Hall et al.

August 20, 2015

A bombshell study was just published in the journal Cell Metabolism by Kevin Hall and colleagues and many journalistic publications are—as they are wont to do—misconstruing the results for the sake of flashy headlines and pageviews. This post isn’t meant to be a full-blown literature review, but I do want to bring the study to everyone’s attention.

You can read the full text of the study here if you want (you don’t,) but do realize that this is probably one of the most in-depth, expensive, and meticulously-crafted studies on human energy balance ever done. Nineteen obese adults were locked in a metabolic ward for upwards of two weeks, allowing for precise caloric measurements (among other things) to be taken. The participants were put on a baseline diet for a while, then divided into low-fat and low-carb groups for six days; we return them to baseline, crossover (low-carbers become low-fat, vice-versa), and repeat.

Both diets were what we call isocaloric: all 19 dieters were eating exactly 1918 calories each day, which was exactly a 30% reduction from baseline. The only variables that changed were carbohydrate and fat intake; protein was kept constant. With this setup, Hall could measure changes in body composition knowing precisely what brought them about.

Many journalists are skewing this study as a nail in the coffin for low-carb dieters, citing the authors’ conclusions:

“Calorie for calorie, dietary fat restriction results in more body fat loss than carbohydrate restriction in people with obesity.”

This probably shouldn’t have been the title of the study, because it wasn’t the hypothesis Hall et al. were testing, which instead was:

“Any diet that succeeds does so because the dieter restricts fattening carbohydrates …Those who lose fat on a diet do so because of what they are not eating—the fattening carbohydrates.”

This is a quote out of Gary Taubes’ book Good Calories, Bad Calories, which is worth a read despite the outcome of this study (spoiler alert?) The study, in essence, was designed to test the assertion that for fat loss to occur, it is necessary for carbohydrates to be reduced. Ultimately, this was not the case. Instead, the low-fat group lost a solid pound more of body fat than the low-carb group.

A few obvious criticisms that have already been made:

  1. The low-fat diet was way too low-fat. Sure: only 17 grams of fat per day is fairly unsustainable for most normal human beings. However, the study wasn’t designed to test real-life diets, and to keep calories and protein equal among both populations, this was the necessary step. Which leads to complaint #2…
  2. The low-carb diet wasn’t really low-carb. Also true: the low-carb group was eating 140 grams of CHO/day (few of which were from sugars,) which was indeed a significant reduction from the 350 grams/day of the baseline diet, but still falls well above the cutoff for ketosis and most low-carb diets (which often restricts CHO-intake to below 100g or even 50g/day.) So it shouldn’t be called low-carb, it should instead be called reduced-carb, or moderate-carb. Again, however, this was necessary to keep both diets isocaloric; any further reduction of carbohydrate would end up putting the low-fat group into the negatives!
  3. The study wasn’t long enough. Yeah, six days with each diet is a relatively short time, but it was short enough to measure some changes. More importantly, keeping people locked up in a metabolic ward for two weeks is already difficult enough, not to mention expensive. Hall et al. did what they could here. I’d love to see the same study done over months, but that would require insane monetary commitments and superhuman test subjects. It probably won’t happen.

So again, this study ought not to be interpreted as “low-carb doesn’t work,” nor as “low-fat is superior.” The authors did, however, succeed in proving their original hypothesis: that carbohydrate-reduction is not required for fat-loss to be achieved.

In my rant about calories last month, I wrote that insulin via carbohydrate intake was the only way for fat to get into our adipocytes:

So, how does one reduce fat accumulation if calories aren’t the answer? It turns out that there is but one way for fat to be deposited in your body: the hormone insulin, a hormone released by the pancreas when (and only when) carbohydrates enter the bloodstream (unless, of course, you are diabetic, but that’s another conversation.)

I was, admittedly, being over-simplistic with this statement, and was called out on it. All foods are insulinogenic, otherwise one could eat a zero-carb diet ad libitum and never gain fat at all. The fact of the matter remains that carbohydrates provide a fast-track to insulin overload and the resistance to insulin that accompanies most metabolic diseases. But I was wrong in asserting that this was the only possible way, as this study has demonstrated. (Remembering throughout, of course, that this study didn’t measure real-life diets.)

So why did the low-fat group “win”? My theory, and many others’, is glycogen, the sugars stored in our muscles and livers for use when carb-intake is low (think “carbo-loading” before the big 5K, or some other silliness.) When the reduced-carb group started on their diet, their bodies likely noticed the change and started burning through glycogen to remain “carb-happy.” The low-fat group, on the other hand, continued burning at “high-fat mode” for a while as fat oxidation was downregulated over a few days. (I’m being simplistic here as well for the sake of space. Stephan Guyenet, a man much smarter than myself, goes into more detail on this subject here.)

If you’ve made it this far into the post, you can relax, it’s almost over. Realize again that I am fairly well-versed in scientific reading and writing, and there are others out there smarter than myself with their own interpretations. The case in point is this: it pays to investigate new studies and draw conclusions based on the science, not a journalist’s hasty interpretation of it. The study has not “finally put the low-carb craze to rest” or whatever other crap you’re reading out there.


Update: Some supplemental reading, if you happen to be crazy and enjoy delving into other analyses of this study. My favorite critique is this one from Metabolism and Medicine. It points out some valuable flaws in the study—for instance, accidentally feeding participants the wrong diets on one day, and completely leaving out one woman’s data—and does a much better job of articulating the same points I’ve tried to. This blog post outlines an alternative baseline diet that would’ve put the low-carb arm under 100g/day, more closely resembling a keto-diet and perhaps producing different results.