This article was published in Scientific American’s former blog network and reflects the views of the author, not necessarily those of Scientific American
Weight-loss research has generated headlines lately, leading me to wonder what my pal Gary Taubes is up to. Over the past dozen years, Taubes has transformed himself from a mere journalist into a major player in dietary science, who has helped raise millions for research. In Good Calories, Bad Calories (2007), Why We Get Fat (2010) and countless articles, he argues that carbohydrates, and especially refined grains and sugars, make us fat, more than fat itself or sheer quantity of food.
I voiced doubts about Taubes' hypothesis in a 2011 post. (See also my chat with him on Bloggingheads.tv.) In 2012, I reported that Taubes had co-created a nonprofit organization, Nutrition Science Initiative, or NuSI, to perform rigorous tests of different diets, including the one he recommends. I emailed Taubes to ask him a few questions.
Horgan: What's the status of research funded by NuSI?
On supporting science journalism
If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
Taubes: In our first fund-raising effort, we raised approximately $40 million to fund three studies that are now up and running. One is a pilot – described in a recent article in Wired -- the primary purpose of which is to gather the preliminary data necessary and to test the methodology for a full-scale trial that will constitute a rigorous test of the competing hypotheses of obesity and weight regulation, which you and I have discussed in the past. The researchers working on that pilot study are now in the process of going through the data accumulated and doing the necessary analyses. They're also beginning the process of designing the full-scale trial. Of the two other trials, the Stanford study has now run two thirds of its subjects, and the Boston Children's Hospital study has just been launched. We're also in the process of meeting with researchers to design the next set of trials that we'll fund, and we're meeting with philanthropists and philanthropic foundations to raise the necessary money. Our plan is ambitious, and the next year should give us a reasonable idea of whether we can accrue the support necessary to make it all happen.
I just re-read your 2011 blog post, and it prompted me to make an additional point before answering your other questions. In your post, you missed the point I was endeavoring to make in both my books on nutrition science, which is also crucial to understanding the goal and purpose of the Nutrition Science Initiative. The point is not about whether an Atkins diet is a healthy diet or whether it's the best dietary intervention for obese, overweight and diabetic individuals -- although that is one question that needs to be answered. It's about establishing the dietary trigger of the obesity and diabetes epidemics. What makes us fat? What makes us diabetic? And what causes the chronic diseases with which obesity and diabetes associate.
We know something has changed in our diets and/or lifestyle to trigger these dramatic increases in the prevalence of obesity and diabetes. The conventional thinking is that obesity is a disorder of energy balance and we've gotten steadily fatter over the past fifty years simply because we eat more than ever and exercise even less. So this is the cause of obesity and the obesity epidemic, and then something about being obese or at least getting fatter is what is driving the diabetes epidemics. My books argue that a more likely suspect, an alternative hypothesis that should perhaps be the null hypothesis, is that these dramatic increases in obesity and diabetes prevalence have been caused by a change in diet quality -- in particular, the increased consumption of easily digestible carbohydrates and sugars. The sugars, in particular -- sucrose and high fructose corn syrup -- are key here, as I argued in this 2011 New York Times Magazine article. It's a simple explanation for why Asian populations eating high-carbohydrate diets have remained relatively lean and healthy, at least until recently. They've had exceedingly little sugar in their diets.
