“Even as you get near the upper reaches of the normal weight range, you begin to see increases in chronic diseases,” said JoAnn Manson, chief of the Division of Preventive Medicine, Brigham and Women’s Hospital, HMS Michael and Lee Bell Professor of Women’s Health, and HSPH professor of epidemiology. “It’s a clear gradient of increase.”
Yeah, she would say that.
Thin people.
And then there’s Frank Hu, professor of nutrition at Harvard:
The studies that Flegal [the author of the original study finding a negative correlation between body mass index and mortality] did use included many samples of people who were chronically ill, current smokers and elderly, according to Hu. These factors are associated with weight loss and increased mortality. In other words, people are not dying because they are slim, he said. They are slim because they are dying—of cancer or old age, for example. By doing a meta-analysis of studies that did not properly control for this bias, Flegal amplified the error in the original studies.
Hu ain’t no fatty either:
This all came pretty fast. It was only last month that the much-publicized (here and here, for example) study came out, claiming a protective effect of being overweight. Post-publication peer-review is a good thing.
If you are overweight but not obese, eat 4 blueberries a day, an apple every Tuesday, 1tablespoon of olive oil with lunch, dark chocolate, and 4 pounds of tofu per day, your live expectancy increases by 0.6 years. The study involved 100 million people, so you’d better believe it.
All that tofu just makes it seem longer.
The signal-to-noise ratio among stories in the news about nutrition is pretty low. Blame the media. Blame the fact that these stories get ratings. Blame the university PR departments. Blame the researchers trying to get enough notoriety to get resources. It’s not a system set up to find truth easily.
And then the industry I’ve worked in nearly all my career gets trashed for nutrition sins in a long NYT article:
http://www.nytimes.com/2013/02/24/magazine/the-extraordinary-science-of-junk-food.html?pagewanted=all
(from my point of view, the article is not a balanced presentation of the facts, but it is a presentation of facts that are worth understanding and which are not terribly positive).
Not that organics are the answer. Behind the wholesome demeanor, organic farmers often think other organic farmers are cheating — an easy thing to do since there’s no real inspection of veggies (harder with milk). Much organic produce now comes from Mexican farms; easy to be a bit skeptical about that.
Still, it’s encouraging to see that post-pub peer review can work in at least a few cases that could affect health.
That study is one example where the NYT actually published a Letter to the Editor from an academic who is an expert on the matter. I found the comment from a postdoc at Columbia illuminating:
http://www.nytimes.com/2013/01/05/opinion/the-link-between-weight-and-health.html?ref=todayspaper&_r=1&
“…recent meta-analysis showing 6 percent lower mortality among those who are overweight compared with those of normal weight. But epidemiologists have long suggested that this paradoxical finding has other explanations. For example, people often lose weight because of the illnesses that ultimately take their lives, and smoking is known to lower body mass index. Both artificially inflate death rates among those of normal weight and help explain this paradox. Studies effectively controlling for these issues have shown that the overweight, in fact, have a higher mortality risk than their normal-weight counterparts.”
Here is an example of such a study:
http://www.nejm.org/doi/full/10.1056/NEJMoa055643
That study is interesting but it is kind of bullshit.
The abstract says that “An analyses of BMI during midlife (age of 50 years) among those who had never smoked, the associations became stronger, with the risk of death increasing by 20 to 40 percent among overweight persons” but the article makes it clear that they are comparing the full overweight range BMI = 25-29.9 with a restricted normal BMI range of 23.5 to 24.9. The real normal range is typically given as 20-24.9.
Basically, the ideal BMI for nonsmokers is something close to 25 going up pretty symmetrically as you go down to 20 or up to 30.
I misread the chart. There is a recalled weight at age 50 figure for never smokers that does have the ideal weight at a BMI of about 24.
My own personal pet hypothesis, not well supported by evidence due to difficulties with measurement, is that fitness may trump weight as the most important factor. But given how tightly these two things often correlate, it is challenging to prove things either way. Maybe one day I will try one more time to look at this.
