Category Archives: gary taubes

Things That Won’t Kill You Volume 4: Saturated Fat, Part I

I know this is misleading because lard is mostly unsaturated, but it's been one of the major icons of "bad" fat and also, how graet is this image? from

This is probably going to be an even harder sell than MSG, but I swear I’m not just trying to be contrary. It’s true that all the major sources of public health and nutrition advice, including the Harvard School of Public Health, Mayo Clinic, CDC, and American Heart Association continue to refer to saturated fats as "bad fats" and suggest that people avoid them as much as possible, limit them to <7-8% of their total caloric intake, and replace them with "good," i.e. unsaturated fats, whenever possible. It’s also true that there are a few studies that suggest that increased saturated fat consumption is correlated (albeit weakly) with cardiovascular disease (CVD).

However, many other studies have found no increase in CVD associated with saturated fat consumption. And several recent review articles have concluded that the evidence for a connection between saturated fat consumption and CVD is inconsistent, insufficient or nonexistent.

There are basically three things that have convinced me that saturated fat isn’t independently responsible for heart disease or death:

1) groups of people who eat vastly more saturated fat than most Americans frequently have lower rates of CVD—or no apparent CVD at all

2) the study that first inspired many people to think saturated fat was a bad thing had a lot of obvious flaws and has been thoroughly discredited

3) there’s no good evidence that the proposed mechanism actually works—briefly: saturated fat supposedly causes heart disease by raising serum cholesterol levels because cholesterol is what clogs arteries and causes heart attacks and strokes, but serum cholesterol turns out to be a really poor predictor of CVD 

1) The French (and Polynesian and Melanesian and Masai and Fulani and Sri Lankan) Paradox 

photo by Arun Ganesh most confounding phenomenon for the theory that saturated fat consumption causes heart disease (sometimes called the lipid hypothesis or lipid-heart hypothesis or diet-heart hypothesis) is the virtual non-existence of CVD in multiple populations that eat way more saturated fat than most Americans. This has primarily been documented in the Pacific islands where coconuts, which are very high in saturated fat, are a staple food. For example, before the 1970s, the inhabitants of the island Tokelau got an estimated 55% of their calories from saturated fat, but heart disease was virtually unknown (according to Gary Taubes and Stephen Guyenet; notably, since their diet has shifted to include less saturated fat but more sugars and refined carbohydrates, many health indicators have worsened).

Many proponents of "natural" foods have already embraced the coconut as a source of "healthy" fat, but the confounding phenomenon isn’t unique to coconut-eaters. The Masai of Kenya and Tanzania and Fulani of Nigeria, whose traditional diet is composed primarily of cow’s milk, meat, and blood and estimated to be 33% saturated fat, also had virtually no incidence of CVD—or didn’t until they started eating refined carbohydrates. Red meat and dairy sources of saturated fat are about as classically "unhealthy" as you can get, and yet don’t necessarily cause heart attacks and death.

The differences can’t be attributed, at least entirely, to genetics, and not just because the phenomenon is consistent among so many far-flung populations. The Masai are known to be highly genetically diverse due to conflict/intermarriage, and all of the populations have experienced changes in their health as their diets have changed.

Nor are the changes likely due to higher activity levels, cyclical feast and famine, average lifespan, or the absence of other risk factors like cigarette use in these populations. A 1993 study of diet and health in Kitava, the Trobriand Islands, and Papua New Guinea found that Kitavans get an average of 17% of their calories from saturated fat (compared to about 10% for Americans or the <7% recommended by the AHA), tend to be about as active as moderately active Swedes, and have virtually no experience of food scarcity or shortage. Furthermore, 75% of the population smokes. Nonetheless, stroke and heart attack basically don’t exist even among the elderly there, and their fasting insulin levels are significantly lower than Swedes’ in every age group.  

