Group Fitness Classes Fitness Class Descriptions Personal Training/Private Instruction Health, Wellness & Nutrition Articles

Weight and Death

Dr. Ralph Ofcarcik, Ph.D.
Director of Nutrition Services

The 2005 Smokin’ Gun Award for controversial research has to go to Katherine Glegal and David Williamson (Centers for Disease Control and Prevention, i.e. CDC) who reported a link between (slightly) overweight Americans and reduced risk of death*. Findings like this rattle cages, especially among those whose similar projects have rendered different results. Wranglers at Harvard and the National Cancer Institute, for instance - their research now being challenged, quickly went for their gums, er-r-r guns. Vocal gunfire erupted at several scientific meetings with claims of uncompromised biases, design flaws, and frequent references to the established(?) paradigm, i.e. normal body weight supports optimum health and longevity. Never to back down from a fight, CDC researchers likewise brought their gums to town, formidably defending their JAMA-published, multi-year survey of 37,000 Americans.

So who is right? In an age when obesity is rampant, suggesting that extra padding = extra life years can ruffle feathers among health educators, not to mention sending a misinterpreted message to an already overweight populace untrained in multivariate analysis. In this situation, we have leading researchers from prestigious organizations (CDC, Harvard, NCI) generating reams of useful data, statistically significant trends, but conflicting outcomes. Could both sides of the debate be correct? Well, possibly (with this thought being entertained by at least one of the gunfighters, then quickly dismissed). Nonetheless, scientific squabbling in any discipline is the norm, it’s healthy, and a part of better understanding our complex world. In the current weight-death debate, researchers on both sides of the table have been forced to probe deeper into population variables, looking for logistics and data supportive of their sacred cows.

The crux of the issue centers around a single CDC finding: Americans with a Body Mass Index (BMI) of 25 up to (but less than) 30 have a lesser chance of dying than normal weight people, i.e. BMI = 18.5 up to (but less than) 25. For a 5’4” woman, for example, chances of dying are significantly less if she weighs between 146-175 lbs, compared to a normal weight (108-145 lbs). Similarly, a 5’10” man weighing 174 to 209 lbs will have a lesser chance of dying than if he weighed 129 to 174 lbs.

However statistically significant, nice, tidy results like these will never get to first base in the scientific community unless project leaders sharpen their pencils and mathematically compromise known biases from lifestyle and health. For example, knowing that not all slender people are healthy, Harvard and the NCI legitimately challenged CDC researchers about the inclusion of people in their study who were of normal weight, but poor health (smokers, alcoholics, cancer patients, etc.). Strategic challenges – certainly – but impressively addressed by the CDC research team. Specifically:

Illness Lowers Body Weight When people contact a serious illness (cancer, alcoholism, Alzheimer’s disease, emphysema, congestive heart failure, etc.), weight loss occurs. Normally, people with known illnesses are excluded from population mortality studies. However, at the beginning of any cohort investigation, some people will have illness but are unaware that they are sick. This is precisely the reason that most epidemiologists will exclude participants who die within the first 5-10 years following the onset of the project. Harvard and the NCI argue that at least a 10-year span is necessary before recording mortality data to screen out (by death) those participants who were unknowingly ill at the beginning. The CDC research excluded participants who died within the first 3-5 years. Assuming the validity of the recommended 10-year waiting period, it is possible that a significant percentage of the leaner CDC participants were unknowingly ill to begin with but was included in the tally of lean deaths. [Note: Although the CDC questioned the validity of waiting 10 years before recording, they plan on re-accessing the data (the project is ongoing) in the years ahead.]

Smoking Lowers Body Weight Smokers weigh less. Just ask any ex-smoker who watched an extra 10-20 lbs accumulate within the first few months after quitting. In the CDC project, smokers and ex-smokers were statistically filtered out of the analysis. The result: lean non-smokers had a higher risk of dying than non-smokers who were overweight. A surprising similar trend (low weight = higher death) was found in the Nurses Health Study (Harvard) which tracked 115,000 women for 16 years. However, the trend included smokers. When smokers were excluded (as well as ex-smokers, women who knew they were sick with a life-threatening disease, women who died within the first 10 years of the project onset, and women who had recently and involuntarily lost weight – a sign of illness), participants who had a normal body weight had the lowest risk of death. Risk increased with body weight.

In Support of the CDC Study

With the statistical validity of the CDC’s shorter prescreening period in question, the nod (so far) has to go to NCI and Harvard for the strongest weight-death argument (i.e. lean people have a lesser risk of dying). However, the CDC study actually weighed the participants whereas the Nurses Health Study and the NCI study relied on questionnaires. Could a prevailing sense of modesty have caused participants to understate their weights and skewed the outcomes? In addition, healthcare in this country has improved dramatically. For example, over the past four decades, hypercholesterolemia in America has dropped more among the overweight and obese than among lean individuals. Could a trend toward more effective healthcare among ill, overweight individuals offset a past life expectancy advantage of being lean? CDC says possibly. No, says Harvard. Per Frank Hu, a Harvard researcher associated with the Nurses Health Study, “There’s no evidence for a substantial decline in the impact of obesity on mortality over time”. And per Walter Willett, Chair of the Nutrition Department at the Harvard School of Public Health, (with reference to hypertension), “Treatment doesn’t bring back the risk to that of someone without hypertension. High blood pressure raises the risk (of heart attack or stroke) three-fold and treatment reduces it by 25%”.

Best Advice

Undoubtedly, the weight-death feud will continue, perhaps for years. However, the CDC, recognizing the possible bias of unidentified lean but ill people in their analysis, is having the best mathematical minds re-evaluate the data. And even though population studies never establish cause and effect, they provide us with useful clues – helpful for NIH and private organizations in financing research at “the next level”. Realistically, there could be dozens, perhaps thousands of variables that could affect epidemiological investigations. Factors such as race, diet trends, stress, sleep patterns, genomics, time span of the study (could butterfly-like subtle changes in environment – food composition, air & water quality -have profound effects upon life expectancy?), and locations(s) (residences) of the participants (pollution, sociological issues, local food preferences, etc.), may, someday, independently or in combination, be integrated into the equation.

Of utmost importance, disregard the advice of food service lobbyists (quick to exploit the CDC study) who now claim that our current obesity epidemic is all “hype”. Per NCI’s Michael Thun, “. . . it’s irresponsible for commercial interests to trivialize what has become a major public health problem in the U.S. and globally”.

Until absolute certainty is established, trust Harvard and the NCI. Lose weight, feel better, and (possibly) live longer.

*Flegal KM, Graubard BI, Williamson DF, Gail MH. Excess deaths associated with underweight, overweight, and obesity. JAMA 2005: 293(15):1861-7 (ISSN: 1538-3598.