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 | BMC Public Health

Study population

The longitudinal data of participants in the present study were from the CHNS. The CHNS is a nationwide survey on risk factors, nutrition, and health-related outcomes in the Chinese population from 15 provinces and autonomous districts, including Beijing, Chongqing, Guangxi, Guizhou, Heilongjiang, Henan, Hubei, Hunan, Jiangsu, Liaoning, Shaanxi, Shandong, Shanghai, Yunnan, and Zhejiang. The CHNS is an ongoing open cohort and an international collaborative project announced in 1989 and subsequently conducted in 1991, 1993, 1997, 2000, 2004, 2006, 2009, 2011, and 2015. In this project, a multistage, random cluster process was used to draw the samples surveyed in each of the provinces. Details of the cohort and sampling process have been published elsewhere [17].

To analyze the correlation between AHIs and the risk of hypertension, we used CHNS data from 1993 to 2015. Waist circumference (WC) and hip circumference (HC) were not collected in 1989 and 1991. Adults aged ≥ 18 years at the first wave with integrated data on sex, anthropometric indicators (including height, weight, WC, and WC), systolic blood pressure (SBP), diastolic blood pressure (DBP), ethnicity, and smoking and drinking status were suitable for this analysis. Participants with the following criteria were excluded: those with missing data for the abovementioned indices; those who had just one medical record over the years; those who were pregnant or lactating at the time of the survey; and those with extreme values (e.g., height < 120 cm; weight > 150 kg; WC < 50 cm or > 150 cm; and HC < 50 cm or > 150 cm). In addition, participants with hypertension at baseline (including those who had a self-reported diagnosis of hypertension, antihypertensive medication, and average SBP ≥ 140 mmHg or DBP ≥ 140 mmHg) were also excluded. Stata/SE version 15.1 was used for original data merging, calculation, cleaning and conversion before the statistical analysis, according to the inclusion and exclusion criteria as previously mentioned.

Demographic parameters

Information on age, gender, race (Chinese Han or other), urban or rural residence, and current smoking and drinking status were gathered from the questionnaires at each follow-up survey. Current smoking was defined by whether participants themselves reported that they had still smoked cigarettes at each survey. Nonsmokers and former smokers were defined as noncurrent smokers. Drinking status was evaluated according to the frequency of alcohol consumption by self-report. In the present study, a participant who did not drink in the past year was defined as nondrinking, and the others (including drinking daily, 3–4 times/week, 1–2 times/week, and less than once a week) were defined as drinking status.

Blood pressure measurements

The trained staff performed blood pressure measurements following the standard protocol and using appropriately sized cuffs at each follow-up survey. Before measurements, all participants were required to have a 10-min seated rest. Triplicate measurements of blood pressure on the right arm were conducted by using mercury sphygmomanometers, with at least 1 min between recordings [18]. The average of the three blood pressure measurements was calculated for the final analysis.

Anthropometric measurements

The anthropometric measurements were administered by well-trained research staff in a private and comfortable room. All participants were requested to remove bulky clothing and shoes before measurement. Weight, height, WC, and HC were obtained using the calibrated equipment according to a standard procedure. ABSI was estimated as the WC divided by the BMI raised to two-thirds and by the square root of the height. In addition, the BRI was based on WC and height.

The specific formulas of ABSI and BRI are as follows [8, 9]:

Study outcome

The outcome of the present study was new-onset hypertension during follow-up. Participants were identified if they had an average SBP ≥ 140 mmHg or DBP ≥ 90 mmHg, self-reported a diagnosis of hypertension, or were currently taking antihypertensive at any one of the follow-up visits [19, 20]. The first time for diagnosis with hypertension was considered the time when the end event occurred. For those free of hypertension in all follow-up surveys, the final survey date was used to calculate the follow-up time.

Statistical methods

All statistical analyses were performed using R software (Version 4.2.1). Continuous data with a normal distribution were expressed as the mean ± standard deviation (SD). Categorical variables were expressed as the frequency. Continuous data were compared using the independent-samples t test. Differences in categorical variables were compared among the groups using the chi-squared test. The AHIs were converted into z scores and quartiles. In the etiological analysis section, Kaplan-Meier curves were used to evaluate the cumulative incidence for AHIs categories, and the log-rank test was utilized to examine the significance of the differences between groups. Univariate and multivariate Cox regression models were applied to analyze the association between each anthropometric measurement and the incidence of hypertension. The confounders in multivariate Cox regression were selected as those with significance in univariate analysis or reported by previous studies. Stratified analysis and interaction tests were conducted according to age (< 40 and > = 40 years old), sex, ethnicity (Chinese Hans and non-Hans), residence (urban and rural), current smoking status, current drinking status, SBP level (< 120 and > = 120 mmHg), and DBP level (< 80 and > = 80 mmHg). The hazard ratio (HR) of hypertension incidence and the 95% confidence intervals (95% CI) were calculated. In addition, we preliminarily analyzed the discrimination of hypertension incidence for ABSI and BRI, respectively. First, the ability to discriminate the incidence of hypertension was compared between ABSI and BRI using time-dependent receiver-operating characteristic (ROC) curve analysis. Second, owing to not collecting hypertension family history and serum biochemical indices in the early stage of CHNS, we could not directly use the existing prediction model in the literature. Furthermore, the integrated discrimination improvement (IDI) and net reclassification index (NRI) were calculated to appraise the incremental discrimination value of new-onset hypertension beyond the traditional factors based on age, sex, ethnicity, residence, smoking, drinking, SBP and DBP.