Have you heard the great news? According to the latest data from the Centers for Disease Control, obesity probably only kills about 112,000 Americans every year. John Tierney and David Brooks can hardly contain themselves.
[The new research has] taken away the favorite cudgel of the scolds who used the "obesity epidemic" as an excuse to attack the flabby. The supposedly deadly consequences of fat provided the scientific rationale for the last politically correct form of prejudice.
The fatophobes are fighting on, disputing the new study and arguing that it still shows the fatal dangers of being seriously obese. But they have lost the scientific high ground. Not only do people of "normal" weight die younger than the moderately overweight, the study shows, but thin people die even younger than those of normal weight.
Not coincidentally, the food and restaurant lobby is also helping itself to the latest CDC data. Last week the industry-sponsored Center for Consumer Freedom issued a press release berating the CDC for releasing "flawed" research earlier this year (namely, the study that estimated the excess deaths at 400,000 per annum).
The Center for Consumer Freedom is a parasitic form of think tank life. Instead of producing its own trumped-up studies like the anti-global warming groups, the CCF prefers to hijack other people's work.
In the press release, the Center for Consumer Freedom basically accuses the CDC and the authors of the earlier study of propagandizing the public, if not of outright scientific malpractice. The data were there in the Center's computers all along, they press release claims. Well, yes. Of course data sets used by Flegal and her colleagues have been around for awhile. They're the three National Health and Nutrition Examination Surveys (NHANES) compiled by the Department of Health and Human Services and used for countless studies over the years. But that doesn't mean that it should have occurred to the scientists who made the earlier estimates of obesity-related mortality to use the same statistical methods as the Flegal team. There is such a thing as scientific progress. We expect later research to supersede earlier work by capitalizing on its insights and overcoming its limitations. In their article Flegal and colleagues describe how they adapted a statistical technique designed to estimate breast cancer risks to the study of obesity. I get the impression that this approach is an original contribution and not a standard technique that their CDC colleagues should have anticipated.
Sorry, guys, the but size-16 print giveth, and the size-2 print taketh away...
Flegal KM, Graubard BI, Williamson D, et al. Excess deaths associated with underweight, overweight, and obesity. JAMA.2005;293:1861-1867.
These researchers argue that earlier studies may have overestimated the impact of obesity per se because their statistics didn't fully control for the subtly confounding effects age, gender, smoking, and drinking. Flegal and her colleagues tried to overcome the limitations of previous studies by using newer survey data and more sophisticated statistical methods.
The study itself looks pretty solid to me. It's the secondary spin that's insidious. Third parties are trying to give the public a false sense of security about the obesity crisis. Being overweight is good for you, the flacks insist. Well, not exactly... (Being overweight is probably healthier than yo-yo dieting, eating disorders, or extreme bariatric surgery, but this study doesn't bear on those important issues.)
Here's the first of several catches:
The new study attributes 111,909 deaths to obesity, but then subtracts the benefits of being modestly overweight, and arrives at the 25,814 figure. [CNN]
The Center for Consumer choice has arbitrarily seized upon the 25,814 figure because it's a good stick to beat the authors of the 400,000 estimate. But why subtract the number of lives "saved" by mere overweight from the number of deaths attributable to obesity? (Flegal compared recorded-deaths-in-2000 to a what the death rate would have been in an otherwise identical parallel universe where everyone was at a "normal" weight.) It's likely that the current standards for "normal" weight are too narrow. They may even be too low for the average American. In which case we should reckon excess deaths by comparing the real world to one in which everyone is either "normal" or "slightly overweight." Even so, it wouldn't follow that obesity (as currently defined) is less common or less dangerous than we thought.
Obesity apologists are equivocating about the difference between overweight and obesity. The effects of being overweight are uncertain, but there is no doubt that obesity increases the risk of death and ill-health. Furthermore, the number of excess deaths per year is a function of the prevalence of obesity, and no one disputes that the condition is becoming increasingly common.
But does being slightly overweight actually improve health? Or is this finding some kind of artifact? The intuitive explanation is that our current height-weight tables are wrong (either for the entire population or for certain sub-groups). Maybe it's healthier to have a little extra body fat, especially late in life. Besides, some heavier people may be at lower risk because they are leaner, fitter, and more muscular than many of their lighter counterparts.
