This post is a detailed critique of the following study:
The Hullabaloo over Hazen and Koeth’s Red Meat Study
A few days ago, a brand new study by Dr. Stanley Hazen’s group at the Cleveland Clinic was published, incriminating an unfamiliar ingredient within red meat as the cause of heart disease. The New York Times trumpeted:
“The real culprit, they proposed, was a little-studied chemical that is burped out by bacteria in the intestines after people eat red meat. It is quickly converted by the liver into yet another little-studied chemical called TMAO that gets into the blood and increases the risk of heart disease.” (Gina Kolata, New York Times 4/7/13)
This article received a lot of attention, so I was asked by readers and friends to comment on it. This new study is actually a conglomeration of mouse experiments, human clinical studies, and human epidemiological studies, which the authors then weave together to make their case against red meat. Straightforward it is NOT. It is a many-headed beast, armed with tentacles and suckers, but is smart, elegant in many ways, and deserving of detailed scrutiny.
Much like a game of Twister, the authors stretch so far to try to connect the dots between their studies that their sophisticated theory ultimately collapses into a sad, lifeless heap. The bottom line is that there are no new reasons to fear red meat and no proof within this study (or any other study, for that matter) that red meat causes heart disease. But if you want to know how I came to this conclusion, trudge forth, gentle reader. And Godspeed.
The Authors’ Beliefs
(or why they conducted these studies in the first place)
The article opens this way (emphases mine):
“The high level of meat consumption in the developed world is LINKED to CVD (cardiovascular disease) risk, PRESUMABLY owing to the large content of saturated fats and cholesterol in meat. However, a recent meta-analysis of prospective cohort studies showed NO ASSOCIATION between dietary saturated fat intake and CVD, prompting the SUGGESTION that other environmental exposures LINKED to increased meat consumption are responsible.” [Koeth 2013]
If after more than 40 years of studying meat and heart disease, this is the strongest statement intelligent scientists can make condemning meat, then maybe, just maybe, the theory is weak…or…heaven forbid…completely false.
Allow me to summarize the authors’ line of reasoning:
- Red meat must cause heart disease somehow, because epidemiological studies (which have no power to demonstrate cause and effect) suggest that people who eat more red meat are at higher risk for heart disease. Epidemiological studies also suggest that vegetarians have lower rates of heart disease than omnivores [click HERE to read why this may have nothing to do with meat.]
- But a recent meta-analysis showed no connection between dietary saturated fat and heart disease, so fat is not the culprit after all…
- We need to find a new villainous ingredient within red meat to blame for heart disease, because the possibility that red meat might NOT cause heart disease either hasn’t occurred to us, or sounds preposterous to us. [Perhaps they are not aware that some groups of people eating very high-meat diets, like the Inuit, had very low rates of heart disease. I will never understand how intelligent people can believe that an ancient food is causing a modern disease.]
- Hey…red meat contains more L-carnitine than white meat, and plant foods are extremely low in carnitine, so maybe L-carnitine causes heart disease!
- Oh…but wait…L-carnitine is a vital nutrient in our bodies. L-carnitine is used to transport fat into the mitochondria of our cells so it can be burned for energy. It is so essential that if we don’t eat enough of it, we go out of our way to make it from scratch. Since L-carnitine itself is innocent, we have to work extra super hard to connect it to heart disease. How can we can pin the tail on this red-meat-filled donkey?
But first, a map of the forest, so we won’t get disoriented in the trees.
Carnis is the Latin word for meat. Because carnitine is vital to animal energy production, all animal foods contain carnitine, but red meat has a LOT more than other foods because dark muscle fibers rely most heavily on fat for energy:
Table 1: Selected food sources of carnitine
|Beef steak, cooked, 4 ounces||56-162|
|Ground beef, cooked, 4 ounces||87-99|
|Milk, whole, 1 cup||8|
|Codfish, cooked, 4 ounces||4-7|
|Chicken breast, cooked, 4 ounces||3-5|
|Ice cream, ½ cup||3|
|Cheese, cheddar, 2 ounces||2|
|Whole-wheat bread, 2 slices||0.2|
|Asparagus, cooked, ½ cup||0.1|
Carnitine in meat exists as “acetyl-L-carnitine esters”, not as free carnitine. All this means is that meaty carnitine has special little attachments that make it easier for the small intestine to absorb (i.e. more “bioavailable”) than free L-carnitine. If we eat meat, we absorb a large percentage of the natural form of carnitine it contains. Vegans and vegetarians absorb more carnitine (66-86%) than meat-eaters (54-72%), because the body needs carnitine and it’s easier to get it from food than to make it from scratch. By contrast, if we swallow a supplement containing free L-carnitine, we only absorb about 10 to 20% of it.
