Shots - Health News
Don't Count On Extra Weight To Help You In Old Age
Originally published on Tue February 19, 2013 8:47 am
Wouldn't it be great, considering how many of us are overweight, if carrying a few extra pounds meant we'd live longer?
A recent analysis of nearly 100 published studies involving almost 3 million people found, surprisingly, that being a little overweight was associated with a lower risk of death than having a normal weight or being obese.
The sweet spot, as it were, appears to be a body mass index ranging from 25 to less than 30. The findings were published in early January in JAMA, the Journal of the American Medical Association.
But is the finding real? An editorial in the same journal pointed to problems with BMI as tool for assessing obesity. And it suggested that artifacts in the data might be another factor behind the results.
The lowest death rates in most studies have been seen for people with a BMI between 22 and 25, the editorial point out. The authors suggested that the most important findings from the analysis were that death rates were higher for the obese (BMI of 35 or more) and people who were quite underweight (BMI less than 18.5).
As luck would have it, Ryan Masters, a demographer at Columbia University's Mailman School of Public Health, has been looking into some of the same data that have been cited in support of the so-called obesity paradox.
"I sort of gingerly came into this field and was blown away by the debate surrounding it," he tells Shots. Some of the theories advanced to explain the paradox include a beneficial metabolic effect from modest fat reserves for the elderly and cushioning in case of falls, he says.
Masters decided to look at the data, and he found problems in plain sight.
Some of the studies excluded people who lived in institutions, like nursing homes, skewing the results toward healthier people. Frail people would be less likely to participate in surveys and studies, too, he says.
He also found a problem that he said reminded him of a report about falling cats in Manhattan. Researchers found, paradoxically, that cats falling from windows on the highest floors of apartment buildings were more likely to survive than those who stumbled out of window on middle floors, say the fifth or sixth floors.
The theory (as explained in an episode of Radiolab) is that the cats falling from high floors reached terminal velocity and had time to turn themselves and land feet first — something the cats falling from middle floors wouldn't have time to do.
But Masters says there's a better explanation. A so-called selection effect that is biasing the data, he says. All the reports about the falling cats came from local vets who treated them. The cats that fell from the high floors were far more likely to die on impact and never be taken to a vet, Masters says, but a few survived. Those that fell from middle floors and were in bad shape had a chance at survival. The vets saw more of these cats, but many more died.
In the surveys of elderly people, the overweight and obese are outliers like the cats that survived a fall from a high floor. "They're probably not representative of what's normal," he says. "They're probably representative of the best."
"The population we're most interest in studying — the older aged obese population — is simply not there," he says.
An analysis Masters and colleagues did challenges the paradox head-on. "The risk of mortality from obesity compounds and grows stronger as you age," he says.
Indeed, their paper, published online by the American Journal of Epidemiology this month, calls for a rejection of the paradox and a clear-eyed assessment of the hazards the obesity epidemic poses to our aging population.
"[O]ur results caution against framing the obesity epidemic as a 'moral panic' in which the harm administered stems more from ill-advised hyperbole than from the real-life consequences of obesity," Masters and his colleagues write. "In light of our findings, we are, on the contrary, much more concerned about inappropriate denial of the epidemic's consequences for US mortality. Our results suggest that major public health consequences will track with the epidemic and that efforts to stem its growth are probably well worth the investment."