Over the years ideal or desirable weights have been associated with th
e lowest mortality and defined in a number of ways. The widely used he
ight-weight tables of the Metropolitan Life Insurance Company, develop
ed in the 1940s, have been supplanted in the last decade by new weight
standards based on findings of several population-based studies that
compared body weight to mortality. The Quetelet or body mass index (BM
I), now used as the de facto criterion for defining a desirable weight
index, indicates relative fatness and is only minimally correlated wi
th height. However, the optimal BMI or weight for longevity remains to
be defined for a number of methodological reasons. This article revie
ws the strength of the evidence for increased mortality in adults who
are overweight or underweight based on standard BMI ranges and in thos
e who reported a change in body weight. Epidemiological studies show t
hat excess body weight is associated with increased mortality, dependi
ng on fat patterning, gender, and age. A similar increase is shown for
subjects with body weight below the relative weight range, although h
ere associated life-style factors are greater contributors. Preliminar
y data suggest that change in body weight may also be associated with
increased mortality. Overall, carefully measured weight and height rem
ain the most easily performed and useful determinants of nutritional s
tatus and predictors of mortality for the general population.