Tuesday, October 19, 2010

Can You Be Fat and Not Know It?

The short answer is yes, you can be fat and and not realize that you're fat.

A recent article in the Archives of Internal Medicine used data from the 2000-2002 Dallas Heart Study to describe this phenomenon. The full article (Body Size Misperception: A Novel Determinant in the Obesity Epidemic) is not yet available, but news descriptions of it are here, here and here. The authors have presented the data in powerpoint form here.

Misperception of body size
Researchers asked 2056 obese men and women to look at line drawings like those pictured above. The study participants were asked to choose the figure that looked the most like them. Although all of the participants met criteria for obesity, eight percent of them (14% of blacks, 11% of Hispanics and 2% of whites) did not consider themselves to be obese.

Let's take a moment to define our terms. Study patients did not have the benefit of the BMI values that are printed underneath each series of pictures. I've added those because I suspected that readers would like to take the test themselves and see (1) which figure they would choose and (2) which figure actually corresponds to their body size. (You can click on the pictures to enlarge them, if necessary.) If you don't know your BMI, it can be calculated here.

As obesity grows more common, it also becomes harder to pick out which figures are thin, normal and fat. Here are the definitions according to BMI:
  • Underweight = BMI less than 18.5
  • Normal weight = 18.5–24.9
  • Overweight = 25–29.9
  • Obese = BMI of 30 or greater

In the figures above, the only underweight figure is women's drawing 1. The overweight figures are drawings 5 & 6 of both sexes. The obese figures are drawings 7, 8 & 9 of both sexes. (Note that, in this context, "overweight" and "obese" have very specific definitions.)

Behavioral associations
Back to the Dallas Heart Study. Among the eight percent of obese study participants who misperceived their body size, 66% believed they were at low risk for obesity, even though they were already obese. Although those with and without body size misperception had equal probabilities of developing diabetes, heart disease and high blood pressure, those with body size misperception were significantly less aware of their risks for these conditions.

In practical terms, this meant that fifty six percent of obese subjects with body size misperception had seen a doctor in the past year. Of those, only 38-45% had discussed diet/habits, exercise or weight loss with their physician. By contrast, among the obese patients with accurate body size perception, 74% had seen a doctor in the past year and 64-68% had discussed such lifestyle issues with their physician. This suggests that obese people who do not realize their condition will get poorer medical care as a result of their misperceptions.

BMI and health risks
Not all causes of mortality can be related to obesity. A 2007 article in JAMA did an extensive analysis of the association of BMI with particular causes of death for the year 2004 in the United States. Researchers found that obesity was not related to increased deaths from cancer, respiratory disease or injuries. However, obesity was associated with increased mortality from cardiovascular disease and from diabetes and kidney disease. When all causes of mortality are considered, the graphs look like this (source article here):

Interestingly, these graphs show that a BMI of about 25 (between normal weight and overweight) seems to be the healthiest for long-term survival, and being underweight is actually unhealthy. However, the graphs do show that as obesity (BMI=30.0 or more) increases, all-cause mortality risk increases. For that reason, it pays to have an accurate idea of our body size, and to be willing to recognize that dealing with obesity could have a beneficial effect on our personal lifespan.


Jake said...

BMI is the most useless health measure in medicine today.
It tells you nothing about body composition and proper body comp is extremely important for longevity and health.

My BMI is 25.9 which puts me in the overweight category. But my body fat percent is 11.2 using skin fold calipers.

The medical industry should move to a 3 site skin fold measurement to determine body fat percentage. It is just as easy to learn as proper blood pressure measurement and takes about as long.

Yeah I know it is not super accurate, but it is 100 times more accurate than BMI in determining the health status of an individual.

Steve Parker, M.D. said...

For Americans over 65, the longest lifespans are seen at BMI between 25 and 30.

Thanks for reviewing this info for us.

Anonymous said...

This is so funny to me! I have ALWAYS had a fatter image than what I am. When I saw these pictures I thought, for sure I am a 9. I'm betting my bmi is off the charts. However when I did the bmi calculator, I am a 6. Still I think I am somewhat fatter than that 6 because i carry my weight around my waist, with small hips etc. I KNOW that even if I was in the "normal" range of bmi, i would probably still have a waist that was bigger than what they say it should be.

That said I am just not super stressed about it anymore. When I was 118 lbs and had a bmi that showed I was underweight, I still felt fat because my waist was never small.

I think more people have the opposite body image issue, they think they are much fatter than what they are.

deon10 said...

Actually the answer is NO...

BMI is a big bullshit tool that doesn't say anything about being fat or about health. It doesnt take into account bodyfat or musculature or bone weight.

Stargazey said...

Hi, Jake! I agree that BMI leaves much to be desired, but as somebody who supervises a medical office, I'd say that investing in skin fold calipers and doing 3-site skin fold measurements probably won't catch on. It's hard enough to get an accurate height, weight, temperature, pulse and blood pressure. And you wouldn't get paid extra for the skin fold measurements. (Sorry that it works that way, but it does.) So I think we'll probably have to make do with BMI and all its flaws.

Hi, Steve! Your observation about the longer lifespan for people over 65 with a BMI between 25 and 30 sounds correct, but I couldn't find a reference. Do you have one?

Hi, Nancy! I agree that people who frequent health-related websites may think of themselves as less healthy than they really are. However my interactions with people in everyday life lead me to believe that many people are in a state of denial about their weight and their health. They see fat people all around them, and they see doctors working medical miracles. With those things in mind, they figure that whatever they're doing is probably normal and they don't need to change anything. But I could be wrong.

Ned Kock said...

Hi Stargazey.

I like your blog!

These BMI-mortality findings are consistent with an analysis I conducted recently on the China Study II data:


In simple terms, the China Study II data seems to suggest that those who eat well, but not too much, live the longest. Those who eat little have slightly lower longevity (a bit of a blow to the calorie restriction theory of longevity). Those who eat too much seem to have the lowest longevity, perhaps because of the negative effects of excessive body fat.

BMI is not such a bad measure, because muscular but lean people are somewhat rare. The influence that this minority has on statistical tests tends to be relatively small. The larger the sample, probably the smaller the influence.

Stargazey said...

Hi, Ned! I like your blog, too. Thanks for introducing me to it!

Anonymous said...

Those conclusions about BMI are based on the SAD, not a diet free from the Neolithic agents of disease.

Anne said...

I just read the JAMA study that was linked. The underweight deaths were associated with respiratory disease particularly in Koreans who had TB ! I think the overweight and underweight in that study should have been more like for like and from a Western country such as the US or UK and not from one where infectious diseases such as TB are endemic as it skews the results of this study imho.