Wednesday, August 26, 2015

How Much Does Sugar Contribute to Obesity?

Last week, the British Medical Journal published a review article titled "Dietary Sugars and Body Weight", concluding that "free sugars" and sugar-sweetened beverages contribute to weight gain.  But what are "free sugars", and why does the scientific literature suggest that the relationship between sugar intake and body weight isn't as straightforward as it may initially appear?

In a new review paper (meta-analysis), Lisa Te Morenga and colleagues review the studies evaluating the link between certain types of sugar intake and body weight in adults and children.  These studies include both observational studies and randomized controlled trials.  They conclude that the intake of "free sugars" and sugar-sweetened beverages are linked with higher body weight in both groups (1).

This conclusion appears sound and I have no quibbles with it.  But what are "free sugars"?  And why does this conclusion seem to be at odds with an older literature suggesting that people who eat more sugar tend to be leaner?

What are "free sugars"?

Here is the definition of "free sugars" they provide in table 1 of the paper, which is a standard definition used by the World Health Organization and the Food and Agriculture Organization:
All monosaccharides and disaccharides added to foods by the manufacturer, cook, or consumer; sugars naturally present in honey, syrups, and fruit juices.
So the term  refers to added sugars and fruit juices, but excludes the sugar that occurs naturally in fruit.  Importantly, it doesn't refer to total sugar intake, but rather to a major component of total sugar intake.

In plain language then, what the authors found is that added sugar and sweet beverage consumption are associated with a higher body weight in observational studies.  In controlled trials, these sugars increased body weight when calorie intake wasn't held constant, and had no effect on body weight when calorie intake was held constant.  For me, this conclusion is consistent both with the scientific literature I've read, and with common sense.

They do state in the paper that their result applies to specific types of sugar intake, rather than total sugar intake, but at certain points it sounds as if they're referring to total sugar intake.  For example, the title of the paper doesn't specify that the paper is specifically about added sugars and sweet beverages.  This could easily lead to misunderstandings about what they actually found.

The relationship between sugar intake and body weight is more complex than you may realize

When I first skimmed through the paper, I thought it was about total sugar intake, and I was surprised to see that they found an association between sugar intake and a higher body weight.  Why?  Because most of the observational studies that have examined the association between total sugar intake and body weight have found that people who eat more total sugar weigh less.  And the remaining studies found no association.  There is virtually no observational evidence that people who eat more total sugar weigh more than people who eat less, or gain more weight over time.

The hypothesis that sugar intake could be linked to weight gain is a pretty obvious one, and it's been around for a long time.  Consequently, many observational studies have evaluated it, beginning in the 1970s.  James Hill reviewed these studies back in 1995 (2), concluding:
Carbohydrates, particularly refined sugars, are still widely assumed to be fattening. However, there is now a substantial body of epidemiologic evidence refuting this view...  Almost all of the above studies support the contention that a high-carbohydrate, high-sugars diet is associated with lower body weight and that this association is by no means trivial.
Whether or not you agree with this hypothesis, it's still pretty interesting to note the marked difference between this conclusion and the result recently published by Te Morenga and colleagues, particularly since many of the same studies were available to be included in both review papers.

A major difference between the two review studies is that Hill was concerned with total sugar intake, while Te Morega was concerned with added sugars and sweet beverages specifically.  Today, few diet-health observational studies focus on total sugar intake as an outcome; usually the outcomes are focused on sugar-sweetened beverages, fruit intake, or other subsets of sugar intake.  Maybe people got tired of seeing that total sugar intake is associated with leanness?  Maybe the finding isn't novel enough anymore?  Or maybe it just makes the sugar-health story a little too complicated?  I find it a bit strange, personally.

These are observational studies we're talking about here, so we do have to be cautious about interpreting them.  Who knows how well self-reported sugar intake, and components of sugar intake, actually correspond to real intake.  And who knows to what extent differences in body weight are caused by differences in sugar intake, rather than other things that are associated with sugar intake.

Still, I think when we look at all of the data together, including the two papers I discussed in this post, a fairly logical and consistent picture emerges: added sugars and sweet beverages tend to be fattening because they lead us to eat too many calories, but whole fresh fruit isn't fattening and probably actually tends to be slimming.  When it comes to body weight, it's not so much the sugar itself, but the way in which it's packaged.

