Description
Background With the global rise in obesity, measuring the adiposity status of individuals is of growing importance, especially across different ethnicities and nationalities. An increasing amount of research is being performed to determine which adiposity measures are better in screening for cardiometabolic risk factors and diseases. However, there are conflicting results. This cross-sectional study was designed to compare the associations of body mass index (BMI), waist circumference and waist-to-height ratio with hyperglycemia, hypertension and hypercholesterolemia in a community sample of Mexican adults. The study was also intended to determine optimal cutoff values for the adiposity measures in predicting at least one cardiometabolic risk factor. Methods Participants were Mexican adults residing in or near the border city of Tecate who attended a yearly community health fair between 2005 and 2009. Linear regression and logistic regression were performed to assess the association between each standardized adiposity measure and each cardiometabolic outcome. Receiver operating characteristic (ROC) curve analysis was employed to compare the adiposity measures in how well they predicted each outcome and to determine optimal cutoff values for the adiposity measures. Results Of the 1523 observations, 760 were unique adult observations, 266 of which contained enough information to include for analysis (98 men, 168 women). BMI, waist circumference and waist-to-height ratio were similarly statistically associated with log blood glucose in men (p=0.001, p=0.029, p=0.004, respectively). BMI was significantly associated with mean arterial pressure in women (p=0.013). The area under the curve (AUC) for BMI was significantly higher than the AUCs for waist circumference and waist-to-height ratio in predicting hypercholesterolemia (p=0.016, p=0.026, respectively). The AUC for waist-to-height ratio was significantly higher than the AUC for waist circumference in predicting hypertension in women (p=0.008). The optimal cutoff values for predicting at least one cardiometabolic outcome were a BMI of 28.8 kg a waist circumference of 91.8 cm for men and 91.6 cm for women, and a waist-to-height ratio of 0.551 for men and 0.583 for women. Sensitivities for these cutoffs ranged from 45 to 58% and specificities ranged from 55 to 75%. Discussion BMI tended to perform marginally better than either waist circumference or waist to height ratio for being associated with the outcomes, but was not significantly or consistently so. The confidence intervals for the statistics measuring the strength of association overlapped in each model, suggesting that the adiposity measures were similarly associated with the outcomes. However, the lack of strength and consistency of these associations precluded definite results. The adiposity measure cutoff values derived from this study were similar to published ranges but showed only moderate predictive value.
