Nick Jikomes, PhD, is a neuroscientist and metabolic health researcher investigating how food and drugs affect our bodies — he covers these topics on his Mind & Matter podcast. Jikomes completed his PhD research at Harvard, studying feeding behavior in the department of endocrinology, diabetes, and metabolism at Beth Israel Deaconess Medical Center. He previously worked in cannabis tech running research and data strategy.

Today in Pirate Wires, he explains how scientific ideas evolve into public dogmas through the process of “narrative selection” — and how a false, manufactured scientific consensus can obfuscate the truth. This post was originally published on Jikomes’ Substack, Mind & Matter, where you’ll find all of his writing and podcasts.

--

A reply to a Tweet from Elon Musk last year illustrates two opposing beliefs about human health:

Musk: Wow, steak & eggs with coffee in the morning really feels like a powerup!

Brian Krassenstein: Sounds like a heart attack in the long-term.

Whether he realizes it or not, Krassenstein is espousing the “diet-heart hypothesis.” The theory goes: dietary saturated fat and cholesterol increase blood cholesterol, driving heart disease. Consequently, a steady diet of steak and eggs will lead to a heart attack.

Bill Gates is another proponent of the diet-heart hypothesis. In 2019, he said:

Gates: Of all the categories, the one that has gone better than I would have expected five years ago is this work to make what’s called artificial meat. And so you have people like Impossible or Beyond Meat, both of which I invested in.

David Rubenstein: Do you eat it as well? Do you like it?

Gates: Absolutely. You can go to Burger King and buy the Impossible Burger.

Rubenstein: Is it healthier for you, or just healthier for the atmosphere?

Gates: It’s slightly healthier for you in terms of less cholesterol.

What does Gates mean when he says plant-based artificial meat is “healthier for you in terms of less cholesterol”? He means it contains less cholesterol than animal meat and will therefore reduce your cardiovascular disease risk — standard thinking among devotees of the diet-heart hypothesis.

But there’s a problem: dietary cholesterol intake has only a modest impact on cholesterol levels. And, there’s a “diminishing returns” effect — the higher your baseline level of dietary cholesterol, the less your serum cholesterol will rise in response to consuming more. It’s likely that negative feedback mechanisms in our bodies reduce cholesterol synthesis as we consume more of it. The fact that our bodies synthesize cholesterol de novo (i.e., without relying solely on dietary intake) and that negative feedback mechanisms regulate overall levels tells us the body “wants” cholesterol levels to be neither too high nor too low (more on that later).

Like I once did, Gates appears to believe the diet-heart hypothesis without knowing the relevant biology. Saturated fat intake can increase serum cholesterol, but the plant-based meats he invested in have similar saturated fat levels to animal meat. Artificial plant meats do have lower cholesterol content than real meat, which is why Gates said what he said — but dietary cholesterol alone doesn’t reliably impact blood levels.

Why would a smart, “pro-science” person believe a biological idea without a solid grasp of the underlying biology? Perhaps because Mr. Gates did not arrive at his beliefs from a thorough examination of the science. Perhaps, like mine, his beliefs were given to him by fiat.

Fiat Science: How the public assimilates scientific ideas

As an old Millennial born in the late 1980s, I grew up at the height of the diet-heart hypothesis’ influence on American culture. The low-fat craze was at its peak. I Can’t Believe It’s Not Butter! We were told to eat plenty of grains, but culture’s focus on low-fat dieting distracted us from how much added sugar was in our “heart healthy” cereal. Demonization of saturated fat made us feel good about spreading fake butter loaded with industrial trans fats onto our white bread. Ideas spread when they make people feel good about themselves.

Government institutions set dietary guidelines in consultation with credentialed experts. Beginning in 1992, official guidelines were distilled into the food pyramid and diligently uploaded to the minds of American schoolchildren. We heeded the guidelines, as you can see in food consumption trends (dissected more thoroughly here). In the 1990s of my childhood, everyone was preaching the “low cholesterol” gospel.

I wish we could blame Fabio, but every adult I naturally viewed as an authority said the same things for my entire childhood and adolescence. Today, many of those adults are either dead, obese, or shooting up Ozempic. Those alive probably still believe eating cholesterol is unhealthy. We received these ideas by fiat — credentialed experts told us it was The Science™.

It doesn’t feel like a cult while you’re in it. But I’m getting ahead of myself.

As an adolescent, I remember explaining to my mother that Silk was healthier than milk. Like most teens, I believed this without ever reflecting on my belief’s origins. I had no knowledge of human metabolism, but believing the “right” thing made me feel good. It also felt good telling other people what to believe. I was a vector for a meme.