Another potentially confounding fact that some low-carb proponents point to is the alleged decline in the average American’s consumption of saturated fat. It’s true that the consumption of some primary sources of saturated fat like butter and lard declined for most of the 20th Century; however, per capita red meat consumption grew until the 1970s, and according to the USDA, the proportion of saturated fat in the U.S. food supply has remained pretty steady for the last century. What is clear, at least, is that increasing rates of cardiovascular disease were not caused by increasing per capita consumption of saturated fats.

So why the hell have the institutions we trust to tell us how our diets affect our health spent the last half-century trying to convince us that saturated fat clogs your arteries, causes heart attacks, and might kill you?

2) It All Began With Some Falsified Data January 13, 1961 cover of Time magazine

I know, I sound like a climate change skeptic, but this is way more blatant than some sketchy e-mails.

The research that first convinced many Americans and public health institutions like the AHA that saturated fats cause heart disease was the Seven Countries Study led by Ancel Keys. Keys and his colleagues carried out surveys of men (only men because the rate of CVD in women was considered too low to merit attention) in eighteen rural regions in Italy, the Greek Islands, Yugoslavia, the Netherlands, Finland, Japan, and the U.S. The regions were supposedly chosen because they were areas with stable and widely-contrasting diets. The researchers asked the men about their eating habits, performed chemical analysis on meals prepared by randomly-selected families, and compared their findings to the rate of disease and death in the different populations, with special attention given to CVD and "risk factors" like serum cholesterol levels. They concluded that higher levels of dietary saturated fat cause elevated serum cholesterol and heart disease. The especially low rates of CVD in Crete, despite relatively high proportions of dietary fat, were a large part of the impetus behind the "Mediterranean Diet" and the idea that olive oil is nutritionally sacrosanct.

However, as many people later pointed out, it seems more than coincidental that Keys only included regions where both saturated fat consumption and heart disease were high or where they were both low. Apparently, data was available for a number of other countries (anywhere between 15 and 27, according to various critics), perhaps most notably France, which would have been just as easy to study as Finland and Japan. Keys seemingly ignored, or systematically excluded, all confounding data from his analysis.  

Even for the countries that were included, according to statistician Russell H. Smith:

The dietary assessment methodology was highly inconsistent across cohorts and thoroughly suspect. In addition, careful examination of the death rates and associations between diet and death rates reveal a massive set of inconsistencies and contradictions…. It is almost inconceivable that the Seven Countries study was performed with such scientific abandon. It is also dumbfounding how the NHLBI/AHA alliance ignored such sloppiness in their many "rave reviews" of the study…. In summary, the diet-CHD relationship reported for the Seven Countries study cannot be taken seriously by the objective and critical scientist.

The problem seems to be that Keys had already decided that saturated fat was the problem. According to the Time magazine cover article (click the picture to read it), Keys had treated a heart disease patient with large knobs of what turned out to be cholesterol under his skin and very high serum cholesterol. When he put that patient on a low-fat diet, his cholesterol went down. Since cholesterol is what clogs arteries and causes heart attacks, and foods high in saturated fat are often also high in cholesterol, Keys concluded that the high rates of heart disease in American men were caused by the saturated fat and cholesterol content of their diets. In the Seven Countries study, he set out to prove that theory (as opposed to testing it).

Fifty years of subsequent research have not found a consistent or reliable correlation between saturated fat consumption and heart disease or death. This entry got a little out of hand, so I split it into a couple of parts. More coming soon on dietary cholesterol and the dreaded LARD, neither of which, it turns out, are likely to kill you or even probably hurt you a little bit. Also, an epilogue of sorts on trans-fats, which might actually kill you.

Things that won’t kill you Vol. 2: Fruit juice

This may seem like a strange thing to argue about, because the popular consensus still seems to be that juice is healthy. Jamba Juice markets itself as "the category-defining leader in healthy blended beverages, juices, and good-for-you snacks." They even use Jamba as an adjective to mean the opposite of high fructose corn syrup and trans-fats (adding those things to juice ""just wouldn’t be Jamba"), which again, constructs the brand as healthy vs. the demon poisons that make people fat. Even if it’s foolish to go looking for truths in advertising, I don’t think Jamba Juice’s branding generally occurs to people as a massive irony or lie. Advocates of banning or restricting soda vending machines in schools often claim that the soda should be replaced with 100% fruit juice with no added sugars, and for many people, a glass of orange juice still represents "part of a nutritious breakfast" strongly with desirable nutrients like Vitamin C.

