Insulin does not control weight loss.

Kevin Hall performs both theoretical and experimental studies in human metabolism at the NIDDK of NIH. This week, he and colleagues published an article in Cell Metabolism entitled, "Calorie for Calorie, Dietary Fat Restriction Results in More Body Fat Loss than Carbohydrate Restriction in People with Obesity."

I think it's going to piss some people off.

For this reason, I decided to get my facts straight before writing about the study. Dr. Hall gave me some time on the phone earlier today. But before I get to his explanations, let me set up why, in diet research circles, this study is likely to have a large impact.

The central tenet of the carbohydrate hypothesis of obesity causation is that insulin mediates the growth of fat tissue. It is a smarter way of looking at obesity than the "eat less, exercise more" philosophy espoused by your doctor, because at least the carbohydrate-insulin explanation tries to answer why fat tissue grows. If you think about it, getting heavier isn't the problem, it's getting more fat tissue that bothers us. So, lets begin with this: the carbohydrate hypothesis may be right or wrong, but it is at least going to be right or wrong about an important question. It is proposing a solution to a regulation problem, which is a good start.

Photo credit: my midnight snack while writing this.

The hypothesis asserts that high carbohydrates in the diet cause a chronically high insulin state. Since insulin's main role as an energy regulator is to take sugar from the blood and cause it to be stored, it is cast as the villain in the weight gain story. This one hormone simultaneously: increases fat production, reduces fat burning, increases glucose storage as glycogen and stops production of new glucose from protein. Thus, to lose weight, we must reverse the process: lower carbs, to lower insulin, to reduce storage and allow fat burning. It's a pretty logical proposal. But it only makes sense if insulin is far and away the most important regulator of our metabolism and fatty tissues. The study by Hall et. al. suggests that it isn't.

This was not a diet intervention study, but a mechanistic test of whether altering macronutrients might differentially affect metabolism. The researchers took 19 volunteers and studied their metabolism in the highly controlled circumstances of a "metabolic ward." Staying in the ward is like living in a Tokyo apartment, except that the apartment has the ability to record your motion and heat response, while you are fed fixed meals and attendants run tests on you periodically. Each participant did both diets, which were 30% reductions of total calories (individualized to each subject's metabolic expenditure-measured on normal diet during a run in phase) either through carb reduction or fat reduction. A host of variables were checked before and after the six day diet periods including Dexa scan of body fat, insulin, glucose, several of the hormones that mediate hunger in the gut (such as Ghrelin and GIP), but most importantly, they computed the oxidation of nutrients.

This is something that can't be done in the usual diet study which pits one diet concept against another and decides success based on participants' weight at six months. This study was run in these highly controlled conditions in order to test some specific contentions of the carbohydrate hypothesis: 

Does low carb dieting alter nutrient oxidation?
Does that response lead to greater fat loss?
Can weight loss occur in circumstances of normal insulin amount?

What the authors concluded, is that you can lose weight on a low fat diet, without changing insulin. In fact, they found, by looking specifically at fat oxidation, that more fat is burned by restricting fat intake than restricting carb intake. While on low fat diets, the volunteers had a net fat loss of 89 grams per day, vs. 53 g/day for low carb. 

These rates of fat oxidation were calculated by a combination of validated methods using indirect calorimetry (oxygen in...CO2 out is used to estimate metabolic rate) combined with measures of 24 hour nitrogen losses in urine.

During the low fat phase, participants had no change from baseline in the metabolism of any macronutrient: rate of fat oxidation, carb oxidation and protein oxidation did not significantly adapt to cutting 800 calories from fat. As a result, during the low fat weeks, the subjects lost very close to the theoretical maximum of 800 calories worth of fat per day (765 +/-37). 

During the low carb phase, on the other hand, there were significant metabolic adaptations. Subjects were found to have ramped up fat oxidation by 403 cal/day and decreased carb oxidation by 520 grams per day. This would seem like validation of the low carb hypothesis, right? Everybody switched to "fat burning mode" when the carbs were cut. But the switch wasn't efficient. The slow down in carb burning overbalanced the increase in fat burning, such that overall rate of oxidation of nutrients was slower while on a low carb diet than baseline. They still lost weight, but not as much as theoretically possible from cutting 800 calories, because of the inefficiency.

The low carb diet produced less fat loss than the low fat diet. But that is probably not the most important point in the paper, since the difference was pretty small. Because this was a study of how we lose weight, specifically looking at low carb diet recommendations, a much more interesting point is that insulin was entirely unchanged in the low fat group.

