Does fatness kill?

For years, we were told being fat makes us die prematurely. This “fact” suffered a seemingly fatal blow last year when Katherine Flegal of the Centers for Disease Control published the first careful analysis of what the data on obesity really showed.

Ms. Flegal’s study found there were more premature deaths among the underweight and those of normal weight than among the overweight. Overweight Americans were actually most likely to live the longest. This result is no statistical fluke, as controlling for such things as smoking status or weight loss due to illness did not change the finding.

These results directly challenged the public health establishment’s cherished belief that fatness is life-threatening. In response, a Harvard conference sought to “explain” Ms. Flegal’s findings. The conference asserted that, whatever Ms. Flegal’s results, being overweight is still deadly.

But the Flegal data is not so easily dismissed. Many studies also contradict the claim that being overweight leads to a shorter life. These studies show the overweight face less risk for premature death than those of ideal weights.

Now, we have another well-publicized study, this time from the National Cancer Institute, which warns Americans, particularly Baby Boomers, that their waistlines can kill them. According to The Washington Post’s alarmist Page One headline, “Just a few extra pounds could mean fewer years.” A closer look at both the study and the data suggest the Post’s headline is wildly misleading. Neither the Boomers nor anyone else has much to worry about from being overweight.

First, there are enormous problems with the design of the National Cancer Institute study, itself, problems that make any of its conclusions highly suspect. For example, study subjects were drawn from only six states and the 18 percent of AARP members who bothered to fill in questionnaires.

Hence, there is no reason to assume the study is representative of the American population. If the study’s sample is unrepresentative, then its shaky results can hardly be generalized to the entire population.

Second, the study failed to control for many factors that affect mortality. The study specifically claims excess weight accounted for almost 8 percent of all premature deaths among men, an astonishing claim since it has no information on what killed the participants; the database does not include cause of death.

None of the key facts about any of the participants was validated, including, most importantly, their body weight and height. All such facts were self-reported. Most surprisingly, the study relied on “participants recalled weight at the age of 50 years,” a highly dubious measure given most of us cannot recall what we had for lunch last week, let alone what we weighed at age 50.

In short, this study makes headline-generating claims about the supposed risks of being overweight but fails the most basic of scientific tests: ensuring that its data and measurements are, in fact, reliable.

It is not simply the study’s design that undermines its credibility. It is also its data, which fail to support its central claim that “even moderate elevations in BMI [Body Mass Index, the conventional yardstick for determining if one is underweight, normal weight, overweight, or obese] conferred an increased risk of death.”

For example, among the men in the study, there was no statistically significant increase in relative risk for death for overweight men. When the men were broken into five-year age cohorts, there was no statistically significant increase in relative risk for death for overweight men except for those aged 61 to 65, with BMIs of 28 to 29.9 (a BMI between 25 and 29.9 is considered overweight). Similarly, with overweight women, there was no statistically significant increase in relative risk for death in any five-year age cohort.

Authors failed to note their study confirms what other researchers have found, that is, the highest risk for premature death comes not from being too fat but from being too thin. In fact, normal weight men had a greater risk of premature death than grossly obese men, a finding that completely contradicts the study’s claims about the risks of being overweight.

Astonishingly, the National Cancer Institute researchers began with the conclusion that “excess body fat has long been recognized as a harbinger of disease and early death,” for which they provide not a single bit of scientific evidence. They proceeded to shoehorn the data to fit their conclusion. Had The Washington Post analyzed the study more carefully, it might have found how wildly its headline diverged from the real story.

Patrick Basham is director and John Luik is a senior fellow of the Democracy Institute. This fall, London’s Social Affairs Unit publishes their book, “Diet Nation: Exposing the Obesity Crusade.”