Big Fat Facts Big Fat Index

Now they're saying that average BMIs lead to early death

There are two fat and health items in the news right now.

The first - the request from lap band manufacturers to the FDA to lower the BMI where lapbands are allowed - was covered by AndyJo in the previous post.

Another study, published by New England Journal of Medicine with Amy Berrington de Gonzalez as the first author, is being reported on in articles such as ABC New's "Higher Body Mass Index Linked to Greater Mortality Risk" and touted as "proof" that people in the 25-30 BMI ("overweight") and the 30-35 BMI (obesity type one) categories are at increased risk of death because of their weight. This study is a data dredge, that is, it's built on data from older studies that could have been cherry-picked to get the desired results. It's obviously meant to refute Katherine Flegal's 2005 study and a recent Canadian government study, both based on population-level mortality statistics and both of which indicate that people in the the 25-35 BMI range have an average or above average life expectancy.

The flaws in the new de Gonzalez study? Well, I haven't read it myself, but I'm hearing from medical professionals who have read it that it doesn't control for level of physical activity - or health insurance status (!!). It's interesting that it was published shortly before this request to expand the market for lapbands. Lucky for the lap band manufacturers, eh?

There's a great post on both studies on Suethsayings: Push for weight loss surgery even if you have a lower BMI follows study about obesity.

Health and science reporters? If any of you are reading this, please look at the quality of the research before suggesting that a data dredge invalidates large, comprehensive and well designed studies like "BMI and mortality: results from a national longitudinal study of Canadian adults" (Orpana et al, 2009) and "Excess Deaths Associated With Underweight, Overweight, and Obesity." (Flegal et al, 2005)

There is nothing they will not do to make a buck... | FDA Panel recommends lowering BMI requirements for lap band

D-Man's picture
December 4th, 2010 | Link | This just in: Living found

This just in: Living found to increase mortality risk!

I have to say, I haven't seen as many of these hairbrained pointless fattylulz studies as I used to. I sincerely hope that's an indication that they're dying out.


Meowzer December 4th, 2010 | Link | So where's their study

So where's their study comparing the health and longevity prospects of fat people who lost 50+ pounds, became "normal" BMI, and kept it off permanently, versus fat people who stayed fat?


I find it very, very interesting that nobody has that data set anywhere. There must not be enough former fat people for them to study.

Tobysgirl December 5th, 2010 | Link | I am so glad I have read all

I am so glad I have read all the stuff about WLS on this site. I have an abdominal mess and the third surgeon I saw, who was actually experienced at the surgery I need, was in private practice with bariatric surgeons and pushed lapband surgery. Which I will not have.

Interestingly, the fourth, and most recent, surgeon I saw also does bariatric surgery, but she works for a teaching hospital and has NO NEED to push surgery of any kind, as it does not increase her income. I do need to get smaller for the surgery I need, but I refuse to diet, talk to a nutritionist (the folks who know less about food than MDs, and that's saying something), worry about my weight, etc. Just keep trying to feel better and be the active woman I used to be (have always been fat as an adult and been very physically active), not getting into that deprivation mentality.

And for health professionals, if you would bother to learn something about the body and nutrition, you would be able to help your patients lower their cholesterol, blood pressure, and blood sugar, with NO SURGERY. How many doctors or nurses are telling people that if they eat carbs they must eat some protein with it, in order for the liver to secrete the enzyme to tell the pancreas to stop pumping insulin? Simple, huh? And unknown by these idiots. Instead, millions of women eat popcorn (very high-glycemic) with no butter for a snack and think they're being super-healthy.

BigLiberty's picture
December 5th, 2010 | Link | Here's the link to the

Here's the link to the article proper (which lives behind a paywall);

One thing I notice right away --- the data neatly concludes that the BMI categories that currently exist are EXACTLY correct (up to overweight, underweight is written off largely): 20 - 24.9 is the least mortality.

That seems rather unlikely, frankly.

Also noting lots of NIH grants for this, grants from Cancer societies. The author's other work is largely based on showing the link between BMI and pancreatic cancer (from what I can tell).

And as other people have noticed already, this study came out in a 'press release' fashion all over the major media directly before the FDA was petitioned to lower the thresholds for the lapband. So I smell blood in the water on this.

And a giant SECOND to everything Meowzer said.

EDIT: As a note, while they took smoking into account to jack the numbers in their favor (I mean, uh, make the results more rigorous), they didn't take into effect diseases that have been shown to *cause* weight gain like PCOS and insulin resistance (though there's a chicken-and-egg debate concerning the latter). Also, they didn't take into effect early mortality linked to conditions like depression, the meds of which often cause significant enough weight gain that one might start 'normal' and end up 'obese.'

So until they do all that -- which is unlikely, since the going thing in obesity research is to screw up causation and say fat *causes* depression, PCOS, diabetes --- then I'm not going to take their results seriously.

pani113's picture
December 5th, 2010 | Link | Interesting that at the very

Interesting that at the very bottom of the NEJM abstract, one researcher disclosed ties to Inovate Health Services. If one googles them, one finds out they market nutritional supplements. They also were in hot water for false claims about what their products (including weight loss products) can do. Of course, disclosure is voluntary so we don't know what other ties the researches have!

"Fat can be beautiful. Intolerance is ALWAYS ugly!"

DeeLeigh's picture
December 6th, 2010 | Link | That is a really excellent

That is a really excellent point, and one that we should keep bringing up. Many of the tools that people employ in their weight loss attempts are harmful.

BigLiberty's picture
December 6th, 2010 | Link | Debra, Yeah, paywalls


Yeah, paywalls outsmart me every time Sad But I figured people might still want a look at the stated abstract and conclusion, and the author's names and so on. My Google Scholar -fu wasn't at its best.

It's interesting you mention Metabolife - I took a whole lot of that back when it had ephedrine. They were my dad's diet pills; he said I should take some, too (I was 15) and in four months I'd lost 80 lbs. Ten years later I have orthostatic hypotension and arrhythmia. Aaand I'm a lot fatter. So there's that.

Mentioning death due to diet-related behavior (not to mention that those people are also likely fatter due to the diet-related behavior as well) is very important, thanks for making that addition.

There are so many ways statistics are used to massage data to conclude anything the authors really want it to conclude that we should all be extremely wary of data dredges. Even when the data is coming from a reliable source, by making adjustments and using certain regression models they can skew results fairly widely. And the case of BMI, when you're only dealing in small percentage increases (13 - 44% increase in early mortality risk based on BMI as opposed to thousands of percent for smokers, for instance) it's pretty easy to find ways to shift that J-curve to where you want it to go.

This of course means that we need to resist trumpeting data dredges that are in our favor, as well. Best to most loudly promote the more rigorous, double-blind studies; it's much harder to poke holes in their results.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

© 2000-2020 Big Fat Blog and its authors, all rights reserved. Big Fat Blog, Big Fat Facts, and Big Fat Index are our trademarks.