How the Egg Industry Tried to Bury the TMAO Risk

“Metabolomics is a term used to describe the measurement of multiple small-molecule metabolites in biological specimens, including bodily fluids,” with the goal of “[i]dentifying the molecular signatures.” For example, if we compared the metabolic profile of those with severe heart disease to those with clean arteries, we might be able to come up with a cheap, simple, and noninvasive way to screen people. If heart patients happened to have something in their blood that healthy people didn’t, we could test for that. What’s more, perhaps it would even help us understand the mechanisms of disease. “To refer to metabolomics as a new field is injustice to ancient doctors who used ants to diagnose the patients of diabetes” (because the ants could detect the sugar in the diabetics’ urine).

The first modern foray discovered hundreds of substances in a single breath, thanks to the development of computer technology that made it possible to handle large amounts of information—and that was in 1971, when a computer took up nearly an entire room. “[N]ew metabolomics technologies [have] allowed researchers to measure hundreds or even thousands of metabolites at a time,” which is good since more than 25,000 compounds may be entering our body through our diet alone.

Researchers can use computers to turn metabolic data into maps that allow them to try to piece together connections. You can see sample data and a map at 1:28 in my video Egg Industry Response to Choline and TMAO. Metabolomics is where the story of TMAO started. “Everyone knows that a ‘bad diet’ can lead to heart disease. But which dietary components are the most harmful?” Researchers at the Cleveland Clinic “screened blood from patients who had experienced a heart attack or stroke and compared the results with those from blood of people who had not.”

Using an array of different technology, the researchers identified a compound called TMAO, which stands for trimethylamine N-oxide. The more TMAO people had in their blood, the greater the odds they had heart disease and the worse their heart disease was.

Where does TMAO come from? At 2:19 in my video, you can see a graphic showing that our liver turns TMA into TMAO—but where does TMA come from? Certain bacteria in our gut turn the choline in our diet into TMA. Where is the highest concentration of choline found? Eggs, milk, and meats, including poultry and fish. So, when we eat these foods, our gut bacteria may make TMA, which is absorbed into our system and oxidized by our liver into TMAO, which may then increase our risk of heart attack, stroke, and death.

However, simply because people with heart disease tend to have higher TMAO levels at a snapshot in time doesn’t mean having high TMAO levels necessarily leads to bad outcomes. We’d really want to follow people over time, which is what researchers did next. Four thousand people were followed for three years, and, as you can see in the graph at 3:10 in my video, those with the highest TMAO levels went on to have significantly more heart attacks, strokes, or death.

Let’s back up for a moment. If high TMAO levels come from eating lots of meat, dairy, and eggs, then maybe the only reason people with high TMAO levels have lots of heart attacks is that they’re eating lots of meat, dairy, and eggs. Perhaps having high TMAO levels is just a marker of a diet high in “red meat, eggs, milk, and chicken”—a diet that’s killing people by raising cholesterol levels, for example, and has nothing to do with TMAO at all. Conversely, the reason a low TMAO level seems so protective may just be that it’s indicative of a more plant-based diet.

One reason we think TMAO is directly responsible is that TMAO levels predict the risk of heart attacks, strokes, or death “independently of traditional cardiovascular risk factors.” Put another way, regardless of whether or not you had high cholesterol or low cholesterol, or high blood pressure or low blood pressure, having high TMAO levels appeared to be bad news. This has since been replicated in other studies. Participants were found to have up to nine times the odds of heart disease at high TMAO blood levels even after “controll[ing] for meat, fish, and cholesterol (surrogate for egg) intake.”

What about the rest of the sequence, though? How can we be certain that our gut bacteria can take the choline we eat and turn it into trimethylamine in the first place? It’s easy. Just administer a simple dietary choline challenge by giving participants some eggs.

Within about an hour of eating two hard-boiled eggs, there is a bump of TMAO in the blood, as you can see at 4:51 in my video. What if the subjects are then given antibiotics to wipe out their gut flora? After the antibiotics, nothing happens after they eat more eggs. In fact, their TMAO levels are down at zero. This shows that our gut bacteria play a critical role. But, if we wait a month and give their guts some time to recover from the antibiotics, TMAO levels creep back up.

These findings did not thrill the egg industry. Imagine working for the American Egg Board and being tasked with designing a study to show there is no effect of eating nearly an egg a day. How could a study be rigged to show no difference? If we look at the effect of an egg meal (see 5:32 in my video), we see it gives a bump in TMAO levels. However, our kidneys are so good at getting rid of TMAO, by hours four, six, and eight, we’re back to baseline. So, the way to rig the study is just make sure the subjects hadn’t eaten those eggs in the last 12 hours. Then, you can show “no effect,” get your study published in the Journal of the Academy of Nutrition and Dietetics, and collect your paycheck.


