Polyphenols, Hormesis and Disease: Part I

What are Polyphenols?

Polyphenols are a diverse class of molecules containing multiple phenol rings. They are synthesized in large amounts by plants, certain fungi and a few animals, and serve many purposes, including defense against predators/infections, defense against sunlight damage and chemical oxidation, and coloration. The color of many fruits and vegetables, such as blueberries, eggplants, red potatoes and apples comes from polyphenols. Some familiar classes of polyphenols in the diet-health literature are flavonoids, isoflavonoids, anthocyanidins, and lignins.

The Case Against Polyphenols

Mainstream diet-health authorities seem pretty well convinced that dietary polyphenols are an important part of good health, due to their supposed antioxidant properties. In the past, I've been critical of the hypothesis. There are several reasons for it:

1. Polyphenols are often, but not always, defensive compounds that interfere with digestive processes, which is why they often taste bitter and/or astringent. Plant-eating animals including humans have evolved defensive strategies against polyphenol-rich foods, such as polyphenol-binding proteins in saliva (1).
2. Ingested polyphenols are poorly absorbed (2). The concentration in blood is low, and the concentration inside cells is probably considerably lower*. In contrast, essential antioxidant nutrients such as vitamins E and C are efficiently absorbed rather than excluded from the circulation.
   
3. Polyphenols that manage to cross the gut barrier are rapidly degraded by the liver, just like a variety of other foreign molecules, again suggesting that the body doesn't want them hanging around (2).
4. The most visible hypothesis of how polyphenols influence health is the idea that they are antioxidants, protecting against the ravages of reactive oxygen species. While many polyphenols are effective antioxidants at high concentrations in a test tube, I don't find it very plausible that the low and transient blood concentration of polyphenols achieved by eating polyphenol-rich foods makes a meaningful contribution to that person's overall antioxidant status, when compared to the relatively high concentrations of other antioxidants in blood (uric acid; vitamins C, E; ubiquinone) and particularly inside cells (SOD1/2, catalase, glutathione reductase, thioredoxin reductase, paraoxonase 1, etc.).
   
5. There are a number of studies showing that the antioxidant capacity of the blood increases after eating polyphenol-rich foods. These are often confounded by the fact that fructose (in fruit and some vegetables) and caffeine (in tea and coffee) can increase the blood level of uric acid, the blood's main water-soluble antioxidant. Drinking sugar water has the same effect (2).
6. Rodent studies showing that polyphenols improve health typically use massive doses that exceed what a person could consume eating food, and do not account for the possibility that the rodents may have been calorie restricted because their food tastes horrible.
   

The main point is that the body does not seem to "want" polyphenols in the circulation at any appreciable level, and therefore it gets rid of them pronto. Why? I think it's because the diversity and chemical structure of polyphenols makes them potentially bioactive-- they have a high probability of altering signaling pathways and enzyme activity, in the same manner as pharmaceutical drugs. It would not be a very smart evolutionary strategy to let plants (that often don't want you eating them) take the reins on your enzyme activity and signaling pathways. Also, at high enough concentrations polyphenols can be pro-oxidants, promoting excess production of free radicals, although the biological relevance of that may be questionable due to the concentrations required.

A Reappraisal

After reading more about polyphenols, and coming to understand that the prevailing hypothesis of why they work makes no sense, I decided that the whole thing is probably bunk: at best, specific polyphenols are protective in rodents at unnaturally high doses due to some drug-like effect. But-- I kept my finger on the pulse of the field just in case, and I began to notice that more sophisticated studies were emerging almost weekly that seemed to confirm that realistic amounts of certain polyphenol-rich foods (not just massive quantities of polyphenol extract) have protective effects against a variety of health problems. There are many such studies, and I won't attempt to review them comprehensively, but here are a few I've come across:

• Dr. David Grassi and colleagues showed that polyphenol-rich chocolate lowers blood pressure, improves insulin sensitivity and lowers LDL cholesterol in hypertensive and insulin resistant volunteers when compared with white chocolate (3). Although dark chocolate is also probably richer in magnesium, copper and other nutrients than white chocolate, the study is still intriguing.
• Dr. Christine Morand and colleagues showed that drinking orange juice every day lowers blood pressure and increases vascular reactivity in overweight volunteers, an effect that they were able to specifically attribute to the polyphenol hesperidin (4).
• Dr. F. Natella and colleagues showed that red wine prevents the increase in oxidized blood lipids (fats) that occurs after consuming a meal high in oxidized and potentially oxidizable fats (5).
• Several studies have shown that hibiscus tea lowers blood pressure in people with hypertension when consumed regularly (6, 7, 8). It also happens to be delicious.
  
• Dr. Arpita Basu and colleagues showed that blueberries lower blood pressure and oxidized LDL in men and women with metabolic syndrome (9).
• Animal studies have generally shown similar results. Dr. Xianli Wu and colleagues showed the blueberries potently inhibit atherosclerosis (hardening and thickening of the arteries that can lead to a heart attack) in a susceptible strain of mice (10). This effect was associated with a higher expression level of antioxidant enzymes in the vessel walls and other tissues.

Wait a minute... let's rewind. Eating blueberries caused mice to increase the expression level of their own antioxidant enzymes?? Why would that happen if blueberry polyphenols were themselves having a direct antioxidant effect? One would expect the opposite reaction if they were. What's going on here?

In the face of this accumulating evidence, I've had to reconsider my position on polyphenols. In the process, and through conversations with knowledgeable researchers in the polyphenol field, I encountered a different hypothesis that puts the puzzle pieces together nicely.


* Serum levels briefly enter the mid nM to low uM range, depending on the food (2). Compare that with the main serum antioxidants: ~200 uM for uric acid, ~100 uM for vitamin C, ~30 uM for vitamin E.

Tropical Plant Fats: Palm Oil

A Fatal Case of Nutritionism

The concept of 'nutritionism' was developed by Dr. Gyorgy Scrinis and popularized by the food writer Michael Pollan. It states that the health value of a food can be guessed by the sum of the nutrients it contains. Pollan argues, I think rightfully, that nutritionism is a reductionist philosophy that assumes we know more about food composition and the human body than we actually do. You can find varying degrees of this philosophy in most mainstream discussions of diet and health*.

