A good friend of mine recently emailed me about a guy named Brad Marshall who has a fascinating take on why we get fat, which he calls the ROS (Reactive Oxygen Species) Theory of Obesity. Inspired by the French Paradox, Brad claims that ROS signaling as a result of eating saturated fats is critical for satiety. Here’s what my friend had to say:
“..the more I read the more I’m pretty sure he is onto something. Joel Greene thinks Brad is crazy on this point but I have a feeling he will end up being right on most if not all of it. You can read more here: “The ROS Theory Of Obesity.“
Naturally, I just had to get in touch with this guy, and you can tune in to my podcast this Saturday to hear straight from the horse’s mouth how all of this could be true.
But for today’s article, Brad has graciously offered to fill you in on how The French Paradox compelled him to experiment with a “Croissant Diet,” (and the shocking results of his experiment), precisely what the ROS Theory of Obesity is, what all of this means for burning fat, and more.
The French Paradox
One often-overlooked reason you may be struggling with your weight, particularly in the abdominal area, could be the quality and ratio of the fats you’re eating.
Specifically, I’m referring to the ratio of long-chain saturated fats (which you should be eating more of) to unsaturated fats (which you should be eating less of as they will make your fat cells mildly, and reversibly, insulin resistant).
See, insulin sends a signal to fat cells to stop releasing fat to feed the rest of your body and takes your body out of fat-burning mode. Keeping your dietary fat highly saturated prevents insulin from locking away your fat stores, thereby enabling your fat cells to continue feeding your body. This results in greater satiation immediately after a meal and less hunger hours later.
Widely varying ratios of saturated to unsaturated fats can help explain the differences in obesity rates that can be seen around the world and over time.
A perfect example of this is the case of the French remaining lean while eating butter croissants. As an off and on again keto dieter—mostly off the last 10 years as I ran a farm and butcher shop and watched my waistline swell—I’ve always been amazed and astounded that the French were able to stay thin on a diet of butter croissants, baguettes with cheese, and bonbons. While this has changed somewhat in the last 20 years as their diets have vastly changed, urban Parisians in 1970 stayed slim while combining processed white flour and white sugar with butter. Conversely, the Middle Eastern diet combines white flour, white rice, and sugar with polyunsaturated vegetable oils, and today Kuwait is one of the few countries on earth with a higher obesity rate than America.
The following table is extracted from “The Anti-French Diet” based on food disappearance data from the Food and Agricultural Organization of the United Nations. Food Disappearance is the amount of food a country “goes through.” It does not account for waste so the caloric amounts might not be accurate, but the data gives a really good snapshot of the quality of the diet by illustrating the ratios of the foods that are eaten. I’ve included the US as a reference point.
Americans are in their own category with their jaw-dropping consumption of sugar and soybean oil, but the French and Kuwaiti diets look more similar than different: total calories, total calories from carbohydrates and alcohol, and sugar consumption are all about the same.
The main difference between the French diet of 1970 and the Kuwaiti diet of 2013 is that French sources of fat were predominately from meat and dairy whereas Kuwaiti sources of fat were predominately from polyunsaturated vegetable oils.
How could this affect their waistlines so dramatically?
The main argument I hear people say when I mention this is that, “the French walk a lot.” People in New York City walk a lot, too. I lived there for two years—10 blocks to the subway, take the subway to work, 6 blocks to work, walk 12 blocks round trip for lunch to that new hip spot, do it all again in the evening. It certainly didn’t prevent me from gaining weight, and when I go to the city now the obesity epidemic (while perhaps less severe than in other parts of the US) is on full display.
Needless to say, I think there’s more to it than just walking, which brings me to the ROS Theory Of Obesity.
The ROS Theory Of Obesity
In the winter of 2019, I realized that I had gained quite a bit of weight after neglecting my diet for some time. I struggled to take the weight off using my usual fail-safe keto diet, and soon realized it wasn’t as easy to lose weight at the age of 43 as it had been when I was 25. No shocker there… but it was frustrating.
So, I began re-reading the protons thread from a blog called Hyperlipid. Hyperlipid is brilliant, controversial, and extraordinarily dense with science and jargon. While absorbing it, I realized that the Protons Theory actually explains the French Paradox—why the French stay thin and the Kuwaitis become obese. Shortly thereafter, I published The ROS Theory of Obesity to explain why this was so.
