Obesity-and lipid-related indices as a predictor of hypertension in Mid-aged and Elderly Chinese: A Cross-sectional Study

Objective Middle-aged and elderly people in China probably suffer from hypertension. There is a close relationship between obesity-and lipid-related index and hypertension, which is recognized by recent studies. However, these studies have not systematically compared the relationship between the two. We aim to find the most effective obesity-and lipid-related index for predicting hypertension. Method A total of 9488 middle-aged and elderly people in China participated in this study. In this study, the subjects were divided into male and female groups by the definition of the 2018 Chinese Guidelines for Prevention and Treatment of Hypertension. Searching for the best predictors among 13 obesity-and lipid-related indicators through binary logistic regression analyses and receiver operator curve (ROC). These 13 indicators are body mass index (BMI), waist circumference (WC), waist-height ratio (WHtR), conicity index (CI), visceral adiposity index (VAI), Chinese visceral adiposity index (CVAI), lipid accumulation product (LAP), a body shape index (ABSI), body roundness index (BRI), triglyceride glucose index (TyG-index) and its correlation index (TyG-BMI, TyG-WC, TyG-WHtR) Results After adjusting bias, all 13 indexes are risk factors for hypertension. In ROC curve analysis, thirteen obesity-and lipid-related factors can predict the occurrence of hypertension. Among them, CVAI has the best prediction effect (male: AUC = 0.660, female: AUC = 0.699). AUC for WHtR was equal to that for BRI and TyG - WHtR in identifying hypertension in male. Similarly, AUC of TyG-BMI and BMI were the same. In females, AUC for WHtR and BRI were the same when predicting hypertension. AUC of ABSI was much lower than other test indexes. Conclusion In predicting hypertension, thirteen obesity-and lipid-related factors are effective. In addition, in males and females, CVAI is the best indicator to indicate hypertension. TyG-WHtR, WHtR, and BRI performed well in predicting metabolic syndrome in both males and females. ABSI has a poor ability to predict hypertension.


Introduction
Hypertension is the most common risk factor for cardiovascular disease among middle-aged and elderly people in China. Hypertension is de ned as systolic blood pressure (SBP) ≥ 140mmhg, and/or diastolic blood pressure ≤ 90mmhg, and/or taking antihypertensive drugs within two weeks [1]. In a study of 977 participants, the incidence of hypertension was as high as 29.49%, while only 43.2% of hypertension patients are fully aware of their overall disease condition [2]. According to a national survey on high blood pressure among adults in China conducted since 2012, the prevalence of high blood pressure among adults in China is about 23.2% [3]. However, the prevalence rate of hypertension the middle-aged and elderly people in China is as high as 56.1%, and Materials And Methods Participants Participants in this cross-sectional study were China community residents over the age of 45. All participants were from China Health and Retirement Longitudinal Study (CHARLS). CHARLS is a nationally representative longitudinal survey. Every two years, CHARLS conducted computer-assisted personal interviews (CAPI) and structured questionnaires with participants. In the survey, 17,284 participants were 45 years of age or older. CHARLS collected data from 2011, 2013 and 2015, and we used the data of 2011. We excluded participants who were not followed up, as well as any standard individuals without data on age, sex, education, smoking history, activity participation, regular exercise, and chronic disease. The number of people who completed both baselines without hypertension symptoms was 9488.

Hypertension symptom
According to the 2018 Chinese Guidelines for Prevention and Treatment of Hypertension [22], Clinical systolic blood pressure≥ 140mmHg or diastolic blood pressure≥ 90 mmHg was de ned as hypertension. Blood pressure is usually measured with an international standardized upper arm medical electronic sphygmomanometer or a mercury sphygmomanometer that meets measurement standards (at least 5 minutes of sitting in a quiet environment). Diagnosis of hypertension is divided into three categories: 1. Clinic systolic BP≥ 140 mmHg or diastolic BP≥ 90 mmHg (antihypertensive drugs were never used in three different visits).

Covariates
In this study, we divided the participants into a male group and a female group. At the same time, we regard age, education level, marital status, current residence, current smoking, drinking, chronic diseases, participation in activities, and regular exercise as covariates of this study. We counted 14 chronic diseases and grouped them according to the number of diseases, which is the same as our previous studies [23,24]. The nine covariates are shown below.

Measurements
Waist circumference (WC) is the circumference of the line connecting the lowest point of the rib to the midpoint of the upper edge of the iliac crest before the end of expiratory breath [33]. It should be noted that the other 12 indicators need to be calculated. At the same time, some indicators need to be invasive to obtain TG and HDL. Body mass index (BMI) is the value of weight (kg) divided by the square of height (m) [34]. Waist height ratio (WHtR) is the ratio of WC (m) to height (m) [35]. The calculation of visceral adiposity index (VAI) differs between males and females but is based on WC, BMI, TG, and HDL  [40]. Conicity index (CI) is obtained through WC, weight, and height [41]. Triglyceride glucose index (TyG-index) is a lipid index calculated from TG and glucose [42]. At the same time, TyG index combined with BMI, WC, and WHtR constitutes TyG-BMI, TyG-WC, and TyG-WHtR [43][44][45]. We have listed the calculation formula for the 12 indicators below: (1) BMI = Weight/Height 2 (2) WHtR = WC/Height