Carnitine that is not absorbed by the small intestine can make it all the way down to the colon, where some types of bacteria can break it down, releasing a by-product called “TMA” (trimethylamine). TMA is a gas that smells like rotten fish. TMA wafts into the bloodstream and makes its way to the liver, where special enzymes convert it into TMAO (trimethylamine oxide)—an odorless substance that is easily excreted in the urine. TMAO is the molecule that Dr. Hazen’s group thinks causes heart disease.
Why would our bodies contain an enzyme that would go out of its way to deliberately turn a rotten gas (TMA) into a substance that can kill us (TMAO)? I doubt it would…While TMA itself does not seem to be toxic, it can combine with proteins to form potentially cancer-causing nitrosamines. This may be just one reason why the liver works to transform it into something the body can get rid of.
Hey, what about CHOLINE?
Carnitine is not the only nutrient that bacteria can turn into TMA. Another common food molecule—choline—can also be turned into TMA. Choline is found in all kinds of foods, not just in meat:
[mg choline per 100g of food]
- Egg yolks 680
- Egg whites 270
- Liver 195 – 333
- Toasted wheat germ 152
- Meat/seafood 34 – 103
- Nuts 29 – 72
- Brussels sprouts/broccoli 40
“The normal human diet contains approximately 500 mg of free choline/d and humans do not usually smell “fishy” unless they are treated with large oral doses of choline. Supplementary choline is ingested in “health food” preparations by many individuals, and choline supplements…” [Zeisel]
If TMAO is bad for us, we should also stop eating saltwater fish. We don’t even need bacteria to generate TMAO from fish—fish naturally contain TMAO; they use it to regulate their fluid balance so they won’t shrivel up in the salty waters of the sea:
Flounder: 400 mg/100 g
Alaskan Pollack and Cod: 1000 mg/100 g
So, why pick on red meat, when TMAO can result from the digestion of other foods, such as egg whites, fish, wheat germ, nuts, and cruciferous vegetables?
The authors acknowledge that choline can form TMA and TMAO but do not mention that choline comes from plant foods as well as animal foods, and they do not mention fish TMAO at all. Either they did not do their homework, or their biases have blinded them to the facts. Such is human nature—believing is seeing.
The Authors’ Arguments
(or how they think TMAO might cause heart attacks)
In order to draw the conclusion that red meat causes heart disease, you have to buy the following series of arguments. If you read only these 6 points, and you are someone who wants to believe that red meat is bad for you, I can see how you would come away feeling as if your beliefs are validated.
- Point #1. People with heart disease tend to have higher TMAO levels [Wang 2011].
- Point #2. Vegans and vegetarians naturally have lower TMAO levels.
- Point #3. Vegans and vegetarians produce less TMAO after consuming steak and L-carnitine than meat-eaters do.
- Point #4. Vegans and vegetarians have different kinds of bacteria living in their colon than omnivores do, and this may explain why they produce less TMAO.
- Point #5. L-carnitine supplements increase atherosclerosis in genetically-altered mice.
- Point #6. TMAO interferes with “Reverse Cholesterol Transportt” (RCT) in genetically-altered mice, so it’s harder for their bodies to get rid of excess cholesterol.
ONE GIANT LEAP FOR MANKIND
If you eat red meat, your body will fill up with cholesterol and you will have a heart attack.
