front |1 |2 |3 |4 |5 |6 |7 |8 |9 |10 |11 |12 |13 |14 |15 |16 |17 |18 |19 |20 |21 |22 |23 |24 |review

To determine the trajectory of weight loss with aging, we used a random effects model accounting for between-subject variation in baseline weight and the change in body weight with age, while adjusting for baseline age, sex, and housing strata. A second regression model was fit to test the effect of physical activity on weight change during the follow-up. This model was also adjusted for baseline age, sex, height, race, education, housing, functional disability score, mobility, smoking, and the prevalence of chronic conditions. In the preliminary analyses, we observed no effect modification by sex of the relation between physical activity and weight loss; therefore, the data for men and women were combined. Since the longitudinal relation between physical activity and weight change is most likely influenced by chronic disease and survival of 12 years, the statistical modeling was performed on the entire cohort, stratified by survival status, and then stratified again by survival status and chronic disease status (no/yes).