Ethnicity Disparities in the Prevalence, Awareness, Treatment, and Control Rates of Hypertension in China

Objectives Previous studies reported that there were disparities in hypertension management among different ethnic groups, and this study aimed to systematically determine the prevalence, awareness, treatment, and control rates of hypertension in multiple Chinese ethnic groups. Methods We searched Embase, PubMed, and Web of Science for articles up to 25 October, 2022. The pooled prevalence, awareness, treatment, and control rates of hypertension were estimated with 95% confidence intervals (CI). The heterogeneity of estimates among studies was assessed by the Cochran Q test and I2 statistic. Meta-regression analyses were conducted to identify the factors influencing the heterogeneity of the pooled prevalence, awareness, treatment, and control rate of hypertension. Results In total, 45 publications including 193,788 cases and 587,826 subjects were eligible for the analyses. The lowest prevalence was found in the Han group (27.0%), and the highest prevalence was in the Mongolian population (39.8%). The awareness rates ranged from 24.4% to 58.0% in the four ethnic groups. Both the highest treatment and control rates were found in the Mongolian population (50.6% and 16.0%, respectively), whereas the Yi group had the lowest control rate (8.0%). In addition, the study year, the mean age of subjects, mean body mass index of subjects, tobacco use (%), alcohol use (%), residence (urban%), and education (primary school%) had varied effects on heterogeneity. Conclusions These findings highlight the disparities in prevalence, awareness, treatment, and control rates of hypertension in a different ethnic population of China, which could provide suggestions for making targeted prevention measures.


Introduction
Hypertension is an important risk factor for cardiovascular diseases and other chronic diseases which impose a huge burden on economic and human development globally [1,2]. Nevertheless, low awareness, treatment, and control of hypertension will handle the prevention progress of cardiovascular diseases [3][4][5]. In China, the median prevalence of hypertension was 32% in females and 37% in males [6], which were higher than the average level of 20% of women and 24% of men around the world [1]. Te awareness, treatment, and control rates reported by a systematic review were only 53%, 44%, and 17%, respectively, among the female population, and these rates were even lower in men [6].
Some studies reported that the prevalence rate of hypertension may be diferent in each ethnic group, which could be attributed to the diferent genetic and environmental conditions [7][8][9][10]. China is a unifed, multinational country, with 56 ethnic groups in all. It is evident that the status and management of hypertension are varied in these groups [11]. Understanding the status of the ethnic groups was benefcial for hypertension control.
Most of the individual studies were conducted for one ethnic group, however, systematically collected data to compare the prevalence, awareness, treatment, and control rates among varied ethnic populations were insufcient. Terefore, this study investigated the disparities in the management of hypertension among ethnic groups to improve hypertension prevention in order to achieve the global noncommunicable disease targets.

Data Sources.
We searched the three databases (Embase, PubMed, and Web of Science) up to 25 October, 2022. Te medical subject heading (MeSH) terms were used for the search strategies. Te search strategy was as follows: (blood pressure OR hypertension) AND (epidemiology OR prevalence OR incidence OR awareness OR treatment OR control) AND (China). To maximize the yield of articles, we further screened the references of all included studies or meta-analyses in the database to select additional eligible articles.

Inclusion and Exclusion
Criteria. Te eligible articles were under the following criteria: (1) articles that were published in peer-review journals in both English and Chinese; (2) articles that defned hypertension as SBP ≥ 140 mmHg or/and DBP ≥ 90 mmHg; or selfreported pharmacological treatment for hypertension; (3) articles that reported the prevalence or awareness or treatment or control rate; (4) articles that provided the cases and sample size or sufcient data to calculate these rates. Some exclusion criteria were set before data extraction: (1) articles that provided limited information; (2) articles that were not the original research articles, such as review, comment, or letter; (3) articles that focused on the treatment or medication of hypertension. Two investigators (G.DQ and Z.ZQ) evaluated each study and checked the eligibility independently.

Data Extraction.
Data were extracted by two researchers (Y.YN and Z.ZQ) using the standard information forms, including frst author name, publishing year, study year, study design type, ethnicity, case number, sample size, prevalence rate, awareness rate, treatment rate, control rate, mean age of the population, age range of population, mean body mass index (BMI), tobacco use (%), alcohol use (%), gender (female%), residence (residence%), and education (education%). When the data from the same survey were published in more than one publication, we used the largest one.

Quality
Assessment. Te quality of included studies was independently assessed by two investigators (H. KL and Y. YN) using the adapted Newcastle-Ottawa Scale (NOS) [12]. If there were any disagreements, it was resolved by consultation with another author (G. DQ). Nine items were applied to assess the included studies, and each study received a total score (TS) for methodological quality. Te score ranged from 0-10: TS ≥ 7 ranks as "high quality," 7<TS ≤ 6 ranks as "moderate quality," and TS < 6 ranks as "low quality."