Statistical analysis
In this study, SPSS version 25.0 (IBM SPSS, Armonk, NY, USA) was used for statistical analysis of all obtained sample data. The chi-square test was applied to classi ed variables and the t-test was applied to continuous variables to determine the strength of differences between the variables. The odds ratio (OR) of obesity-related and lipid-related indices to hypertension before and after adjustment of the bias was calculated. These biases included age, education level, marital status, current residence, current smoking, alcohol drinking, taking activities, regular exercise, and chronic diseases. The ability of these indicators to predict hypertension was compared by receiver operator curve (ROC). Table 1 shows the baseline characteristics of participants according to gender difference. The number of participants was 9488. 45.89% were male, and 54.11% were female. 12.45% were single. 92.40% were living in rural. 24.69% of males do not smoke, 92.19% of females do not smoke, 13.71% of males have never received education, while 42.56% of females have never received education. Among them, 24.69% of males have never smoked, and 92.19% of females have never smoked. 58.45% of males are smoking, and 5.96% of females are smoking. There are signi cant differences in education level, smoking and drinking between males and females. These 13 obesity-and lipid-related indices are different between male and female groups (P < 0.05).     Table 4 shows the ROC analysis between the 13 indicators and hypertension. Figure 1 and Fig. 2 show the ROC curve and AUC for males and females, respectively. In male, the largest AUC was observed for the CVAI ( It should be noted that these 13 indicators have signi cant signi cance for predicting the incidence of hypertension in both men and women (P < 0.001)  [3], the prevalence of high blood pressure among adults in China is about 23.2%, among which the prevalence of hypertension increases with age, the prevalence of hypertension in the population aged 65 and over is over 55%, and the prevalence rate in men is higher than that in women. Therefore, it is necessary to nd a more suitable obesity-and lipid-related index to predict the incidence of hypertension in middle-aged and elderly people in China. Table 3 shows that the ABSI is signi cantly related to hypertension in men, after adjusting the individual characteristics, but it is not statistically signi cant in women. Some studies [37,[46][47][48] show that ABSI is lower than other indicators in relation to hypertension, which is consistent with the results of this study. It is worth noting that the data in Table 4 shows that ABSI still has reference signi cance in predicting the incidence of hypertension among middle-aged and elderly people in China. However, ABSI's prediction ability is far lower than the other 12 indexes. Cheung, Y.B [47] study showed that compared with WC and BMI, ABSI was less associated with middle-aged and elderly morbidity. Because the deviation of ABSI from its average value is low, ABSI is not signi cant in predicting the incidence of hypertension [49]. After adjusting the individual characteristics, the odds ratio (OR) and 95% CI of various obesity-and lipid-related indices to hypertension were calculated, and all reached statistical signi cance (P < 0.05).

Results
ROC analysis of obesity-and lipid-related indices showed that the AUC of all indexes of male and female was statistical meaning (P < 0.05). Among the 13 indexes, CVAI, WHtR, BRI, and TyG -WHtR have received special attention.
In a cross-sectional study [50] of 14,573 participants, VAI was a better predictor of hypertension in both men and women. VAI proposed by Amato M.C can effectively evaluate visceral fat function [36]. A study conducted by Fiorentino, T.V [51] to test the risk factors related to the progression of hypertension in patients with prehypertension showed that VAI was an independent risk factor for the progression of hypertension. However, the signi cant overlap of con dence intervals of VAI did not mean that VAI was superior to other obesity indicators in predicting the development of hypertension. Notably, the combination of VAI and WC shows the highest predicted values for hypertension. This conclusion is consistent with the results of a cohort study [52] conducted in Chengdu, Sichuan Province, China. It can be analyzed for the following reasons: the difference of body fat distribution between different races is also obvious. According to the distribution law of body fat in Asian population, Xia, M. F. put forward the index of CVAI based on VAI to calculate the visceral fat area of China people [41]. In this study, compared with VAI, CVAI is more prominent in evaluating the prevalence of hypertension in middle-aged and elderly people in China.
A cohort study [53] of 10,304 Chinese adult residents showed that CVAI outperformed other measures of visceral obesity in predicting the incidence of hypertension in either men or women. A study by Li, B [54] has shown that CVAI is more effective in discriminating hypertension and prehypertension among the general population in China. Similarly, Lin, M [55], in a 2022 cohort of 2, 033 participants, showed that CVAI performed best in predicting hypertension. In this study, ROC analysis was performed on obesity-and lipid-related indicators, maximum AUC (male: AUC = 0.660, 95% CI = 0.643-0.676 and optimal cut-off = 111.142, female: AUC = 0.699, 95% CI = 0.685-0.713, and optimal cut-off = 113.022) was observed for the CVAI. CVAI showed the best ability to predict the occurrence of hypertension in both men and women. This conclusion has corresponded with the conclusion of the previous study.
WHtR and BRI have obvious advantages in forecasting the incidence of hypertension in middle-aged and elderly people [56]. In this study, WHtR and BRI are indicators for predicting the occurrence of hypertension, and their prediction ability is second only to CVAI. Lee, J. W [57] study showed that WHtR and WC were superior to BMI as screening tools in forecasting the incidence of hypertension in middle-aged and elderly Koreans. At the same time, A prospective cohort study [58] of 812 participants showed that WHtR had a signi cant advantage over WC, and BMI in predicting the development of hypertension. Similarly, in a cross-sectional study by Saeed, A. A [59], WHtR was the best predictor of hypertension among various anthropometric measures. These studies are basically consistent with the view of this paper that WHtR has a good performance in predicting the occurrence of hypertension.
Some obesity-and lipid-related indices provide a more reliable basis for predicting the occurrence of hypertension. In this study, the association strength between four lipid indexes, such as TyG index, TyG-body mass index, TyG-WC and TyG-WHtR, and hypertension was evaluated. In fact, a large number of Table   Table 3 is available in the Supplementary Files section. Figure 1 See image above for gure legend.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. Table3.doc