idea!) or (c) got bored with Snickers and stopped finding them rewarding. And of course they disconnected their Actiheart monitors at the gym.

Otherwise how you can eat 2000kcal over your energy expenditure, equivalent to nearly 200g of fat gain per day, and gain a kilo of fat in the first week, then continue to eat an excess 2000kcal/d for a second week and only gain half a kilo of fat? Calories in, calories out, you know the rules. Hmmm, in the second week there are 14,000 excess calories-in, 5,000 stored, very interesting.

We all know the obese lie about calories. It seems probable that so too must experimental subjects, in direct proportion to the duration of their over eating! Now we know. Bit of a milestone paper this one.

So what is really going on? What appears to be happening is the insulin system working exactly as it should do. Insulin resistance protects cells from caloric excess, when forced in to the body by a study protocol. Think of it in these terms, with thanks to Dr Guyenet from back when I used to read him. The vast majority of free radicals will be generated at complex I.

The mitochondria say they have too many calories. It's easy for mitochondria to refuse calories from glucose by using insulin resistance, working at the whole cell level. In the presence of massive oral doses of glucose this must elevate insulin to maintain normoglycaemia. The elevated insulin diverts calories from dietary fat in to adipocytes, away from muscle cells. And inhibits lipolysis at the same time, look at the FFA levels in Table 3 on days 7 and 14, waaaay down from pre and post study values. I wonder why they didn't measure FFAs on day 3? So insulin goes up to maintain normal blood sugar levels, overcomes insulin resistance to run cells on a reasonable amount of glucose and shuts down FFA release to counterbalance its action in facilitating the entry of glucose in to cells.

Core to this is (a) there is no hyperglycaemia, insulin still successfully controls glucose flux and (b) insulin inhibits lipolysis. So you store fat. These subjects are both young and healthy. They do not have insulin resistant adipocytes, mitochondrial damage or a fatty liver. The system works as it should.

As time goes by fasting insulin levels fall and weight gain slows. Calorie intake doesn't drop. The only plausible explanation is that the subjects generate more heat and radiate that heat during the second week of the study. Total energy expenditure was estimated using the Actiheart device. You have to wonder how well its computer algorithms coped with the massive overfeeding. It looks like the weakest link in the protocol, assuming the the subjects really ate their muffins and Snicker bars. The device is supposed to be very good but where else did those calories-in go? So let's consider uncoupling proteins. These decrease the inner mitochondrial membrane potential and so decrease free radical production, which decreases insulin resistance. You can afford to allow more glucose in to cells when the UCPs are in place and working. More posts on this to come when we get back to the electron transport chain and free radicals. Free radical generation at complex I is VERY dependent on the inner mitochondrial membrane voltage.

Unfortunately the clamp studies were only performed on days 0 and 14, so all we can say about the insulin sensitivity by clamp (gold standard) is that it was worse on day 14 than day 0. Who knows what the results would have been on days 3 and 7? This protocol is understandable as clamps are a pain to do, but there is no way of ascertaining what glucose disposal per unit insulin was through the body of the study.

So forced weight gain fits quite nicely with the role of insulin in fat storage. It's a nice study because it looked at insulin before fat gain had occurred, and kept on looking too. But how much does forcing people to overeat tell us about "accidental" weight gain in people who have spent good money on some slimming plan to lose weight temporarily with enormous difficulty? Under these circumstances calories are offered to cells at very reasonable levels, mitochondria are already dysfunctional and signal (using free radicals) excess calorie warnings (free radical leakage) and so induce insulin resistance inappropriately. So calories, especially those from dietary fat, get diverted to adipocytes through the subsequent hyperinsulinaemia. And are kept there due to fasting hyperinsulinaemia.

The situations are quite different, I'm not sure that overfeeding healthy subjects tells us too much about accidental obesity, except that they both seem to work through insulin and mitochondria. But this study does tell us a great deal about the idea of reward.

This study is the ultimate affirmation of the Reward Hypothesis of obesity:

If you reward people with enough Danish Krone for over eating, they gain weight.



LeonRover said...

Brilliant conclusion, Petro.

The "F" in FR hypothesis stands for FINANCIAL.

Only Billy Connolly has better one-liners.

I chortled.


"I can resist everything except temptation".
Verbal in your (& my) case.

Howard said...

"The only plausible explanation is that the subjects generate more heat and radiate that heat during the second week of the study."

Uh, no. Unless you actually weight and measure *everything* that comes out (as well as everything that goes in), you can't know that for sure. For instance, how many calories passed through undigested? If you haven't measured that, you are guessing. And that guess is GUARANTEED to precisely reflect your preconceived ideas, much like the guesswork that permeates most of what passes for "science" in the field of nutrition.

