FLEGAL ET AL. RESPOND

Abstract

Hu et al. speculate that the number of deaths attributable to obesity in the United States may be underestimated when relative risks are calculated on the basis of current body mass index (BMI). They cite no data to support their speculations, but instead invoke the notion of “reverse causality.” They hypothesize that relative risks are lowered by obese people who become ill, lose weight because of this illness to become normal weight, and die shortly thereafter of the underlying illness, surviving just long enough to be included in the study. However, this reverse-causation hypothesis is unlikely to be the correct explanation for the lower relative risks in the elderly for 2 reasons: first because data show that exclusion of preexisting illness has little effect on relative risk estimates,1 and second because weight loss from obesity to normal weight is relatively uncommon. Nationally representative data on measured weight change in US adults comes from the National Health and Nutrition Examination Survey (NHANES) I Epidemiologic Follow-up Study (NHEFS), in which body weight was measured on 2 occasions, approximately 10 years apart. Our analyses of these data show that among older adults (aged 65 years and older) with normal BMIs (18.5–24.9) in the early 1980s, only 2% had been obese (BMI ≥30) 10 years previously. Of older adults who were obese in the early 1970s, only 4% had normal BMIs 10 years later. The probability of large weight loss among obese adults is likely to be even lower today.2 Our analyses of NHANES 1999–2002 self-reported past-weight data for adults 65 years of age and older show that only 6% of non-obese older adults (BMI < 30) were obese 10 years previously and 2% of normal-weight older adults were obese 10 years previously. Hu et al. raise the same specter of reverse causation to assert that individuals with pre-existing illness should be excluded from cohorts used to derive relative risks applicable to the whole US population. However, these and other common exclusions (such as exclusion of the older elderly and of current or former smokers) not only make the remaining cohort less representative of the US population but in fact exclude the people more likely to die, thus making the deaths in the cohort even less representative of mortality in the US population. Relative risks based only on a subgroup of the US population defined by exclusions are not necessarily valid for inference to deaths outside that subgroup.