Should You Trust That Body Fat Test?

Depending on which Japanese conglomerate you believe, either I have the body of a 25-year-old or I'm pushing 70. Which is disconcerting either way, because I was a mess when I was 25, and I'd prefer to let 70 wait its turn.

But according to the statisticians at such companies as Omron and Tanita, my "metabolic age" lies at one of those extremes.

What's "metabolic age"? It is a statistical construct that, along with a growing number of other bells and whistles, is being built into the "body composition" monitors proliferating on store shelves.

There are a good half-dozen or so of these machines on the market now, at places including Target, with prices from about 30 bucks to nearly 300. Typically similar in shape and use to bathroom scales, their main purpose is to measure body fat.

They use a technology called bioelectrical impedance, which passes a small current through conductive foot pads or handheld electrodes (and, in some cases, both). The current can pass easily through water-rich muscle fiber, but it bogs down in fat. Based on a measure of impedance (how much of the current gets through from one electrode to the other), the machines use mathematical models to estimate the amount of fat that got in the way en route.

Why is this a good idea? It is pretty widely acknowledged that people should be less concerned with what they weigh than on whether that weight comes from fat or muscle.

The Mayo Clinic uses the (frightening) term "normal-weight obesity" to capture the issue. The scales may treat you kindly, in other words, but if too much of your weight comes from fat, as opposed to lean tissue, you run some of the same health risks as those who are obese.

For men, the aim is to keep body fat under about 20 percent of total weight; for women, under about a third, though the numbers vary with age.

Not content with that statistic, however, competing companies have begun loading their machines with lots of other stuff: estimates of how much muscle you have, how many pounds of bone, hydration levels, the amount of "visceral fat" larded around your vital organs, how many calories you need to eat in a day, and, based on all of the above, how "old" you are.

Which prompts the question: If the age estimate given by two monitors can be so divergent, what about the rest of the stuff the machines are supposed to measure?

To get a sense of the accuracy of retail-grade monitors, at least when it comes to that basic measure of body fat, I gathered five models from three companies and matched them against two clinical methods for taking the same measurement: a hydrostatic "dunk tank" test often used in research and the hand-calipers pinch test often performed in health clinics and gyms.

The two clinical measures, taken by doctoral student Andy Ludlow at the University of Maryland, raised a point that representatives for the monitor companies like to emphasize: that even accepted standards such as the dunk tank, which uses formulas related to the displacement of water and the comparative density of muscle and fat, are still only estimates of body fat. In the case of the monitors, the mathematical models involved have many built-in assumptions -- particularly when it comes to things such as visceral fat -- and if your body strays from that norm, results will be less accurate.