Simpson JAM, R. J.; Peeters, A.; Hopper, J. L.; Giles, G. G.; English, D. R. (2007) A comparison of adiposity measures as predictors of all-cause mortality: the Melbourne Collaborative Cohort Study. Obesity (Silver Spring) 15:994-1003
OBJECTIVE: Our goal was to examine five different measures of adiposity as predictors of all-cause mortality. RESEARCH METHODS AND PROCEDURES: Subjects were 16,969 men and 24,344 women enrolled between 1990 and 1994 in the Melbourne Collaborative Cohort Study (27 to 75 years of age). There were 2822 deaths over a median follow-up period of 11 years. BMI, waist circumference, and waist-to-hip ratio were obtained from direct anthropometric measurements. Fat mass and percentage fat were estimated by bioelectric impedance analysis. RESULTS: Comparing the top quintile with the second quintile, for men there was an increased risk of between 20% and 30% for all-cause mortality associated with each of the anthropometric measures. For women, there was an increased risk of 30% (95% confidence interval for hazard ratio, 1.1-1.6) observed for waist circumference and 50% (1.2-1.8) for waist-to-hip ratio, but little or no increased risk for BMI, fat mass, and percentage fat. Waist-to-hip ratio was positively and monotonically associated with all-cause mortality for both men and women. There was a linear association between waist circumference and all-cause mortality for men, whereas a U-shaped association was observed for women. DISCUSSION: Measures of central adiposity were better predictors of mortality in women in the Melbourne Collaborative Cohort Study compared with measures of overall adiposity. We recommend measuring waist and hip circumferences in population studies investigating the risk of all-cause mortality associated with obesity. The use of additional measures such as bioelectric impedance is not justified for this outcome.