Who Knew? You Can Be Too Thin! Of Course, Most People Aren't...

Abstract & Commentary

By Barbara A. Phillips, MD, MSPH, Professor of Medicine, University of Kentucky; Director, Sleep Disorders Center, Samaritan Hospital, Lexington. Dr. Phillips is a consultant for Cephalon, and serves on the speaker's bureaus for Resmed and Respironics.

Synopsis: In a large study of American adults, all-cause mortality was lowest for those with a body mass index (BMI) of 20.0-24.9 kg/m2.

Source: Berrington de Gonzalez AB, et al. Body-mass index and mortality among 1.46 million white adults. N Engl J Med 2010;363:2211-2219.

This report results from a pooled analysis of 1.46 million white adults who were enrolled in 19 different prospective studies funded by the National Cancer Institute. To be included in this analysis, a study had to have started in 1970 or later, and had to include data about height, weight, and smoking status. Most studies also included information about pre-existing health conditions, alcohol consumption, cardiac disease, education, marital status, and physical activity. The investigators formatted these variables to be consistently classified across studies into standard categories. The cause of death was determined from death certificates or medical records. Analyses used proportional-hazards models, with age as the underlying time variable, and adjusted for alcohol intake, educational level, marital status, and physical activity. These investigators defined a BMI of 22.5-24.9 kg/m2 as the referent category because of previous studies demonstrating this BMI was usually associated with the lowest mortality.

In the large, combined (1.46 million people) cohort, 58% were women, median age was 58 years, and median BMI was 26.2 kg/m2. Smoking was inversely correlated with BMI; only 13% of the cohort were active smokers, though only 53% reported that they had never smoked. Pre-existing cancer and emphysema were more common in thinner people, but physical inactivity and lack of a college degree were both associated with a higher BMI.

The rate of death from any cause was lowest among those with a BMI of 22.5-24.9 kg/m2, and the risk of death increased with both higher and lower levels of BMI. When the investigators excluded smokers and those with heart disease or cancer at baseline, the risk of death increased for a BMI > 25 kg/m2, but decreased for a BMI < 22.5 kg/m2. The hazard ratios by BMI for death for healthy women who never smoked are presented in Table 1 (right); results were similar for men.

Table 1. Hazard ratios by BMI for death in healthy women who never smoked.

BMI (kg/m2) Risk Hazard Ratio for Death

25-29.9 1.13

30-34.9 1.44

35-39.9 1.88

40-49.9 2.51

Table 2. Items that contain about 250 calories.

2 1/2 apples

5 strips of bacon

20 oz of beer (1 3/4 cans)

10 carrots

1/6 quart ice cream

1/2 cheeseburger

10 oranges

5 chocolate chip cookies

20 oz of cola beverage (1 3/4 cans)

3 glasses of skim milk

2 1/2 fried eggs

1 fried chicken breast

Table 3. Activities that consume about 250 calories for a 150-lb person (more for a heavier person).

Walking 45 minutes

Bike riding 30 minutes

Swimming 22 minutes

Running 13 minutes

Sitting on the couch watching TV 3 1/4 hours

Climbing 12 flights of stairs

Further adjustment for alcohol consumption, physical activity, educational level, and marital status only slightly reduced the hazard-ratio estimates associated with a BMI ≥ 25 kg/m2. The risk of a higher BMI was greater for younger people. The cause of death for a BMI ≥ 25 kg/m2 was highest for heart disease and lowest for cancer.

The risk of death for those with lower (< 20 kg/m2) BMIs declined as the length of follow-up increased, and was only significant for the very thinnest people (BMI < 18.4 kg/m2) at 15 or more years of follow-up.


This is not the first paper to show a relationship between obesity and death, of course, but it is the largest. Further, the authors were able to adjust for smoking and prevalent illness, which has not always been possible in many prior studies of the relationship between BMI and mortality.1-4 The current study also confirms that being "merely" overweight (defined as a BMI between 25 kg/m2 and 30 kg/m2) is associated with an increased risk of death.1,4,5 The authors note that although the data suggest that very thin people have increased risk of death in the short run, this risk is weak in the long run, and the association between low BMI and death was much weaker in physically active people. They interpret this to mean that the association between low BMI and death probably results from pre-existing disease.

