8 The Journalist Gary Taubes 8: Anomaly Hunter 2
With Framingham, Puerto Rico, and Honolulu behind us, Chicago is the next stop on our anomaly tour.
The Chicago reference Gary Taubes uses is one of the research papers to come out of what is commonly called the Western Electric Study, which was so named because of the company where its subjects worked. The research from the Western Electric study is very well-known and is considered to be supportive of diet-heart, so once again, Taubes seems to have chosen poorly if he is trying to piece together evidence that supports his views.
Here is the line from this reference that Taubes must have liked. “Worthy of comment is the fact that of the 88 coronary cases, 14 have appeared in the high-fat intake group and 16 in the low-fat group.” With more coronary cases in the low-fat group, this does seem puzzling at first glance.
That was not the only finding published in this paper. The authors also wrote that “a definite relationship was demonstrated between higher blood cholesterol levels as found the first year, and subsequent incidence of clinical coronary heart disease.” “This relationship is highly significant,” the authors wrote. The authors found cholesterol to have a comparable influence on heart disease as was found for cigarette smoking. I’m just curious, Mr Taubes. Do you dispute their findings that high cholesterol was associated with heart disease risk and that it was found to be as dangerous as smoking? Shall we take your use of this reference as an endorsement of its observation that high cholesterol is linked to heart disease?
Notice the line on the left of your screen in reference to the observed relationship between cholesterol and heart disease. “To those who decry the significance of blood cholesterol findings in population groups, this should be food for thought.” It didn’t prompt much thought from Taubes, apparently.
So what about that line from the beginning about the high- and low-fat groups? Was a low-fat diet shown to be worse for heart disease? Well, notice how these groups were defined. “Low fat” here meant 39% of calories from fat. That’s a bit of a laugher. Do you think there were any whole food vegans in that bunch of men at that plant in Chicago in the 1950s? “Low fat” can mean a lot of things, and back then, it probably didn’t mean a varied diet based on fresh whole plant foods. Gary Taubes may not understand that. Take a moment to notice the reference at the bottom to “regression toward the mean.” I’ll talk about a similar concept in a later video.
Our first problem here, as you saw, was the definition of “low fat” in this paper. Another problem was with the way their data on food consumption was collected. The men self-reported during interviews. Here’s a great section illustrating what’s wrong with studies based on interviews. The men said they were embarrassed to admit to all the food they really ate. We shouldn’t believe their food records were accurate. All studies have their weaknesses but that is a pretty serious one for this study. Taubes found a superficial anomaly that can be explained by their very soft data for actual food consumed and by their arbitrary and inappropriate definition of “low fat.” The finding that high cholesterol is dangerous was the headline, but Taubes didn’t retain that information.
This is another paper that came out of the Western Electric study much later than the previous one. Taubes did not reference it for good reason. The consumption of saturated fat, the levels of cholesterol in their blood, and the risk of heart disease all joined to favor the diet-heart idea. This research covered 19 years. The authors said, “The results support the conclusion that lipid composition of the diet affects serum cholesterol concentration and risk of coronary death in middle-aged American men.” This was a better set of data, but the journalist Gary Taubes didn’t like it. Here you see the Western Electric Study without the Taubes Filter.
This second paper ran down the different problems with prospective studies of cholesterol such as the potential for inaccurate food records. This is worth reading. We live in a complicated world. These issues must all be accounted for somehow if we are more interested in science than selling books.
Interestingly, in this study dietary cholesterol stood out more than saturated fat as a factor associating with death from heart disease. Be sure to watch my video about dietary cholesterol later in this playlist.
Here is one last piece of research from Western Electric that the journalist Gary Taubes didn’t want in his book. This paper came from the same lead author of his chosen reference. Inconveniently for Taubes, it addresses the Yudkin hypothesis regarding sugar and heart disease. The sugar hypothesis received no support here. They called any association with heart disease insignificant and likely confounded by smoking behavior. The beliefs of Yudkin failed to predict reality again.
Mr Taubes, don’t you think Francis Bacon would find you to be a rather stubborn character?
After Chicago comes Tecumseh, Michigan.