This is the same conclusion I reached in my 2012 post, Is Sugar Fattening?


Unknown said...

All I know is that just a little bit of sugar makes pure butter rather more-ish thus it must be hard to separate out the effects of sugar on foods or the sugar itself in observational studies. Personally I'm not that surprised by total sugar not being associated with obesity.

hcferris said...

So, this aligns with Dr. Robert Lustig's view that high blood sugar elevates insulin, which blocks your satiety hormone leptin. Thus, sugar contributes to obesity due to hormone cascade effects that lead to overeating. No?

Stephan Guyenet said...

Hi hcferris,

These studies weren't on the hormonal effects of sugar. I've followed every reference I could find in Lustig's talks and papers, and even asked him personally for references, and I have yet to see any evidence supporting his hypothesis that elevated insulin is the cause of leptin resistance in obesity.

Anonymous said...

I’m reluctant to reduce this to another “a calorie is a calorie” argument. I’m not a big fan of the discourse among health experts, going back & fourth on diet superiority.

That said, in terms of isoenergetic exchanges, how would dietary fat compare/contrast to bodyweight changes (compared to sugar)? Reading your post elicited this counter-question.

I know there is no simple answer. To that end, do you think there would be significant differences to BMI if the macronutrient source were fat, not sugar?

Table 4 and Figure 5 of the Te Morenga et al paper didn’t address this question adequately as it only looked at other forms of carbs, not fats.

Metabolically, sugar worse than fat? Or is the true concern, sugar's ubiquitous nature and sweet & effortless dose frequency?

Coconut guy said...

I come from low carb, Gary taubes style. Lost weight, but never thought it was because i avoid carb: quite frankly, just a look at mcdougall or other vegan high carb diets or japanese ones, lead me to the idea that there some associations of foods that are bad. Starch and fat are probably a bad idea if you want to lose weight. But starch alone or fat alone, no problem, you'll lose weight. I dont agree with professor guyenet and his idea of palatability, though. Yes,if you reduce greatly the paltability, you'll certainly lose weight, but for how long? Frankly, i really enjoyed my salami coated with my melted cheese or my almond butter and it was a very good reason to continue my diet (and i imagine that the so called "durianrider" still love his sugar-banana smoothie. I changed my diet after reading Ray peat articles and lost more weight and now, i'm perfectly lean (this guy is really fascinating). I dont believe sugar is fattening anymore, unless you call a doghnut or an oreo, sugar.

So, too be clear, based on my own experience:
-if you want to lose weight and eat starch, ditch the fats or eat low fat. Be an asian.
-if you want to lose weight and eat protein and fats, ditch the starch.
-but, what i found very interesting is that protein, fat, and simple sugars did not lead me to gain weight, but the opposite.

Eventually, the nutrition world remembers me the Medieval problem of univerals. Take a donut. Gary Taubes call that "a carb", mcdougall call that a "fatty food", so they both forbid it. Then S. Guyenet comes to say "no, that's food reward, high in calories and whatever the reason, both low carb or high carb eaters will avoid it". But to me, it's not that. It's refined starch, coated with polyunsaturated fat Omega 6, and sugar, and additives, and probably trans fat, and everybody should know that it's obesinogenic. But it's possible, like me, to have a food reward high in calories and lose weight, like organic ice cream with coconut oil with chocolate on it: high fat, high sugar, high palatability, lose weight. Why? Well, it's so satisfaying, that i eat less after. But i found that high starch and high fat recipe lead me to eat more.

Just a testimony, and thank you very much to the great Stephan Guyenet who remembered me the truth, too often neglected, "in the end, it's all about the calories" (although coconut oil lead me to doubt ahah!).

Thanks stephen for your wonderful website! Julien from France

Aaron said...

I still go back to the fact that if you eat a lower fat diet (we are talking less than 15% of calories) sugar is just not going to make you fat unless you are sucking down tons of it. The only people that can eat lots of fat and stay slim are low carb people (so of course they are going to bash sugar/starch). Only thing I might be worried about on high sugar is potential mineral balance issues and or a raising of uric acid if your levels are already borderline high.

Unknown said...

Easy for me to believe that a high-carb diet is associated with low weight, because when i ate in that fashion i had indeed a low BMI. However, that meant little muscle as well as almost no fat. I don't see any mention of the balance between muscle and fat in this post.
PS: this is my first ever comment on this blog!

