The comparison of percent body fat estimated by different anthropometrics to predict the incidence of hypertension.

Percent body fat (%BF) is associated with the development of hypertension. However, the power of %BF estimated by different anthropometrics to predict incident hypertension was unknown. This study was from the China Health and Nutrition Survey (CHNS). %BF was calculated using the equations with BMI, WC, and skinfold thickness and divided into low and high %BF according to ROC. Cox regression was employed to evaluate the power of different %BFs to predict the development of hypertension. When not adjusting for covariates, %BFs defined by BMI, WC, and ST were the significant predictors of the development of hypertension (all P < 0.0001; crude HR: 2.238, 3.243, and 1.574; and HR 95% CI: 2.098-2.387, 2.905-3.619, and 1.464-1.692). When three %BFs entered into model simultaneously, the significance in %BF estimated by ST disappeared (P = 0.0765; adjusted HR: 1.124; and HR 95% CI: 0.988-1.280). For males, %BFs by BMI, WC, and ST significantly affected the incidence of hypertension as they were separately analyzed (all P < 0.0001; crude HR: 2.445, 2.335, and 1.828; and HR 95% CI: 2.220-2.693, 2.011-2.712, and 1.636-2.042, respectively). For females, %BFs estimated by BMI, WC, and ST were the determinants of the development of hypertension whether covariates were adjusted or not (all P < 0.0001). In conclusion, there was a poor and ineffective association of %BF estimated by triceps skinfold thickness with the development of hypertension, especially when three %BFs were analyzed together. High %BFs estimated by BMI and WC were the true and effective predictors of the incidence of hypertension.