Holy cow. It's a small world! One of the main guys behind this is actually a great friend of mine. I've known him since kindergarten and he's now amongst the top researchers in the world. I wish I had something to do with it! He recently moved back to Canada from Copenhagen.
He has no clue what Hacker News is, but I'll email him this thread, perhaps he'll chime in if any of you have questions.
Very true, the fist thing I wanted to look for - "how did my favourite high-fat, low-carb diet rate in this study? Oops, how are the results of any use then ..."
Maybe it's just overzealous copywriting, but it seems like all they figured out is that a low GI, high protein diet is better than the other combinations they tested, and are far from solving anything rigorously.
Bodybuilders have been saying this for a while. There are always arguments about thermodynamics and what not, and there is good reason behind those, but everyone in a gym who builds up a lean body (sans steroids or not) follow a similar dietary pattern: lean meats, veggies, no flour/sugar/starchy veggies, etc for cutting and then eating more protein and some complex carbs (like from potatoes) to build lean mass.
I have trouble taking such studies seriously. I have a life-threatening medical condition which typically causes people to be underweight and I was quite thin as a child but put on a lot of weight as an adult (probably in part due to being prescribed steroids for my health issues). After finally getting an appropriate diagnosis late in life, I did some research and changed my diet to address my health issues. I have lost several dress sizes, even though that was not my goal. I was getting plenty of positive male attention when I was a size 24-26 and my doctor was telling me to work on my health issues and not worry about my weight (because people like me are typically underweight, which is more immediately life-threatening than being overweight), so I had no goal of losing weight. I was just trying to get healthier because I was so sick, on a lot of medication and miserable.
My experience tends to fly in the face of the pronouncements such studies make, so I just have a really hard time taking it seriously that you can come up with a generic suggestion for diet that will work equally well for 6 billion people.
I applaud you for staying strong throughout your struggle. However, your exception proves the rule: people who don't have a reason shouldn't be overweight. Or underweight.
I don't see how it proves that at all. Presumably everyone who is overweight (or underweight) is so for a reason. Given how many people wish they weighed less and can't seem to make it happen, often in spite of dieting, exercising, etc, I think something more complicated is going on than these studies assume.
Prof. Arne Astrup, a colleague of mine, was the principal investigator of this study when I was in Denmark. Interestingly, I will see him this weekend in Quebec City for the PhD defense of my wife. I agree with you guys that the title is too strong and we have to be careful when interpreting studies. Only few people fully realize that the obesity problem is complex and multifaceted. Nobody is immune. The causes and paths to weight gain are manifold and a narrow focus on the Big Two factors (unhealthy diet and physical inactivity) is not enough. For some people, factors such as short sleep duration, mental stress, socioeconomic status, drugs, etc. can cause weight gain and thereby should be included in the big picture. The main problem lies in the current 'obesogenic' environment that encourages overeating and sedentariness, and people are not able to swim upstream against those strong pressures. A change in the environment is more than needed at this point, because telling people to eat less and exercise more does not really work.
I've thought the argument in "Sugar: The Bitter Truth" is very insightful. Viewable at http://www.youtube.com/watch?v=dBnniua6-oM. Pointing out that the Atkins diet (no carbs, fat okay) and the Japanese diet (mostly carbs, low fat intake) have in common that they eliminate sugar, the major cause of obesity.
These studies are nutty. You can't separate diet from activity; you can't treat wide categories of food so indiscriminately; you can't use BMI or weight instead of body composition, and get definitive results.
Just for example, raw foodists tend to have low-protein diets that may be low OR high-glycemic index, and tend to be very thin.
Without a comprehensive listing of specific meals, I would guess that people in certain groups in this study would be much more likely to cheat on their diets. The selection of low-protein, low-GI foods for most people used to typical Western diets would lead to very disappointing meals and thus more cheating over a 6-month period.
Further, gaining a pound of fat over a period of 6 months works out to less than 20 extra calories per day. This is essentially noise. Cheating with 150 extra calories once per week (e.g. a single sugary treat) could account for this.
