Obesity has, in a number of studies, been found to correlate to disability and mortality, primarity due to diseases of the circulatory and musculoskeletal systems. In addition, an excess mortality among underweight subjects has been observed in previous studies. Five complete birth-year cohorts (1926–1930) of male residents in Malmō (n = 7697) were invited to the survey at the Department of Preventive Medicine, Malmō General Hospital, and 5926 (77%) attended with complete data. Each subject was followed from inclusion, defined by the date of examination, until the end of the calendar year when he turned 58, a total study period of approximately 11 years. Data on about 300 questionnaire items and laboratory tests were determined at the health survey visit. Nationwide Swedish data registers were used for surveillance. Of the participants, 4.7% were underweight, 37.7% overweight, 7.3% obese and 50.3% normal weight; 849 (14.3%) had been granted disability pension at the end of follow-up, 717 after screening. After adjustment for smoking there was a J-shaped relation between body mass index (BMI) and incidence of disability pension, the relative risk (with the normal group as reference) among underweight men being 1.9. For the overweight subjects it was 1.3 and for the obese 2.8, all differences were significant. Diseases of the musculoskeletal and circulatory systems and mental disorders accounted for 67.2% of all main diagnoses resulting in disability pensions during follow-up. A total of 377 (6.4%) men died during follow-up. Diseases of the circulatory system, neoplasms, injury/poisoning and diseases of the respiratory system accounted for 91.8% of the deaths. Both underweight, overweight and obesity were related to risk of disability pension, with a J-shaped risk relationship.