Anthony Colpo recently posted a discussion of one of my older posts on seed oils and body fat gain (1), which reminded me that I need to revisit the idea. As my knowledge of obesity and metabolism has expanded, I feel the evidence behind the hypothesis that seed oils (corn, soybean, etc.) promote obesity due to their linoleic acid (omega-6 fat) content has largely collapsed.
There are many studies in which animals are fed refined high-fat diets to promote body fat gain under controlled conditions. Some of them compared different types of fats to one another. I've collected many of these over the years, and they are not consistent with one another. Some support the hypothesis that seed oils rich in linoleic acid cause more fat gain than other fats such as animal fats (2, 3, 4, 5), others show no difference (6, 7), and still others support the hypothesis that other fats cause more fat gain than seed oils (8, 9, 10). Perhaps unsurprisingly, studies in the latter category are most often cited. The only finding that is relatively consistent is that refined high-fat diets in which the fat comes mostly from omega-3 fatty acids (fish oil, flax oil, etc.) typically don't cause fat gain (11). However, these diets also cause a failure to thrive and pancreatic problems (11). Keep in mind these diets are the equivalent of getting 30-50% of your calories from fish or flax oil-- lower doses do not cause those problems.
Together, this broadly suggest to me that long-chain saturated, monounsaturated and omega-6 fatty acids are pretty much equivalent in their effects on body fatness in animals, and that differences observed in some studies are due to other factors such as the animals' flavor/texture preferences, minor constituents of the fats, the dietary background, or species/strain differences. In other words, seed oils are probably not inherently more fattening under controlled conditions than other types of fat.
Because of this, I think the idea that linoleic acid (the main fatty acid in most seed oils) interferes with thyroid function in a significant way is on shaky ground. There is some evidence from the rodent and cell culture literature that seed oils interfere with certain effects of thyroid hormone (12, 13, 14), however the overall physiological relevance of this is uncertain. Therefore, I've removed my old post, but kept the references (above) for those who are interested.
Another piece of evidence was the five year controlled human study in which volunteers ate a normal diet or a diet in which the animal fat was replaced by seed oils (15). Subjects in the seed oil group ended up weighing roughly 5% more than those in the normal diet group by about three years, although it wasn't designed to study the effects of dietary fat composition on weight change. There are a few problems with the study, including the fact that the investigators didn't measure body composition (so we don't know if the weight difference was in fat) and that many participants dropped out. It remains an interesting observation, but nothing to hang your hat on.
Now let's discuss the piece de resistance: the two studies that reported that feeding mice a diet with a poor omega-6:3 balance (excessive omega-6) caused a multi-generational obesity phenotype (16, 17). I was planning to discuss those studies in my AHS talk, so I revisited them to make sure they were solid. Unfortunately, I discovered serious flaws in both papers that fatally undermine the authors' conclusions. The short version is that neither study was properly controlled to come to the conclusion that omega-6:3 balance influenced fat gain. Both studies compared diets that differed in many ways, and that fact was not reported straightforwardly. In retrospect, I feel misled by these papers. I should have read them more carefully, and I apologize to my readers for that. As there is already too much misinformation on the internet, I've taken the post down.
So does this exonerate seed oils? Not necessarily. Our seed oil intake has increased dramatically in the last century, far exceeding that of our recent ancestors. Omega-3 intake has also increased, but the omega-6:3 ratio has nevertheless roughly doubled (18). As opposed to most other fatty acids, linoleic acid accumulates to a large degree in fat tissue (19, 20), and the amount depends on the proportion of linoleic acid in the diet. The linoleic acid content of body fat has increased tremendously in the US over the last 50 years, as shown in the following graph, based on a number of different studies, each of which is represented by an orange dot (21, 22, 23, 24, 25, 26, 27).
It's so linear that I can understand why someone might be suspicious of how I selected the studies, but I can assure you I included every relevant study I came across and was very surprised to see such a consistent trend. There are probably other studies out there, but the ones I found are sufficient to illustrate the change. I also included a study that looked at the linoleic acid content of chimpanzee fat, for reference (28). The caveat is that the paper doesn't specify if the chimp was captive or wild, so you can take it with a grain of salt. Linoleic acid in human breast milk has also increased quite a bit over the same time period (29).
I don't know what the consequences of this dramatic increase in the linoleic acid content of body fat and breast milk are, but it certainly gives me pause. We are truly living an uncontrolled experiment in the modern world. I think seed oils promote obesity because they're used to increase the reward/palatability value of processed food, but this would be equally applicable to animal fats and more saturated plant oils if they were used in the same foods. Do seed oils promote obesity by another mechanism as well? It's possible, but I'm not aware of any compelling evidence to support that idea at this point.
A true scientist is always open to revisiting previous hypotheses. Keep up the great work.
Fantastic that you put this all together, someone has to! Thank you :)
What do you think about the idea that increased omega-6 consumption disrupts liver functions, which can contribute to fatty liver disease and eventually to metabolic syndrome?
Well, I can't pass up the opportunity to bring up the endocannabinoid system and dietary PUFA. On the one hand, you certainly have obesity researchers interested in blocking ECS (if they can do so without horrible side effects) and on the other, you have at least one potential pathway where LA can be converted to anandamide (especially in those with chronically high insulin).
I've been wanting to check out the Kim paper (Endocannabinoid signaling and energy metabolism: a target for dietary intervention) that Emily Deans discussed here. Would love to hear your take on that.
Referring to the following post - http://wholehealthsource.blogspot.com/2010/01/body-fat-setpoint-part-iii-dietary.html
Pulling out a few select quotes -
"Inflammation in the hypothalamus, with accompanying resistance to leptin signaling, has been reported in a number of animal studies of diet-induced obesity"
"Any time a disease involves inflammation, the first thing that comes to my mind is the balance between omega-6 and omega-3 fats. The modern Western diet is heavily weighted toward omega-6, which are the precursors to some very inflammatory substances (as well as a few that are anti-inflammatory). These substances are essential for health in the correct amounts, but they need to be balanced with omega-3 to prevent excessive and uncontrolled inflammatory responses. "
So is the mechanism : Seed oils --> distorted O3/O6 --> inflammation of hypothalamus --> leptin resistance --> obesity.