My own pet hypothesis is that lack of serious theory and warped incentives result in fat chance science.
One of the best metrics is body fat percentage. But it’s not very popular because it’s much harder to measure than BMI. There are studies showing that individuals with “healthy” BMI but elevated body fat percentage (because they have very little muscle mass) are at the elevated risk of cardiovascular disease and diabetes. Google “normal weight obesity”.
Another metric that seems to work even better than body fat % is waist circumference, either as an absolute number or as a ratio of waist circumference to height. Current recommended thresholds are 102 cm for men and 88 cm for women. Anything above that indicates a significantly increased risk of CVD. (They are fairly conservative, 102 cm in men roughly corresponds to BMI of 30.) The idea is that, while body fat % measures total body fat, waist circumference measures mostly abdominal fat, which is worse.
Level of physical activity is correlated with health outcomes independently of fat levels. http://eurheartj.oxfordjournals.org/content/25/24/2212/F1.expansion.html
Sarcasm?! Tongue-in-cheek?! I guess so. But it’s hard to tell from just reading the post… (Just bein’ too thin skinned maybe!)
X
Oui, I’m just being silly because I have nothing statistical to add to the reports.
There is at least one statistical point one might make from the blog post:
“Most of the scientists who claim being thin is healthy are thin themselves”
may be true. But that is not so interesting if it is also true that
“Most of the scientists who do not claim being thin is healthy are also thin themselves.”
Oh darn! I thought you were endorsing Fat Studies.
Pingback: Thin scientists say it’s unhealthy to be fat | Fifth Estate
JoAnn Manson does not actually deny that there is a protective effect to being moderately overweight.
She says that there is an increase in chronic diseases. This is perfectly compatible with an improvement in mortality as many “chronic diseases” are even completely asymptomatic from the POV of the patient and only defined by changes in biomarkers (elevated blood pressure, for example, does not always lead to more heart attacks).
A couple of meta (i.e., not precisely statistical but worth noting if we’re going to talk about them) beefs with all these studies: (1) There’s no adequate way to control for stigma (in particular, the well-documented bias against fat people on the part of health care workers), even in studies that do adequately control for access to nutrition, access to health care, stress levels, socioeconomics, etc. Stigma has two effects: it is stressful, which increases mortality, and it means that fat people do not go to the doctor as often, which, you guessed it. (2) Even if we take as read that fatness is bad for your health, I have yet to see any convincing study that suggests there is a good way to permanently alter fatness. Also note: dieting typically leads to weight cycling which is much more dangerous than a consistent high weight. (And again, most of the studies that we’re thinking about in this category don’t or can’t control for dieting and weight cycling.) (3) Most meta of all: given that there’s a significant debate about the per se effects of fatness, but no significant debate about the per se effects of fitness, why are we (this is the public ‘we’) still focusing on fat? http://www.ncbi.nlm.nih.gov/pubmed/15942543
Amelia:
Good points. #2 is especially relevant from a statistical perspective. Different potential interventions that affect weight can have different effects on health.
Bringing up these points, though valid, anew in this discussion may (like AG’s title) create the false impression that there is a controversy in the field where there is none or little (but a lot of cognitive dissonance to cope with for fat people). Controlling for stigma in cultures and social networking or for dieting behaviour and weight cycling is possible and it’s not like the effects are so small that such studies are impossible.
But yes, losing weight is hard, so it’s better not to gain that much, ie prevention.
Uh, are you actually saying these scientists are credible because they’re skinny, or is there some subtle humor here that I’m missing? I think you just showed your ass. . .
There was not-that-subtle humor that you were missing.
Andrew is well known for flippant humor in his blog setting. Also for publishing papers about indirect interviewing of Zombies.
Pingback: Somewhere else, part 35 | Freakonometrics
You are joking but I suspect that it is no coincidence that those guys are jumping through hoops to prove that being slightly overweight is bad for you. Is the “credibility of science” really threatened by those who point out that it is better to have a BMI of 26 than a BMI of 20?