1. Introduction
In 2010, the American Heart Association (AHA) designed and promoted a set of seven cardiovascular health (CVH) metrics (Life’s Simple 7 [LS7]) aimed at reducing cardiovascular disease (CVD) [
1]. The metrics are divided into four health behaviors (smoking, weight, physical activity, and diet) and three health factors (blood pressure, total cholesterol, and glucose). Higher attainment of ideal LS7 metrics is associated with a lower risk of incident type 2 diabetes (diabetes), CVD and all-cause mortality [
2,
3,
4]. There is currently limited exploration of LS7 with biological hormones that contribute to adverse cardiometabolic outcomes. Higher aldosterone has been associated with higher incidence of metabolic syndrome, diabetes, CVD and all-cause mortality among African Americans (AA) and thus represents one such potential hormone [
5,
6,
7]. Additionally, there is limited data on the association of the individual LS7 metrics with aldosterone among AAs. Previously, studies have shown positive associations of blood pressure (BP) [
7,
8], fasting plasma glucose (FPG) [
5], body mass index (BMI) [
8], and cholesterol with aldosterone in majority non-Hispanic white (NHW) studies [
8,
9,
10,
11,
12]. There is a paucity of data regarding the association of smoking, physical activity (PA), and diet with aldosterone and no studies assessing the association of the combination of modifiable risk factors assessed by the LS7 with aldosterone in any racial/ethnic group. Given the higher prevalence of diabetes, CVD and mortality among AAs compared to NHWs [
3], it is critical to determine the relationship of LS7 with biological hormones important in cardiometabolic disease. Thus, the goal of this study was to examine the association of LS7 with aldosterone, characterizing the association of combined metrics of LS7 with aldosterone as our primary outcome and the association of individual metrics of LS7 with aldosterone as our secondary outcomes. We hypothesized an inverse association of attainment of LS7 metrics with serum aldosterone.
3. Results
Among 4095 adults (mean age 55 ± 13 years, 65% female), there were 1132 (28%) participants with poor LS7 scores, 2288 (56%) with intermediate LS7 scores and 675 (16%) with ideal LS7 scores (
Table 1). Participants in the ideal LS7 score category had a more optimal cardiovascular profile with lower glucose, SBP, DBP, BMI, total cholesterol and higher levels of educational attainment, professional occupational status, and eGFR compared to intermediate and poor categories (
p-value for comparisons < 0.0001). Median serum aldosterone was 4.90, 4.30, and 3.70 ng/dL in the poor, intermediate and ideal categories, respectively [
p for trend = 0.004]. Serum potassium, chloride and potassium intake were modestly higher, whereas serum sodium and creatinine were modestly lower across categories of poor, intermediate, and ideal LS7.
Median serum aldosterone levels stratified by age, sex, diabetes status, and eGFR are shown in
Table 2. Median aldosterone was higher in men (4.80 ng/dL) compared to women (4.10 ng/dL), in participants with diabetes (4.70 ng/dL) compared to participants without diabetes (4.20 ng/dL) and in participants with eGFR below median (5.00 ng/dL) compared to participants with eGFR above median (3.80 ng/dL) [
p < 0.0001 for all comparisons]. The findings remained significant after exclusion of participants taking medications that alter aldosterone (
Table 3). Sensitivity analysis conducted using the total LS7 score revealed similar findings for median aldosterone (
Table S4).
The associations of the categorical and continuous LS7 scores with aldosterone are shown in
Table 2. Aldosterone in the ideal and intermediate categories of LS7 score was 33% (Mean Ratio: 0.67, 95%CI: 0.61, 0.75) and 15% (Mean Ratio: 0.85, 95%CI: 0.80, 0.90) lower compared to the poor category. For continuous LS7 score, each additional ideal LS7 metric attained [a one-unit increase in score (0–7)] was associated with 11% (Mean Ratio: 0.89, 95%CI 0.86, 0.91) lower aldosterone, with similar estimates after adjustment for aldosterone altering medications. Sensitivity analysis conducted using the total LS7 score revealed significant mean ratios (
Table S4). Additionally, excluding participants taking medications that alter aldosterone revealed similar findings where aldosterone in the ideal and intermediate categories of LS7 score was 38% (Mean Ratio: 0.62, 95%CI: 0.55, 0.70) and 27% (Mean Ratio: 0.73, 95%CI: 0.67, 0.79) lower compared to the poor category (
Table 3). For continuous LS7 score, each additional ideal LS7 metric attained [a one-unit increase in score (0–7)] was associated with a 14% (Mean Ratio: 0.86, 95%CI 0.83, 0.89) lower aldosterone.
