The Impact of Body Mass Index on All-Cause Mortality Worldwide: A Mathematical Demonstration of Clinical Bias Regarding Causal Risk from Body Composition Measures

Body mass index (BMI) defined obesity as characterized by excess fat mass (FM). In associating BMI with mortality, different J-shaped curve patterns have been repeatedly observed. However, BMI refers not only to FM but also to fat-free mass (FFM). We aimed to demonstrate that differences between FM and FFM can mask the use of any mean BMI. Novel mathematical and anthropometric considerations about relationships between FM and FFM were explored. A review of studies on millions of participants associating BMI and newer variables, such as the ratios and mean differences between the two components, with all-cause mortality was conducted. The means of all variables were newly pooled in each of six BMI categories, separately by sex. If, after checking for associations, systematic differences between FM and FFM remained, this indicated that body components differed, and causality could not be assumed. A linear dose–response analysis of FM and FFM with all-cause mortality was found. A significant difference in all associations was demonstrated ( <0.001). A risk equivalence between BMI and FFM only in underweight was demonstrated in either sex (FFM >90%; FM <10%; <0.001). In women with class III obesity, the same sensitivity for BMI and excess FM was proven (FM% > FFM%; < 0.001). In all means of BMI for the associations FM and FFM were significantly different ( <0.001). In any BMI-associated risk, a protective overestimate of FFM over FM consistently occurred. In all BMI cut-offs repeatedly observed for defining population risk, the true causal risk always appeared biased confusing clinicians for a long time.