The battle for the hearts and thighs of Americans reached a fever pitch last year among obesity researchers: How fat is too fat as a matter of health? Peace may be at hand.
On New Year’s Day last year, one of the biggest celebrity catfights of all time broke out—if you follow high-profile epidemiologists, that is. The Journal of the American Medical Association published a Centers for Disease Control (CDC) meta-analysis of 97 studies, suggesting that people who are overweight by government BMI standards (though not obese) live longest.
That’s right, fat people live longer. The results were counterintuitive but not really “new news” in the world of nutrition; a meta-analysis is, by definition, a statistical summary of already-published research.
The paper’s tone was matter-of-fact, the verbiage dry. The authors didn’t call people who toiled to stay trim “suckers” nor suggest that we storm Cinnabon to pack on a few. But in response, scientists at Harvard’s School of Public Health went, to use the scientific term, cuckoo for Cocoa Puffs. As the intellectual leaders in the war on obesity, they railed against the study at academic meetings, organized a symposium to denounce it (at which they reportedly considered placing a Clint Eastwood–style empty chair on the dais to represent the study’s lead author, Katherine Flegal, who’d declined to join them), and by one account a pro-Harvard propagandist passed out anti-Flegal pamphlets at a lecture she gave—a tactic usually reserved for ugly political campaigns.
Heading the negative-ops team was Walter Willett, MD, chair of Harvard’s nutrition department, who presides over the famed Nurses’ Health Studies and is considered the grandfather of modern nutrition epidemiology. “This study is really a pile of rubbish,” he told National Public Radio, “and no one should waste time reading it.”
In person, Willett is charming. With a bushy white mustache and twinkling blue eyes, and wearing a gray-blue Mr. Rogers–style cardigan the day I meet him at his Harvard School of Public Health office, a year after his battle royale with Flegal, he hardly seems the type to eviscerate his foes. I’m enjoying his company because he coauthored a book coming out this month, Thinfluence.
But his affable mien disguises sharp elbows. Even after having a year to cool off, he stands by his “rubbish” remark. “I think it is complete rubbish,” he confirms, somewhat gleefully. He goes on to characterize a recent meta-analysis published in the Annals of Internal Medicine that found no link between higher consumption of saturated fat and heart disease as “a disaster on the order of Flegal.” He’d already written a critical commentary for the Annals, but as he discussed the flaws he saw in the work, he mused, “I think I’m going to write another [one].”
There’s always a flood of nitpickers when a high-visibility journal runs a juicy study from a bigwig scientist like Flegal, who was among the CDC epidemiologists who “discovered” the obesity epidemic in 1994. After all, one of the points of publishing studies is to open up findings to competing ideas and data. And the Harvard group’s critiques of Flegal’s work weren’t beyond the pale: They didn’t feel she’d corrected adequately for the effects on the data of smoking, illness, and age, for instance.
Harvard’s own analyses, using the Nurses’ Health and Health Professionals Follow-Up Studies, had repeatedly demonstrated that the thin—and in some cases, the very thin—enjoy the greatest longevity. Flegal’s meta-analysis (which, by the way, included the Harvard studies) showed, by contrast, that people with a BMI between 25 and 30, considered overweight by government standards, enjoy the lowest mortality rate; for a 5'4" woman, for instance, that’s between 146 and 174 pounds. An earlier Willett-coauthored meta-analysis of 19 studies had identified the optimal BMI as between 20 and 25, or 117 to 145 pounds for a 5'4" woman; and in a 1995 study based on just the Nurses’ Health data, the Harvard team found the greatest longevity at a BMI below 19, which would put our 5'4" woman under 110 pounds—borderline anorexic.
Underlying this altercation was a disagreement about the best use of data. Willett starts with a socioeconomically homogenous subject pool and then further cleanses it to eliminate those who might confound the results. The goal is to isolate weight and mortality as much as possible from other factors associated with death (low income or education, smoking, illness, etc.). The CDC crew, on the other hand, works from a more inclusive data set and then employs statistical techniques to correct for confounders. This, they believe, yields the most widely applicable results.
That’s right, fat people live longer. The results were counterintuitive but not really “new news” in the world of nutrition; a meta-analysis is, by definition, a statistical summary of already-published research.
The paper’s tone was matter-of-fact, the verbiage dry. The authors didn’t call people who toiled to stay trim “suckers” nor suggest that we storm Cinnabon to pack on a few. But in response, scientists at Harvard’s School of Public Health went, to use the scientific term, cuckoo for Cocoa Puffs. As the intellectual leaders in the war on obesity, they railed against the study at academic meetings, organized a symposium to denounce it (at which they reportedly considered placing a Clint Eastwood–style empty chair on the dais to represent the study’s lead author, Katherine Flegal, who’d declined to join them), and by one account a pro-Harvard propagandist passed out anti-Flegal pamphlets at a lecture she gave—a tactic usually reserved for ugly political campaigns.
Heading the negative-ops team was Walter Willett, MD, chair of Harvard’s nutrition department, who presides over the famed Nurses’ Health Studies and is considered the grandfather of modern nutrition epidemiology. “This study is really a pile of rubbish,” he told National Public Radio, “and no one should waste time reading it.”
In person, Willett is charming. With a bushy white mustache and twinkling blue eyes, and wearing a gray-blue Mr. Rogers–style cardigan the day I meet him at his Harvard School of Public Health office, a year after his battle royale with Flegal, he hardly seems the type to eviscerate his foes. I’m enjoying his company because he coauthored a book coming out this month, Thinfluence.
But his affable mien disguises sharp elbows. Even after having a year to cool off, he stands by his “rubbish” remark. “I think it is complete rubbish,” he confirms, somewhat gleefully. He goes on to characterize a recent meta-analysis published in the Annals of Internal Medicine that found no link between higher consumption of saturated fat and heart disease as “a disaster on the order of Flegal.” He’d already written a critical commentary for the Annals, but as he discussed the flaws he saw in the work, he mused, “I think I’m going to write another [one].”
There’s always a flood of nitpickers when a high-visibility journal runs a juicy study from a bigwig scientist like Flegal, who was among the CDC epidemiologists who “discovered” the obesity epidemic in 1994. After all, one of the points of publishing studies is to open up findings to competing ideas and data. And the Harvard group’s critiques of Flegal’s work weren’t beyond the pale: They didn’t feel she’d corrected adequately for the effects on the data of smoking, illness, and age, for instance.
Harvard’s own analyses, using the Nurses’ Health and Health Professionals Follow-Up Studies, had repeatedly demonstrated that the thin—and in some cases, the very thin—enjoy the greatest longevity. Flegal’s meta-analysis (which, by the way, included the Harvard studies) showed, by contrast, that people with a BMI between 25 and 30, considered overweight by government standards, enjoy the lowest mortality rate; for a 5'4" woman, for instance, that’s between 146 and 174 pounds. An earlier Willett-coauthored meta-analysis of 19 studies had identified the optimal BMI as between 20 and 25, or 117 to 145 pounds for a 5'4" woman; and in a 1995 study based on just the Nurses’ Health data, the Harvard team found the greatest longevity at a BMI below 19, which would put our 5'4" woman under 110 pounds—borderline anorexic.
Underlying this altercation was a disagreement about the best use of data. Willett starts with a socioeconomically homogenous subject pool and then further cleanses it to eliminate those who might confound the results. The goal is to isolate weight and mortality as much as possible from other factors associated with death (low income or education, smoking, illness, etc.). The CDC crew, on the other hand, works from a more inclusive data set and then employs statistical techniques to correct for confounders. This, they believe, yields the most widely applicable results.
No comments:
Post a Comment