Effect modification of the association of LS7 with aldosterone was significant for age (
p < 0.001), sex (
p < 0.001 for continuous and
p = 0.0021 for categorical LS7 scores), eGFR (
p < 0.001) and diabetes (
p = 0.016 for categorical LS7 score) among all participants (
Table 2). Sex stratified analyses among all participants (
Table 2) revealed that women had 39% (Mean Ratio 0.61, 95% CI: 0.53, 0.70) whereas men had 19% (Mean Ratio 0.81, 95% CI: 0.69, 0.96) lower aldosterone in ideal vs. poor LS7 score categories. For continuous LS7 score association with aldosterone, for each additional LS7 metric attained, women had 15% lower aldosterone (Mean Ratio: 0.85, 95% CI: 0.82, 0.88) vs. 5% lower aldosterone (Mean Ratio 0.95, 95% CI: 0.91, 0.98) in men. These findings remained significant after excluding participants taking aldosterone altering medications (
Table 3), whereas findings for age, eGFR and diabetes status were attenuated and non-significant after excluding individuals taking aldosterone altering medications (
Table 2 and
Table 3). Excluding individuals below the lower detection limit of the aldosterone assay did not alter the significance of our findings (
Table S3).
Among the individual LS7 metrics (
Table 4), ideal categories of smoking, BMI, blood pressure, total cholesterol, and fasting plasma glucose were associated with lower aldosterone compared to poor categories (all comparisons
p < 0.05). The intermediate vs. poor categories of individual LS7 metrics were significant for fasting plasma glucose (
p < 0.0001). There was significant effect modification by sex, and in stratified models, generally a greater magnitude of mean ratios for the ideal vs. poor category comparisons for women as compared to men, except for BMI. Notably, the greatest difference was for ideal vs. poor categories of smoking with 4% lower aldosterone in men (mean ratio 0.96, 95% CI: 0.83, 1.10) vs. 42% lower aldosterone in women (mean ratio 0.58, 95% CI: 0.53, 0.64), which remained significant in sensitivity analysis excluding individuals taking aldosterone altering medications (
Table S5).
4. Discussion
The JHS is an AA community-based cohort study with well-characterized data on LS7 metrics and aldosterone. This first report of the association of LS7 with aldosterone demonstrates a strong inverse association of LS7 scores with aldosterone, with a greater magnitude of effect among women. Among the individual metrics attaining ideal vs. poor categories of glucose, blood pressure, BMI, smoking status, and total cholesterol were associated with lower aldosterone, with a generally greater magnitude in women except for BMI. This is also the first study to reveal higher serum aldosterone in AA men compared to women, independent of medications that alter aldosterone.
4.1. The Association of LS7 Score with Aldosterone
Although the individual metrics of LS7 have been associated with lower long-term CV risk in multiple studies, the biological underpinnings remain understudied. Here we show that more optimal levels of five of the seven metrics and the overall score are significantly associated with lower aldosterone, independent of aldosterone altering medications. We are aware of no other studies examining the association of similar combinations of lifestyle modifiable risk factors with aldosterone. Although the difference between the aldosterone values is small (highest to lowest 1.20 ng/dL), prior data revealed significant increases in the risk of diabetes with similar changes in aldosterone emphasizing the biological significance [
5]. Racial differences in aldosterone were noted previously with lower aldosterone levels among AAs compared to NHWs, Hispanics, and Chinese Americans [
21]. However, this study reveals novel sex differences with higher serum aldosterone among men, but a greater magnitude of association of higher LS7 score with lower aldosterone among women. We initially hypothesized these findings may be secondary to higher aldosterone in women compared to men as had been shown previously among NHWs in the Framingham Heart Study, which collected aldosterone using similar parameters [
10]. Discordant with the Framingham Heart Study findings, we found a 17–18% higher aldosterone among men in the JHS independent of aldosterone altering medications. A recent study by Toering et al among whites showed that men on a high-salt diet (4.6 grams/day) had higher serum aldosterone than women, with no aldosterone sex differences on a low-salt diet [
22]. In the JHS, the average dietary sodium intake was 3.5 grams per day [
5], much closer to the high-salt diet in the study, and thus our finding of higher aldosterone in men is consistent with this study, but does not explain the greater magnitude of association of higher LS7 score with lower aldosterone among women. Toering et al also evaluated the adrenal response to exogenous angiotensin II which was significantly higher in women [
22]. Thus, there may be sex differences in the regulation of aldosterone with greater biological adrenal responsiveness among women. This hypothesis is supported by a recent finding of 15% greater aldosterone response to Ang II on both restricted and liberal salt diets among women in the majority NHW Hypertensive Pathotype (HyperPATH) consortium [
23]. In a rodent model of aldosterone-mediated cardiovascular disease, myocardial damage and proteinuria were greater in female rats despite having similar blood pressure responses [
23]. Thus, our finding of sexual dimorphism in relationship of LS7 score with aldosterone is potentially secondary to greater responsivity of aldosterone to endogenous aldosterone agonists in women.
