Bad Health Journalism, an addendum

The Atlantic defends seed oils on the basis that Americans "stopped cooking with tallow for a reason." But a good reason? I have to take exception.

Housekeeping note: I was hoping to keep these posts generally short. My apologies for failing on that (so far). Quoting the French mathematician Blaise Pascal, who gets first credit for this notion, “If I had more time, I would have written a shorter letter.”

The addendum to that first post

Among the consequences of writing about the problematic state of health reporting is that it will prompt readers to ask about other articles they have recently read. This article, for instance, from The Atlantic—“America Stopped Cooking with Tallow for a Reason”—raised a host of issues I had not even begun to discuss.

The “reason” alluded to in the headline is that “[c]onventional medical guidance has long recommended” that people eat “less solid fat, more plant oils” and that this conventional guidance is correct.

Now we have Robert F. Kennedy Jr and his MAHA (Make America Healthy Again) allies proposing that it’s the plant oils that are toxic and a major cause of chronic disease in America. By this thinking, the solution is rejecting the conventional medical guidance and embracing those same solid fats—saturated fat-rich animal fats, like tallow, lard, and butter—that we’ve been told for decades will kill us prematurely. Egad, RFK Jr even deep-fried his Thanksgiving turkey in tallow and posted the video on X. Hence, the subtext/sub-head of the Atlantic article: RFK Jr is motivated by his affinity for “bucking tradition,” not because on this subject he might be right.

On this subject, though, I side with RFK Jr.

The Atlantic article explicitly contradicts what I’ve been arguing in my books and articles, dating back to this 2001 investigation I did for the journal Science--“The (soft) Science of Dietary Fat.” (FWIW, that article was awarded one of the two major annual awards in science journalism, the only one for which it was eligible. It was also included in the 2010 anthology, The Best of The Best American Science Writing.) After a year of research and writing, after reading much of the copious (even then) relevant literature and interviewing over 160 sources in academia and government, I concluded that Americans stopped cooking with tallow for reasons that were unjustified by the existing evidence.

This investigation was my red pill: it turned me into a convention-bucker, albeit, I hope, not a knee-jerk one. I came out of the research convinced that the subject of nutrition and chronic disease cannot be reported naively or uncritically.

Part 1 (of 3) of my 2007 book Good Calories, Bad Calories went into further detail on the history and the evidence supporting this notion that it’s the solid fats in our diets—saturated fats—that cause heart disease, and they do so by raising LDL cholesterol. My conclusion is then a subtext of all my subsequent books. Their arguments depend on whether I am right about this one. Most obviously, to suggest that low-carbohydrate/high-fat/ketogenic diets may be beneficial, if not ideal for a significant portion of the population, as I do, then it had better be true that saturated fat is not the dietary evil that Tayag and her chosen authorities assert in this article.

Simply put, if Tayag’s take is correct, then I’m the last person who should be opining about the sorry state of health reporting.

Hence, this addendum to my first post. It’s lengthy because I’ll discuss both the evidence presented by Tayag and offer up some journalistic principles that might guide reporting in these controversial fields in the future.

Tayag herself is a staff writer for The Atlantic, a publication for which I’ve written in the past as a freelancer and hope (perhaps naively) to do so again in the future. Tayag has considerable experience reporting and editing articles in the science, medicine & health (SM&H) beat. Her brief Atlantic bio says she focuses on “science and the future of food,” so she should know what she writes. But if I’m right about the science (always a critical caveat), she lacks the necessary experience to cover this controversy.

This might not matter in most of the subjects Tayag covers, but as discussed in the last post, reporting on any subject even tangentially related to obesity and diabetes must assume the existence of controversy. If nothing else, we have the obesity and diabetes epidemics as evidence of mismanagement, a public health failure on a seemingly unprecedented level. It’s not as though the authorities have this situation well under control and we can trust their judgment because of it.

Why the editors of major publications like The Atlantic, don’t see this issue as critically important is a conversation I’ve always wanted to have with them. (They seem less interested in chatting.) As the obesity and diabetes epidemics have become ever more dire, these publications have continued to assign young writers to the subject who have little to no idea of the long history of the controversy, the traps they might be walking into, or why they might have reason to be skeptical of their sources.

