Catching up on lost time – the Ancestral Health Symposium, food reward, palatability, insulin signaling and carbohydrates… Part II(b)

When last we left off, the subject was the food reward/palatability hypothesis of obesity and why I find it so unconvincing and problematic.

This post is going to address that issue by first discussing three new papers published on the subject of sugar.  Then I’m going to ask a lot of questions, hoping at least some of the answers will be obvious. All three papers are from Peter Havel and his collaborators at U.C. Davis and Vanderbilt; all three looked at the metabolic effect of the fructose-component of sugar, while one of them one also looked at the effect of sugar itself in the form of high fructose corn syrup (HFCS).

Paper number one describes an experiment in which rhesus monkeys were fed their usual monkey chow diet supplemented by a daily 300-calorie ration of fructose-sweetened water. After a year, every last one of the 29 monkeys had developed “insulin resistance and many features of metabolic syndrome, including central obesity, dyslipidemia and inflammation.” Four of the monkeys progressed to type 2 diabetes.

Worth noting is that the monkeys apparently drank all the fructose-sweetened drinks they were given, but they adjusted for the 300 additional calories by cutting back significantly on the chow. All in all, they averaged only 26 calories per day more with the fructose than they did without it. This suggests that the negative sequelae observed over the course of the year were indeed caused by the fructose itself and not an increased intake of total calories (unless we’re willing to accept that increasing calories by just a couple a dozen a day is sufficient to do some very bad things to monkeys and do them relatively quickly.)

The second and third papers, both published in October, were randomized controlled trials in humans — rather than primates or rodents — which is  always a nice attraction in nutrition research since there’s no guarantee any animal model really reflects the physiological situation in, well, us.

The second paper reported that overweight and obese older adults (40 to 72 years-old) getting a quarter of their calories from fructose-sweetened beverages used less fat for fuel and actually expended less energy than did the same subjects when they were getting the equivalent calories from glucose drinks. So the calories were the same; the metabolic effects were different. For the fructose, the effects were what we’d expect (okay, what I’d expect) if the fructose beverages were causing or exacerbating insulin resistance, an observation that Havel et al had published earlier. The more insulin resistant these people became, the less fat they used for fuel—hence, we can assume, the more fat they stored—and the lower their basal metabolic rate. (The rhesus monkeys, too, had decreased their energy expenditure in response to drinking fructose.)

The third study was a trial in younger subjects (18 to 40-year-olds), some lean, some not. These subjects were given three beverages to drink every day, constituting again about a quarter of their day’s ration of calories. The beverages were sweetened either with fructose, glucose or high fructose corn syrup. For the HFCS group, this 25 percent of calories was compatible with what the U.S. Dietary Guidelines considers a safe upper limit for sugar consumption. Still, after only 12 days—less than two weeks—subjects in both the HFCS and fructose groups, but not the glucose group, saw a significant increase in heart disease risk. Triglycerides went up; LDL cholesterol went up, and ApoB concentrations, a measure of the number of LDL particles, increased.

What can we take away from these studies? Well, these three papers certainly support the contention that the sugars consumed in western diets have very specific deleterious metabolic effects, and that maybe these sugars are the, or at least proximate cause of insulin resistance and metabolic syndrome, and so, we can assume, obesity and type 2 diabetes and perhaps all the other chronic diseases that associate with these two conditions (cancer anyone?). This was the thesis of my April New York Times Magazine cover article “Is Sugar Toxic.” Here’s how Havel et al phrased this notion in the monkey paper:

The incidence and prevalence of insulin resistance and metabolic syndrome in the United States has increased dramatically over the past several years, and we and others have proposed that this increase may in part be attributable to increased consumption of fructose derived from dietary sugars, principally sucrose and high-fructose corn syrup (HFCS).

And, yes, the three papers certainly add fuel to this possible fire.