The Case Against Juice

But a number of health trends have begun cast suspicion on juice, especially the (impartial and incomplete) shift from primarily low-fat to primarily low-calorie and low-carb dieting in mainstream weight-loss culture, and the growing concern about the role sugars (especially fructose) play in personal and national obesity.

On the low-calorie front, people who believe that losing weight or maintaining a healthy weight is all about the basic algebra of calories-in vs. calories-out often end up axing all caloric beverages from their diets because they have a bad satiety-to-calorie ratio—I mean, obviously, right? Fruit juice is just fruit with some or all of the filling fiber removed. If the goal is maximum satiety on minimum calories, you’re better off eating whole fruit and drinking water or artificially sweetened beverages.

On the low-carb front, people who believe that what’s important is not how many calories you eat but what kind are also going to see juice (and sometimes most fruits and vegetables as well) as "unhealthy." It does seem to be true that diets high in carbohydrates drive up insulin levels, slowing metabolism and encouraging the body to store fat. And the overwhelming majority of the calories in most fruit juices are in the form of carbohydrates. Some green vegetable juices have protein content approaching 50% of the carbohydrate content, but that just makes it 75% bad rather than 100% bad, at least as true carbophobes are concerned.

And finally, there are some non-carbophobes who might avoid juice because they’re wary of sugar qua sugar, rather than sugar qua carbohydrate. The carbohydrates in fruit juice primarily take the form of fructose—wikipedia has a handy chart of the kinds of sugars in common plant foods. It doesn’t seem like there’s a true consensus yet about whether or not fructose is especially bad—despite recent studies linking fructose to obesity, even within the medical community, some people still advocate fructose as a "low glycemic" sugar that’s better for diabetics. It basically all comes down to whether you think the fact that fructose is digested in the liver and doesn’t trigger insulin production is a good thing or a bad thing. To link it to other sugar purveyors: pro-agave nectar people should also think that fruit juice is healthy and people who think hfcs is bad because they think it’s "high fructose" compared to other sugars are, well, a) wrong, but b) should also be advocating hfcs-sweetened sodas over fruit juices, which are even richer in sugar.

Personally, I think the evidence that fructose in large amounts causes equivalent blood sugar spikes to other sugar, increased "bad" cholesterol and triglycerides and signs of insulin resistance compared to glucose, and can cause non-alcoholic fatty liver disease consumed in vast quantities suggests that it is certainly no better and possibly much worse for human health than glucose or sucrose. But "worse for human health" is relative, not absolute, and depends a lot on amount, kind, and context. 

What is health?

I’m generally convinced by the argument made by people like Gary Taubes that a diet composed of almost-exclusively proteins and fats might better represent the pre-agricultural, hunter-gatherer diet (as would cyclical feast and famine) and prevent carbohydrate-induced insulin resistance and fat storage. Jared Diamond makes some of the same points in Guns, Germs, and Steel. But the benefits of agriculture ultimately outweighed the costs—both for the species as a whole and measured by individual health metrics. In the immediate aftermath of the transition to agriculture, lifespans and average height decreased, but after a few thousand years, people depending on rice, corn, and wheat began to get healthier again

Does that mean carbohydrates are a healthier basis for a diet than proteins and fats? No. But it does mean that people can (and do) live very long lives uninterrupted by diet-based disease during which they are strong and energetic enough to physically do anything they want to do while eating a diet consisting substantially of carbohydrates. And I think that’s not a bad working definition of "healthy."

