On BMI and carbs

Posted 2013-08-07

Last week at OHM2013, Bert Hubert had an interesting talk: What you need to know about what you eat: health & weight. He started with a slightly simpler version of this graph that show how we all started getting a lot fatter since the mid-1980s:

Source: Prevalence of overweight, obesity and extreme obesity among adults: United States, trends 1960-62 through 2005-2006.

I have a big problem with these kinds of statistics. Basically what they do here is count how many people go from a BMI of 24.9 to 25.1 (normal weight to overweight) and 29.9 to 30.1 (overweight to obese). And obviously a BMI of 24.9 is perfectly healthy while 25.1 is problematic </sarcasm>.

Still, it's interesting and somewhat alarming to see how the obese line started going up around 30 years ago. But Bert's graph didn't include the overweight line, which curiously has been quite stable over the past 50 years.

Another issue with these statistics is that the BMI itself is a very flawed measure. The idea is that if you're twice as tall, you get to weigh four times as much: BMI = kg/m². But we live in a three-dimensional world. If you make an object twice as big, its volume gets eight times larger. So all else being equal, it should be eight times heavier: kg/m³. However, that's not exactly how things work with humans. Using a power of 2.3 to 2.7 would be more appropriate, and not "punish" tall people like the current BMI does. As the population grows taller, this could be an issue.

But after crunching some numbers collected by the US National Center for Health Statistics, it turns out that the effect of people getting taller isn't very significant. According to the 1960-1962 numbers, males in the US age 20-74 had a mean BMI of 25.1 and females 24.8. Nearly 40 years later in the 1999-2002 figures this was 28 and 28.3. The mean height of men 20-74 had increased from 173.4 to 176.2 cm, women from 160.2 to 162.5 cm. (Curiously, younger people are actually smaller than they used to be; in the 1999-2002 figures the age group 40-49 is the tallest.)

So the mean BMI for men increased by 11.2% and for women by 14%. 1.2% of that increase in BMI can be explained by the increase in height and goes away if we use kg/m^2.7. A 10 - 13 % increase in height-adjusted BMI is still very significant, but not quite a "graph of doom".

So now that we have a better idea of the size of the problem, what's the cause?

Bert argues that we've been given some very bad advice the past 30 years, and that eating less fat (like meat) and more carbohydrates (bread, pasta, potatoes, rice) has made us all fatter. The reason to reduce our fat intake was to avoid cardiovascular disease, a major cause of death. But it turns out that many fats aren't really bad for you (saturated animal fat) or even good for you (in oily fish, olive oil). You may still want to avoid processed vegetable oils (margarine), especially trans fats, though. But an overdose of carbohydrates is quite problematic.

However, it seems to me that the pendulum is now swinging too far in the opposite direction. Yes, eating too much sugar or carbs in general makes you fat and will likely give you diabetes. But eating too much of anything makes you fat and increases your risk of various diseases.

Grains and potatoes have been a staple of our diets for thousands of years, and we managed to get a lot of things done eating them. The fact that people got smaller and didn't live as long when we switched from hunting/gathering to agriculture is fairly meaningless: when something new comes around, it's almost always inferior to the old thing that it'll be replacing soon—at first.

So for thousands of years we managed to have a healthy weight eating lots of grains and potatoes. This suggests to me that carbohydrates can be part of a healthy, balanced diet. However, there are many people who swear by a low-carb diet. Good for them. It's just that what works for them isn't necessarily the best thing for everyone.

If anyone is interested in my advice, here it is. Don't drink your calories (or joules). Stick to water if you can, but if you need something sweet to drink, find something using an artificial sweetener that you like. Eat bread, pasta, rice, potatoes in moderation, and combine them with foods that have proteins and fats. Minimize candy and snack food. These are optimized to taste good and be cheap to make, so they have lots of sugar, flour and processed vegetable oils that you don't need, and it's sooo easy to overeat this stuff. Eat fruit instead and don't worry about the sugar it contains. Even Bert says 120 grams of carbohydrates a day is ok. That's 10 apples.