Prediction capability of visceral adiposity index for hypertension

Objective To investigate the prediction capability of visceral adiposity index（VAI） for hypertension. Methods A total of 711 subjects were included from communities in Chengdu, Sichuan province, China, as a part survey of "Programme on trends and determinants of cardiovascular disease in a multinational population". All subjects underwent examinations including body height, weight, waist circumference, blood pressure and blood biochemical indicators in 1992 and 2007, respectively. Multivariate Cox regression analysis was used to assess the prediction capability of VAI, waist circumference and body mass index（BMI） for hypertension, and the receiver operating characteristic curve（ROC） were further performed to examine their discrimination. Results After adjusting for confounders, compared with the lowest quartile group, the hazard ratios（95%CI） of VAI, waist circumference and BMI in the highest quartile group were 1.535（1.075-2.193）, 2.245（1.550-3.251） and 2.385（1.673-3.398） for the incidence of hypertension, respectively.ROC analysis showed that compared with VAI area under the curve（AUC）=0.589, P＜0.001, waist circumference（AUC=0.649, P＜0.001） and BMI（AUC=0.648, P＜0.001） had better performance in discrimination for hypertension. Conclusions VAI can predict the incidence of hypertension, yet its discrimination for hypertension is not better than BMI and waist circumference.