Only after years of study did I renounce my faith in the tenets of the diet-heart creed. At the same time, it’s only continued to spread. I may have lost faith in the diet-heart hypothesis, but Bill Gates has not. Your family doctor almost certainly hasn’t either, which is why cholesterol-lowering drugs (statins) are among the most-prescribed pharmaceuticals ever. In fact, most medical experts — people with stellar credentials and high-level positions in government, global health organizations, and hospitals — believe the diet-heart hypothesis. Should we not trust the experts, especially when there’s consensus among them?

But this is not settled science, and I’m certainly not the only one who has lost faith. To understand why the diet-heart hypothesis took hold, we have to reckon with how scientific ideas become dominant within institutions, enshrined in public policy, and exported to the masses in narrative form.

We’ll see that scientific ideas do not spread solely based on the quality of supporting evidence, and that expert consensus is not merely a function of dispassionate truth-seeking. Our ideas serve social functions and spread based on their social utility. As social animals, our lot in life is determined by our position in social status hierarchies. We live in groups. Some of these are created through spontaneous interaction, such as your circle of personal friends. Others, such as institutions, are formally structured and explicitly designed to assemble and disseminate ideas into coherent narratives. The purpose of such narratives is to organize human group behavior, not to disseminate knowledge.

In the Darwinian battle of ideas, a meme will spread insofar as it can be integrated into a social narrative that justifies the growth and expansion of institutions. Institutions are designed to communicate ideas to people en masse, and are composed of humans who adopt and promote ideas that justify their social goals (e.g. career advancement). I call this process “narrative selection,” by analogy with natural selection.

The truth of an idea is just one of many factors determining its social utility. People routinely adopt and spread ideas for reasons other than truth. This is usually what we mean by politics: the spread of untrue ideas to help humans climb social hierarchies.

Illusion of Consensus

Our society is ruled by consensus mechanisms. Elected officials are chosen by a majority consensus of voters. Institutions like the USDA set public health policies and dietary guidelines in consultation with experts. As we’ll see, these experts are also humans, with the same self-interested social goals as the rest of us.

Social crises increase the pressure to reach consensus

The diet-heart hypothesis was championed by Dr. Ancel Keys, who wasn’t just a hard-working physiologist — he was socially successful, too, spreading his ideas via relationships with decision-makers in influential institutions.

Historically, public health crises create intense pressure for policy solutions, and those who come up with them are granted much social status. In the 1950s, the US was in the midst of a cardiovascular disease epidemic, brought to the fore of public consciousness when President Eisenhower suffered a massive heart attack in 1955.

According to the CDC and the American Heart Association, heart disease deaths were rising by 1950, peaked by 1990, then mostly declined. In contrast, cancer rates have steadily increased over the same time period.

If high cholesterol driven by saturated fat intake is the major driver of heart disease, we’d expect to see a correlated change in saturated fat consumption over this interval. Likewise, because polyunsaturated fats (PUFAs) from vegetable oils and other processed foods lower cholesterol, we’d expect to see an inverse relationship with heart disease under the diet-heart hypothesis.

What do we observe?

Saturated fat consumption was flat, even declining somewhat, from 1940 into the 1970s. In contrast, PUFA consumption has risen since the early 1900s. These high-level trends were largely driven by the replacement of animal-based fats (higher in saturated fat) with PUFA-rich vegetable (seed) oils beginning in the 1940s. In other words, during the heart disease panic of the 1950s and 60s, when Keys’ ideas were adopted, saturated fats had already begun to be replaced by PUFAs.

PUFAs are deemed “heart healthy” by the AHA and other major institutions. Coincidentally, back in 1948 AHA was paid over $20 million (in today’s dollars) by Procter & Gamble, makers of the PUFA-rich seed oil Crisco. That was a few years before Eisenhower’s heart attack. A few years later, in 1961, AHA recommended limiting saturated fat intake. Diet journalist Nina Teicholz calls this “arguably the single-most influential nutrition policy ever published, as it came to be adopted first by the US government in 1980, as official policy for all Americans, and then by governments worldwide as well as the World Health Organization.” In the following years, virtually all major health institutions coalesced around one narrative, ostensibly because a scientific consensus had formed.

Public health crises demand solutions now, not later. Policymakers want to implement expert-approved solutions ASAP — jobs, promotions, and credibility are on the line. The experts they consult may therefore be motivated to sell their best ideas as more ironclad than they really are. Experts closely connected to the policymaking apparatus are biased to prematurely coalesce around the ideas that lend themselves to implementable policy solutions. The greater the social pressure, the greater the chance of an illusory consensus.