What say you?
Aravind beat me to it. I second the motion to ask if n-6 oils' role in inflammation could promote obesity.
Me too - re inflammation and PUFA. Your comment that omega 3 didn't cause weight gain but metabolic problems with pancreas.
Stephan, did the omega 3 give pancreatic problems or was it omega 6 and sat fat in that study you quoted?
I think I misunderstood you.
Wow, I look forward to future posts on this topic! The thinking on Omega 6 EFA really seems to be going through quite a transition. Good job revisiting the data and questioning the results of these studies.
Regarding breastmilk: infant formula manufacturers boast of their products being closer than ever to breastmilk, but the breastmilk they refer to is the average American breastmilk. They have a very high w-6/w-3 ratio. Hardly anything to aspire to, based on the fatty acid composition. In the last decade they have started adding DHA and ARA, but also in relatively low levels. There is also the issue that the refining process, particularly the deodorization step, of the vegetable oils in the formula leaves a significant level of trans fat in infant formula.
Going back to breastmilk, the same is probably true, as the vegetable oils, such as soybean and canola, that the mothers consume all contain unintentional trans fat from refining process.
Stephan, I find this interesting and it will take awhile for me to digest. My personal experience has been when I did a whole foods version of Atkins but used seed oils, my hunger and setpoint issues stayed around and I gained the weight back. However, when I switched over to saturated fats, I've lost weight and kept it off. I feel pretty strongly that the high omega 6 oils definitely are screwing with something in a major way. The unnatural process of it's processing is a major red flag from the start. Could you please speak on omega 6:3 imbalance and it's relation to systemic inflammation.
From a holistic perspective, n-6 doesn't seem to cause weight gain, per this post. But not mentioned are updated thoughts on heart disease or liver toxicity. Stephan has been focusing on weight (adiposity) lately, but I suspect there is decent evidence to support damage in other ways from n-6 fats.
Appreciate your honest analysis Stephan. Getting n6 n3 back to a normal range helped my family enormously. i have also been experimenting/exposing friends and family to oil improvement with some humorous results. I look forward to your hypotheses with great interest.
I remember my joy in reading about low carb diets fixing obesity, about the horrors of wheat lectins causing lepton resistance and leaky gut, as well as seed oils increasing fat gain and promoting inflammation. Lately there have been many revisions: carbs aren't bad, wga and other lectins are actually very heat sensitive, and now seed oils aren't as bad as before? I don't know, maybe seed oils really are bad, but at the moment the only harmful substance I see is fructose.... But then again, maybe in a couple weeks we'll find out it's not that important ( after all, the foremost crusader against fructose, Robert lustig, seems to be pushing overweight status himself).
Maybe the most important factor is maintaining nutrient repletion. Out of everything it seems the most logical...
Actually, to expand on the more salient point from my last comment:
the increase in linoleic acid is a proxy for the increase increase in vegetable oil consuption. Much of that vegetable oil was hydrogenated, and even most of the oil that isn't hydrogenated is deodorized, which unintentionally creates unnatural trans fats.
Can you construct a graph of the increase in trans fatty acids in body fat? (forgive me if you already have, I can't remember and a quick search didn't find it) What are the effects of these novel and unnatural trans fats on inflammation, leptin resistance, neuronal function, eicosanoid signaling, etc., and ultimately, for the purpose of this post, adiposity?
I remember my joy in reading about low carb diets fixing obesity, about the horrors of wheat lectins causing lepton resistance and leaky gut, as well as seed oils increasing fat gain and promoting inflammation. Lately there have been many revisions: carbs aren't bad, wga and other lectins are actually very heat sensitive, and now seed oils aren't as bad as before? I don't know, maybe seed oils really are bad, but at the moment the only harmful substance I see is fructose.... But then again, maybe in a couple weeks we'll find out it's not that important...
I share this sentiment. It seems that the entire Paleo edifice is crumbling daily. And that we're back at Calories in Calories out and eat-less-move-more. I don't know what to think anymore... Or what to eat.
I guess I'll just stick to medium carb Paleo as its working for me or maybe I'll go hang out with Matt Stone and Chalen Wychoff. Who knows anymore.
Thanks for this post Stephan. I really respect and appreciate your honesty and rigor and willingness to revise previous positions.
Agree that we're living in an uncontrolled experiment and it's hard to say what the effect may be.
My thought about n6 fat and obesity was along the lines of what Ray Peat argues- polyunsaturated fats reduce metabolic rate, which may contribute to a mis-match between appetite and expenditure, resulting in obesity for some.
Peat shares the story of farmers a couple generations ago inducing hypothyroidism by feeding pigs corn and soy, increasing the PUFA content in their tissues, and allowing them to gain fat it seemed a plausible mechanism for humans as well. Any thoughts on that?
The dogs have been called off of fructose as well, at least in quantities of less than 25-50g/day. (See Ned Kock's Health Correlator blog for more info.) In my opinion, the dangers of fructose depend on quantity and context.
There's still gluten to kick around - although there are naysayers on that, too (Chris Masterjohn). (Not the same thing as WGA.) A diet that regularly includes large amounts of phytic acid still probably isn't a great idea. And I don't think we can give n-6 oils a full pass yet, either.
Hi Stephan, are you saying then that PUFAs do not cause liver toxicity? Thanks very much for dealing with the obesity issue here, but I think there are lots of other harmful effects of PUFAs that aren't addressed here.
I appreciate you slaying some of the paleo demons (such as insulin) so we can hone in on weight loss, but that doesn't leave us with many culprits for diseases of civilization. In fact, I'd say only fructose is left now, and maybe its days as a paleo demon are numbered too.
So do you even believe in diseases of civilization any more?
Why do you, personally, think modern people are so much sicker than hunter gatherers?
I think Stephan here is approaching this matter strictly from the viewpoint of body fat and obesity.
I'm pretty sure excess insulin, high Omega-6 consumption, fructose, etc. can be harmful in other ways besides making us fatter.
When the studies and science are contradicting each other, it's always a good idea to view these things through an evolutionary lens.