The observation that people with heart disease are more likely to have higher TMAO levels does not mean that TMAO causes heart attacks (and the authors do not claim that it does). It might or it might not. TMAO could simply be an innocent bystander—a red (meat) herring. We have no idea why people who have heart disease tend to have higher TMAO levels. Most importantly, researchers did not ask these people what they eat. Did they eat healthy whole foods diets or junky diets? How many of them were vegans? Vegetarians? Omnivores? Were the ones with the highest TMAO levels all omnivores? Were the ones with the lowest TMAO levels vegans? That would have been very interesting and helpful information.
Not all vegans and vegetarians have lower TMAO levels compared to omnivores. From this figure, taken from Hazen’s study, it looks like only a small percentage of them do:
[The lowest horizontal line = 10th percentile. The bottom of each box = 25th percentile. The line inside each box = 50th percentile. The top of each box = 75th percentile. The highest horizontal line = 90th percentile].
The fact that there is so much overlap in TMAO levels between meat-eaters and non-meat-eaters suggests that there is something else about diet–something other than the presence or absence of meat– that is playing a significant role in TMAO levels. Dietary choline is the most logical and likely explanation, but may not be the only one.
a. The form of carnitine found in red meat—acetyl-L-carnitine— was not used in the study; free L-carnitine supplements were used in the study instead. Because these supplements contain the form of carnitine that is hardest to absorb, the majority of it bypasses the small intestines and makes it down to the colon, where bacteria can turn it into the maximum amount of TMA (and TMAO) possible. This study does not tell us whether acetyl-L-carnitine—the form of carnitine found in red meat— raises TMAO levels. Even if TMAO turns out to be the scourge of the West, all you will have shown with this study is that people who want to avoid heart attacks should not take L-carnitine supplements with their steak.
b. NONE of the subjects was fed steak alone—they all received either L-carnitine supplements alone or in combination with steak. There is no proof anywhere in this paper that simply eating steak all by itself will raise anyone’s TMAO levels. In order to convince people that steak raises TMAO levels, you have to include a steak-only experiment. Any high school science student could tell you that.
Here’s an example of what the scientists did:
Experiment: Give one male vegan and one female omnivore each an 8-oz steak (which contains 180 mg of acetyl L-carnitine) plus a capsule containing 250 mg of free L-carnitine and watch what happens to their TMAO levels. [Poor vegan man–he really took one for the team. Give that man a year’s supply of tofu!]
Result: The vegan man produced virtually no TMAO after the steak + carnitine supplement whereas the meat-eating woman generated a TMAO spike in her bloodstream.
Given that studies have shown vegans absorb more carnitine than meat-eaters, isn’t it possible that the vegan man, whose body may have been pining for some long-overdue, prefabricated carnitine, absorbed much more carnitine than the omnivorous woman, so that much less of it made it to his colon to be turned into TMA? Isn’t it possible that the vegan man didn’t generate a TMAO spike because his liver for some reason couldn’t turn TMA into TMAO? In this case, he would have generated a TMA spike, but TMA was not measured. Chris Masterjohn wrote an excellent review of this study and proposes other interesting possibilities, including gender differences and vitamin deficiency issues. He also does a beautiful job of analyzing the omnivore vs vegan/vegetarian experiments.
Yes, what we eat determines the types of bacteria in our colon, but we have no idea which mix of bacteria is healthiest for us, we have no idea which diet encourages the best mix of bacteria, and we have no idea whether any of this has anything to do with heart disease.
The researchers detailed the interesting differences in bacterial patterns they discovered between meat-eaters and non-meat-eaters. One type of intestinal bacteria called Prevotella was strongly associated with higher TMAO production. Four of the human volunteers were found to have bacterial colonies in their colons that were rich in Prevotella. Were all four of these individuals meat-eaters? No. Three of them were omnivores (5.9% of all omnivores in the study), and one was not (3.8% of all non-meat-eaters). To quote the authors, this suggested:
“more complexity in the human gut microbiome than anticipated…Indeed, other studies have demonstrated variable results in associating human bacterial genera…including Prevotella, to omnivorous and vegetarian eating habits.”
Translation: Even if you are convinced that having Prevotella colonies setting up house in your innards is bad for you, eating a vegan/vegetarian diet will not guarantee you a low-Prevotella colon.
