The article seems to mix "thin" and "normal weight" all the time, being just poorly written. Including underweight people and normal people in the "thin" category is a major mistake.
Also, this quote is troubling:
"That people at higher weights are going to be OK."
Not only NOTHING in the article supports that, almost every single piece or research shows the opposite: that in general, people at higher weights have much higher mortality rates. The article just shows that, when some chronic diseases are present, overweight people seem to fare better than non-overweight (which I take includes underweight people).
It really glosses over this with only a brief mention that's immediately dismissed:
"It’s true that these groups are slightly more likely to suffer from heart disease and some other life-threatening conditions in the first place. But many factors influence the likelihood of a person getting heart disease."
From what I've read obesity does cause heart disease but a much stronger factor in the correlation is that lack of exercise causes both obesity and heart disease.
Those higher mortality rates start above what qualifies as "mildly obese," though. Even discounting factors such as existing illness and smoking, people with higher weights (weights below the "obese" definition) in general have longer lives.
Once you go beyond the "overweight" (or perhaps mildly obese, which has a negligible uptick in mortality) then the curve starts changing very rapidly.
Body-mass index and mortality among 1.46 million white adults.
With a BMI of 22.5 to 24.9 as the reference category, hazard ratios among women were 1.47 (95 percent confidence interval [CI], 1.33 to 1.62) for a BMI of 15.0 to 18.4; 1.14 (95% CI, 1.07 to 1.22) for a BMI of 18.5 to 19.9; 1.00 (95% CI, 0.96 to 1.04) for a BMI of 20.0 to 22.4; 1.13 (95% CI, 1.09 to 1.17) for a BMI of 25.0 to 29.9; 1.44 (95% CI, 1.38 to 1.50) for a BMI of 30.0 to 34.9; 1.88 (95% CI, 1.77 to 2.00) for a BMI of 35.0 to 39.9; and 2.51 (95% CI, 2.30 to 2.73) for a BMI of 40.0 to 49.9. In general, the hazard ratios for the men were similar.
And the conclusion:
In white adults, overweight and obesity (and possibly underweight) are associated with increased all-cause mortality. All-cause mortality is generally lowest with a BMI of 20.0 to 24.9.
This is going to shock you, but there are a great many studies on this, not all of which come to the same conclusion. From a meta-analysis:
Relative to normal weight, both obesity (all grades) and grades 2 and 3 obesity were associated with significantly higher all-cause mortality. Grade 1 obesity overall was not associated with higher mortality, and overweight was associated with significantly lower all-cause mortality. The use of predefined standard BMI groupings can facilitate between-study comparisons.
This is going to shock you, but I guess you didn't read the Harvard link addressing that specific study that I posted on my first post. But here it goes:
I've read this in the past, yes. The initial study you posted suffers from the same complaint. Which is not shocking given that the JAMA results are a meta-analysis (which include your study).
Please don't use "I guess you didn't read" as a rhetorical device. That sort of point-scoring only makes comments more abrasive and doesn't add information.
I can basically guarantee that the low BMI cohort in this study had significantly reduced lean body mass (skinny-fat), which greatly predisposes you to injury by things like falls and what not as you age.
Yes, but it makes it trivial to get a good p-value, so people use it. The state of nutritional epidemiology is abysmal, and has been for much longer than that of social psych.
The thing is, we have better measures than BMI, many of which are dead simple (e.g. % body fat as measured by calipers, or waist measurement). I also wish nutritional epidemiologists would stop using linear models all the time -- it's well known that mortality curves are U-shaped, as excessively thin people have high rates of all-cause mortality.
Returning to your original point, I'm 6'3" and 220 lbs. I have 12% body fat. Somehow this makes me fat, according to BMI.
BMI also seems to have the problem that it correlations are ethnically-dependent. The big outlier are East Asians, who tend to start having obesity-related issues in the upper end of "normal" and have less issues in the upper end of "underweight". Those categorizations were based on statistical studies of Europeans.
There are studies showing that people taking vitamin supplements are in better health condition than those that don't, even though it has also been shown that taking vitamin supplements doesn't beat placebo. So when looking at people taking vitamin supplements, the studies are biased simply because people taking vitamin supplements are the kind of people more inclined to care about their health in general.
So one has to wonder, for studies making the claim that overweight people are more unhealthy, is it the actual weight and body fat? Or maybe it's something they eat more of, causing organs to malfunction? Or is it that such people tend to care less about their health than other people?
I don't want to defend the claims in this article, as it's a pretty shallow one. But personally I take all nutritional advice with a grain of salt, including the one you've just made. This is because nutritional studies cannot be made double-blind, the study groups cannot be controlled and for any significant conclusion you'd have to do it for a large period of time and it's basically impossible to control the variables.