What it takes to get that experience and knowledge is a critical question, of course (and a subject for a later post). After all, as I suggested in the last post, if you do the necessary legwork and conclude that the authorities are unreliable, how do your editors or your readers (or you, if you’re properly skeptical) have faith that you came to the right conclusion?

Imagine if Tayag, for instance, had written an article taking RFK Jr’s side on this one. Need I say anything more (let alone the thousands of words to follow)?

The Atlantic proposition: polyunsaturated fats are good because saturated fats are bad

When government functionaries half a century ago concluded for the first time ever that they had an obligation to disseminate dietary guidelines aimed at preventing chronic disease, they did so based largely on this fundamental proposition that saturated fats are the link between diet and heart disease and that link goes through LDL cholesterol. Essentially all other dietary guidelines given by any authoritative institutions from the 1980s onward have had to be reconcilable with that notion.

It’s this proposition that Tayag assumes to be true and defends. Hence, the logic of her article:

1. The seed-oils-are-toxic argument (i.e., what RFK Jr believes) is nonsense.

2. We know it’s nonsense because we know that saturated fat is harmful.

3. So, if nothing else, consuming the polyunsaturated fats in seed oils is better than consuming saturated fats.

Researchers (loosely speaking) started making this argument in the 1950s. The American Heart Association took it up in 1960, and eventually the U.S. Department of Agriculture and the U.S. Congress, beginning in 1977, and the National Heart, Lung, and Blood Institute, NHLBI, and other government institutions in 1984.

While these institutions fell into lockstep, the clinical trials—the tests of this dietary-fat-causes-heart-disease hypothesis—never actually panned out. By the late 1970s, however, so much money, effort, and hope had been poured into this research—largely by NHLBI—that the authorities in these agencies and health institutions decided that the proposition was likely enough to be right that they could assume it was and act accordingly. Medical, health, and wellness reporters of the era dutifully followed along.

Hence, the post-1984 messaging: We should avoid saturated fat because it will raise LDL cholesterol and kill us prematurely. Instead, we should consume the polyunsaturated fats in seed oils because they won’t. The food industry was content because it could now sell seed oils to the nation as a health food, and we went from there.

This is the history, as described in GCBC and Nina Teicholz’s 2014 book The Big Fat Surprise. Tayag’s article (with length restrictions, I assume) does not get into it. Rather she picks up the story with the assumption that this ubiquitous messaging of the dangers of SFA has a compelling evidence base; that the science is worthy of our trust and worth defending.

Before I get to that issue, the important one, I want to discuss one of the many lesser quibbles I have with Tayag’s reporting.

Her lead describes “cardiologists shudder[ing] at the thought” of RFK deep-frying his turkey in tallow. Clever as this is as a transition, it speaks to one of the many conceptual problems with SM&H reporting in general.

Let’s call it rule 1 of half a dozen I’m going to suggest (in no particular order of importance):

Rule 1: A medical doctor is not a scientist with occult powers of observation, and should not be treated as one.

Practitioners of medicine possess no first-hand evidence about the cause of a chronic disease. Those shuddering cardiologists, for instance, specialize in the diagnosis and treatment of heart problems, not in establishing the evidence base of the dietary causes of heart disease. Their understanding is based not on accumulating the relevant data, but on what they’ve been taught in medical school or have read in the journals. If those sources are wrong, then the doctors are also.

A related problem: Tayag could have written “nutritional epidemiologists shuddered at the thought,” and she would have seemed on safe ground. The nutritional epidemiologists, after all, are the ones who generated much of the evidence on which this SFA->CHD link is based. Were they right?

The Atlantic itself has reported on the troublesome glitches (i.e., bad signs) in this nutritional epidemiology business, so its writers should be particularly skeptical, and perhaps hesitant to quote nutritional epidemiologists as experts. (That article is a good read.) These complications are why a reporter might want to spend more than a week or two reporting these issues.

I admit, though, this is only a digression. I’m bringing it up because I once had to listen to the same logic perpetrated by Mehmet Oz on his Dr. Oz show. He knew with certainty that saturated fat caused heart disease, he said, because he is (was?) a cardiothoracic surgeon (he wore his scrubs, presumably so viewers would not forget) and could somehow divine this cause and effect by having had his hands quite literally in his patients’ chests. The clotted arteries apparently whispered to him that it was saturated fat that did it. I remain skeptical.

Rule 2: If you find yourself writing something that doesn’t make sense or is internally inconsistent, pause and reflect.

