Body Weight and Survival: The US experience

The expanding waistlines of the American population have stirred immense research and public interest in how secular changes in body weight affect population health. Big controversies and gaps remain in our understanding about excess weight and health. It remains unclear whether obesity-related excess mortality declines over age, and whether people should control and maintain weight for optimal survival. Despite overall health improvement, we do not know if the tide has lifted all boats, and people in all weight groups, fat or lean, are living longer. This dissertation analyzes the weight-mortality relationship in the US from three different perspectives. First, I use recalled weight at age 25 to classify weight status, and document for the US female population changes in the last thirty years in a variety of indicators of mortality differentials by age-25 weight. Second, I re-examine the age patterns of mortality differentials by weight concurrent with the baseline, using an age- and cohort-specific framework. Third, I apply the marginal structural model (MSM) to study the dynamic temporal process of weight, health conditions and mortality. The first two analyses use the National Health and Nutrition Examination Surveys and national mortality data, and the last analysis, the 1992-2006 Health and Retirement Study. In contrast with previous findings based on concurrent weight, mortality does not differ between the age-25 underweight and normal-weight, and is equally elevated for the age-25 overweight and obese women. Between 1976 and 2004, mortality differentials widened, and mortality reversals were observed for the overweight/obese. Overall life expectancy continued to rise because lean mortality rates declined at an unabated pace, and the majority of the population was still characterized as age-25 lean. Based on concurrent weight, the weight-mortality relationship is constant over age, but varies depending on the age of weight measurement. Excess mortality is larger for more recent cohorts. The MSM analysis suggests the diverse health implications of weight change trajectories that may not be captured by one single measure. The implications of this research are discussed.

Mortality

20490