Consider this graph, taken from my post “Long Chain Saturated Fat Causes Fat Loss in Mice,” that shows the amount of fat mass in mice fed a low-fat diet (10-week chow). As you can see, the low-fat chow diet mice maintained their fat mass over the ten weeks, the long-chain saturated fat stearic acid (long-chain saturated fat found in beef and chocolate) mice lost close to half of their body fat, and the mice fed oleic acid (long-chain monounsaturated fat found in olive oil and avocados) had a significant gain in body fat. What makes this more striking is that this experiment was done on wild-type (normal) mice, not some strain that had been artificially bred to become easily diabetic or obese.
The difference between the mice who lost body fat and the mice who gained body fat is a single double bond.
Stearic acid is an 18-carbon saturated fat found in foods such as beef suet and cocoa butter, and oleic acid is an 18-carbon monounsaturated fat famously found in olive oil (but even most beef fat is around half oleic acid). So the only difference between stearic acid and oleic acid is that single double bond. To learn more about all of the different fats, you can check out my primer The What’s, Why’s and Where’s of Fat.
A Billion-Year-Old Molecular Bottleneck In The Mitochondria
Mitochondria were at one point free-living organisms that were taken into other cells in an amoeba-like fashion, except that instead of being digested, they lived there. Mitochondria have their own DNA, and their job is to oxidize glucose and fat, with the ability to switch from burning one to the other.
Inside of the mitochondria is where the real action is, metabolically speaking:
If a cell is burning glucose, the glucose is broken into two molecules of pyruvate (which are shuttled into the mitochondria) by a process known as glycolysis.
If a cell is burning fats, they are broken down and shuttled into the mitochondria via a process known as beta-oxidation.
Ultimately, the pyruvate and the acetyl-CoA produced from beta-oxidation enter the Krebs cycle, AKA the citric acid cycle. In this cycle, electrons are steadily removed from carbon and hydrogen-rich molecules and passed through something called the electron transport chain, which is a series of protein complexes in the inner mitochondrial membrane. The electron transport chain turns the chemical energy stored in the carbohydrates and fats into adenosine triphosphate (ATP), which is used to power your muscles and other bodily activities.
The electrons are moved to the electron transport chain by two carrier molecules: nicotinamide adenine dinucleotide + hydrogen (NADH) and flavin adenine dinucleotide (FADH2). NADH hands its electrons off to Complex I (a very large enzyme catalyzing the first step of the mitochondrial electron transport chain) of the electron transport chain and FADH2 hands its electrons off to Complex II (transfers electrons from FADH2 onto Coenzyme Q).
Both Complex I and Complex II hand their electrons to Coenzyme Q, and this is the bottleneck. When ATP levels are high, and NADH and FADH2 are also both high, this bottleneck becomes overwhelmed with traffic and some of the electrons ping back out and combine with molecular oxygen (O2) to form a free radical called superoxide.
As I mentioned, one primary difference between burning glucose and fat is that glucose goes through glycolysis and the fat goes through beta-oxidation. During glycolysis, nearly all of the electrons are transported to the electron transport chain by NADH. Conversely, a saturated fat produces a molecule of FADH2 for each round of beta-oxidation, which happens one time for every two carbons in the fat. A saturated fat produces a much higher ratio of FADH2:NADH than does burning glucose. Whenever a fat has a double bond, no FADH2 is produced during that cycle of beta-oxidation. Therefore, the more unsaturated a fat is, the lower the ratio of FADH2:NADH is.
After a meal high in saturated fat, energy levels are high and ATP is high. In a cell that is burning saturated fat from said meal, FADH2 and NADH will both be trying to hand electrons off to Coenzyme Q. In this scenario, a significant number of electrons will ping back out of the electron transport chain and form superoxide. Superoxide is rapidly turned to hydrogen peroxide by a molecule called superoxide dismutase, which is a crucial part of your body’s inbuilt antioxidant system. It is literally the fastest enzyme that has ever been found.
Hydrogen peroxide is unique in that it is an oxidant, but it is not a free radical. Therefore, it is not a powerful enough oxidant to damage DNA or protein. It also is soluble in water but can pass through lipid (fat) membranes. So if the fuel burning in the mitochondria produces a lot of FADH2, a lot of superoxide is produced, which is rapidly converted to hydrogen peroxide, which can diffuse out of the mitochondrial membrane. These molecules, superoxide and hydrogen peroxide are known collectively as reactive oxygen species (ROS).
