To establish the actual serial changes in body weight in Japanese people and to elucidate the influence of changes in BMI on morbidity, we conducted a historical cohort study of university graduates from 1955 to 1990 using questionnaires and BMI data. The subjects of this study were 3,675 university graduates aged 26-62 years in whom BMI was determined at the time of enrollment in the university (Pre-BMI), 5 to 40 years earlier. Morbidity (one or more system diseases or obesity-related system diseases) was analyzed according to current age, sex, current BMI, deltaBMI (difference between current BMI and pre-BMI), and various lifestyle variables. The proportion of overweight subjects at enrollment to university was higher in recent male students compared to old students, but not in female graduates, and the BMI in both genders increased progressively after graduation, especially in recent male graduates. Pre-BMI correlated negatively and significantly with deltaBMI. The percentages of obese (BMI > or = 30 kg/m2) males and females were 1.6% and 0.5%, respectively, and high morbidity was observed in 56.1% and 42.2% of males and females, respectively. Stepwise regression analysis showed that in subjects with normal BMI at enrollment, prospective morbidity was dependent on ABMI in addition to age. Our results indicate that in subjects with normal body weight, prospective morbidity is determined by increment of ABMI, and suggest that maintenance of BMI at the late adolescence level is an important factor in preventing future disease.