As for the important stuff, the evidence supporting the SFA-causes-heart-disease notion, here’s Tayag’s money paragraph, proposing that we know with sufficient certainty that the SFA-PUFA switch led to a healthier America:

The shift from saturated to polyunsaturated fats—not just in restaurants but in home kitchens—corresponded with major health gains in the United States. In 1962, Americans began to consume more vegetable fats, largely in the form of margarine; four years later, cardiovascular deaths began a decades-long decline. From 1940 to 1996, deaths from heart disease fell by 56 percent, and they continued falling through 2013, albeit at a lower rate. Although the decline can be partly attributed to factors such as better blood-pressure control and lower rates of smoking, “the increase in polyunsaturated fat is probably one of the primary factors, if not the primary factor, in dramatically reducing heart-disease death” as well as lowering the risk of diabetes, dementia, and total mortality, Walter Willett, a Harvard professor of nutrition and epidemiology, told me.

Tayag’s paragraph itself is full of inconsistencies and inaccuracies, which should have been a sign, a reason to ask Walter Willett some tough questions if nothing else. For instance, she writes that Willett told her that the increase in polyunsaturated fat consumption may likely have been “`one of the primary factors, if not the primary factor, in dramatically reducing heart disease death’ as well as lowering the risk [my italics] of diabetes…”

But the risk of diabetes has not lowered since the 1960s. Quite the opposite: diabetes prevalence (and so risk) has exploded, increasing by 600 percent. That’s the diabetes epidemic.

Now, Willett might be right that increasing PUFA consumption has profoundly affected diabetes risk. But, if it has, the effect is in the opposite direction that Willett seemingly implies and Tayag writes.

That’s what the seed-oils-are-toxic proponents are arguing: eating more PUFA associates with an explosion in diabetes risk. Obesity rates also exploded from the 1960s onward—i.e., corresponding with the shift from saturated to polyunsaturated fats—so Tayag’s description of the shift as corresponding, by implication, only with “major health gains” is a bizarre one.

We’ll get back to this. It’s a critical observation.

Moreover, it’s because of Willett’s research and advocacy that the public health community came to think that trans fats, the stuff of margarine, are deadly, per this 1995 review when he was beginning to make this case.

That could make him an ideal source to discuss this subject, but it could also make him a very biased source.

Either way, if Willett was right about trans fats—and artificial trans fats have been banned since 2018 (Harvard is quite proud of that)—then Tayag should not have been so quick to imply that the switch to “largely” margarine had anything to do with the decline in CHD mortality that begins four years later.

Ultimately, we’re dealing with the association-is-not-causality problem. We have trends in heart disease mortality and we’re looking for interventions that happened around that time that may have been responsible. We’d always be guessing which ones might be causal and how significant the role, but we can still say what is more or less likely to be true. We also have to be exceedingly wary when playing this association game that we may be missing something that wasn’t measured.

Let’s look at the article Tayag references about the 56% decline in heart disease deaths that could correlate with the shift from SFA to PUFA. Here’s the graph showing the decline from 1950 to 2014:

Here’s the table in which the authors discuss the “major advances” that played critical roles:

Rule 3: You can’t know what you don’t know, particularly if you’re new to the story

That looks like a comprehensive list of major advances, but is it missing anything? By referencing the article, Tayag implies that these authors knew what they were doing and could be trusted to include all the “major advances” that may have contributed to reducing CHD mortality.

But they don’t.¹

What the authors neglect to mention, and so neither does Tayag, (I’m assuming this is new to her) are the remarkable advances in emergency medical services, EMS, and emergency room care that also date to the 1960s. These undeniably had a major impact on heart disease mortality, although how major is impossible to quantify accurately. That the authors, all affiliated with NHLBI, don’t mention this is mind-boggling. (And, yes, you’d have thought I would have lowered my expectations by now.) Willett should know this, too, although Willett is a nutritional epidemiologist and this is not directly a subject of his research.

What did emergency medical services and emergency room care have to do with heart disease mortality? Before the 1960s, if you had a heart attack you were far less likely to survive long enough to make it to the hospital where the surgeons and cardiologists could get to work trying to keep you alive. Once in the ER, you’d be less likely to survive because you hadn’t been treated promptly enough. So having a heart attack in the 1950s was significantly more likely to be deadly for these reasons than it is, for instance, today.