If any of what I just covered is unclear, remember that I’m going to be unpacking all of this with Ben this Saturday on his podcast, which you can find here once it goes live. The takeaway message though is that when your cells are burning saturated fat, there is an electron bottleneck that produces ROS. What is ROS?
ROS Is The Signal
Despite having somewhat of a bad rap, ROS are actually crucial molecules that send a signal to the cell, from the mitochondria, that the mitochondria are burning fat. Hydrogen peroxide does its job by oxidizing the sulfur atoms of specific cysteine amino acids of what are known as “redox-sensitive proteins.” These sulfur atoms are like light switches.
Hydrogen peroxide turns these genes on or off. Some of the genes turned on by hydrogen peroxide are transcription factors, which go into the nucleus and turn on genes that are involved in fat metabolism. The nucleus of the cell responds to what the mitochondria are doing, and the messenger is ROS.
In my post, Two Elegant Experiments Demonstrating That ROS Is the Signal, I discuss a study done on C elegans, which are tiny nematodes that live in the soil and eat exudate from plant roots, typically carbohydrates. Glucose is a signal to a C elegans worm that life is good, energy is plentiful, and they should go on reproducing and pass away. (They only live for about 18 days.) The study showed that if you deprive them of glucose, they will live an additional three days, which gives them longer to find a nice source of carbohydrates. During glucose restriction, they burn their own body fat, creating ROS, which is the signal to the organism that the mitochondria are burning fat, not glucose.
Here’s the crucial point about the C elegans: If you deprive them of glucose and give them an antioxidant, the antioxidant mops up the ROS, no message is delivered to the nucleus, the worm doesn’t know that glucose is in short supply, and therefore it doesn’t live the extra three days—because ROS is the signal.
Physiological Insulin Resistance
Some of the signaling proteins that are turned off by hydrogen peroxide include those in the insulin signaling pathway. When these proteins are turned off, insulin cannot signal as strongly.
This creates a mild, reversible form of insulin resistance in tissues that are burning saturated fat.
One of the main functions of insulin is to shut off lipolysis, which is the process by which fat cells release their stored energy as free fatty acids, or fuel for the rest of your body. Fat cells that are physiologically insulin resistant continue to release fat in the presence of insulin and therefore tend to get smaller in time.
For every double bond in a fat, one less molecule of FADH2 is formed. Animals can only create saturated or monounsaturated fat, which is stored as a blend of around 50:50 saturated to monounsaturated (along with a little polyunsaturated fat, which has to come from the diet and is usually less than 10% and under 3% in stored adipose tissue). The system is very tightly controlled to recognize this blend of fat. If mitochondria are burning pure monounsaturated fat, like in the mouse study above, the FADH2:NADH ratio is too low to elicit a large ROS response from the mitochondria.
Polyunsaturated fats, like those in soybean or corn oil, have a second double bond and produce even less FADH2 (and therefore less ROS) than do monounsaturated fats. The predominant polyunsaturated fat in modern diets is called linoleic acid. In a tightly controlled system where small changes in FADH2:NADH ratios can have huge effects on organismal energy balance as demonstrated by the feeding study in mice, introducing an evolutionarily novel fat such as linoleic acid into the system—which dramatically lowers the FADH2:NADH ratio—can really throw a wrench into the gears of your underlying metabolism.
France Vs. Kuwait Revisited
Let’s think again about the diet of lean French people in the 1970s and obese Kuwaitis in 2013. The primary difference between their diets was the FADH2:NADH ratio.
The French ate butter croissants, which forced their visceral fat to be physiologically insulin resistant and enabled them to burn their own stored body fat throughout the day.
The Kuwaitis, on the other hand, ate pastries made out of soybean oil, which caused their fat cells to become very insulin sensitive and respond to the insulin released from the meal by shutting down lipolysis. Three hours after a meal, their blood glucose would have returned to normal, but their free fatty acids would still be below pre-meal levels and hunger would ensue. They couldn’t continue to burn their stored body fat because their fat cells responded to insulin and were no longer releasing their stored energy to the rest of the body. They ended up with less circulating energy in their blood precisely because they ate. They had lots of stored energy in their fat cells, but they couldn’t access it because their fat cells had become insulin sensitive.