For example, for us to really measure the effects of being overweight and be correct about it, we'd have to have at least 2 groups with people of similar weights on average. And then you'd have to fatten one group, while feeding the other group a normal calorie intake. But the catch is that you'd have to give them the exact same diet, just different quantities. And you'd have to monitor them all the time, because people tend to lie about their diet. And then you'd have to do this for at least 10 years to produce significant results. And even then you can't be sure about the cause - as maybe there was something in that diet that in larger quantities caused negative or positive health effects. And you can see how such a study would raise ethical questions - I mean you'd have to intentionally cause harm to people.
That said, being overweight is not natural. Plus we know that being overweight is probably a sign for insulin resistance or a malfunctioning pancreas, or in other words type 2 diabetes. Many overweight people are on their way to type 2 diabetes. And that definitely can't be healthy.
So I tend to agree that the article is bullshit. But on the other hand the current practice of reductionist science really needs to stop.
AFAIK a plausible thesis for why overweight is correlated with better health is that the extra fat means greater energy stores in times of crisis. Basically, if you're thin and you get sick or injured, your body doesn't have the resources to heal itself effectively.
The article also mentions that there are some confounding factors such as people smoking having lower weight and disease itself being a reason to lose weight.
Not sure about that. Just one pound of fat has the caloric value to sustain you for over a day. 10% BF for males (and 15% for females) is low enough that your mother will probably say that you're too thin. So you could go for a week without ingesting any calories for a week if you were to lose 5% BF while you're sick.
I wonder if anyone has run the data using body fat percentage instead of BMI. BMI is a very coarse metric that will sometimes label short and/or muscular people as overweight/obese when they are quite healthy (think football players or weight lifters).
I came to post more or less this exact same comment. BMI does not take muscle mass into account at all. The Rock has a BMI of 31 (260lbs at 6'5") which is classified as obese. Hafþór Björnsson has a BMI of 41 (419lb at 6'9").
I agree with some of the other commenters too in that a full nutritional/lifestyle analysis over a long period of time is next to impossible because it relies on self-reporting which is known to be flawed [1][2]. People under-report their intake and overreport their exercise frequency and duration.
Uf you don't fit into any of the categories of people that would normally measure their own Body fat percentages (athlete of some description, normally), your BMI is probably a good estimate of your "shape". Sure it's not perfect, but it's a perfectly good estimate for a large number of people.
I wouldn't call it perfectly good. It assumes that weight should increase as the square of height when really you should be using something like the 2.6th power. That means that taller people will have a systematically higher BMI than they ought to which can be misleading when populations get taller due to better nutrition, say.
Speaking as someone who is a tad over 6ft tall, and not a body builder, it fits me quite well. My bmi when I started paying attention was roughly 26, which I was shocked at, and I got defensive about very quickly saying "it's incorrect for tall people" but the fact of the matter was I was about 20 pounds (10kg) overweight. It also quite well describes my partner, who is about 5ft5-6
> The Rock has a BMI of 31 (260lbs at 6'5") which is classified as obese
Why shouldn't he be classified as obese? If he was that heavy because he was working out then that would be one thing, but the reason he's that heavy is due to using steroids to put on low quality muscle mass. I don't see why that would be any less unhealthy than just sitting around eating doughnuts or whatever.
If it were as unhealthy as eating donuts (which you would need to prove), it would most likely be unhealthy by a different mechanism. So it would make sense to distinguish this if you want to find out what causes illnesses.
I also don't understand what you mean by "low quality muscle mass".
> I also don't understand what you mean by "low quality muscle mass".
Muscle with a low weight to power ratio, and/or with an unfavorable type I to type IIa/b ratio. E.g. you can quickly bulk up by lifting 3 x 3 at a high weight (or whatever Starting Strength recommends), but you're going to develop much lower quality muscle than if you just do 6 x 10 or whatever in terms of power. And similarly if you're not doing cardio in at least 90 minute increments then on a regular basis then your type II muscles aren't going to be able to properly utilize fat, because they'll never get past the point of just relying on stored glycogen or whatever.
Would you mind explaining "low quality muscle mass"? That doesn't seem like a real thing. Do you know how steroids typically work? They allow you to work out more and recover quicker. You're putting on "real" muscle, just faster than you could un-aided.
> They allow you to work out more and recover quicker.
They also act as anabolics, so you could argue that the muscle mass gained due to their anabolic effect is "low quality" (that's not what I'm doing, though).
That there are people like The Rock, Hafþór Björnsson who have a high BMI but are otherwise healthy* does not take anything away from the fact that a high BMI is a pretty good indicator that you're overweight or obese. It's an imperfect measure but realistically if you're at those Obese or above levels you're either overweight and you know it, or an outlier and you know it.
* = I suspect that a good number of guys like this are taking substances that would damage them in other ways that could lead you to describe them as unhealthy, but as another commenter said that's maybe not for BMI to measure.