Anonymous said...

It's really all quite simple. It's not about high or low carb, high or low fat. Both are perfectly healthy, and there are dozens and dozens of examples for both. From Okinawa to Kenya to the Inuit. That is, they are all healthy as long as they are low sugar. Why is sugar a problem? Because of the fructose. Why is fructose a problem? Because it induces ROS damage, inflammation, high uric acid, insulin and leptin resistance, and it drives overeating by targeting neuronal reward mechanisms while muting satiety. Besides, fructose is the main driver for cavities and digestive problems. Why is fruit much less of a concern? Because polyphenols and other iingredients counteract the effects of fructose.

There is only more factor that aggravates the metabolic damage inflicted by fructose: the inflammatory n6/n3 PUFA imbalance in modern diet.

The normal insulin response in a healthy person never has been a concern. Therefore (non-sugar) carbs never have been a concern: in fact a high-carb (low sugar) improves insulin sensitivity and fasting blood glucose. The problem is insulin resistance. Induced by fructose.

Unknown said...

Sugar can be addictive. Sugar also makes foods taste better, especially when combined with fat. Combine these factors and the result is that people eat more (i.e. sugar, fat, calories, whatever..) and gain more weight.

Teech said...


I remember an old post of yours where you detailed how much sugar consumption has increased since the 1800s. I believe you stated that Americans were consuming 13 lbs of sugar manually, whereas today we are consuming over 100lbs annually. How the heck are we consuming so much and doesn't it seem like sugar is more palatable and less satiating than fat and protein? Wouldn't you see the same for fat if it was equally palatable?

des said...

Spent last 12 months reading up on Nutrition and concluded, that unprocessed food is the only thing you should eat. Sugar for me means chocolate, doughnuts, cakes, biscuits/cookies, sweets,ice cream. I could easily consume over 2500 calories in one sitting in one hour if sat in front of the TV. They are addictive and not satiating. This is the calories in, calories out paradigm. I now eat only 2 main meals per day filled with vegetables and fruit (covered in Greek yogurt) + protein and some carbs (quinoa, beans, sweet potatoes). I don't snack. This works most of the time, but on occasion, I still binge on chocolate.

Tim said...

That picture of sugar crystals doesn't make me salivate, like the Food Reward Friday pictures sometimes admittedly do (not always, mind you! ;-). Thus it is not sucrose itself that makes poeple overeat (I have never seen anyone eating table sugar by the spoon) but all the highly addictive processed food of which sugar is an essential ingredient.

Aaron said...

This post makes me think fructose really isn't so bad given certain context.

Richard said...

The issue for me in this context is not exactly whether "sugar" (meaning table sugar mainly) is evil or grossly deleterious in any amount or use, the issue is whether large amounts of sugar are harmful, or not really good, in a nutritional sense. After all, sugar is optional, not a nutritional obligation. (The body can make sugar from protein, after all.) Taking sugar consumption to extremes...well, that's not really the point, to eliminate any and all sugars. That's probably a lot of work with a minimal return, and from an evolutionary point of view, I would suppose that at any opportunity primitive people ate all the honey they could locate and make off with. So the human body is meant to be able to cope with occasional consumption of sugar.

Considering this as a preamble, I would refer to the easily accessible work and popular presentations of Dr. Fung in "Intensive Dietary Management" in which he describes the effects of continual insulin production. The concept is, based on the work of Unger and Fung, that insulin and (considering the work of Unger) glucagon are bodily "pulses" of a powerful hormones, part of an effective negative feedback system to control/maintain blood proper sugar levels and fat storage. With continual (sugar) stimulation of the system that process/mechanism produces insulin on a continual basis (as opposed to pulses) and as a result the various cellular receptors begin to resist the insulin. This is the same drug resistance that is seen with drug addicts and with other drugs. It's not very mysterious, really.

But the result is terrible.

The idea is that people can and do eat sugar, and they have done so (in an evolutionary sense) from the time they could get their hands on it.

The point is whether the sugar is the problem, and in some sense it is, if it results in a sweet/starchy snack late at night, on a regular basis. It's not just sugar per se, it's when you eat it. In this same sense, the fat people are trying to lose weight, but cannot, because they are eating too many calories. From the perspective of insulin resistance the excess weight (fat, not muscle) is not just excess calories, it is a symptom of insulin resistance. You do not treat insulin resistance by cutting calories. Cutting calories is a side-effect of proper dietary management. Don't eat all the time.

