Publishing bias stirs controversy
Controversy abounds in the obesity field. Much of the difficulty stems from the way early research of the 1970s and 1980s was allowed to be conducted and reported, so that it generally supported the interests of the weight loss industry. Research in the field was, and continues to be, intimately bound up with special interests at a number of universities and even at the federal policy level.
Thus, much of the well-publicized research not only claims that obesity is associated with a long list of severe health risks, but that it causes them. Although this is not supported by research, the idea that obesity is causal has been repeated so often that it is common belief among health professionals and the general public.
Even today the suggestion that obesity causes an array of diseases and conditions is encouraged on the CDC website, which says, “ Being overweight or obese increases the risk of many diseases and health conditions, including the following: hypertension (high blood pressure), osteoarthritis (a degeneration of cartilage and its underlying bone within a joint), dyslipidemia (for example, high total cholesterol or high levels of triglycerides), type 2 diabetes, coronary heart disease, stroke, gallbladder disease, sleep apnea and respiratory problems, and some cancers (endometrial, breast, and colon).” (www.cdc.gov/nccdphp/dnpa/obesity/index.htm)
However, the relationship is not as clear-cut as it once seemed from that earlier research. Critics are deconstructing much of that research and finding it wanting.
One can ask such questions as:
n Why are there no federal statistics on eating disorders? Why are eating disorders and disturbed eating being left out of national health programs, such as Healthy People 2010? Who might be impacted by an open and objective health policy on this severe health problem?
n Since the category of Overweight is shown by Flegal's CDC research and numerous other studies to be the category associated with lowest risk of death, and Underweight is associated with high death rates, why is CDC still fighting the "war against obesity," instead of focusing on health for people of all sizes?
Flegal research confronts
health risk claims
On April 20, 2005, a study published by Centers of Disease Control (CDC) researchers in the Journal of the American Medical Association (p1861-1867) shook the foundations of two popularly-held assumptions: first, that obesity causes the diseases commonly associated with it and, second, whether there is an association at all between obesity and higher mortality risk exept perhaps at the higher end.
The study, led by Katherine Flegal, PhD, senior research scientist with CDC’s National Center for Health Statistics, looked at actual deaths, unlike previous highly publicized CDC studies which were based on abstract mathematical models, and far more rigorous than those studies. This new and more inclusive study found:
- Somewhat MORE total annual deaths in the two LEANEST categories (underweight and normal weight) than in the two heaviest categories (overweight and obese).
- Thus, FEWER deaths were associated with the HIGHER weights of body mass indexes of 25 and over than with BMIs under 25.
- The group defined as OVERWEIGHT (BMI 25 to 29.9) had the LOWEST death rate of all.
Backlash and rebuttal
But no sooner had this news hit the media than a backlash began. The Klegal paper was immediately attacked by the “Harvard group” of obesity specialists, who called a news conference to criticize the findings. As quoted in the New York Times, Dr. Frank Hu, an associate professor of nutrition and epidemiology at Harvard, citing their own long-running and well-publicized Nurses’ Health Study, insisted that as body mass index increases, “the death rate increases dramatically.”
He and Dr. JoAnn Manson, also of Harvard, said the new federal analysis had failed to exclude smokers and people who were already ill. (This was untrue.) “That can lead to serious underestimates of mortality linked to overweight and obesity,” Manson said. Dr. Walter Willett, chair of nutrition at the Harvard public health school, agreed, calling the new analysis “deeply flawed.”
Dr. Katherine Flegal replied that she and her colleagues had analyzed their data in a variety of ways, both with and without current or former smokers and people who had chronic diseases. She said the results always came out the same: There was no mortality risk from being overweight and little from being obese, except for the extremely obese.
Flegal’s co-author Dr. David F. Williamson added that one reason for the discrepancy between their results and other findings might be the populations under study. While the Harvard group polled mostly-white nurses in a few states, self-reported, and the cancer society looked at volunteers, “We have data sets that are truly nationally representative of the U.S. population,” said Williamson.
Another reason for the differing conclusions, Flegal said, may be that the Harvard and cancer society researchers excluded large numbers of subjects from their analysis for one reason or another. One analysis of the Nurses’ Study excluded nearly 90 percent of the deaths, she pointed out. (Although Flegal didn’t say it, this is a technique commonly called “cherry picking,” the sorting out and focusing on a smaller subset that gives desired results.)
