A recent article published in Population studies has changed our thoughts about obesity and health risks1. Indeed, it is one of the first studies directly analysing the relation between BMI and excess mortality. In this article, the author warns that BMI alone is an insufficient indicator, but also that current studies underestimate the inherent risks of a high BMI, introducing here the notion of “mortality bias”.
What we thought until now about BMI and excess mortality
Most articles published until now analyzed diseases associated with obesity, such as heart disease, cardiovascular diseases and diabetes, which have been effectively proven to be associated with excess mortality.
However, when it comes to analyzing directly BMI and excess mortality, landmark studies have evidenced that obesity elevates adult mortality risk only at extreme levels of body mass index. Some even consider that being overweight (BMI 25-30) significantly lowers mortality risk, but that above a BMI of 30, the risk is increased2. Other research suggests a near-linear relationship between high BMI and mortality risk and that over 15 per cent of adult deaths in the United States may be attributable to overweight and obesity3.
Challenges in estimating excess mortality when using BMI
Observational data pose several challenges to estimating the mortality consequences of overweight and obesity. The BMI-mortality association might reflect differences in sampling design, age composition, and availability of measures to control for important confounders, such as smoking status and physical activity, as well as patients’ individual weight stories4.
The new evidence uncovered by this publication
With a cohort of nearly 18 000 individuals and 4 468 deaths, the author has managed to demonstrate that the consequences of being overweight and obese have likely been underestimated, especially in older ages.
In his initial premise, the author argues that BMI can also be a confounding variable, generating survival bias (Also see our article “Is BMI a big fat lie?”). This survival bias can be positive, i.e. underestimating excess mortality, in higher BMIs due to recent weight gain, and can likewise be negative i.e. overestimating mortality in case of lower BMIs, specifically when it comes to recent weight loss.
The author also recommends that when it comes to studying the correlation between BMI and excess mortality, it is important to consider recent weight changes against BMI level and composition of BMI samples. This is particularly important when it comes to recent changes in weight, for example in recent illness-associated weight loss. As for weight gain, recent weight gain is considered less dangerous than being overweight or obese in the long term.
In conclusion, there is a bias when analyzing BMI against excess mortality, and other factors should be considered, requiring a detailed study of the weight history of the patient, in order to draw conclusions. Researchers should therefore be very cautious when using simply BMI in excess mortality studies.
What do you think about BMI? Should it still be used as a predictive tool or should we consider other measurements?
Reference
- Ryan K. Masters (2023) Sources and severity of bias in estimates of the BMI–mortality association, Population Studies, 77:1, 35-53
- McGee, D. L. and Diverse Populations Collaboration. (2005). Body mass index and mortality: A meta-analysis based on person-level data from twenty-six observational studies, Annals of Epidemiology 15(2): 87–97
- Berrington de Gonzalez, A., P. Hartge, J. R. Cerhan, A. J. Flint, L. Hannan, R. J. MacInnis, S. C. Moore, et al. (2010). Body-mass index and mortality among 1.46 million white adults, New England Journal of Medicine 363: 2211–2219.
- Abdullah, A., R. Wolfe, J. U. Stoelwinder, M. De Courten, C. Stevenson, H. L. Walls, and A. Peeters. (2011). The number of years lived with obesity and the risk of all- cause and cause-specific mortality, International Journal of Epidemiology 40(4): 985–996.