Let's not get too lost in the science, as well all know modern science is often corrupt and inconclusive.
And in the end it is the END RESULTS that matter. People lose weight on low-carb, they lose weight when they cut back on fructose and their health improves when reducing Omega-6 consumption, type II diabetics can reverse their disease on very low-carb, etc. etc.
The end results are clear, whether the mechanism is known or not is up for debate.
It's clear processed food takes a toll on our health. Pinpointing the worst ones is hard since we all eat a variety of foods. Last week it was hot dogs, but people tend to eat hot dogs in buns and wash them down with soda, and you don't want to know what's for dessert. Blaming carbs, or oils, or animal foods, or grains might be useful, if it helps us avoid processed foods, but I think it's mostly guesswork at this point.
i've slept on it and i have some more things i want to bring up. i understand that you see no mechanisms by which seed oils lead to obesity- in my mind, i always thought that fructose played the biggest role in adiposity- but how do you see the role of seed oils in systemic inflammation and as a major catalyst to various diseases of civilization. does it have a role in inflammation of the hypothalamus leading to leptin receptor damage?
"Let's not get too lost in the science, as well all know modern science is often corrupt and inconclusive.
And in the end it is the END RESULTS that matter. People lose weight on low-carb, they lose weight when they cut back on fructose and their health improves when reducing Omega-6 consumption, type II diabetics can reverse their disease on very low-carb, etc. etc."
People lose weight on any monotonous whole-food diet, such as low-carb. They lose weight when they cut back on refined sugar not fructose, and industrial vegetable oil not omega 6. Had they kept their honey and peanut butter they would've been just fine, perhaps better off.
You tell us not to get lost in science, yet use these highly reductionist bits of reasoning that resulted from people getting very 'lost in the science' and throwing the baby out with the bath water.
I think we should pay strict attention to the science - but in a field as complex and ambiguous as nutrition we should also take it with a dose of common sense.
One of your previous posts (http://wholehealthsource.blogspot.com/2009/05/eicosanoids-and-ischemic-heart-diseas.html) states that omega 6 concentrations in tissue plateau after intake gets above 4% of calories.
However the graph in your current post showing steadily rising linoleic acid content of body fat over time, presumably long after omega 6 intake had surpassed 4%.
Can you clarify whether or not this is a contradiction?
Roberto: there's more to health than just weight loss.
The end results are clear, whether it is the PUFAs, fructose, sugar, vegetable oils or something else that is leading to the results is up for debate.
Most people do lose weight on low-carb, despite the fact that low-carb diets are often very palatable.
Nothing Stephan or you have said has changed that fact, although the mechanism may not be lack of insulin but something else.
Approaching this whole matter from a strict weight loss sense is wrong in my opinion. What about heart disease, diabetes and others? They are correlated with obesity but not always.
i agree with you about the low-carb being an integral part in a great number of people losing weight- at least initially. my hunch is that the decreased fructose gives the leptin receptors a chance to heal therefore allowing the body's natural weight regulation system to recallibrate itself. i do think once it's recallibrated, starchy carbs can be ramped back up to normal levels. i also think seed oils fit into this matrix in some way also because even when taking out gluten and fructose, the weight always came back. it wasn't until i took out the o6 oils that the weight loss was sustainable and effortless.
When you first started writing about the food reward theory of obesity, in trying to reconcile it with my previous understanding based on NAD, it occurred to me that maybe they were the same theory; i.e. foods like wheat (with its opiate receptor effects) and linoleic acid (which is an essential fatty acid) might be causing inflammatory states as well as increasing the reward value of food.
In the specific case of linoleic acid, I think of it in a similar way to how I think about fructose: we've evolved a preference for it because its incidence in the wild was relatively limited (wild animal fats have relatively low levels of it). This is consistent with the idea that we would preferentially store up excess linoleic acid in the fat cells for times of scarcity.
As such, it is not clear to me that if Doritos or McDonalds french fries were made in animal fats, that they would be as fattening as they are now, or even that they are fattening at all. Maybe the reason that nuts and peanuts and peanut butter and chinese food, etc. are so easy to binge on is because they are hyperrewarding in and of themselves as a result of the simple fact that they have excess linoleic acid in them. Then maybe, downstream of this effect, the overconsumption of these NAD causes inflammation, which disrupts leptin signaling, causing a greater consumption, etc. What do you think?
Aravind and a couple of other people mentioned this, but I wonder about the long-term connection between the inflammatory effects of n-6 LA and weight gain.
As I outlined in this post, several lines of evidence suggest that inflammation is at least a contributing factor to obesity. We have a lot of evidence that n-6 LA causes inflammation.
We have one 3-year study in humans that showed a very slight weight gain in people who replaced saturated fat with LA. Are there any other long-term studies?
In any event, seems somewhat of a moot point from a clinical perspective since there are several other reasons to avoid industrially processed seed oils. I do appreciate the clarification and applaud your willingness to change your stance based on the available data. I wish all scientists had the same sense of integrity.
I was just thinking about Chris's articles on inflammation and obesity. It really does look like various cytokines disrupt the endocrine system in various ways. But the issue is probably that a high linoleic acid diet doesn't necessarily cause higher inflammation under all circumstances. A good discussion would be under which circumstances does it cause inflammatory problems and under which circumstances doesn't it?
Stephan. Thank you for continuing to update us as you change your views.
I was wondering, though, whether you had considered prepending a disclaimer to your old posts (with links to your new thinking), rather than removing them.
You have to manage your own reputation your own way, and I know full well what it's like to have some embarrassment over old views (and the ways they were arrived at), but it seems like it would be a wonderful way to watch a young scientist learn from his errors were you to leave the posts up with addenda added.
I will continue to be a fan either way.
@luckybastard - "does it have a role in inflammation of the hypothalamus leading to leptin receptor damage?"
Look at comment #5 from Aravind. You guys had the identical line of thought and the same exact question!
