Obesity and body fat distribution (FD) are established risk factors for chronic diseases. The body mass index (BMI) and the waist/hip circumference ratio (WHR) are used conventionally as indices of obesity and FD in epidemiological studies. Although some general limitations of these indices are recognized, others that affect their use in relative risks for disease are not well recognized. These include effects of sex, ethnicity, and especially age on the relationships between these indices and body composition, which can result in substantial misclassification of obesity and FD. There is considerable variability in body composition for any BMI, and some individuals with low BMIs have as much fat as those with high BMIs. This results in poor sensitivity for classifying levels of body fatness (e.g., too many "false negatives," or overweight individuals classified as not overweight), and relative risks are attenuated across all categories of BMI. A more serious problem, however, is that at different ages the same levels of BMI correspond to different amounts of fat and fat-free mass. Data from the Rosetta Study and the New Mexico Aging Process Study show that older adults have, on average, more fat than younger adults at any BMI, due to the loss of muscle mass with age. As a result, the sensitivity of BMI cutpoints with respect to body fatness decreases with age, and the use of a fixed cutpoint for all ages results in "differential misclassification bias." Taken together, these issues suggest that the increases with age in the prevalences of overweight and obesity, and in the risks for chronic diseases, may be mis-estimated using BMI. Similar issues may affect the use of WHR for estimating prevalences and associated risks of FD. New field methods for estimating body composition are available that can be applied in large, epidemiologic follow-up studies of chronic diseases. These methods will allow epidemiologists to consider, for example, whether it is increased fat, or the replacement of fat-free mass with fat, with age that is associated with risk for chronic disease.