With the revolution in EMS services that began in the 1960s, not only was ambulance service made far more widely available, but the EMS techs on the ambulance were far better trained, as this article explains. They were taught how to resuscitate heart attack patients if necessary (the famous defibrillator scenes we’re so familiar with now from TV and movies), to do CPR, and to keep patients alive and viable until they got to the ER. The ER physicians and nurses in turn were also trained to deal with heart attacks. From the article linked to above:

The 1960s were a time of rapid improvement in emergency care. During that decade, the importance of cardiopulmonary resuscitation, defibrillation, cardioversion, and new pharmaceutical therapies was demonstrated.1 The American Heart Association and the American Red Cross accepted these techniques and began to train health care providers, although EMS providers were initially excluded. Advances in trauma care also occurred, including the development of specialty trauma centers, such as the University of Maryland’s Shock Trauma Center. Techniques that had reduced the mortality of injured soldiers reaching medical facilities began to be applied to civilian trauma patients.10 As a result, physicians and politicians began to treat traumatic deaths as an abnormality, not an inevitable event.25 In fact, R. Adams Cowley, a leader in trauma and critical care, estimated that a quality emergency health system could cut the accident death rate by 50%.25

Many of these newest technologies were being applied by EMS with immediate, quantifiable benefits….

In 2004, Daniel Levy, a director of the Framingham Heart Study, and co-authors tried to use the Framingham data to assess what factors may have reliably contributed to the observed decline in heart disease mortality in that cohort. “Some but not all of the decline in [sudden cardiac death]” rates, they conclude, could be explained by resuscitation alone – EMS personnel defibrillating patients back to life. That still left open all the other interventions carried out by the EMS techs at the site of the patient’s heart attack and in the ambulance on the way to the hospital and in the ER once the patient arrived.

The reason I know about this and Tayag (apparently) doesn’t is because at some point in the five years I spent on my book, I was told by one of the 600-some sources I interviewed that the great bulk of the decline in CHD mortality in this era was due to these remarkable advances in emergency care. Changes in diet may still have played a role, but not the primary factor that Willett is suggesting in the quote. And because I wasn’t working on a deadline, I had time to check to see if this was right.

Here the problem is not so much with Tayag’s reporting, but the NHLBI authors who left the EMS advances from this era out of their article. They should have known better.

Rule 4: Make sure you’re talking about the right variable.

When discussing the benefits of prevention, mortality is the wrong variable to be discussing, in any case. This is another issue that should be simple, but the authorities tend to avoid it, which means the reporters do as well.

Since the end of World War 2, the medical and pharmaceutical establishments have invested enormous effort into both preventing and treating heart disease—keeping people alive and so minimizing mortality. It’s fair to assume they’ve had considerable success. The medical community has also put enormous effort into diagnosing heart disease risk factors earlier than ever, and treating those—hence all the money spent on testing, on cholesterol and blood pressure medications, on stents, etc. And we can assume those, too, have had an overall beneficial effect.

The reason to change our diets, though, to replace SFA with PUFA, is to prevent heart disease, not treat it (even as cardiologists and dietitians will advise it for treatment, too, albeit with the usual absence of clinical trial evidence). Preventing heart disease means we should see less of it, regardless of any effect on mortality: the average 50-year-old should be less likely to be diagnosed with CHD because the average 50-year-old has been, in theory, eating to prevent it.

Now I’m just going to quote from the prologue of my 2007 book Good Calories, Bad Calories because having written this once, it seems silly to rewrite (references are available in the book or on request):

Consider, for instance, that most reliable evidence suggests that Americans have indeed made a conscious effort to eat less fat and particularly less saturated fat since the 1960s. According to the U.S. Department of Agriculture, we have been eating less red meat, fewer eggs and more poultry and fish; our average fat intake has dropped from 45 percent of total calories to less than 35 percent, while National Institutes of Health surveys have documented a coincident fall in our cholesterol levels. Between 1976 and 1996, there was a 40 percent decline in hypertension in America, and a 28 percent decline in the number of individuals with chronically high cholesterol levels. But the evidence does not suggest that these decreases have improved our health. Heart disease death rates have indeed dropped over those years, and perhaps the risk of suffering a severe heart attack, what physicians call an acute myocardial infarction, but the incidence of heart disease has not declined, as would be expected if eating less fat made a difference. This was the conclusion, for instance, of a seven-year study of heart disease mortality published in The New England Journal of Medicine in 1998, which suggested that the death rates are declining largely because doctors and emergency medical service personnel are treating the disease more successfully.² American Heart Association statistics support this view: Between 1979 and 2003, the number of inpatient medical procedures for heart disease increased 470 percent. In 2003 alone, more than a million Americans underwent cardiac catheterizations; more than a quarter million had coronary artery bypass surgery.

