Alcohol and ideal cardiovascular health: The Multi‐Ethnic Study of Atherosclerosis

Background Alcohol consumption is associated with cardiovascular disease (CVD), with moderate drinkers having decreased CVD risk compared to non‐ and heavy drinkers. However, whether alcohol consumption is associated with ideal cardiovascular health (CVH), assessed by the American Heart Association's (AHA) Life's Simple 7 (LS7) metrics, and whether associations differ by sex, is uncertain. Hypothesis Heavy alcohol consumption is associated with worse CVH. Methods We explored associations between alcohol consumption and CVH in a multi‐ethnic population including 6506 participants free of CVD, aged 45 to 84 years. Each LS7 metric was scored 0 to 2 points. Total score was categorized as inadequate (0‐8), average (9‐10) and optimal (11‐14). Participants were classified as never, former or current drinkers. Current drinkers were categorized as <1 (light), 1 to 2 (moderate) and >2 (heavy) drinks/day. Multinomial logistic regression models assessed associations between alcohol and CVH, adjusted for age, sex, race/ethnicity, education, income, and health insurance. Results Mean (SD) age was 62 (10) years, 53% were women. Compared to never drinkers, those with >2 drinks/day were less likely to have average [0.61 (0.43‐0.87)] and optimal CVH [0.29 (0.17‐0.49)]. Binge drinking was also associated with unfavorable CVH. Overall, there was no independent association for light or moderate drinking with CVH. However, women with 1 to 2 drinks/day were more likely to have optimal CVH [1.85 (1.19‐2.88)] compared to non‐drinking women, which was not seen in men. Conclusion Heavy alcohol consumption was associated with unfavorable CVH. Although light or moderate drinking may be associated with a more favorable CVH in women, overall, the association was not strong.


| INTRODUCTION
Globally, heavy alcohol consumption is responsible for considerable socioeconomic burden causing 3.3 million deaths in 2012. 1,2 In the United States, an estimated 62 000 men and 26 000 women die annually from alcohol-related causes 1,3 which makes alcohol the third leading preventable cause of death, behind tobacco use, and poor dietary habits/physical inactivity. 1,4 Alcohol misuse in the United States cost approximately $250 billion in 2010 and 3/4 of this was attributable to binge drinking. 1,5 However, literature is replete with research suggesting an inverse relationship between light to moderate alcohol consumption and cardiovascular disease (CVD), as well as all-cause mortality. 6,7 For example, a review of 84 studies reported that alcohol consumption of 1 drink or less per day was consistently associated with a 14% to 25% reduction in the risk of all the health outcomes evaluated, which included CVD, compared to those who abstained from alcohol. 7 Although the effects of alcohol on CVD has been well researched and documented, 8,9 research exploring the association between alcohol consumption and intermediate measures of cardiovascular health (CVH), assessed using the American Heart Association's Life's Simple 7 (LS7) metrics, 10,11 is sparse. 12 Even less is known about sex, age, and race/ethnic differences in these associations. The LS7 metrics were introduced to measure and monitor the CVH of individuals and populations where ideal CVH is defined as the presence of both ideal health behaviors (non-smoking, body mass index [BMI] <25 kg/m 2 , physical activity at goal levels and diet consistent with guidelines) and ideal health factors (untreated total cholesterol ≤200 mg/dL, untreated blood pressure <120/mg and untreated fasting blood glucose <100 mg/dL). 10,11 We examined the cross-sectional associations between alcohol consumption and ideal CVH using data from the Multi-Ethnic Study of Atherosclerosis (MESA), a large multicenter prospective cohort study. We hypothesized that heavy alcohol consumption will be associated with worse CVH independent of sociodemographic factors, which include age, sex, race/ethnicity, education, income, and health insurance status.

| METHODS
A detailed methodology of MESA has been previously described. 13 Briefly, the MESA study recruited 6814 adults between July 2000 and were used to obtain information from participants. We included 6506 participants from MESA baseline data after excluding those without complete information on the LS7 metrics (n = 308). Details of the assessments of the Life's Simple 7 metrics can be found in Appendix S1, Supporting Information.
Alcohol consumption was assessed from questions in a personal history questionnaire. Participants were asked the following questions: "Have you ever consumed alcoholic beverages?" If yes, they were also asked "Do you presently drink alcoholic beverages?" Answers given to these questions were used to classify participants into never, former, and current drinkers. Current and former drinkers were asked about the number of years of drinking, and the usual number of drinks consumed per week (before they stopped drinking in the case of former drinkers). In addition, current drinkers were asked about the number of drinks consumed in the past 24 hours and the largest number of drinks consumed in 1 day in the past month. These questions were used to know participants who engage in binge drinking, defined as the consumption of 5 or more drinks on 1 occasion in the past month. 13,14 Information was collected on sociodemographic variables, which included age, sex, race/ethnicity, educational attainment, income, and health insurance. 13 In our stratified analysis, age, and sex had two categories, <65 and ≥65 years; men and women, respectively. Race/ethnicity had four categories, White, Chinese American, Black, and Hispanic. Education was classified as ≥bachelor's degree and < bachelor's degree. Income was divided into participants who made ≥$40 000 and < $40 000 while health insurance was grouped into "Yes" and "No."

