Body Weight and Risk of Early Death


  • P.W. Franks

    Corresponding author
    1. Department of Nutrition, Harvard School of Public Health, Boston, Massachusetts, USA
    2. Genetic Epidemiology & Clinical Research Group, Department of Public Health & Clinical Medicine, Section for Medicine, Umeå University, Umeå, Sweden
    • Department of Clinical Sciences, Genetic and Molecular Epidemiology Unit, Lund University Diabetes Center, Skåne University Hospital Malmö, Malmö, Sweden
    Search for more papers by this author

  • Disclosure: The authors declared no conflict of interest.

Correspondence: P. W. Franks (

Although most acknowledge that the relationship between body mass index (BMI) and mortality is J-shaped in its distribution, where the nadir of this curve resides is contentious. Many evidence-based clinical practice guidelines and position statements identify the lowest risk BMI category as 22.0-24.9 kg/m2 in non-Asian populations (1-3). However, in a recent systematic review and meta-analysis of 2.88 million adults, of whom 270,000 died during follow-up, Flegal et al. [4] challenged this view, concluding that mortality rates are lowest in persons who would be clinically defined as “over-weight” (BMI of 25.0-29.9 kg/m2). Flegal et al.'s study was widely publicized through news headlines that emphasized the controversy of the study's conclusions [5] and triggered vigorous debates (6-8) about the veracity of the results and the impact they might have on how clinicians and the public view overweight as a risk factor for morbidity and premature mortality.

BMI is a proxy measure of adiposity, used widely because of its strong correlation with adiposity [9] and because it can be calculated from the easily and accurately obtained measures of height and weight. The amount, distribution, composition, and subtype of adipose tissue determine the quantity and type of secreted peptide hormones (often termed “adipokines”) and the consequential impact of adiposity and adipokines on health and disease. Adipokines regulate many important processes including physical maturation, cognitive development, fecundity, cellular signaling, and bone health. Adipose tissue also provides a sink for blood lipids and a reserve for energy. The complete or near-complete absence of adipose tissue is associated with life-threatening complications and premature death [10]. Conversely, a chronic excess of adipose tissue is also detrimental to health and longevity [1, 3].

Because obtaining detailed measures of adiposity at a population-level is impracticable, surrogate anthropometric measures such as BMI are typically favored. Studies that define where on the distribution of BMI the risk of disease and death begin to rise are needed for accurate clinical practice guidelines focusing on the prediction and prevention of these diseases; for this reason, Flegal et al.'s study [4] and others on this topic are important. No less important, though, are the pragmatic discussions and debates that should accompany all widely publicized, controversial medical findings. This may be especially true when the findings emanate from major, well-respected public health organizations such as the Centers for Disease Control and Prevention (CDC), as did Flegal et al.'s study [4].

In this edition of the Journal, we publish two Perspective articles [11, 12] that discuss the methods, results, and conclusions of studies on BMI and mortality, with a third Perspective describing the relevant concept of “Metabolically Healthy Obesity” [13]. These perspectives offer valuable insights that aid in the interpretation of Flegal et al.'s work [4] and other related studies, whilst also reminding us why focusing on population-level data on body size and mortality can sometimes conceal important subgroup effects, knowledge of which might prove useful for the clinical translation of obesity research.


The author thanks Dr. Jeremy Pomeroy (Phoenix Epidemiology & Clinical Research Branch, NIDDK) for his helpful comments on this editorial.