@Jack - LuckyB and I spoke earlier today. I think he was just reiterating the question because the general consensus with several of us is that inflammation is still a smoking gun that impugns vegetable oils. We figured if we all gang up on Stephan that he will yield :-)
I don't think you really needed to retract the old post Stephan, I mean the evidence wasn't super compelling to begin with. I got the jist from you that you were just hypothesising and thinking 'well maybe..'.
I still think seed oils have some part to play in obesity through messing with endocannabinoid system.
I get a really lovely calming effect from seed oils. I don't know if anyone else has the same experience but combine that with white flour and sugar and I'm practically sedated. Don't get this from mashed potato with butter unfortunately, though food reward could be hard to tease apart with that one.
a lot of lists of goitrogens include canola oil -- has this been exonerated, too?
While the emphasis of this post was on seed oils and obesity, I (along with some of your other readers judging from comments) am wondering what your thoughts are regarding w6's and inflammation. Apparently Walter Willett's opinion is now that w6's are not pro-inflammatory, and as you know it has been suggested that replacing SFAs with PUFAs has benefits re: CVD.
I'm interested in your thoughts on this aspect. Thanks.
The huge problem with any obesity research and diet is that you will never get anything consistent studying adults.
I am sure you are aware of research of Dr Jules Hirsch and Dr Ethan Simms that pretty much shows that weight/BMI can be considered as an elastic body attribute.
For those not familiar with this, then I suggest you read Gina Kolata's article:
Have you blogged on this previously?
But given the above, we are always left with the question of, why is society in general producing children slightly fatter than their parents? The obvious answer then is the magic buzz word 'epigenetics'.
A question I posed to Chris Masterjohn, and awaiting some interesting response on is as follows: If diet is the big promoter of obesity, why does W.A Prices research: Nutrition and Physical Degeneration not mention the word 'obesity' once. Did Dr Price fail to notice it? was the effect too small? or was it non-existent?
But, given the sketchy inconsistent results of obesity observations, there does seem to be dietary behaviours that can promote short-term weight gain in people - some probably harmless (like diet change to more calories) and other maybe associated with harm - thyroid damage? liver damage?
If epigenetics is happening, and making children express higher weight set-points, could we perhaps make an educated guess that a lifestyle habit that can induce short-term weight-gain/loss in adults, will have a long-term effect in infants?
One, behaviour that may cause infants to develop higher weight set-points seems to be calorie restriction, ie The Dutch Famine. Could habitual dieting be a factor in epigenetic obesity?
If linoleic acid, or fructose or whatever plays a role in obesity, then we need to be looking at the diets of pregnant mothers, babies and toddlers for the smoking gun.
Interesting turn of events. I am still wary of seed oils though.. Then again I have definitely jumped off the fructose bandwagon. I think the fructose-phobia is a distraction for all but maybe the most metabolically screwed. For someone who is physically active every day, fructose and simple sugars are a godsend.
I have also noticed a relaxed, calm attitude when eating seed oils. I don't really interpret that as a good thing though.
It is an interesting finding that in six years of clinically testing humans who are also obese I have never found one with a O6/O3 ratio below 12 to 1. And not one have had a clinically normal thyroid symptomotology. Moreover, they all have elevated acute phase proteins to some degree. Again this is correlative data but the fact that the trend is strongly inflammatory points the smoking gun to an overweighted series two prostaglandin issue in all tissues. This is supported by the high O6/3 ratios. And most have a diet loaded with omega six oils, refined carbs especially fructose products. Not all O6 oils are bad either. Some O6 oils are quite helpful in series three and series one reactions in tissues. Obesity is a disease of multiple axis's acting simultaneously on the hypothalamus at once. I think stimulus intensity and duration are critical factors and focusing in on macro nutrient ratios futile. This is modeled in the hippocampus with cortisol levels in PTSD and in learning problems and sleep deprivation. Dr. K
Further to my vote for epigenetics. Has anyone looked at this study:
I he originally promoted his earlier work as being evidence that saturated fat was the trigger. I sent him an email to ask how he figured that and he sent me his paper - which mentioned Saturated fat only once, and not in relation to anything of importance.
Asking him again, he made some hand-waving argument about Omega 3 levels.
I'm glad you took the time to read the manuscript. The interpretation of the sat fat argument just so you know is as follows:
All the mothers on the HF diet had an omega-6/omega-3 ratio in plasma (a steady state measure of saturated fat) that was 3 times higher compared to chow fed control mothers, as expected when you feed this type of diet. The same was true in the fetal plasma compartment.
Now consider that although the HF diet resistant (HF-DR) mothers in our study ate this high calorie high saturated fat diet they showed a metabolic profile very similar to chow fed control mom's, e.g. normal wt gain, insulin, etc during pregnancy with the exception of the omega 6/3 ratio.
You might conclude that the fetuses from HF-DR mothers were exposed to less calories (we dont know that) or were just getting the normal maternal metabolic responses that happened to have an extremely high sat fat content. What is most interesting is all the high fat fed offspring weighed slightly less than the fetuses from the chow fed mothers (table 3) but all had evidence of severe liver damage. Knowing from the literature that sat fat produces more liver damage in rat studies than equivalent number of calories from carbs or starches, it's a pretty safe bet that sat fat is the bad actor here. The next paper will report this sat fat story in more detail.
Nevertheless, is this not evidence that liver damage during pregnancy is an epigenetic trigger for obesity?
Friedman is saying the O6/O3 ratio correlates, but assumes it is a marker for SUFA! (Oh the irony here!)
What is in the junk food that is causing this liver damage?
"high-fat diet of potato chips, peanut butter and chocolate"
Now this does not sound particular high in fructose* (apart from chocolate)? but is high in linoleic acid! maybe trans-fats also?
* - wouldn't most monkeys have a high tolerance to fructose?
I understand this post is mainly focused with weight loss, but I have a quick question regarding health and quality of life. Obviously many populations enjoyed great health on variable macronutrient ratios. But after hearing about depression arising in those individuals consuming a low fat diet, both HDL and LDL decreasing and trigs increasing with too little fat present, as well as a kind of cognitive haze that arises (anecdotally, at least). I'm just wondering, weight loss aside, would you recommend a higher fat (from whole foods, of course) diet for superior quality of life?
