Hypertension is the major modifiable cause of premature death worldwide and in China. According to a recent survey, the standardized prevalence of hypertension reached 24.7% in 2018 and high systolic blood pressure (BP) was one of the leading risk factors contributing to deaths in mainland China [1,2]. Using the new hypertension definition of a systolic BP ≥130 mm Hg or a diastolic BP ≥80 mm Hg [3], the prevalence of hypertension in Chinese adults has nearly doubled (46.4%), whereas the awareness and control of hypertension remain substantially low [4].

Sex differences exist in hypertension, and it is widely known that men have higher BP levels and hypertension prevalence than women before the age of 60 years, after which the trend reversed between men and women [5,6]. Based on the Global Burden of Diseases, Injuries, and Risk Factors Study, the annual increase in global hypertension prevalence from 2000 to 2019 was only found in women but not in men, while hypertension-related mortality rate decreased more pronouncedly in women than in men [7]. In addition, previous studies have shown that unfavorable socioeconomic status (SES) is associated with increased hypertension risk [8-11]. However, most previous epidemiological studies reported the contribution of a single SES indicator, such as education or income without fully considering a comprehensive set of SES indicators, which might affect health in an aggregated way. Ignoring the synergistic effects of various social risk factors undoubtedly undermines efforts to identify socially disadvantaged people and implement targeted interventions to address social inequities [12]. Moreover, the role of SES in the sex differences of hypertension remains underinvestigated, and limited studies have explored the complex associations between SES, sex, and hypertension among general populations.

Therefore, using data from the China Noncommunicable Disease Surveillance study with a wide coverage of geographical areas and SES across mainland China, we aimed to evaluate the potential impacts of SES on sex differences in BP levels and prevalent hypertension in a general population.

General characteristics of study participants stratified by sex are shown in Table 1. Men were more likely to have a longer duration of education, a higher annual household income, and higher proportions of an overall high SES, current smokers, and current drinkers than women. Systolic and diastolic BP levels, and prevalence of hypertension were significantly higher in men than women (all P values <.001). Stratified by sex and each SES indicator, similar trends are shown in Tables S1-S5 in Multimedia Appendix 1. Because significant multiplicative interactions between all included SES indicators and sex in associations with BP characteristics were found, subsequent analyses were stratified on them (P_interaction<.05).

Table 2 shows the distribution of BP characteristics in Chinese men and women stratified by SES categories. Systolic and diastolic BP levels, and prevalence of hypertension were generally higher in men than women in each group stratified by SES indicators. Using men as the reference, the associations between sex and BP characteristics in different categories of SES were examined both in unadjusted (Table S6 in Multimedia Appendix 1) and multivariable-adjusted models (Table 3). There were no significant sex differences in BP characteristics in participants with duration of education<6 years. However, women were associated with significantly lower BP levels and hypertension prevalence compared with men in SES categories with longer durations of education. For example, in participants with >9 years of education, women were associated with a significantly lower level of systolic BP (β coefficient=−7.93, 95% CI −8.83 to −7.04) and diastolic BP (β coefficient=−3.28, 95% CI −3.82 to −2.75), and 37% lower probabilities of having prevalent hypertension defined by BP ≥140/90 mm Hg (PR 0.63, 95% CI 0.58-0.68) compared with men. Similar trends were found in groups stratified by other SES indicators with more evident sex differences in BP characteristics in participants having more favorable SES (Table 3). Similar results were found in hypertension defined by BP ≥130/80 mm Hg (Table S7 in Multimedia Appendix 1).

Mean systolic BP levels, mean diastolic BP levels, and prevalence of hypertension in Chinese men and women as well as the difference between Chinese men and women in subcategories with cross-tabulation of education and another SES indicator are shown in Figure 1 and Figures S2-S5 in Multimedia Appendix 1. By labeling with shades of colors, a clear pattern of sex differences in BP characteristics was seen. In participants with longer duration of education and more developed economic status, women had lower BP levels and hypertension prevalence than men, while in participants with less education and less developed economic status, women tended to have higher BP levels and hypertension prevalence than men. The same pattern was seen for multivariable-adjusted β coefficients and PRs demonstrating associations between sex and BP characteristics (Figures S6-S9 in Multimedia Appendix 1).

Figure S10 in Multimedia Appendix 1 showed the percentages of different composite SES scores in men and women. Women tended to have higher percentages of low SES scores and lower percentages of high SES scores than men. Using the overall low-SES group as the reference, a 9% lower probability (PR 0.91, 95% CI 0.83-0.98) and a 30% lower probability (PR 0.70, 95% CI 0.63-0.78) of having hypertension defined by BP ≥140/90 mm Hg were found in women with an overall intermediate SES and high SES, respectively after multivariable adjustment (Figure 2). In contrast, no significant reduction in the probability of having hypertension was observed in the intermediate or the high-SES group versus the low-SES group in men. Women-to-men ratios of PRs showed that there were significant differences between men and women in the reduced probabilities of having hypertension in the intermediate-SES group or the high-SES group versus the low-SES group (Figure 2). Similar trends were found in hypertension defined by BP ≥130/80 mm Hg (Figure S11 in Multimedia Appendix 1).