One conspicuous way nutritionism manifests is in the idea that saturated fat is harmful. Any fat rich in saturated fatty acids is typically assumed to be unhealthy, regardless of its other constituents. There is also apparently no need to directly test that assumption, or even to look through the literature to see if the assumption has already been tested. In this manner, 'saturated' tropical plant fats such as palm oil and coconut oil have been labeled unhealthy, despite essentially no direct evidence that they're harmful. As we'll see, there is actually quite a bit of evidence, both indirect and direct, that their unrefined forms are not harmful and perhaps even beneficial.

Palm Oil and Heart Disease

Long-time readers may recall a post I wrote a while back titled Ischemic Heart Attacks: Disease of Civilization (1). I described a study from 1964 in which investigators looked for signs of heart attacks in thousands of consecutive autopsies in the US and Africa, among other places. They found virtually none in hearts from Nigeria and Uganda (3 non-fatal among more than 4,500 hearts), while Americans of the same age had very high rates (up to 1/3 of hearts).

What do they eat in Nigeria? Typical Nigerian food involves home-processed grains, starchy root vegetables, beans, fruit, vegetables, peanuts, red palm oil, and a bit of dairy, fish and meat**. The oil palm Elaeis guineensis originated in West Africa and remains one of the main dietary fats throughout the region.

To extract the oil, palm fruit are steamed, and the oily flesh is removed and pressed. It's similar to olive oil in that it is extracted gently from an oil-rich fruit, rather than harshly from an oil-poor seed (e.g., corn or soy oil). The oil that results is deep red and is perhaps the most nutrient-rich fat on the planet. The red color comes from carotenes, but red palm oil also contains a large amount of vitamin E (mostly tocotrienols), vitamin K1, coenzyme Q10 and assorted other fat-soluble constituents. This adds up to a very high concentration of fat-soluble antioxidants, which are needed to protect the fat from rancidity in hot and sunny West Africa. Some of these make it into the body when it's ingested, where they appear to protect the body's own fats from oxidation.

Mainstream nutrition authorities state that palm oil should be avoided due to the fact that it's approximately half saturated. This is actually one of the main reasons palm oil was replaced by hydrogenated seed oils in the processed food industry. Saturated fat raises blood cholesterol, which increases the risk of heart disease. Doesn't it? Let's see what the studies have to say.

Most of the studies were done using refined palm oil, unfortunately. Besides only being relevant to processed foods, this method also introduces a new variable because palm oil can be refined and oxidized to varying degrees. However, a few studies were done with red palm oil, and one even compared it to refined palm oil. Dr. Suzanna Scholtz and colleagues put 59 volunteers on diets predominating in sunflower oil, refined palm oil or red palm oil for 4 weeks. LDL cholesterol was not different between the sunflower oil and red palm oil groups, however the red palm oil group saw a significant increase in HDL. LDL and HDL both increased in the refined palm oil group relative to the sunflower oil group (2).

Although the evidence is conflicting, most studies have not been able to replicate the finding that refined palm oil increases LDL relative to less saturated oils (3, 4). This is consistent with studies in a variety of species showing that saturated fat generally doesn't raise LDL compared to monounsaturated fat in the long term, unless a large amount of purified cholesterol is added to the diet (5).

Investigators have also explored the ability of palm oil to promote atherosclerosis, or hardening and thickening of the arteries, in animals. Not only does palm oil not promote atherosclerosis relative to monounsaturated fats (e.g., olive oil), but in its unrefined state it actually protects against atherosclerosis (6, 7). A study in humans hinted at a possible explanation: compared to a monounsaturated oil***, palm oil greatly reduced oxidized LDL (8). As a matter of fact, I've never seen a dietary intervention reduce oxLDL to that degree (69%). oxLDL is a major risk factor for cardiovascular disease, and a much better predictor of risk than the typically measured LDL cholesterol (9). The paper didn't state whether or not the palm oil was refined. I suspect it was lightly refined, but still rich in vitamin E and CoQ10.

As I discussed in my recent interview with Jimmy Moore, atherosclerosis is only one factor in heart attack risk (10). Several other factors are also major determinants of risk: clotting tendency, plaque stability, and susceptibility to arrhythmia. Another factor that I haven't discussed is how resistant the heart muscle is to hypoxia, or loss of oxygen. If the coronary arteries are temporarily blocked-- a frequent occurrence in modern people-- the heart muscle can be damaged. Dietary factors determine the degree of damage that results. For example, in rodents, nitrites derived from green vegetables protect the heart from hypoxia damage (11). It turns out that red palm oil is also protective (12, 13). Red palm oil also protects against high blood pressure in rats, an effect attributed to its ability to reduce oxidative stress (14, 15).

Together, the evidence suggests that red palm oil does not contribute to heart disease risk, and in fact is likely to be protective. The benefits of red palm oil probably come mostly from its minor constituents, i.e. the substances besides its fatty acids. Several studies have shown that a red palm oil extract called palmvitee lowers serum lipids in humans (16, 17). The minor constituents are precisely what are removed during the refining process.

Palm Oil and the Immune System

Red palm oil also has beneficial effects on the immune system in rodents. It protects against bacterial infection when compared with soybean oil (18). It also protects against certain cancers, compared to other oils (19, 20). This may be in part due to its lower content of omega-6 linoleic acid (roughly 10%), and minor constituents.

The Verdict

Yet again, nutritionism has gotten itself into trouble by underestimating the biological complexity of a whole food. Rather than being harmful to human health, red palm oil, an ancient and delicious food, is likely to be protective. It's also one of the cheapest oils available worldwide, due to the oil palm's high productivity. It has a good shelf life and does not require refrigeration. Its strong, savory flavor goes well in stews, particularly meat stews. It isn't available in most grocery stores, but you can find it on the internet. Make sure not to confuse it with refined palm oil or palm kernel oil.


* The approach that Pollan and I favor is a simpler, more empirical one: eat foods that have successfully sustained healthy cultures.