This is his pick for Tecumseh. Cholesterol, as well as triglycerides, were unrelated to diet as determined by the 24-hour recall method. Now are you saying that the diet can’t influence triglycerides either, Mr. Taubes? As you will see later on, Mr. Taubes think carbs are bad because they raise your triglycerides. Once again, he can’t seem to get his story straight.
The authors said that when their subjects were divided up according to their blood test results, their various diets didn’t relate well to their numbers on those tests. Well, would you expect to see nice, clean relationships if they were all eating similar diets?
We’re back to his admission that “this might have been true” that everyone in these places ate pretty much the same way.
This study was done in the 1970s in Tecumseh, MI. Even in the 2010 census, Tecumseh was 96% white. This town probably lacked much cultural, and therefore dietary, diversity. Any variations in cholesterol would therefore have had to have been influenced by other factors like age and genetics.
Otherwise, Taubes is once again contradicting his statement that saturated fats raise cholesterol. If you are just a cholesterol confusionist, it is not necessary to be consistent.
The authors of this study that Taubes picked were very clear about the lack of dietary variation in their participants. All the people there ate a lot of fat. For this reason, the authors reported that the “results of this study do not refute the striking correlations between fat consumption and serum cholesterol levels found” in cross-cultural studies. Taubes had his explanation right there in this very study and yet he pretended that these findings could not be explained.
Despite the homogeneity of their population, the researchers did detect a faint indication that carbs and fiber showed an inverse relationship with cholesterol, exactly as we might expect.
What is more important for you, dear viewer, to understand is that in Tecumseh, just as in France, high blood cholesterol was a risk factor for coronary heart disease.
The authors of this paper Taubes picked did come up with a funny conclusion. “The absence of correlation between dietary intake and lipid levels indicates that variations of dietary intake are less significant determinants of serum lipid levels among free-living individuals than would be anticipated from the findings of metabolic ward studies conducted under controlled conditions.” This is quite silly if you think about it. Apparently people must magically change once they set foot inside a research facility. When people aren’t being observed, there cholesterol metabolism changes, you see. In this study as in so many others that have created confusion, only a cross-sectional approach was used. It represented just a snapshot so it couldn’t account for changes over time. I’ll explain why cross-sectional studies can be weak evidence when not interpreted correctly in my video called The Measurement Problem.
The next anomaly was found in Evans County, GA.
In this one, the researchers found some men with high cholesterol and some men with low cholesterol, only 26 in all, and matched them with randomly chosen age-matched men. They weren’t matched for anything else beside age. Between that and the small number of men they studied, this was one weak study.
They took only two cholesterol measurements, and they only used a 24-hour diet recall, so their data was quite soft.
Notice that the ages of their participants went up to 74. This was back in 1965. Are you going to tell me they had low cholesterol because there were a bunch of old vegans hanging out in the Deep South in the 1960’s, or will you consider that some of these guys were probably dealing with health problems that would have lowered their cholesterol levels? This study was just nothing.
The authors warned against the over-interpretation of their findings. They had only a small group of men. They didn’t want too much made out of this paper. Apparently Taubes can’t distinguish a strong study from a weak one.
Some of the same authors also looked at both whites and blacks in Evans County in another paper published that same year. This one might give us a chance to see some diversity of diet. They found something we don’t often see in the old research. In comparing whites of high social class with whites and blacks of lower social class, the upper class whites were found to consume less animal fat. Think of some of the fatty traditional foods among the poor in the South in those days and this actually makes a lot of sense. Physical activity was a more obvious predictor of heart disease than diet under these circumstances.
However, they did find something very interesting about diet and cholesterol. Blacks ate more carbs, less protein, and had lower serum cholesterol. They also had lower heart disease prevalence. Didn’t you see this one, Mr Taubes?
As you see, we are now up to the last anomaly on this slide. Don’t worry, we’ll look at plenty more of these going forward. But for now just notice that I haven’t yet even made it through a single sentence composed by the journalist Gary Taubes.
Taubes found an oddball study from Israel to finish out his sentence.
This is it. The major finding, the authors said, was that diet didn’t seem to relate to serum cholesterol. Again, is Taubes trying to say that diet cannot affect cholesterol? What point is he trying to make with all these?