Gordon I have been beating the epigenetic drum for a while. I think it explains all the messy macronutrient studies we see. The real conundrum is that we dont have tests that tell us the epigenetic switch position of what influences it. We do have to use clinical hormone testing as the Rosetta Stone to deciphering how we partition/react to different foods. We also can make adjustments to the diet and retest serially. Its time consuming and tedious but I do it a lot. Most patients don't have the stomach or pocketbook for it but It reveals some interesting correlations and clinical pearls about diets and how our ferrari really works. Dr. K
"I discovered serious flaws in both papers that fatally undermine the authors' conclusions...Both studies compared diets that differed in many ways, and that fact was not reported straightforwardly."
I've obtained the datasheets for both the diets in Hanbauer et.al.: Teklad Custom Research 00522 and 7912 (the irradiated version of 7012).
Both diets contain ground corn and ground wheat as their two primary constituents, and both have roughly equivalent macronutrient ratios: 23/60/17 vs. 29/56/15 PRO/CHO/FAT.
The main difference appears to be the substitution of soy meal, fish meal, soy oil, and "wheat middlings" in 7912 for casein, safflower oil, and "cellulose" (which may be the same thing as "wheat middlings") in 00522. This would be necessary in order to make the mix n-3 deficient, since soybeans and fish meal contain a meaningful amount of n-3, particularly ALA.
Have I missed some other vital difference between the two diets? Because while I put limited stock in mouse studies, those differences don't appear to merit throwing Hanbauer et.al. on the trash heap.
"Do seed oils promote obesity by another mechanism as well? It's possible, but I'm not aware of any compelling evidence to support that idea at this point."
What about insulin resistence? On "The Carbohydrate Hypothesis of Obesity" you wrote:
"This brings us to a core concept that Taubes fails to incorporate in his thinking: the idea that insulin signaling in fat tissue depends both on the concentration of insulin around, and on the sensitivity of cells to that insulin. In obesity, fat tissue is insulin resistant."
Here we have a crossover study on diabetics on high PUFA (30ml of sunflower oil) versus high MUFA (30 ml of olive oil) diet. Two weeks on one diet and then two weeks on the other. Fasting glucose (11.8%) AND fasting insulin (13.5%) were both higher in the PUFA fase.
Rats and mice don't equal humans and never will. They are great for science to form hypothesis but very little can be generated between species. If you want to know the essence of human obesity you need to focus on human studies in my view. That however opens a new set of issues. Great point j Stanton on those diets and points about rodents! Dr. K
I've always felt that the whole "O-6" thing was because of (micro) nutrient paucity. We've essentially taken the three macronutrients and divorced them from any significant micronutrient content. We have isolated soy protein, whey and all their buddies. We have refined oils. And finally we have the sugars and refined flours. You could make a meal out of all three of those and contribute very little toward micronutrient intake. I think that high fructose corn syrup has a little manganese in it.
It may not have anything to do with anything, but that's why I avoid them. Edible foods with lots of micronutrients taste good to me (until I'm full). It's that simple.
Why does Walter Willet from Harvard assess the liturature and come up with the believe that "the statement that 0-6s are proinflammatory is a myth".
I follow Paleo, but I still read everything, even the J of American Dietetics. I am scratching my head at all the conflicting expert opinion RE 0-6s and inflammation. On page 674-675 of the May 2011 J of ADA there is a Q&A with Walter Willett MD at the Fall 2010 ADA convention.
RD Q “There’s been much agreement about the need to increase polyunsaturated fats. But there is also a concern about n-6 fatty acids being inflammatory due to their production of inflammatory eicosanoids. I’m wondering about the message of increasing polyunsaturated fats without really saying a lot about the n-3s being a necessary component of that.”
Willett: “Yes it is amazing how far a myth can go withjout any data to support it. And I agree with Dr Kuller that the encouragement to increase n-6s during the 1960s and 1970s had an enourmous benfit on reduction of cardiovascular disease. We went from about 3% of energy to about 7% of energy from O-6s. And so if we try to go backwards – there are people out there wanting to reduce n-6s because of this hypothetical inflammation – that would increase the risk of cardiovascular disease, we know that very well. It would definitely lead to higher LDL. There have been carefully controlled studies that have fed different levels of O-6s and looked at inflammation markers, and they dont show any adverse effects. And that is partly because there are many other pathways by which n-6s operate, besides possible competing with n3s in the elongation and desaturation pathways. For example, n-6s downregulate NF-kappaB which is a major controller of inflammation. And they are also an insulin sensitizer, acting at PRAR-y. And if you reduce insulin resistence, ie, improve sensitivity – that will be anti-inflammatory. So n-6s are not proinflammatory.”
Because Walter Willet should have been an English major.
There is some evidence that n-6 can contribute to liver damage, but that probably also requires a calorie excess.
Hi Aravind, Helen and Sue,
It's a plausible mechanism, but in the absence of any direct evidence that n-6 promotes obesity, it's speculative. Excess n-6 may have other harmful effects besides obesity.
Hi Rob A,
I think that anecdote about farmers feeding their animals is not that useful, personally. I'd like to see some hard evidence rather than a story. The other thing is that if I recall, the story involved animals that were leaner eating coconut fat. Coconut fat is a medium-chain fat and so it isn't really a fair comparison with linoleic acid. A better comparison would be olive oil, lard or some other fat made with long-chain fatty acids.
I'm not saying that seed oils aren't harmful, this post was specifically about body fatness. I still believe in the diseases of civilization.
Hi Dim Sum massacre,
That graph represented omega-6 in plasma membrane. This graph is LA in fat tissue. They don't follow the same kinetics.
Hi Chris K,
Honestly, the basis for saying that n-6 causes inflammation is mostly hypothetical. I've come to realize that the case based on eicosanoid formation has been overstated. The strongest piece of evidence is that excess n-6 seems to suppress n-3 elongation and accumulation, and we know that n-3 is important in resolving inflammation.
Hi J Stanton,
Those differences are significant enough that I don't feel the study was adequately controlled.
Thanks for the link. I think there's some evidence that seed oils degrade glycemic control and could contribute to cardiovascular problems in diabetics.
