Skip to content

Abstract Search

Nutrition/Obesity

Comparing anthropometric measures from the Cancer Prevention Study II Nutrition Cohort Kierstin Faw* Kierstin Faw Emily Deubler Lauren R Teras Alpa V Patel

Introduction: Recent commentaries question whether body mass index (BMI) is a good proxy for excess body fatness in predicting disease risk. Other anthropometric measures that may better measure visceral fat such as, waist circumference (WC) or waist to height ratio (WtHR), have been postulated to better predict disease risk. This study aims to understand the relationship between BMI, WC, and WtHR in a large population of men and women.

Methods: In the Cancer Prevention Study II Nutrition Cohort, self-reported height was collected in 1992 and self-measured weight and WC were ascertained in 1997 among cancer-free individuals and used to calculate BMI, WtHR, and WC. These measures were compared continuously, using correlation coefficients (r) and categorically, using weighted kappa statistics (k, percent agreement). The WtHR categories were <0.50, 0.50-0.55, and ≥ 0.55. BMI categories were 18.5-25 (healthy weight), 25-30 (overweight), and ≥ 30 kg/m2 (obese). WC cut points were <31.25, 31.25-35.75, and ≥ 35.75 cm for women and <36.75, 36.75-40.00, and ≥ 40.00 cm for men.

Results: Among women (N= 50,618), based on BMI, 50% were healthy weight, 33% were overweight, and 17% were obese. Continuous BMI was strongly associated with WtHR and WC (r=0.81; r=0.79). The kappa statistics revealed similar results with categorical BMI having moderate agreement with WtHR and WC (k=0.54; k=0.49). Among men (N=43,783), based on BMI, 36% were healthy weight, 49% were overweight, and 15% were obese. Continuous BMI was strongly correlated to WtHR and WC (r=0.81; r=0.78). Categorical BMI had fair to moderate agreement with WtHR and WC (k=0.34; k=0.47).

Conclusion: In the Cancer Prevention Study II Nutrition Cohort, continuous BMI, WtHR, and WC were all strongly correlated and categorical measures had fair to moderate agreement for men and women. Although BMI is a proxy for general adiposity, this study suggests it’s generally a reliable, easy to measure way to ascertain body size.