Obesity may be a big problem without necessarily being responsible for as many excess deaths as we thought. The authors found that excess risk is highest for younger obese people. But most deaths occur after age 70. So, even obesity is associated withe a large relative increase in mortality for younger people, this increase won't necessarily translate into a large absolute number of excess deaths. Whereas if weight causes even small changes in the death rate of older people, these will produce large absolute numbers of excess deaths (or lives saved). But it's also worth noting that the preventable death of a younger person probably "costs" its victim more years of life. (As Amp suggests, no death is preventable. So-called "preventable" deaths are just the ones we can postpone. Many early obesity-related deaths are eminently postponable.)
Of the 111,909 estimated excess deaths associated with obesity (BMI>30), the majority, 84,145 excess deaths, occurred in individuals younger than 70 years. In contrast, of the 33,746 estimated excess deaths associated with underweight, the majority, 26, 666 excess deaths, occurred in individuals aged 70 years and older. (Flegal K, et al., 2005)
Maybe the latest results mean that doctors should encourage older people to bulk up--but not necessarily. A "normal" weight in a person over 70 may be a sign of a more serious health problem. After a certain age, people tend to get lighter for reasons that are more sinister than simple fat loss. Weight loss in the elderly is more likely to be associated with muscle wasting, bone loss, protein/calorie malnutrition, and a host of other serious health problems that directly increase a person's risk of dying, more or less irrespective of their body fat levels. It certainly doesn't follow that it's better to be overweight throughout one's life, rather than just during the critical years. (It should be emphasized that the current study doesn't claim to establish any cause and effect relationships. It is merely a sophisticated accounting of correlations between selected variables. Controlled trails and longitudinal studies will be needed to determine the precise causal relationship between weight and mortality.)
Brooks, Tierney, and the restauranteurs also overlook the fact that being overweight is itself a risk factor for future obesity. The more people who are overweight today, the more are likely to become obese in the future. So, even if overweight isn't associated with ill health, we should still be concerned about the potential for large numbers of overweight people to become obese. And this study suggests that if obesity rates increase, excess deaths will increase as well. The more fat someone has already, the smaller their "safety margin." People tend to gain weight with age until relatively late in life. Given that so many of the Baby Boomers are already overweight, many more will surely cross the line into obesity in the next few years. Besides, there is overwhelming evidence that excess fat actually causes even more weight gain through the vicious cycle known as metabolic syndrome.
The latest data don't take into account the fact that people are getting fat much earlier in life, nor that overweight and obesity are becoming more common every year.
Children and teenagers are the fastest-growing group of obese Americans. The effects of the childhood obesity boom are already manifesting themselves in historically unprecedented rates of type 2 diabetes and high cholesterol in childhood. The current study doesn't reflect the most recent trends. The study only includes people who were at least 25 when they participated in one of the three NHANES surveys (between 1971 and 1994), so it can't give us even a cross-sectional glimpse of the effects of early obesity.
Today's obese adults are more likely to have been overweight or obese since childhood than were obese people of previous generations. We don't know what the long-term effects of earlier onset will be, but we can only assume they will be worse. How could more years of high blood pressure, high cholesterol, and diabetes be a good thing?
Even if obesity "only" kills 100,000-odd people per year, that's still a lot of preventable death. Moreover, the researchers hypothesize that excess weight is less dangerous than it used to be primarily because of better treatments for high blood pressure, high cholesterol, and other illnesses caused or exacerbated by excess body fat. All things being equal, it would be better to prevent these diseases than to maintain people on expensive and potentially dangerous drugs for life. The study didn't even attempt to measure the detrimental effects of excess weight on general health or quality of life.
Tierney and others imply that if mere overweight is good for people, then our public health programs must be misguided. But if being overweight really extends people's lives, then we should redouble our public health efforts to stop millions of overweight Americans from drifting into obesity. For these people, even a small weight gain could have dire consequences. Maybe relatively modest weight loss will also turn out to be a huge benefit for people who are already obese. Even preventing further weight gain in obese people might save many lives. If so, perhaps obesity interventions are even more cost-effective than we thought. (Strategies to reduce the health impact of obesity should include efforts to fight size discrimination, which probably has a lot to do with the negative health outcomes associated with the condition.)
Pundits and interest groups are exploiting the Flegal study as an excuse to sneer at public health iniatives, malign the CDC, and to downplay the seriousness of the obesity epidemic.
Update: I notice that Brian Leiter is as big a Tierney fan as I am.
More interesting blog-based discussion of the Flegal from Alas, a Blog, Left Oblique, and Kevin Drum, and Jay Sennett.