** Some Nigerians are also pastoralists that subsist primarily on dairy.

*** High oleic sunflower oil, from a type of sunflower bred to be high in monounsaturated fat and low in linoleic acid. I think it's probably among the least harmful refined oils. I use it sometimes to make mayonnaise. It's often available in grocery stores, just check the label.

Nitrate: a Protective Factor in Leafy Greens

Cancer Link and Food Sources

Nitrate (NO3) is a molecule that has received a lot of bad press over the years. It was initially thought to promote digestive cancers, in part due to its ability to form carcinogens in the digestive tract. As it's used as a preservative in processed meats, and there is a link between processed meats and gastric cancer (1), nitrate was viewed with suspicion and a number of countries imposed strict limits on its use as a food additive.

But what if I told you that by far the greatest source of nitrate in the modern diet isn't processed meat-- but vegetables, particularly leafy greens (2)? And that the evidence specifically linking nitrate consumption to gastric cancer has largely failed to materialize? For example, one study found no difference in the incidence of gastric cancer between nitrate fertilizer plant workers and the general population (3). Most other studies in animals and humans have not supported the hypothesis that nitrate itself is carcinogenic (4, 5, 6). This, combined with recent findings on nitrate biology, has the experts singing a different tune in the last few years.

A New Example of Human Symbiosis

In 2003, Dr. K. Cosby and colleagues showed that nitrite (NO2; not the same as nitrate) dilates blood vessels in humans when infused into the blood (7). Investigators subsequently uncovered an amazing new example of human-bacteria symbiosis: dietary nitrate (NO3) is absorbed from the gut into the bloodstream and picked up by the salivary glands. It's then secreted into saliva, where oral bacteria use it as an energy source, converting it to nitrite (NO2). After swallowing, the nitrite is reabsorbed into the bloodstream (8). Humans and oral bacteria may have co-evolved to take advantage of this process. Antibacterial mouthwash prevents it.

Nitrate Protects the Cardiovascular System

In 2008, Dr. Andrew J. Webb and colleagues showed that nitrate in the form of 1/2 liter of beet juice (equivalent in volume to about 1.5 soda cans) substantially lowers blood pressure in healthy volunteers for over 24 hours. It also preserved blood vessel performance after brief oxygen deprivation, and reduced the tendency of the blood to clot (9). These are all changes that one would expect to protect against cardiovascular disease. Another group showed that in monkeys, the ability of nitrite to lower blood pressure did not diminish after two weeks, showing that the animals did not develop a tolerance to it on this timescale (10).

Subsequent studies showed that dietary nitrite reduces blood vessel dysfunction and inflammation (CRP) in cholesterol-fed mice (11). Low doses of nitrite also dramatically reduce tissue death in the hearts of mice exposed to conditions mimicking a heart attack, as well as protecting other tissues against oxygen deprivation damage (12). The doses used in this study were the equivalent of a human eating a large serving (100 g; roughly 1/4 lb) of lettuce or spinach.

Mechanism

Nitrite is thought to protect the cardiovascular system by serving as a precursor for nitric oxide (NO), one of the most potent anti-inflammatory and blood vessel-dilating compounds in the body (13). A decrease in blood vessel nitric oxide is probably one of the mechanisms of diet-induced atherosclerosis and increased clotting tendency, and it is likely an early consequence of eating a poor diet (14).

The Long View

Leafy greens were one of the "protective foods" emphasized by the nutrition giant Sir Edward Mellanby (15), along with eggs and high-quality full-fat dairy. There are many reasons to believe greens are an excellent contribution to the human diet, and what researchers have recently learned about nitrate biology certainly reinforces that notion. Leafy greens may be particularly useful for the prevention and reversal of cardiovascular disease, but are likely to have positive effects on other organ systems both in health and disease. It's ironic that a molecule suspected to be the harmful factor in processed meats is turning out to be one of the major protective factors in vegetables.

Copper and Cardiovascular Disease

In 1942, Dr. H. W. Bennetts dissected 21 cattle known to have died of "falling disease". This was the name given to the sudden, inexplicable death that struck herds of cattle in certain regions of Australia. Dr. Bennett believed the disease was linked to copper deficiency. He found that 19 of the 21 cattle had abnormal hearts, showing atrophy and abnormal connective tissue infiltration (fibrosis) of the heart muscle (1).

In 1963, Dr. W. F. Coulson and colleagues found that 22 of 33 experimental copper-deficient pigs died of cardiovascular disease. 11 of 33 died of coronary heart disease, the quintessential modern human cardiovascular disease. Pigs on a severely copper-deficient diet showed weakened and ruptured arteries (aneurysms), while moderately deficient pigs "survived with scarred vessels but demonstrated a tendency toward premature atherosclerosis" including foam cell accumulation (2). Also in 1963, Dr. C. R. Ball and colleagues published a paper describing blood clots in the heart and coronary arteries, heart muscle degeneration, ventricular calcification and early death in mice fed a lard-rich diet (3).

This is where Dr. Leslie M. Klevay enters the story. Dr. Klevay suspected that Ball's mice had suffered from copper deficiency, and decided to test the hypothesis. He replicated Ball's experiment to the letter, using the same strain of mice and the same diet. Like Ball, he observed abnormal clotting in the heart, degeneration and enlargement of the heart muscle, and early death. He also showed by electrocardiogram that the hearts of the copper-deficient mice were often contracting abnormally (arrhythmia).

But then the coup de grace: he prevented these symptoms by supplementing the drinking water of a second group of mice with copper (4). In the words of Dr. Klevay: "copper was an antidote to fat intoxication" (5). I believe this was his tongue-in-cheek way of saying that the symptoms had been misdiagnosed by Ball as due to dietary fat, when in fact they were due to a lack of copper.

Since this time, a number of papers have been published on the relationship between copper intake and cardiovascular disease in animals, including several showing that copper supplementation prevents atherosclerosis in one of the most commonly used animal models of cardiovascular disease (6, 7, 8). Copper supplementation also corrects abnormal heart enlargement-- called hypertrophic cardiomyopathy-- and heart failure due to high blood pressure in mice (9).