Statistical Analysis.
A pooled prevalence, awareness, treatment, and control rates of hypertension were estimated with 95% confdence intervals (CI). Te heterogeneity of estimates among studies was assessed by the Cochran Q test and I 2 statistic. Meta-regression analyses were conducted to identify the factors infuencing the heterogeneity. In regression models, the study year (<2010, ≥2010), mean age (years), mean BMI, tobacco use (%), alcohol use (%), gender (female%), residence(urban %), and education (primary school%) were included as covariates. Te sensitivity analysis was also conducted to explore the potential heterogeneity. Te Egger test and Begg test were performed for publication bias. Statistical analyses were conducted using Stata 14 software (https:// www.stata.com/) p ≤ 0.05 was defned as statistical signifcance.

Characteristics of Included Studies.
A PRISMA fowchart was used to illustrate the selection of studies ( Figure 1) [13]. We screened a total of 3,618 articles, and fnally, 45 publications of 193,788 cases and 587,826 subjects were included in the analyses. According to the number of datasets, a total of four ethnic groups (Han, Tibetan, Yi, and Mongolian) were considered to conduct the meta-analyses. Tere were 14 datasets used for the

Quality of the Studies.
Te quality of included studies was relatively high. According to the results of the quality assessment, 31 (68.9%) studies were categorized as "high quality" and 14 (31.1%) studies were classifed as "moderate quality." Te details of the quality assessment are shown in Supplementary Table 2.

Prevalence, Awareness, Treatment, and Control of
Hypertension. All prevalence rates were calculated using the cut-of value (140/90 mmHg), and the awareness, treatment, and control rates were calculated among cases with hypertension. Te pooled prevalence, awareness, treatment, and control rates of the total population were 31.7%, 42.5%, 33.4%, and 12.2%, respectively. Te lowest prevalence rate was found in the Han group (27.0%), and the highest prevalence was in the Mongolian population (39.8%). Te awareness rates ranged from 24.4% to 58.0%% in four ethnic groups. Te highest treatment and control rates were found in the Mongolian population (50.6% and 16.0%, respectively). Te Yi group had the lowest control rate (8.0%), see Table 1.

Results from Meta-Regression Analyses.
Te metaregression analyses were conducted for the four ethnic groups, respectively. In the meta-regression model for the Han group, the results of mean age and mean BMI were statistically signifcant, explaining 40.96% (p � 0.015) and 35.16% (p � 0.032) of the heterogeneity in prevalence between studies, respectively. Te mean BMI, alcohol use, and study year explained 54.21% (p � 0.006), 66.83% (p � 0.001), and 54.20% (p � 0.003) of the heterogeneity in awareness between studies, respectively. And the mean BMI, alcohol use, study year, and tobacco use were statistically signifcant, explaining 48.44% (p � 0.011), 53.19% (ws), 40.19% (p � 0.009), and 26.99% (p � 0.041) of the heterogeneity in treatment or control between studies (See Table 2). In the meta-regression model for the Mongolian group, the source of heterogeneity for prevalence was mean age (Adj. R 2 � 69.12%, p � 0.004), and education (Adj. R 2 � 46.36%, p � 0.039) for control rate (see Table 3). In addition, the residence (Adj. R 2 � 100.00%, p � 0.048) afected the awareness rate of the Tibetan group (see Supplementary Table 3). Te mean BMI (Adj. R 2 � 76.68%, p � 0.034) and study year (Adj. R 2 � 57.40%, p � 0.036) also had efects on the prevalence of the Yi group, respectively (see Supplementary Table 4).

Bias and Sensitivity
Analysis. 31 (68.9%) studies were judged to be of high quality with a low risk of bias, while 14 (31.1%) studies were at moderate risk of bias. In the sensitivity analyses, no individual study was found to have any major impact on the pooled prevalence, awareness, treatment, and control rate of hypertension (see supplementary . Te Egger test was not statistically signifcant for testing the publication bias, except in analyses of pooled awareness of the Han population (p � 0.0002) and the pooled treatment of the Tibetan population (p � 0.0005). Te Begg test showed no small-study bias in these metaanalyses, all the p values were greater than 0.05.

