Relationship between novel anthropometric indices and the incidence of hypertension in Chinese individuals: a prospective cohort study based on the CHNS from 1993 to 2015

Xin Zhang, Runyu Ye, Lirong Sun, Xueting Liu, Si Wang, Qingtao Meng, Xiaoping Chen

Recently, novel anthropometric indices (AHIs), including the body roundness index (BRI) and a body shape index (ABSI), were proposed to evaluate a subject's nutritional status and metabolic disorders. In the present study, we mainly analyzed the relationship between AHIs and the incidence of hypertension and preliminarily compared their abilities to discriminate hypertension incidence in the Chinese population from the China Health and Nutrition Survey (CHNS). A total of 12,154 participants were included in this longitudinal study. The age range of this cohort was 18-94 years old (mean age: 40.73 ± 13.85 years old). 4511 participants developed hypertension during a median of 7.00 years of follow-up. Cox regression analysis, stratified analysis, and interaction tests were used to analyze the relationship between AHIs and the incidence of hypertension. Time-dependent receiver operating characteristic (ROC) curves, integrated discrimination improvement (IDI) and net reclassification index (NRI) were calculated to appraise the AHIs' discrimination value of new-onset hypertension. Kaplan‒Meier curves demonstrated that the participants in higher quartiles of AHIs (ABSI or BRI) at baseline were at greater risk of hypertension incidence during the follow-up. After adjusting for confounding factors, multivariate Cox regression models showed that the quartiles of BRI were significantly associated with an increased risk of hypertension in the whole cohort but were relatively weak for ABSI quartiles (P for trend = 0.387). In addition, ABSI z score (HR = 1.08, 95% CI: 1.04-1.11) and BRI z score (HR = 1.27, 95% CI: 1.23-1.30) were positively associated with increased incident hypertension in the total population. Stratified analysis and interaction tests showed a greater risk of new-onset hypertension in those < 40 years old (HR = 1.43, 95% CI: 1.35-1.50) for each z score increase in BRI and a higher incidence of hypertension in participants who were drinkers (HR = 1.10, 95% CI: 1.04-1.14) for each z score increase in ABSI. In addition, we observed that the area under the curve for identifying hypertension incidence for BRI was significantly higher than that for ABSI at 4, 7, 11, 12, and 15 years (all P < 0.05). However, the AUC of both indices decreased over time. Furthermore, the addition of BRI improved the differentiation and reclassification of traditional risk factors with a continuous NRI of 0.201 (95% CI: 0.169-0.228) and an IDI of 0.021 (95% CI: 0.015-0.028). Increased ABSI and BRI were associated with an increased risk of hypertension in Chinese individuals. BRI performed better than ABSI in identifying the new onset of hypertension, and the discrimination ability of both indices decreased over time.