The percentage of Americans who smoke cigarettes has dropped considerably—from 33 percent of Americans over eighteen in 1979 to 25 percent fifteen years later. This should have also significantly reduced the incidence of heart disease. That it hasn’t, strongly suggests we’re doing something that counteracts the beneficial effect of giving up cigarettes. Indeed, if the last few decades were considered a test of the fat/cholesterol hypothesis of heart disease, the observation that the incidence of heart disease has not noticeably decreased could serve in any functioning scientific environment as compelling evidence that the hypothesis is wrong.

Rule 5: If you’re going to speculate from association to causality, you must be an equal opportunity speculator.

Recall what I said above: The switch from SFA to PUFA corresponded not just with “major health benefits,” as Tayag wrote, but with the emergence of the obesity and diabetes epidemics—quite the opposite of major health benefits.

If a reporter (or the authorities she’s citing) is going to associate a decrease in CHD mortality with a switch from SFA to PUFA and speculate as to possible causality, then she (and they) should do that with all other relevant associations as well.

How do we know that the SFA-to-PUFA dietary guidelines—the stop-cooking-with-tallow advice—disseminated by the authorities did not have unintended consequences? All interventions do, whether medical, dietary, public health, or otherwise.

Indeed, the reason I’m still writing about these subjects is largely because I was naïve enough to take that possibility seriously when it was first suggested to me by a National Institutes of Health administrator early in my research. Here’s the paragraph from GCBC:

I first heard this notion in 1998 when I interviewed William Harlan, then associate director of the Office of Disease Prevention at the National Institutes of Health. Harlan told me that public health experts like himself assumed that if they advised all Americans to eat less fat, with its densely-packed calories, weights would go down. "What we see instead," he said, "is actually weights have gone up, the portion sizes have gone up, the amount we eat has gone up…. Foods lower in fat became higher in carbohydrates and people ate more.”

Yes, the conversion from SFA to PUFA might correlate with a decrease in heart disease mortality, but it also correlates with those skyrocketing rates of obesity and diabetes—the epidemics that we should find so disconcerting. That correlation cannot be ignored. It, too, might be causal.

A reporter cannot speculate as to one and ignore the other, much as her sources (who in this case are largely responsible for the shift) would like to do so.

Let’s go back to that GCBC prologue:

Throughout the world, in fact, the incidence of obesity and diabetes is increasing at an alarming rate. Obesity levels in the United States remained relatively constant from the early 1960s through 1980, between 12 and 14 percent of the population; over the next twenty-five years, coincident with the official recommendations to eat less fat and so more carbohydrates, it surged to over 30 percent. By 2004, one in three Americans was considered clinically obese. Diabetes rates have increased apace. Both conditions are associated with an increased risk of heart disease, which could explain why the incidence of heart disease is not decreasing. It is also possible that obesity, diabetes and heart disease all have a single, underlying cause. The surge in obesity and diabetes occurred as the population was being bombarded with the message that dietary fat is dangerous and that carbohydrates are good for the heart and for weight control. This suggests the possibility, however heretical, that this official embrace of carbohydrates might have had unintended consequences.³

Tayag can dismiss this possibility (or at least its implications) in her Atlantic article because, she writes,

The research has continued to bear out the dangers of saturated fats—and, crucially, the benefits of replacing them with polyunsaturated ones.

Obvious question: Has it?

Rule 6: If you’re going to cite an article supporting your assertion, read the article first and ensure that it does.

Curiously enough, I wrote a sentence in my most recent book, Rethinking Diabetes, that made essentially the opposite assertion of Tayag’s:

The latest reviews of the evidence continue to reiterate how equivocal it [the evidence for the dangers of saturated fat and replacing them with PUFA] remains,.

Tayag and I cited the same review to make our contradictory claims, a 2017 American Heart Association “Presidential Advisory.” She says the evidence “continued to bear out” the dangers of saturated fat, I say the evidence never did and continued to be, at best, equivocal.