| Statistical analysis
The characteristics of the study participants were reported by categories of alcohol consumption. Categorical variables were presented as frequencies with percentages, and continuous variables were presented as means with standard deviation (SD). Descriptive statistics were used to compare the baseline characteristics of all participants by alcohol consumption categories, using ANOVA for continuous variables and χ 2 tests for categorical variables.
The LS7 metrics were each categorized into ideal, intermediate, and poor, 10 as shown in Table S1. Points were awarded to each category with 0 indicating poor; 1, intermediate; and 2, ideal. The points were summed, yielding a total CVH score ranging from 0 to 14. 15 As previously reported, CVH scores of 0 to 8, 9 to 10, and 11 to 14 were considered inadequate, average, and optimal, respectively. [16][17][18][19][20][21] We reported the proportions of the individual LS7 metrics by alcohol consumption categories. Multinomial logistic regression modeling was used to examine the association between alcohol consumption and CVH score. Two separate regression models were fitted. Model 1 was unadjusted and model 2 was adjusted for sociodemographic factors (age, sex, race/ethnicity, education, income, and health insurance status). Odds ratios (ORs) and their 95% confidence intervals (CIs) were calculated for average and optimal CVH score across the categories of alcohol consumption. Consistent with prior studies, 14,16,18 the reference groups were the inadequate score for the CVH categories and the "never" categories for alcohol consumption and binge drinking.
We examined the interaction of alcohol consumption with sex, race/ethnicity, and age using the likelihood ratio χ 2 test, by including interaction terms in model 2. The association between alcohol consumption and CVH score was stratified by sex, race/ethnicity, and age. All analyses were performed using STATA version 14.1 (StataCorp LP, College Station, Texas) and a two-sided P-value <0.05 was considered statistically significant.

| RESULTS
The baseline characteristics of the 6506 study participants varied across alcohol consumption categories as reported in Table 1. Fiftythree percent of participants were women, and the mean age (SD) was 62 (10) years. Among participants, 20% were never drinkers, 24% were former drinkers and 56% were current drinkers. Of the current drinkers, 78% consumed <1 drink/day (mild); 16%, 1 to 2 drinks/ day (moderate) and 6%, >2 drinks/day (heavy). In addition, 14% of current drinkers reported binge drinking in the past month. Of note, healthy diet score decreased as alcohol consumption increased. For the overall cohort, 47%, 33%, and 20% met criteria for inadequate, average and optimal CVH scores, respectively. The proportion of participants with optimal CVH scores were lowest in the >2 drinks/ day and binge drinking categories.
The distribution of the individual LS7 metrics by alcohol consumption in the study population is reported in Table 2. With the exception of physical activity and blood glucose, the proportion of never drinkers who met the ideal criteria for the LS7 metrics was larger than that for current drinkers who consumed more than 2 drinks/day. None of the participants who reported more than 2 drinks/day or binge drinking met the ideal criteria for diet.
The associations between alcohol consumption and CVH in the overall cohort and stratified by sex, age, and race/ethnicity are reported in Table 3, Table S2    Abbreviation: OR, odds ratio. *Adjusted for socio-demographic factors: age, sex, race/ethnicity, education, income and health insurance status; OR < 1 is interpreted as decreased odds of having an optimal or average cardiovascular health score. Results in italicized font are statistically significant; P < 0.05.

FIGURE 1
Mean CVH score (and SE) for each alcohol intake category. First bar: Total alcohol consumption; second bar: Never drinkers; third bar: Former drinkers; fourth bar: <1drink/day; fifth bar: 1 to 2 drinks/day; sixth bar: >2 drinks/day; seventh bar: Binge drinking in the past month. CVH score ranges from 0 to 14. CVH, cardiovascular health drinking: P = 0.024 for optimal scores). However, the interaction between alcohol consumption and race/ethnicity or age was not statistically significant.
The distribution of the mean CVH score by alcohol consumption categories are shown in Figure 1. Mean CVH score was slightly lower for former drinkers and participants who consumed more than 2 drinks/day as well as for those who reported binge drinking ( Figure 1). The associations of alcohol consumption with each LS7 metric are shown in Table S4. Regardless of quantity of alcohol consumed, participants had lower odds of achieving the ideal criteria for smoking. Participants with <1 and 1 to 2 drinks/day had higher odds of achieving the ideal criteria for physical activity, BMI, and blood glucose. For total cholesterol, those who reported >2 drinks/day were 54% less likely to meet the ideal criteria, while for blood pressure, those who reported light intake (<1 drink/day) were 38% more likely to meet the ideal criteria.