A great experiment done in Spain illustrates this very clearly: Groups of healthy volunteers were given an 800-calorie meal comprised mostly of fat, little starch, and only 160 calories from carbohydrates. The fats were either butter, olive oil, or vegetable oils. The only group that had the same amount of free fatty acids 3 hours after the meal was the group given butter (shown as black squares in the graph). Even eight hours after the meal, the group given butter had significantly more fat available from their own fat cells than did the groups whose adipocytes were insulin sensitive.
Another reason frequently given for why the French could remain thin is that they don’t snack, to which I reply, “Maybe they don’t snack because they’re not hungry.”
After considering all of these concepts, I started to wonder if I could actually lose weight by eating croissants, so I set off on creating my very own Croissant Diet.
Real croissants get upwards of 70% of their calories from fat, all from butter. I wanted my Croissant Diet to work, so I gamed the system by doping butter with stearic acid(the same long-chain saturated fat fed to the skinny mice) before making my own croissants. My stearic acid-enhanced butteroilwas VERY high in long-chain saturated fat, quite low in monounsaturated fat, and very low in polyunsaturated fat.
I started making croissants with the butteroil and eating exclusively croissants, black coffee, and red wine—it’s a diet based on French concepts, after all. Amazingly, I started losing inches off of my waist right away!
I noticed several things while eating croissants: First, they were intensely satiating. I couldn’t eat more than one and a half croissant sandwiches for dinner. For me, that’s not a lot of food! I’ve always struggled to get real satiation. I would usually go for thirds and then stop eating because my stomach hurt. This was different.
The other thing that happened was suddenly I could go longer between meals. I started skipping meals, sometimes by mistake. That is definitely something that never happened before. So I lost inches, found satiation, and went longer between meals despite the fact that I reintroduced white flour into my diet! (Admitedly not all that much as a croissant only has about 25-30g of carbohydrates, and I was eating less than three per day.)
One of the other areas of the body that expresses LPL, and therefore is targeted by dietary fat, is the hypothalamus, the gland in your brain that signals satiation. A key component of the satiety signal is ROS production in the hypothalamus. One of the reasons I was failing to feel true satiation earlier in life may have been that my dietary fat was insufficiently saturated, and I was unable to drive sufficient ROS production in my hypothalamus.
So What Types Of Fat Should You Eat?
The main premise of The Croissant Diet is that, when thinking about a weight loss or weight control diet (especially if you have excess abdominal fat), you should be thinking about fat quality in addition to macro ratios.
Independent from the amount of carbohydrates, fats, protein, and ethanol you consume is the FADH2:NADH ratio of your diet.
The table you’ll see below, excerpted from The Croissant Diet Specification, lists the ratio of long-chain saturated to unsaturated fats in a variety of fat sources. It doubly penalizes polyunsaturated fats. It also lists the amount of stearic acid in each fat. The best fats would have a high ratio and be high in stearic acid. The chart also shows that some of what many in the keto community consider to be “safe” fats—such as olive oil and nuts—actually do very poorly by this metric.
Another trojan horse for polyunsaturated fats is pork and chicken, especially in America. You are what you eat, and this is no different for pigs and chickens. In the US, what they routinely eat is corn—a relatively oily grain. Additionally, chickens are routinely given supplemental soybean oil in their ration only because it gets them to market a few days faster. Unfortunately, this is true whether or not the chicken is pastured, non-GMO, or organic. All of the feed mills have their diets designed by university-trained animal nutritionists who tell them they have to have soybean meal in their diets. Consider sourcing chicken and pork that is low in polyunsaturated fat!
If you are struggling with your weight, especially in the abdominal area, it would surely benefit you to give some serious thought to the types of fats you’re consuming.
Eating highly saturated fats will allow your fat cells to keep your body in fat-burning mode, releasing their stored energy to keep you running on your own stored energy all day.
And when you consider the fact that the traditional French way of consuming highly-saturated fats includes eating plenty of butter and chocolate, it’s a fairly easy diet to get behind.
To recap, here’s a list of fats, from best to worst:
How about you? Have you tried a highly-saturated fat diet? Let me know in the comments section below your successes or failures with fats, or if you have any questions, comments, or thoughts, and I’ll respond!