In 2005, a team of investigators at the Center for Disease Control and Prevention, or CDC, published a paper that shook the foundations of our long-held beliefs; they found that death rates due to overweight (BMI 25-30) and obesity (BMI >30) were actually lower than death rates due to underweight. The paper came under withering criticism by scientists and nutritionists who had been preaching the gospel of weight control. “Libertarian” organizations funded by the food industry, such as “Food Freedom“, piled on with glee. Their basic message: You can’t trust those scientists; just let the consumer beware (and, I presume, let Darwin and the forces of the “free” market weed out those who did not).
The most serious criticism of the study was that it was not even a prospective study, let alone a controlled one. It had a retrospective design, the least reliable of epidemiological studies. Furthermore, it relied on death certificate reports, notoriously inaccurate and many times misleading. Just as one example, a patient dying of a heart attack would normally have his cause of death listed as myocardial infarction, without mention of the underlying type 2 diabetes. It also flew in the face of a huge body of literature showing excess mortality due to obesity and its link to type 2 diabetes and its cardiac and kidney complications, as well as some of the biggest cancer killers such as colon, breast, esophageal, uterine, ovarian, kidney, and pancreas—all obesity-related cancers. I might add that recently prostate cancer was added to the list.
In a paper published in the Nov 7th issue of JAMA, the authors of the 2005 study went back to the same databases they had used, in order to determine which causes of death are associated with underweight and which are associated with overweight and obesity.
The envelope, please…
The group with normal weight (BMI 20-25) were considered the baseline and the groups over or below this range were compared to them.
The diseases associated with overweight and obesity were not surprising; the usual suspects were identified, again: heart disease, kidney disease, type 2 diabetes, and cancers related to excess weight.
What is intriguing is the relationship of underweight (BMI less than 18.5) to disease; to my knowledge, this kind of information has not been available before this study was published. These people suffered excessive mortality rates from acute and chronic respiratory disease, injury, as well as some cancers that are not related to excess weight, and miscellaneous other diseases (Alzheimer’s, Parkinson’s).
The study is revealing, especially in its identification of underweight as not healthy. Jewish and Italian mothers, please stand up and say in unison: I TOLD YOU SO.
What could explain the lower rate of mortality associated with overweight than that associated with underweight? No answer is offered by the survey, but what comes to mind is that modern medicine is simply too good at warding off death. We can now keep patients with heart disease and kidney disease alive with all kinds of wonderful drugs and procedures. Even colon and breast cancers, major killers in the not-too-distant past, are now more like chronic diseases, thanks to early detection, chemo- and biological therapy. And bear in mind: The study measured death rates only; it did not attempt to measure the prevalence of disease that has not resulted in death during the study period.
What could explain the underweight association with disease? A study of this kind cannot establish causality, but one can speculate (especially when not subjected to the jaundiced eyes of peer reviewers). Two things come to mind: muscle mass and immunity. BMI below 18.5 almost by definition means that some of the lost weight comes from loss of muscle. One of the most important muscles, when it comes to infectious diseases, is the diaphragm. When this muscle is weak, respiration is weak, lungs are not ventilated completely, and before long pneumonia ensues. Bedridden patients, elderly people, patients with AIDS, cachectic (wasted) patients with advanced cancer—all are susceptible to respiratory infections. In fact, this is the most common cause of death in such patients. Likewise, low nutritional status is associated with defective immunity to infectious diseases. But, as I said, this is sheer speculation.
What about the validity of the study as a whole? None of the deficiencies that plagued the original paper, namely retrospective design and reliance on death certificates, have been cured. The methodology is identical, the database is identical—the only difference is that here, we get an analysis of the diseases associated with out-of-the-normal body weights.
Dr, Flegal, the senior author of the study, was quite cautious in her assessment of the study. According to Dr. Flegal, “The take-home message is that the relationship between fat and mortality is more complicated than we tend to think.” On the other hand, experts like Walter Willett, professor of epidemiology and nutrition at the Harvard School of Public Health, “dismissed the findings as fundamentally flawed, saying [that] an overwhelming body of evidence has documented the risks of being either overweight or obese.” He called the findings “rubbish.”
Well, well, I wouldn’t go that far. The study did contribute some valuable information on underweight and its relationship to disease. And it provoked controversy—which is great; this is how science is done and how progress is made.