OUP user menu

Serum aldosterone and its relationship to left ventricular structure and geometry in patients with preserved left ventricular ejection fraction

Frank Edelmann, Andreas Tomaschitz, Rolf Wachter, Götz Gelbrich, Manuela Knoke, Hans-Dirk Düngen, Stefan Pilz, Lutz Binder, Raoul Stahrenberg, Albrecht Schmidt, Winfried März, Burkert Pieske
DOI: http://dx.doi.org/10.1093/eurheartj/ehr292 203-212 First published online: 19 August 2011

Abstract

Aims Cardiac remodelling might be an important mechanism for aldosterone-mediated cardiovascular (CV) morbidity and mortality. Previous studies relating aldosterone to left ventricular (LV) structure however revealed conflicting results.

Methods and results We aimed to evaluate the relationship of serum aldosterone concentration (SAC) and aldosterone-to-renin ratio (ARR) with echocardiographic parameters of LV remodelling in CV risk patients with preserved left ventricular ejection fraction (LVEF). We studied 1575 participants (54.1% female) with CV risk factors and LVEF >50% (61.7 ± 6.1%). Of the total, 94.7% of patients had no overt heart failure. All patients underwent measurement of SAC, ARR, and comprehensive echocardiographic analysis. Overall, multivariate adjusted analysis of covariance (ANCOVA) showed a significant increase in LV mass (P= 0.001), LV mass index (P= 0.001), relative wall thickness (P= 0.011), and LV posterior wall thickness (P< 0.001) with increasing SAC. This overall association of SAC and LV remodelling was driven by a statistic significant effect exclusively in women. In multivariate logistic regression analysis higher SAC levels were independently related to concentric LV hypertrophy [odds ratio (OR; with 95% CI) by comparing SAC levels in the third gender-specific tertile with the first tertile: 1.87; 95% CI: 1.31–2.68; P= 0.001]. Higher SAC levels were positively related to concentric LVH in either sex. We observed no significant associations between the ARR and echocardiographic parameters of LV remodelling.

Conclusion Circulating aldosterone but not ARR levels are independently related to echocardiographic parameters of LV structure, particularly in women. Higher SAC however was related to concentric LVH in either sex. Our findings in a large CV risk cohort with preserved LVEF indicate aldosterone-mediated pro-hypertrophic effects as a potential pathway for structural alterations of the left ventricular myocardium.

  • Serum aldosterone
  • Left ventricular structure and geometry

Introduction

The adrenal hormone aldosterone regulates salt and water homeostasis via binding to the mineralocorticoid receptor (MR) in the kidney. Beside of the classical binding sites of aldosterone to MR such as the renal collecting duct and the salivary glands, MRs have been additionally localized in non-epithelial tissues, i.e. in the myocardium. Absolute aldosterone excess in terms of primary aldosteronism has been associated with higher risk of arterial hypertension, kidney damage, and left ventricular (LV) hypertrophy.1 However, an increasing body of evidence points to aldosterone as a major cardiovascular risk factor in the absence of primary aldosteronism and independent of its primary stimulator angiotensin II.2

In the last years various experimental and clinical studies revealed that apart from neurohormonal activation in the setting of heart failure (HF), pre-existing tissue damage, high dietary salt intake, increased reactive oxygen species-generation and presumably lower kidney function might be important pre-requisites for the initiation of aldosterone-mediated deleterious effects.3 Accordingly, it has been recently shown that circulating aldosterone levels within the physiological range are strongly related to increased risk of cardiovascular mortality, fatal stroke, and sudden cardiac death.4 Consistent with this finding Pitt et al. and Zannad et al.57 impressively demonstrated an improvement of survival after MR-blockade in patients with severe HF but presenting with normal aldosterone levels. Further interventional studies revealed that MR-blockade effectively attenuates LV hypertrophy in hypertensive patients and in patients with mild to moderate HF, respectively.8,9

Several experimental investigations in salt fed animals documented profibrotic and pro-hypertrophic effects of aldosterone independent of arterial blood pressure (BP) and circulating plasma volume.10 In view of LV hypertrophy as a major independent risk factor of HF, sudden cardiac death and cardiovascular mortality studies in humans revealed conflicting results regarding the relationship of circulating aldosterone levels to cardiac structure. Most of these studies however were limited by small sample size, selection bias, and inappropriate consideration of confounding factors.

In view of the inconsistency regarding the associations between circulating aldosterone levels and cardiac structure, we aimed to evaluate the gender-specific relationship of serum aldosterone and aldosterone-to-renin ratio (ARR) levels, which reflect inappropriate aldosterone secretion, to LV structure and pathological patterns of LV geometry.

