Many of the observational epidermiologic studies suggest that the relationship between BMI and mortality weakens with increasing age, especially among persons aged 75 and above (287-290).
Several factors have been proposed to explain this observation. Older adults are more likely than younger adults to have diseases that both increase mortality and cause weight loss leading to lower body weight (291-293). In addition, as people age, they tend to have larger waist circumferences that increase their risk of mortality even at lower BMIs (294). Also, weight in middle age is positively related to risk of mortality in old age (292). The impact of smoking on body weight and mortality is likely to be much stronger in older adults because of the cumulative health effects of smoking (295).
BMI, which is an indirect estimate of adiposity, may underestimate adiposity in older adults whose BMI is similar to younger adults' (296). It is also possible that persons most sensitive to the adverse health effects of obesity are more likely to have died before reaching older ages, resulting in older cohorts that are more "resistant" to the health effects of obesity. Recently, a 20-year prospective study of a nationally representative sample of U.S. adults aged 55 to 74 years suggested that lowest mortality occurs in the BMI range of 25 to 30 (297, 298). After adjusting for smoking status and pre-existing illness, lowest mortality occurred at a BMI of 24.5 in white men, 26.5 in white women, 27.0 in black men, and 29.8 in black women.