Look at the subgroups involved in this study. They were broken up by ethnicity. You can see that the groups all ate pretty much the same amount of saturated fat. That’s not my main point with this slide. Look at this comparison of the Israeli group and the Asian group. Look at the last little box I highlighted in each column. That is for weight to height ratio. Do you see how the Israelis are 245 and the Asians are 232? The Asians were lighter. Now look to the next box up. That’s calories. The Asians ate 100 more daily calories than the Israelis, and yet they were thinner. How did they do that? Move up to the next box. They ate more bread. Move up again. They ate many more starches and carbs. Now look at the top box. They ate less fat. Mr Taubes, don’t hide this amazing information from your readers! You just found a study that shows us that you can eat more calories and weigh less if you choose carbs over fats! You did notice that, didn’t you? I mean, you’re such a thorough researcher!
So back to the matter upon which Taubes would prefer we focus. The authors emphasized that they were comparing individuals, not groups in this one. They acknowledged that group averages would give us a better comparison for the effect of diet on cholesterol because individuals vary so much.
They said they wanted to emphasize that they were not studying the relationships between their groups' diets and their cholesterol levels. Taubes is not using this study the way they wanted it to be used, although they must have known some people would look at it and make group comparisons anyway. More on that in a moment.
They did say, for what it’s worth, that blood cholesterol among the Africans was lower than among the Eastern or Central Europeans.
And if you look at our table again, you can see that the Africans consumed the least saturated fat and the Eastern and Central Europeans the most. This is exactly what we would expect to see. Again, Mr Taubes, did you not look at this table? Do you not find this interesting? It affirms diet-heart.
I should point out that they also were not commenting on any of this in relation to heart disease incidence in this paper. This paper had a lot of limitations and was a bit confusing.
Nevertheless, the authors did leave the impression that cholesterol levels in groups were unrelated to diet in their study, and that made them an obvious target for the merciless scrutiny of Ancel Keys. Keys pointed out, as I did, that there in fact was a relationship among the various groups for saturated fat and serum cholesterol. There were no surprises in this paper and so there were no true anomalies, either. Keys had criticisms of their methodology, and I find it very interesting that the authors apparently never responded to these criticisms. They never defended their work, but Taubes decided to use it anyway.
Keys said they weren’t thorough enough with their cholesterol measurements. Many blood samples are needed because cholesterol does not hold steady from day to day.
He said that metabolic ward experiments, which were rigidly controlled and during which many measurements were taken, showed very clearly that different fats have different effects on cholesterol. Isn’t that the sort of controlled experimentation Taubes says he likes? Keys knew that intraindividual variability must be considered in study design. “Intraindividual variability” is just a fancy way of saying that in the same person, cholesterol can go up and down naturally. It doesn’t hold steady all the time.
Keys rightly pointed out that the authors of the paper from Israel did not deal with this important factor at all. Keys finds their methods to be an object lesson in faulty research practices.
Here you see Keys wrote about a very common problem with studies that you will see even today. Investigators will decide to look at one food, like eggs, and they will ignore other dietary sources of cholesterol and saturated fat. Then they will make a conclusion saying that eggs don’t affect cholesterol. This seems like a completely obvious point, but you will see in my video about dietary cholesterol that researchers make this error even now.
Taubes found a couple other anomalies that I learned about through this news article. We are told that he said there was a study in Trinidad showing that women were obese despite low caloric intake, and we learn that Mexican-American oil-field workers in Texas became overweight without eating fast food. I had to look into this more to figure out what he was talking about.
This is where this nonsense appears in his recent book, Why We Get Fat. Let’s start with the one on the bottom. That is the one where women in Trinidad were obese even though they were consuming less than the minimum recommended number of calories by the FAO.
We see in the study he cited that obese and control subjects in Trinidad did not differ in their caloric intake. So what do you think happened here? Choice A: The difference between the normal and the obese here had nothing to do with their diets. The fat people were special. Or Choice B: The fat people did not report their calories accurately? This isn’t a hard one.
They obtained their diet information from food diaries and from interviews which is bad enough, but as you can see in the first box, the women were not necessarily literate. If they couldn’t read and write, we at least have the assurance that someone in the family could read and write. So now, in addition to the usual accuracy issues with diet reporting that Taubes knows are problematic, we add the additional layer of illiteracy. The women admitted to “light snacking of high-carbohydrate foods”. What does that mean? Fruit or candy? Additionally, the obese women were said to be less physically active than the normal weight women. This is the study Taubes is using to convince the clueless people in his audiences that the mainstream message on calories is wrong?