Methods

Study characteristics

The ongoing, multi-centre observational trial DIAST-CHF on prevalence and clinical course of diastolic dysfunction (DD) and HF with preserved ejection fraction (HFpEF) is part of the nationwide German Competence Network Heart Failure project.11 Patients (aged between 50 and 85 years) who revealed at least one risk factor for DD or HFpEF (defined as history of arterial hypertension, diabetes mellitus, sleep apnea syndrome, or atherosclerotic disease) or a documented history of HF were prospectively included into the DIAST-CHF study after reference to primary care physicians. Only patients with an inability to participate or give written consent, i.e. due to severe concomitant diseases or geographic reasons were not included. For the present analyses, we excluded patients with systolic dysfunction [left ventricular ejection fraction (LVEF) <50%] or with atrial fibrillation. Heart failure with preserved ejection fraction was diagnosed in accordance with recent ESC guidelines and with Framingham diagnostic criteria.12,13

The majority of 1935 patients (1728) who were included in DIAST-CHF were from Goettingen and Berlin, Germany. Serum aldosterone concentration (SAC) was measured in 1575 patients (78.8%) who revealed an LVEF ≥50% and sinus rhythm at presentation. No statistical differences were seen for age, gender distribution, systolic and diastolic BP, LV mass, N-terminal pro B-type brain natriuretic peptide (NT-proBNP) levels, intake of antihypertensive medication, and the presence of chronic HF in participants with and without SAC measurements, respectively. Out of n= 1557, in n= 78 (5.3%) patients HFpEF was diagnosed. Since associations reported below were not significantly different (P> 0.05) between patients with or without overt HFpEF, this group was not listed separately.

The study complied with the Declaration of Helsinki, the protocol was approved by the responsible Ethics Committee and all patients provided written informed consent.

Comprehensive echocardiographic analysis of cardiac function and structure was performed by experienced physicians on Sonos 5500 device (Hewlett-Packard, Andover, MA, USA) according to current guidelines of the American Society of Echocardiography. In addition, randomly chosen echo examinations were reviewed by the echo core laboratory at the University of Essen for quality assurance. Dimensions were recorded by standard techniques. Interventricular septum, LV posterior wall thickness, LV end-diastolic diameter (LVEDD), and end-systolic diameter (LVESD) were determined by M-mode or anatomically, as appropriate. Left ventricular end-diastolic volume (LVEDV), left ventricular end-systolic volume (LVESV), and LV ejection fraction (EF) were measured by the modified Simpson's method. Left ventricular mass (g) was calculated by the Devereux formula and also expressed as LV mass index (g/cm2).14 Relative wall thickness was calculated by the formula LV posterior wall thickness x2/LVEDD.

Left ventricular hypertrophy was defined as an LV mass index exceeding 110 g/m2 in women and 125 g/m2 in men, respectively, and the cut-off point for abnormal relative wall thickness was 0.44.15 Concentric remodelling (CR) was defined as elevation of relative wall thickness but normal (gender-specific) LV mass index, concentric LV hypertrophy was defined as a gender-specific elevation of LV mass and relative wall thickness and eccentric LV hypertrophy as a gender-specific elevation of LV mass but normal relative wall thickness.16

Diastolic dysfunction was characterized according to the ASE guidelines and graded by an algorithm [(i) normal diastolic function; (ii) mild DD; (iii) moderate DD; and (iv) severe DD], which has been defined in the study protocol and described previously in detail.17,18

Definition of covariates

Body mass index (BMI) was calculated as weight in kilogram divided by height in meter squared. History of coronary heart disease (CHD), hyperlipidaemia, current smoking status, chronic obstructive pulmonary disease, alcohol abuse, history of cancer, and sleep apnea syndrome were evaluated at presentation through a patient provided medical history. Detailed ongoing medication [ACE-inhibitors (ACE-I), angiotensin-II type-1 receptor blockers (ARB), beta-blockers, calcium channel blockers, diuretics, and statins] was recorded at baseline. In addition, only 14 participants (0.9%) took MR-blockers at baseline. Diabetes mellitus was defined according to the current recommendations of the American Diabetes Association as fasting glucose >7.0 mmol/L or 2 h glucose >11.1 mmol/L after performing an oral glucose tolerance test.19 Arterial hypertension was defined according to the JNC VII report (Seventh Report of the Joint National Committee for Prevention, Detection, Evaluation, and Treatment of Hypertension) as systolic/diastolic BP >140/90 mmHg and/or ongoing intake of antihypertensive medication. Systolic and diastolic BP was measured at baseline by experienced physicians under standardized conditions. The heart rate was recorded after resting via electrocardiography at baseline. Estimated glomerular filtration rate (eGFR) was calculated according to the MDRD (modification of diet in renal disease) study formula to estimate GFR. All patients at presentation were on a normal Western (sodium) diet.

