Hypertension and Obesity: A Cross-National Comparison of BMI and Waist-Height Ratio

Hypertension has been identified as a prominent co-morbidity of obesity. To improve clinical intervention of hypertension, it is critical to identify metrics that most accurately reflect risk for increased morbidity. Two of the most relevant and accurate measures for increased risk of hypertension due to excess adipose tissue are Body Mass Index (BMI) and Waist-Height Ratio (WHtR). Previous research has examined these measures in cross-national and cross-ethnic studies, but has most often relied on secondary means such as meta-analysis to identify and evaluate the efficacy of individual body mass measures. In this study, we instead use cross-sectional analysis to assess the cross-ethnic discriminative power of BMI and WHtR to predict risk of hypertension.

Using data from WHO SAGE surveys, which include anthropometric and biometric data from people in six middle-income countries (China, Ghana, India, Mexico, Russia, South Africa), we implement logistic regression to examine the discriminative power of measured BMI and WHtR for hypertension risk.  We stratify by gender and age to identify whether optimum cut-off points that are adequately sensitive as tests for risk of hypertension may differ between groups. We report results for OR, RR, and ROC curves for each age/sex group in each of the six SAGE countries.

We find that cut-off points for WHtR are relatively dependent upon gender, age and country. While an optimum omnibus cut-point for WHtR may be 0.55, our results suggest that the gender and age relationship with WHtR warrant the development of sub-group specific cut-offs to optimize health outcomes. Between-country comparisons indicate that clinical outcomes would be significantly impacted if aggregate WHtR cut-points were used. Trends through multiple countries further show that the optimum cut-point for WHtR increases with age while the area under the curve (AUROC) decreases with age, for both men and women. Comparison between BMI and WHtR indicate that BMI may remain more robust as a risk for hypertension than WHtR. Implications for public health policy are discussed.