Deficiencies in BMI as a Benchmark

[Guest guest]

Hi Warren:

A personal observation regarding 'BMI'. I am afraid it is going to

be a long post:

BMI is essentially an approximation of the old Metropolitan Life

weight tables, but in a much more convenient form than a big table of

data. With BMI, a single number can be promoted as the appropriate

one for people of all heights, rather than having to have a different

number for all possible gradations of height from the shortest to the

tallest.

But I found the Metropolitan Life weight tables were really pretty

much useless for a couple of reasons. One problem was the huge range

of weight considered appropriate for an individual of any given

height. It was so wide in fact that one person could look like a

skeleton at the lower end of his range, while another would look

obese at the other end of his. In my case, as an example, I started

to look 'gaunt' a few pounds below the TOP of the range. At the

bottom of the range I doubt if I would have still been alive.

My understanding of this issue improved when around 1980 I read one

of Pritikin's books (I will quote from it below). It had a method

for calculating appropriate weight for which the input was not just

height and gender, as it is for both the Met Life tables and BMI, BUT

IN ADDITION WRIST CIRCUMFERENCE (WC). The idea was that WC is

pretty well correlated with the mass of your bone structure. And

also that when people gain weight they experience relatively minor

changes in WC, so it remained a relatively good measure even for

people who were considerably overweight. When I calculated my 'ideal

weight' using Pritikin's method I came out with numbers that made

sense. When I had negligible excess fat I weighed pretty close to

what the Pritikin formula suggested, because of the adjustment for WC.

[Of course all these (conventional wisdom) calculations of ideal

weight were determined before CR was as widely understood as it is

today. So I now realize, of course, that my real ideal weight is a

lot lower than I had believed before reading Walford.]

Nevertheless, that said, the criticism that BMI does not in any way

reflect varying skeletal mass among individuals remains valid. And

in my opinion it represents a serious drawback to its use as a

benchmark. I believe that an assessment of a person's ideal BMI

needs to take into account skeletal mass. The appropriate BMI of

someone with a substantial bone structure will be significantly

higher than that of a person with much smaller skeletal mass.

Here is the way Pritikin organized the data: (I am looking at pages

52 and 53 in my, now much dog-eared, paperback copy of THE PRITIKIN

PERMANENT WEIGHT LOSS MANUAL). Where Metropolitan Life had one table

for all possible heights for males and one for females, Pritikin had

three tables for males and three for females. Of the three tables

one was for those with 'small frame', a second for 'medium frame' and

a third for 'large frame'.

Here is an example quoted from the tables in front of me: For a male

5' 10 " tall with a small frame the ideal weight is 144 to 154 lbs;

for medium frame it is 150 to 165 lbs; and for a large frame 159 to

179 lbs. These ranges are much smaller than those in the

Metropolitan Life ideal weight tables. But note how big the

difference is from the bottom of the small frame range to the top of

the large frame range - 35 lbs. The high number is 24% greater than

the low number - all because of a difference in frame size, OF WHICH

BMI TAKES NO ACCOUNT WHATEVER. Even from the middle of the small

frame range to the middle of the large frame the difference is 20 lbs.

There is a consequence to be derived from this, in my opinion. Do

not assume that one BMI fits all. It clearly doesn't. (And we have

not even gotten into the issue of large framed MUSCULAR people!) It

looks like it would be advisable, when deciding what BMI to aim for,

to take account of your frame size. If you have a large frame, as

defined below, aim for the higher end of the range of BMIs. If you

have a small frame perhaps you should aim for the lower end to get

the same effect as the large-framed person would achieve at the high

end. The optimal BMI for a person with especially large frame size

might be 25% higher than that of a notably small-framed person. In

other words if a small-framed person's ideal BMI is 18, then for a

large-framed person it might be 22 1/2. This is an appreciable

difference.

Here are Pritikin's frame size criteria for males: Small = < 6 1/4 " ;

medium = 6 1/4 " - 7 " ; large = > 7 " .

For females: Small = < 5 1/4 " ; medium = 5 1/4 " - 6 " ; large = > 6 " .

This is far from a perfect science. But we are learning more every

day. This, I guess, is just another wrinkle. But perhaps an

important one.