- Yes, L-carnitine supplements caused an increase in atherosclerosis in genetically altered mice. However, L-carnitine is not red meat. It is not even the form of carnitine found in red meat (acetyl-L-carnitine).
- The dose of L-carnitine used in these mice was extremely high. According to Chris Masterjohn, it was the equivalent of a human eating 1000 sirloin steaks.
- For reasons I don’t understand, the mice used in these studies were genetically-altered so that they were missing a gene (apoE) required for normal cholesterol processing. These mice are popular with scientists who study heart disease because they are very good at developing atherosclerosis. It’s already a big stretch to apply information from mouse studies to human health, why widen the gap by using unnaturally defective mice? Absurd.
Now here’s where the authors really jumped the shark. They fed their miscreant mice standard mouse chow supplemented with L-carnitine, choline, or TMAO itself. Then a trusty lab assistant was charged with the delightful task of measuring how much cholesterol the mice…um…released. Makes me nostalgic for the 7 years I spent as a lab tech…not. They found that mouse pellets from animals fed TMAO contained 30% less cholesterol than those fed unsupplemented chow. This is how they propose TMAO causes human heart disease. There is so much I could say about these studies, but the take-home message for your refrigerator magnet is this:
Just because the cage droppings of mutant mice deliberately fed TMAO contained less cholesterol than those who were not fed TMAO does not mean that a human who eats a steak is going to have a heart attack.
My BIGGEST BEEF with this study
Like most studies that try to connect red meat to human disease, this collection of studies does not control for refined carbohydrate intake. We do not know how healthy the diets of the human volunteers were before this study began. Were the vegans and vegetarians health-conscious types who avoided junk foods? What about the omnivores? What if it turned out that everyone who had a higher TMAO level (including the veg/vegans with higher TMAO) also happened to eat a lot more sugar and flour?
After all, if there is one clear culprit emerging from the piles of studies about diet and heart disease, it is refined carbohydrate—not saturated fat, not cholesterol, not red meat. So, in this day and age, in my humble opinion, if you are going to conduct a study of diet and heart disease, you simply must take refined carbohydrate into consideration in order to be taken seriously.
Let’s look at two clear examples in this study of how refined carbohydrate intake was completely ignored by the authors.
The authors measured TMAO levels in 2,595 people and concluded that there was a dose-dependent relationship between TMAO levels and cardiovascular disease (CVD) risk, after adjustment for “traditional CVD risk factors”, which were age, gender, diabetes, smoking, blood pressure, cholesterol levels, and medications. Notice the absence of refined carbohydrate intake, fasting blood glucose, hemoglobin A1C or any other related values.
But hope was kindled when, within a description of one of their mouse studies, I came across this promising line:
“Of note, the increase in atherosclerotic plaque burden with dietary L-carnitine occurred in the absence of proatherogenic changes in plasma lipid, lipoprotein, glucose, or insulin levels…”
Excellent–they acknowledged glucose and insulin as potential risk factors! Had they actually determined that carnitine caused atherosclerosis despite healthy glucose and insulin levels? The sentence above, which is found in the primary article, would lead you to believe this, but if you dig through the 40-page supplementary information, you unearth this telling table:
As you can see, the triglyceride, cholesterol, glucose and insulin levels of mice being fed normal chow are already pretty high (insulin levels at baseline convert to 16.3 microunits/ml). When carnitine is added, there is no statistically significant change in any of these levels (as indicated by P values greater than 0.05). The authors phrased their sentence carefully so that what they say is true—there were no proatherogenic (artery-clogging) changes in these profiles. They were already proatherogenic to begin with.
These mice, like most laboratory animals, are fed unbelievably junky diets. In this case, they were eating a “standard chow control diet”, called Teklad 2018:
Ingredients (in descending order of inclusion)- Ground wheat, ground corn, wheat middlings, dehulled soybean meal, corn gluten meal, soybean oil, calcium carbonate, dicalcium phosphate, brewers dried yeast, iodized salt, L-lysine, DL-methionine, choline chloride, kaolin, magnesium oxide, vitamin E acetate, menadione sodium bisulfite complex (source of vitamin K activity), manganous oxide, ferrous sulfate, zinc oxide, niacin, calcium pantothenate, copper sulfate, pyridoxine hydrochloride, riboflavin, thiamin mononitrate, vitamin A acetate, calcium iodate, vitamin B12 supplement, folic acid, biotin, vitamin D3 supplement, cobalt carbonate.