She and her colleagues used actual measured weights and heights, not self reported ones, she added, so they may have had more accurate numbers to work with.
“I don’t know what to say” about the attack on the paper, Flegal said. “I don’t have a problem with people at a conference talking about their data, but I do have a problem with their talking about our data and saying we should have found the same things that they found.”
Flegal and co-authors, Williamson and Elsie Pamuk, concluded, “Hu et al. speculate that the number of deaths attributable to obesity in the United States may be underestimated when relative risks are calculated on the basis of current body mass index (BMI). They cite no data to support their speculations, but instead invoke the notion of ‘reverse causality.’ They hypothesize that relative risks are lowered by obese people who become ill, lose weight because of this illness to become normal weight, and die shortly thereafter of the underlying illness, surviving just long enough to be included in the study. However, this reverse-causation hypothesis is unlikely to be the correct explanation for the lower relative risks in the elderly for two reasons: first, because data show that exclusion of preexisting illness has little effect on relative risk estimates, and second, because weight loss from obesity to normal weight is relatively uncommon.”
Controversy apparently simmered within CDC as well: Director Dr. Julie Gerberding seemed to repudiate the Flegal findings and said the agency will continue its fight against obesity. However, the agency did publish online Katherine Flegal’s rebuttal through an analysis of the data online. It may be reviewed at the following site: http://www.cdc.gov/nchs/products/pubs/pubd/hestats/excess_deaths/excess_deaths.htm
(Flegal KM, Graubard BI, Williamson DF, Gail MH. Excess Deaths Associated With Underweight, Overweight, and Obesity. JAMA. 2005;293:1861-1867 (April 20, 2005). jama.ama-assn.org/cgi/reprint/293/15/1861.pdf.)
Are weight categories set too low?
The following weight categories were set in 1998 by the NHLBI Guidelines, defining health risks at a body mass index (BMI) of 25 and above:
ADULTS |
| Weight category |
BMI |
Underweight |
|
<18.5 |
Normal weight |
|
18.5-24.9 |
Overweight |
|
25 - 29.9 |
Obesity |
|
30 or more |
|
This distinction is controversial because it suggests that health risks, if any, begin at levels much lower than supported by research.
The NHLBI Guidelines (Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults: The Evidence Report) was developed by a 24-member panel of specialists convened by the National Heart, Lung and Blood Institute. (1)
The report analyzed 236 randomized controlled trials and found the lowest mortality is at a body mass index (BMI) of 24.8 for white men, 27 for African-Americans, and 25 to 30 for adults age 55 and older. Obviously, this was about at the level that eventually became the overweight category.
But then, oddly, someone threw Harvard’s Nurse’s Health Study into the mix. Unlike the other 236 studies, the Nurse’s Study is nonrandomized and uncontrolled. It was a study of nurses who were mostly white women, lived in 11 states, and self-reported their weight and health through a mailed-in questionnaire. It was hardly a study that scientists would consider to represent the U.S. population of men and women, or even extrapolate to just women in those 11 states. The key: it reported that there were higher risks above a BMI of 22 for the people in the study – or at least a subset of them.
Apparently this was enough to convince the 24 members of the panel, and they proceeded to set overweight at the very level linked to the lowest death rate.
There was no reasonable justification for this. And since then other research has amassed further evidence to question those official weight categories. Most notable is CDC’s own Flegal study. (2) Unfortunately, despite this powerful study that was published in the fall of 2005, the categories remain unchanged, and CDC continues its “war against obesity.”
The question is asked: How many people on that 24-member panel had vested interests? The answer is obscure. The panel, chaired by F. Xavier Pi-Sunyer, was not asked to disclose financial affiliations. However, Pi-Sunyer himself had previously recorded his own financial support from these seven companies, all involved in the weight loss industry, Weight Watchers, Lilly, Genentech, Hoffman-LaRoche, Neutrogen, Wyeth-Ayerst and Knoll Pharmaceuticals.
1. National Heart, Lung and Blood Institute, NIH. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults: The Evidence Report, 1998.
2. Flegal KM, Graubard BI, Williamson DF, Gail MH. Excess Deaths Associated With Underweight, Overweight, and Obesity. JAMA. 2005;293:1861-1867 (April 20, 2005). jama.ama-assn.org/cgi/reprint/293/15/1861.pdf.
3. Berg FM. NIH-NHLBI Guidelines: An Evaluation. Healthy Weight Journal 1999;13:2:26-29
|