It seems to me that the debate comes down to a difference between ideas about nutritional "health" based on what might be theoretically optimal (for a very limited set of criteria), ideas about health based on potential pathology, and ideas about health based on actual health outcomes

Fear of juice is based on the first two—the idea that either people should eat as few carbohydrates as possible in pursuit of some sort of optimal diet, or that the fructose in juice will cause fatness (an aesthetic problem, not a medical problem) or disease and eventually death. Based on actual health outcomes, I think it would be almost impossible to make a case for the claim that drinking fruit juice—occasionally or regularly—is categorically unhealthy or the direct cause of disease or death.

In fact, things like fruit juice and hamburgers and Doritos, which can each be constructed as "unhealthy" are hard to entirely rule out of a "healthy diet." Even proponents of a soda tax generally agree that the only reason soda is a reasonable target is because it has no identified nutrients (what would happen if they fortified them, I wonder?).

The nutrient-density of juice is the primary reason for the long tradition of juice being regarded as a health food. If your concern is about essential vitamins and minerals (like many older models of nutrition, which people like Marion Nestle stand by) or consuming carbohydrates for fuel, which many physically active people still do, it’s hard to argue with the healthfulness of juice. I agree with Michael Pollan’s claim that popular beliefs about health often fall prey to "nutritionism," or the attempt to reduce food and nutrition to scientifically-identified nutrients and vitamins. At the same time, I don’t think you have to be brainwashed by the continued prevalence of nutritionism to believe there’s good evidence that many of the nutrients that scientists have identified are actually valuable or promote health and well-being (even if they’re not the only valuable aspects of food).

All juices are not created equal

The person who requested this entry was concerned specifically about fresh juices being portrayed as unhealthy, because they seem to have been smeared by concerns about packaged juices being just other source of dietary sugar.

While not all fruits lend themselves as readily to the production of refined sugars as sugar beets, some like apples, pears, and grapes can be turned into a nutrient-poor sweetener without most of the fruits’ color, flavor, or minerals and many fruit juices marketed as 100% natural fruit juices, like Juicy Juice, are sweetened with fruit juice that’s basically been turned into a sugar syrup. The nutritional distinction between those drinks and hfcs-sweetened soda is probably negligible regardless of whether your primary concern is calories, carbohydrates, sugar, or vitamins.

But the reason packaged juices often combined with fruit-based sugar is that many fruit juices aren’t actually that sweet on their own, or their sweetness is offset by the intensity of the flavor, as anyone who’s ever tried 100% cranberry or concord grape or cherry or blueberry juice knows. The fresh juices you can get at juice bars or make at home are calorie-dense, but they’re also extraordinarily nutrient dense and not likely to be consumed in quantity or alongside meals. They’re more often enjoyed on their own, like a snack, particularly after a workout—basically just like fruit. When you leave in some of the pulp, it becomes even less nutritionally distinct from fruit, and when you include vegetable juices or the juices of things like wheatgrass and ginger which are difficult or unpleasant to eat raw, you may be enjoying something that could, by some criteria, be healthier than a piece of fresh fruit.

Some juices even have pretty well-established medicinal uses. Cranberry juice, for example, can help prevent and cure urinary tract infections (that study notes the existence of "diet" cranberry juice, which I’d never heard of, but now that I have I wonder why there aren’t more "diet" juices sweetened with artificial sweeteners rather than pear or grape juice-sugar. Not that those would necessarily be "healthy" by everyone’s standards, especially given the links between saccharin and cancer and suspicions about the healthfulness of aspartame and sucralose…)

Ultimately, while I don’t think even the occasional hfcs-sweetened Capri Sun is incompatible with a "healthy" diet and life, I think it’s unreasonable to conflate fresh juice without added sweeteners with juices sweetened with refined juice-sugar. I guess people trying to eat an "optimal" diet a la Gary Taubes should avoid all juices, fresh or no, but I don’t envy them their carbohydrate-less life, nor am I convinced that the total deprivation of many foods that have aesthetic, gustatory, social and/or cultural value is necessarily "healthy" or "optimal" either. For the vast majority of people who think fruit and vegetables are part of a healthy diet, fruit juice and especially fresh fruit juice should also pass muster as a "healthy" choice especially when consumed in moderation, which I suspect fresh fruit juices usually are.