Gary Taubes: The Complications of Cholesterol

In 2005, my wife and I interviewed Gary Taubes at his apartment in New York City. We were planning on doing a documentary on diet and health, and during this time (January 2005-May 2005) I went on a fast food diet. I was basically doing Fat Head before (I think) Fat Head. (Tom Naughton did an incredible job.)

Whether the documentary will eventually see the light of day is unclear, but I wanted to share some clips on the blog that we produced along the way.

"In the case of diet and heart disease, Ancel Keys's hypothesis that cholesterol is the agent of atherosclerosis was considered the simplest possible hypothesis, because cholesterol is found in atherosclerotic plaques and because cholesterol was relatively easy to measure," writes Gary Taubes in Good Calories, Bad Calories (GCBC). "But as the measurement technology became increasingly more sophisticated, every one of the complications that arose has implicated carbohydrates rather than fat as the dietary agent of heart disease." In this 2005 interview, before the publication of GCBC, Taubes discusses the cholesterol story. A more comprehensive account can be found in Chapter 9 of GCBC: Triglycerides and the Complications of Cholesterol




Below is a transcription:

[00:02:30.06] It's so bizarre, that when you realize what's happening -- I'm going to try and tell it. See if I can get this straight. They come up, they have this guy John Gofman, Berkeley, has this device that can fractionate lipoproteins. Basically, it can take the lipoproteins, which are the carriers of the cholesterol, from your blood. Say you've got so much LDL, so much VLDL, very low density, so much IDL. 

[00:03:02.17] And look, VLDL seems to be as much of a risk factor if not more. If you believe his equations, it's a greater risk of heart disease than LDL. And VLDL carries the triglycerides. Then another group of scientists, Pete Ahrens and some other people, Margaret, whose name I'm going to forget at the moment, who really did great work and I shouldn't forget her name, and it's embarrassing.

[00:03:30.20] Margaret [Albrink] got marginalized. She was at Yale and she took a job at West Virginia and it was like 'poof' -- West Virginia, who cares? Her husband got a job there, we don't care what you have to say. Anyway, they pushed. They said triglycerides, they're as great a risk factor as cholesterol. The studies show it. So in '67, this guy Donald Fredrickson and Robert Levy and a guy named Lees, they're at NIH and they put this five part series in the New England Journal of Medicine on -- it's called lipoprotein disorders, basically, and they classify them as five different lipoprotein disorders.

[00:04:04.02] Some of them are low LDL and some of them are low VLDL or high VLDL or high triglycerides. And they want to know how these things, the portion in the population at large, because they've just been studying patients who were referred to them at NIH. Frederickson and Levy are big deals at NIH. Back then it was the National Heart Institute instead of NHLBI. 

[00:04:26.08] So they get funding to give money to Framingham and five other big population studies, to see how much LDL -- how these lipoprotein disorders appear in the population. For the first time ever, these studies are going to measure LDL and VLDL cholesterol, total cholesterol. Nobody's ever measured VLDL or triglycerides in large populations. It's never been done. Now for the first time ever these big studies like Framingham have the money to do it.

[00:05:01.23] Frederickson, Levy and Lees also give them a ... they come up with a technology that makes it relatively inexpensive to measure these things. You don't need this huge equipment that Gofman had at Berkeley. So they go off to do this study and the problem is you can't measure LDL directly. Don't know the details why, but it can't be done. What you have to do is measure total cholesterol, triglycerides, and HDL and you do this equation and you calculate LDL.

[00:05:31.28] So they have to measure HDL also. Now, since 1950 or '51, I forget which, people have been saying HDL is the single most important risk factor for heart disease, and it's been ignored because if HDL is high, that means it's good for you. It's part of total cholesterol and Ancel Keys has been saying that total cholesterol is bad for you and that has to be low. And how do you make sense of part of total cholesterol being good, when total cholesterol is bad. And this is very confusing.

[00:06:01.05] So HDL has been ignored. Now they have to measure HDL in order to come up with the number for LDL. The head of -- the guy who does the biostatistics for Framingham and NIH and all these studies [Tavia Gordon] decides he's got the HDL data, this is now mid-70s, let's look at it. So he looks to see what kind of risk factor HDL is and lo and behold, HDL is four times a better predictor of heart disease than LDL. Total cholesterol, he finds out, is meaningless. 

[00:06:30.23] Total cholesterol doesn't predict heart disease. This is what's being measured. It's meaningless. LDL is a "marginal predictor." This is the phrase they use. And HDL, four times better. Four times more accurate at predicting heart disease. If HDL is low, that's a powerful indicator that you're going to get heart disease. And on top of it, for women, HDL is the only number that matters. So in 1976 they publish these articles. Two articles in particular.

[00:07:01.23] One from Framingham alone. And one from Framingham and the other five groups together. Two different types of studies. Both of them say the same thing. Total cholesterol is meaningless. LDL is a marginal predictor of heart disease risk. HDL is the single best predictor. And triglycerides are also a good predictor. Now the problem is, you have a half billion dollars of studies dedicated to lowering total cholesterol, the LRC, Lipid Research Clinics trial, and MRFIT trial, are both aimed at lowering total cholesterol.

[00:07:36.22] So what they say, and you see this in the papers. They say total cholesterol is meaningless but LDL makes up the biggest part of total cholesterol and we know that LDL is a marginal predictor. So we're going to play up how good LDL is. And as you watch the papers the adjectives used to describe it as a risk factor get more and more ... zealous, as the time ... it goes from being marginal to robust, without the data ever changing.

[00:08:08.12] And we're going to say everything we had said about total cholesterol, all the studies from the 1950s and 1960s, the famous seven countries studies, had all measured total cholesterol. They say "everything we say about total cholesterol is now going to apply to LDL cholesterol. And the MRFIT?? study and the Lipid Research Clinic study we're going to say are aimed at lowering LDL cholesterol.

[00:08:29.23] That's what we're trying to do because LDL is a predictor." Now the problem is, that in the clinic, where doctors all around the world are measuring total cholesterol, they're measuring something that's meaningless. But you can't measure LDL cholesterol. They don't have a test for it. You've got to measure all these other things and calculate LDL cholesterol which is too difficult to do.

[00:08:56.20] And it's not the kind of thing that they can tell doctors to do, because they're going to have all kinds of problems with mistakes and errors, and all these things that have to be worked out that they can't do. Meanwhile they have the doctors measuring total cholesterol. They're doing it all over the world. They say, in the clinic -- and you read this in the articles, I'm not making this up -- in the clinic, total cholesterol is a good enough substitute for LDL cholesterol that people can keep measuring total cholesterol and use that as their guide. So the study says that total cholesterol is meaningless.

[00:09:29.27] In the scientific discussion we're going to replace LDL ... total cholesterol with LDL cholesterol, because that's a marginal predictor, and then in the clinic we're going to re-replace LDL cholesterol with total cholesterol going back to the meaningless predictor on the basis that it's a close enough substitute for LDL that it'll suffice. So within three years, they can say that doctors should continue measuring total cholesterol, even though they now know that total cholesterol is meaningless. Meanwhile HDL and triglycerides just still continue to get left out totally because what do you do about that?

[00:10:00.15] And it's four times the predictor. [interviewer] And the problem ... the interesting thing is ...