So on one hand, we can explain the metabolic abnormalities observed in these three studies—the insulin resistance, metabolic syndrome, raised triglycerides and increases in ApoB and LDL cholesterol—as all caused by the metabolism of the fructose in the liver in concert perhaps with the insulin-stimulating effects of glucose.  This is not a particularly controversial position. The sugar industry is not a fan, but research has consistently demonstrated these effects and has since the 1960s, and the biochemistry behind them is well understood.

So now we have three clear propositions and we’ll follow them up with the obvious question:

Proposition #1: The metabolism of fructose in large quantities by the liver results in a system-wide hormonal disruption — insulin resistance, in particular.

Proposition #2: Insulin resistance is intimately associated with obesity and is the underlying defect in type 2 diabetes.

Proposition #3: Insulin is the fundamental hormonal regulator of fat accumulation in fat cells.

Obvious question: Considering these three propositions, isn’t it likely that insulin resistance and the chronically elevated levels of insulin that go with it are causal factors in the obesity seen in the monkeys and, well, in humans who consume significant sugar, too?


This line of reasoning is one of the principle reasons that I’ve argued in my books and this blog that the carb/insulin hypothesis should be considered the null hypothesis of obesity, the one we believe until remarkable, unambiguous evidence comes along to refute it.

We have a disorder, obesity, that associates with a host of chronic diseases of western diets, all of which are associated as well with metabolic syndrome – i.e.,  insulin resistance — and type 2 diabetes. So if insulin plays a major role in regulating fat accumulation in individual fat cells, which is not a point of controversy, and if dysregulation of insulin signaling is a major component if not the causal factor in the diseases that associate with obesity, also not particularly controversial, it seems like a very good assumption that obesity, too, is caused by a the same dysregulation in insulin signaling.

Now the salient questions are these: How remarkable and unambiguous does the evidence have to be before we toss all this out in favor of the idea that we get fat because rewarding foods stimulate neurochemical changes in brain centers that raise our adiposity set point and ultimately make us eat more than we expend? And what do we gain by doing it?

I will argue that tossing out the role of insulin on fat accumulation to embrace the idea that rewarding foods stimulate brain centers seems inherently foolish, because the evidence is by no means unambiguous and in the process we disassociate weight gain from all these other insulin-mediated and very closely related disorders.

Let’s first try to address this cost-benefit ratio — what we gain from the food reward/palatability hypothesis compared to what we lose — by narrowing our scope to the three sugar papers from Havel’s group. We want to answer these two questions: 1). what does the food reward/palatability hypothesis of obesity add to our understanding of the fructose-related effects documented in these papers? 2). Is it a valuable addition?

As I said, few of us, if any, would doubt that sugar-sweetened beverages (SSBs, as they’re now fashionably known) or fructose-sweetened beverages have a high-food reward value. The monkeys seemed content to drink all their 300 calories-a-day of fructose, and to do so at the expense of other more nourishing food (the chow). Put simply, they apparently found the fructose more rewarding. Havel told me when I interviewed him for the NYT Magazine article that they had difficulty finding any of the younger (human) subjects for their experiments who didn’t already drink at least two daily sugar-sweetened beverages, and quite a few were drinking six.

So it’s certainly a reasonable assumption that SSBs are hyper-rewarding, among the low-hanging fruit in the food reward/palatability world, as , as Dr. Guyenet of Whole Health Source has categorized them in his posts on this subject. Easy targets in other words—indisputably rewarding.

Now these highly rewarding beverages are given in a randomized-controlled trial to human subjects, as Havel and colleagues did, and the amount given is fixed.  They’re given daily to monkeys as part of their dietary regimen, as Havel and colleagues did, again a fixed amount. In both species, they cause the metabolic/hormonal effects we would expect on the basis of a few decades of biochemical and endocrinological research on insulin and on the metabolism of fructose in the liver.

