Fat and Carbohydrate: Clarifications and Details

The last two posts on fat and carbohydrate were written to answer a few important, but relatively narrow, questions that I feel are particularly pertinent at the moment:

• Was the US obesity epidemic caused by an increase in calorie intake?
• Could it have been caused by an increase in carbohydrate intake, independent of the increase in calorie intake?
• Does an unrestricted high-carbohydrate diet lead to a higher calorie intake and body fatness than an unrestricted high-fat diet, or vice versa?
• Could the US government's advice to eat a low-fat diet have caused the obesity epidemic by causing a dietary shift toward carbohydrate?

However, those posts left a few loose ends that I'd like to tie up in this post.  Here, I'll lay out my opinions on the relationship between macronutrient intake and obesity in more detail.  I'll give my opinions on the following questions:

• What dietary macronutrient composition is the least likely to cause obesity over a lifetime?
• What dietary macronutrient composition is best for a person who is already overweight or obese?
• Is fat inherently fattening and/or unhealthy?

From the beginning

The first question is: what dietary macronutrient composition favors leanness throughout life?  In other words, if we were able to assign "the average person" to a diet from infancy to old age, what macronutrients would be the least likely to promote obesity?  My answer, of course, is speculation, but there are indirect ways of getting at the question.

Animal studies in a variety of species generally suggest that fat is the most fattening macronutrient, carbohydrate is in the middle, and protein is the least fattening. The ability of fat to cause fat gain seems to depend in large part on its ability increase the calorie density of the diet, suggesting that fat isn't necessarily fattening if it's coming from foods that are lower in calorie density like vegetable dishes and dairy.  Refined carbohydrate is more fattening than unrefined carbohydrate, and presumably the same applies to fat although I've never seen it tested directly.  

Since protein can't reasonably supply the majority of calories in most species, the bulk of calories end up coming from some combination of carbohydrate, fat, and short-chain "fatty acids" produced by the intestinal fermentation of fiber.  The latter are a relatively minor source of calories for humans, who lack the intestinal fermentation capacity of chimpanzees for example, so that leaves us with carbohydrate and fat.  Rodents maintained on unrefined high-carbohydrate diets and without the ability to exercise do gain fat over the course of their lives-- many of them become overweight eventually, and a few become obese.  However, maintaining them on calorie-dense high-fat diets causes them to gain far more fat over the same period of time.  Rodents aren't humans, but the long-term effects of fat-rich diets on body fatness do seem fairly consistent over a number of different species.  

In humans as well, unrestricted diets rich in high-fat foods often lead to a higher intake of calories and higher body fatness than unrestricted diets rich in high-carbohydrate foods, at least in studies lasting weeks (1, 2, 3, 4, 5, 6, 7).  

There are countless examples of traditionally-living cultures that remain quite lean throughout life eating diets that are primarily carbohydrate, although there are also a few examples of cultures eating higher-fat diets that are also lean.  It's important to remember that these diets are eaten in a very different context than in the modern world today: abundant physical activity, absence of highly palatable foods, absence of food advertising, etc.  

In my opinion, it's probably possible to design a diet that favors lifelong leanness and centers around either carbohydrate or fat, however it's likely easier to pull off if the diet is more focused on carbohydrate.  If the diet is high in fat, it would have to be carefully planned so it's not too high in calorie density and palatability.  Likewise, the high-carbohydrate diet wouldn't include low-fat Snackwells; rather, sweet potatoes, potatoes, fruit, beans/lentils, oatmeal, and other simple unrefined carbohydrate foods.  

Already overweight

Now, let's consider a different scenario: we're already adults who are overweight or obese, and we want to lose weight and maintain the loss.  What dietary macronutrient composition is best now? 

We have a lot of evidence we can apply to this question.  It's relatively clear at this point that a carbohydrate-restricted diet is a more effective fat loss tool than a fat-restricted diet, at least over periods up to a year.  Some people find that their appetite normalizes and their positive relationship with food is restored by a low-carbohydrate diet.  Typically, on a low-carbohydrate diet, the proportion of fat and protein increase, although in an absolute sense the increase may not be large because total calorie intake declines.  