4.2. The Association of Individual LS7 Health Factors and Health Behaviors with Aldosterone: Human Studies and Mechanisms
4.2.1. Cholesterol
Total cholesterol is the overall amount of cholesterol in the blood including a combination of high-density lipoprotein (HDL), low-density lipoprotein (LDL), and triglycerides, calculated as HDL + LDL + (triglycerides/5). There is limited data on the association of aldosterone with total cholesterol. Among majority NHWs, aldosterone has been positively associated with total cholesterol/HDL ratio [
10], triglycerides, and inversely associated with HDL [
11]. In smaller studies of AA adults, there have been inconsistent findings for the association of cholesterol with aldosterone including positive associations with triglycerides but not total cholesterol [
24], positive associations of total cholesterol and triglycerides with standing but not supine aldosterone [
8], and an inverse association of HDL with aldosterone with no reported association with total cholesterol or triglycerides [
9]. In the JHS, aldosterone was positively associated with change in triglycerides, but not HDL over 4 years with no previous assessment of total cholesterol [
7]. From a mechanistic perspective, cholesterol regulates aldosterone synthesis and regulation via various lipoprotein components. Very low-density lipoprotein through various signaling pathways (STaR and aldosterone synthase) induces aldosterone synthesis [
12]. LDL provides substrate (cholesterol) for aldosterone synthesis, thus increases aldosterone levels [
25]. Aldosterone production in human adrenocortical cells is stimulated by HDL2 through increased expression of aldosterone synthase (
CYP11B2) [
26]. Consistent with regulatory function of cholesterol, statins lower aldosterone levels [
20]. Thus, our finding of a positive association of aldosterone with total cholesterol is the first report in a large population of AAs and is supported by mechanistic studies.
4.2.2. Glucose
In this study, ideal (FPG < 100 mg/dL) and intermediate (FPG 100–125 mg/dL) glucose categories compared to poor (≥126 mg/dL) were associated with lower aldosterone. Aldosterone excess impairs insulin secretion and insulin sensitivity [
27]. Aldosterone increases insulin resistance through inhibition of insulin signaling and insulin-stimulated glucose uptake via glut-4 translocation in adipocytes, skeletal muscle, and vascular smooth muscle cells [
28]. Additionally, aldosterone impairs adipokines and nuclear receptors that improve insulin sensitivity through adipose tissue inflammation including adiponectin and peroxisome proliferator-activated receptor-gamma [
29]. Among AAs in the JHS, aldosterone was positively associated with insulin resistance, glucose, and change in glucose over 4 years, consistent with studies in other racial/ethnic groups [
5,
7].
4.2.3. Blood Pressure
Aldosterone is associated with higher blood pressure, increases in blood pressure over time, and predicts hypertension among AAs [
7,
8,
30] and whites in single race/ethnicity studies [
31,
32]. In studies including both AAs and whites, the findings have been mixed, correlations of blood pressure with aldosterone have been more consistent and striking in AAs throughout the lifespan with some studies showing no association in whites [
30,
33]. In the Multi-Ethnic Study of Atherosclerosis, there was no association of aldosterone with blood pressure among NHWs or AAs [
21]. Consistent with the majority of studies in AAs, ideal vs. poor blood pressure categories were associated with lower aldosterone in this analysis.
4.2.4. Body Mass Index
Measures of adiposity including BMI, visceral adipose tissue, and waist circumference have been positively associated with aldosterone among NHWs [
9,
34]. The findings from smaller populations of AA adults support a cross-sectional association of anthropometric measures of adiposity including body mass index, waist circumference and waist:height ratio with aldosterone [
8,
9]. Among AAs in the JHS, aldosterone was not associated with change in waist circumference over 4 years, but cross-sectional measures have not been assessed [
7]. This study reveals a positive association of body mass index with aldosterone, consistent with the findings among NHWs and provides further evidence for a significant relationship among AAs. Potential mechanisms for the association of BMI with aldosterone are adipokines. Higher levels of BMI are associated with higher levels of leptin and lower levels of adiponectin [
35]. Leptin is a direct regulator of
CYP11B2 (aldosterone synthase) expression and aldosterone production in adrenal zona glomerulosa cells [
36]. Adiponectin on the contrary decreases steroidogenesis including aldosterone production [
37]. There is also a reciprocal interaction between aldosterone and adiposity with aldosterone induced mineralocorticoid receptor activation promoting inflammation, adipocyte differentiation, and alteration of adipokine expression (higher leptin/lower adiponectin) leading to further production of aldosterone [
38,
39]. Additionally, adipocytes possess aldosterone synthase and can produce aldosterone independently and stimulate production of a liver-derived adrenal aldosterone secretagogue [
39]. The BMI-aldosterone association among AAs in the JHS suggest that studies showing weight loss lowers aldosterone levels, may also be relevant for AAs [
34,
40,
41].
