Monday, June 28, 2010

Hitting the gym harder for a decade won't do a thing for your weight.

I would have had the headline read, "Exercising exercise's confirmation bias" but figured that wouldn't be as grabby.

From the only publishable because the world has such a huge crush on exercise impacting on weight file comes the, Effect of change in physical activity on body fatness over a 10-y period in the Doetinchem Cohort Study published ahead of print in the American Journal of Clinical Nutrition.

The study is just one of many in a long string of studies that fail to show any dramatic benefit on weight from long term exercise.

This one?

The authors followed the 4,944 adult participants of something called the Doetinchem Study and they tracked weight and waist circumference change as a function of physical activity - both at baseline and if physical activity increased or decreased over a 10 year period.

Amazingly this thing got published. I say "amazingly" not because it was a particularly bad study and not because the results weren't particularly impressive, but rather because the conclusions drawn by the authors stand in stark contrast with their results.

Once again with this paper the authors themselves did all the heavy lifting on why exercise isn't the be-all (or perhaps even the be-any) of weight. Here are their comments:

"Random mixed-effects models showed that a single measurement of physical activity was not clearly related to change in body weight and WC over a 5-y period."
Translation? In their analysis of the data, how much you exercised at the start of this study didn't impact on your likelihood of overweight or obesity, or your waist size 5 years down the line.
"Analyses of repeated measures showed that compared with those who maintained their activity level, those who increased their physical activity over a 5-y period had less gain in WC and possibly in body weight. Most importantly, these effects were sustained (although not statistically significant) in the consecutive 5 y for WC and for body weight."
Translation? People who upped their exercise from baseline had their waist circumferences grow less (though they still grew) than those who didn't, but they weren't significantly lighter. Also, while this lesser gain in waist circumference was found to be significant over the course of the first 5 year period study, it wasn't found to be significant over 10.

Statistical significance or insignificance aside, what type of spectacular results are we talking about? The folks who reported a marked increase in exercise over a decade found themselves a whole 1.2lbs lighter ten years later than the folks who didn't and had waist circumferences half a centimetre (roughly a fifth of an inch) smaller.


This of course leads me to conclude that this study is in fact consistent with the bulk of the evidence which suggests that in the absence of dietary interventions exercise does not dramatically impact on weight over time. It also leads me once again to beg researchers to stop focusing on weight as an exercise study's primary endpoint and instead focus on those things more likely to demonstrate the incredible benefits of exercise - things such as high blood pressure, cholesterol levels, arthritis, cardiovascular fitness and overall quality of life. Oh, and also beg authors that when using weight as an exercise study's endpoint, to draw conclusions that don't fuel the fire of the fallacy of exercise being a primary driver of weight just because that's what people want to hear.

What did it lead these authors to conclude?
"An increase in physical activity was associated with a statistically significant lower gain in body weight and in WC, which was maintained during the following 5 years. These findings support the need for public health programs that promote physical activity."
Yeah, that's what these statistically insignificant and basically inconsequential results support. We should fund public health programs designed to help with weight by promoting increases in physical activity so that in a decade people will gain 1.2 fewer pounds and have 0.5cm smaller waists. That sounds like a great way to spend limited resources.

Good grief.

Again I've got to ask, what the frak is up with the peer reviewers for the American Journal of Clinical Nutrition?

May, A., Bueno-de-Mesquita, H., Boshuizen, H., Spijkerman, A., Peeters, P., & Verschuren, W. (2010). Effect of change in physical activity on body fatness over a 10-y period in the Doetinchem Cohort Study American Journal of Clinical Nutrition DOI: 10.3945/ajcn.2010.29404