Keeping your blood sugar in check is a major factor in preventing binge eating and fat storage. There are various ways to do this, including low-GI foods, exercise, and eating small meals frequently. The recent standard for Hollywood stars and starlets is working with a personal trainer and eating every 2 hours. This helps keep blood sugar steady.
On the flip side, people engaged in sustained cardiovascular exercise must eat high-GI foods or their performance drops. Riders in the Tour de France chow down on every sugary snack and drink known to man (not just "performance formulas" of energy bars and drinks), and Michael Phelps eats 12,000 calories a day while training.
You don't need that level of Herculean effort before this becomes important. Even if you are completely sedentary, if you are engaged in a mentally taxing task, your brain is burning through a lot of glucose. Eating sucrose means half glucose and half fructose, and only the glucose can be metabolized immediately. Fructose can be stored by the liver and converted, but if the liver reserves are full (e.g. late in the day, as opposed to waking up from fasting during sleep), fructose will end up as fat. Fructose has a low glycemic index, so guess which sugar is more likely to be included in a low-GI diet. (Glucose is of course used by diabetics in order to raise their blood sugar quickly. It is also called Dextrose.)
Finally, without body composition analysis, they haven't shown anything about fat loss. High protein diets require more water to process, and a difference of a pound or two on a human body can easily be down to differences in fluid retention. High protein diets can lead to slightly more dehydration.
The study is impressively well done, as are most articles in the New England Journal. Other researchers will be studying their methods as much as their results.
> Without a comprehensive listing of specific meals
Actually, they did better and developed local indexes for all foods for each center and had the patients keep food diaries.
> Further, gaining a pound of fat over a period of 6 months works out to less than 20 extra calories per day. This is essentially noise.
First, this is why we enroll more than one patient. The noise averages out. Two: the unit of measure was kilograms, the average arm gained more than 1 kg. Three, see figure 2B in the article: they quite nicely demonstrate sufficient granularity to appreciate signal.
> Keeping your blood sugar in check is a major factor in preventing binge eating and fat storage
> people engaged in sustained cardiovascular exercise must eat high-GI foods or their performance drops
Your liver has glycogen stores on board that will keep your blood sugar rock steady for the first 20 miles of a marathon. Unless you are exercising at a heart rate of 160-180 for 2.5 or more hours, this is useless trivia.
> Riders in the Tour de France chow down on every sugary snack and drink known
> Michael Phelps eats 12,000 calories a day while training.
Some of the most over-cited factoids on the Internet. I swam competitively for 7 years and raced bikes competitively for four, and still ride. I'm also a doctor who spent a fair amount of a surgical internship with the bariatrics unit. These facts about elite athletes do not bear on anyone who has a BMI of 34 (Patients are generally ineligible for bariatric surgery below a BMI of 35)), the average human, or, almost certainly, you. A far more useful factoid is that your glycogen stores are depleted overnight, so a 2-3 hour cardio session in the morning will force you to burn fat. The body has no other choice. (No, it won't burn muscle).
> if you are completely sedentary, if you are engaged in a mentally taxing task, your brain is burning through a lot of glucose
Normal adult human brain uses, 120 g to, maybe, 220 g of glucose per day. At the outside, that's 750 Calories.
> sucrose means half glucose and half fructose, and only the glucose can be metabolized immediately. Fructose can be stored by the liver and converted, but if the liver reserves are full (e.g. late in the day, as opposed to waking up from fasting during sleep), fructose will end up as fat. Fructose has a low glycemic index,
True. Unfortunately, your implicit conclusion . . .
> guess which sugar is more likely to be included in a low-GI diet
Is what? Are you saying they loaded these cats with fructose? If they did, then these results are even more impressive, because we would expect a high fructose diet to promote fat gain. Are you saying they loaded them with glucose? That would make no sense, because glycemic index measures "blood glucose".
> Finally, without body composition analysis, they haven't shown anything about fat loss.
Even so, the media described in the abstract makes no mention of body composition; without that, there is no way this study can be considered adequate, let alone "impressively well done".
> Your liver has glycogen stores on board that will keep your blood sugar rock steady
Your reply is confused. There are two different points that you are conflating.
One is in regard to keeping blood sugar steady to prevent binging by AVOIDING high GI foods; the other is performance-related, and the NECESSITY of using high GI foods in specific instances.
