The trusted source for
healthcare information and
Abstract & commentary
Synopsis: A BMI between 20-25 is optimal and is associated with the longest survival for any causes of death.
Source: Calle EE, et al. N Engl J Med 1999;341: 1097-1105.
The relationship between excessive body weight and mortality is problematic. In addition to the questions of the optimal weight at different ages and gender, there has been the suggestion of a U-shaped curve regarding body weight with respect to survival; the latter has stimulated much discussion regarding the potential downside of being underweight or lean, perhaps due to unrecognized medical conditions. The Cancer Prevention Study II, a prospective study of mortality among U.S. men and women, initiated in 1982 by the American Cancer Society, recently reported 14-year follow-up data. By 1996, 20% of the entire cohort of 457,785 men and 588,369 women had died. All participants were divided into 12 categories of body mass index (BMI). The primary end points were all-cause death as well as cardiovascular disease mortality. Smoking history, status of disease at entry, and race and gender were examined. The results indicate a relationship between increasing BMI and mortality that differed by smoking status and the presence of any disease. Obesity was more strongly associated with decreased survival in nonsmokers and in those without a disease history. Leanness was most strongly associated with decreased survival in smokers with a history of disease. Intermediate survival was noted for smokers without a history of disease and for nonsmokers with concomitant disease. At a BMI of 28 and higher, the relative risk of death began to increase most steeply for the nonsmokers without a history of disease. The highest mortality rates occurred in obese men, with a relative risk of 2.7 vs. 1.9 in women. There was a small increase in risk in the leanest men and women. The nadir of the BMI curves and mortality was at a BMI between 23.5 and 25 in men and 22.0 and 23.4 in women. In nonsmokers without a disease history, the association between a high BMI and increased mortality was stronger in whites than in blacks. BMI and cancer death demonstrated a positive relationship; there was no elevation in risk among lean individuals for cancer. The cardiovascular curves were U-shaped; there was no increased risk of dying in lean men and women. A high BMI was predictive of death from cardiovascular disease; it was greater in men (relative risk 2.9) than in women; risk began to increase at a BMI of more than 25 in women and 26.5 in men. The relative risk associated with high BMI diminished with increasing age. Overall death rates increased throughout the range of moderate to severe overweight in both men and women, less so in blacks and particularly black women (about one-third lower than white women). Cancer deaths increased by 40-80% in the heaviest groups of men and women, without a concomitant increase risk in lean subjects. Calle and associates conclude that a BMI between 20 and 25 for men and women of all ages is optimal and is associated with the longest survival for any causes of death. "These data offer support for the use of a single recommended range of body weight throughout life."
Comment by Jonathan Abrams, MD
These data, while not surprising, are of great interest because of the enormous size of the study population (> 1 million) and the long-term (14-year) follow-up. That approximately one-third of American adults meet WHO criteria for a grade 1 overweight (BMI 25-30) and 22% are even more overweight confirms that obesity is a substantial health problem. Only 8% of the adults in the United States have a BMI less than 20. The health care burden of a high BMI is well demonstrated in this study. It is unclear why blacks carry a lower burden of mortality for comparable degrees of obesity; however, it must be stressed that moderate to high levels of BMI are adverse for all causes of death unrelated to age, gender, or ethnicity. This study did not include coronary artery risk factor measurements; one cannot conclude that increase in mortality among the healthy overweight was related to coronary events, although this is likely. Obesity is associated with insulin resistance or the "metabolic syndrome," which includes dyslipidemia, hypertension, impaired glucose tolerance, and overt diabetes. For high-BMI individuals, particularly nonsmokers without disease, the BMI curves in this study are truly alarming.
Nonsmoking, healthy, significantly overweight men and women have a lower mortality rate related to BMI than smokers.