Hi LisaO and Garry,
I respect Walter Willett and I think he has made important contributions. However, he is not a "mechanism guy" and his statements here are really not well founded IMO. He claims that the reduction in CHD deaths in recent years are attributable to an increase in LA intake, but what about the 50% reduction in smoking, improvements in emergency treatment, and preventive drugs? I've seen no evidence that increasing LA intakes have contributed to the reduction in CHD deaths, and I think the controlled trials do not support that mechanism.
I think Dr. Willett is correct that the evidence has been exaggerated that higher LA will lead to higher production of inflammatory eicosanoids. At least above a certain threshold, it doesn't seem to matter much. Bu what he doesn't discuss is how excess LA displaces omega-3s in tissue and may reduce the production of omega-3 derived eicosanoids that resolve inflammation. This has been shown to lead to insulin resistance, at least in rodents, but it operates on long timescales.
He claims that LA inhibits inflammatory signaling through NF-kB and sensitizes to insulin through PPAR gamma, but there's no evidence of that in vivo that I'm aware of (in other words, that evidence is from cell culture rather than in actual animals as far as I know). To the contrary, as Mario mentioned, seed oils seem to impair glycemic control and insulin sensitivity in diabetics relative to oils rich in oleic acid.
I agree with Dr. Willett on a lot of things, but the seed oil idea isn't one of them.
Reading all of this is just like reading the report of the committee of blind men describing the elephant. Each report sounds good but none of the blind 'experts' agrees with the others. Not only that, they seem to seldom even listen to the others. The excerpt from Kolata's book posted by Gordon Rouse is particularly disheartening, especially since that work was done years ago and we are still being told to 'just eat a little less and exercise a little more'. Oh, and now it will help if the food doesn't taste very good. Yeah, right.
Seed oils make me feel terrible. Ingesting a significant amount makes my heart pound. I feel bloated too.
Stephan, I have always loved your blog but reading this post reminded me of why Michael Pollan was so brilliant in his critique of "nutritionism." We should eat natural wholes foods and not get caught up in the variables that make up the food, about which we know little.
I eat nuts because they make me feel good. French fries make me feel awful. Both are high in LA, but I doubt that makes them both unhealthy.
One more thought:
Can't vegetable oils be bad for some reason other than the presence of LA? Didn't Udo Erasmus document that during the manufacture of vegetable oils much of the fat becomes toxic due to high heat during processing?
Vegetable oils may be bad for reasons beyond LA, no?
Do you still believe that keeping intake of LA below, say, 4% of calories likely leads to substantial reductions (e.g., 50-90%) in the risk of getting cardiovascular disease and cancer?
The evidence in favor of this hypothesis included a small number of RCTs (like the Lyon diet heart trial), animal evidence, ecological/observation evidence (kitivans, Inuit, Japanese, etc), and mechanistic plausibility.
Recently, it has been discovered, however, that eating more LA does not result in more AA in red blood cells. Findings like this eroded the mechanistic plausiblity of the theory to some extent, right?
That said, I still strongly suspect that excess LA is one of the principal causes of CVD and cancer...
Do you agree with my assessments?
Thanks for your responses, Stephan, they're very informative.
What do you think of the argument that omega 6 is problematic because it is excessively prone to oxidation?
Incidentally, my own (admittedly half-informed) feeling on fats is that the most important distinction to make is that of traditional fats vs. processed. If you are consuming fat in a form that existed prior to the 20th century (whether it's olive oil, dairy/lard/beef tallow/duck fat from naturally raised animals, coconut, nuts, etc.), don't worry about it. But vegetable and seed oils, be wary of.
It seems to me like maybe this is the real key to the success of the mediterranean diets... the countries in this region all tend to have in common a strong attachment to traditional fat sources.
Well, I don't know about seed oils, but I'm comfortable taking them out of my diet. I basically will eat anything that I could have either picked or killed but have kept wheat out. All fresh vegetables and fruits are fair game; all fresh meat and fish; eggs; nuts; and I've left in cheese and butter. No pasta, no seed oils, no processed or manufactured food. I have no idea how this has changed my O6-O3 profile. Anybody have any hypotheses? I really like the food reward idea because my cravings have totally subsided. I think trying to pinpoint a specific substance as the "cause" of the obesity epidemic is a mistake - our diets have changed so drastically from those of our parents and grandparents that trying to single out a particular item as the "cause" is bound to end up failing.
One idea (can't take credit for it) about the potential health difference between n-6 fatty acids naturally occurring in whole foods and those found in processed vegetable oils is that when you eat them in whole foods you are also eating the plant's own antioxidants (such as vitamin E and other "phytochemicals") that protect them from getting rancid in the plant, which, in turn, protect them from getting rancid (oxidizing) in you. Maybe with these fellow travelers they are good for you and without them they are not so much. (This also applies to n-3 oils in plants, such as flaxseed.)
The other changes they go through when processed also dim the picture for refined vegetable oils, as others have noted.
Diabetics basically oxidize everything more quickly (I know I'm probably making an absurdly unscientific generalization, but this is how I understand it in laymen's terms), so maybe n-6 oils are worse for them for this reason. I wonder if dietary cholesterol is worse for diabetics for this reason, as well - I just read a study showing egg consumption correlated with increased CVD in diabetics, but not in healthy people. (This was contradicted by another study, but it's suggestive.)
I am still having a very hard time with the food reward issue, and to a lesser degree with e set-point approach.
It seems pretty obvious to me that a high saturated fat meal, with few calories from starch, and none from sugar, leads to much reduced extremes in blood sugar levels. The result is a much smoother transition and different feelings of hunger later, a hunger without the sense of desperation.
As for palatability, chicken liver pate tastes pretty good, properly made it's pure paleo, and I cannot imagine anyone getting fat on that. Or for that matter bacon or a properly cooked steak.
And if eating a "Paleo" diet changes the set-point...well, why?