Discussion
To our knowledge, this study is the frst to systematically summarize the rates of hypertension management in multiple Chinese ethnic groups. In this study, we fnally selected four main ethnic groups (Han, Tibetan, Yi, and Mongolian) in our analyses according to the number of databases. Our results indicated that the status varied in these four groups, in which Mongolian had the highest prevalence, treatment, and control rates. However, Yi people had a diferent story of increasing the prevalence rate and relatively lower the The number of datasets less than three (n=11) International Journal of Hypertension control rate. Te results denoted that understanding the diferences in varied ethnic groups could inform targeted prevention measures to improve hypertension management and reduce the disparities. Based on previous fndings, the prevalence rates of hypertension would vary in diferent ethnic populations due to genetic and environmental factors [14,15]. In our study, the combined prevalence among four ethnic groups was 31.7%, which was slightly higher than the results from the Chinese Hypertension Survey in 2012-2015 which found that the average rate of the Chinese population was 27.9% [16]. In these ethnic groups, the rate of Han nationality was the lowest (27.0%), while the rates of Mongolian and Tibetan were more than 30%. Tese results were consistent with previous studies [17]. Most of the Tibetan population in China are living in the highlands, and some publications indicated that chronic hypoxia exposure may cause a higher prevalence of hypertension [18]. In addition, according to our previous studies, hypoxia exposure may also increase the risk of other diseases, such as sleep disorders, which were regarded as the risk factors for hypertension [19,20]. For the Mongolian population, the prevalence was high in the recent thirty years in some large-scale surveys [21]. Tis phenomenon may be attributable to the changes in their diet culture because Mongolian people prefer to consume meat and milk as the principal food, especially among people living in pastoral regions [22]. Yi population was found to have a low prevalence of both overweight and obesity [23], which may contribute to a lower rate of hypertension (28%). However, the population attribution fraction (PAF) of overweight or obesity in this population was also increasing in recent years. For example, in 1996, the PAF was 27.66% and the rate of hypertension was 5.33% in 1996, while the PAF increased to 33.26% in 2015 with a prevalence of hypertension of 17.2% [24].
Furthermore, among people with hypertension, the awareness, treatment, and control rates were also analyzed in this study. In China, the average rates were 51.5%, 46.1%, and 16.9% reported by a nationwide survey in 2015 [25], which were higher than the rates reported in other developing countries, such as Ghana [26]. In this study, only the rates from the Mongolian population were close to this level, and the Tibetan and Han population slightly lower than that level, while these rates in the Yi group were the lowest. Two possible reasons might account for these results. Firstly, compared to the other groups, Yi people had a relatively lower prevalence of hypertension over a long-term period, which may make them pay less attention to the increase of risk factors at the population level. For example, some studies had reported that Yi people's dietary patterns changed largely in the past two decades, which caused the gradual growth of the prevalence of hypertension [27,28]. Secondly, the previous studies showed that numerous Yi people migrated from rural to urban areas, which imposed heavy pressure on the prevention of these two distinct groups. Yi migrants usually had a higher prevalence of hypertension than Yi farmers, and it was useless to implement the same measures to improve their awareness, treatment, and control situation among Yi people. More targeted strategies and policies are needed in the Yi population.
Tere are some strengths in our study. Te pooled estimates of the prevalence, awareness, treatment, and control of hypertension in four main Chinese ethnic groups were reported comprehensively. Te search strategy involved plenty of studies and subjects based on the standard of PRISMA [29]. Our results were robust and corroborated by further sensitivity analyses and the tests of small-study bias and publication bias. Nonetheless, there are also some limitations. Firstly, high heterogeneity was found in our meta-analyses, which were familiar with other meta-analytic studies focusing on the pooled rates because of the diferences in the study population, inclusion and exclusion criteria, and assessment [30,31]. In this systematic review, we also included some common covariates to explore the source of heterogeneity, such as BMI, gender, tobacco use, alcohol use, age, study year, and education. Of which, mean age, mean BMI, study year, alcohol use, and tobacco use were signifcantly associated with the rates of hypertension management. Secondly, some other Chinese ethnic groups, such as Zhuang, Uygur, and Hui, were not included in the meta-analyses because of insufcient datasets. International Journal of Hypertension 5

Conclusions
Te systematic review included four main Chinese ethnic groups and conducted meta-analyses for prevalence, awareness, treatment, and control rates of hypertension. Te fndings suggest that hypertension management varied in these Chinese ethnic groups which Mongolian and Tibetan had higher prevalence rates. Notably, the Yi population was regarded as the low-risk population for hypertension, and it should pay more attention to prevention because the prevalence rates were rising in recent years, while their awareness, treatment, and control rates were still low. We call for further studies to better understand the specifc risk factors and mechanisms of hypertension in the Tibetan and Mongolian populations to make more targeted preventive measures. In addition, for Yi people, it is urgent to implement interventions to improve their awareness and treatment, and fnally achieve the goal of hypertension control.

Supplementary Materials
Supplementary  Table 3: meta-regression analysis of the Tibetan group to explore potential sources of heterogeneity, prevalence, awareness, treatment, and control. Supplementary Table 4: meta-regression analysis of the Yi group to explore potential sources of heterogeneity and prevalence. Supplementary