| Main findings
In this multi-ethnic cohort study of adults free of CVD at baseline, we found that after adjusting for sociodemographic characteristics, most study participants who reported <1 drink/day (light) and 1 to 2 drinks/day (moderate) were more likely to have average and optimal CVH scores, but the association was only statistically significant for women. In contrast, in the overall cohort and in the stratified analyses by sex, age, and race/ethnicity, participants who consumed more than 2 drinks/day (heavy) or reported binge drinking in the past month had lower odds of having average and optimal CVH scores compared to never drinkers or those who did not report binge drinking, respectively. However, because of the cross-sectional design of this study, we cannot prove a causal relationship between alcohol consumption and CVH.

| Comparison to previous studies
Previous studies have examined the association between alcohol and CVD. 6 reported excessive alcohol consumption is a common cause of reversible hypertension, non-ischemic dilated cardiomyopathy, atrial fibrillation, and stroke. The authors also reported that though light to moderate alcohol consumption (≤1 drink/day for women and 1-2 drinks/day for men) was associated with lower risk for all-cause mortality, coronary heart disease, type 2 diabetes, heart failure, and stroke, the risk-benefit ratio of drinking may be less favorable in younger individuals, because they are more likely to abuse alcohol which is a leading cause of premature death among people between 15 and 59 years. 22

| Possible mechanisms
The main finding of our study showed light, and in some participants moderate, alcohol consumption was associated with favorable CVH scores and most ideal metrics. Participants who drank any alcohol did not meet the ideal criteria for smoking. This is not surprising because alcohol consumption and smoking are closely related behaviors. 23 Among alcoholics, the rate of smoking is approximately 90% and may be attributable to neurobiological mechanisms, such as crosstolerance and cross-sensitization to both nicotine and alcohol. 23  Prior research has demonstrated heavy drinking increases blood pressure mainly through stimulation of the sympathetic nervous system and increase in intracellular calcium in the vascular smooth muscle. [27][28][29] Another interesting finding of our study was that sex was an effect modifier of the association between alcohol and CVH. Similar to prior research, 30 we found that women had a larger proportion of abstainers compared to men (29% vs 10%) while men had a larger proportion of heavy (6% vs <1%) and binge drinkers (14% vs 3%). These differences in drinking habits may be driven by psycho-socio-cultural factors, such as the perceived differences in traditional gender roles observed in different ethnic and age groups. [30][31][32][33][34] However, recent studies show that the differences reported in alcohol consumption among men and women is diminishing as the socioeconomic and cultural characteristics of younger people become more similar regardless of sex. 30 37 This is consistent with our findings of an association with a favorable CVH profile for light drinking but an unfavorable CVH profile for heavy or binge drinking.

| Strengths and limitations
Our study has many strengths including the large, ethnically diverse population that allowed for stratification by sex, age, and race/ethnicity. In addition, data collection for the measurements of CVH and alcohol were performed using standardized methods and procedures.
This study also has limitations. First, the use of self-report questionnaires may have introduced recall bias for the collection of data on alcohol, smoking, physical activity, and diet. Second, because some participants may have given socially acceptable responses to questions about alcohol consumption, there is a possibility that alcohol consumption may have been under-reported which may lead to nondifferential misclassification and bias the associations toward the null.
Third, this study is observational, and although we adjusted for confounding factors such as demographics and socioeconomic status, there may be residual confounding factors that distinguish people who chose to drink vs not. Fourth, the study is cross-sectional so we are unable to determine temporality of the associations seen. Fifth, many of the associations examined may not have been statistically significant because of multiple comparisons. We were unable to make inferences regarding ethnic differences for Chinese Americans and Blacks with optimal CVH who reported >2 drinks/day and for Chinese Americans who reported binge drinking due to the relatively small sample sizes in these categories.

| CONCLUSION
In summary, using the relatively new health metric of the LS7 criteria, 10 this study explores the relationship between alcohol and CVH, which represents a paradigm shift from a focus on CVD prevention to promotion of CVH and wellness. 11 Given that alcohol consumption is common in developed countries, our study was designed to further the understanding of the relationship of various levels of alcohol consumption with CV health and behavioral factors. We showed that light alcohol consumption was associated with favorable CVH, whereas heavy alcohol consumption was unfavorable. However, we do not recommend that abstainers start drinking because alcohol has a high addiction potential and several adverse health consequences. Current drinkers should limit consumption to ≤100 g per week in line with recently published research that shows this threshold to be associated with a lower risk of CVD and all-cause mortality. 37 Future research could investigate the association between alcohol consumption and CVH in a younger cohort because they consume a higher amount of alcohol and binge drinking is more commonly reported among them. 22 As suggested in prior research, 8 biomarkers of alcohol, such as PEth can be employed to corroborate self-reported use of alcohol and differentiate between light, moderate, and heavy consumption. 8 In addition, the role sex plays as an effect modifier of the association between alcohol consumption and CVH could be further examined in future studies.