While Keys and others did amass evidence for their claims that cholesterol causes heart disease, it had major shortcomings.¹ The diet-heart hypothesis gained favor nonetheless.

Competition for social status distorts the “marketplace of ideas”

The marketplace of ideas is not flat — it’s a fitness landscape of hills and valleys. In genetic evolution, fitness landscapes visualize how traits are more or less adaptive within a given environment. Hills represent high fitness, valleys low fitness.

In memetic evolution, a similar concept applies, but the landscape is social instead of physical. The topology of the social fitness landscape is set by high-status individuals, especially those concentrated within prestigious institutions, who have an outsize ability to set social norms and constrain the flow of information. They use their influence to bias peaks in the social fitness landscape toward their preferred ideas. Naturally, people prefer ideas that secure them high status within social hierarchies.

To give you a sense of just how much social status was at stake with regard to the diet-heart hypothesis: in 1961, Ancel Keys made the cover of TIME, earning him a reputation as a national hero for “solving” heart disease. This put him in league with famed scientists like Jonas Salk.

His fame was not simply the result of being a prolific and hardworking researcher. He was also a charismatic, well-connected, ambitious marketer. After his Seven Countries Study — the world’s first longitudinal study examining the relationship between lifestyle, diet, and heart disease — he published Eat Well and Stay Well, a book coauthored with his wife in which they essentially sold the Mediterranean diet his research claimed to support. It would shape diet trends in the US for the next half-century.

The stronger the competition for social status, the greater the incentive to cheat

It was some time before anyone realized Keys cherry-picked the data in his Seven Countries Study. As pediatric endocrinologist Dr. Robert Lustig explained on my podcast:

Jikomes: Is it established fairly well that he intentionally cherry-picked this data?

Lustig: Well, intent is complicated. I never asked him. I don’t know that anyone ever asked him if he intended. But there were 22 countries. And when he published it, there were only seven. We have the data on the other 15 and they don’t fit [his study’s conclusions]. So I don’t know. You tell me. Did he or didn’t he?

To this day, the American Heart Association and other prestigious institutions promote the idea that lower cholesterol is always “heart healthy” and that one should swap saturated fat intake for polyunsaturated fatty acids (PUFAs). It’s why you see cereal boxes containing ultra-processed grains with added sugar proudly advertising themselves as “AHA Certified.” Institutions continue to advance the diet-heart hypothesis, though it’s based on shoddy data.

When hypotheses become policy, ideas become dogma

The social conditions propelling Keys’ hypothesis forward were so strong that competing evidence was suppressed from publication. Case in point: the Framingham Heart Study, conducted from the 1940s-1960s to test the diet-heart hypothesis.

I spoke with Dr. Orrin Devinsky, a board-certified neurologist and NYU professor, about this study’s results. He explained:

In 1961, they published their first paper on risk factors — that’s where the term “risk factor” comes from — and that was the study that showed, especially men middle-aged or older who have total cholesterols over 260, had higher rates of cardiovascular disease and death. And no doubt, I think that’s true. What no one talks about is that that same Framingham Study, and many, many other studies since, have found that very low cholesterol levels have been associated with higher rates of cancer, accidental death, and suicide deaths.

Overall, even in Framingham for men whose cholesterol was under 160, total mortality was not terribly different than men whose cholesterol was over 260. But in 1960 it was “heart disease, heart disease, heart disease.” No one talked about nutrition and cancer. And when they eventually did, they just lumped saturated fat with heart disease and said, “Well, it must be causing cancer, too. Because it’s got to, right? There’s more cancer now than there used to be, it’s gotta be saturated fat!”

Remember the trend in cancer rates, which have risen steadily over the last century? Low cholesterol is a factor, which suggests the possibility that our myopic focus on lowering cholesterol may be having unfortunate side effects. But, by the time the Framingham results went public in the early ’60s, Keys’ diet-heart hypothesis was not just an ascendant idea, it was dogma.

A telltale sign of dogma? The suppression of contradicting evidence.

And that’s exactly what happened, according to Devinsky. Results that directly contradicted Keys’ diet-heart hypothesis went unpublished, collecting dust for decades before seeing the light of day.²

Dr. George Mann, who ran one of these contradictory studies (a supplement to the Framingham study), describes how this played out in his book. It’s a crushing example of how the illusion of consensus can emerge. Mann was denied funding and forbidden from publishing his research in mainstream journals.