I did read the paper, however, that we both referenced.

Here’s what I wrote in Rethinking Diabetes, with the key sentences in bold italics responding to Tayag’s assertion that meaningful evidence continued to accumulate.

When the American Heart Association assembled a dozen authorities to publish a 2017 review—an “AHA Presidential Advisory on Dietary Fats and Cardiovascular Disease”—the authors, many of them long-time proponents of the fat-cholesterol hypothesis, were no longer addressing the issue of whether avoiding fat in the diet would make anyone live longer, but only whether or not it might prevent heart disease. They acknowledged that only four clinical trials had ever satisfied their criteria for adequately testing the hypothesis, all dating to the 1960s and early 1970s, and had produced at least some evidence that the hypothesis was right.⁴

When they treated the data from those four trials together— a meta-analysis—they concluded that diets that replaced saturated fat with unsaturated fats could reduce heart disease risk, albeit not those that replaced saturated fat with carbohydrates, which was the essence of the AHA (and U.S. government) dietary guidelines for all those years. The AHA authorities also found reasons to exclude from their analysis all those trial results, including those of the Women’s Health Initiative, that did not align with their thinking. “Readers may wonder,” they added, “why at least 1 definitive clinical trial has not been completed since then [the 1960s],” and they reiterated all the reasons why, but did not mention that the absence of such a definitive trial should still leave their interpretation open to question. Three years later, a collaboration of a dozen international authorities published a “State-of-the-Art” review in the Journal of the American College of Cardiology proposing a halt to public health recommendations calling for the restriction of saturated- fat-rich foods. “The totality of available evidence does not support further limiting the intake of such foods,” they concluded.

I wrote up my initial assessment of this 2017 AHA presidential advisory a few days after it was published, and the journalist Larry Husten posted it on his CardioBrief website. You can read it here. It’s worth it (or so I think) and goes into great detail about how the AHA and its chosen authorities have thought about this issue. (Having now re-read it, I’m thinking it’s worth reposting on this Substack in the future if you’d like to wait.)

A key point is that the half-billion-dollar dietary arm of the Women’s Health Initiative (2006) that I mention in the above paragraphs was one of two major relevant trials that were completed since the 1970s, that provided evidence speaking to this dangers-of-saturated-fat hypothesis, and that the AHA authorities concluded were unworthy of inclusion in their analysis. These two trials were among the most ambitious and expensive clinical trials ever done: the WHI was the first, the $200 million Look AHEAD trial (2013) was the second. Look AHEAD targeted diabetes but was nonetheless relevant to the SFA->CHD argument. Neither observed any heart-healthy benefits from counseling subjects to restrict SFA and (mostly) consume PUFA instead.

Tayag could have used the Cochrane Collaboration here to make her point; it would be interesting to know why she didn’t. The Cochrane Collaboration publishes (relatively) unbiased systematic reviews of the safety and efficacy of clinical interventions and has also assessed this evidence. (The Cochrane reviews, however, also have to be read critically, because a recent trend is to simplify and so overinterpret in the “author’s conclusion” what was concluded in the review itself.)

Allow me to return to Rethinking Diabetes:

Since the late 1990s, the most authoritative reviews on the benefits and risks of medical interventions such as low-fat or cholesterol-lowering diets have been published under the imprimatur of an international organization known as the Cochrane Collaboration, which was established to provide a standardized methodology that would minimize the effect of bias. Implicit in the Cochrane Collaboration’s existence is the assumption that a considerable number of medical authorities believe analyses like those of the AHA [the Presidential Advisory] are unreliable, that authors will create a rationale for choosing the trials and results they consider meaningful, consciously or unconsciously, based on whether or not the trials resulted in outcomes that align with the authors’ beliefs.

Since 2000, researchers from the University of Bristol in the United Kingdom have been using the Cochrane methodology to review the evidence for the benefits of low-fat or cholesterol-lowering diets [[i.e., those that replace SFA with PUFA]]. They have updated the reviews every half-dozen years, but their conclusions have changed little because no new meaningful trials have been conducted since the Women’s Health Initiative published its results in 2006.⁵ In the 2020 edition of their review, the Bristol researchers suggested in their summary statement that the existing evidence is still sufficient to recommend low-fat diets for the population. But they also concluded that they could find “little or no effect of reducing saturated fat on all-cause mortality”—i.e., the chance of dying prematurely—or even on cardiovascular disease mortality, which would be the chance of dying prematurely of a heart attack. “The effects on total (fatal or non- fatal) myocardial infarction, stroke and CHD events (fatal or non- fatal) were all unclear,” they added. The authors also stated that the data “are too limited to be able to answer [the] question” of whether or not the public health pronouncements to avoid dietary fat and specifically saturated fat have been beneficial.