Unfortunately, this appears to be part for the course for the egg industry. For more on their suspect activities, see:

For more on the TMAO story, see:

In health,
Michael Greger, M.D.

PS: If you haven’t yet, you can subscribe to my free videos here and watch my live presentations:

Is Type 1 Diabetes Triggered by the Bovine Insulin in Milk?

The tight correlation among countries between the incidence of type 1 diabetes in children and cow’s milk consumption didn’t account for Iceland. Indeed, studies correlating dairy intake in children and adolescents with the incidence of type 1 diabetes deliberately excluded the Icelandic data. Why? Is it because of genetics? Perhaps, yes and no. The people of Iceland are similar genetically to other Nordic countries, but their cows are not. As I discuss in my video Does Bovine Insulin in Milk Trigger Type 1 Diabetes?, there are two main types of the cow milk protein casein: A1 and A2. Icelandic cattle, who “have been isolated from interbreeding with other cattle breeds for over 1,100 years,” are unusual in that they produce mostly A2 milk, which may explain the lower incidence of type 1 diabetes in Iceland.

Unlike A2 casein, A1 casein breaks down into casomorphin, which has opioid properties that may alter immune function, perhaps increasing susceptibility to infections that may themselves trigger type 1 diabetes. That’s what’s in the milk from the classic black-and-white patterned Holstein cows, who make up about 95 percent of the U.S. dairy herd and much of the global herd—A1 casein. This issue has even caused dairy boards to begin taking out patents on methods for selecting “nondiabetogenic” milk to avoid triggering of Type 1 diabetes. Indeed, looking only at A1 casein consumption certainly restores that tight linear relationship between milk intake and type 1 diabetes and you can see at 1:47 in the video.

These so-called ecological, or country-by-country, studies, however, primarily serve to suggest possibilities that then need to be put to the test. For example, a study was designed where hundreds of siblings of type 1 diabetics were followed for about ten years and found that those who drank a lot of milk did have about five times the risk of coming down with the disease, too. By the mid-1990s, more than a dozen such studies were done.

Overall, researchers found that early cow’s milk exposure appears to increase the risk of type 1 diabetes by about 50 percent. Those data were enough for the American Academy of Pediatrics to decide that “cow’s milk protein may be an important factor” in the initiation of the process that destroys our insulin-producing cells. The organization went on to say that the avoidance of cow’s milk protein may reduce or delay the onset of type 1 diabetes. As such, the American Academy of Pediatrics emphasizes that breast milk is best and, for those at higher risk of the disease, strongly encourages the avoidance of products containing cow’s milk protein that is intact, as opposed to hydrolyzed formula where the milk proteins are broken up into tiny pieces.

Typically, hydrolyzed formula is given to children with dairy allergies and could potentially make it less risky, but we don’t know until we put it to the test. Based on the population studies and meta-analyses of antibody studies, which suggested that “cow’s milk may serve as a trigger of Type 1 diabetes,” a pilot study was initiated the following year. Researchers wanted to see if babies at high genetic risk for the disease would be less likely to develop antibodies that would then attack their own pancreas if they drank hydrolyzed casein—that is, casein that was chopped up. The hydrolyzed formula did seem to reduce the appearance of at least one autoimmune antibody, but not two or more, which is much more predictive of the development of the disease.

Nevertheless, that was enough for the investigators to embark on the ambitious Trial to Reduce Incidence of Diabetes in Genetically at Risk, also known as the TRIGR study. This multinational, randomized prospective trial involved randomizing thousands of newborns across 15 countries. In 2010, preliminary data suggested the hydrolyzed formula may have helped, but they didn’t quite reach statistical significance, approximately meaning there was greater than a 1 in 20 chance the findings could have just been a fluke. Indeed, when the final autoimmune antibody results were published, the special hydrolyzed formula didn’t seem to help at all.

The researchers only looked at a special group of children, though—ones who were at high genetic risk with diabetes running in the family—whereas the great majority of children who get type 1 diabetes do not have any afflicted close relative. Perhaps most importantly, however, as the researchers themselves emphasized, their study wasn’t designed to test whether cow’s milk is or is not a trigger for the disease. Instead, it aimed to analyze the potential effects of the hydrolyzed casein formula. Maybe it’s not the casein, though. Maybe it’s the bovine insulin.