For more than three decades, Dr. Klevay has been a champion of the copper deficiency theory of cardiovascular disease. According to him, copper deficiency is the only single intervention that has caused the full spectrum of human cardiovascular disease in animals, including:

• Heart attacks (myocardial infarction)
  
• Blood clots in the coronary arteries and heart
• Fibrous atherosclerosis including smooth muscle proliferation
• Unstable blood vessel plaque
  
• Foam cell accumulation and fatty streaks
  
• Calcification of heart tissues
  
• Aneurysms (ruptured vessels)
• Abnormal electrocardiograms
• High cholesterol
• High blood pressure
  

If this theory is so important, why have most people never heard of it? I believe there are at least three reasons. The first is that the emergence of the copper deficiency theory coincided with the rise of the diet-heart hypothesis, whereby saturated fat causes heart attacks by raising blood cholesterol. Bolstered by some encouraging findings and zealous personalities, this theory took the Western medical world by storm, for decades dominating all other theories in the medical literature and public health efforts. My opinions on the diet-heart hypothesis aside, the two theories are not mutually exclusive.

The second reason you may not have heard of the theory is due to a lab assay called copper-mediated LDL oxidation. Researchers take LDL particles (from blood, the same ones the doctor measures as part of a cholesterol test) and expose them to a high concentration of copper in a test tube. Free copper ions are oxidants, and the researchers then measure the amount of time it takes the LDL to oxidize. I find this assay tiresome, because studies have shown that the amount of time it takes copper to oxidize LDL in a test tube doesn't predict how much oxidized LDL you'll actually find in the bloodstream of the person you took the LDL from (10, 11).

In other words, it's an assay that has little bearing on real life. But researchers like it because for some odd reason, feeding a person saturated fat causes their LDL to be oxidized more rapidly by copper in a test tube, even though that's not the case in the actual bloodstream (12). Guess which result got emphasized?

The fact that copper is such an efficient oxidant has led some researchers to propose that copper oxidizes LDL in human blood, and therefore dietary copper may contribute to heart disease (oxidized LDL is a central player in heart disease-- read more here). The problem with this theory is that there are virtually zero free copper ions in human serum. Then there's the fact that supplementing humans with copper actually reduces the susceptibility of red blood cells to oxidation (by copper in a test tube, unfortunately), which is difficult to reconcile with the idea that dietary copper increases oxidative stress in the blood (13).

The third reason you may never have heard of the theory is more problematic. Several studies have found that a higher level copper in the blood correlates with a higher risk of heart attack (14, 15). At this point, I could hang up my hat, and declare the animal experiments irrelevant to humans. But let's dig deeper.

Nutrient status is sometimes a slippery thing to measure. As it turns out, serum copper isn't a good marker of copper status. In a 4-month trial of copper depletion in humans, blood copper stayed stable, while the activity of copper-dependent enzymes in the blood declined (16). These include the important copper-dependent antioxidant, superoxide dismutase. As a side note, lysyl oxidase is another copper-dependent enzyme that cross-links the important structural proteins collagen and elastin in the artery wall, potentially explaining some of the vascular consequences of copper deficiency. Clotting factor VIII increased dramatically during copper depletion, perhaps predicting an increased tendency to clot. Even more troubling, three of the 12 women developed heart problems during the trial, which the authors felt was unusual:

We observed a significant increase over control values in the number of ventricular premature discharges (VPDs) in three women after 21, 63, and 91 d of consuming the low-copper diet; one was subsequently diagnosed as having a second-degree heart block.

In another human copper restriction trial, 11 weeks of modest copper restriction coincided with heart trouble in 4 out of 23 subjects, including one heart attack (17):

In the history of conducting numerous human studies at the Beltsville Human Nutrition Research Center involving participation by 337 subjects, there had previously been no instances of any health problem related to heart function. During the 11 wk of the present study in which the copper density of the diets fed the subjects was reduced from the pretest level of 0.57 mg/ 1000 kcal to 0.36 mg/1000 kcal, 4 out of 23 subjects were diagnosed as having heart-related abnormalities.

The other reason to be skeptical of the association between blood copper and heart attack risk is that inflammation increases copper in the blood (18, 19). Blood copper level correlates strongly with the marker of inflammation C-reactive protein (CRP) in humans, yet substantially increasing copper intake doesn't increase CRP (20, 21). This suggests that elevated blood copper is likely a symptom of inflammation, rather than its cause, and presents an explanation for the association between blood copper level and heart attack risk.

Only a few studies have looked at the relationship between more accurate markers of copper status and cardiovascular disease in humans. Leukocyte copper status, a marker of tissue status, is lower in people with cardiovascular disease (22, 23). People who die of heart attacks generally have less copper in their hearts than people who die of other causes, although this could be an effect rather than a cause of the heart attack (24). Overall, I find the human data lacking. I'd like to see more studies examining liver copper status in relation to cardiovascular disease, as the liver is the main storage organ for copper.

According to a 2001 study, the majority of Americans may have copper intakes below the USDA recommended daily allowance (25), many substantially so. This problem is exacerbated by the fact that copper levels in food have declined in industrial nations over the course of the 20th century, something I'll discuss in the next post.

Eicosanoids, Fatty Liver and Insulin Resistance

I have to take a brief intermission from the heart disease series to write about a very important paper I just read in the journal Obesity, "COX-2-mediated Inflammation in Fat is Crucial for Obesity-linked Insulin Resistance and Fatty Liver". It's actually related to cardiovascular disease, although indirectly.

First, some background. Polyunsaturated fatty acids (PUFA) come mostly from omega-6 and omega-3 sources. Omega-6 and omega-3 are precursors to eicosanoids, a large and poorly understood class of signaling molecules that play a role in basically everything. Eicosanoids are either omega-6-derived or omega-3-derived. Omega-6 and omega-3 compete for the enzymes that convert PUFA into eicosanoids. Therefore, the ratio of omega-6 to omega-3 in tissues (related to the ratio in the diet) determines the ratio of omega-6-derived eicosanoids to omega-3-derived eicosanoids.

