Obesity in Capsule
Overweight and obesity have long been associated clinically with early death and disability, and with susceptibility to heart attack, stroke, diabetes, and plethoric habitus. The classic build and blood pressure studies of the early 20th century insurance industry provided quantitative estimates of the excess mortality risk mainly of gross overweight. Clinical studies found frequent correlates of obesity with metabolic disorders, mainly glucose intolerance and hypertension.
Epidemiological studies of the modern era have gradually thrown light on the nature and degree of cardiovascular, chronic diseases, and overall mortality risk to body mass. The issue has been central to CVD epidemiology from its origins. Despite the precision of weight, the difficulty to measure fat composition of the body and the complex interrelations of overweight and metabolism have made study difficult. Gender and age differences, duration of follow-up, consideration of smoking habit and other risk factor correlates of smoking complicate the study. Cultural differences in the habitual diet by which individuals and populations achieved overweight play a major role in different findings about risk and obesity, as does the duration of follow-up of cohorts; the risk of overweight affects longer term risk of those measured in health.
Ancel Keys in particular found that the distribution of CVD events in relation to fractiles of body mass index or skinfold thickness varied greatly among populations, that overweight had to be at the extreme to affect risk, and that maintained weight or even weight gain in middle-ages of men of many cultures was associated with favorable mortality from all causes. 
“Patient epidemiology,” with longer term studies, exclusion of early deaths, and adjustment for confounding correlates of overweight, all tend to enhance the relationship found for body mass and CVD risk. They tend to explain the under-estimates of risk found in some studies among healthy cohorts and representative population samples. (Henry Blackburn)
 Keys, A. 1980. Seven Countries: A Multivariate Analysis of Death and Coronary Heart Disease, Cambridge: Harvard University Press.