Skip to Content
Discovering the causes of cancer and the means of prevention

NCI Cohort Consortium Investigates the Role of Obesity

July 2011 - Linkage Newsletter

In many developed countries today, more than half the population is overweight or obese, and the potential impact of this situation on total mortality has drawn considerable attention. In 2008, investigators from more than two dozen prospective studies in the NCI Cohort Consortium formed a collaboration to measure the mortality effects of body mass index (BMI, kg/m²) and related factors. The initiative is led by Patricia Hartge, Sc.D., Deputy Director of DCEG's Epidemiology and Biostatistics Program; Dr. Michael Thun of the American Cancer Society in Atlanta, Georgia; and Dr. Walter Willett of the Harvard School of Public Health in Boston, Massachusetts.

The first in a series of studies combined the data from 1.46 million persons of European ancestry, of which 160,087 died during a median follow-up of 10 years. Among the healthy never-smokers (i.e., no self-reported history of cancer or heart disease at baseline, deaths = 35,369), a J-shaped relationship was seen between BMI and all-cause mortality, with the lowest mortality at normal weight (BMI = 20.0–24.9). Men and women showed similar effects, which were strongest for BMI measured at younger ages and weakest for BMI measured after age 70. Amy Berrington de González, D.Phil., Radiation Epidemiology Branch, lead author of the paper published about this study in the New England Journal of Medicine, noted, "In this study, we focused on whites living in developed countries due to their large numbers in our study populations. As the relationship between BMI and mortality may vary across racial and ethnic groups, our next priority is to study the effects of BMI on minority populations of the United States."

The African American population, in particular, warrants investigation, with 1 in 10 adults in this group now in the morbidly obese category (BMI ≥ 40). Yikyung Park, Sc.D., Nutritional Epidemiology Branch (NEB), and colleagues are pursuing earlier findings from the NIH-AARP Diet and Health Study and other cohorts, which surprisingly showed that mortality rose less steeply with higher BMI among African Americans than among other populations. The Cohort Consortium also is investigating the influence of waist circumference on mortality among African Americans. Dr. Julie Palmer of the Boston University School of Public Health in Massachusetts is leading an effort to increase the study populations by adding cohorts with substantial numbers of African Americans. Members of the Cohort Consortium also are exploring the relation of BMI to mortality among Asian Americans.

In addition to examining the mortality effects of BMI in various populations, the Cohort Consortium is seeking to disentangle the role of obesity-related factors, including waist circumference, height, and physical activity. Although we know that physical activity affects the risks of obesity and mortality, the interplay of obesity, activity, and mortality remains uncertain. Dr. I-Min Lee of the Harvard School of Public Health, Steven C. Moore, Ph.D. (NEB), Dr. Alpa V. Patel of the American Cancer Society, and colleagues have recently pooled data from six cohorts in the Cohort Consortium that assessed physical activity for a total of 654,827 individuals and 82,465 deaths. Within each BMI category, they saw clear benefits of activity, even at the modest levels recommended in current guidelines (see Figure 1). The investigators are now extending the analysis to include additional cohorts in order to study risks of specific cancers.

Figure 1 shows the estimated hazard ratios of death in relation to joint categories of body mass index (BMI) and physical activity. MET is the metabolic equivalent task. BMI is the weight in kilograms divided by the square of the height in meters. Obese class I is defined as a BMI of 30.0-34.9, overweight is a BMI of 25.0-29.9, and normal weight is a BMI of 18.5-24.9.