It so happens that I think you're right here, but the only way to know for sure is to measure *everything*. Which essentially nobody ever does.

Nigel Kinbrum said...

"The only plausible explanation is that the subjects generate more heat and radiate that heat during the second week of the study."

Uh, no. kcals burned due to NEAT/SPA can change, even if kcals due to AT/PA = zero.

Also, overall body Insulin Sensitivity (IS) is a function of adipose IS, muscular IS and hepatic IS. Muscular & hepatic IS can change without a significant change in adipose mass/IS.

Also, what Howard said.

john said...



...Why would high fat rats (81%f, 19%p), with IR adipocytes, have similar pad weight as mixed-diet group? There is a mention of an association between adipocyte IR and fat storage, but the latter causing the former is an obvious explanation (I think).

Susan M. said...

I have been reading your posts for a couple of years now, and while I have absolutely no training in the sciences beyond H.S. biology, I can still get the "gist" of what you are saying, some of the time. This is like going back to school for me and I am loving it!

Thanks for your great posts Peter.
I sure hope you can do posts more often now.

Susan M.

Peter said...

Howard, true, but they weren't adding 2000kcal of wheat straw or lignin to the diet. It was Snickers all the way.

They didn't actually measure much that mattered, but this is obesity research.

Leon, ah, The Big Yin...


Peter said...

John, I had the paper on my hard drive already but I think I gave up on it when I twigged they were comparing massive sucrose intake with massive protein, with massive hydrogenated vegetable oil, when you couldn't tell if the hydrogenation was partial or complete... Is there a fundamental here I'm missing? I notice the adipocytes are very insulin resistant on hydrogenated vegetable oil but, as you point out, this should be expected on zero carb intake... I probably need to think about this one a bit more!


Peter said...


Actiheart would probably pick up NEAT rather well.


john said...


Those strange diets do make things tricky. I was thinking that the rats gain the fat before developing adipocyte IR, but I think that even long term, ketogenic animals still have around average fat storage.

Does the adipocyte IR encourage a decrease in adipose, or does it just prevent more? What is happening in the adipocytes without glucose/nutrient uptake?

Peter said...

Hmmm, I'm never expecting ketogenic diets to lead to zero fat storage, eventually some sort of balance has to be reached. I remember the free choice mice in the control group (wild type) for brown adipose tissue KO mice ate a ton of fat (which was pretty well Food, coconut oil plus German lard, as opposed to CIAB). They ended up somewhat heavier than the CIAB mice but non-significantly MORE insulin sensitive. Bearing in mind the sensitivity to insulin was peforemd in the fasted state which probably makes the CIAB mice more insulin resistant than they would be when fed, without affecting the mice which chose to eat LC without any obligation...

From here:


I had a bit of a think about it here


What is an ideal weight?????


George Henderson said...

To do prove that the first law of thermodynamics does in fact apply in these experiments, and that calories in do in fact equal calories out to the last joule, one would have to measure the caloric value of all excreta, as well as heat loss, and/or heat maintenance (in which case all environmental temperature fluctuations would also need tracking).
My guess is that, if you are honest, you can never get it to balance exactly, however carefully you measure.
It's not proof of an anomoly in the rules of the universe, it's just bloody hard to measure exactly.

js290 said...

Energy input and output are coupled in the mathematical sense. You can change one without affecting the other.

ItsTheWooo said...

Peter, this is trivially easy to explain in the context of FR hypothesis.

Explanation of findings:

We know this because BRAINZ RESISTED DISEASE.


Maybe an extra cup of mountain dew was the trick to make it sufficiently rewarding?

Or, another explanation: they were puking up everything after a few days. Basically, the researchers observed bulimia from forced obesity. Certainly I know prolonged overfeeding of relatively trival overconsumption produces an uncomfortable feeling of nausea (it's the wonderful sound of fat lingering in the blood that can't go anywhere... don't even SHOW ME a calorie or else I"ll upchuck. Food stimulus, in this situation, produces a similar reaction as being exposed to the stimulus of vomit. Instant nausea.)

I could imagine if I was forced to gain SO MUCH WEIGHT in such a short period of time, even I, a weight reduced fatass, would probably end up losing my lunch a few times.