By making the issue about Atkins and Ornish, vegetarian vs. omnivorous diets, etc., journalists and public health authorities and the diet book authors themselves direct attention away from the issue of utmost importance, which is why we get fat and why we get diabetic. What's driving these epidemics? Is it because of a failure of will power -- we eat too much and exercise too little -- or the carbohydrate (or fat or something else entirely) content of our diets? By making it about Atkins as you did in 2011, you missed the larger point. The Atkins diet plays it into it because it's one piece of anomalous evidence (and there are many) that suggests that the current belief system, or at least the belief system circa 2002 and up until very recently -- is wrong. The studies we're funding with the non-profit are not trying to establish whether an Atkins diet is healthier or more effective than an Ornish diet or a Zone diet or a Paleo diet or an AHA diet, etc., but are using these extreme dietary interventions as tools to probe the effect (if any) of the macronutrient content of the diet on fat accumulation and metabolic signaling. If easily-digestible carbohydrates and sugars have an effect on fat accumulation independent of their caloric content, as I suggest is likely in my books, understanding this is absolutely critical to understanding the genesis of the obesity and diabetes epidemics, and our public health and medical interventions have to take this into account. Up until now, they've assumed that a calorie is a calorie is a calorie, and that the only dietary factor that influences fat accumulation is the balance of calories in and out. Whether this is really true, and I've argued that it's bizarrely naive, it's one of the primary issues that the research we're funding is attempting to settle.
Horgan: An NIH-funded study reported in the Annals of Internal Medicine last month favored low-carb over low-fat diets. Comment?
Taubes: Regarding this NIH-funded trial at Tulane, it has serious methodological problems, but was nonetheless consistent with perhaps two dozen other trials that have been done in the last 15 years, perhaps more. Typically these trials, though, compare subjects who are advised to adhere to calorie-restricted low-fat diets to those who are advised to adhere to eat-as-much-as-you-want (technically ad libitum) low-carb diets. One problem with all such "free-living" studies, and one that the NuSI-funded research will try to ameliorate, is that they get poor compliance to the diets. After three to six months typically, it's clear few of the subjects are adhering to the diets. Hence, you can think of these tests as comparing what happens when you advise subjects to eat a low-carb diet and compare that to what happens when you advise subjects to eat a low-fat diet. What you want to know -- okay, what I want to know -- is what happens when people really do eat those diets and keep it up for years at a time. Not what happens when they're advised to do it, and most of them don't. Another problem with this study and most others like it is that both the low-fat and low-carb diets improve the quality of the carbohydrates consumed. The subjects in both arms restrict sugars and highly refined grains -- they exchange sugary beverages for sugar free, they drink less beer or drink lite beer (reduced in carbs and calories); they avoid desserts, which they might think they do for the fat content in the low-fat arm, but they're avoiding flour and sugar as well. So you can think of these studies as comparing a low-fat carbohydrate-restricted calorie-restricted diet, to a high-fat, carbohydrate restricted diet. The researchers don't think of it like this, though, and instead simplistically think they're comparing low-fat diets to low-carb diets.
Finally, these studies typically calorie-restrict the low-fat arm, but not the low-carb. This actually wasn't the case in this Tulane trial, as I'll get to shortly, but it usually is. The catch is that when the investigators report the results, they typically ignore the fact that one diet was prescribed calorie-restricted (the low-fat diet) and one wasn't (the low-carb). This study published in 2010 is a good example. It was funded by the NIH and was the largest such trial ever done up until then. The investigators mention in the intervention section that the low-fat diet consisted of a "limited energy intake" and the low-carb diet did not, but that's effectively the only time they mention it. So what they're doing is comparing a low-fat diet that is calorie-restricted and, as I said, carbohydrate-restricted, to a low-carb diet that is neither fat or calorie-restricted, but you'd never know it from reading the rest of the paper. And because the low-carb diet does better, you have to ask yourself whether the calorie restriction was necessary or the fat restriction. The researchers don't address this question, because they're oblivious to or ignoring these details. They certainly don't consider them relevant.
