Monday, October 27, 2008

Saturated Fat and Health: a Brief Literature Review, Part I

Even years ago, when I watched my saturated fat intake, I always had a certain level of cognitive dissonance about it. I knew that healthy non-industrial cultures often consumed large amounts of saturated fat. For example, the Masai of East Africa, who traditionally subsist on extremely fatty milk, blood and meat, do not appear to experience heart attacks. Their electrocardiogram readings are excellent and they have the lowest level of arterial plaque during the time of their lives when they are restricted (for cultural reasons) to their three traditional foods. They get an estimated 33% of their calories from saturated animal fat.

Then there are the Pacific islanders, who often eat large amounts of highly saturated coconut. Kitavans get 17% of their calories from saturated fat (Americans get about 10% on average), yet show
no trace of heart disease, stroke or overweight. The inhabitants of the island of Tokelau, who I learned about recently, eat more saturated fat than any other culture I'm aware of. They get a whopping 55% of their calories from saturated fat! Are they keeling over from heart attacks or any of the other diseases that kill people in modern societies? Apparently not. So from the very beginning, the theory faces the problem that the cultures consuming the most saturated fat on Earth have an undetectable frequency of heart attacks and other modern non-communicable diseases.

Humans have eaten saturated animal fat since our species first evolved, and historical hunter-gatherers subsisted
mostly on animal foods. Our closest recent relatives, neanderthals, were practically carnivores. Thus, the burden of proof is on proponents of the theory that saturated fat is unhealthy.

There have been countless studies on the relationship between saturated fat and health. The first studies were epidemiological. Epidemiological studies involve collecting data from one or more populations and seeing if any specific factors associate with the disease in question. For example, the Framingham Heart study collected data on diet, lifestyle and mortality from various diseases and attempted to connect diseases to lifestyle factors. This type of study is useful for creating hypotheses, but it can only determine associations. For example, it can establish that smokers tend to die more often from heart disease than non-smokers, but it can't determine that smoking is actually the cause of heart disease. This is because multiple factors often travel together. For example, maybe smokers also tend to take care of themselves less in other ways, sleeping less, eating more sugar, etc.

Epidemiological data are often incorrectly used to demonstrate causality. This is a big problem, and it
irritates me to no end. There's only one way to show conclusively that a diet or lifestyle factor actually causes something else: a controlled trial. In a controlled trial, researchers break participants into two groups: an intervention group and a control group. If they want to know the effect of saturated fat on health, they will advise the participants in each group to eat different amounts of saturated fat, and keep everything else the same. At the end of the trial, they can determine the effect of saturated fat on health because it was the only factor that differed between groups. In practice, reducing saturated fat also involves either increasing unsaturated fat or decreasing total fat intake, so it's not perfect.

I'm not going to review the epidemiological data because they are contradictory and they are "lesser evidence" compared to the controlled trials that have been conducted. However, I will note that Dr. Ancel Keys' major epidemiological study linking saturated fat consumption to heart disease, the "Seven Countries" study, has been thoroughly discredited due to the omission of contradictory data (read: the other 15 countries where data were available). This was the study that sparked the anti-saturated fat movement. Older epidemiological studies and those conducted internationally tend to find nonexistent or weak links between saturated fat and health problems, while more recent American studies, such as the Nurses' Health study, have sometimes found strong associations. I'll address this phenomenon in another post.

In the next post, I'll get into the meaty data: the controlled trails evaluating the effect of saturated fat on health.

Thanks to Rockies for the CC photo.

No comments:

Post a Comment