Yes. As are you. Put it this way, BMI is used to calculate that 35% of people in the USA are obese and 34% are overweight. Is that because there are lots of people who are borderline ripped and BMI is flawed, or is it because there are lots of overweight people? Which do you see more of when you walk down your average street. If it's the former then I will concede that I am incorrect and that you're statistically normal, but we both know it's the latter and it's to your credit that you're likely to be very healthy.
Note: I'm making an assumption you're in a western country like the USA or the UK
BMI seems to only be a useful metric if you're interested in estimating the general health of, say, the entire human population of the state of Wisconsin.
For assessing individuals (or even small groups), I can't imagine it being very helpful.
It seems like these kinds of "studies" are only made to write headlines and to write books (notice how half of the people quoted in the article have some surprising new "facts" that they reveal in a book they wrote?).
The only paradox is why researchers still latch onto a calculation of health based on just height and weight that was concocted in over 150 years ago. What other field would rely on such a primitive, outdated formula and use it as the base of research?
> Another paradox is why the media always seems to portray meta studies (correlation) as fact (causation).
correlation vs. causation is a completely different axis of differentiation from meta-study vs. (direct) study, which is a completely different axis of variation from (something that isn't fact) to fact.
So, you've conflated at least three different distinctions there.
And low birth-weight children born to smoking mothers have a lower infant mortality rate than the low birth weight children of non-smokers [1]. Clearly that must mean that smoking while pregnant can "protect" your baby, and totally not a misuse of statistics.
Looking at either side of optimal BMI: People with a BMI of 26 (BMI 25-30 'overweight') have mildly better survival than those with 21 (18.5-25 'normal'). Push BMI up a little bit to above 28 and death rates are worse than with a BMI of 21. Not a huge paradox. Maybe we should just shift 'normal' BMI range to 20-27.
39 comments
[ 86.9 ms ] story [ 674 ms ] threadAlso, this quote is troubling: "That people at higher weights are going to be OK."
Not only NOTHING in the article supports that, almost every single piece or research shows the opposite: that in general, people at higher weights have much higher mortality rates. The article just shows that, when some chronic diseases are present, overweight people seem to fare better than non-overweight (which I take includes underweight people).
Even if they fare better under certain very specific conditions, in general they fare quite worse. Source: http://www.hsph.harvard.edu/nutritionsource/questions/ask-th...
"It’s true that these groups are slightly more likely to suffer from heart disease and some other life-threatening conditions in the first place. But many factors influence the likelihood of a person getting heart disease."
many factors, especially obesity?
Once you go beyond the "overweight" (or perhaps mildly obese, which has a negligible uptick in mortality) then the curve starts changing very rapidly.
http://www.independent.co.uk/life-style/health-and-families/...
http://healthland.time.com/2013/01/02/being-overweight-is-li...
Body-mass index and mortality among 1.46 million white adults.
With a BMI of 22.5 to 24.9 as the reference category, hazard ratios among women were 1.47 (95 percent confidence interval [CI], 1.33 to 1.62) for a BMI of 15.0 to 18.4; 1.14 (95% CI, 1.07 to 1.22) for a BMI of 18.5 to 19.9; 1.00 (95% CI, 0.96 to 1.04) for a BMI of 20.0 to 22.4; 1.13 (95% CI, 1.09 to 1.17) for a BMI of 25.0 to 29.9; 1.44 (95% CI, 1.38 to 1.50) for a BMI of 30.0 to 34.9; 1.88 (95% CI, 1.77 to 2.00) for a BMI of 35.0 to 39.9; and 2.51 (95% CI, 2.30 to 2.73) for a BMI of 40.0 to 49.9. In general, the hazard ratios for the men were similar.
And the conclusion:
In white adults, overweight and obesity (and possibly underweight) are associated with increased all-cause mortality. All-cause mortality is generally lowest with a BMI of 20.0 to 24.9.
Relative to normal weight, both obesity (all grades) and grades 2 and 3 obesity were associated with significantly higher all-cause mortality. Grade 1 obesity overall was not associated with higher mortality, and overweight was associated with significantly lower all-cause mortality. The use of predefined standard BMI groupings can facilitate between-study comparisons.
http://www.ncbi.nlm.nih.gov/pubmed/23280227
http://www.hsph.harvard.edu/nutritionsource/questions/ask-th...
Shocking, right?
https://news.ycombinator.com/newsguidelines.html
I can basically guarantee that the low BMI cohort in this study had significantly reduced lean body mass (skinny-fat), which greatly predisposes you to injury by things like falls and what not as you age.
Being lean is healthy, being frail is not.
Yes, but it makes it trivial to get a good p-value, so people use it. The state of nutritional epidemiology is abysmal, and has been for much longer than that of social psych.
