Obesity is not a myth

There is a great conversation going on at Megan McArdle's blog with Paul Campos, author of The Obesity Myth. I say great because it give me the opportunity to show how astonishingly wrong Campos in suggesting that the obesity at the lower end of the BMI spectrum -- not just morbid obesity -- is not bad for your health. Frankly, contrary to his statements, the data is definitive, and I am amazed about how he gets away with publishing this trash.

A core argument that Campos makes in the interview is that there is no negative consequences for going over what he considers the very arbitrary line of a Body Mass Index of 30 (the standard dividing line between overweight and obese):

Obesity is defined completely arbitrarily as a body mass index of 30 or higher (175 pounds for an average height woman). Now body mass follows more or less a normal distribution, whiich means if the the mean body weight is in the mid to high 20s, which it has been for many decades now, then tens of millions of people will have BMIs just below and just above the magic 30 line. So if the average weight of the population goes up by ten pounds, tens of millions of people who were just under the line will now be just over it.

This might be meaningful if there was any evidence that people who have BMIs in the low 30s have different average health than people with BMIs in the high 20s, but they don't. At all. So the "obesity epidemic" is 100% a product of tens of millions of people having their BMIs creep over an arbitrary line. It's exactly as sensible as declaring that people who are 5'11 are healthy but people who are 6'1" are sick.

Adding to the absurdity of all this, people with BMIs in the mid to high 20s actually have the best overall health and longest life expectancy -- ,more so than those in the so-called "normal" BMI range.

The notion that the 30 BMI line is arbitrary is just nonsense.

In fact, huge prospective studies have demonstrated that individuals with greater than 30 BMI do have greater all-cause mortality than individuals who are simply overweight (BMI 25-29). Here is the money-figure. The Prospective Studies Collaboration compiled 57 studies with nearly 900,000 participants (a mammoth study!) to look at the relationship between different BMIs and overall mortality. The study confirmed the inverted U effect long argued for by scientists and physicians that having a BMI either below or above the normal range confers greater risk of death. Here is Figure 2 of that paper:

i-93be04e02f38569d6afd4eaa2e8915aa-PIIS0140673609603184.gr2.lrg-thumb-500x589-16771.jpg

You can see that above BMI 30, mortality increases above an average mortality maintained for BMI 20-29 -- a range that includes both individuals defined as of "normal" weight and overweight.

What specifically are the causes of the observed all-cause increase in mortality? The PCS identified several. HR stands for hazard ratio which is a measure of the risk relative to the general population.

  • Heart attack (HR 1.39)
  • Stroke (HR 1.39)
  • Diabetes (HR 2.16)
  • Kidney disease (HR 1.59)
  • Cancer (HR 1.10) including liver, kidney, breast, endometrial, prostate, and colon
  • Lung disease (HR 1.20)

Now, it is possible that Campos is not aware of this data. The study was published in 2009, and his book was released in 2004. It is possible that he has been living in a cave or refuses to acknowledge this evidence. But this data is as definitive as you can possibly get: becoming obese as defined by having a BMI >30 is bad for your health.

it has been a source of considerable controversy in the medical community whether BMI is the best indicator of how we should measure whether someone is at risk or not.

More like this