When inducting a ketogenic diet from a position of years and years of hyperinsulinemia/ hypertrophic adipocytes, I pretty much could NOT EAT A MEAL without feeling like I would vomit. This sort of happens when FFA are just flooding your blood stream at any given time because you are cartoonishly fat but your blood insulin is relatively normal now in ketosis.
I imagine these people felt something like that, given they are obesity resistant and being forced to over eat for days. They probably felt like I did 10 years ago, whale sized, eating 20 carbs per day. Eating made me more nauseated, unless the food was a high carb item, then it went away totally (insulin to fat tissuez: OH HI, did you lose your way little FFA? Come back inside my dearz...)

ItsTheWooo said...

This just in at WHS:


Highly technical studies involving emotionally abused tortured rodents and no humans ever lead us to conclude unanimously this ground breaking knowledge:
Over the past few years, people have been getting fatter, and they also have been eating more energy.

:O :O :O :O :O!!!

The only obvious conclusion to be drawn here is that eating more calories has caused obesity. You can all go home now.

Where's my lil'jon-esque chalice, I need to celebrate this triumph of SCIENZ right!

Get krunk! Get my doctorate of the neurobiology of obesity on!

Nigel Kinbrum said...

Peter said...
Actiheart would probably pick up NEAT rather well."
I looked at the Actiheart site. The device is chest-worn, so its accelerometers won't detect arm & leg movements e.g. fidgeting, or activities where the torso doesn't "jiggle about" e.g. cycling.

john said...


Ahh yes, I do remember that. Body fat tends to not be too consistent across studies anyway I suppose. Protein is perhaps even the best associated (negatively) macro.


Despite Stephan's footnote acknowledgement of an alternative hypothesis, one has to wonder about his article naming, "Calories Still Matter." It looks like a strategy to set up a strawman and hence get a bunch of unintelligent commenters exclaiming, "Yea! Those low carb idiots think you can eat 10,000 calories per day!"

Peter said...


You been doin' the ondansetron again!


Elliot said...


"I looked at the Actiheart site. The device is chest-worn, so its accelerometers won't detect arm & leg movements e.g. fidgeting, or activities where the torso doesn't "jiggle about" e.g. cycling."

The Actiheart also measures heart rate, so I tend to doubt your theory that the subjects decided to take up cycling in the second week.

Nigel Kinbrum said...

@Elliot: I merely pointed out (for completeness) that Actihearts under-report CO when cycling. I didn't theorise that the subjects took up cycling. Thank you for the straw man. I shall add it to my collection!

There's still the issue of Actihearts not reporting CO due to constant arm & leg fidgeting, which doesn't significantly raise HR.

Elliot said...

Nigel, there's a difference between a straw man and a joke!

Stephan Guyenet said...

You are utterly confused Peter. These people only gained 0.8 kg of fat mass. Over the course of the study, they went from lean to slightly less lean.

If you look at studies where overfeeding produced greater fat gains, you see a consistent increase in fasting insulin that corresponds with fat gain, just as those silly obesity researchers would predict:


In animal models of diet-induced obesity (rodents and dogs), blocking the hyperinsulinemia has no effect on the rate of fat gain.

The carbohydrate-insulin-obesity hypothesis is dead and buried, and all that remains are rearguard attempts to salvage it using increasingly complex theories. Just let go of the cognitive dissonance man.

Andrew S said...

Peter says: calories, especially those from dietary fat, get diverted to adipocytes through the subsequent hyperinsulinaemia. And are kept there due to fasting hyperinsulinaemia.

Stephan says: blocking the hyperinsulinemia has no effect on the rate of fat gain

Are these the same 'hyperinsulinemia'? Does Peter mean 'elevated insulin, ie insulin above baseline', where Stephan means 'excessively high insulin'?

ie, all it takes is a little insulin to store excess calories; it doesn't take excessive insulin. And excessive insulin is a sign that tissues are IR, ie that the body is desperately trying to flush out blood glucose.

Is Stephan arguing that insulin isn't required to store calories in adipocytes? Or is Peter arguing that excessive insulin is required?

Sam Knox said...

"The carbohydrate-insulin-obesity hypothesis is dead and buried, and all that remains are rearguard attempts to salvage it using increasingly complex theories."

Only in your mind, Stephan.

nothing91 said...

Stephan keeping it classy as usual.

Though, with his most recent bombshell report that (prepare yourselves for this one, folks) the obesity epidemic correlates with increased calorie consumption (!!!), I can certainly see why he's a little conceited.

Bill said...

@ Andrew S,
under normal circumstances, mildly elevated insulin levels completely suppress lipolysis.

Bill said...