4.2.5. Physical Activity
Prior studies analyzing the relationship between aldosterone and physical activity have been inconclusive. Some studies reveal that physical activity in the form of aerobic exercise decreases aldosterone levels [
42,
43], while others have shown no effect of aerobic exercise on aldosterone [
44,
45]. Interestingly, there may be racial/ethnic differences with one study revealing lower aldosterone after aerobic exercise training among whites but not AAs [
44]. Long-term endurance training has not been associated with reductions in aldosterone in the young or elderly [
46]. Physical activity was not associated with aldosterone in our study. Further, epidemiologic and clinical trial research with objective measurement of physical activity is needed to improve the understanding of acute and chronic effects of physical activity on aldosterone.
4.2.6. Smoking
Ideal vs. poor smoking had the greatest magnitude of association with lower aldosterone in this study. Previous data on the association of acute and chronic smoking with aldosterone are mixed. One study noted that aldosterone levels rise acutely after smoking with a peak at 30-minutes [
47]. Laustiola et al noted that baseline aldosterone levels in chronic smokers are higher compared to non-smoking identical twins [
48]. However, some population based studies have not shown significant differences in aldosterone levels between smokers and non-smokers [
49]. In this study, notable sex differences were found with a 4% non-significant lower aldosterone in the ideal (non-smoking) compared to poor group (current smoking) among men and 42% significantly lower aldosterone among women. Cigarette smoke contains over 4000 chemicals; one of the major chemicals is nicotine. In cultured human endothelial cells, nicotine increases the expression and activity of angiotensin-converting enzyme [
50,
51]. Angiotensin converting enzyme converts angiotensin I to angiotensin II and angiotensin II stimulates secretion of aldosterone from the adrenal cortex. Thus, nicotine increasing angiotensin II through higher angiotensin converting enzyme activity leading to higher aldosterone in smokers, particularly women, due to the blunted adrenal response to angiotensin II in men, may provide one explanation for the sex difference in the smoking–aldosterone association [
22,
23]. Given the immense number of chemicals in cigarettes, further epidemiological and preclinical studies examining the effects of smoking on aldosterone are needed to clarify this relationship and potential sexual dimorphism.
4.2.7. Dietary Intake
The role of dietary intake consistent with the AHA ideal diet, including fruits and vegetables, fish, fiber-rich whole grains, low sodium diet, and limited intake of sugar sweetened beverages has not been examined in relation to aldosterone. Prior studies have shown that low sodium diets increase plasma aldosterone levels [
52]. The ideal diet according to AHA LS7 was not associated with lower aldosterone levels in our study. One of the challenges is accurately measuring dietary intake via subjective measures such as food frequency questions. Further studies focusing on dietary intake and aldosterone may be helpful in delineating the relationship.
4.3. Strengths and Limitations
The strengths of our analysis include a large, contemporary, population-based cohort with rigorously ascertained physiologic and laboratory measures. Additionally, the JHS used validated questionnaires and has a comprehensive documentation of medications, which are critical to categorizing participants into the correct LS7 categories. Despite these and other strengths, there are some potential limitations. First, this is a cross-sectional study, so neither temporality nor causation can be ascertained. Second, the self-reported diet and physical activity were not associated with aldosterone potentially due to misclassification across the categories. Also, only 19% and 0.9% achieved ideal levels of physical activity and diet, respectively, which may have limited the ability to detect significant differences. Third, some molecules that may potentially impact with aldosterone were not collected (blood urea nitrogen, plasma and urinary catecholamines) or were not available in the full cohort (plasma renin activity). Fourth, renin–angiotensin–aldosterone medications may have impacted the aldosterone levels. Thus, sensitivity analysis was performed excluding participants on medications that influence aldosterone including ACE-inhibitors, angiotensin receptor blockers, mineralocorticoid receptor antagonists, and statin medications, which revealed persistent associations of LS7 score with aldosterone. Lastly, the participants in the JHS are from one geographic area in the southeastern United States and may not be representative of all AAs.