Overall, these trials suggest that the primary "active ingredient" of low-carbohydrate diets (at least moderate low-carb) is their high protein content.  In other words, high-protein diets yield a similar fat loss outcome whether they're low in carbohydrate or low in fat.  The low-carbohydrate diet concept seems to be a useful heuristic for getting people to eat a more protein-centric diet, dropping some of the junk foods, and getting people to pay more attention to what they're eating.  

Do the results of low-carb diet studies mean overweight people respond in a fundamentally different way to dietary fat than lean people?  Actually, no.  The shorter-term studies show that high-fat foods lead to overeating whether a person is lean, overweight, or obese-- if anything, overweight people overeat more on high-fat diets, and gain more weight (8).  

The difference isn't in the physiology; it's in the intervention.  A low-carbohydrate diet is a restrictive diet, whereas the high-fat diets used in these shorter-term studies aren't restrictive.  They take a normal, unrestricted diet, and shift it to include more fatty foods.  These diets are higher in fat, higher in calorie density, and higher in palatability, without increasing protein intake.  As opposed to the high-fat diet studies, in which people immediately begin eating more calories, when a person is placed on a restrictive low-carbohydrate diet, he immediately begins eating fewer calories.

Health implications of macronutrients

I don't think fat or carbohydrate are inherently unhealthy.  Many cultures have thrived on carbohydrate-rich diets, and although there are fewer examples of cultures eating fat-rich diets, overall they seem to have stayed relatively healthy as well.  Thinking about this from the point of view of evolution, it makes no sense to design a human that can only eat carbohydrate or fat.  Humans evolved in an environment that contained carbohydrate-rich and fat-rich foods, and some weeks our ancestors got more of one than the other.

It is worth noting that the ancestral African hunter-gatherer diet was probably not high in fat on most days, at least not animal fat.  African game is characteristically extremely lean, and the only African hunter-gatherer group I'm aware of that gets a fair amount of fat is the !Kung, and most of that fat comes from mongongo nuts (although the mongongo fruit/nut is mostly carbohydrate by calories, a fact that Staffan Lindeberg recently pointed out to me).  Most African game just doesn't contain much fat, even if you include the brain and marrow, and the primary exceptions, like hippos, are extremely dangerous to hunt with stone-age weapons (9).  I have yet to see a single credible account of an African hunter-gatherer group that regularly eats a diet high in animal fat.  If you know of one, please cite it in the comments.

I've come across a lot of arguments that the ancestral human diet was typically high in fat, but these invariably strike me as wishful thinking.  One argument I frequently encounter is that the plant foods we ate were mostly "fat", due to the fact that the calories they provided were mostly via fatty acids produced by the intestinal fermentation of fiber.  Therefore, we should eat a lot of lard to replicate this.  However, the short-chain "fatty acids" that are produced by intestinal fermentation are not at all analogous to what we normally think of as dietary fat.  These are technically fatty acids, but is vinegar (acetic acid, one of the primary "fatty acid" products of intestinal fermentation) equivalent to lard?  Of course not.  It's completely different on almost every level, from the role it plays in our diet, to the way in which it's absorbed, to the way in which it's metabolized*.

Fat probably has the greatest potential to be fattening among macronutrients, yet it isn't necessarily fattening.  Nuts, avocados, and dairy are probably not very fattening relative to other foods, although it's possible to overeat almost anything.  Lean meat is one of the most slimming foods, despite the fact that most lean meats still contain a fair amount of fat.

In the end, I believe the best diet is the one that keeps you relatively lean and healthy.  That diet might differ based on your background, current lifestyle, genetic makeup, and goals.  A diet's macronutrient composition is one variable that determines body fatness, although it's probably not the most important variable.  It's simply one of the easiest to understand.

* Acetate and butyrate are absorbed by the liver from the portal circulation as they exit the digestive tract.  In the liver, they're used as building blocks to produce glucose and fatty acids.