So now we’re going to take a leap of faith and assume that the hyper-rewarding nature of these beverages explains why the monkeys got fatter drinking them, and maybe the humans, too, if any of them did indeed gain fat in such a short period on the beverages, but what about the metabolic and hormonal effects observed?  Does the fact that SSBs have a high food reward/palatability value tell us anything meaningful about why these sugar- or fructose-sweetened beverages cause peripheral effects in addition to increased adiposity—insulin resistance, metabolic syndrome, type 2 diabetes, raised triglycerides and ApoB number, etc.?

Here’s another way to phrase that same question: If consuming these different sugar waters have very well-specified effects in the body (peripheral effects, as they’re known in the lingo), and these effects are sufficient to explain all the observations – increases in adiposity, insulin resistance, dyslipidemia, type 2 diabetes, etc. — why evoke a couple of arguably vague concepts like food reward and palatability to explain one of them alone, the increase in adiposity?

Let’s agree for the moment, that it’s worth it. We’re going to accept this hypothesis because it’s just right and the carbohydrate/insulin hypothesis is not. It’s the hyper-rewarding nature of sugar- and fructose-sweetened beverages causes fat accumulation, as seen in the monkeys in paper number one, and in free-living humans drinking SSBs. If we decide this reward value thing is at least a major contributor, we’re still stuck with having to explain the metabolic and hormonal disorders. We can’t just ignore those, so we have to complicate the hypothesis to do so — add epicycles, as historians or philosophers of science would put it. We have to explain how these SSBs also cause insulin resistance and metabolic syndrome, why they raise triglycerides and LDL cholesterol and ApoB etc..

Researchers who study obesity and insulin resistance have been confronted with this same problem since the early 1960s, once they decided that obesity had to be caused by taking in more calories than we expend. The way they’ve traditionally done it is to say that we get fatter because we take in more calories than we expend, and then we get insulin resistant because we are getting fatter. Two different mechanisms.

This rationale can help proponents of the food reward/palatability hypothesis out of the same hole. In this case, blame the excess fat accumulation on the hyper-rewarding nature of sugar and the neurochemical effects it induces in the brain; this raises set point and so we eat more than we did and get fatter. Then we can blame the insulin resistance and all that goes with it on the process of getting fatter.

Not too bad, but then we have to explain why the same metabolic effects happen in individuals who remain lean, who do not accumulate the offending fat, and we can’t, or at least not without adding yet another mechanism (another epicycle) for the lean people. Because lean people can also be insulin resistant and dyslipidemic and get type 2 diabetes and all the associated chronic diseases. So one way or the other (or so it seems to me), we need our knowledge of the peripheral effects of sugar consumption – the metabolic and hormonal effects in the liver, the pancreas and other organs – to understand at least some of the observations, if not all of them.

In its current form, the palatability/food reward hypothesis explains only fat accumulation. It explains why the monkeys gained weight, and I guess why we do too when we drink a lot of sugared beverages, but it offers no explanation for the the metabolic consequences that accompany the weight gain, both in the studies discussed and in real life.

Here’s yet another question to always keep in mind when thinking about the food reward/palatability hypothesis of obesity: Do we gain anything from it other than the obvious: that if these foods weren’t highly rewarding or palatable, we probably wouldn’t eat them, or at least not enough of them to make us fat and cause the deleterious metabolic effects?

And even if we do gain something meaningful from acknowledging that the food reward/palatability value of foods causes us to consume more of them, this benefit to our understanding would still come with a very important caveat: Maybe the reason these foods have a high reward value or are so palatable is because of the metabolic and hormonal effects of eating them. Maybe a food that can cause insulin resistance or hyperinsulinemia can also cause us, through its various peripheral effects, to crave it more, to find it more rewarding.

Anything’s possible, and this is one hypothesis, a very important one, we should always keep in mind. And it’s one I’ll return to shortly, or at least shortly by my standards—i.e., over a few more posts and several thousand words more words.

(And when I’m done with these food reward hypothesis posts, I’ll get to the evidence against the insulin hypothesis. Lord knows when that will be, but I’ll get there.)