In this 1995 paper, this very study chosen by Taubes was examined along with others that used self-reported food intake and came up with counterintuitive results. These researchers wondered about the reliability of records of food intake reported by obese individuals.
By using more precise methods of measuring caloric intake, they were able to assess the accuracy of the food diaries of 16 individuals over a four year period who were obese or who reported unexplained weight gain. It was found that these individuals were underreporting their energy intake by more than 100%!
This is another study that referenced the Trinidad paper Taubes used. These researchers used similar methods to what you just saw as they studied reported and actual food intake by 27 obese young people and reached a similar conclusion despite training the subjects in proper food measurement and despite the use of financial penalties for inaccuracy. They said, “our data strongly indicate that reported metabolizable energy intakes in obese or nonobese adolescents are not representative of energy expenditure or energy requirements.” Taubes is going around talking about this pathetic study to play to the excuse-making, victimhood mentality of many obese individuals. This argument is pure low-carb low standards and it’s how Taubes has made a nice living for himself.
Now we look at the other piece of anecdotal nonsense Taubes has unearthed. He tells us that William Mueller measured and weighed 1100 Mexican-Americans. A huge proportion of them were obese. He reports that Mueller said there was only one restaurant in the county where they lived. What the heck is this one about? Is Taubes saying that one restaurant fed all the 1100 people Meuller weighed every day for all their meals? Is he saying there was no other source of food in the county? What is his point?
I looked up Mueller’s research from that time and place. This paper is about how people who carried their fat mostly around their stomachs had worse risk factors for cardiovascular disease. There is nothing in here about people getting fat without eating.
Here’s another. Cortisol is said to increase the degree of fatness around the stomach. That's not good. Cortisol is a stress hormone. This one reminds me of another paper Taubes likes.
Taubes wrote in the New York Times about a study conducted by David Ludwig that showed a higher resting metabolic rate for subjects eating a low carb diet. According to Mr Taubes' warped priorities, that's supposed to be good somehow.
Nowhere in his Times piece does Taubes mention that the low carb diet in that study also raised the stress hormone cortisol, a very worrying development. It also increased C-reactive protein, a marker of inflammation. Let’s listen to what Dr Ludwig had to say about that.
LUDWIG: … But there seemed to be a catch. That is that we saw … a tendency for chronic inflammation to increase.
TOM ASHBROOK: Not good for heart health.
LUDWIG: And the other perhaps more concerning change was that the stress hormone cortisol – that’s the ... body’s version of the medication cortisone that you buy in (the) supermarket – that went up very substantially on the low-carb diet. And both chronic inflammation and cortisol are key long term predictors of cardiovascular health and overall mortality … It may be that when you try to push so many calories through a limited metabolic pathway it stresses the body … There’s never a downside to switching from highly processed to unprocessed carbs, and from switching from predominantly saturated fats to more unsaturated fats.
That is Dr Ludwig, also speaking without the Taubes Filter.
Interestingly, it has been found in a study of the children of women who practiced a low-carb diet during their pregnancies that those children had higher cortisol levels in later life. They were apparently damaged by their mothers’ diets.
It is also interesting that the original low carbers, the Inuit, were noted to have a higher basal metabolic rate long ago. That didn’t make them healthy. A fever from an infection will increase your metabolic rate and your markers of inflammation, too.
Recall that we started by looking at the writings of Meuller about these Mexican-Americans who produced obesity from nothing. Higher cortisol was found to contribute to a more dangerous body fat distribution. That would suggest that we should avoid dietary practices that raise our cortisol. Therefore, it also suggests we should skip the low-carb foolishness.
That’s all I could find from William Mueller on this topic. I don’t know what is being referenced by Taubes.
That news story that got my attention included this line. Apparently Taubes has found a hedge fund operator with more money than brains who is impressed with all his cobbled together distortions and fake anomalies. Taubes has been able to make a lot of money off of this nonsense, and someone who couldn’t see through all this nonsense has plenty of money to waste on him. No one said life is fair.
Taubes has padded his book with many other tortured references. I’ll get back to digging through this junk in the next video.