Tuesday, September 02, 2008

Physiological insulin resistance; Alzheimers

I'd just like to run through the effects of saturated fat vs unsaturated fat on the development of Alzheimers disease in a mouse model genetically engineered to be AD prone. Here is the opening sentence of the paper (thanks to Gary for the pdf):

"In Western Society, diets have gradually changed since World War II, with an increase in total caloric intake, saturated fat, and hydrogenated fat, leading to a decrease in the healthier, unsaturated fats."

This statement is frankly wrong. Good start. BTW There's accurate information on changes in fat intake in the American diet here on Stephan's blog.

Next is the methods. They added 2% by weight of chemical grade cholesterol to the diet of the saturated fat group only, none to the PUFA group. This is called failing to control your variables. If I was scrutineering this paper I would have thrown it out at this point. BTW I roughly translated this amount of cholesterol to mean >50 eggs per day for an adult human, but not as eggs, as chemical grade cholesterol. Anyone fancy trying this?

If you look through the table you will see that just under 24% of the "food" by weight was Alphacel. I think this is cotton wool, near enough.

So the "food" components of the diet are 27% by weight fat, 26% by weight starch and 32% by weight, err, sugar (dextrinised starch=sugar). Yummie. So any mouse eating any food must eat LOTS of sugar.

Why is this a problem if you mix it with saturated fat but not with soya oil? Because saturated fat induces insulin resistance.

Why should this be? Because palmitic acid is the primary NEFA released from human adipose tissue during fasting. Think of palmitic as a signal molecule to tell the muscles that inhibition of glucose uptake is needed and to tell the liver that increased gluconeogenesis is required because there is no food coming in.

Because linoleic acid comprises only a small portion of adipose tissue in humans there is no reason why an increase in this fatty acid should signal the need for physiological insulin resistance. Over an evolutionary time scale elevated linoleic acid levels mean nothing except perhaps you found a few nuts. This is not starvation and insulin resistance doesn't need to happen. Other saturated fats seem to do the same as palmitic, certainly coconut oil, fully hydrogenated, seems to do this in these mice. Lipoprotein lipase certainly spills some of the NEFA it releases from chylomicrons in to the general circulation, especially in muscle beds.

So these poor mice are being made insulin resistant, while being fed sugar. Each mouthful combines saturated fat and sugar. Non stop. Whenever they eat anything. Almost as much sugar as someone on the upper end of the sugar intake in the SAD, but with more fat (27% by weight is 54% of calories). Combining insulin resistance with sugar when you are genetically engineered to get Alzheimers does not embody bad luck. Dementia is guaranteed. For a human sugar junkie there is more genetic chance involved and you might just get lucky enough to have a heart attack sooner rather than dementia later...



Dave said...

Interesting. If SFA induce insulin resistance, that might provide the (weak) epidemiological association between dietary saturated fat and other diseases. Of course you'd also need lots of carbs in the diet, so when you observe a low-carb population like the Masai warriors or traditional Inuit, a "paradox" appears. Similarly, it sounds like the Kitavans have relatively little SFA in their diets. Despite the relatively high carb intake, perhaps this allows them to stay insulin sensitive. Do the Kitavans get much fructose compared to the SAD? I would guess not.

And we haven't even considered the effects of wheat . . .

Drs. Cynthia and David said...

Fascinating. I have been wondering whether insulin resistance is the pathological condition it is made out to be, and coming to the conclusion that it cannot be pathological per se, but must be intended to serve a physiological purpose, and only becomes pathological under conditions of carbohydrate (and food in general?) excess. I wonder if any of these researchers ever bother to read your blog (or those of other educated and thinking individuals). I assume not, because after all, they are the experts, writing grants and reading papers and thinking deeply about these issues all day (hah!).

I wonder what happens in the case of endurance events. We know that trained individuals increase both blood glucose and free fatty acids during endurance exercise (and break down muscle protein to provide amino acids for fuel as well). In this situation, you don't want your muscle cells to remain insulin resistant do you? Do they remain sufficiently sensitive to insulin to take in glucose (at a slower but still useful rate)? and if so, does this require exogenous glucose to trigger insulin release? or is this where the non-insulin mediated glucose uptake comes in handy (i.e., you don't really need insulin anyway)? The answer may be out there (too many papers to read!)

Dave- are you entered in the Stevens Creek 50K this weekend? If so, I'd like to say hi.


Stephan Guyenet said...

It never ceases to amaze me how very intelligent people can get the basic facts so so wrong. Saturated fat and animal fat consumption have decreased in the US, and the "healthy" n-6 poyunsaturates are being eaten in large amounts for the first time in human history! These data are widely available to anyone with an internet connection.