Peter, what makes you say that CIAB-fed mice would be more insulin sensitive in the fasted state relative to the fed state? From my experience, 18 hour fasted mice are more insulin sensitive than 6 hour fasted mice (this becomes a very tricky experiment in "fed" mice).

Bill said...

@ Howard,
"Which essentially nobody ever does."

indeed, and this is very unfortunate.

ItsTheWooo said...


"The carbohydrate-insulin-obesity hypothesis is dead and buried, and all that remains are rearguard attempts to salvage it using increasingly complex theories. Just let go of the cognitive dissonance man."

Right, spot on! It's dead and burried, and we know absolutely know FR is correct, because we have finally cured obesity. There are so many people out ther ewho have lost dozens of pounds via controlling food reward, after all... and NONE out there who have lost dozens of pounds by reducing their insulin requirements via calorie/carb restriction. So smart!

Look, there is a linear correlation between being a glutton, errr food addict, and having your ass expand out of chairs. We know this, because you posted the two parallel graphs and they match. Gluttony up, fat assery up. What more evidence do you need that the gluttony -> fat assery? Clearly, being a food addict and eating more, caused the ass expansion problem.

Here is the cause of obesity succintly described: food is excessively finding its way into our mouths in spite of the fact we fat slobs are NEVER HUNGRY, we don't feel physical hunger, and we don't get hypoglycemia...we just eat compulsively, like a heroin addict shoving a dirty needle in his arm, or a drunk with shaky hands guzzling $2 vodka in the morning. We have no hunger, no hypoglycemia, but we shove down and on with the food because nestle crackle bars are irresistable to our primitive tribal brains, and we are hopelessly addicted. As for those who reisst becoming obese? Nervous system sophisticates such as yourself don't have this weakness/flaw.

Again: my body blew up like a cartoon character circa puberty (from a prior steady state of mild chubbiness) because of the fact I developed a food addiction at that time. The fact that puberty leads to enhanced growth and can trigger IR, similar to pregnancy or stress, has nothing to do with it.

I was completely misinformed about the shaking/crying 2 hrs after my starchy breakfast, and being away from food for 2 hrs leading to impulsive hunger and irritability. I made it all up, it never happened. The time I guzzled dextrose and my sugar went from 140 to 45 within an hour also was made up. Imaginary. The countless times I've taken my sugar after eating 'Normal things" and it was 120, only to be about 60 a few minutes later... yea, not real, can't happen, hypoglycemia is a myth. Besides, 60 isn't hypoglycemia and going from 120 -> 60 totally won't make you feel hungry or shaky.

When I went on a ketogenic diet at 20 and all of these problems went away magically it was because I lowered the reward value of my diet. The effortless loss of dozens of pounds of flab was because my diet was not rewarding anymore.

When I eat the most delicious glucose free food, and easily feel full/energetic/stay skinny, it's becuase it is not high reward even though it tastes ridiculously yummy. When I eat stale crackers and feel shaky/hungry, it's becuase stale crackers and dry pasta are high reward.

Thanks for all your work Dr Guyenet, the attentive way you observe real obesity and help us control our disorder will go down in history. That rotten Taubes / meddlesome kids are good for nothing bad science promoters who have never helped anyone understand and control severe health problems which have plagued them for years. Thanks to the FR hypothesis, I have maintained a weight loss of 160 pounds for over 10 years and controlled nebulous health problems... but the insulin / endocrine / metabolic hypothesis of obesity? Pure bullshit! All weight / eating behavior except for rare extreme examples are explained by food reward. End discussion, deal with it.

ItsTheWooo said...

Green tea 24/7 helps me lose/maintain weight because it reduces food reward.

Magnesium helps me lose/maintain weight because it reduces food reward.

Vinegar dips and sauces added to plainer foods help me lose/maintain weight because it reduces food reward. So it is less novel/rewarding to eat meat with sugar tasting bbq sauce or ketchup or mustard , than it is to eat the meat plainly.

Fiber rich bread products which are low starch help me lose/maintain weight compared to NO bread products, because it reduces food reward.

Fresh berries and cream and chocolate with butter and granular erythritol is less rewarding than cold pasta, and I know this because I don't eat anymore from the first meal, but after the second meal, I am quivering and ravenous at 1.5 hrs.

Broiled steak with fat and broccoli sauteed in butter/garlic with salt is less rewarding than white rice, and again I know this because the first meal makes me feel awesome and full and energetic and I move around easily, but the second meal turns me into a food vacuum and will NEVER make me feel as well as the first meal.

Food reward is SO OBVIOUS, I don't know how people still persist with this insulin / endocrine / metabolic hypothesis. DEAD UND BURRIED~.