Omega-6 eicosanoids are very potent and play a central role in inflammation. They aren't "bad", in fact they're essential, but an excess of them is probably not good. Omega-3 eicosanoids are generally less potent, less inflammatory, and tend to participate in long-term repair processes. So in sum, the ratio of omega-6 to omega-3 in the diet will determine the potency and quality of eicosanoid signaling, which will determine an animal's susceptibility to inflammation-mediated disorders.

One of the key enzymes in the pathway from PUFA to eicosanoids (specifically, a subset of them called prostanoids) is cyclooxygenase (COX). COX-1 is expressed all the time and serves a "housekeeping" function, while COX-2 is induced by cellular stressors and contributes to the the formation of inflammatory eicosanoids. Non-steroidal anti-inflammatory drugs (NSAIDs) such as aspirin and ibuprofen inhibit COX enzymes, which is why they are effective against inflammatory problems like pain and fever. They are also used as a preventive measure against cardiovascular disease. Basically, they reduce the excessive inflammatory signaling promoted by a diet with a poor omega-6:3 balance. You wouldn't need to inhibit COX if it were producing the proper balance of eicosanoids to begin with.

Dr. Kuang-Chung Shih's group at the Department of Internal Medicine in Taipei placed rats on five different diets:

1. A control diet, eating normal low-fat rat chow.
2. A "high-fat diet", in which 45% of calories came from a combination of industrial lard and soybean oil, and 17% of calories came from sucrose*.
3. A "high-fat diet" (same as above), plus the COX-2 inhibitor celecoxib (Celebrex).
4. A "high-fat diet" (same as above), plus the COX-2 inhibitor mesulid.
5. An energy-restricted "high-fat diet".

The "high-fat diets", besides being high in sucrose (table sugar), also presumably had a poor omega-6:3 ratio, in the neighborhood of 10:1 or possibly higher. Weight and fat mass in rats and humans increases with increasing omega-6 in the diet, and also increases with a high 6:3 ratio. I wrote about that here. Rats eating the high-fat diets (groups 2- 4) gained weight as expected**.

Rats in group 2 not only gained weight, they also experienced increased fasting glucose, leptin, insulin, triglycerides, blood pressure and a massive decline in insulin sensitivity (seven-fold relative to group 1). Rats in groups 3 and 4 gained weight, but saw much less of a deterioration in insulin and leptin sensitivity, and blood pressure. Group 2 also developed fatty liver, which was attenuated in groups 3 and 4. If you're interested, group 5 (energy restricted high-fat) was similar to groups 3 and 4 on pretty much everything, including insulin sensitivity.

So there you have it folks: direct evidence that insulin resistance, leptin resistance, high blood pressure and fatty liver are mediated by excessive inflammatory eicosanoid signaling. I wrote about something similar before when I reviewed a paper showing that fish oil reverses many of the consequences of a high-vegetable oil, high-sugar diet in rats. I also reviewed two papers showing that in pigs and rats, a high omega-6:3 ratio promotes inflammation (mediated by COX-2) and lipid peroxidation in the heart. Are you going to quench the fire by taking drugs, or by reducing your intake of omega-6 and ensuring an adequate intake of omega-3?

*Of course, they didn't mention the sucrose in the methods section. I had to go digging around for the diet's composition. This is typical of papers on "high-fat diets". They load them up with sugar, and blame everything on the fat. This kind of shenanigans wouldn't fly in a self-respecting field, but it's typical of nutrition-health papers.

**Rats gain fat mass when fed a high-fat diet (even if it's not loaded with sugar), although when the fat is butter or coconut oil, they gain less than if it's vegetable oil. But humans don't gain weight on a high-fat diet (i.e. low-carb diet); to the contrary. What's the difference? It may have to do with the fact that rats eat more calories when they have ad libitum access to high-fat food, while humans don't. In fact, most low-carbohydrate diet trials indicate that participants spontaneously reduce their caloric intake when eating high-fat food.

The Tokelau Island Migrant Study: Cholesterol and Cardiovascular Health

Let's get right to the meat of this study. It's a direct test of the idea that saturated fat is a cause of cardiovascular disease. If you were to design the perfect experiment to determine if saturated fat causes heart disease, and ethics were not a concern, how would you do it? You would stuff one group of people with as much saturated fat as they would eat for their entire lives, while feeding far less to a genetically identical group. Ideally, you would keep everything else about the diet and lifestyle the same. Then, you would measure some marker of cardiovascular disease, or even better, count actual heart attacks.

The Tokelau Island Migrant study isn't a perfect experiment, but it's about as close as we're going to get. Tokelauans traditionally obtained 40-50% of their calories from saturated fat, in the form of coconut meat. That's more than any other group I'm aware of, even topping the roughly 33% that the Masai get from their extremely fatty Zebu milk.

So are the Tokelauans dropping like flies of cardiovascular disease? I think most of the readers of this blog already know the answer to that question. I don't have access to the best data of all: actual heart attack incidence data. But we do have some telltale markers. In 1971-1982, researchers collected data from Tokelau and Tokelauan migrants to New Zealand on cholesterol levels, blood pressure and electrocardiogram (ECG) readings.

The Tokelauan diet, as I've described in detail in previous posts, is traditionally based on coconut, fish, starchy tubers and fruit. By 1982, their diet also contained a significant amount of imported flour and sugar. Migrants to New Zealand had a much more varied diet that was also more typically Western: more carbohydrate, coming chiefly from wheat, sugar and potatoes; more processed sweet foods and drinks; more red meat; more vegetables; more dairy and eggs. Sugar intake was 13 percent of calories, compared to 8 percent on Tokelau. Saturated fat intake in NZ was half of what it was on Tokelau, while total fat intake was similar. Polyunsaturated fat intake was higher in NZ, 4% as opposed to 2% in Tokelau. I don't have data to back this up, but I think it's likely that the n-6:n-3 ratio increased upon migration.