If contradictory information is barred from publication, it looks like all the experts agree. If you’re a scientist working in an environment where contradicting dogma could mean the end of your career, you may simply choose to remain silent. Do you want to rock the boat, or get tenure?

Who, exactly, buried the Framingham results contradicting the diet-heart hypothesis? Dr. Devinsky’s word choice is telling:

The “heart mafia” — the American Heart Association, the NIH. The National Health Service, I believe, paid for the Framingham Heart Study. Our tax dollars paid for the study. And they buried it. They would not allow it to come out.

Dr. Devinsky is as decorated and mainstream as they come. He’s been a board-certified neurologist and professor at NYU for decades, and he’s Director of its Comprehensive Epilepsy Center. This is not a fringe character. Heart mafia. Those are his words. They indicate the nature of the social forces influencing public policy behind the scenes.

The heartbreaking case of Dr. Mann is part of a larger pattern that played out through the 20th century. As described by Teicholz:

Reviews and books critical of the diet-heart hypothesis were not unknown in the 1960s and 1970s, including a publication by a former editor of the Journal of the American Heart Association and articles by other prominent scientists. They argued that the hypothesis was not supported by the available data and was contradicted by numerous observations. Over time, however, these critics were effectively marginalized and silenced.

(Notice that a contrary publication came from a former editor of the AHA. One wonders: why did they publish only after leaving the AHA?)

A literal, Italian-style criminal mafia uses social force (backed by physical force) to extract unearned rents and ensure conformity to their demands. Dr. George Mann didn’t have his kneecaps busted with a crowbar, but he was disappeared. For a scientist, that’s what losing funding can mean.

Notice, again, the analogy with evolution by natural selection: the social environment (funding agencies) selectively culls the population (pulls funding), leaving certain lines of evidence (diet-heart hypothesis) behind to keep reproducing.

This is a mechanism by which a false consensus can emerge, forming the basis for a narrative used to formulate “scientific” public policy.

Two types of evidence contrary to the diet-heart hypothesis: published & personal

Since Dr. Mann’s buried study, further evidence has stacked up against the diet-heart hypothesis.³ Let me be clear: no one is saying that arbitrarily high cholesterol is a good thing. But it does not impact heart health as directly as we once thought. The more important question: how high is too high?

We know that both very high and very low cholesterol levels are clearly linked to negative health outcomes, including increased all-cause mortality risk. We also know that the medical definition of “too high” has moved lower and lower over time, leading to a wider window of cholesterol levels within which doctors reflexively prescribe statins — the most profitable pharmaceuticals ever.

Statins are often prescribed to people whose cholesterol levels are in the range with the lowest odds of all-cause mortality. I’ve witnessed this directly with family members. Family doctors simply recommend statins whenever cholesterol levels are above official guidelines’ magic threshold. In my experience, physicians are often woefully ignorant as to what the data actually shows regarding cholesterol’s non-linear relationship to mortality, including elevated risk of death when cholesterol levels are too low:

Extensive research has failed to support a role for dietary cholesterol intake in the development of cardiovascular disease. Plenty of eminent researchers disagree with public health guidelines that recommend limiting saturated fat intake, and comprehensive reviews have concluded, “The totality of available evidence does not support further limiting the intake of such foods.”

The ultimate study is not published—it’s you

At the end of the day, you are your own best source of health data. That’s why I get regular bloodwork done. Even if a hypothesis proves statistically true at the population level, you may not conform to the pattern individually. The diet-heart hypothesis has not proven true based on my personal bloodwork data. Here are my total cholesterol levels at three points in time:

Between the second and third points, six months elapsed. In that period, staples of my diet were: four pasture-raised eggs almost every morning; grass-fed cuts of beef several nights per week; wild-caught salmon with liberal amounts of organic avocado mayonnaise 2-3 nights per week; a solid amount of cholesterol-rich seafoods, like shrimp. At home, I cook with either butter or coconut oil, and actively minimize my consumption of cholesterol-lowering omega-6 PUFAs (abundant in seed oils). This is a diet high in saturated fat and cholesterol, both of which became more prominent components of my diet after the second point. Nonetheless, my total and LDL-cholesterol levels got lower, contrary to the diet-heart hypothesis. (My triglycerides, APOB, and inflammatory markers also failed to rise).

Your life is ultimately a stand-alone “n=1” experiment. You can take your health into your own hands: get regular bloodwork done, especially before and after making significant lifestyle changes. It is now possible to affordably do this from the comfort of your home.

Narrative Selection: Transforming the process of science into The Science™ of public policy

We have seen how the American Heart Association and policymaking institutions in the US came to promote an illegitimate consensus on the dietary causes of heart disease. That illusion of consensus persists to this day, driving both dietary guidelines and physicians’ behavior.