Laboratory methods

Serum aldosterone concentration (pmol/L; conversion factor pg/mL × 2.744 = pmol/L) was determined according to manufacture's instructions by the Direct Aldosterone ELISA kit (Aldosterone ELISA, IBL GmbH, Hamburg, Germany). The intra-assay and inter-assay coefficients of variation of this assay were 4.1–10.4 and 9.4–9.7%, respectively. Plasma renin concentration (PRC, µIU/mL) was measured using a fully automated chemiluminiscent assay for the direct measurement of renin (LIAISON® Direct Renin, DiaSorin, Saluggia, Italy). This newly developed assay reportedly has suggested to have an improved functional sensitivity of up to <2 mU/L (measuring range: 0.5–500 µIU/mL; referenced to WHO68/356). The analytical sensitivity and functional sensitivity was 0.5 and < 1.96 µIU/mL, respectively. The ARR (SAC in pmol/L/PRC in µIU/mL) was calculated as a recommended screening method for the detection of primary aldosteronism. A recently published evaluation of cut-off values for positive primary aldosteronism screening based on the DiaSorin Liaison automated chemiluminiscent immunoassay for direct renin revealed a 100% sensitivity and 93% specificity for detecting primary aldosteronism by use of an ARR cut-point >35 pmol/L/µIU/mL in patients with circulating aldosterone levels >300 pmol/L.20 N-terminal pro B-type brain natriuretic peptide (pg/mL; conversion factor: pg/mL × 0.118 = pmol/L) was measured by an electrochemiluminiscent assay (ECLIA®, Roche Diagnonstics GmbH, Mannheim, Germany). High sensitive C-reactive protein (mg/L) was measured by the new C-reactive protein high-sensitive assay on COBAS INTEGRA® (Roche Diagnostics GmbH, Mannheim, Germany). LDL-cholesterol (mg/dL) was determined by a selective direct single measurement by the LDL-C Select FS method (DiaSys Diagnostic Systems GmbH, Holzheim, Germany) and HDL-cholesterol (mg/dL) was measured by an in vitro assay (Wako Chemical GmbH, Neuss, Germany). Thyroid stimulating hormone (TSH, µU/mL) was determined by an electrochemiluminiscent assay (ECLIA®, Cobas, Roche Diagnostics GmbH, Mannheim, Germany).

Statistical methods

Clinical and anthropometric characteristics of the study participants are reported according to levels of SAC (in tertiles) and were presented as percentages for categorical data and as medians with inter-quartile ranges or as means with SD for continuous data. Comparisons between SAC groups were performed using the χ2 test with P by linear-for-linear test for categorical data and by using analysis of variance for continuous data. All continuous parameters following a skewed distribution (PRC, ARR, high sensitive C-reactive protein, TSH, NT-proBNP, and LV mass) were logarithmically transformed before being used in parametric statistical procedures.

Pearson correlation analysis was used to analyse the correlations between SAC, ARR, and echocardiographic variables of LV structure and function overall and separately in women and men, respectively.

To better adhere to suggested sex-related relationship of SAC to LV mass, LV mass index, relative wall thickness, LV-PW, and interventricular septum levels over a broader range of data, we also categorized patients into gender-specific tertiles of SAC. Analysis of Covariance (ANCOVA) followed by Bonferroni's post hoc test was used to compare mean values of LV mass, LV mass index, relative wall thickness, LV-PW, and interventricular septum across gender-specific tertiles of SAC. Each model was adjusted for relevant key elements in regard to the variable of interest [age, gender, BMI, HDL-cholesterol, LDL-cholesterol, systolic BP, serum sodium, serum potassium, NT-proBNP, detailed antihypertensive medication (ACE-I use, ARB use, beta-blocker use, calcium-channel blocker use, diuretic use, and MR-blocker use), statin use, eGFRMDRD, diabetes mellitus, history of CHD, high sensitive C-reactive protein, current smoking status, TSH, alcohol abuse, heart rate, history of cancer, history of sleep apnea syndrome, chronic obstructive pulmonary disease, and PRC].

Logistic regression analysis (with backward elimination) was performed to investigate the association of SAC for patients presenting with pathological patterns of LV geometry (concentric LV remodelling, concentric LV hypertrophy, and eccentric LV hypertrophy). We calculated odds ratios (OR) with 95% confidence interval (95% CI) relating different pathological patterns of LV geometry to gender-specific tertiles of SAC (pmol/L) by comparing high SAC levels within the third and second gender-specific tertile, respectively, with SAC levels within the first (reference) gender-specific tertile. The analysis was performed overall and separately in women and men, respectively. According to the adjustment for important confounders, three different models were computed: model 1: crude; model 2: age-adjusted; model 3: multivariate adjusted [age, gender, BMI, HDL-cholesterol, LDL-cholesterol, systolic BP, serum sodium, serum potassium, NT-proBNP, detailed antihypertensive medication (ACE-I use, ARB use, beta-blocker use, calcium-channel blocker use, diuretic use, and MR-blocker use), statin use, eGFRMDRD, diabetes mellitus, history of CHD, high sensitive C-reactive protein, current smoking status, TSH, alcohol abuse, heart rate, history of cancer, history of sleep apnea syndrome, chronic obstructive pulmonary disease, and PRC]. Additionally, ORs (with 95% CI) for 1 SD increment in log-SAC (derived from multivariate adjusted model 3) for each abnormal LV geometric pattern were calculated.