I remain puzzled by making a complicated structure to explain obesity, when the answer, or solution, is to simply eat a more saturated fat based diet. If that is the cure, then the problem is excessive consumption of starches and carbohydrates. Now, it may be that obese people eat too many carbs, because they taste so good, but again, as some of us understand it, what we may be seeing is the low sat-fat diet run amok. And while before these poor folks were told to watch what they eat, and avoid saturated fat, which resulted in excessive starches and sugars and (healthy) vegetable oils, now they are told to lose weight by eating bad tasting food. I have never been obese, or even 15 pounds overweight, but I think I might just throw in the towel if I had a lifetime of bad tasting food to look forward to.
I mean, what's the point?
Again, however, a complicated analysis may be based on reality, in the sense of two or more overlapping systems regulating body weight. The body is not a machine, it is a self-regulating organism.
Still, I think this is losing track of the more basic reality. Are there some obese Paleo-diet people? Or are Paleo people self selecting for thin body types?
Your blog has truly inspired me, especially the current article. thank you for sharing your thoughts and the impact of them has been tremendous to me.Thank you more time and keep up with the good work.
thanks loss weight quick
Its depend on the great thoughts and good knowledge,,I am so hapy to read your article and getting some extra.
There are two hypothetical statements. One is that seed oils cause inflammation. The other is the suggested pathway that they could do so. If you invalidate the pathway, the first part of the statement may still ring true. I'm reminded of your (Dr. Guyenet) post on seed oils being strongly connected to osteoporosis.
Resolution of Adipose Tissue Inflammation Ana González-Périz and Joan Clària Well illustrated review paper free online here that may help some people.
Oils would be expected to increase obesity risk due to the energy density (or calorie density) of oils.
Oils contain 9 calories per gram. They contain no fiber. So, one tiny tablespoon of oil contains over 100 calories.
The human body is very efficient at converting dietary fat into body fat. Very few calories are burned in the dietary fat to body fat conversion process. Less than 5 percent.
But with carbohydrates, the body has to burn about 25 percent of the consumed calories in order to convert dietary carbohydrate to body fat.
This is why people who live on rice and vegetables are trim, while those who live on meet and cheese are fat.
Bottom line: if you want to lose weight, eat foods that are loaded with water and fiber.
Aspirin blocks inflammatory factors. The same inflammatory factors produced by arachidonic acid.
Aspirin causes fat loss by lessening inflammation.
Vitamin D lowers parathyroid hormone, lessening the inflammatory effect. This could in turn cause weight loss.
Polyunsaturated fatty acids could cause weight loss by lowering inflammation. In contrast, polyunsaturated fatty acids could cause weight gain if they are converted to arachidonic acid, which causes inflammation.
Sugar could cause weight gain by inducing inflammation and/or oxidative stress. In contrast, sugar could cause weight loss by replenishing liver glycogen, which activates the conversion of thyroxine to triiodothyronine.
This is all theory, and their empirical effects are contradictory from one study to another.
hai Stephan Guyenet, im send email to you, plis cek. thanks
Your observation and analysis is so honest. Your post is full of lots of information. Thanks for sharing.
So. Two things:
1) Do the studies comparing fats in mice account for the fact that 1 g fat != 9 kcal? Different length fatty acids have different calorie content. Add to that that it takes 5-10% of the energy to digest the fat (more energy to digest protein, less to digest carbohydrate).
2) What about hexane? Cheap seed oils are extracted with hexane. Hexane doesn't seem to accumulate in humans (see studies on shoe gluers) but it could extract all sorts of interesting compounds that might never disassociate from the indigestible fraction otherwise.
Love your thoughtful posts.
I haven't read the other comments carefully, so please disregard the following question if it proves redundant . . .
Do you think the "increase" in these fats found in fat tissue may have something to do with changing phylogenetics (e.g., increased immigration)? I guess what I'm asking is WHO was included in these studies over time? Caucasian Americans, mostly men, women, etc.? Or did they start out with the latter and shift to studies conducted primarily in SE or E Asia? Just trying to understand a potential source of variation among the studies you've included in your graph. . .
Helen wrote ...
when you eat them in whole foods you are also eating the plant's own antioxidants
cuts both ways, but the bulk of the evidence as I'm seeing it is that extra antioxidants are more likely to hurt than help.
The recent studies debunking some of the earlier vitamin C and E studies, the failed (spectacularly) intervention studies of Beta Carotene in smokers, this stuff is adding up.
some even looser and woolier thinking: from the plant's side, the protective chemicals are optimized to work in the milieu of the plant
from the animal's side: how could two such completely different entities have similar enough kinetics in the stomach, intestines & liver to end up in the same compartment?
Even if both the oil and the antioxidant survive the digestive tract and the liver's first pass (remember nutrients are delivered to the liver first, and if cytochrome doesn't like the looks of the antioxidant ... ) one does not know where the oil and the antioxidant will end up ... Certainly if the body is smart enough to recognize an external protective chemical as useful and "tag" it to the target susceptible/protected molecule, or make sure they end up in the same compartment, it would be "smart enough" to make its own
 shortcut for "if there was sufficient evolutionary pressure for ... "
I wasn't trying to convince you BTW, It's a question for experiments to decide, not loose and woolly arguments.
The studies I've seen showing antioxidants to be harmful were using antioxidants as supplements. I think any food component taken out of context can be a bad thing. Too much antioxidant vitamins can be deleterious in a variety of ways.
As for the other chemicals in nuts - I think small amounts of phytochemicals (to lump them together) can have pharmacological or (as Stephan has posted) hormetic (sp?) effects. In large doses, yes, they are nearly all - or perhaps all? - toxins.
For a while I got freaked out by the "vegetables have toxins!" as well as "fruits have fructose *and* toxins!" warnings. I decided those are two food groups I am willing to take my chances on. I'm trusting our body has generally learned to deal with - and use to best advantage - stuff we've been eating for hundreds of thousands of years. Yes, I know, I know with the cultivated year-round fruit, but I have just not seen *any* study showing anything bad happening, but rather, often good things happening, with regular fruit consumption. I have scrubbed PubMed for anti-fruit evidence, but haven't found it. Certainly, if it makes a person feel bad, they shouldn't eat it.
