Science Thursdays: Fat Chance

I'm trying my hand at a recurring piece here in ONAF: science writing every Thursday. Let me know if my idea is a good one or not.

Either way, today I rant about obesity measurements.

The News and Observer recently reported that North Carolina is the 14th fattest state in the country, with 63% of our residents tipping the virtual scales as officially overweight or obese. Mississippi took top dishonors with 67%, while outdoors-friendly Colorado was at the relatively shapely bottom with 55%. Combined with a recent report that being just a little bit overweight can have serious health effects.

Then again, it may not be such a bad thing. A recent study from a Brazilian team demonstrated that people classified as "overweight" are actually at lower risk for heart disease than those classified as normal. So should we all break out the Twinkies and save our hearts? Or is there something else at work here?

The Brazilians certainly think so. The problem, they say, lies in the instrument used to classify people as "obese" or "overweight"; the body mass index (BMI).

Few serious doctors doubt the link between fattiness and diseases such as type 2 diabetes, heart disease, hypertension, and other ugly stuff. In an ideal world, we'd be able to find out how much fat we have in our bodies and diagnose ourselves from there. This function is served by the body fat percentage, which is the mass of fat in your body divided by your total mass. It sounds simple, but it's very difficult to measure. Some relatively less annoying methods for estimating body fat percentage are chronicled in the Wikipedia article, but as the article states, the only truly accurate way of going about it is by using dual energy X-ray absorptiometry, or a bone density scan. You probably don't have one of those machines lying around. Most people don't - it's costly, and doctors don't generally want to waste a DXA on someone who doesn't have osteoporosis.

You probably do, however, have a scale, and this is where the BMI comes in. Invented by Belgian Adolphe Quetelet during the mid-19th century, it was intended as an estimate of how overweight someone was. The BMI is simply your weight in kilograms divided by the square of your height in meters. If that's greater than 25, you're overweight. Over 30, you're obese. Nice and simple.

And it works, too, provided you make a simple assumption: everyone has the same ratio of muscle to fat. And this is where the BMI breaks down. Since muscle is more dense than fat, someone with more muscle than average can easily tip the scales as "overweight" even if they don't have any more fat than someone who is "normal." Similarly, someone who has a good deal of fat but very little muscle could sneak in under the limit and be mistakenly classified as "normal." Dallas Cowboys wide receiver Terrell Owens, for example, has a BMI of 28.2 - substantially overweight. Saints running back Reggie Bush tips the scales at 27.9. Poor 'Skins running back Clinton Portis has a BMI of 30 - he's obese.

Of course, no one in their right mind would accuse these athletes of being fatasses. No one would think to believe the BMI diagnosis where they're concerned. What the Brazilian study demonstrated was that the BMI index was so deeply flawed that its usage was suspect even among everyday people. Perhaps the BMI works among generally sedentary Americans. But if you take a group of people who have gained muscle through more regular physical activity - say, Brazilians - the BMI loses its effectiveness.

Furthermore, the BMI ignores the differences between acceptable fat levels for men and women. Women are generally supposed to be fattier than men; men should limit themselves to a body-fat percentage of 17%, while women can get away with 24% and still be considered fit. This is a distinction that is lost on most people who use BMI to determine risk factors, however, and one that isn't used in determining statistics like the ones I cited earlier.

We need a better way to estimate one's risk of disease from obesity. Fortunately, there are several options. Scientists have proposed the use of waist circumference, waist circumference-to-height ratio, and the waist circumference-to-hip circumference ratio. A German group found that of these measures, waist circumference-to-height ratio was the best at predicting risk of heart disease. So divide your waist circumference by your height. Men, if this number is higher than 0.55, you're in trouble. Ladies, your magic number is 0.53.

This isn't perfect, mind you; shorter people and people with thicker builds are more likely to be flagged by this method. It has its problems, just like BMI. However, in the absence of a DXA machine in every house, quick-and-dirty methods like the waist-to-height ratio and BMI will have to do. Just take them for what they are: an approximation. And most importantly, know their limitations. You'll be fine as long as you don't think that BMI is the alpha and omega of fitness. Now if someone would only tell that to the BMI-crazed statkeepers out there...