Finally, interaction terms were introduced into the fully adjusted (model 3) logistic regression models to evaluate whether the nature of the association between SAC and concentric LV hypertrophy differs according to further parameters related to LV structure.

A P-value <0.05 was considered statistically significant and all statistical tests were two-sided. Data were analysed using SPSS 17.0 statistical package (SPSS, Inc., Chicago, IL, USA).

Results

At baseline, SAC (mean: 232.5 ± 208.9 pmol/L) was determined in 1575 study participants (mean age: 65.5 ± 8.4 years; 54.1% women). The distribution of demographics, co-morbidities, laboratory characteristics, and echocardiographic parameters according to levels of SAC (in tertiles) of patients at baseline is shown in Table 1. Higher SAC levels were associated with higher diastolic arterial BP and lower eGFR values. In addition, SAC levels were related to higher values of interventricular septum, LV posterior wall thickness, LV mass, LV mass index, and relative wall thickness values. Baseline SAC levels were not associated with varying severity of DD, age, BMI, hyperlipidaemia, arterial hypertension, high sensitive C-reactive protein levels and prevalence of abnormal geometric LV patterns.

View this table:
Table 1

Baseline characteristics according to serum aldosterone concentration tertiles

VariableTertile 1 (<164.8 pmol/L) n= 525Tertile 2 (164.8–224.8 pmol/L) n= 525Tertile 3 (>224.8 pmol/L) n= 525P Trend
SAC, pmol/L139.0 (116.3–152.3)193.1 (179.4–206.1)287.9 (249.0–370.9)
PRC, µIU/mL13.6 (6.4–33.8)16.5 (7.5–43.1)19.0 (8.6–54.7)0.001
ARR, pmol/L/µIU/mL9.4 (3.6–19.8)11.7 (4.5–25.4)15.9 (6.0–42.0)<0.001
Age, years65.9 ± 8.465.3 ± 8.464.8 ± 8.40.238
Female sex, %67.652.042.1<0.001
Body mass index, kg/m228.7 ± 5.028.8 ± 4.828.7 ± 4.60.880
LDL cholesterol, mg/dL124.8 ± 33.6127.1 ± 32.1125.9 ± 31.50.483
HDL cholesterol, mg/dL56.6 ± 16.653.2 ± 17.053.2 ± 15.70.016
Hyperlipidaemia, %39.839.836.80.311
Statin use, %28.123.725.10.271
Current smokers, %10.914.110.10.774
Alcohol abuse, %10.512.515.80.693
History of cancer, %9.17.07.40.303
Sleep apnea syndrome, %5.15.76.50.355
Chronic obstructive pulmonary disease, %7.28.25.30.224
Diabetes mellitusa, %22.724.821.90.770
Arterial hypertension, %77.177.379.20.413
Systolic blood pressure, mmHg147 ± 21147 ± 21146 ± 230.520
Diastolic blood pressure, mmHg82 ± 1284 ± 1284 ± 120.046
Heart rate, b.p.m.66 ± 1067 ± 1167 ± 120.170
History of CHD, %15.614.115.40.931
Antihypertensive treatment
 ACE-inhibitors, %36.737.939.80.301
 ARBs, %16.514.615.90.778
 Beta-blockers, %39.840.844.10.164
 Calcium-channel blockers, %17.717.118.80.648
 Diuretics, %36.244.446.70.001
 MR-blocker, %00.22.7<0.001
eGFR, mL/min per 1.73 m²75.6 (64.5–88.1)74.7 (65.1–86.5)71.4 (63.3–82.1)0.018
Serum sodium, mmol/L140.5 ± 2.4140.5 ± 2.4140.3 ± 2.70.397
Serum potassium, mmol/L4.3 ± 0.54.3 ± 0.54.4 ± 0.70.119
High sensitivity C-reactive protein, mg/L1.7 (0.8–4.1)1.7 (0.9–3.4)1.6 (0.8–3.5)0.685
Thyroid stimulating hormone, µU/mL0.97 (0.62–1.47)1.01 (0.66–1.60)1.07 (0.73–1.54)0.170
NT-proBNP, pg/mL102.0 (57.3–187.4)82.4 (44.4–154.7)83.9 (43.2–165.4)0.005
NYHA-classification, %0.871
 090.692.489.5
 I1.32.92.7
 II5.54.35.7
 III2.51.32.1
Echocardiographic parameters
 LV end-diastolic diameter, mm48.2 ± 5.549.0 ± 5.649.0 ± 5.70.032
 LV end-systolic diameter, mm29.9 ± 5.330.5 ± 5.430.4 ± 5.50.182
 LV end-diastolic volume, mL110.4 ± 28.3114.7 ± 29.9115.1 ± 29.50.033
 LV end-systolic volume, mL36.4 ± 15.938.2 ± 16.138.1 ± 16.90.182
 LV ejection fraction, %61.6 ± 6.261.8 ± 6.261.5 ± 6.10.675
 Interventricular septum, mm12.0 (10.3–13.0)12.0 (11.0–13.0)12.0 (11.0–13.0)<0.001
 LV posterior wall, mm10.9 ± 1.611.3 ± 1.711.4 ± 1.7<0.001
 LV mass, g124.5 (85.5–180.1)146.8 (98.1–205.4)149.1 (101.2–217.5)<0.001
 LV mass index, g/cm2109.4 ± 24.3116.9 ± 25.9116.7 ± 26.6<0.001
 Relative wall thickness0.46 ± 0.10.47 ± 0.10.47 ± 0.10.046
Patterns of LV-hypertrophy
 Normal LV, %68.562.762.9<0.001
 LV concentric remodelling, %14.415.915.70.567
 LV concentric hypertrophy, %4.24.45.50.377
 LV eccentric hypertrophy, %12.916.915.90.308
DIAST-CHF grade of DDb0.810
None, %20.218.123.2
Mild, %60.060.456.5
Moderate, %19.621.420.1
Severe, %0.20.20.2
  • Values are given as mean with SD or as median (25th, 75th percentile) for continuous variables, and percentage for categorical data. Group differences (P for trend) were calculated by ANOVA for continuous and χ2 test for categorical variables. Tertiles are from baseline serum aldosterone concentration.