The simplest and easiest way to avoid insulin resistance is with fasting. Intermittent Fasting is fasting. Eat dinner (mainly protein and fat) at 6PM or 7PM, and don't eat anything, don't drink anything except water, coffee, or broth until noon or 1PM the next day. That mid-day meal is breakfast. Do that for a few days, or a month. See the difference. You won't even have to weigh yourself. Just look in the mirror.

JasmineJohend said...

I am LCHF, my own N=1 reaffirms that this is the only effective diet plan for me, nothing anyone says will convince me in any way, everyone needs to find what works for them. However your work/theories on food reward are spot on. Whether we overeat sugar, starches, crap-in-bag or cream/bacon/eggs/nuts etc, it's not always about actual physical hunger, it does seem to stem from that Hungry Brain as you call it (great title). Some may not get this as food is just fuel to them and can't understand why we overeat... fine..but I personally am very grateful that you do. Nothing to add to conversation really except I am eagerly awaiting your book...but mid 2016 uggh...can't we get a sneak preview and scraps like Lyle McD is throwing to us with his upcoming book for women? :) You said in the Sigma podcast it draws a lot on David Kessler's End to Overeating which happens to be one of my favourite books on the subject. He offers strategies of control that actually work; sorry but "call a friend/have a bath/paint your nails" is done and dusted. Thank you Stephan

Jouker Sat said...

Sugar can be addictive. Sugar also makes foods taste better, especially when combined with fat. Combine these factors and the result is that people eat more (i.e. sugar, fat, calories, whatever..) and gain more weight.

Unknown said...

It depends. Individual has different absorbency, and someone also gain weight by "free sugar" --Some gene expression of them decided they are more easy to store fat( theories only suit for some people.

elizabeth192 said...

Stephen, are you aware that James O. Hill, professor at University of Colorado School of Medicine, whom you cite above, is president of a “non-profit” funded by Coca Cola called Global Energy Balance Network? Global Energy Balance Network was formed to promote the idea that the key to maintaining healthy weight is exercise and not diet.

Stephan Guyenet said...

Hi Elizabeth,

Yes, I'm familiar with the GEBN controversy. Dr. Hill seems to have gotten himself into some hot water lately. However, the papers I cited were published two decades ago, long before he was associated with Coke.

Unknown said...

Studies that examine the effect if HFCS-sweetened beverages neglect (I assume, as I don't read most if them) to account for the high percentage content of starch in HFCS. Starch requires much more insulin than sucrose. Sucrose has no starch. This is likely the reason of the ill effects of HFCS - sweetened beverages. Number of calories consumed is also a factor, off course.

Tim said...


Why do people always come up with funny hypothesis in order to make the facts fit their fancy (be it paleo, vegan or whatever) without even bothering to get the most basic things right - let alone to read the papers they critize?

No, HFCS does not contain starch, except for insignificant trace amounts. No, there is no evidence to suggest that HFCS is any worse than sugar. It's the same.

Unknown said...

Hi TIm
Take a look at this

Take a look at the nutritional label of table sugar

Then, at the nutritional label if HFCS

The amount of "carbohydrates" listed on a label if sucrose equals the amount of "sugars"
The amount of of "carbohydrates" listed on the label of HFCS is 3 times higher than that of "sugars". No fiber is listed and generally the difference, at least in part corresonds to the amount of starch. There may also be some moisture as the numbers don't add up.
Am I missing anything here? is the starch content insignificant?

I expected that the metabolism of a higher glucose content (starch) to require more insulin and more micronutrients. It seems you are right, from the few papers that I looked at briefly, it seems HFCS does not require more insulin than sugar
Then again it could be that the ill-effects attributed to HFCS are simply due to just consumig too many calories (i.e large size soda drinks etc.) and not the biochemical properties of HFCS.
Thanks for your comment.

thhq said...

@elie, as I understand it, HFCS is generated by enzymatic treatment of corn syrup, which is glucose made by treating corn starch with amylase. If the corn syrup isn't 100% pure glucose, the unconverted starch left in the corn syrup would be the source of starch in HFCS.