> for the first 20 miles of a marathon
All you're saying is that if you carb load in advance, you will have glycogen until you run out of glycogen.
You're incorrect about "rock steady". Typically, blood sugar will actually spike at first, as the liver releases glycogen / glucose at the onset of exercise. And guess what happens following a spike...
> Unless you are exercising at a heart rate of 160-180 for 2.5 or more hours, this is useless trivia.
I specifically mentioned sustained cardiovascular exercise AND the Tour de France in regard to this. I give specific examples, and you scoff and say it wouldn't matter -- unless someone were doing something I specifically mentioned as an example. Are you objecting just for the sake of objecting?
By the way, heart rates are generally lower (in the 140 range, not counting time trials and climbs) and eating is fairly continuous in the TdF.
> Normal adult human brain uses, 120 g to, maybe, 220 g of glucose per day. At the outside, that's 750 Calories.
Is this supposed to be an objection? Also, your math is wrong, unless you're using some kind of magic low-calorie glucose (220 x 3.81 = 838).
If you're eating sucrose to get glucose, double the calories. That's a lot of sugar.
Also, things like playing chess or programming all day are a lot more mentally taxing than "normal" activities.
If you know of studies with either group, I'd love to see them.
> These facts about elite athletes do not bear on
You apparently missed where I specifically cited Phelps's herculean effort.
The only difference is the scale; I used Phelps as an illustration of what I stated at the top: You can't separate diet from activity. I'm not sure how you managed to miss every point and misconstrue every example.
> A far more useful factoid is that your glycogen stores are depleted overnight,
I mentioned this already in passing.
> so a 2-3 hour cardio session in the morning will force you to burn fat.
Anything increasing metabolic demands will result in increased oxidation of fatty acids even if you aren't fasted.
> (No, it won't burn muscle).
That's sheer nonsense. The body is already in a catabolic state after fasting, and adding 2-3 hours of cardio will absolutely burn muscle.
> Are you saying they loaded these cats with fructose?
Low-protein, low-GI diets can well be high in fructose. Do you have the food diaries?
> If they did, then these results are even more impressive, because we would expect a high fructose diet to promote fat gain.
They gained weight.
> First, this is why we enroll more than one patient. The noise averages out.
Once again, you missed the point. If a certain diet makes it very likely that people will cheat, you can end up getting a lot of dropouts (as they did) and the ones who stay in will be shifted toward the higher end. This makes the entire result set suspect.
> Two: the unit of measure was kilograms
Do you get points now for stating the obvious?
From the article: "The average weight regain among all participants was 0.5 kg". I took the liberty of translating that into about a pound.
> they quite nicely demonstrate sufficient granularity to appreciate signal
I already proposed the factor of water retention which would quite plausibly account for a kilo of difference.
> See Table 2 of the study.
I can't see it, but I can read the description, if you mean Figure 2:
"Panel B shows the change in weight for each of the dietary groups during the weight-maintenance intervention, adjusted for body-mass index at randomization, weight loss during the low-calorie-diet phase, sex, family type (single-parent family, two-parent family with one parent as participant, or two-parent family with both parents as participants)"
Absolutely nothing about body composition. It's just weight and BMI. They don't determine body fat percentage, lean mass, water weight gain.
It's also amusing how all the points got sucked out of my post and went into yours. I assume people saw "I'm also a doctor" and assumed that an argument from authority was valid. And you're not even an authority when it comes to sports medicine, which is the bulk of what I was referring to.
That's pretty disappointing, but I suppose the average reader of HN has even less knowledge about these things than you. Even so, your reply was obviously objecting spuriously and the rest was either repeating my points or stating the obvious. Very weak.
Being a bariatric intern has no bearing on sports medicine; it's practically the opposite end of the scale. You're concerned with gross measures like BMI. Do you even care about a kilo of difference on an obese subject? Yet for this study, with figures that small, body composition is the only way to figure out if it's relevant. A couple pounds of "weight" is meaningless.
He has no clue what Hacker News is, but I'll email him this thread, perhaps he'll chime in if any of you have questions.