Abstract
BACKGROUND: Obesity is a well-established independent risk factor for hypertension and other cardiometabolic disorders. However, the best anthropometric index of obesity that predicts or associates strongly with hypertension and related conditions remains controversial and inconclusive. OBJECTIVE: This study compared the performance of eight anthropometric indices of obesity: body mass index (BMI), ponderal index (PI), waist circumference (WC), hip circumference (HC), waist-hip ratio (WHR), waist-height ratio (WHtR), body adiposity index (BAI) and conicity index (CI) as correlates and potential predictors of risk of hypertension and prehypertension in a Nigerian population, and also the possible effect of combining two or more indices in that regard. METHODS: This church-based, cross-sectional study was conducted in Anambra state, south-eastern Nigeria from 2012 to 2013. A total of 912 persons (436 male and 476 female) drawn randomly from three major cities (Awka, Onitsha and Nnewi) in the state participated in the study. Information on demography, medical history and lifestyle were obtained using a well-structured and validated questionnaire. The systolic/diastolic blood pressure and anthropometric measurements were taken by well-trained personnel. The resulting data were analysed using descriptive statistics, logistic regression, Poisson regression and receiver operating characteristic curve analysis. RESULTS: The mean values of all the anthropometric indices studied increased from normotension, through prehypertension to hypertension in both genders. BMI, WC, HC and CI were significantly higher (p < 0.05) in females than males. All the anthropometric indices studied were significantly (p < 0.001 except for CI) correlated with systolic and diastolic blood pressure. BMI, WHtR, WC and PI (with higher correlation coefficients for blood pressure) showed the best potential to predict hypertension and prehypertension in the study: BMI (cut-off = 24.49, AUC = 0.698; cut-off = 23.62, AUC = 0.659), WHtR (cut-off = 0.55, AUC = 0.682; cut-off = 0.5, AUC = 0.636), WC (cut-off = 91.44, AUC = 0.692; cut-off = 82.55, AUC = 0.645), PI (cut-off = 14.45, AUC = 0.670; cut-off = 13.69, AUC = 0.639), in males; and BMI (cut-off = 24.44, AUC = 0.622; cut-off = 28.01, AUC = 0.609), WHtR (cut-off = 0.51, AUC = 0.624; cut-off = 0.6, AUC = 0.572), WC ( cut-off = 96.62, AUC = 0.616; cut-off = 96.52, AUC = 0.584), PI ( cut-off = 16.38, AUC = 0.619; cut-off = 17.65, AUC = 0.599), in females for hypertension and prehypertension, respectively. In predicting hypertension risk, WC and WHtR did not significantly improve the performance of BMI in the models when included using our decision rule. Overall, CI had a very poor discriminatory power for both conditions in this study. CONCLUSION: BMI, WHtR, WC and PI emerged the best predictors of hypertension risk, and BMI, WC and PI of prehypertension risk in this study. The combination of high-performing anthropometric indices in a model did not improve their performance. Therefore we recommend the simultaneous but independent use of BMI and either WC or WHtR for predicting hypertension, and BMI and WC for prehypertension risk, bearing in mind that both types of index (abdominal and general obesity) account for different forms of obesity.