Influence of age on the association between various measures of obesity and all-cause mortality

To determine whether the association between various simple measures of obesity and risk for all-cause mortality differs between younger and older men and women.

Prospective cohort study with 8.7 +/- 0.2 years of follow-up for mortality linkage.

Third National Health and Nutrition Examination Survey, 1988 to 1994.

Four thousand, four hundred thirty-seven men and 5,166 women.

Measures of obesity included body mass, waist circumference, waist-to-hip ratio, hip circumference, sum of skinfolds, and bioelectrical impedance.

Overall and abdominal obesity are associated with greater mortality risk in younger adults (<65) (P<.05), whereas the associations between obesity and mortality are null or inverse in older adults (>65). In general, the association was stronger with measures of abdominal obesity than with measures of overall obesity or fat-free mass.

The adverse effects of obesity on mortality risk are apparent only in adults younger than 65. Obesity as characterized using several different measures was not generally associated with greater mortality risk in older adults. Although weight loss is beneficial for reducing morbidity in obese adults of any age, it is unclear whether weight loss is equally beneficial for reducing mortality risk in older adults.