NNEB analyses showed that increasing every 13 men of their durations of education from <6 years to >9 years could avoid one case of hypertension (if defined by BP ≥140/90 mm Hg). In contrast, only 4 women were needed to increase their durations of education from <6 years to >9 years to avoid one case of hypertension.

Using a nationally representative sample of 98,658 Chinese adults across mainland China, we found that there were significant sex differences in BP levels and hypertension prevalence, and SES had a significant influence on these sex differences. Women had an increasingly favorable BP level and prevalent hypertension versus men with an improving SES. NNEB analyses indicated that a much smaller number of women are needed to be educated to avoid one case of hypertension than men. Our findings provided evidence of the potential effects of SES on sex differences regarding hypertension, and suggested that targeted public health policies in men and women are needed and interventions to alleviate the socioeconomic inequities, especially in women are warranted to reduce the hypertension burden.

Sex differences in prevalent hypertension found in our study were consistent with results from most but not all previous studies [4,26,27]. A pooled analysis of 135 population-based studies from 90 countries revealed global disparities in hypertension prevalence in men and women, and that the overall prevalence of hypertension was almost identical in men (n=520.1 millions, 31.7%) and women (n=518.8 millions, 31.2%) in low- and middle-income countries though still higher in men than in women (n=174.2 millions, 31.6% vs n=174.7 millions, 25.3%) in high-income countries [28]. These numbers further highlight the sex- and region-specific differences in hypertension prevalence. Previous studies suggested that a higher prevalence of hypertension in men than women might result from the androgen-specific effects of increased BP through the renin-angiotensin system or the potentially protective effects of estrogen on the vasculature and the sympathetic nervous system [29]. Unhealthy lifestyle behaviors including smoking and drinking as revealed in this study and higher energy, fat, and sugar intakes in men, which were associated with an increased risk of hypertension [30-33], might also count.

Using single, cross-tabulated, and composite SES, our study indicated that SES played an important role in the sex differences of hypertension, and women with an improving SES were significantly associated with lower BP levels and hypertension prevalence, which is partly consistent with previous studies [34-37]. A cross-sectional population-based study in France showed that education and income inequalities were stronger among women than among men [34]. However, the China Kadoorie, Biobank study showed no obvious sex difference in the association between hypertension and income in different GDP per capita areas [36]. Another large cohort study of Chinese population demonstrated an inverse association between SES and hypertension observed only in women [37]. However, sex differences were not compared directly in these studies [34-37] and often single SES indicators were included without considering regional disparities in China.

It has been recognized that SES indicators are associated with health outcomes partly through their impacts on lifestyle behaviors [38]. Previous studies have demonstrated that lower SES was associated with unhealthy lifestyle behaviors including lower physical activity, lower fresh food intake, and higher sodium intake [39]. In this study, the proportions of current smoking and excess drinking decreased while the leisure-time physical activity and diet score improved with increasing education levels in both men and women (Table S1 in Multimedia Appendix 1). It is worth noting that with increasing durations of education, levels of BMI and waist circumference increased in men but decreased in women. This was also found in previous studies which demonstrated that higher education levels were more likely to be associated with overweight or obesity in men while this association was reversed in women [40]. This could partly contribute to the widened differences in BP levels between women and men with increasing durations of education found in this study.

Participants with more favorable SES such as those living in the urban area or living in the developed economic region had higher levels of leisure-time physical activity and diet scores though at the same time higher BMI in both men and women (Tables S2 and S3 in Multimedia Appendix 1). Previous studies have elucidated that the protective impact of leisure-time physical activity on hypertension tended to be more evident in women, especially in those with overweight or obesity [41]. Dietary intervention also lowered systolic BP more in women than in men, which might lead to more significant sex differences in participants with favorable SES status [42]. Consistent with previous studies [43,44], we found that participants living in the north were more likely to have higher BMI and waist circumference than participants living in the south in both men and women (Table S5 in Multimedia Appendix 1). Emerging evidence showed that there was a stronger association between obesity or BMI and hypertension in women than in men, which might potentially account for the less significant sex difference in the north than in the south in this study [45,46]. However, the potential mechanisms of SES affecting sex differences in hypertension warrant further investigation.

The major strength of this study was the large study sample recruited from 162 study sites located across mainland China and representative of Chinese adults aged ≥18 years with a wide coverage of SES evaluated using 5 SES indicators. There are also several limitations. First, BP levels were measured at a single point in time, although standard procedures were followed and 3 measurements were conducted for each participant. Second, information on some SES indicators was self-reported and this may lead to misclassifications such as for annual household income. Third, causality cannot be inferred due to the cross-sectional nature of this study. However, sex is a born attribute and SES indicators were generally determined at early stages of life. The temporality of their associations with BP or hypertension is somehow clear. Finally, the generalizability of our findings to other populations in countries with different levels of development might be limited.

In conclusion, we found significant sex differences in BP levels and hypertension prevalence in Chinese men and women, and SES played an important part in the sex disparities. Women with favorable SES were inversely associated with BP levels and prevalent hypertension compared to men. Policies at increasing SES in women such as fostering longer durations of education to alleviate socioeconomic inequalities can be both effective and efficient toward a reduced burden of hypertension. Future studies are needed to illustrate the interplay of biological, social, and behavioral factors in sex differences in health outcomes.