Body Volume Index, a novel obesity metric, and mortality: comparison with established obesity measures

BACKGROUND Obesity is a major public health concern, yet conventional anthropometric indicators such as body mass index (BMI) show inconsistent associations with mortality, including the obesity paradox. OBJECTIVE This study introduces the Body Volume Index (BVI)—a novel 3D ellipsoid shape-based obesity indicator capturing central body dimensions relative to body mass—and evaluates its performance in predicting mortality against established measures. METHODS We analyzed data from 288,331 Korean adults in the National Health Insurance Service–National Sample Cohort (2009–2019). Five obesity indicators—BMI, waist-to-height ratio (WHtR), a body shape index (ABSI), body roundness index (BRI), and BVI—were categorized into sex-specific deciles (Q1-Q10, with Q5 as the reference). Associations with all-cause and cancer-specific mortality were assessed using multivariable Cox proportional hazards models adjusted for sociodemographic factors, health behaviors, and comorbidities. RESULTS In fully adjusted models, BMI showed a reverse J-shaped association with all-cause and cancer-specific mortality. WHtR and BRI paradoxically exhibited inverse or non-monotonic associations with mortality outcomes (e.g., for WHtR among men, adjusted hazard ratio [aHR] of all-cause mortality:1.27, 95% CI 1.13–1.42 for Q1 vs. Q5). Conversely, ABSI and BVI showed consistent positive linear trends with all-cause mortality in both sexes (e.g., for men in Q10 vs. Q5, aHR 1.58, 95% CI 1.41–1.77 for ABSI; and aHR 1.20, 95% CI 1.09–1.32 for BVI). Although ABSI showed steeper gradients for all-cause mortality than BVI, its association with cancer mortality was more variable. CONCLUSIONS BVI demonstrated stable positive associations with mortality outcomes alongside ABSI. Given its straightforward geometric formulation and wider numerical variability, BVI may offer favorable practical feasibility for risk stratification in large-scale epidemiologic contexts.