The thing is, we have better measures than BMI, many of which are dead simple (e.g. % body fat as measured by calipers, or waist measurement). I also wish nutritional epidemiologists would stop using linear models all the time -- it's well known that mortality curves are U-shaped, as excessively thin people have high rates of all-cause mortality.
Returning to your original point, I'm 6'3" and 220 lbs. I have 12% body fat. Somehow this makes me fat, according to BMI.
So one has to wonder, for studies making the claim that overweight people are more unhealthy, is it the actual weight and body fat? Or maybe it's something they eat more of, causing organs to malfunction? Or is it that such people tend to care less about their health than other people?
I don't want to defend the claims in this article, as it's a pretty shallow one. But personally I take all nutritional advice with a grain of salt, including the one you've just made. This is because nutritional studies cannot be made double-blind, the study groups cannot be controlled and for any significant conclusion you'd have to do it for a large period of time and it's basically impossible to control the variables.
For example, for us to really measure the effects of being overweight and be correct about it, we'd have to have at least 2 groups with people of similar weights on average. And then you'd have to fatten one group, while feeding the other group a normal calorie intake. But the catch is that you'd have to give them the exact same diet, just different quantities. And you'd have to monitor them all the time, because people tend to lie about their diet. And then you'd have to do this for at least 10 years to produce significant results. And even then you can't be sure about the cause - as maybe there was something in that diet that in larger quantities caused negative or positive health effects. And you can see how such a study would raise ethical questions - I mean you'd have to intentionally cause harm to people.
That said, being overweight is not natural. Plus we know that being overweight is probably a sign for insulin resistance or a malfunctioning pancreas, or in other words type 2 diabetes. Many overweight people are on their way to type 2 diabetes. And that definitely can't be healthy.
So I tend to agree that the article is bullshit. But on the other hand the current practice of reductionist science really needs to stop.
I agree with some of the other commenters too in that a full nutritional/lifestyle analysis over a long period of time is next to impossible because it relies on self-reporting which is known to be flawed [1][2]. People under-report their intake and overreport their exercise frequency and duration.
[1]http://www.ncbi.nlm.nih.gov/pubmed/2082216 [2]http://ajcn.nutrition.org/content/76/4/766.full
Why shouldn't he be classified as obese? If he was that heavy because he was working out then that would be one thing, but the reason he's that heavy is due to using steroids to put on low quality muscle mass. I don't see why that would be any less unhealthy than just sitting around eating doughnuts or whatever.
I also don't understand what you mean by "low quality muscle mass".
Muscle with a low weight to power ratio, and/or with an unfavorable type I to type IIa/b ratio. E.g. you can quickly bulk up by lifting 3 x 3 at a high weight (or whatever Starting Strength recommends), but you're going to develop much lower quality muscle than if you just do 6 x 10 or whatever in terms of power. And similarly if you're not doing cardio in at least 90 minute increments then on a regular basis then your type II muscles aren't going to be able to properly utilize fat, because they'll never get past the point of just relying on stored glycogen or whatever.
They also act as anabolics, so you could argue that the muscle mass gained due to their anabolic effect is "low quality" (that's not what I'm doing, though).
* = I suspect that a good number of guys like this are taking substances that would damage them in other ways that could lead you to describe them as unhealthy, but as another commenter said that's maybe not for BMI to measure.
As of this moment I'm 5'9", 187, and borderline ripped. People often compliment me on my physique.
But my BMI is 27.6, which makes me "overweight."
Note: I'm making an assumption you're in a western country like the USA or the UK
For assessing individuals (or even small groups), I can't imagine it being very helpful.
But you can make a lot of money failing.
The only paradox is why researchers still latch onto a calculation of health based on just height and weight that was concocted in over 150 years ago. What other field would rely on such a primitive, outdated formula and use it as the base of research?
This just in: Margarine Causes Divorce in Maine [1]
1. http://tylervigen.com/spurious-correlations (99.3% correlation between Maine's divorce rate and margarine consumption)
correlation vs. causation is a completely different axis of differentiation from meta-study vs. (direct) study, which is a completely different axis of variation from (something that isn't fact) to fact.
So, you've conflated at least three different distinctions there.
[1] https://en.wikipedia.org/wiki/Low_birth-weight_paradox
Article makes lots of qualitative statements. Much easier to look at the data (900,000 person study): http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2662372/figure/f...
Optimal BMI appears to be 23-25 (at least for minimising the risk of death).
Another large study (1.46 million people) shows a similar result: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3066051/figure/F...
Looking at either side of optimal BMI: People with a BMI of 26 (BMI 25-30 'overweight') have mildly better survival than those with 21 (18.5-25 'normal'). Push BMI up a little bit to above 28 and death rates are worse than with a BMI of 21. Not a huge paradox. Maybe we should just shift 'normal' BMI range to 20-27.