Blood pressure did not change significantly over time in Tokelau from 1971 to 1982, if anything it actually declined slightly. It was consistently higher in NZ than in Tokelau at all timepoints. Men were roughly three times more likely to be hypertensive in NZ than on Tokelau at all timepoints (4.0% vs. 12.0% in the early 1970s). Women were about twice as likely to be hypertensive (8.1% vs. 15.0%).

On to cholesterol. Total cholesterol in male Tokelauans was a bit lower on average than in New Zealand, but neither was particularly elevated (182 vs. 199 mg/dL). LDL was also a bit higher in NZ males (119 vs. 132 mg/dL). Get these guys on Lipitor!! Triglycerides were lower in Tokelauan men than in NZ (80 vs. 114 mg/dL). There were no differences in total cholesterol, LDL cholesterol or triglycerides between Tokelauan and NZ women.

These data would make Dr. Uffe Ravnskov smile (actually I'm sure he's aware of them). Much of the hoopla surrounding saturated fat is due to the fact that in controlled clinical trials, it seems to elevate blood cholesterol (by elevating both LDL and HDL). What Dr. Ravnskov and others have pointed out is that the correlation between saturated fat intake and blood cholesterol is weak, and in any case, so is the correlation between blood cholesterol and cardiovascular disease. This study lends support to the idea that saturated fat is not a major determinant of total cholesterol or LDL.

But does it cause heart attacks? The best data I have from this study are ECG readings. These use electrodes to monitor the electrical activity of the heart. There are certain ECG patterns that suggest that a person has had a heart attack (Minnesota codes 1-1 and 1-2). The data I am going to present here are all age-standardized, meaning they are comparing between groups of the same age. On Tokelau in 1982, 0.0% of men 40-69 years old showed ECG readings that indicated a probable past heart attack. In NZ in 1980-81, 1.0% of men 40-69 years old showed the same ECG readings. In Tecumseh U.S.A. in 1965, 3.5% of men 40-69 years old showed the same ECG pattern. I don't have data for women.

These data don't prove that no one ever has a heart attack on Tokelau. They do sometimes, and they also have strokes (at least in modern times). But they do allow us to compare in quantitative terms between genetically similar people living in two different environments.

This is consistent with what has been observed on Kitava and other traditional Pacific island cultures: a vanishingly small incidence of cardiovascular disease while they retain their traditional diet and lifestyle (and sometimes even when some processed Western food has been introduced). When diets and lifestyles become modern, there is invariably a rise in the incidence of chronic disease.

I don't believe that saturated fat contributes to cardiovascular disease. The best data available have never supported that hypothesis, even from the very beginning. The Tokelau Island Migrant study, among many others, should have put it out of its misery long ago. Tokelau underlines the fact that the most important determinant of health is a diet based on whole, natural foods that are familiar to the human metabolism, prepared in traditional ways that maximize their digestibility and nutritional value.

Unless otherwise noted, the data in this post are from the book Migration and Health in a Small Society: the Case of Tokelau.

Health is Multi-Factorial

Thanks to commenter Brock for pointing me to this very interesting paper, "Effects of fish oil on hypertension, plasma lipids, and tumor necrosis factor-alpha in rats with sucrose-induced metabolic syndrome". As we know, sugar gives rats metabolic syndrome when it's added to regular rat chow, probably the same thing it does to humans when added to a processed food diet.

One thing has always puzzled me about sugar. It doesn't appear to cause major metabolic problems when added to an otherwise healthy diet, yet it wreaks havoc in other contexts. One example of the former situation is the Kuna, who are part hunter-gatherer, part agricultural. They eat a lot of refined sugar, but in the context of chocolate, coconut, fish, plantains, root vegetables and limited grains and beans, they are relatively healthy. Perhaps not quite on the same level as hunter-gatherer groups, but healthier than the average modernized person from the point of view of the diseases of civilization.

This paper really sheds light on the matter. The researchers gave a large group of rats access to drinking water containing 30% sucrose, in addition to their normal rat chow, for 21 weeks. The rats drank 4/5 of their calories in the form of sugar water. There's no doubt that this is an extreme treatment. They subsequently developed metabolic syndrome, including abdominal obesity, elevated blood pressure, elevated fasting insulin, elevated triglycerides, elevated total cholesterol and LDL, lowered HDL, greatly increased serum uric acid, greatly elevated liver enzymes suggestive of liver damage, and increased tumor necrosis factor-alpha (TNF-alpha). TNF-alpha is a hormone secreted by visceral (abdominal) fat tissue that may play a role in promoting insulin resistance.

After this initial 12-week treatment, they divided the metabolic syndrome rats into two groups:

• One that continued the sugar treatment, along with a diet enriched in corn and canola oil (increased omega-6).
• A second that continued the sugar treatment, along with a diet enriched in fish oil (increased omega-3).

The two diets contained the same total amount of polyunsaturated fat (PUFA), but had very different omega-6 : omega-3 ratios. The first had a ratio of 9.3 (still better than the average American), while the second had a ratio of 0.02, with most of the omega-3 in the second group coming from EPA and DHA (long-chain, animal omega-3s). The second diet also contained four times as much saturated fat as the first, mostly in the form of palmitic acid.

Compared to the vegetable oil group, the fish oil group had lower fasting insulin, lower blood pressure, lower triglycerides, lower cholesterol, and lower LDL. As a matter of fact, the fish oil group looked as good or better on all these parameters than a non-sugar fed control group receiving the extra vegetable oil alone (although the control group isn't perfect because it inevitably ate more vegetable oil-containing chow to make up for the calories it wasn't consuming in sugar). The only things reducing vegetable oil and increasing fish oil didn't fix were the weight and the elevated TNF-alpha, although they didn't report the level of liver enzymes in these groups. The TNF-alpha finding is not surprising, since it's secreted by visceral fat, which did not decrease in the fish oil group.

I think this is a powerful result. It may have been done in rats, but the evidence is there for a similar mechanism in humans. The Kuna have a very favorable omega-6 : omega-3 ratio, with most of their fat coming from highly saturated coconut and cocoa. This may protect them from their high sugar intake. The Kitavans also have a very favorable omega-6 : omega-3 ratio, with most of their fat coming from coconuts and fish. They don't eat refined sugar, but they do eat a tremendous amount of starch and a generous amount of fruit.

