To the Editor: In the study by Dr Flegal and
colleagues,1 it seems likely that their estimates
of excess mortality due to overweight and obesity would have been substantially
higher, and their estimates of excess mortality due to underweight would have
been substantially lower, if they had accounted for 2 confounders that they
did not address. The first is caused by elderly persons tending to lose weight
prior to death.2 The second is caused by potential
regression-to-the-mean patterns in the BMI data.2,3 Corrections
for these confounders can be made with longitudinal data on body weight, which
are available in the NHANES I Epidemiologic Follow-up Study (NHEFS) database.
This approach was used successfully by the Prospective Studies Collaboration
to adjust for confounders in order to estimate mortality associated with different
levels of blood pressure.3 It would be valuable
if Flegal et al could use the NHEFS data to apply these corrections to their
original estimates of excess mortality.
Underweight, Overweight, Obesity, and Excess Deaths. JAMA. 2005;294(5):551-553. doi:10.1001/jama.294.5.552-a