This report: 1) examines the importance of long-term follow-up (25 years) in BMI-cardiovascular mortality associations in eight age-gender subcohorts from the Chicago Heart Association Detection Project in Industry; and 2) addresses the impact of previously identified methodologic problems in studies on weight-mortality associations, which include no adjustment for cigarette smoking, adjustment for physiological effects of obesity, and no exclusion of early deaths. Associations were adjusted for age and ethnicity only, then for cigarettes/day, and for systolic blood pressure, serum cholesterol, and diabetes, for three periods of follow-up, 0 to 15, 6 to 15, and 16 to 25 years. In 25-year follow-up with exclusion of deaths for 0 to 15 years, there was a positive association in all subcohorts, with seven significant, both with and without adjustment for smoking or obesity-related risk factors. In contrast, with adjustment for obesity-related risk factors and no exclusion of early deaths in shorter-term (15-year) follow-up, there were two non-significant positive, three non-significant inverse, one significant positive, and one significant quadratic association. Failure to address potential methodologic problems can substantially alter associations in BMI-mortality studies and may contribute to observed differences among studies. The long-term outlook with overweight or obesity is adverse and deserves concerted efforts in prevention and treatment.