Dave, the Kitavans actually eat a lot of saturated fat (17% of calories), despite having a low-fat diet overall. That's because most of their fat comes from coconuts. So saturated fat (from coconuts) can't cause significant insulin resistance in the long term, at least if the lifestyle is otherwise healthy. Kitavans also eat a fair amount of fructose because fruit is a major source of calories for them.

Peter, there are several things that always get me about this type of study. First, like you said they don't control variables. There are so many rodent studies that proclaim fat causes obesity, and then when you look at the diet it's actually fat + sugar.

The second thing is that high-fat diets actually do cause obesity in mice, even when the fat comes mostly from lard. So they aren't a good model for studying diet composition and health in humans right off the bat, because high-fat diets more often cause weight loss in humans.

These things are all ignored, because they have an animal model that agrees with their politically correct nutrition dogma. It takes a willful ignorance of some basic historic, biological and medical facts.

By the way Peter, I enjoyed the last two posts.

Charles R. said...

Between Stephan and Peter, I'm getting an amazing education here. And to me, this seems like a good example of a real acceleration of knowledge through the internet as a result of the indirect but immediate collaboration that is made possible.

But yes, insulin resistance from saturated fat. That's another unexpected addition to the puzzle. So as the drs. above state, it's not the ultimate evil that we have thought it was, and only becomes devilish when combined with large amounts of carbohydrates?

But then we have Stephen's demonstration that saturated fat consumption has gone down, while (presumably) insulin resistance has gone up. So is the increase in carbohydrates, particularly wheat as Stephan notes, the more significant factor?

I'd also suggest that we maybe are not seeing just one syndrome here, but a number of them that look similar in effect.

donny said...

Articles like this; http://www.webmd.com/heart-disease/heart-failure/news/20030331/high-fat-diet-ups-dangerous-hidden-fat

Containing mantras like this;

"Saturated fat is the worst. This is hard fat like lard, butter, or fat on meats," says Stewart. People who said that more than 30% of their calories came from fat were most likely to have measurable visceral fat."

Oy. I was looking for something I read about linoleic acid being preferentially stored in visceral tissue, when not simply oxidized, but it turned out to be a rat study, and what I could find on humans had it spread more evenly around the body.

I thought this was interesting, though. Rats put through a form of intermittent fasting tended to oxidize a greater amount of polyunsaturated fatty acids, and ended up with less polyunsaturates in their tissues. I wonder if any of the benefits of fasting come from the cleansing of polyunsaturates from the body? Detox?


Anonymous said...

Because saturated fat induces insulin resistance.


Would you mind expounding on this a bit? Is this a de facto argument for avoiding saturated fats?

Brad Reid said...


And, in addition to Varangy's request, would you comment on the supposed postprandial arterial constriction after a high-fat meal? Is this accurate? Does it actually reflect, more accurately, a high-fat, high-carbohydrate meal? So often, it seems, what is being passed along as evidence against high-fat/sat-fat as bad diet, is actually high-fat AND high-carb. Bad science, I guess.

Cheers! Brad

Thackray said...

Hi Peter,

You’ve probably seen this but I’ll throw it in here for those who haven’t.

Here is a test tube study (full text available) about high glucose and saturated fatty acids:


The authors state that this supports the “glucolipotoxicity hypothesis” which I gather from looking at the references is (in 2003) replacing the (just) “lipotoxicity hypothesis”. This is good – right??

From looking at the titles in the reference section I see how saturated fat could get a bad name in some communities!

Philip Thackray

Dave said...

Hi Stephan. I really doubt the Kitavans get anywhere near the amount of fructose as many Westerners. One super-sized soda contains about 100g of sugar, probably all in the form of high-fructose corn syrup. So that's a minimum of 55g of fructose, which is about the same amount as in a kilogram of pineapple. I doubt the Kitavans are eating a kilogram of pineapple each day, but you probably don't have to look very hard to find Americans knocking back multiple super-sized sodas on a daily basis.

Many physiological responses are very nonlinear, showing a characteristic "sigmoid" response. Insulin dose-response as a function of plasma glucose is a good example, exhibiting a sensitive range, and flattening out on either side. I would infer that this means the body can tolerate carbohydrate intake in a certain range, above which the pancreas fails to keep up.

The same might (and I emphasize the "might") be true of insulin resistance induced by saturated fat. In a "normal" range of carbohydrate/SFA intake, things operate well. Push outside of this range, and the regulatory mechanisms struggle to keep up. Note also that high fructose intake (e.g. 3 or 4 of those super-sized sodas) is essentially equivalent to high SFA intake, because once liver glycogen is topped off, the rest of that fructose is rapidly converted to palmitate. To really compare with the Kitavans, we'd have to consider the total SFA contribution from all sources, as well as any other effects from diet and lifestyle.

