Transcript: The Saturated Fat Studies: Set Up to Fail
Where did these consensus guidelines to dramatically lower saturated fat consumption come from? From literally hundreds of metabolic ward experiments, which means you don’t just ask people to change their diets, you essentially lock them in a room, for weeks if necessary, and have total control over their diet. You can then experimentally change their level of saturated fat intake however you want and see the corresponding change in their cholesterol levels. And the results are so consistent you can create an equation, the famous Hegsted Equation, where you can predict how much their cholesterol will go up based on how much saturated fat you have them eat. So if you want your LDL cholesterol to go up 50 points all you have to do is eat like 30% of your calories from saturated fat. When you plug in the numbers, the change in cholesterol shoots up, right as predicted. The experiments match the predictions. You can do it at home with one of those home cholesterol testing kits, eat a stick of butter every day and watch your cholesterol climb—it’s not rocket science. And look at this, this was 1965; we’ve known about this for 50 years that even if you keep calorie intake the same, increases in saturated fat intake are associated with highly significant increases in LDL bad cholesterol. Now your good cholesterol goes up a bit too, but that increase is smaller than the increase in bad, which would then translate into increased heart disease risk overall.
So if you feed vegetarians meat even just once a day, their cholesterol jumps nearly 20% within a month. To prevent heart disease, ideally we would need to get a total cholesterol under 150, which you can see these vegetarians were, but then even just once a day with the meat and their cholesterol shot up 19%. But the good news is that within just 2 weeks of returning to their meat-free diet, their cholesterol dropped back down into the safe range. Note that their HDL good cholesterol hardly moved at all, so their ratio went from low risk of heart attack to high risk in a matter of weeks with just one meat-containing meal a day. And indeed randomized clinical trials show that dietary saturated fat reduction doesn’t just appear to reduce cholesterol levels, but subsequent cardiovascular events like heart attacks as well.
So we have randomized clinical trials, controlled interventional experiments, our most robust forms of evidence—no wonder there’s a scientific consensus to decrease saturated fat intake. This is going to be a toughie for Big Cheese and Chicken. You’ll note, though, that the Y-axis here is not cholesterol, but change in cholesterol. That’s because everyone’s setpoint is different. Two people eating the same diet, the same amount of saturated fat, the same number of chicken nuggets a day can have very different cholesterol levels. One person can eat 10 chicken nuggets a day and have an LDL cholesterol of 90; another person eating 10 a day could start out with an LDL of 120. It depends on your genes. But while our genetics may be different, our biology is the same, meaning the rise and drop in cholesterol is the same for everyone. So if both folks cut out the nuggets, the 90 might drop to 85, whereas the 120 would be expected to drop to 115. Wherever we start, we can lower our cholesterol by eating less saturated fat, but if I just know what your saturated fat intake is—how many nuggets you eat–I can’t tell you what your starting cholesterol is. All I can say with certainty is that if you eat less, your cholesterol will likely improve.
But because of this extreme “interindividual variation,” this wide variability in baseline cholesterol levels for any given saturated fat intake, if you take a cross-section of the population, you can find no statistical correlation between saturated fat intake and cholesterol levels, because it’s not like everyone who eats a certain set amount of saturated fat is going to have over a certain cholesterol. So there’s like three ways you could study diet and cholesterol levels: controlled feeding experiments, free-living dietary change experiments, or cross-sectional observational studies of large populations. As we saw, there is a clear and strong relationship between change in diet and change in serum cholesterol in the interventional designs, but because of that interindividual variability, in cross-sectional designs, you can get zero correlation. In fact, if you do the math, that’s what you’d expect you’d get. In statistical parlance, one would say that a cross-sectional study doesn’t have the power for detecting such a relationship. Thus because of that variability, these kinds of observational studies would seem an inappropriate method to study this particular relationship. So since diet and serum cholesterol have a zero correlation cross-sectionally, an observational study of the relationship between diet and coronary artery disease incidence will suffer from the same difficulties. So again, if you do the math, observational studies would unavoidably show nearly no correlation between saturated fat and heart disease. These prospective studies can be valuable for other diseases, but the appropriate design demonstrating or refuting the role of diet and coronary heart disease is a dietary change experiment. And those dietary change experiments have been done; they implicate saturated fat, hence the lower saturated guidelines from basically every major medical authority. In fact, if we lower saturated fat enough, we may even be able to reverse heart disease, opening up arteries without drugs, without surgery. But wait a second. Let’s put our Big Cheese and Chicken hat back on. Observational studies would show no correlation, mathematically could show no correlation. We’ve known since 1979 that observational studies simply don’t have the power to show the relationship. Bingo!
All we need now is a friendly researcher. How about Ronald M. Krauss? Funded by the National Dairy Council since 1989, also the National Cattlemen’s Beef Association, as well as the Atkins Foundation—perfect. Then you just combine together all the observational studies that don’t have the power to provide significant evidence and what do you know, no significant evidence was found.
This 2010 meta-analysis was basically just repackaged for 2014, using the same and similar studies. As the chair of Harvard’s nutrition department put it, their conclusions regarding the type of fat being unimportant are seriously misleading and should be disregarded, going as far as suggesting the paper be retracted, even after the authors corrected a half dozen different errors.
But it’s not like they falsified or fabricated data—they didn’t have to. They knew beforehand the limitations of observational studies; they knew they’d get the “right” result, and so they published it, helping to “neutralize the negative impact of milkfat by regulators and medical professionals.” And it’s working, brags the dairy industry: "Perceptions about saturated fat in the scientific community are changing. This is a welcome message to consumers, who may be tired of hearing what they shouldn’t eat." They don’t need to convince consumers, just confuse them. Confusion may easily be misused by the food industry to promote their interests.
It’s like that infamous tobacco industry memo that read “doubt is our product.” “Doubt is our product since it’s the best means of competing with the body of fact that exists in the mind of the general public.” They don’t have to convince the public that smoking is healthy to get people to keep consuming their products. They just need to establish a controversy. Some science says its bad, some says it’s not bad. Conflicting messages in nutrition cause people to become so frustrated and confused they may just throw their hands up in the air and eat whatever they want, which is exactly what saturated fat suppliers want–but at what cost to the public’s health?
To see any graphs, charts, graphics, images, and quotes to which Dr. Greger may be referring, watch the above video. This is just an approximation of the audio contributed by Katie Schloer.
Please consider volunteering to help out on the site.