Going back to my argument about n-6 oils in their natural packaging vs. refined oils - even if the body doesn't use the plant's own protective chemicals, at least they are protecting the oils in the plant until consumed. All manner of unpleasant things can happen with refined vegetable oils, and adding Vitamin E or whatnot are added back later may not help. Like any commercially refined food.
I think if you're looking for evidence against eating nuts and seeds because of their n-6 content, you'd have to look at the evidence from actual consumption, not theories about the components of the nuts. (From your comment I think you'd agree?) I've seen mostly good things about nuts, but I haven't really dug into the abstracts & studies, like I did with fruit. Maybe because my kids eat fruit, but one's allergic to nuts. ;)
Agree with your comment in its entirety
Helen wrote ...
were using antioxidants as supplements. I think any food component taken out of context can be a bad thing
A point I wrote initially but edited out for length. The point that experiments have to be the final arbiter
Helen wrote: unpleasant things can happen with refined vegetable oils
A point I was playing upon in half seriously responding to Todd on carbsanity: if insulin locks fat into fat cells then in hyperinsulinemics the polyunsaturates would eventually become tar in the adipocytes.
Re I think any food component taken out of context can be a bad thing. Too much antioxidant vitamins can be deleterious in a variety of ways." We must not forget Vitamin D3 and Melatonin are both antioxidants.
Although both are available to some extent from food sources they are primarily created in humans via sunlight exposure, UVB to skin (Vitamin D3) bright light to pineal gland (melatonin secretion)
The natural ability to create these antioxidants not only declines with ageing but insufficiency status is now common globally at all ages given the environmental pollution (particulate/light/noise) modern urban living creates.
Really Fantastic that you have done
Regarding the adverse effects of linoleic acid, a study found that leukocyte telomere length (LTL) is inversely associated with linoleic acid but not total fat. In other words, increased linoleic acid intake is associated with shortened lifespan.
Association of Marine Omega-3 Fatty Acid Levels With Telomeric Aging in Patients With Coronary Heart Disease
The higher the blood levels of fish-derived omega-3 acids in patients with coronary heart disease, the longer the telomeres.
Rob A. said:
"Peat shares the story of farmers a couple generations ago inducing hypothyroidism by feeding pigs corn and soy, increasing the PUFA content in their tissues, and allowing them to gain fat it seemed a plausible mechanism for humans as well. Any thoughts on that?"
That's not an argument. Soy has phyto-estrogens and goitrogens. There are multiple variables so we should not identify PUFAs as the cause.
Ironically, Peat suggests consuming a quart of milk per day plus loads of cheese, all which are loaded with estrogens.
"As there is already too much misinformation on the internet, I've taken the post down."
More respect to you, Stephan!
Over eating is a big problem we have to control it for remain healthier.
with regards: hemorrhoids treatment
Nice post is that baby fatness is a common problems today we need to take care of it in advance. It could become the cause of many problems as nail fungus
Thanks for the honest and great analysis on this post. I agree that omega 3 or oil from fish typically don't cause fat gain and it's proven by many researchers. Maybe obesity are comes form many factors like inherent from your parents, lack of exercise and eating not healthy foods. Some people are using essential oils because they do believe that it's also helpful to prevent any diseases especially when you are obese to stay healthy.
Linoleic acid n-6 intake inversely related to C-reactive protein levels in Japanese men: "The results suggest that increased intake of not only alpha-linolenic acid (n-3 PUFA) but also linoleic acid (n-6 PUFA) has a beneficial effect on systemic inflammation in men."
In Americans, "n-3 alpha-linolenic acid and n-6 cis-linoleic acid were not significantly related to the inflammatory markers" and "n-6 fatty acids do not inhibit the antiinflammatory effects of n-3 fatty acids and that the combination of both types of fatty acids is associated with the lowest levels of inflammation. "
I have not yet found any studies substantiating the claim that dietary linoleic acid increases inflammation, on the contrary the evidence I have so far found points the other direction. Of course that could change. But right now I think Willet might be right, LA is not inflammatory.
I certainly agree with the others that the effects of LA depend on the context of consumption. Foods rich in LA (e.g. nuts) have a different effect from extracted oils rich in LA, largely because of innumerable other phytochemicals present in the whole food.
"As there is already too much misinformation on the internet, I've taken the post down."
Well, you are aware that referencing any one of your posts is asking for a possible inconsistent entry (or book) in a near future, aren't you? I think —obviously my opinion— that you should suppose we —your readers— are able to think by ourselves and let those pages available with all the new information/interpretation/forward-references you wish available between brackets. I know it is your choice, I simply want to be certain that you are aware also of the bad side of your decision.
Maybe Willetts is right about LA's inflammatory propensity having been exaggerated. This study caught my eye recently.
::Elevated tissue AA (arachidonic acid) levels are believed to be positively associated with eicosanoid formation and risk for a variety of chronic diseases, including cardiovascular disease, cancer and inflammation. The literature expresses concern over the fact that increasing dietary LA can potentially enrich tissues with AA due to their metabolic link. The results of this study do not support this concern. Whereas AA levels in blood phospholipids is increased by GLA (gamma LA) or AA supplements, intervention studies bring no evidence to suggest that changes in dietary LA will modify tissue AA content in an adult population consuming a Western-type diet."
However, seed oils were not a substantial part of ancestral diets, as far as I can tell, so avoiding them seems unlikely to do harm (and may do some good for reasons yet to be fully explored and explained).
You may want to consider the HEXANE issue with Soy/sunflower/corn oils.
Hexane in the “natural” soyfoods industry
I am just wondering if the curve of the increasing subcutaneous pufa runs parallel with the increase in skin cancer. I think prof (em.) Homer Black did some work on that. If that is the case, I think I'd rather pursue a different hypothesis
I have been reading your blog and trying to understand all the developments in nutrition lately. Very informative. I was surprised to see this in the news, from India:
"When a recipe calls for ghee or butter, substitute it with canola or olive oil. It can help you shed weight and reduce liver fat, lipids and triglycerides, says a study."
So the Indians are being told to give up ghee for canola oil by their health experts? If this blog is correct, sounds like a recipe for disaster.
I found this article while searching the news for one I remembered that states fatty liver does not increase mortality.
Post a Comment