I asked Kevin Hall whether this meant that he'd disproved the carbohydrate hypothesis:

"The reduced carb diet does everything that the proponents of low carb diets claim: it reduces insulin secretion, it increases fat oxidation. AND it leads to fat loss. But this study was designed to look at another diet in comparison: If you don't decrease insulin because you don't decrease carbs, can you also get fat loss?

"Some folks, Gary Taubes in particular, have made the very strong claim that if you don't reduce insulin by reducing carbs, in particular refined carbs, then you can't lose fat from the body. And this study clearly demonstrates that that is not true. Insulin secretion went down in the low carb group by 22% and stayed the same in the low fat group."

I suggested to Dr. Hall that if we stop thinking like a low carb person and just start from scratch, it shouldn't seem weird that reducing fat caused a loss of fat from the body. It needs to make up the deficit. I suggested that the study results were unsurprising (don't say that to a researcher that just finished a study that took like three years. He didn't seem to mind, but still).

He answered: "A surprising feature of the study was the lack of change in fat oxidation in the low fat group. If you reduce fat intake, but otherwise leave protein and carbohydrate unchanged, why does the body go on burning fat like it did before? Why didn't it switch to a higher mixture of oxidation of carbs or protein? Why would the body say, 'Ya know what, I'm going to go on burning my fat just the same, even though I've cut 800 calories from the diet and I'm eating 8% fat in the current diet?' I find that interesting."

Asked to explain what mathematical modeling added to this research project:

"...before we brought in a single patient, we made some predictions that were in the clinical protocol. For example, the primary outcome of the study, that there would be no shift in the metabolic fuel mixture with the reduced fat diet, but that there would be a robust shift toward fat oxidation with the reduced carb diet-that was all in the clinical protocol, based on the model. And those both turned out to be correct."

I would think that feels pretty satisfying?

"Yeah, it is, although at the same time, one of the reasons that we did the study, was because, I kind of thought that the model might be wrong. I kinda thought, maybe hoped, secretly, that when we reduced fat in the diet, that there might be a shift in metabolic fuel utilization and that we would somehow be able to find some signal in the circulation that was orchestrating the process of changing fuel selection in reduced fat diets. Some effect analogous to how insulin changes dramatically when you change carbohydrates. We found no such effect. So one possible conclusion is that there is no such factor that responds to dietary fat. There are things that change in response to dietary fat (such as CCK) but they don't affect how many calories you're burning or where the fuel comes from."

So, is a calorie a calorie, based on this?

"Well, one of the main findings in the paper is that virtually nothing changes in the metabolism when you cut 800 calories of fat out of the diet. But all we can say is that we did these two different diets and we saw very small differences in fat loss. One of the diets gave a small reduction in sleeping metabolic rate and total energy expenditure, while the other didn't, so in that case, a calorie is not exactly equal to a calorie. But - it's pretty close. And some of the model simulations that we're showing here would suggest that these two diets are the two that would show the most difference. Specifically, if you were to keep the calories and protein constant, but reduce the carbs even more and add back fat to maintain calories, that you would start to see results more similar to the reduced fat diet we used. The point there is that this is probably the maximum effect you can see in terms of trying to find a difference in the rates of fat loss in the body.

"The 'calorie is a calorie' prediction is that there shouldn't be any difference on equal calorie diets when you look at rate of fat loss, or energy expenditure and our model suggests that the small differences that we did see are probably the maximum you would expect."

In response to my question about the backlash that is likely to come from the low carb advocates out there, Dr. Hall laughs and adds:

"I do want to be clear that this is only one of many pieces to the whole carbohydrate-insulin hypothesis. It doesn't take down the concept of a low carb diet: they make you less hungry and feel more full, people may stick to them better, there all sorts of things that may occur on a low carb diet. This research is only taking down that one proposal, which is that the only way you can lose body fat is by reducing carbs, through changing insulin. I think it's pretty darn clear in the study that we have two diets of equal calories in the same people: one of them reduces carbs and reduces insulin and causes less fat loss than the other diet that only cut fat and doesn't cause any change in insulin."

Kevin Hall, et. al. Calorie for Calorie, Dietary Fat Restriction Results in More Body Fat Loss than Carbohydrate Restriction in People with Obesity. Cell Metabolism. Published online August 13, 2015.