Body mass index and disease-specific mortality in an 80-year-old population at the 12-year follow-up.

Although many investigations examined the relationship between body mass index (BMI) and mortality, little is known about the possible associations between BMI and disease-specific mortality in very elderly people. Here we evaluated this association in an 80-year-old population. In 1998, 675 residents in Japan's Fukuoka Prefecture participated. They were followed up for 12 years after the baseline examination; 37 subjects (5.5%) were lost to follow-up. The subjects were divided into six groups by their BMI values: <19.5 (most-thin), 19.5 to <21.1 (relatively thin), 21.1 to <22.5 (thin/normal), 22.5 to <23.8 (normal/overweight), 23.8 to <26.0 (relatively obese), ≥26.0 (most-obese). The most-thin group had the highest mortality from all-causes, and from respiratory disease. The normal/overweight group had the lowest overall mortality among the six BMI groups. These associations were found in the men, but not in the women. The most-obese group did not have higher mortality from all-causes or cardiovascular disease compared to the normal/overweight group. Respiratory disease-related mortality was lowest in the most-obese group. No association was found between BMI group and mortality from cancer. In conclusion, in an 80-year-old Japanese population, mortality from all-causes or respiratory disease was highest in the most-lean group (BMI <19.5), and mortality from all-causes or cardiovascular disease was lowest in the group with BMI 22.5 to <23.8.