Epidemiol Health.  2024;46(1):e2024066. 10.4178/epih.e2024066.

Comparative analysis of body mass index and obesity-related anthropometric indices for mortality prediction: a study of the Namwon and Dong-gu cohort in Korea

• Kim YR
• Shin MH
• Lee YH
• Choi SW
• Nam HS
• Yang JH
• Kweon SS

• KMID: 2564568
• DOI: http://doi.org/10.4178/epih.e2024066

Abstract

OBJECTIVES
This study investigated the associations between several obesity-related anthropometric indices and mortality in middle-aged and elderly populations to compare the indices’ predictive ability with that of the body mass index (BMI).
METHODS
We analyzed data on 12 indices calculated from 19,805 community-based cohort participants (average age, 63.27 years; median follow-up, 13.49 years). Each index was calculated using directly measured values of height, weight, waist circumference (WC), and hip circumference (HC). We calculated hazard ratios (HRs) and 95% confidence intervals (CIs) for each index using Cox regression and evaluated mortality prediction with the Harrell concordance index (c-index).
RESULTS
Adding anthropometric indices to the basic mortality model (c-index, 0.7723; 95% CI, 0.7647 to 0.7799) significantly increased the predictive power of BMI (c-index, 0.7735; 95% CI, 0.7659 to 0.7811), a body shape index (ABSI; c-index, 0.7735; 95% CI, 0.7659 to 0.7810), weight-adjusted waist index (WWI; c-index, 0.7731; 95% CI, 0.7656 to 0.7807), and waist to hip index (WHI; c-index, 0.7733; 95% CI, 0.7657 to 0.7809). The differences between the BMI model and the other 3 models were not statistically significant.
CONCLUSIONS
In predicting all-cause mortality, the ABSI, WWI, and WHI models based on WC or HC had stronger predictive power than conventional risk factors but were not significantly different from the BMI model.

Keyword

Obesity; Weights and measures; Mortality