The paper also suggests that the metabolic syndrome is largely reversible.

I believe that both excessive sugar and excessive omega-6 from modern vegetable oils are a problem individually. But if you want to have a much bigger problem, try combining them!

How to Give a Rat Metabolic Syndrome

I was doing my usual journal rounds today when I came across an article in the American Journal of Hypertension that caught my eye. It's called "Metabolic Syndrome: Comparison of the Two Commonly Used Animal Models." Metabolic syndrome is a cluster of symptoms including large waist circumference, elevated triglycerides, elevated blood pressure, and insulin resistance. It's the quintissential modern metabolic disorder, and it affects 24% of Americans (NHANES III). So what are the two most commonly used animal models of metabolic syndrome?

• A strain called the spontaneously hypertensive rat (SHR), fed a high-sucrose (table sugar, 50% fructose) diet.
• Sprague-Dawley (generic lab strain) rats fed a high-fructose diet.

When fed sugar, these rats develop insulin resistance, impaired glucose tolerance, elevated triglycerides and hypertension. Fructose causes leptin resistance in rats. Leptin resistance causes metabolic syndrome in rats. These studies trace a line directly from sugar to the metabolic syndrome.

On to humans. Total sugar and fructose consumption have been increasing in the U.S. in recent decades, along with metabolic syndrome. I think the average numbers may hide some important information, because there is a fraction of the population that consumes far more than the average amount of sugar through soda. Leptin resistance seems to be central to the metabolic syndrome, and typically precedes the other symptoms. The evidence suggests that the rat research on metabolic syndrome is applicable to humans.

I don't think sugar acts alone in causing the metabolic syndrome in humans. I believe the liver is a central player in the disorder, as many of the markers used to diagnose it are measures of processes that occur in the liver (triglyceride synthesis, glucose and insulin disposal). Insulin resistance in the liver is sufficient to cause many of the hallmarks of the metabolic syndrome in mice. The fructose portion of sugar and high-linoleic (omega-6) vegetable oils act synergistically to cause liver dysfunction in rats and probably humans.

I also believe wheat contributes to the process, perhaps through its ability to cause hyperphagia (overeating) or intestinal damage. So we're back to the three big killers in the modern diet:

• Refined vegetable oils
• Sugar
• Wheat
  

Masai and Atherosclerosis

I've been digging deeper into the health of the Masai lately. A commenter on Chris's blog pointed me to a 1972 paper showing that the Masai have atherosclerosis, or hardening of the arteries. This interested me so I got my hands on the full text, along with a few others from the same time period. What I found is nothing short of fascinating.

First, some background. Traditional Masai in Kenya and Tanzania are pastoralists, subsisting on fermented cow's milk, meat and blood, as well as traded food in modern times. They rarely eat fresh vegetables. Contrary to popular belief, they are a genetically diverse population, due to the custom of abducting women from neighboring tribes. Many of these tribes are agriculturalists. From Mann et al: "The genetic argument is worthless". This will be important to keep in mind as we interpret the data.

At approximately 14 years old, Masai men are inducted into the warrior class, and are called Muran. For the next 15-20 years, tradition dictates that they eat a diet composed exclusively of cow's milk, meat and blood. Milk is the primary food. Masai cows are not like wimpy American cows, however. Their milk contains almost twice the fat of American cows, more protein, more cholesterol and less lactose. Thus, Muran eat an estimated 3,000 calories per day, 2/3 of which comes from fat. Here is the reference for all this. Milk fat is about 50% saturated. That means the Muran gets 33% of his calories from saturated fat. This population eats more saturated fat than any other I'm aware of.

How's their cholesterol? Remarkably low. Their total serum cholesterol is about half the average American's. I haven't found any studies that broke it down further than total cholesterol. Their blood pressure is also low, and hypertension is rare. Overweight is practically nonexistent. Their electrocardiogram readings show no signs of heart disease. They have exceptionally good endurance, but their grip strength is significantly weaker than Americans of African descent. Two groups undertook autopsies of male Masai to look for artery disease.

The first study, published in 1970, examined 10 males, 7 of which were over 40 years old. They found very little evidence of atherosclerosis, even in individuals over 60. The second study, which is often used as evidence against a high-fat diet, was much more thorough and far more interesting. Mann et al. autopsied 50 Masai men, aged 10 to 65. The single most represented age group was 50-59 years old, at 13 individuals. They found no evidence of myocardial infarction (heart attack) in any of the 50 hearts. What they did find, however, was coronary artery disease. Here's a figure showing the prevalence of "aortic fibrosis", a type of atherosclerotic lesion:


It looks almost binary, doesn't it? What could be causing the dramatic jump in atherosclerosis at age 40? Here's another figure, of total cholesterol (top) and "sudanophilia" (fatty streaks in the arteries, bottom). Note that the Muran period is superimposed (top).


There's clearly a pattern here. Either the Masai men are eating nothing but milk, meat and blood and they're nearly free from atherosclerosis, or they're eating however they please and they have as much atherosclerosis as the average American. There doesn't seem to be much in between.

Here's a quote from the paper that sums it up well:

We believe... that the Muran escapes some noxious dietary agent for a time. Obviously, this is neither animal fat nor cholesterol. The old and the young Masai do have access to such processed staples as flour, sugar, confections and shortenings through the Indian dukas scattered about Masailand. These foods could carry the hypothetical agent."

I know this blog is starting to sound like a broken record, but I'll say it again: you can eat a wide variety of foods and be healthy, except industrial grain products (particularly wheat), sugar, industrial vegetable oil and other processed food. The Masai are just one more example of a group that's healthy when eating a traditional diet.

What to do if Your Study Contradicts Conventional Wisdom

I just read a recent paper from the British Journal of Sports Medicine, "Daily Energy Expenditure and Cardiovascular Disease Risk in Masai, Ruran and Urban Bantu Tanzanians". The study caught my eye because I think we have a lot to learn from healthy traditionally-living populations worldwide.

