Obesity decreases longevity. Some new data involving body mass index (BMI;kg/m2) and mortality rate (MR) tentatively support three suppositions 1) the number of annual deaths attributable to obesity is roughly 1/3 of that previously estimated;2) overweight (as opposed to obesity) does not result in an increased MR and may even confer a survival advantage;and 3) the deleterious effect of overweight and obesity on MR has declined over calendar time, and obesity and overweight are less harmful than they once were. Other data and much recent discussion highlights that we may be mischaracterizing the impact of obesity by reliance on BMI as opposed to measures of total or regional fat mass. Each of the three suppositions listed above is important. The first helps place the magnitude of society's obesity problem in perspective. Regarding the second point, the extent to which overweight influences MR has important implications on how severe a health problem overweight is and how we should respond to it. The third point is quite novel and, if true, has the important implication that continuing to develop better treatments for the negative sequelae of obesity might dramatically lessen the need to directly address obesity per se. Yet, there are reasons to question each of these propositions and we propose to thoroughly evaluate them. We also aim to bring more direct measures of adiposity to bear on such questions.
In Specific Aim 1, we will separately analyze (i.e., not pool) raw data from multiple well-characterized US datasets, each with a sufficiently long follow-up, to test whether the effect of obesity on MR has changed over calendar time within studies. We will then conduct a comprehensive meta-analysis to address this same question across studies. By using both approaches to address Specific Aim 1 we counterbalance the different strengths and limitations of these approaches, thereby allowing us to rigorously address this important question.
In Specific Aim 2 we will use the bio-impedance analysis measurements of body fat and waist and hip measurements in NHANES III to estimate deaths attributable to obesity, years of life lost because of obesity, and mortality hazard ratios. Measuring obesity as body fatness and fat distribution instead of just using BMI should provide a more solid basis for public health and clinical recommendations by focusing on the more relevant variables of body fatness.