  • SAC, serum aldosterone concentration; PRC, plasma renin concentration; ARR, aldosterone to renin ratio; CHD, coronary heart disease; ACE, angiotensin converting enzyme; MR blocker, mineralocorticoid receptor blocker; ARB, angiotensin-II type-1 receptor blocker; eGFR, estimated glomerular filtration according to the MDRD (modification of diet in renal disease) formula to estimate GFR; LV, left ventricular.

  • aDiabetes was classified according to the recommendations of the American Diabetes Association; bDIAST-CHF grade of DD (Grade of diastolic dysfunction).

Across the entire cohort, SAC correlated with echocardiographic parameters except for LVEF (Table 2). Overall no associations were found between ARR and echocardiographic parameters of the LV. The correlation analysis was further stratified by gender. In women, positive correlations were seen between SAC and LV mass, relative wall thickness and LV posterior wall thickness. No associations were observed between aldosterone and echocardiographic parameters of the LV in men. No relationship was found between aldosterone and internal LV dimension in either gender.

View this table:
Table 2

Correlation coefficients between serum aldosterone levels, aldostereon-to-renin ratio levels, N-terminal pro B-type brain natriuretic peptide, and echocardiographic parameters of left ventricular function and structure

VariablesOverall (n = 1575)Women (n = 849)Men (n = 726)
SACARRSACARRSACARR
SAC
ARR0.236**0.231**0.257**
NT-proBNP−0.085*0.132**−0.0670.092**−0.0540.168**
LV ejection fraction−0.035−0.009−0.004−0.027−0.0340.010
LV mass0.133**−0.0200.076*−0.0120.015−0.017
LV mass index0.109**0.0400.0630.0430.0340.076
Relative wall thickness0.060*−0.0270.075*−0.0400.044−0.010
Interventricular septum0.107**−0.0170.058−0.0710.0500.067
LV posterior wall0.125**−0.0300.096**−0.0330.038−0.018
LV end-diastolic diameter0.061*0.007−0.0040.030−0.028−0.009
LV end-systolic diameter0.052*0.0020.0200.017−0.006−0.006
LV end-diastolic volume0.060*0.008−0.0040.031−0.027−0.009
LV end-systolic volume0.052*0.003−0.0210.019−0.006−0.006
  • SAC, serum aldosterone concentration; ARR, aldosterone-to-renin ratio; LV, left ventricular.

  • *P< 0.05; **P< 0.001.

In an attempt to further assess the association between the SAC and echocardiographic parameters, the overall cohort was stratified according to gender-specific tertiles (T; each comprising 525 patients) of SAC. In fully adjusted ANCOVA, we noted a significant increase in LV mass (P= 0.001), LV mass index (P= 0.001), and relative wall thickness (P= 0.011) values across increasing values of SAC in the entire cohort (Figure 1). Moreover, mean values of LV posterior wall thickness and interventricular septum increased significantly from the first to the third gender-specific T of SAC (P< 0.001 for both).

Figure 1

(A–C) Distribution of left ventricular mass, left ventricular mass index and relative wall thickness values according to gender-specific tertiles of serum aldosterone concentration levels overall, in women and men, respectively (n= 1575). Mean values for echocardiographic parameters of each ANCOVA-analysis are illustrated. ANCOVA models were adjusted for: age, gender, body mass index, HDL-cholesterol, LDL-cholesterol, systolic blood pressure, serum sodium, serum potassium, N-terminal pro B-type brain natriuretic peptide, detailed antihypertensive medication (ACE-inhibitor use, angiotensin II type-1 receptor blocker use, beta-blocker use, calcium-channel blocker use, diuretic use, and mineralocorticoid receptor-blocker use), statin use, eGFRMDRD, diabetes mellitus, history of coronary heart disease, high sensitive C-reactive protein, current smoking status, thyroid stimulating hormone, alcohol abuse, heart rate, history of cancer, history of sleep apnea syndrome, chronic obstructive pulmonary disease, and plasma renin concentration. *P< 0.05.