This is a classic example of narrative selection, the social process by which institutions select ideas (such as those published in scientific literature) for integration into narratives justifying social goals. In genetic evolution, natural selection drives genes offering strong survival and reproductive advantages to fixation (high, stable frequency) across a population. In memetic evolution, public policy and education — based largely on narratives influenced by social status hierarchies — drive ideas to fixation.

High-status individuals (elites) have a natural tendency to fit ideas together into narratives that justify the status quo, offering them security in their high-status positions. Lower-status individuals who aspire to higher status but are unable to achieve it (counter-elites) will attempt to propagate ideas that support counter-narratives. These counter-narratives aim to destabilize the existing social status hierarchy in an attempt to create a new one, with them at the top. Competition among social factions generates competing narratives, culminating in politics as we know it.

This process evokes what the Italian political scientist and sociologist Gaetano Mosca called a “political formula”: a set of beliefs, ideologies, or myths that justify the power and authority of the ruling class. Ideas are not selected into widely disseminated narratives based simply on their ability to predict outcomes in the external world — they are selected for social reasons, by human minds competing for social status. All too human, we never truly escape high school.

—Nick Jikomes, PhD

--

FOOTNOTES

¹ Shortcomings in Keys’ evidence for the heart-health hypothesis:

• Much of it was correlational, lacking clearly established cause-and-effect relationships between dietary saturated fat, cholesterol, and heart disease.

• Animal research, causally showing that dietary cholesterol (a) spiked blood cholesterol, and (b) led to cardiovascular disease, was often done in species metabolically distinct from humans. For example, many early studies in this field used rabbits. As herbivores, rabbits respond to dietary cholesterol radically differently than meat-eating omnivores like humans. Their bodies did not evolve for cholesterol intake.

• Many of the human epidemiological studies showing a correlation between dietary fat consumption and heart disease were either highly undersampled, not representative of the populations they drew from, or composed of cherry-picked data. The most infamous example is Keys’ “Seven Countries Study,” the world’s first longitudinal study examining the relationship between lifestyle, diet, and heart disease. The associations between cholesterol and heart disease in that study often disappeared when more robust datasets were analyzed, and have been contradicted by larger, more robust analyses since.

² Dr. Devinsky described this unsettling episode in medical history:

Just to show how politics get ugly: a very famous doctor in the Framingham Heart story, George Mann, who studied the Maasai, one of those populations where they eat milk, blood, and meat — very, very, very high saturated fat diets. Extraordinarily high fat diet and low heart disease rates. And he got lambasted. He became a persona non grata and eventually lost all of his funding.

As a doctor at Framingham, he did a nutritional study that went for four years, from 1956 to 1960. And in a cohort of like 2,000 people, he and a dietician went to the houses, interviewed homemakers about what they cooked, measured things, weighed things, and looked in great detail at what people consumed and then followed them and looked at what their heart disease rate was, mortality, etc. What they found was no correlation between how much saturated fat you ate and what your cholesterol was, nor any relationship between how much saturated fat you ate and your heart health.

What happened to that study? He was forbidden from publishing it. It exists as a supplement to the Framingham Study that you could not get at any medical library in the US. I was able to get a [copy] from the Framingham library, and amazingly it got published by a statistician who said, “I found this incredible study. It’s basically been buried. The data is immaculate. The study was beautifully performed, and I feel like it’s my duty to write this up. And he wrote it up.”

So, they did the best study they could do in the late 1950s, and because it didn’t agree with the dogma of Ancel Keys and his crew, they buried it.

³ Here’s a summary of evidence contradicting the diet-heart hypothesis, courtesy of Teicholz:

• Large, randomized controlled trials were conducted in the 1960s and 70s, replacing saturated fats with PUFAs from seed oils in ~67,000 people (reviewed here). Despite the fact that reduced saturated fat intake did lower cholesterol levels, most of these studies did not observe reductions in cardiovascular or total mortality, despite what diet-heart predicted. Again, these results simply weren’t enough to combat the inertia of Keys’ success in driving policy.

• Between 2003 and 2013, the largest epidemiological cohort study ever conducted found that saturated fat intake was not associated with cardiovascular disease mortality. In fact, it found that saturated fat intake was associated with lower total mortality and lower risk of stroke.

• More than 20 papers reviewing data relevant to the diet-heart hypothesis have been published since the mid-20th century. The vast majority concluded, like this 2020 review, that “the totality of available evidence does not support further limiting the intake of such foods [saturated fat-rich foods].”