(Thanks for reading this far!)

Rodney.

> Please do CR seriously. Please put your Body Mass Index (BMI) in

the

> conservatively safe range of 18.0 to 20.0 -- which is medically

sound

> and non-controversial. It is not too late, if you are still alive.

[Guest guest]

Perhaps this is an argument in favour of 'body fat percentage' as a

better measure of ideal weight?

Rodney.

--- In , " Rodney " <perspect1111@y...>

wrote:

> There is a consequence to be derived from this, in my opinion. Do

> not assume that one BMI fits all. It clearly doesn't. (And we

have

> not even gotten into the issue of large framed MUSCULAR people!)

It

> looks like it would be advisable, when deciding what BMI to aim

for,

> to take account of your frame size. If you have a large frame, as

> defined below, aim for the higher end of the range of BMIs. If you

> have a small frame perhaps you should aim for the lower end to get

> the same effect as the large-framed person would achieve at the

high

> end. The optimal BMI for a person with especially large frame size

> might be 25% higher than that of a notably small-framed person. In

> other words if a small-framed person's ideal BMI is 18, then for a

> large-framed person it might be 22 1/2. This is an appreciable

> difference.

[Guest guest]

Regarding the measurement of wrist circumference. You can come up

with varying numbers. If you pull really tight whatever it is you

are using to measure it, obviously the number will come in smaller.

Pritikin didn't specify how to take the measurement. But the best

way seems to be to find the narrowest point on your wrist, pull the

tape tight and then release it to the point where no wrinkles are

evident from the tightness of the tape.

My wrist measurement in the past few years has dropped by about one

quarter of an inch, from 7 1/2 " to 7 1/4 " . So the numbers are not

necessarily static. Exercise, or the lack of it, is one factor

affecting bone mass.

Rodney.

> > There is a consequence to be derived from this, in my opinion.

Do

> > not assume that one BMI fits all. It clearly doesn't. (And we

> have

> > not even gotten into the issue of large framed MUSCULAR people!)

> It

> > looks like it would be advisable, when deciding what BMI to aim

> for,

> > to take account of your frame size. If you have a large frame,

as

> > defined below, aim for the higher end of the range of BMIs. If

you

> > have a small frame perhaps you should aim for the lower end to

get

> > the same effect as the large-framed person would achieve at the

> high

> > end. The optimal BMI for a person with especially large frame

size

> > might be 25% higher than that of a notably small-framed person.

In

> > other words if a small-framed person's ideal BMI is 18, then for

a

> > large-framed person it might be 22 1/2. This is an appreciable

> > difference.

[Guest guest]

Hopefully my last contribution on this topic:

I just calculated that for me to drop to a BMI of 20 I would have to

lose 28 lbs from my 'gaunt' weight, mentioned earlier.

To drop to a BMI of 18 I would have to lose 43 lbs from that same

level.

I have a tough time persuading myself that a drop of anywhere near

that much would be healthy.

When my body fat drops to 8% I will no doubt be relating here what my

BMI turns out to be. But that will not be for a year or two based on

current plans. And I doubt I will be letting my body fat get below

8%.

Rodney.

> >

> > > There is a consequence to be derived from this, in my opinion.

> Do

> > > not assume that one BMI fits all. In

> > > other words if a small-framed person's ideal BMI is 18, then

for

> a

> > > large-framed person it might be 22 1/2. This is an appreciable

> > > difference.

[Guest guest]

I share that concern, until I see more studies that say HOW, at our age, to do a healthy weight drop without heart problems.

Regards.

----- Original Message -----

From: Rodney

Sent: Saturday, November 29, 2003 6:41 PM

Subject: [ ] Re: Deficiencies in BMI as a Benchmark

Hopefully my last contribution on this topic:I just calculated that for me to drop to a BMI of 20 I would have to lose 28 lbs from my 'gaunt' weight, mentioned earlier.To drop to a BMI of 18 I would have to lose 43 lbs from that same level.I have a tough time persuading myself that a drop of anywhere near that much would be healthy. When my body fat drops to 8% I will no doubt be relating here what my BMI turns out to be. But that will not be for a year or two based on current plans. And I doubt I will be letting my body fat get below 8%.Rodney.