Gender specific analyses revealed a significant association between higher SAC levels and LV mass, LV mass index, relative wall thickness and interventricular septum exclusively in women. Left ventricular posterior wall thickness increased significantly from the first (reference) gender-specific T of SAC to the third T in both, men (P= 0.046) and women (P= 0.008), respectively. Detailed multivariate adjusted analyses showed no significant associations between ARR and echocardiographic parameters (data not shown) overall and in either sex, respectively.

Left ventricular CR, concentric LV hypertrophy, and eccentric LV hypertrophy were present in 30.5, 26.1, and 16.9% of the study population. Table 3 shows the associations between higher SAC levels and abnormal geometric LV patterns across the whole cohort and stratified according to gender distribution derived from logistic regression models. Across the whole cohort and in gender-specific analysis, SAC levels in the highest (third) gender-specific T were strong correlates of concentric LV hypertrophy compared with the first (reference) T of SAC. After multivariate adjustment (model 3; adjusted according to ANVOCA models as stated above), patients in the highest gender-specific T of SAC were at a 1.9-fold increased risk of presenting with a concentric LV hypertrophy (adjusted OR: 1.87, 95% CI: 1.31–2.68, P= 0.001) compared with patients in the first gender-specific T of SAC. Gender-specific analysis revealed a significant relation of higher SAC levels to increased risk of concentric LV hypertrophy in both, women and men, respectively.

View this table:
Table 3

Logistic regression analysis relating plasma aldosterone levels (stratified in gender specific tertiles) to abnormal geometric left ventricular patterns

LV geometric patternOverallWomenMen
OR (95% CI) by comparing SAC gender-specific tertile 2 with tertile 1OR (95% CI) by comparing SAC gender-specific tertile 3 with tertile 1OR (95% CI) by comparing SAC gender-specific tertile 2 with tertile 1OR (95% CI) by comparing SAC gender-specific tertile 3 with tertile 1OR (95% CI) by comparing SAC gender-specific tertile 2 with tertile 1OR (95% CI) by comparing SAC gender-specific tertile 3 with tertile 1
Concentric remodelling
 Model 11.01 (0.77–1.34)0.95 (0.72–1.26)0.99 (0.68–1.45)1.05 (0.72–1.53)1.04 (0.69–1.56)0.84 (0.55–1.29)
 Model 21.02 (0.77–1.35)0.96 (0.72–1.27)0.99 (0.69–1–.45)1.04 (0.71–1.52)1.05 (0.70–1.58)0.84 (0.55–1.29)
 Model 30.91 (0.65–1.24)0.84 (0.60–1.18)0.96 (0.59–1.56)0.94 (0.58–1.52)0.92 (0.56–1.51)0.80 (0.49–1.31)
 SAC per log-SD0.94 (0.82–1.08)0.95 (0.78–1.15)0.90 (0.72–1.13)
Concentric hypertrophy
 Model 11.49 (1.10–2.02)1.74 (1.28–2.35)1.70 (1.12–2.58)1.77 (1.17.2.67)1.29 (0.82–2.01)1.70 (1.10–2.65)
 Model 21.52 (1.12–2.07)1.75 (1.29–2.36)1.69 (1.11–2.57)1.74 (1.15–2.63)1.35 (0.86–2.11)1.76 (1.13–2.75)
 Model 31.63 (1.14–2.35)1.87 (1.31–2.68)1.56 (0.94–2.57)1.62 (0.98–2.67)1.72 (1.01–2.92)2.06 (1.23–3.46)
 SAC per log-SD1.22 (1.06–1.40)1.22 (1.01–1.48)1.22 (0.98–1.52)
Eccentric hypertrophy
 Model 10.94 (0.67–1.32)0.80 (0.56–1.13)1.23 (0.76–2.01)0.91 (0.55–1.52)0.73 (0.45–1–16)0.72 (0.44–1.16)
 Model 20.94 (0.67–1.32)0.80 (0.56–1.13)1.23 (0.76–2.01)0.91 (0.55–1.52)0.73 (0.45–1.16)0.72 (0.45–1.16)
 Model 31.08 (0.73–1.61)0.86 (0.58–1.29)1.53 (0.87–2.69)0.93 (0.51–1.69)0.64 (0.36–1.12)0.69 (0.39–1.20)
 SAC per log-SD0.94 (0.80–1.12)0.89 (0.70–1.14)0.94 (0.72–1.24)
  • Data represent odds ratios (OR) with 95% confidence interval (95% CI) relating different pathological pattern of LV geometry to gender-specific tertiles of serum aldosterone concentration (SAC in pmol/L) by comparing high SAC levels within the second and third gender-specific tertile, respectively with SAC levels within the first (reference) gender-specific tertile. Model 1: crude. Model 2: age-adjusted; Model 3: multivariate adjusted (according to ANCOVA models (Figure 1)).