This chow is comprised almost entirely of processed wheat, corn, and soy—nearly 100% refined carbohydrate, held together with soybean oil. This is not a diet that exists in nature. Poor little mice. No wonder manufacturers have to supplement this stuff with 18 vitamins and minerals. Look familiar? Is anyone reminded of the ingredient list on the side of your typical cereal box? Coat these mouse pellets with chocolate frosting or sprinkles and you’ve got yourself a yummy breakfast treat 🙂
The authors provide evidence that vegans/vegetarians have a different mix of bacteria living in their colon than omnivores, and they do an excellent job of convincing us that you can’t generate TMAO without bacteria (humans do not possess this capability). Then, because they believe that red meat is bad for us, they try to connect red meat to bacterial metabolism of carnitine to TMAO. Now allow me to do the same thing with refined carbs. It took me about 3 minutes to find this reference:
“Phosphotransferase systems involved in microbial processing of carbohydrates also were found to be over expressed in the intestines of obese people, notably from Prevotellaceae…” [Kallus]
I am not saying that carbs increase Prevotella activity or that Prevotella causes heart attacks. All I’m saying is that refined carbohydrates may also play an important role in what kind of bacteria elbow their way to your colonic buffet table.
GUILT BY ASSOCIATION
Based on the thousands of scientific articles I have read about food and health, in combination with my personal experience, clinical experience, and common sense, I find nothing to suggest that red meat is bad for people, and plenty of evidence that red meat is good for people (see my meat page). Every article I’ve ever read that tries to blame meat for our health problems fails to to take refined carbohydrate into consideration. In my opinion, damning nutritious meat for your heart attack is like pouring sugar into your gas tank and then cursing the gasoline when your car breaks down.
If you want to scare me away from eating meat using mouse experiments, you are going to have to raise a slew of mice on a healthy whole foods vegetarian diet, prove to me that those mice show no signs of cardiovascular disease, then start feeding them one miniature sirloin steak for dinner every night until they all clutch their furry little chests, keel over, and die.
HAS THE FAT LADY SUNG?
As far as I can tell, the authors’ theory that red meat provides carnitine for bacteria to transform into TMA which our liver then converts to TMAO, which causes our macrophages to fill up with cholesterol, block our arteries, and cause heart attacks is just that–a theory–full of sound and fury, signifying nothing.
To read my critique of the World Health Organization’s 2015 proclamation that red and processed meats cause cancer: WHO Says Meat Causes Cancer?
Bennett BJ et al. Trimethylamine-N-Oxide, a Metabolite Associated with Atherosclerosis, Exhibits Complex Genetic and Dietary Regulation. Cell Metabolism 2013; 17: 49–60.
Kallus SJ and Brandt LJ. The Intestinal Microbiota and Obesity. J Clin Gastroenterol 2012;46:16–24.
Koeth RA et al. Intestinal microbiota metabolism of l-carnitine, a nutrient in red meat, promotes atherosclerosis. Nat Med Apr 7 2013 [Epub ahead of print].
Rebouche CJ. Kinetics, pharmacokinetics, and regulation of L-Carnitine and Acetyl-L-carnitine metabolism. Ann NY Acad Sci 2004; 1033: 30–41.
Reuter SE and Evans AM. Carnitine and Acylcarnitines: Pharmacokinetic, Pharmacological and Clinical Aspects. Clin Pharmacokinet 2012; 51(9): 553-572.
Sikorski ZE. Seafood: Resources, Nutritional Composition, and Preservation. 1990 CRC Press.
Wang Z et al. Gut flora metabolism of phosphatidylcholine promotes cardiovascular disease. Nature 2011; 472 (7341): 57-63.
Zeisel SH, et al. Conversion of Dietary Choline to Trimethylamine and Dimethylamine in Rats: Dose-Response Relationship. J Nutrition 1989; 119 (5): 800-4.