Part V Fat: No More Fear, No More Contempt

2016 Edited to Add: When I first began writing here I generated a searchable database for references and I have since done away with that option. This entire series remains to be edited to include the full references within each piece. As time allows, this series will be edited to include complete references at the end of each part in the series.

Body Mass Index

Most of you who have followed my posts know that I reference the body mass index range often.

But what is it?

Part V: Body Mass Index (BMI) explained.

Part V: Body Mass Index (BMI) explained.

Well, in fact it's a pretty obscure measurement that was originally created by a Belgian astronomer in the 1830s. Adolphe Quetelet was interested in whether he could apply mathematical laws of probability to human beings. He measured the heights and weights of army conscripts and when he applied the calculation of mass over height squared to his results, it resulted in a bell-shaped curve. He determined that the middle point of the bell-shaped curve denoted "normal" or "average" and those on either side of that point were either under or over weight.

From 1830 to 1940 life insurance companies made use of Quetelet’s averages. Generally, the older someone is, the more likely they will die, but in the interests of trying maximize profits, Quetelet’s averages enabled insurance companies to avoid covering another group that might be liable to require a policy payout: those who were underweight relative to their height.

Age of the Microbe

In the 1830s if a young adult was underweight (relative to the average) chances were he was part of the undernourished working class and/or he was suffering an infectious disease: tuberculosis, cholera, etc. And those diseases would likely kill him in a time frame that would require a policy payout. Of course, he may have been part of a small number of naturally thin people, but an insurance company certainly had no interest in taking that risk.

But by the 1940s in North America and Europe, outbreaks of malaria, cholera and typhoid fever all but disappeared due to treatments, vaccination and improved hygiene and sanitation. Even tuberculosis had viable antibiotic treatment by the mid-1940s as well.

Not much was made of this until Metropolitan Life Insurance, after slightly over 100 years of using Quetelet’s averages, decided to chart the death rates of policyholders.

It appeared that heavier people died sooner than lighter people. From there MetLife created a table with some ranges of ideal body weight relative to height.

And since that time we have all assumed that heavier weights cause earlier deaths.

"As a result of the educational process [of the life insurance industry with its medical examiners], most physicians gradually accepted the concept that asymptomatic personal characteristics could increase the long-term risk of developing disease." [W. Rothstein, Public Health and the Risk Factor, Hushion House, 2003].

Life expectancy in the U.S. in 1940 was approximately age 60 for men and age 65 for women. MetLife took its predominantly eastern seaboard, European-origin, generally well-off policy-holding population of the U.S., and determined that policies that had to be paid out were for former customers who were heavier than average.

But did Louis Dublin, the chief statistician at MetLife responsible for determining that heavier people were more likely to require a policy payout, miss anything?

Probably. His entire population base had changed out from under him in the past 100 years. A key marker for determining a greater chance of death, namely being underweight, had all but disappeared from the data pool.

His population base was now skewed – almost all were well nourished and unlikely to die suddenly of many infectious diseases that had previously wiped out past policyholders. But we all have to die sometime, and so was the fact that policy payouts were going to heavier-than-average policyholders actually a correlation, a coincidence, or an ominous causative agent?

Critically the 1940s MetLife tables shifted average weight expectancy by age range (which had been used in the insurance industry up to that point), to publishing ideal weight levels for life (identified from the average weights found between the ages of 20-29) and separated the ranges by three different kinds of poorly defined body frames.

Dublin was convinced that heavier than MetLife-defined ‘ideal’ weights were an ominous causative agent for earlier death.

Nothing in the 70 years since has actually been able to prove Dublin’s theory. Being overweight does not cause early death. Being overweight does not even dependably correlate with early death either. In fact, being overweight is strongly correlated with less likelihood of death when compared to average mortality.

The human form simply has a range in which it can function well.

Too Tall

For very severe forms of dwarfism, life expectancy is approximately 10 years less than average. However for acromegaly the mortality rate is 2-3 times that of those of average height. Because gigantism is extremely rare there are not even dependable mortality data available. The difference between acromegaly and gigantism is that growth occurs after plate cartilage has fused for acromegaly but growth occurs in linear fashion from childhood in gigantism.

The human body is more capable of surviving being excessively under height rather than over height.

The shortest people in the world are approximately 23 inches (58 cm). The tallest people in the world have been recorded at approximately 106 inches (269 cm) [Guinness World Book of Records].

The average woman is 64 inches (162 cm) and the average man is 68 inches (173 cm). There is about 40 inches going out to the shortest and tallest ends on this bell curve of human height. Even though it is equidistant to get to one end of the bell curve or the other, the mortality rate is more severe at the tallest end of curve compared to the shortest end of curve.

Too Thin

Now back to weight. The lowest BMI recorded was 7.5 (she was only 21 inches tall and died of hypothermia), highest BMIs are around 188 [Guinness World Book of Records]. I am using BMI instead of weight because of course height factors into how much we can and do weigh. 

The average BMI for women is approximately 26.8 and for men, 26. The top point of the curve is not right in the middle of the range and we have a very long tail going out to the highest BMIs ever recorded.

The life expectancies of people who have BMIs of 188? 10-12 years less than average. However there are so few people at this range that the data are not statistically valid. 

The 2009 meta-analysis of 57 separate studies on BMI and mortality rates [Lancet Medical Journal] show lowest mortality rates for women and men between BMI 22.5 to 25. And while mortality increased quickly for those under BMI 22.5 (to about 7 years’ less than average life expectancy by BMI 17.5), life expectancy for BMI 35 was reduced to between 2-4 years. It was estimated that life expectancy beyond BMI 35 would be perhaps 8-10 years less than average. 

This particular meta-analysis was funded by: UK Medical Research Council, British Heart Foundation, Cancer Research UK, EU BIOMED programme, US National Institute on Aging, and Clinical Trial Service Unit (Oxford, UK).

Just at face value, there is a commensurate estimated loss in life expectancy if you drop 5 points in BMI levels from 22.5 or if you gain 10 points from BMI 25.

On top of that, the meta-analysis had to estimate life expectancy loss at and beyond BMI 35 and it is therefore likely an over-zealous estimate of mortality for that BMI and above.

A correlation means that the two items being studied appear together in a statistically relevant way so that we can say the two things are linked together. A causation means that there are actual data available to suggest that one thing actually causes another.

When I relay that there is a correlation between weight and increased mortality rates above approximately BMI 38, there is still no data to suggest that the weight causes the increase in mortality.

We are now entering the land of purposeful misrepresentation of data and as much as it is advertised that having excess adiposity (fat tissue) is life threatening, the data do not support it. There is however good data to indicate having too little adiposity is very life-threatening. 

Next: The smoke and mirrors of obesity