Summing up the story with guidance from some earlier authorities

Tayag concludes her article by saying that the arguments favoring saturated fats “can largely be split into three categories.” Let’s go through them:

The first “questions the validity of the research that established the harms of saturated fat” is what we’ve been discussing. If I’m right, Tayag’s interpretation of this evidence is skewed because her sources are conflicted and biased. She trusts the authorities without knowing why she shouldn’t.

She is trapped, though, as health reporters are, in a vicious circle: The authorities (i.e., the “experts” or the scientists who say whatever it is the reporter wants to communicate) tell her to trust their interpretation of the evidence, and the only way she might conclude otherwise is by distrusting the authorities. However, the only reason she should distrust the authorities, is if she has reason to believe their interpretation of the evidence is flawed.

In an ideal world (ok, my ideal world), Tayag would go to journalists like me and Nina Teicholz, who have been arguing that the authorities are wrong. She would want to find out why we’re so skeptical if nothing else. We would then direct her to sources in academia who believe as we do and whom she might consider more reliable sources because they’re not writing books for a living. (Even if their beliefs are based on reading our books.) But Tayag would have to know about our work to do so, and she’d have to have reason to consider us reliable sources, particularly as the authorities she trusts (mostly) do not.

“The second category,” Tayag writes, “alleges the harms of seed oil.”

Now she refers to “tallow truthers,” which is clever but exposes a disdain for the convention-buckers that suggests a prematurely closed mind. It’s an easy hit, but if she’s wrong, the “tallow truthers” got it right, and it’s the SFA-will-kill-you crowd who are more akin to zealots, a suggestion I make in the epilog of GCBC.

As I have written, I also find the evidence that seed oils are a major cause of chronic disease in America less than compelling. I remain skeptical. But I think I know enough not to defend it on the basis that replacing saturated fat with polyunsaturated was, and still is an unequivocally good idea. All benefit, no harm. (We should be so lucky.)

“The third category,” writes Tayag, “which is perhaps the most puzzling, comprises a bona fide enthusiasm for tallow—which, to be fair, makes a delicious french fry.”

Again, this is circular. If the SFA research is as flawed as Nina, I, and others believe we have documented, and as the authors of the AHA scientific statement essentially acknowledge, then bona fide enthusiasm for tallow is understandable. A delicious French fry is worthy of enthusiasm (much as I might think that the carbohydrates in the potato are less than benign for those who struggle with their weight and their blood sugar—i.e., most Americans).

Finally, writes Tayag:

The crux of the anti-seed-oil, pro-tallow position is a belief that the medical consensus on dietary fats is compromised by financial interests—of the seed-oil and medical industries, of universities, of the government...

For Kennedy and his supporters, the science isn’t really the point—bucking convention is. Rejecting the consensus about saturated fats makes a political statement.

Whether or not financial interests have compromised the science, the crux of the position is what it always was: the quality of the evidence, particularly in the context of dietary guidelines for an entire nation.

The evidence is the point, the science, and its translation to public health guidelines.

The epidemiologist Geoffrey Rose, one of the most respected in the field, argued this position back in 1981 when he published a seminal article in the BMJ—“Strategy of prevention: lessons from cardiovascular disease”—that informed much of the thinking on chronic disease prevention ever after.

We can “usefully distinguish two types of preventive measures,” Rose explained. One removes or restricts access to unnatural factors and so restores what he called “`biological normality’—that is, of the conditions to which presumably we are genetically adapted.”

Advising against cigarette smoking, for instance, is an obvious example. (Rose also believed that telling people to eat less saturated fat was justified. As was common then, he thought modern diets were unnaturally saturated fat-rich.) “Such normalizing measures,” he wrote, “may be presumed to be safe, and therefore we should be prepared to advocate them on the basis of a reasonable presumption of benefit.”