Hi Cynthia. Unfortunately, I don't run, so I won't be at the Stevens Creek 50K. I take it you live in the Bay Area, though, so we can probably find a way to bump into each other :-) I'm currently planning to attend the Weston Price Conference in Burlingame, which might be a good opportunity.

Mr Secret said...

Eczema from either 38% full fat cream or 85% cocoa chocklade. Wich one of these two sources gives me extreme itching and eczema on the legs? im also following the Jan K "diet".

emil henry said...

Speaking of insulin: how does cheese fare? I *seem* to get a little tired in the evening from eating moldy goat cheese, which might be due to histamines. Caseine is obviously an issue, as it is insulinotropic. How significant would a nibble of cheese for lunch be on someone eating low carb?


Stephan Guyenet said...


Don't get me wrong, I'm not defending super-sized slurpees. Kitavans get about 10% of their calories from fruit. I think we can safely assume they're getting 3-5% of their calories from fructose, whatever number of grams that works out to.

I'm not convinced about the insulin resistance effects of saturated fat. Has it been shown to cause long-term insulin resistance or is this just in short-term studies? The fact that the Kitavans were able to tolerate 69% carbohydrate while eating 17% saturated fat and remaining thin and healthy is difficult to reconcile with long-term effects on insulin sensitivity.

And these are people whose close neighbors seem to be exquisitely sensitive to the insulin-destroying effects of the modern diet.

Anonymous said...

"The second thing is that high-fat diets actually do cause obesity in mice, even when the fat comes mostly from lard. So they aren't a good model for studying diet composition and health in humans right off the bat, because high-fat diets more often cause weight loss in humans."

Not true. If mice are fed zero-carb they weigh 15% less than the other eating a low-fat diet. They have to be overfed to weigh the same. Here is a study I found recently showing that a zero-carb diet protects mice against cancer better than low-fat, even though the low-fat diet was a lot lower in PUFAs (1.3% vs 8.8%). All mice got the same mix of fats: 10 parts butter fat, 10 parts lard, and 1 part corn oil. (I would have preferred 10 parts butter fat, 10 parts beef tallow, 1 part lard, no parts corn oil. But you can't have everything. This study's results speak for themselves. The mice that ate zero-carbs also had less fatty liver than all the others.


My problem with the study Peter has cited is that they use hydrogenated coconut oil, then blame "saturated" fat. They should have tested normal coconut oil. Many studies show that fresh coconut oil speeds up thyroid function and metabolism and gives a strong protection against oxidative stress (free radicals).

Anonymous said...

JB: "And, in addition to Varangy's request, would you comment on the supposed postprandial arterial constriction after a high-fat meal? Is this accurate? Does it actually reflect, more accurately, a high-fat, high-carbohydrate meal?"

Chris Masterjohn has written about this effect. The truth is that high PUFA constricts arteries, not high saturated fat. PUFAs do the damage, not truly saturated fats.

They blame SFAs, because they fail to isolate variables. Most of the studies call lard a saturated fat. Lard has 3-4 times more PUFAs than beef suet and butter, if not more. Some studies call french fries and doughnuts saturated fat. Don't pay attention to claims against eating saturated fat with carbs. They are all nonsense and they cause modern disease epidemics. I would eat as much saturated fat as possible and as little PUFAs as possible.

The Western Diet is low in SFAs and high in PUFAs and trans fatty acid. People eat low-fat dairy, lean meat with the fat trimmed or cooked off, soybean oil, Crisco, peanuts, fried chicken, etc. Those are all high in PUFAs and trans fats, not saturated fatty acids. The problem comes from the unscientific terminology, where something is called a saturated fat when it's less saturated, or higher in PUFAs than other fats (like beef fat, coconut oil, butter, macadamia oil, foie gras, and so on).

Stephan Guyenet said...


Zero-carb diets aside, feeding mice a high-fat diet that includes some carb makes them fat even if it's lard. And even if they aren't overfed.

A while back, I looked through some mouse obesity studies that claimed a high-fat diet caused obesity and payed attention to the methods section. I had an 'aha' moment because in the first few studies I found, the "high-fat" diets were also high in simple sugars. But I did end up finding several studies where the mice got fat even when sugar wasn't added and the diet was regular ad lib. I don't remember if the mice were a particular strain. I can dig the refs up if you're interested.

rheum101 said...

In response to Dave "the Kitavans have relatively little SFA in their diets."

They certainly have a high complex carb diet but they eat shed loads of coconuts (and fish of course). As you well know, this is both long but especially medium chain triglycerides.

The Kitavans have long relatively healthy lives on their traditional diets. Is there something about coconut that is 'protective'?

Peter said...

Hee hee, not being corn oil is probably what makes coconut oil protective......