  • Additionally, odds ratios (with 95% CI) for 1 SD increment in log-SAC (derived from multivariate adjusted model 3) for each abnormal LV geometric pattern are shown.

When SAC was evaluated as a continuous variable the association between SAC levels and concentric LV hypertrophy remained highly significant (adjusted OR: 1.22, 95% CI: 1.06–1.40, per log SD increase in SAC levels, P= 0.005). Furthermore, these significant results did not materially change after exclusion of 76 (4.8%) participants, who were suggestive for primary aldosteronism (indicated by an ARR >35 pmol/L/µIU/mL and SAC >300 pmol/L). The highest gender-specific T remained related to a 1.7-fold increased risk of concentric LV hypertrophy compared with the first gender-specific T of SAC (P= 0.012).

We found that the association between SAC and risk of concentric LVH is modulated by gender distribution (P< 0.001), age (P< 0.001), BMI (P= 0.006), HDL-cholesterol (P= 0.004), systolic BP (P= 0.006), serum sodium (P= 0.006), NT-proBNP (P= 0.003), use of ACE-I (P= 0.026), use of beta-blocker (P= 0.013), eGFRMDRD (P= 0.001), and sleep apnea syndrome (P= 0.006).

Overall, no significant association was seen for SAC and risk of concentric LV remodelling and eccentric LV hypertrophy, respectively. Furthermore, no associations were found between ARR levels and pathological patterns of LV geometry overall and in either sex.

Discussion

In this large cohort of CV risk patients with preserved LVEF, we demonstrated that higher circulating aldosterone but not ARR levels were associated with increased levels of LV mass, LV mass index, relative wall thickness, interventricular septum, and LV posterior wall thickness. These associations were stronger in women and remained stable after consideration of several confounders such as plasma renin levels, arterial hypertension, current medication, diabetes mellitus, BMI, and kidney function. In addition, this is the first study to show in a large cohort of patients that higher circulating aldosterone levels are independently correlated with concentric LV hypertrophy in both men and women. This relationship did not change materially after exclusion of those participants who were suggestive for primary aldosteronism and were similar in patients with and without overt HF. Finally, gender distribution, age, BMI, HDL-cholesterol, systolic BP, serum sodium, NT-proBNP, use of ACE-I, use of beta-blocker, eGFRMDRD, and sleep apnea syndrome are crucial effect modifiers for SAC-related risk of concentric LV hypertrophy.

Our findings indicate that excessive adrenal aldosterone secretion in the setting of primary aldosteronism is not mandatory for aldosterone-mediated alterations of the LV structure. This is in line with previous findings of Tomaschitz et al.4,21 who documented that plasma aldosterone levels within the normal range predict higher risk of fatal cardiovascular events and sudden cardiac death. Thus, aldosterone-mediated alterations of the LV structure indicate an important role for mineralocorticoids in arrhythmogenesis.5 This concept is supported by significant LV hypertrophy regression and antiarrhythmogenic effects of MR-blockers in hypertensive and CHD patients.8,22

Most of the previous investigations regarding the associations between higher circulating aldosterone levels and LV structure in patients without confirmed primary aldosteronism revealed conflicting results. El-Gharbawy et al.23 noted no associations between circulating aldosterone levels and cardiac mass in hypertensive Caucasians. Other cross-sectional studies showed a positive correlation between circulating aldosterone levels and LV mass index.24,25 However, these investigations might have been limited by small sample sizes and lacking of gender-specific analyses. In a large community-based investigation, Vasan et al.26 documented a significant association between plasma aldosterone levels and LV posterior wall thickness and relative wall thickness exclusively in women. Interestingly, no association was seen for aldosterone and LV mass. In contrast to Vasan et al., we found robust associations between echocardiographic parameters of LV structure including LV mass and LV mass index and SAC across the whole cohort, but consistent associations were found exclusively in women. However, the participants of the DIAST-CHF study were older, showed higher systolic/diastolic BP values and suffered more frequently from diabetes compared with subjects from the Framingham study. Aldosterone-related alterations of the LV structure when comparing different studies might be hampered by the heterogeneity of study participants, i.e. differences in dietary salt intake, investigated. Accumulating evidence mainly derived from animal studies suggests that the combination of dietary salt excess and high aldosterone contributes to the development of cardiac fibrosis and hypertrophy independent of arterial BP and ventricular loading.27 Two recent observational studies in humans confirmed an increase in LV mass index in normo- and hypertensive patients with both increasing urinary sodium and aldosterone levels.28,29 Inappropriate downward adjustment of aldosterone synthesis during high dietary salt intake, which typically occurs in modern societies, may contribute to the development of a relative aldosterone excess induced cardiac remodelling.30 In addition, gender differences in LV adaption to pressure overload might be in part due to oestrogen signalling in the myocardium resulting in gender-specific differences of aldosterone-related variances of cardiac structure.31 Accordingly, a previous study showed that MR-blockade reduces ischaemia-induced cardiac remodelling and phenotypic alterations of gene expression to a greater extent in females than in male rats.32