The Masai have a very unique diet consisting almost exclusively of whole cow's milk, cow's blood and meat. As you might imagine, they eat a lot of fat, a lot of saturated fat and a modest amount of carbohydrate (from lactose). They also have low total cholesterol, low blood pressure, and virtually no overweight. They have been a thorn in the side of the lipid hypothesis for a long time.

The Bantu are an agricultural population that traditionally eat a diet low in fat and high in carbohydrate. Their staples are root vegetables, corn, beans, fish and wild game. The paper also describes a group of urban Bantu, which eats a diet intermediate in fat and carbohydrate. Incidentally, the investigators describe it as a "high-fat diet", despite the fact that the percentage fat is about the same as what Americans and Europeans eat, shamelessly exposing their bias.

The investigators recorded the three groups' diets, activity levels, physical characteristics and various markers of cardiovascular disease risk. Here's what they found: only 3% of Masai were obese, compared to 12% of rural Bantu and 34% of urban Bantu (they'd fit right in here!). The Masai, despite smoking like chimneys, had generally lower CVD risk factors than the other two populations, with the urban Bantu being significantly worse off than the rural Bantu.

Overall, the Masai came out looking really good, with the rural Bantu not too far behind. The urban Bantu look almost as bad as Americans. How do we make sense of these two conflicting facts? 1) The urban Bantu eat an amount of fat and saturated fat that's right in the middle of what the Masai and the rural Bantu eat, yet they seem the most likely to keel over spontaneously. 2) Saturated fat KILLS!! Answer: keep digging until you find something else to blame your results on.

They certainly did find something, and it's the reason the study was published in the British Journal of Sports Medicine rather than the American Journal of Clinical Nutrition. The Masai exercise more than either of the other two groups. I don't have too much trouble believing that. However, the authors used a dirty trick to augment their result: they normalized calorie expenditure to body weight. They present their data as kcal/kg/day. In other words, the fatter you are, the lower your apparent energy expenditure! It makes no sense to me. But it does inflate the apparent exercise of the Masai, simply because of the fact that they're thinner than the other two groups.

Due to this unscrupulous number massaging, here's what they got (data re-plotted by me):


I'm going to try to un-massage the data. Here's what it looks like when I factor bodyweight out of the equation. Calories expended (above resting metabolic rate) is on the Y-axis. The bars look a bit closer together...



Here's what it looks like when you add back resting metabolic rate. I assumed 1500 kcal/day. This graph is an approximation of their total energy expenditure per day:



Hmm, the differences keep getting smaller, don't they? I'm not challenging the fact that the Masai exercise more than the other two groups, but I do have a problem with this kind of manipulation of the data in misleading ways.

Their conclusion is that exercise is protecting the Masai from the deadly saturated fats in their diet. A more parsimonious explanation is that saturated fat per se doesn't cause heart disease. It's also more consistent with other healthy cultures that ate high-fat diets like the Inuit, certain Australian aboriginal groups, and some American Indian groups. It's also consistent with the avalanche of recent trials of low-carbohydrate diets, in which people consistently see improvements in weight, blood pressure, and CVD markers, among other things. Not that I have much faith in blood lipid markers of CVD.

My conclusion, from this study and others, is that macronutrients don't determine how healthy a diet is. The specific foods that compose the diet do. The rural Masai are healthy on a high-fat diet, the rural Bantu are fairly healthy on a low-fat, high carbohydrate diet. Only the urban Bantu show a pattern really consistent with the "disease of civilization", despite a daily energy expenditure very similar to the rural Bantu. They're unhealthy because they eat too much processed food: processed vegetable oil, processed grain products, refined sugar.

Thanks to kevinzim for the CC photo

Scientist Discovers that Only Pills can Control Hypertension

I went to a presentation today by a prominent hypertension researcher. His talk began with a slide that had two pictures side-by-side: one of the late fitness advocate Jim Fixx, and the other of Winston Churchill. Fixx was a marathon runner, while Churchill was inactive, overweight and had a famous appetite. Fixx died of a sudden heart attack at 52, while Churchill lived to 90. The presenter went on to state that this is an example of how genes control CVD risk, implying that despite Fixx's exemplary lifestyle, his genes had condemned him to an early death.

I wanted to jump up and yell "I think you're leaving out the alternate hypothesis: running marathons and stuffing yourself with grains isn't healthy!" But instead I suffered quietly through what ended up being an inane yet informative presentation.

His lab looks for gene variations that affect blood pressure (BP). There's a huge amount of money and research going into this. His lab and others have come up with two classes of mutations:

• Common allele variants that have an insignificant but measurable effect on blood pressure.
• Rare genetic mutations that have a significant effect on BP. The most common affects 1 in 2,000 people in the US.

Despite truckloads of funding and research, they have yet to uncover any gene or combination of genes that accounts for even a fraction of hypertension in Americans. So what's the next step? Keep looking for genes.

I suspect they will never find anything interesting. The reason? Hypertension is tightly linked to lifestyle. It's a quintessential aspect of the "disease of civilization". It's highly responsive to carbohydrate restriction, as a number of clinical trials have shown. Remember the Kuna? They don't get hypertension when they live a non-industrial, grain-free lifestyle (despite eating more salt than the average American), but as soon as they move to the city their hearts explode. It's been demonstrated in a number of other similar cases as well. Genetics are clearly not responsible.

Don't get me wrong, I do think genetics can modify a person's response to a poor lifestyle. But when the lifestyle is healthy, the vast majority of these differences fade away. I have a more thorough discussion of this point here.

If you give just the right dose of poison to a group of animals, 50% will die and 50% will survive (called the EC50 dose). You might then conclude that genetics had determined who lived and died. You wouldn't be wrong, but you'd be missing the point that what killed them was the poison.

The thing that really bothers me about this thinking is it's disempowering. The presenter suggested that the reason for the difference between Fixx and Churchill was their genes. If genes have us in such a tight grip, why bother trying to live well? The only logical solution is to pop hypertension pills and eat cake all day.

My guess is that if they had lived a more natural lifestyle, Fixx would have made it to 90 and Churchill would have been fit and lean.