I just ate a massive bowl of greek yogurt, a few fresh strawberries, cream, coconut oil, crushed salted roasted macadamia nuts, raw almonds, flax meal, with as much sucralose and vanilla extract as desired. I feel full and energetic. If I ate an equal calorie amount of processed fat free goop "yogurt" loaded with sugar, I would feel like death. Clearly the meal I just ate was not rewarding, in spite of the fact it tastes one million times better.

ItsTheWooo said...

Sugar is less rewarding than pasta because when I eat sugar, I push away the plate (NO APPETITE; I AM GOD!!) and laugh my head off and ramble as if I took booze instead of candy... BUT if I eat pasta I just feel lethargic right away, with no energy burst. Clearly the sugar is less rewarding, as it produces less food consumption, in spite of the fact the sugar is producing CLEAR CUT signs of high dopamine/endorphin. The fact thin children like sugar a lot and fat kids like food a lot is more evidence!!!

Peter said...


"Peter, what makes you say that CIAB-fed mice would be more insulin sensitive in the fasted state relative to the fed state? From my experience, 18 hour fasted mice are more insulin sensitive than 6 hour fasted mice (this becomes a very tricky experiment in "fed" mice)."

I'm not sure where I would have said this, if you would like to point it out I'll see if it's a typo. Personally I would expect mice to be persistently insulin resistant under starvation as they would only have limited stores of glycogen. That is certainly my impression from reading people who have examine adipocytes from fasted rodents, rats I think. Hard to see how insulin resistance would decrease with time unless proteolysis/gluconeogenesis was providing the glucose in a rather generous manner.

Once you start trying to look at insulin sensitivity under fed conditions you are then in to very interesting questions indeed, paticularly superoxide generation and insulin resistance as an antioxidant defense mechanism.

Very deep questions.


Nigel Kinbrum said...

Elliot said...
"Nigel, there's a difference between a straw man and a joke!"
Dammit man, use a winking smiley next time so that I can tell the difference! :-D

Nigel Kinbrum said...


Insulin increases the amount of glucose & FFAs entering fat cells, muscle cells & the liver.

Insulin decreases the amount of glycerol & FFAs exiting fat cells & the amount of glucose exiting the liver.

Hyperinsulinaemia (which can produce sedation) results when one or more of the following tissues loses insulin sensitivity:- fat cells, muscle cells & the liver.

So, why do people keep saying that hyperinsulinaemia locks nutrients away in fat cells only, thus robbing other cells of nutrients, thus causing lethargy?

The relative insulin sensitivity of tissues determines the relative partitioning of nutrients into those tissues.

When tissues lose sensitivity to insulin, blood glucose control becomes impaired. This results in a roller-coaster blood glucose level after eating high-glycaemic carbohydrates. A rapidly-falling blood glucose level can cause ravenous hunger even in the absence of hypoglycaemia. I have experienced this during medically-monitored tests (OGTTs & an insulin shock test).

Low-carb/ketogenic diets don't result in a roller-coaster blood glucose level and therefore don't cause ravenous hunger. Simples!

Over-eating due to ravenous hunger is NOT gluttony, just as under-moving due to sedation is NOT sloth.

THIS is gluttony.

Andrew S said...

@ Bill,

So this quote: "In animal models of diet-induced obesity (rodents and dogs), blocking the hyperinsulinemia has no effect on the rate of fat gain." becomes completely irrelevant. Hyperinsulinemia? Who cares? A bit of insulin to stop lipolysis and promote fat storage is all it takes to get on the path to obesity.

Or is Stephan talking about blocking insulin altogether? Rodent and human models of total insulin removal (aka T1 diabetes) tend to be thin and dead.

Seems to me that Stephan is addressing the strawman that it takes hyperinsulinemia to produce obesity.

PhilT said...

I had a go with the simulator at http://bwsimulator.niddk.nih.gov/ which looks at the dynamic simulation of body weight. As near as I could get to the test protocol it showed :-

1. Total weight increasing from 71.3 to 78.2kg over 14 days, compared to 72.9 kg in the paper referenced here.

2. Fat mass increasing by 2.7kg rather than the reported 1.5 kg.

3. Fat Free mass increasing by 4.1kg rather than the reported decline in FFM of 0.3 kg.

Hall's model is validated against the Minnesota study (including refeeding effects). It does highlight the loss of FFM reported, but of course doesn't help understand why. Less activity leading to muscle waste, perhaps. That would lead to less RMR which makes the lack of fat accumulation even more imponderable from the point of view of the simple energy balance.