We observed a higher risk of concentric LV hypertrophy in both men and women with increasing SAC levels independent of established LV-hypertrophic mediators. Previous studies however reported inconsistent findings in this field. Muscholl et al.33 found highest values of aldosterone in patients with LV remodelling and eccentric LV hypertrophy. Schunkert et al.34 observed significantly higher aldosterone levels in women with concentric or eccentric hypertrophy compared with women with normal LV patterns. Our investigation points to an association between aldosterone and concentric LV hypertrophy, which is supported by the observed reduction in the LV mass index after addition of MR blockers in patients with concentric LV hypertrophy.35

In contrast to previous investigations, we did not find any association between higher ARR levels and both, LV structure and pathological patterns of LV geometry, respectively.36 In the Framingham Offspring study, ARR was positively related to eccentric and concentric hypertrophy.37 This is in disagreement with our results and it might be speculated that differences in dietary sodium intake, prevalence of co-morbidities and aldosterone-renin determining factors such as ongoing medication might impact aldosterone and renin levels which results in contradicting findings.38 In particular, a slight elevation of the PRC (or activity), which is the main determinant of the ARR is itself considered as an independent correlate of cardiovascular outcome, due to intake of medication or the presence of co-morbidities such as HF, might result in lower ARR levels and presumably in less strong or absent associations with functional and structural pattern of the LV.39 Moreover, inconsistent cut-off values for pathological patterns of LV geometry and different laboratory methods used for aldosterone and renin determination might damper comparability of aldosterone-associated changes of the LV structure between studies. However, previous investigations showed that the ARR is one of the most important predictors of arterial BP, which is in itself a powerful risk factor of LV hypertrophy.40 The consistent association between SAC and echocardiographic parameters of LV structure even after consideration of PRC levels in the current analyses however indicates a renin-independent relationship between aldosterone and LV geometry. Whether aldosterone-induced deleterious effects are related to renin and/or further stimulators of aldosterone secretion such as ACTH, parathyroid hormone, or free fatty acids remains a matter or research. In addition, aldosterone-mediated effects might also be modified by an up-regulation and/or an altered activation of MR receptors, i.e. in the myocardium, in the presence of permissive factors such as dietary salt excess or oxidative stress.3

Overall and in gender-specific analyses specifically, we did not document associations between both aldosterone/ARR levels and systolic LV function reflected by an LVEF. This is in line with previous suggestions that asymptomatic structural alterations of the LV may exist long before the clinical syndrome of HF, cardiac arrhythmia, and cardiovascular events appear. In addition, findings from previous interventional studies that showed beneficial effects of MR-blockade in CHF patients might not necessarily be applicable to rather population-based observational cohorts, such as the DIAST-HF study. Whether aldosterone blockade exerts beneficial effects in diastolic HF (i.e. HFpEF) independent of arterial BP is currently being evaluated.41

Limitations

Findings regarding the association between circulating aldosterone levels and echocardiographic LV structure should be treated with caution. Since we did not measure structural parameters of the right ventricle, no assumptions about the association of aldosterone and the right heart structure and geometry can be made. Local extra-adrenal aldosterone synthesis, a complex cross-talk between aldosterone and angiotensin II and the aldosterone-cortisol interplay at MR-level might not be accurately reflected by a one-time aldosterone measurement in serum or plasma. In addition dietary salt intake has not been evaluated. Several other factors, which have been shown to interfere with aldosterone and renin levels, i.e. ongoing medication, might have confounded our analysis.38 The determination of 24-h urinary aldosterone excretion is superior to the measurement of serum/plasma aldosterone levels in estimating the integrated daily exposure to aldosterone.28 It is also important to stress that the cross-sectional design of our analysis does not enable conclusions with regard to causal relationships. However, the findings of our analyses in a large sample of CV risk patients with preserved LV EF might be extrapolated to a broader population within the related age class and is strengthened by gender-specific analyses and the consideration of a panel of established confounders.

Conclusions

The association of unremarkable circulating aldosterone levels with parameters of LV structure favours the notion of potential aldosterone-mediated effects on the cardiovascular system in the absence of primary aldosteronism. Mechanistic studies however are warranted to elucidate both the permissive milieu and mechanisms beyond aldosterone-driven effects on the myocardium. Our findings further support the idea that MR-blockade may delay or prevent the occurrence of abnormal patterns in patients with preserved LV EF, which warrants further observational and interventional studies in this field.

Funding

This work was supported by grants from the German Federal Ministry of Education and Research (German Heart Failure Network), TP 7 (FKZ 01GI0205) and the German Diabetes Foundation.

Conflict of interest: none declared.

Footnotes

  • These authors contributed equally to this work.

References

View Abstract