But the second type of preventive measure is fundamentally different: not removing an unnatural factor that appears to cause a disease but adding an unnatural factor with the hopes of conferring protection. That’s where he specifically included the kind of PUFA-rich diets we are eating today as a result of the not-cooking-with-tallow advice, and why PUFA-rich diets made him nervous (my bold italics):

“The end result is to increase biological abnormality by an even further removal from those conditions to which we are genetically adapted. For coronary heart disease such measures include a high intake of polyunsaturates and all forms of long-term medication. Long-term safety cannot be assured, and quite possibly harm may outweigh benefit. For such measures as these the required level of evidence, both of benefit and (particularly) of safety, must be far more stringent.”

This remains the crux of the anti-seed oil position: refined seed oils, like ultra-processed foods in general (a subject for a later post), are new foods entirely to the human race. They’re unnatural additions to our diets.

As Rose said, with such industrial-era products, drugs or dietary, we always have reason to be anxious that they will do more harm than good. That’s why we do not have public health guidelines recommending we all take statins and (not yet, anyway) the new obesity and diabetes drugs.

Living, as we are, with those don’t-cook-with-tallow-associated epidemics of obesity and diabetes, I would think we would be particularly anxious today. Half a century of uncontrolled epidemics seems a good reason to remain open to any reasonable hypothesis that might explain them. Even a hypothesis that RFK Jr happens to believe.

Back in 1979, this is what Phillip Handler testified to in Congress, as this very controversial SFA—>CHD ball was getting rolling. This was, I might add, after all four trials that the AHA Presidential Advisory considered so significant had been completed.

Handler was a National Medal of Science winner, who had spent his career studying human and animal metabolism. At the time, he was president of the National Academies of Science. In short, he was the kind of authority Tayag might have trusted (as was Rose).

Handler had summed the situation up memorably. We might conclude, he testified to the assembled Congressman, “as some have,” that there exists a “thinly linked, if questionable, chain of observations” connecting fat and cholesterol in the diet to cholesterol levels in the blood to heart disease.

However tenuous that linkage, however disappointing the various intervention trials, it still seems prudent to propose to the American public that we not only maintain reasonable weights for our height, body structure and age, but also reduce our dietary fat intakes significantly, and keep cholesterol intake to a minimum. And, conceivably, you might conclude that it is proper for the federal government to so recommend.

On the other hand, you may instead argue: What right has the federal government to propose that the American people conduct a vast nutritional experiment, with themselves as subjects, on the strength of so very little evidence that it will do them any good?

Mr. Chairman, resolution of this dilemma turns on a value judgment. The dilemma so posed is not a scientific question; it is a question of ethics, morals, politics. Those who argue either position strongly are expressing their values; they are not making scientific judgments.

When Tayag suggests in her last paragraph that “as far as scientists can tell,” rejecting the consensus about saturated fats “is not going to make anyone healthier,” she is establishing that she and I have very different conceptions about what a scientist is.

I would start my definition with whether they have an open mind and are aware that when the evidence is this tenuous, they are, as Richard Feynman might have said, very much likely to be fooling themselves. This is independent of whether they are affiliated with any respected academic institution, let alone Harvard.

1. Perhaps because they all work at the National Heart, Lung, and Blood Institute, they include only the advances for which the NHLBI can get some credit--those directly related to the circulatory system, the heart, and the coronary arteries, and so cholesterol, hypertension, smoking, coronary stents, etc. That’s what the NHLBI does.

2. I wrote “ten-year study” in the book. That was an error. The same group published ten-year data in 2001, but it was seven in the NEJM article.

3. In my book and work, I have focused on the replacement of SFA and total fat with carbohydrates. The seed-oils-are-toxic argument focuses on the replacement with PUFA.

4. These included the Los Angeles Veterans Administration trial (1969), which suggested that the substitution of PUFA for SFA might increase cancer risk; a British trial published in 1968 in men who had already had heart attacks and replaced the saturated fat in their diets with soybean oil; a non-randomized study in a Finnish mental hospital study (1972); and an unblinded study in Oslo (1966) carried out by a physician who counseled his subjects to avoid saturated fat and eat more unsaturated fats (from fish, for instance). The Oslo findings were compromised by the fact that the physician’s fish-eating patients also reduced their sugar consumption by almost half, which could have been, among many other possibilities, why they seemed to have less heart disease.

5. Because the LOOKAHEAD subjects were obese patients with type 2 diabetes, the Cochrane methodology rejects its relevance.