In this Tulane trial, the dietary interventions were both ad libitum, meaning they did not explicitly advise either group to restrict calories or portion sizes. As a result, they saw slightly larger differences than previous trials in weight loss between the two groups. Otherwise, this trial confirmed what had been seen consistently in the other trials, despite all these methodological issues (and as we discussed the last time you blogged about this). So it's nice that the trials are still seeing consistently better risk factors (i.e. surrogate endpoints for heart disease and diabetes). And this continues to suggest that if physicians and public health authorities are going to be giving evidence-based advice to their patients and the public on what constitutes a healthy diet -- whether for weight loss or weight maintenance -- they should, perhaps, be saying the opposite of what we've been told for the past forty years -- i.e., eat more fat and fewer carbohydrates, not vice verse. What's also clear is that far better trials have to be done, with a far more rigorous understanding of all the variables being affected in these dietary interventions and with the subjects doing a far better job of complying to the dietary advice. Again, what I want to know is not what happens if someone tells me to eat dietary pattern A rather than dietary pattern B and I don't faithfully follow that advice. I want to know what happens if I do. Another way to think about this is that if I'm thinking of moving to Crete because I've heard a Mediterranean diet is the healthiest way to eat and I want to live as long as I can, I want my decision to be informed by unambiguous information about how the Cretan version of a Mediterranean diet will effect my life, not by the result of studies that compared people who occasionally ate at a Mediterranean restaurant to those who didn't.
Horgan: A report in Journal of the American Medical Association recently found little difference between low-fat and low-carb diets and seemed to support the idea that any reduction in food will help reduce weight. Comment?
Taubes: This JAMA article, not surprisingly, also has serious methodological problems because it's more or less a comparison of dietary studies the likes of which I discuss above that are flawed both in how they were carried out and then in their interpretation. So first of all, this JAMA article is not actually comparing low-fat to low-carb diets so much as name diets to each other -- Atkins to Ornish to the Zone, etc. As such it's not addressing what I want to know and I'd argue we all want to know, which is what happens when we restrict carbohydrates -- or specifically, easily digestible carbohydrates and sugars -- and what happens if we don't and, say, only restrict dietary fats.
Second, the analysis only focuses on the effect of these diets (or being advised to consume these diets) on weight loss, not on heart disease or diabetes risk. Back when I wrote "What If It's All Been a Big Fat Lie" in the New York Times Magazine in 2002, the conventional thinking was that an Atkins-like, high fat, high saturated fat diet, was deadly because of the saturated fat content. (When you did your blog post in 2011, this also seemed to be the subtext.) Fear of saturated fat and high fat diets was the genesis of the paradigm that a low fat diet is inherently healthy and the reason why Atkins was considered a dangerous quack by the medical establishment ever since he published his book in the early 1970s. It's the reason why my 2002 article was so wildly controversial (to put it kindly).
So here it is a dozen years later, and the saturated-fat-will-kill-us argument isn't even raised in this JAMA article. Its only concern is whether one popular diet is more or less effective than other popular diets for weight loss. By the same token, when Annals published a meta-analysis in March by British researchers concluding that there was insufficient evidence to recommend restricting the saturated fat content of the diet, the criticism from the old guard -- the types of researchers and authorities who condemned my 2002 NYT Magazine article -- was along the lines of, "Sure, everyone knows that replacing saturated fat with refined grains and sugars is bad," and the debate was whether it might benefit us to replace saturated fat with polyunsaturated fat. So one of the exceedingly contentious points I had made back in 2002 had now become mainstream science and the bar had been raised to a debate about the SFA/PUFA trade-off.
And then, as I said, the JAMA meta-analysis began with a host of studies with the kinds of methodological problems I discussed above, and then compared them as though these problems didn't exist and didn't matter. I think it would have benefitted from being done by researchers who actually had clinical and research experience with obesity, which did not appear to be the case. But as I said, many to most obesity researchers tend not to pay attention to these issues. One who does, Dariush Mozaffarian, dean of the School of Nutrition Science at Tufts, did raise these issues to the Boston Globe in a discussion of this article: "In the JAMA meta-analysis looking at commercial diet plans, participants reduced how much they ate and also altered their diet composition, so it’s impossible to know what actually led to the weight loss." I think Mozaffarian hit the nail squarely here. Again, what's needed is for researchers to pay more attention to the variables they are and are not changing in these trials so better trials get done.