Obesity and life expectancy Lennert Veerman, The University of Queensland 9 June 2011 This is a very interesting paper, and properly taking into account the disability that is associated with the approaching end of life is a worthwhile innovation. The only reservation I have is with respect to the results for overweight and obesity and their impact on life expectancy. The authors discuss the 'obesity paradox', but they could have discussed one more aspect of it: the fact that BMI tends to decrease as a consequence of chronic health problems ('reverse causality'). Many studies have found that overweight and even mild obesity at older age is associated with a lower mortality risk. I think those many studies are wrong to interpret this as an argument that obesity is not very lethal or even improves survival. At old age, people with ‘normal’ BMI will consist of a mix of people who always had a BMI in the 18.5-25 range and people who used to be heavier but lost weight due to (detected or undetected) disease. Conversely, obese persons are a selection of formerly obese out of which some persons with chronic, ultimately lethal disease have been removed. Notoriously, a high BMI indicates a lower risk of dying from chronic respiratory disease. Some studies adjust for baseline morbidity and exclude the first years of mortality in an effort to reduce the bias this gives (although I am not aware of any study that adjusted for respiratory capacity at baseline), but neither strategy was employed in this study. Smokers also tend to have a lower BMI and an increased risk of death. That was not adjusted for in this study, either, again leading to underestimation of the effect of obesity on life expectancy. Studies with longer follow up (e.g. ) tend to find a larger effect of high BMI on mortality, and a lower ideal weight. The follow-up time in this study is rather short (about 8 years), which would lead to a rather large reverse causality effect. A study using offspring BMI as the exposure variable, a method that avoids reverse causality, did not confirm the survival benefit of obesity at higher age and suggests the apparent beneficial effects of excess weight found in other studies are due to reverse causality . Notably, this offspring-BMI study did not find that obesity protects from dying of chronic lung disease. In sum, I think the results with respect to body mass index and life expectancy should be disregarded. That does not diminish the importance of the conclusion that obesity disables, smoking kills, and that a drink or two a day improves life. References 1. Peeters A, Barendregt JJ, Willekens F, Mackenbach JP, Al Mamun A, Bonneux L: Obesity in adulthood and its consequences for life expectancy: a life-table analysis. Ann Intern Med 2003, 138:24-32. 2. Davey Smith G, Sterne JA, Fraser A, Tynelius P, Lawlor DA, Rasmussen F: The association between BMI and mortality using offspring BMI as an indicator of own BMI: large intergenerational mortality study. Bmj 2009, 339:b5043. Competing interests I have no competing interests.