The relevance of these findings is amplified by the fact that they apply to so many people. Two-thirds of United States adults are overweight or obese.6,7 Among non-Hispanic persons in the United States, an estimated 11% of men and 17% of women had a BMI ≥ 35 in 2008.7

Sadly, there were more than five times as many deaths among participants in the highest BMI categories (BMI of 35.0-39.9 kg/m2 and 40.0-49.9 kg/m2) than in previous related studies,1-5 because severe obesity has become more common.

Let's face it: Obesity and overweight are huge medical problems in this country. American ingenuity has responded in predictable ways. Weight Watchers is a $1.4 billion industry, with a vast array of options for the would-be thin person, including books, internet programs, special food products, and branded items on menus of some popular restaurant chains.8 Just in time for the holidays, the highly successful Weight Watchers program just announced a new point system. And Weight Watchers works.9 Of course, most insurance plans won't pay for Weight Watchers. They will, however, often pay for bariatric gastric reduction procedures — one of which has just been recommended by an FDA panel10 to be approved for use in those who have a BMI of just 30 kg/m2 — along with a medical complication. Is there something wrong with this picture?

For those of us on the front lines, the message is clear: Painful and time-consuming as it is, we must deliver the message that being overweight or obese kills. Since most of my patients have sleep apnea (and diabetes and hypertension), I have a great deal of experience with this endeavor. And most of it is pretty negative. Patients expect the doctor to tell them to quit smoking. But they often don't expect to hear that they need to lose weight. In my experience, patients can and will punish the bearer of this news in a variety of ways, including (but not limited to) describing (in excruciating and time-consuming detail) what they eat, explaining (in excruciating and time-consuming detail) why they can't exercise, crying, or complaining to the hospital administrator. It helps to have a handout and a plan. Our handout includes a BMI chart so the patient can "do the math," rather than take my word for it. It also includes calorie charts for food and exercise (see Tables 2 and 3, above), so that patients can get the message that it is easier not to take in calories than it is to expend them. But it's an uphill battle. I would be a much more popular doctor if I were better at this part of my job, or if I simply didn't take this part of my job so seriously.

It is likely that obesity now kills more Americans than cigarettes do. And, on that point, some parallels might be drawn. Just as we didn't make much progress against cigarette smoking until we stigmatized it, limited it, and taxed it, I doubt we will make much headway against obesity as long as we continue to pretend that it is a harmless lifestyle choice instead of a deadly, chronic, largely self-inflicted disease.


1. Prospective Studies Collaboration. Body-mass index and cause-specific mortality in 900,000 adults: Collaborative analyses of 57 prospective studies. Lancet 2009;373:1083-1096.

2. Flegal KM, et al. Cause-specific excess deaths associated with underweight, overweight, and obesity. JAMA 2007;298:2028-2037.

3. Orpana HM, et al. BMI and mortality: Results from a national longitudinal study of Canadian adults. Obesity (Silver Spring) 2010;18:214-218.

4. Calle EE, et al. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med 2003;348:1625-1638.

5. Pischon T, et al. General and abdominal adiposity and risk of death in Europe. N Engl J Med 2008;359: 2105-2120; Erratum published in: N Engl J Med 2010;362:2433.

6. International Obesity Task Force. Global prevalence of adult obesity. Available at: www.iotf.org/database/documents/Global PrevalenceofAdultObesityOctober2009v2.pdf.

7. Ogden CL, et al. Prevalence of overweight and obesity in the United States, 1999-2004. JAMA 2006;295: 1549-1555.

8. Fruit Lovers, Weight Watchers Has Good News: Oreo Fans, Sorry. New York Times, Dec. 4, 2010, p A1.

9. Tsai AG, Wadden TA. Systematic review: An evaluation of major commercial weight loss programs in the United States. Ann Intern Med 2005;142:56-66.

10. Panel Votes to Expand Surgery for Less Obese. New York Times, Dec. 4, 2010, p B1.