A longitudinal study by Stessman et al of more than 1000 individuals indicated that a normal BMI, rather than obesity, is associated with a higher mortality rate in elderly people. The investigators determined that a unit increase in BMI in female members of the cohort could be linked to hazard ratios (HRs) of 0.94 at age 70 years, 0.95 at age 78 years, and 0.91 at age 85 years.

In men, a unit increase in BMI was associated with HRs of 0.99 at age 70 years, 0.94 at age 78 years, and 0.91 at age 85 years. According to a time-dependent analysis of 450 cohort members followed from age 70 to age 88 years, a unit increase in BMI produced an HR of 0.93 in women and in men.

Similar results were found in a Japanese study of 26,747 older persons (aged 65-79 years at baseline). Tamakoshi et al found no elevation in all-cause mortality risk in overweight (BMI 25.0-29.9) or obese (BMI ≥30.0) men; slightly elevated hazard ratios were found in women in the obese group, but not in the overweight group, compared with women in the mid–normal-range group. In contrast, an association was found between a low BMI and an increased risk of all-cause mortality, even among persons in the lower-normal BMI range.