Further thoughts on BMI

Anyone who has taken his child to a pediatrician will have seen a growth chart. Here is one for boys aged 2 to 5 (pdf). It shows, for example, that the median height of a three year old boy is about 96 cm, and the 15th and 85th percentiles are 92 cm and 100 cm, respectively. In other words, the chart recognizes that there is a distribution of heights. The height of most (70%) three year old boys will be between 92 cm and 100 cm. The chart also gives the 3rd and 97th percentiles, which indicate that the height of the vast majority (94%) of three year old boys will be between 89 cm and 103 cm.

Suppose it were up to you to define a normal height range for three year old boys — something that parents could use to determine whether their child was of normal height for his age. What range would you use?

I think I would pick the 15rd and 85th myself. Would anyone choose the 50th and 97th percentiles? Would you be comfortable then advising parents that any child whose height was below the 50th percentile was “short or stunted”?

In effect, this is what the government has done with weight for adults. The NHANES I study measured the height and weight of 23,808 Americans (16,165 adults and 7,643 children) from 1971 to 1975 — before the onset of the obesity epidemic. The “healthy BMI range” of 18.5 to 25 corresponds to the 4th and 53rd percentiles of the adult population. So by today’s standards, 47% of the population in the early 1970s was “overweight or obese”!

If I were to define a normal BMI range, I would choose perhaps the 25th and 75th percentile of the BMI distribution of a reference population. If we use the NHANES I data as the reference population, those percentiles are 22 and 28.


5 Responses

  1. Thanks for your thoughts on this. I would certainly agree that the BMI is an imperfect measure; in fact, I remember hearing a few years ago that if you used the BMI to measure the Duke basketball team, over half would be classified as morbidly obese(!). So I certainly sympathize with your concern over these arbitrary metrics.

    However, I don’t really believe in such a thing as a “healthy weight range”; rather, I believe that the thinner you are, the better (all else equal). The data will support a fairly strong linear relationship between weight and risk for hypertension, heart disease, cancer, diabetes, etc. To some people this is a radical view because how could someone who is little more than skin and bones possibly be healthy? In fact, in nearly every living species studied, animals put on a severely calorie restricted diet (but still getting enough energy for basic bodily functions) will live healthier, longer lives. It’s not well understood why this is the case, but the relationship has been shown time and time again.

    So really my challenge to you is to say why people shouldn’t try to be as thin as humanly possible.

    • Justin: I have read about calorie-restricted diets and while I find the concept very intriguing, I remain skeptical that they could have the same effect on human lifespans that they do on animals. First, I’m not aware of any large-scale clinical trials in humans. There are many things that work in animal studies that do not carry over to humans. I know there are a lot of people practicing calorie-restricted diets, and I’ve read some glowing testimonials, but anecdotes are not data.

      In fact, the data I have seen show that it is just as risky to be underweight as it is to be overweight. Take a look at this paper, especially Figure 1. For most diseases, the curve of risk vs. BMI is U-shaped, with the highest risk at the extremes.

      From what I’ve read, people on calorie-restricted diets have to take vitamin supplements to ensure that they get the nutrients they require. It could be that underweight people in the general population are not getting all these nutrients, and that is why they have higher mortality than heavier people. If that’s the case, then despite any benefits, it would be dangerous to encourage people to be as thin as possible because most people would not take the time to ensure they were getting all the nutrients they require.

      Finally, when I think of it in evolutionary terms, people who live longer have more opportunities to pass on their genes. If the calorie-restricted diet is able to extend the human lifespan, evolution would have selected for those least able to feed themselves and their families: the worst hunters, the worst farmers. So on the whole I find it implausible, but I am keeping an open mind.

  2. Thanks for the thoughtful response.

    I will have to take a look at the paper you linked to. From what I’ve read, most sources citing a U-shaped relationship between BMI and health risks fail to control for appropriate factors. For example, many diseases and conditions cause people to lose lots of weight, thus leading to the assessment that lower weight causes disease, when in fact the reverse is true.

    I like your thinking on the evolutionary perspective. My take is that the reason evolution has not favored caloric restriction is because back in those days people did not have the luxury to be as idle as we are today, and so needed to consume lots of calories so they could go out and obtain more calories for their family. They did, though, have to do lots of intermittent fasting, which has been shown to produce effects similar to caloric restriction. And caloric restriction is not evolutionarily favored today because most of the benefits of caloric restriction do not become apparent until after reproductive age. Remember that evolution does not always produce the most efficient outcome (e.g., male peacocks need big bushy tails to attract mates, but this slows them down in other ways).

    I enjoy this discussion and I think it deserves more public attention.

  3. If a woman doesn’t get her period because of being underweight, that’s definitely not good for her health. So there’s a real limit to CR for women. I tried an 1800-2000 cal diet for awhile – because I get a lot of exercise, that’s a weight loss diet – and when I got down to a BMI of about 17, I lost my period, and gained weight back.
    The “optimal nutrition” part of CR is really important, one’s diet should be packed with vegetables and plenty of protein to avoid getting weak.
    If the CR diet “cracks” at some point and you start binging and gain weight back, it’s too much. It has to be sustainable and you can’t be uncomfortably hungry. If people get medical problems bc of being underweight, they might suddenly realize it’s too much and start binging. For me, losing my period was the trigger.

  4. ps I read that Behnke defined the amount of essential fat. For men, no less than 3%. For women, no less than 12%. So getting your body fat measured might be a good way of determining if you’re dangerously underweight.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: