Demographic and socioeconomic inequalities in ideal cardiovascular health: A systematic review and meta-analysis

Background In 2010, the American Heart Association introduced a new concept of ideal cardiovascular health (CVH) defined as the simultaneous presence of 7 favorable CVH metrics (smoking, diet, physical activity, body mass index, blood pressure, total cholesterol, and fasting blood glucose). The objective of this study was to conduct a systematic literature review and meta-analysis of studies examining the prevalence of ideal CVH, and each of the ideal CVH metrics as well as the relationship between socio-demographic determinants and ideal CVH. Methods A comprehensive literature search was conducted in Medline and Scopus databases for studies published between 1 January 2010 and 30 June 2020. A total of 50 studies including 2,148,470 participants were analyzed. Associations were estimated using DerSimonian-Laird random-effect models. Heterogeneity was investigated through subgroup analyses, Q-test, and I2 statistics. Results This study showed a low prevalence of ideal CVH defining as 6 and 7 ideal metrics (3.3%). Among seven ideal CVH metrics, smoking was the best metric (71%), while the poorest CVH metric was a healthy diet (5.8%). Gender was a statistically significant moderator of ideal smoking (81% in females and 60% in males) and ideal blood pressure (42% in females and 30% in males). Females and young adults had better CVH status compared to males and older adults. Also, more educated and better-off individuals had a greater number of ideal CVH metrics. Conclusions To the best of our knowledge, this is the first systematic review on the relationship between participants’ socioeconomic status and ideal CVH. The results suggest that the prevalence of ideal CVH and most metrics was unsatisfactory. In order to achieve the improvement of the CVH metrics and the overall ideal CVH, nationwide prevention efforts at the population and individual levels are urgently needed.


Introduction
In 2010, the American Heart Association (AHA) introduced a new concept of ideal cardiovascular health (CVH) as part of its efforts to improve the CVH of all Americans while reducing deaths from cardiovascular disease (CVD) [1]. This concept is defined as the simultaneous presence of 7 favorable CVH metrics or "Life's Simple 7": 4 health behaviors (smoking, diet, physical activity, and body mass index) and 3 health factors (blood pressure, total cholesterol, and fasting blood glucose) in the absence of CVD.
Therefore, evaluating associations between socio-demographic characteristics and ideal CVH metrics would be a valuable resource for communities to develop public health and clinical interventions and policies to improve ideal CVH and consequently prevent CVD events. However, studies that synthesize data on the prevalence of 7 ideal CVH metrics, ideal CVH, and their distribution by socio-demographic characteristics are lacking. To the best of our knowledge, there is only one systematic review [15] and one meta-analysis [19] on this topic. Accordingly, we conducted a comprehensive systematic review and meta-analysis with the aim to synthesize data on the prevalence of ideal CVH, and each of the ideal CVH metrics. What is new in our review compared to the previous reviews, is the updating of the population's ideal CVH due to the significant number of recently published studies, and the fact that we examined the relationship between social determinants such as education and income and ideal CVH.

Search strategy for identification of studies
We performed a comprehensive electronic search of published studies from their inception to June 2020 by searching Medline (via PubMed) and Scopus databases with the following terms: "cardiovascular health", "Life's simple 7", and "ideal" in English. The search string used for the review in

Prevalence of ideal CVH metrics and ideal CVH
Meta-analyses of each of seven ideal CVH metrics are presented in Fig 2. The observed proportions for ideal smoking (Fig 2A) ranged from 0.32 to 0.98, and the estimated pooled proportion based on the random-effects model wasm ¼ 0:71 (95% CI: 0.68 to 0.74). For ideal diet ( Fig 2B) the observed double arcsine transformed proportions ranged from 0.003 to 0.569, and the estimated pooled double arcsine transformed proportion based on the random-effects model wasm ¼ 0:234 (95% CI: 0.194 to 0.274). The back-transformed summary proportion was 0.058 (95% CI: 0.040 to 0.080). The observed proportions for ideal physical activity (Fig 2C) ranged from 0.02 to 0.95, and the estimated pooled proportion based on the random-effects model wasm ¼ 0:41 (95% CI: 0.34 to 0.47). For ideal body mass index (Fig 2D), the observed proportions ranged from 0.14 to 0.80, and the estimated pooled proportion based on the random-effects model wasm ¼ 0:41 (95% CI: 0.36 to 0.45). The observed proportions for ideal blood pressure ( Fig 2E) ranged from 0.05 to 0.76, and the estimated pooled proportion based on the random-effects model wasm ¼ 0:34 (95% CI: 0.29 to 0.39).
For ideal total cholesterol (Fig 2F), the observed proportions ranged from 0.22 to 0.86, and the estimated pooled proportion based on the random-effects model wasm ¼ 0:55 (95% CI: 0.51 to 0.59). The observed proportions for ideal fasting blood glucose ( Fig 2G) ranged from 0.14 to 0.97, and the estimated pooled proportion based on the random-effects model wasm ¼ 0:68 (95% CI: 0.62 to 0.73).
In a meta-analysis of 37 studies, for the simultaneous presence of 5, 6, and 7 ideal CVH metrics (Fig 2H), the observed double arcsine transformed proportions ranged from 0.181 to  The prevalence of ideal CVH (defined here as achieving 6 and 7 ideal CVH metrics), intermediate CVH (defined as achieving 2 to 5 ideal CVH metrics), and poor CVH (defined as 0 to 1 ideal CVH metrics) in selected countries are presented in S1 Fig. The prevalence of ideal CVH was low in all observed countries and ranged from 0.5% in the USA Jackson Heart Study [36] to 15% in the Chinese Health Examination Database study [70]. As many as 15 out of the

PLOS ONE
21 studies conducted worldwide reported less than 5% prevalence of ideal CVH. Most participants had intermediate CVH with the prevalence ranging from 70% in the USA Heart SCORE study [29] to 93% in the Serbian Health Survey [48]. The prevalence of poor CVH ranged from 1% in the Korean Seoul Male Cohort Study [51] to 29% in the USA Heart SCORE study [29]. The prevalence of ideal health behaviors and the prevalence of ideal health factors from 10 studies with available data are presented in S2 Fig. The percentage of subjects who achieved the 4 ideal behaviors or lifestyles (smoking, diet, physical activity, and BMI) was lower than the percentage of those with the 4 ideal health or biological factors (blood pressure, total cholesterol, fasting blood glucose, and smoking) in all observed studies, except one [29]. As defined by the AHA, given the importance of abstinence from smoking and smoking cessation to health promotion, this metric appears in both health behaviors and health factors [1].
For the ideal CVH (6 and 7 CVH metrics) a total of 44 studies were included in the metaanalysis (Fig 2I), and the observed double arcsine transformed proportions ranged from 0.015 to 0.415. The estimated pooled double arcsine transformed proportion based on the randomeffects model wasm ¼ 0:188 (95% CI: 0.154 to 0.220). The back-transformed summary proportion was 0.033 (95% CI: 0.022 to 0.046).
For all ideal CVH metrics, as well as for the presence of 5 and more ideal CVH metrics and ideal CVH, the true outcomes (proportions) appear to be heterogeneous indicating the need for moderator (subgroup) analysis.

Prevalence of ideal CVH metrics and ideal CVH according to gender and age
Where data was available, participants were stratified by gender (female and male) and moderator analysis was performed for each of 7 ideal CVH metrics, for the presence of 5 and more ideal CVH metrics, and for ideal CVH (6 and 7 ideal CVH metrics) ( Table 2) A total of 11 studies had available data and were included in the analysis of age and ideal CVH (6 and 7 ideal CVH metrics) (Fig 3). The back-transformed summary proportions were: 0.08 (95% CI: 0.05 to 0.11) for the young age group, 0.03 (95% CI: 0.01 to 0.04) for the middle age group, and 0.01 (95% CI: 0.01 to 0.02) for the older age group. Test of moderators revealed that the difference between the three subgroups is significant (QM = 24.24, DF = 2, p < 0.001). The better CVH was seen in younger and middle-aged adults compared to older participants.

Ideal CVH according to socioeconomic characteristics
The results of 14 studies describing the association between socioeconomic variables (education, income, and Wealth index), and the prevalence of ideal CVH are presented in Table 3.
Results from most studies showed that participants with the highest education had a greater number of ideal CVH metrics, i.e. better CVH in comparison with those with the lowest

PLOS ONE
education [4,9,33,39,48,49,53,56,73]. In a study by Foraker et al. [39] individuals with a bachelor's degree, had on average, a CVH score of 0.67 points higher compared with those with high school or less education. Similarly, persons with graduate/professional degrees had on average a CVH score of 0.93 points higher than their less-than-high educated school counterparts. Machado et al. [55] found that individuals with less than high school education had lower ideal CVH scores than college-educated individuals. In a study by Matozinhos et al. [56] lower levels of education were associated with a lower prevalence of ideal CVH, regardless of gender, compared to those with a higher level of education. In contrast, Ongunmoroti et al. [60] found that lower education was associated with better CVH, while Benziger et al. [30] and Zhao et al. [73] did not find any statistically significant association between education and ideal CVH (Table 3).
Only a few studies assessed the relationship between income and CVH. Foraker et al. [39] found that those with higher individual and neighborhood levels of income had higher CVH scores, i.e. better CVH. There was an average increase in CVH score of 0.31, and 0.19 points associated with each 1-category increase in individual income, and neighborhood income, respectively. Better results, considering the number of ideal CVH metrics were also achieved in persons with more family income living in Hong Kong [33]. Machado et al.
[55] demonstrated that low (<1245 USD) family income was associated with lower ideal CVH scores compared to those with high (�3320 USD) family income. In contrast, in a rural area of Northwest China, no associations were found between family income and ideal CVH [73] (Table 3).
In one study conducted in China, higher socioeconomic status (defined according to both personal income and years of education) was associated with an increasing prevalence of meeting 5 or more ideal CVH metrics in women but not in men [65]. Only two studies included in our review assessed the association between Wealth index (an aggregation of assets and household facilities) and ideal CVH. While Jankovic et al. [48] noted a statistically positive association between Wealth index and ideal CVH, Benziger et al. [30] found the opposite (Table 3).

Quality assessment
According to both applied quality assessment tools (NIH-QAT and tool specifically designed for prevalence studies), almost all included studies were regarded as having a low risk of bias (S3 and S4 Tables). Agreement on the quality assessment between the two reviewers was high.

Discussion
Our study is the second meta-analysis ever done that calculated the pooled prevalence estimates of the seven ideal CVH metrics and overall ideal CVH according to the AHA's guideline, and the first systematic review that examined the association between socioeconomic status (SES) and ideal CVH. Since a considerable number of studies on the prevalence of ideal CVH have been published in the last few years our study provides up-to-date information that could be useful for policymakers, clinicians, researchers, communities, and other stakeholders to understand and implement the most effective approaches to improve CVH in populations.
This study showed a low prevalence of ideal CVH defined as 6 and 7 ideal metrics (3.3%) that is in line with a previous systematic review [15]. When a less strict definition of ideal CVH was applied (5 to 7 ideal metrics) about 16% of our participants had an overall ideal CVH like in recent meta-analysis [19]. In most countries poor CVH (0-1 ideal metrics) was more frequent than ideal CVH (6 and 7 ideal metrics), while intermediate CVH (2-5 ideal metrics) was the most prevalent. Except in one study [29], we also observed a lower prevalence of ideal health behaviors (ranged from 0.1% to 2%) compared with the prevalence of ideal health factors (ranged from 1.4% to 16.4%).
The results of our meta-analyses are summary proportions of ideal CVH and ideal CVH metrics from studies conducted during different study periods, ranging from 1984 to 2017. We are aware that the CVH status has changed during the observed period and several cohort studies assessing trends in ideal CVH and CVH metrics documented that finding. Patel et al. [62] examined the trends in ideal CVH during economic recession (2007-2010) and subsequent economic recovery (2011-2016) among American adults and noted a decline in ideal CVH score that was primarily driven by the increased prevalence of obesity and poor fasting glucose. A Danish study [5], using data from six cross-sectional studies conducted in an adult population aged 30-64 years in Denmark from 1978 to 2006, reported an increasing trend in ideal CVH. Huffman et al. [74] analyzed CVH behavior and health factor changes from 1988 to 2008 and showed modest, further declines in tobacco consumption, high cholesterol, and high blood pressure, offset by increases in obesity and dysglycemia. The high proportions of people with favorable smoking status in this study (71%) could be a reflection of comprehensive tobacco control policies and a significant decrease of smoking worldwide [74]. The prevalence of ideal fasting blood glucose metric (68%) like in reviews by Younus et al. [15] and Peng et al. [19] was also high but trend results suggest its decline [62,74]. This high optimal prevalence of fasting blood glucose may partly be explained by the slight improvement of physical activity and dietary pattern [74][75][76], which were found to be risk factors for elevated fasting blood glucose and diabetes [77]. In our review, 55% of participants had ideal total cholesterol that is in line with the recent meta-analysis [19]. Total cholesterol levels were decreasing, partly due to the increasing use of lipid-lowering drugs [78,79]. We reported that 41% of participants had ideal blood pressure that is consistent with findings from the recent US nationally representative survey (42.3%) [80], but higher in comparison with results of a recent meta-analysis (34.6%) [19] and the China national hypertension survey (35.5%) [81]. Our results of ideal physical activity (41%), and ideal BMI (41%) are also in accordance with a previous meta-analysis by Peng et al. [19]. In our study, the poorest CVH metric was a healthy diet (5.8%) like in both previous reviews [15,19] and almost all individual studies that examined ideal CVH. Although the prevalence of a healthy diet is extremely low, the question of its relevance to ideal CVH does not arise. A healthy diet is one of 7 CVH metrics proposed by AHA, essential for keeping people healthy across the lifespan which could be successfully improved [1]. Unfortunately, physical activity levels and low diet quality scores changed minimally during the observed time [74]. Hence, successful prevention efforts for improvement of healthy diet and physical activity are needed. This should result in improvement in BMI, blood pressure, fasting blood glucose, and overall CVH status and reducing the related CVD burden.
In this study, women had twice higher ideal CVH (6%) than men (3%) and a higher percentage of 5 and more ideal CVH metrics (21% in women vs. 13% in men) that is in line with findings from a recent meta-analysis [19]. Several studies have specifically investigated gender disparities in the distribution of ideal CVH and discussed the potential reasons behind that [7,9,31]. There is evidence that women attended primary care (general practitioner/family physician) and preventive services more often than men [82][83][84]. Simon et al. [9] hypothesized that more frequent contact of women with the health care system throughout their life (e.g. concerning contraception, pregnancy, and child care) is an opportunity to be more sensitive to health promotion and prevention, contributing to a greater willingness to follow public health and medical recommendations. Subgroup analysis also revealed that gender was a statistically significant moderator of ideal smoking, with a higher percentage in females (81%) compared to males (60%). It could partly be explained by less frequent smoking in women than men.
However, epidemiological evidence suggests that smoking is a stronger cardiovascular risk factor in women [85,86]. This finding and increasing trend in female smoking point out that tobacco control policies should include items specifically targeted at women. Also, there was a statistically significant difference in ideal blood pressure between females (42%) and males (30%) which is in concert with a higher prevalence of raised blood pressure in men compared to women worldwide, but this difference was only statistically significant in the Region of the Americas and the European Region [87].
From the beginning of this century, considerable efforts have been made to improve understanding of the sex/gender differences in CVD and to heighten awareness of heart disease in women [7]. Policy and prevention efforts, to be successful, need to be investigated and targeted within genders. A better understanding of sex differences in CVD is needed to prevent and treat CVD more efficiently in both gender. Continued efforts are required to unravel the belief that CVD is a man's disease [88].
Like in themeta-analysis by Peng et al. [19] our results showed that younger adults had higher overall CVH than older counterparts. We found statistically significant differences in the summary prevalence of ideal CVH among younger (8%), middle-aged (3%), and older participants (1%). The highest prevalence of ideal CVH observed in younger subjects could be explained mostly by their better ideal health factors, although this is in contrast to the poor results on health behaviors reported elsewhere [4,89]. Comparing to other two groups, the oldest participants had lower levels of almost all ideal CVH metrics except for smoking and diet [4,49]. Nevertheless, the young people should be a priority objective for primordial prevention that may prevent worsening of their current CVH in the future.
Although literature data highlights the relevance of SES as determinants of health in both high-income and low-and middle-income countries [90][91][92], only a few studies have examined the association between ideal CVH and SES. Accumulating evidence supports the notion that social environment (SES and social networks) shape the personal lifestyle choices such as physical activity, nutritional choices, and smoking which profoundly influence CVH risk factors. However, the cumulative impact of the environment on CVD risk has been difficult to assess and the mechanisms by which some environment factors influence CVD remain obscure [93]. Education, income or aggregation of assets and household facilities known as the Wealth index [94,95] are most commonly used as proxies for SES. Results from many studies included in our systematic review showed that participants with higher education had a greater number of ideal CVH metrics, i.e. better CVH in comparison with those with lower education. Olsen et al. [5] reported an increasing trend in ideal CVH with a more unfavorable risk profile among persons with low educational levels. Another study [96] found that additional years of education are associated with better CVH. The Tromso study [97] demonstrated the most unfavorable CVD risk factors in the lowest educational group. Our results are also in accordance with the results from the world's largest population-based cross-sectional study performed in the 50 states of the USA [3]. Only two studies [30,48] included in our review assessed the association between Wealth index and ideal CVH. Jankovic et al. [48] found a positive association, while the opposite results were noted in the study by Benziger et al. [30]. Several previous studies failed to find any significant association between the Wealth index and CVH score [98,99]. Overall, the results of this review suggest that those with higher levels of SES have better CVH and a possible explanation could be the fact that people with higher education or income are more prone to afford healthier foods and exercise more compared with those with lower education or income. Also, better-off people compared with those worse-off are more likely to afford high-quality health care.
The strength of our study is that we calculated the pooled prevalence estimates for the 7 ideal CVH metrics and overall CVH according to age and sex on a large sample size (over 2 million people). Besides, this is the first systematic review on the relationship between participants' SES and ideal CVH. However, several limitations should be briefly stressed. First, the results of meta-analyses of ideal CVH metrics and ideal CVH showed a great heterogeneity across studies and it could be explained by the variance in the measurements of ideal CVH metrics in assessed studies, differences in the study populations in terms of gender, age, SES, geographic distribution, lifestyle patterns, and a difference in survey years for each of the included studies. Second, some CVH metrics (smoking, dietary intake, and physical activity) were self-reported measures, which carried an inherent degree of bias. Third, not all studies followed the metric definitions of AHA, especially for a healthy diet or physical activity which may preclude comparisons between studies. To address heterogeneity at least a part, we performed subgroup analyses of each of the 7 ideal CVH metrics and their clustering (6 and 7; and 5 and more ideal metrics) by gender and age groups.

Conclusion
Our systematic review and meta-analysis provided evidence that ideal CVH is low worldwide. We found that gender was a statistically significant moderator of ideal CVH, cluster of 5 and more ideal CVH metrics, ideal smoking, and ideal blood pressure (the better values were observed in females). Also, younger, more educated and better-off individuals had a greater number of ideal CVH metrics, i.e. better ideal CVH in comparison with middle or old age participants, less educated and worse-off. To achieve the improvement of the CVH metrics and the overall ideal CVH, nationwide primordial prevention efforts at the population and individual levels are urgently needed and should focus on the diet, as well as those components which showed gender inequalities, such as smoking and blood pressure. Targeting public health interventions to improve CVH status among less educated and more disadvantaged individuals may have substantial societal implications.