Insulin autoantibodies—antibodies our body produces to attack our own insulin—often appear as the first sign in prediabetic children. “Because cow’s milk contains bovine insulin,” around the same time researchers were looking into casein, another team “followed the development of insulin-binding antibodies in children fed with cow’s milk formula.” They found significantly more antibodies to bovine insulin in the cow’s milk formula group compared to the exclusively breastfed group, who may have only been exposed to cow proteins through their mom’s breast milk (if their mothers consumed dairy). Furthermore, the bovine antibodies cross-reacted with human insulin, potentially being that caught-in-the-crossfire cause triggering at least some cases of type 1 diabetes.

Of course, we can’t know for sure until we put it to the test. Researchers ran another randomized, double-blind trial, but, this time, tried a cow’s milk formula from which the bovine insulin had been removed. And, indeed, without the bovine insulin exposure, the children built up significantly fewer autoimmune antibodies. What we don’t know yet is whether this will translate into fewer cases of diabetes.


You may also want to check out:

For more on the concerns with cow’s milk exposure in infancy and childhood, see:

I’ve developed a six-part series on the role of gluten-free, casein-free diets in the treatment of autism:

What’s The Best Baby Formula? Breast milk!

In health,
Michael Greger, M.D.

PS: If you haven’t yet, you can subscribe to my free videos here and watch my live presentations:

Milk Consumption Is Linked to Type 1 Diabetes

Type 1 diabetes, which typically strikes children and young adults, is an autoimmune disease in which our own immune system attacks and destroys the insulin-producing cells of our pancreas. Untreated it’s deadly, but even with well-managed insulin replacement, it may shorten life expectancy by a decade. “Families are devastated when a child receives a diagnosis of type 1 diabetes…Thus, one of modern medicine’s ‘holy grails’” is to understand what causes the body to attack itself, in the hopes that we can cure or prevent the disease. Genetic susceptibility plays an important role, but the “concordance for type 1 diabetes is only about 50% among identical twins.” So, even if someone with our exact same DNA gets the disease, there’s only about a 50 percent chance we will get it, too—meaning there must be external factors as well.

Some countries have low rates of incidence, and others have high rates. Japan, for example, has type 1 diabetes rates 18 times lower than the United States. This disparity isn’t due only to genetics, however, because, when children migrate, they tend to acquire the risk of their new home, suggesting it’s got something to do with the environment, diet, or lifestyle. In fact, the incidence rates vary more than 350-fold around the world. Some countries have rates hundreds of times higher than others, and it is on the rise. Researchers looked at 37 populations from around the world and found that the incidence has been increasing about 3 percent a year—our genes don’t change that fast. In fact, they couldn’t find a single population with decreasing incidence of type 1 diabetes.

Something is going on that started around the end of World War II. “The best evidence available suggests that childhood diabetes [also known as type 1 diabetes] showed a stable and relatively low incidence over the first half of the 20th century, followed by a clear increase that began…around…the middle of the century.” Why the increase, though? A number of factors have been postulated, including vitamin D deficiency, certain infections, or exposure to cow’s milk.

Decades ago, published cross-country comparisons showed a tight correlation between milk consumption and the incidence of  type 1 diabetes. The “analysis showed that 94% of the geographic variation in incidence might be explained by differences in milk consumption.” Investigators in Finland, the country with the highest rates of type 1 diabetes and cow’s milk consumption, led much of the research into this area.

It all started with studies like the one I feature in my video Does Casein in Milk Trigger Type 1 Diabetes?, showing that the less babies are breastfed, the higher the rates of type 1 diabetes, leading to the obvious conclusion that “[b]reast-milk protects the newborn infant.” On the other hand, if babies are not getting breast milk, they’re getting formula, which probably contains cow’s milk proteins. In the first few months of life, our gut is especially leaky to proteins. Is it possible that as our immune system attacks the foreign cow proteins, our pancreas gets caught in the crossfire? This was based on animal experiments, however. In susceptible mice, a diet containing the cow’s milk protein, casein, produced diabetes, but it did not cause diabetes in rats. So, are we more like mice or rats?

Researchers drew blood from children with type 1 diabetes to see if they had elevated levels of antibodies that attack bovine proteins compared to controls. Their finding? Every single child with type 1 diabetes had elevated anti-bovine protein antibodies circulating in their blood compared to much lower levels in the control subjects. That seems pretty convincing, but what about Iceland? They drink more milk in Iceland than in Finland, yet Icelanders have less than half the type 1 diabetes. I explore this paradox in my follow-up video Does Bovine Insulin in Milk Trigger Type 1 Diabetes?.


The vast majority of diabetes cases are type 2, so that’s been my concentration:

I’m pleased to have been able to address type 1 diabetes. For even more on this disease, see:

In health,
Michael Greger, M.D.

PS: If you haven’t yet, you can subscribe to my free videos here and watch my live presentations: