Sunday, January 10, 2016

WTF?: The New Dietary Guildelines and the Actual Literature on Saturated Fat and Heart Disease


After listening to the Nina Teicholz conversation with a government spokesperson (good job Nina!) about the new dietary guidelines, I am a bit mad. The doubling down on saturated fat is really disturbing given the evidence is so weak (and I know it is nothing new, the current entrenchment of that view is an unfortunate fact of life). 

But how weak is the research? Well, I will tell you that it is so weak that if every Phd or MD (that understands the difference between observational studies and experiments) was required to review the literature on the actual experiments as part of their education, the government recommendations would be defunct or withdrawn in no time because the vast majority of said students would consider that the statement the "evidence is weak" is a charitable interpretation.

Specifically:
1) All but one of the experiments showed replacing sat fat with PUFA (which the various experiments had to do since no one would eat a diet with all the fat removed) showed cardiac deaths unchanged (or increased) as a result of removing sat fat.

2) The one study that did find an effect is the notorious LA Hospital study which inadvertently left more smokers in the sat fat group (1.5 times the number of moderate smokers and 2 times the number of heavy smokers) among other errors and possible confounding factors (e.g, the sat fat group was Vitamin E deficient which is somewhat bizarre given their butter intake, etc).

This is also the study that people quote when they refer to the idea taking sat fat out will make you die of cancer instead of a heart attack because even after screwing up the design of the experiment, the death rate between sat fat and PUFA groups were he same. Chris Masterjohn put it well when he suggested a more accurate description of the results is that sat fat protects you from cancer. The other thing found in some experiments of this type is that the death rate from neurologically related causes (accidents, suicides) also goes up.

Even the observational data taken in its entirety is not even strong enough to probably have justified the experiments in the first place. (Which was the conclusion of a recent WHO expert panel after doing a review of the existing observational data). If I were a betting man, I would lay money on the proposition that all this is going to be considered one of the great scientific scandals some day.

The Chris Masterjohn review of the literature is my favorite because it is so targeted
followed by Peter Attia's talk which is much broader than just looking at the experiments, you can also view the video version here.